https://wiki.seas.harvard.edu/geos-chem/api.php?action=feedcontributions&user=Hyung-Min+Lee&feedformat=atomGeos-chem - User contributions [en]2024-03-29T11:45:07ZUser contributionsMediaWiki 1.24.2https://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint&diff=38403GEOS-Chem Adjoint2018-08-14T06:38:40Z<p>Hyung-Min Lee: /* Current GEOS-Chem Adjoint Research Projects (please add yours!) */</p>
<hr />
<div><big><big><strong>Adjoint Working Group</strong></big></big><br />
<br />
== Contact information ==<br />
<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|-valign="top"<br />
!width="300px" bgcolor="#CCCCCC"|Adjoint Model Scientist<br />
|width="600px"|[mailto:daven.henze@colorado.edu Daven Henze]<br />
<br />
|-valign="top"<br />
!bgcolor="#CCCCCC"|Adjoint Model and Data Assimilation<br>Working Group Co-Chairs<br />
|<br />
*[http://spot.colorado.edu/~henzed/ Daven Henze]<br />
*[https://www.engineering.uiowa.edu/faculty-staff/jun-wang Jun Wang]<br />
<br />
|-valign="top"<br />
!bgcolor="#CCCCCC"|GEOS-Chem Adjoint support team<br />
|<br />
*[mailto:yanko.davila@colorado.edu Yanko Davila], <br />
<br />
|-valign="top"<br />
!bgcolor="#CCCCCC"|Adjoint Working Group email list<br />
|<tt>geos-chem-adjoint [at] g.harvard.edu</tt><br />
<br />
|-valign="top"<br />
!bgcolor="#CCCCCC"|To subscribe to email list<br />
|Either<br />
*Send an email to <tt>geos-chem-adjoint+subscribe [at] g.harvard.edu</tt><br />
Or<br />
*Go to [https://groups.google.com/a/g.harvard.edu/forum/#!forum/geos-chem-adjoint Adjoint Google Group], then<br />
*Click on '''Subscribe to this group'''<br />
<br />
|-valign="top"<br />
!bgcolor="#CCCCCC"|To unsubscribe from email list<br />
|Either<br />
*Send an email to <tt>geos-chem-adjoint+unsubscribe [at] g.harvard.edu</tt><br />
Or<br />
*Go to the [https://groups.google.com/a/g.harvard.edu/forum/#!forum/geos-chem-adjoint Adjoint Google Group], then<br />
*Click on the '''My Settings''' button<br />
*Click on '''Leave this group'''<br />
|}<br />
<br />
--[[User:Bmy|Bob Y.]] ([[User talk:Bmy|talk]]) 15:54, 21 August 2015 (UTC)<br />
<br />
== Historical Development ==<br />
Original work on the adjoint of GEOS-Chem v6 began in 2003, focusing on the adjoint of the offline aerosol simulation. By 2005, the adjoint was expanded to include a tagged CO simulation and a full chemistry simulation; an adjoint of GEOS-Chem v7 was also developed in the following years. Each of these branches of the adjoint code were been constructed in a hybrid fashion using a combination of automatic differentiation software ([http://www.autodiff.com/tamc TAMC], [http://people.cs.vt.edu/~asandu/Software/Kpp/ KPP]) and manual coding of both discrete and continuous adjoints. They shared many common elements yet had unique features for different applications. <br />
<br />
During the summer of 2009, the existing branches were merged and updated to bring the adjoint into alignment with the latest release of GEOS-Chem, v8-02-01. This merged adjoint model is now the standard adjoint code into which all further development efforts will be placed.<br />
<br />
== Forward Model Code ==<br />
<br />
The forward model on which the adjoint is based originally corresponded to GEOS-Chem v8-02-01. It was subsequently updated as follows:<br />
<br />
* KPP solver for gas-phase chemistry (as in GCv8-02-03)<br />
* Implement Bond 2007 BC/OC emissions (as in GCv8-02-02)<br />
* Apply bug fixes from GCv8-02-02 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-02#Previous_issues_now_resolved_in_v8-02-02 here]<br />
* Apply bug fixes from GCv8-02-03 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-03#Previous_issues_now_resolved_in_v8-02-03 here]<br />
* Apply bug fixes from GCv8-02-04 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-04#Previous_issues_now_resolved_in_v8-02-04 here]<br />
<br />
All bug fixes and model updates were previous listed at the top of inverse_driver.f. Since then the code has been kept up to date with bug fixes and some forward model updates in v9 and v10. We have now switched to documenting the code development cycle here in the wiki, see the following section.<br />
<br />
== Code Versions, Bug Fixes and Developments == <br />
<br />
=== Current GEOS-Chem adjoint version released ===<br />
*[[GEOS-Chem_Adjoint_v35]] (You will download this version when you check out.)<br />
<br />
=== Previous GEOS-Chem adjoint versions released ===<br />
*[[GEOS-Chem_Adjoint_v34]]<br />
*[[GEOS-Chem_Adjoint_v33]] <br />
*[[GEOS-Chem_Adjoint_v32]]<br />
*[[GEOS-Chem_Adjoint_v31]]<br />
<br />
== Summary of Main Adjoint Code Supported Features ==<br />
<br />
=== Features === <br />
* Meteorological fields<br />
** GEOS-3 <i>needs testing</i><br />
** GEOS-4<br />
** GEOS-5<br />
** GEOS-FP<br />
** MERRA and MERRA2 <i>in progress</i><br />
* model resolution<br />
** 4 x 5<br />
** 2 x 2.5<br />
** 0.5 x 0.666<br />
** 0.25 x 0.3125<br />
** Nested China and North America<br />
** Small Domain China and North America<br />
* Forward model processes<br />
** convection<br />
** advection<br />
** PBL mixing<br />
** dry deposition<br />
** wet deposition<br />
** strat / trop exchange with LINOZ and new GMI strat chem (v9-01-03)<br />
** NOy up fluxes (now replaced with new GMI strat chem)<br />
** aerosols <br />
*** inorganic aerosol thermodynamics with RPMARES<br />
*** inorganic aerosol thermodynamics with ISORROPIA <i> in progress </i><br />
*** sulfate chemistry <br />
*** BC<br />
*** SOA, Dust, sea salt <i> needs doing </i><br />
*** aerosol surface area feedbacks <i> needs updating </i><br />
*** aerosol optical feedbacks <i> needs doing </i><br />
** emissions<br />
*** all standard emissions included<br />
* Simulation modes<br />
** full chemistry<br />
** tagged CO<br />
** tagged Ox <br />
** CH4<br />
** offline aerosols (for BC and dust only)<br />
** CO2<br />
* Observational Operators<br />
** MOPITT CO column<br />
** SCIAMACHY CO column<br />
** AIRS CO column<br />
** IMPROVE BC<br />
** CASTNet (NH4+) <i> needs updating </i> <br />
** GOME / SCIAMACHY NO2 column <i> needs updating </i> <br />
*** using KNMI retrieval (Henze)<br />
*** using Dalhousie retrieval (Shim)<br />
** SCIAMACHY/OMI NO2<br />
*** using Dalhousie retrieval (Bousserez, Padmanabhan)<br />
** TES NH3 <br />
** TES O3 <br />
** GOSAT CO2 <br />
** MLS O3 and TES CO2 <i> in progress </i><br />
* Control parameters<br />
** Initial Conditions scaling factors (linear or log)<br />
** Emissions scaling factors (linear or log) <br />
*** NH3, primary BC/OC, SO2: anthropogenic, natural, bioburn, biomass, ship<br />
*** NOx: soil, aircraft, anthropogenic, biofuel, bioburn<br />
*** Lightning NOx: injection height, yield <i> in progress </i><br />
*** all other gas-phase tracers: anthropogenic, biofuel, bioburn <br />
* Adjoint sensitivities<br />
** w.r.t. all implemented control parameters<br />
** w.r.t Reaction Rate Parameters <br />
** w.r.t all emissions <br />
** of AQ attainment metrics <i> needs updating </i><br />
** of spatiotemporally averaged species concentrations (e.g., arctic O3) <br />
* Other<br />
** Inverse Hessian approximation <br />
** off-diagonal covariance matrices <i> needs updating </i><br />
** 3D-Var <i> needs updating </i><br />
<br />
Features may be qualified as: <br />
* <i>needs testing</i>: an implemented feature that we haven't fully used yet<br />
* <i>needs updating</i>: a feature developed with a previous branch that has yet to be updated to GEOS-Chem v8 and the merged adjoint<br />
* <i>needs doing</i>: a feature nobody has tackled the adjoint of yet<br />
* <i>in progress</i>: a feature currently under development <br />
* <i>in pipeline</i>: a feature which has been submitted and awaiting integration into the CVS repository<br />
<br />
=== Primary code developers ===<br />
Monika Kopacz, Kumaresh Singh, Changsub Shim, Daven Henze<br />
<br />
=== Adjoint model lead scientist ===<br />
Daven Henze<br />
<br />
<br />
<br />
== Resources ==<br />
<br />
=== User's guide ===<br />
A User's Guide v35 is available. [http://adjoint.colorado.edu/~yanko/gcadj_std/gc_adj_man.pdf User's Guide v35] <br />
<br />
Previous version v32-v33 available at [http://adjoint.colorado.edu/~daven/gcadj_std/GC_adj_man.pdf User's Guide v32]<br />
<br />
Quick Introduction to GitLab is available at [http://adjoint.colorado.edu/~yanko/gcadj_std/GitLab_Tutorial.pdf GitLab Tutorial]<br />
<br />
The forward version of the code is based on v9-02 here is the [http://acmg.seas.harvard.edu/geos/doc/archive/man.v9-02/ User's Guide]<br />
<br />
=== Code flowchart ===<br />
Meemong Lee has created a detailed flowchart of the inverse model code structure. http://adjoint.colorado.edu/~daven/gcadj_std/flowchart.pdf<br />
<br />
=== Plotting tools ===<br />
<br />
Some IDL and MATLAB routines for plotting benchmark results. http://adjoint.colorado.edu/~daven/gcadj_std/tools.tar.gz<br />
<br />
=== Background papers and presentations ===<br />
<br />
Several articles and presentations (including a GC adjoint modeling clinic overview from IGC5) providing background information about adjoints. http://adjoint.colorado.edu/~daven/gcadj_std/adj_articles.tar.gz<br />
<br />
== Distribution and Use ==<br />
Code for the adjoint is distributed through GITLAB, a web interface connected to a GIT server located at adjoint.colorado.edu. You can access GITLAB at http://adjoint.colorado.edu:8080 after your account is created. Here is our [[Quick Start Guide]]. <br />
<br />
Even if your office mate has a copy of the code, the best way to obtain the model is to get an account for yourself and download a version from the repository. So please do not copy code directly from others or pass the code along to third parties. This vastly helps with tracking developments and keeping up with model updates. <br />
<br />
Use of the adjoint model code follows standard practice for GEOS-Chem. It is expected that any developments that come of individual applications based on this community model will eventually be given back to the community by incorporation of new developments into the standard adjoint code. New development should be submitted to Daven Henze for inclusion in the standard adjoint model code.<br />
<br />
Using GIT gives the users the ability to change the code and commit their changes without affecting the main repository hosted at adjoint.colorado.edu. Users can work with their modified versions of the code and even create their own tags because GIT acts as a local repository. When ready to submit your update to the community just create a new branch with your modifications. Send an email Daven Henze explaining your contributions and we'll do our best to include them as soon as possible.<br />
<br />
=== Quick guide to GIT === <br />
As of version 34 we started using git versioning system and the GitLab web interface. We prepared [http://adjoint.colorado.edu/~yanko/gcadj_std/GitLab_Tutorial.pdf GitLab Tutorial] to help users get used to the web interface and git. We recommend first taking a look at GIT manual to get a general feel for how this tools works (e.g., [http://git-scm.com/documentation GIT Documentation] or [http://www.kernel.org/pub/software/scm/git/docs/git.html GIT Manual Page]). <br />
<br />
<br />
Useful GIT commands: <br />
<br />
Initial download:<br />
git clone ssh://git@adjoint.colorado.edu/yanko.davila/gcadj_std.git<br />
<br />
Status of project vs the current repository:<br />
git status<br />
<br />
Check difference of files (differences have colors for easy reading)<br />
git diff --color <wildcard> [<wildcard>] <path>/foo_mod.f<br />
<br />
Checkout specific version<br />
git checkout <wildcard><br />
<br />
Replacing a file with the newest version from the repository:<br />
git checkout origin/master -- <path>/foo_mod.f<br />
<br />
Merging changes in a file: ([http://www.kernel.org/pub/software/scm/git/docs/git-merge.html Reference])<br />
git merge -m <wildcard> <br />
<br />
Comitting<br />
git commit -a<br />
<br />
Tagging a version<br />
git tag -a TAGNAME<br />
<br />
Deleting a tag<br />
git tag -d TAGNAME<br />
git push origin :refs/tags/TAGNAME<br />
<br />
Saving changes to repository<br />
git push<br />
<br />
Saving tags to repository<br />
git push --tags<br />
<br />
List the history of a file:<br />
git log -- <path>/foo_mod.f<br />
<br />
Add a file to the repository<br />
git add <file_name><br />
<br />
Delete a file from the repository<br />
git rm <file_name><br />
<br />
Determine current version<br />
git show HEAD [ | grep commit]<br />
<br />
Download remote changes, rewind your local branch, then replays all your changes over the top of your current branch one by one, until you’re all up to date.<br />
git pull --rebase<br />
<br />
There are several wildcards that you can use on git for example:<br/><br />
''"origin/master"'' - Latest version on the repository <br/><br />
''"HEAD"'' - Latest version as of your last download <br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/72773d7b1a7bd757957d2c186acb3dbb107d5323 v33i]"'' - Specific TAG, find all tag names on GitLab<br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/32d5c926ef529c1dfc640c9dec4a925b24db575c 32d5c926e]"'' - Specific COMMIT, find all commit numbers on GitLab <br/><br />
<br />
[[Using_Git_with_GEOS-Chem|Here is the foward model documentation of git.]]<br />
<br />
=== Backward compatibility (CVS) === <br />
<br />
For people using old version of the code we still have active our CVS repository, but note that the latest version on CVS is v33i-patch2. Here you can find our [[Quick Guide to CVS]]<br />
<br />
== Crediting GEOS-Chem adjoint developers ==<br />
<br />
We aim to make distribution of adjoint model code as immediate as possible. A consequence is that many features may not yet be publicly documented. Therefore, giving code developers due credit is of utmost importance. <br />
<br />
Authors of new additions to the standard code should be offered co-authorship on the first round of presentations and publications to come of their development. Features currently falling in this category and their developers are:<br />
<br />
* (v35j) MOPITT CO Observation operator. Zhe Jiang; Yanko Davila, CU Boulder. <br />
* (v35j) OMI SO2, OMI NO2 and TES O3 Observation operators. Martin Keller, U Toronto; Yi Wang, UNL; Yanko Davila, CU Boulder. <br />
* (v35) HTAP Emissions Inventory. Kateryna Lapina, Daven Henze and Yanko Davila, CU Boulder.<br />
* (v35) NEI2008 Emissions Inventory. Katie Travis; Fabien Paulot, Harvard; Hyungmin Lee, and Daven Henze, CU Boulder. <br />
* (v35) Deposition based cost function. Fabien Paulot, Harvard; Daven Henze and Yanko Davila, CU Boulder. <br />
* (v34) Implementation of the sensitivity to reaction rate constants. Developers: Hyungmin Lee, CU Boulder; Thomas ; Fabien Paulot (Harvard); Daven Henze, CU Boulder and Yanko Davila, CU Boulder. <br />
* (v34) ISOROPIA II adjoint. Developer: Shannon Capps, EPA.<br />
* (v34) Off-diagonal covariance error matrices implementation. Developers: Nicolas Bousserez, CU Boulder; Kumaresh; Yanko Davila, CU Boulder. <br />
* (v32) Nested full chemistry adjoint. Developers: Zhe Jiang, University of Toronto; Daven Henze, CU Boulder.<br />
<br />
<br />
Citation of the appropriate [[GEOS-Chem_Adjoint#Journal_Articles|journal articles]] for mature developments is also encouraged, as well as considering aspects of [http://acmg.seas.harvard.edu/geos/geos_credit.html co-authorship for the forward model]. <br />
<br />
Overall, if you have any questions about authorship, even for a conference presentation, please contact Daven Henze.<br />
<br />
== Current GEOS-Chem Adjoint Research Projects (please add yours!) ==<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|- bgcolor="#cccccc"<br />
!User Group <br />
!Description <br />
!Contact Person<br />
|-<br />
<br />
|-<br />
|Anyang University<br />
|Aerosol emission modeling in East Asia <br />
|[mailto:koo@anyang.ac.kr Youn Seo Koo]<br />
|-<br />
<br />
|-<br />
|CRAES<br />
|source attribution of PM in eastern China<br />
|[mailto:weipeng@craes.org.cn Wei Peng]<br />
|-<br />
<br />
<br />
|-<br />
|CU Boulder<br />
|Aerosol precursors, CO2, O3; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|CU Boulder<br />
|Inverse modeling/optimization: variational source inversions, posterior error estimates in high-dimension; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Nicolas Bousserez]<br />
|-<br />
|CU Boulder<br />
|Constraints on Aerosol Sources<br />
|[mailto:zhen.qu@colorado.edu Zhen Qu]<br />
|-<br />
<br />
|-<br />
|Dalhousie University<br />
|Mass balance vs adjoint approaches for top-down constraints on NOx emissions<br />
|Matthew Cooper<br />
|-<br />
|Dalhousie University<br />
|Trends in OC and BC emissions over the US <br />
|Nathaniel Egan-Pimblett<br />
|-<br />
|Dalhousie University<br />
|Sensitivity of global PM2.5-induced mortality to emissions<br />
|[mailto:colin.lee@dal.ca Colin Lee]<br />
|-<br />
|Dalhousie University<br />
|Top-down constraints on NH3 emissions<br />
|[mailto:chi.li@dal.ca Chi Li]<br />
|-<br />
|Dalhousie University<br />
|Global contribution to the black carbon in the Arctic<br />
|[mailto:Junwei.Xu@Dal.Ca Junwei Xu]<br />
|-<br />
<br />
<br />
|-<br />
|Drexel<br />
|ISORROPIA adjoint development; NH3 assimilation; cloud droplet sensitivities<br />
|[mailto:sc3623@drexel.edu Shannon Capps]<br />
|-<br />
<br />
|-<br />
|Harvard<br />
|Methane<br />
|[mailto:maasakkers@seas.harvard.edu Bram Maasakkers], [mailto:mpayer@seas.harvard.edu Melissa Sulprizio]<br />
|-<br />
|Harvard<br />
|Harvard Emissions Component (HEMCO)<br />
|[mailto:ckeller@seas.harvard.edu Christoph Keller]<br />
|-<br />
|Harvard<br />
|VOC emissions in N.A. constrained by satellite HCHO<br />
|[mailto:jkaiser@seas.harvard.edu Jennifer Kaiser]<br />
|-<br />
|Harvard<br />
|Tradeoffs between air quality and economic outcomes<br />
|[mailto:seastham@fas.harvard.edu Sebastian D Eastham]<br />
|-<br />
<br />
|-<br />
| JAMSTEC<br />
|Multiple-constituent satellite data assimilation<br />
|[mailto:kmiyazaki@jamstec.go.jp Kazuyuki Miyazaki]<br />
|-<br />
<br />
|-<br />
|JPL<br />
|3D-var Microwave Limb Sounder (MLS) and Tropospheric Emission Spectrometer (TES) ozone assimilation <br />
|[mailto:meemong.lee@jpl.nasa.gov Meemong Lee]<br />
|-<br />
|JPL <br />
| TES ozone assimilation/attribution of GHG radiative forcing<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|JPL <br />
| Carbon Monitoring System Flux (CMS-Flux) 4D-var and LETKF assimilation<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|JPL <br />
| Adjoint sensitivity analysis for satellite constellation design<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|-<br />
|MIT<br />
|Air quality, aircraft emissions and sensitivities<br />
|[mailto:kingshuk@mit.edu Kingshuk Dasadhikari]<br />
|-<br />
|MIT<br />
|Air quality, aircraft emissions and sensitivities<br />
|[mailto:dedoussi@mit.edu Irene Dedoussi]<br />
|-<br />
<br />
|-<br />
|Nanjing University<br />
|Inverse modeling of terrestrial ecosystem carbon flux<br />
|[mailto:hengmao.wang@gmail.com Hengmao Wang]<br />
|-<br />
|-<br />
|Nazarbayev University<br />
|Emissions influences on NO_2 columns<br />
|[mailto:mehdi.torkmahalleh@nu.edu.kz Mehdi Torkmahalleh]<br />
<br />
|-<br />
|NCAR<br />
|Adjoint analysis for carbon monoxide and ozone<br />
|[mailto:zhejiang@ucar.edu Zhe Jiang]<br />
|-<br />
<br />
|-<br />
|Peking University<br />
|Satellite constraints on VOC emissions <br />
|[mailto:tmfu@pku.edu.cn May Fu]<br />
|-<br />
|Peking University<br />
|Source attributions of tropospheric ozone over North China<br />
|[mailto:linjt@pku.edu.cn Jintai Lin]<br />
|-<br />
|Peking University<br />
|Satellite Health impact analysis of China interregional pollution trade<br />
|[mailto:1300011409@pku.edu.cn Liu Yang]<br />
|-<br />
|Peking University<br />
|Emissions and chemistry of VOCs in China.<br />
|[mailto:qische@gmail.com Qi Chen]<br />
|-<br />
<br />
|-<br />
|Purdue University<br />
|Feedback between terrestrial ecosystem processes and atmospheric CO2 signals<br />
|[mailto:zhuq@purdue.edu Qing Zhu]<br />
|-<br />
|Purdue University<br />
|Feedback between aquatic ecosystem processes and atmospheric CH4 signals<br />
|[mailto:tan80@purdue.edu Zeli Tan]<br />
|-<br />
<br />
|-<br />
|Seoul National University <br />
|VOC emission constraints over Korea using OMI satellite retrieval and DC-8 flight measurements from KORUS-AQ<br />
|[mailto:hyungmin.lee@snu.ac.kr Hyung-Min Lee]<br />
|-<br />
|Seoul National University <br />
|Emissions and chemistry of aerosols in East Asia<br />
|[mailto:ss99@snu.ac.kr Jaein Jeong]<br />
|-<br />
<br />
|-<br />
|Tsinghua University<br />
|Inverse modeling of anthropogenic emissions over East Asia<br />
|[mailto:qiangzhang@tsinghua.edu.cn Qiang Zhang]<br />
|-<br />
|Tsinghua University<br />
|Nested-gird simulations with the adjoint model<br />
|[mailto:nany12@mails.tsinghua.edu.cn Nan Yang]<br />
|-<br />
<br />
|-<br />
|UCLA<br />
|Constrain black carbon emission<br />
|[mailto:qiling@atmos.ucla.edu Ling Qi]<br />
|-<br />
|UCLA<br />
|source attribution of ozone in the western U.S.<br />
|[mailto:gaomei@atmos.ucla.edu Mei Gao]<br />
|-<br />
<br />
|-<br />
|University of Edinburgh<br />
|Ozone sensitivity in biomass burning plumes<br />
|[mailto:d.finch@ed.ac.uk Douglas Finch]<br />
|-<br />
<br />
|-<br />
| University of Florida<br />
|Nitrogen deposition related to climate strategy projections with GCAM<br />
|[mailto:cbaublitz@ufl.edu Colleen Baublitz], [mailto:barronh@ufl.edu Barron H. Henderson]<br />
|-<br />
<br />
|-<br />
|University of Minnesota<br />
|Inverse modeling of VOC sources based on TES and IASI measurements<br />
||[mailto:kcw@umn.edu Kelley Wells], [mailto:dbm@umn.edu Dylan Millet]<br />
|-<br />
|University of Minnesota & CU Boulder<br />
|Inverse modeling of N2O sources<br />
|[mailto:dbm@umn.edu Dylan Millet], [mailto:Daven.Henze@Colorado.EDU Daven Henze]<br />
|-<br />
<br />
|-<br />
|University of Toronto<br />
|Methane inverse modeling<br />
|[mailto:stanevich@atmosp.physics.utoronto.ca Ilya Stanevich]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and CO to precursor emissions<br />
|[mailto:cwhaley@atmosp.physics.utoronto.ca Cynthia Whaley]<br />
|-<br />
|University of Toronto<br />
|Data assimilation of atmospheric CO and CO2<br />
|[mailto:tailong.he@mail.utoronto.ca Tailong He]<br />
|-<br />
<br />
|-<br />
|University of Wisconsin<br />
|CO2 assimilation and forecast & Temperature profile retrieval <br />
|[mailto:wenguang.bai@ssec.wisc.edu Wenguang Bai]<br />
|-<br />
<br />
|-<br />
|University of Wollongong<br />
|Sensitivity of ozone and adjoint analysis of CO over Australasia.<br />
|[mailto:rb864@uowmail.edu.au Rebecca Buchholz]<br />
|-<br />
<br />
<br />
|-<br />
|USTC<br />
|assimilation & sensitivies of CO2<br />
|[mailto:tuang372@gmail.com Li Tuang]<br />
|-<br />
<br />
== Publications ==<br />
<br />
=== Journal Articles ===<br />
<br />
<br />
* '''In press or sumbitted'''<br />
**Cao, H., T.-M. Fu, L. Zhang, D. K. Henze, C. Chan Miller, C. Lerot, G. González Abad, I. De Smedt, Q. Zhang, M. van Roozendael, K. Chance, J. Li, J. Y. Zheng, Y. H. Zhao, Adjoint inversion of Chinese non-methane volatile organic compound emissions using space-based observations of formaldehyde and glyoxal, Atmos. Chem. Phys. Discuss., https://doi.org/10.5194/acp-2017-1136, in review.<br />
**Chen, C., O. Dubovik, D. K. Henze, T. Lapyonak, M. Chin, F. Ducos, P. Litvinov, X. Huang, L. Li, Retrieval of desert dust and carbonaceous aerosol emissions over Africa from PARASOL/GRASP observations, submitted.<br />
**Zhao, H., Q. Zhang, S. J. Davis, X. Li, Y. Liu, G. Geng, M. Li, B. Zheng, H. Huo, L. Zhang, D. K. Henze, K. He, Inequality of household consumption and air pollution deaths in China, submitted.<br />
**Nakarmi, A., B. Sharma, U. S. Rajbhandari, A. Prajapati, C. S. Malley, J. C. I. Kuylenstierna, H. W. Vallack, D. K. Henze, A. Panday, Mitigating the impacts of air pollutants in Nepal and climate co-benefits: A scenario-based approach, submitted.<br />
<br />
* '''2018'''<br />
**Wells, K. C., D. B. Millet, N. Bousserez, D. K. Henze, T. J. Griffis, S. Chaliyakunnel, E. J. Dlugokencky, E. Saikawa, G. Xiang, R. G. Prinn, S. O'Doherty, D. Young, R. F. Weiss, G. S. Dutton, J. W. Elkins, P. B. Krummel, R. Langenfelds, L. P. Steele (2018), Top-down constraints on global N2O emissions at optimal resolution: application of a new dimension reduction technique, Atmos. Chem. Phys., 18, 735-756, https://doi.org/10.5194/acp-18-735-2018.<br />
**Zhang, L., Y. Chen, Y. Zhao, Y. D. K. Henze, L. Zhu, Y. Song, F. Paulot, X. Liu, Y. Pan, B. Huang (2018), Agricultural ammonia emissions in China: reconciling bottom-up and top-down estimates, Atmos. Chem. Phys., 18, 339-355, https://doi.org/10.5194/acp-18-339-2018.<br />
<br />
* '''2017'''<br />
**Lacey F. G., E. A. Marais, D. K. Henze, C. J. Lee, A. van Donkelaar, R. V. Martin, M. P. Hannigan, C. Wiedinmyer, Improving present day and future estimates of anthropogenic sectoral emissions and the resulting air quality impacts in Africa, Faraday Discuss., 200, 397-412.<br />
**Cui, Y., J. Brioude, W. M. Angevine, J. Peischl, S. A. McKeen, S-W. Kim, J. Neuman, D. K. Henze, N. Bousserez, M. Fischer, S. Jeong, H. Michelsen, R. P. Bambha, Z. Liu, G. W. Santoni, B. Duabe, E. Kort, G. Frost, T. B. Ryerson, S. C. Wofsy, M. Trainer (2017). Top-down estimate of methane emissions in California using a mesoscale inverse modeling technique: The San Joaquin Valley, J. Geophys. Res. Atmos., 122, 3686--3699, doi:10.1002/2016JD026398, online.<br />
**Cooper, M., R. V. Martin, A. Padmanabhan, and D. K. Henze (2017), Comparing mass balance and adjoint methods for inverse modeling of nitrogen dioxide columns for global nitrogen oxide emissions, J. Geophys. Res. Atmos., 122, 4718-4734, doi:10.1002/2016JD025985 online.<br />
**Bousserez, N. and D. K. Henze (2017), Optimal and scalable methods to approximate the solutions of large-scale Bayesian problems: Theory and application to atmospheric inversions and data assimilation, Q. J. R. Met. Soc., doi:10.1002/qj.3209, online.<br />
<br />
* '''2016'''<br />
**Lee, H.-M., F. Paulot, D. K. Henze, K. Travis, D. J. Jacob, L. H. Pardo, B. Schichtel (2016), Sources of nitrogen deposition in Federal Class I areas in the US, Atmos. Chem. Phys, 16, 525-540, doi:10.5194/acp-16-1-2016<br />
**Zhang, L., J. Shao, X. Lu, Y. Zhao, Y. Hu, D. K. Henze, H. Liao, S. Gong, Q. Zhang (2016), Sources and processes affecting fine particulate matter pollution over North China: an adjoint analysis of the Beijing APEC period, Environ. Sci. Technol., 50 (16), 8731-8740, doi:10.1021/acs.est.6b03010<br />
**Bousserez, N., D. K. Henze, B. Rooney, A. Perkins, K. J. Wecht, A. J. Turner, V. Natraj, J. R. Worden, Constraints on methane emissions in North America from future geostationary remote sensing measurements, Atmos. Chem. Phys., 16, 6175-6190, 2016, https://doi.org/10.5194/acp-16-6175-2016<br />
**Tan, Z., Q. Zhuang, D. K. Henze, C. Frankenberg, E. Dlugokencky, C. Sweeney, A. J. Turner, M. Sasakawa, T. Machida, Inverse modeling of pan-Arctic methane emissions at high spatial resolution: what can we learn from assimilating satellite retrievals and using different process-based wetland and lake biogeochemical models?, Atmos. Chem. Phys., 16, 12649–12666, doi:10.5194/acp-16-12649-2016.<br />
<br />
* '''2015'''<br />
**Zhang, L., L. Licheng, Y. Zhao, S. Gong, X. Zhang, D. K. Henze, S. L. Capps, T.-M. Fu, Q. Zhang, Source attribution of particulate matter pollution over North China with the adjoint method, Environ. Res. Lett. 10 084011.<br />
**Zhao, Y. H., L. Zhang, Y. P. Pan, Y. S. Wang, F. Paulot, and D. K. Henze, Atmospheric nitrogen deposition to the northwestern Pacific: seasonal variation and source attribution, submitted.<br />
**Lapina, K., D. K. Henze, J. B. Milford, C. Cuvelier, and M. Seltzer, Implications of RCP Scenarios for future changes in vegetative exposure to ozone in the Western U.S., Geophys. Res. Let., Atmos. Chem. Phys., 15, 10905-10924, 2015, https://doi.org/10.5194/acp-15-10905-2015<br />
**Whaley, C. H., K. Strong, D. B. A. Jones, T. W. Walker, Z. Jiang, D. K. Henze, M. Cooke, C. A. McLinden, M. Pommier, R. L. Mittermeier, P. F. Fogal, Improvements to our understanding of urban ozone air pollution: Sources of Toronto-area ozone during poor air quality events, J. Geophys. Res. Atmos., 120, 11,368–11,390, doi:10.1002/2014JD022984.<br />
**Liu, J., Bowman, K., and Henze, D., Source-receptor relationships of column-average CO2 and implications for the impact of observations on flux inversions., J. Geophys. Res. Atmos., 120,5214–5236, doi:10.1002/2014JD022914<br />
**Turner, A. J. and D. J. Jacob, Balancing aggregation and smoothing errors in inverse models, Atmos. Chem. Phys., 15, 7039-7048, 2015, https://doi.org/10.5194/acp-15-7039-2015<br />
**Turner, A. J., D. J. Jacob, K. J. Wecht, J. D. Maasakkers, S. C. Biraud, H. Boesch, K. W. Bowman, N. M. Deutscher, M. K. Dubey, D. W. T. Griffith, F. Hase, A. Kuze, J. Notholt, H. Ohyama, R. Parker, V. H. Payne, R. Sussmann, V. A. Velazco, T. Warneke, P. O. Wennberg, and D. Wunch, Estimating global and North American methane emissions with high spatial resolution using GOSAT satellite data, Atmos. Chem. Phys., 15, 7049-7069, 2015,https://doi.org/10.5194/acp-15-7049-2015<br />
**Zhang, L., D. K. Henze, G. A. Grell, G. R. Carmichael, N. Bousserez, Q. Zhang, J. Cao, O. Torres, C. Ahn, Z. Lu, Y. Mao, Constraining black carbon aerosol over Southeast Asia using OMI aerosol absorption optical depth and the adjoint of GEOS-Chem, Atmos. Chem. Phys., 15, 10281-10308, doi:10.5194/acp-15-10281-2015.<br />
**Patrick S. Kim, Daniel J. Jacob, Loretta J. Mickley, Shannon N. Koplitz, Miriam E. Marlier, Ruth S. DeFries, Samuel S. Myers, Boon Ning Chewf, Yuhao H. Mao, Sensitivity of population smoke exposure to fire locations in Equatorial Asia, doi:10.1016/j.atmosenv.2014.09.045<br />
**Deng, F., D. B. A. Jones, T. W. Walker, M. Keller, K. W. Bowman, D. K. Henze, R. Nassar, E. A. Kort, S. C. Wofsy, K. A. Walker, A. E. Bourassa, and D. A. Degenstein, Sensitivity analysis of the potential impact of discrepancies in stratosphere-troposphere exchange on inferred sources and sinks of CO2, Atmos. Chem. Phys. Discuss., 15, 10813-10851, doi:10.5194/acpd-15-10813-2015.<br />
**Jiang, Z., D. B. A. Jones, J. R. Worden, H. M. Worden, D. K. Henze, Y. X. Wang, Regional data assimilation of multi-spectral MOPITT observations of CO over North America, Atmos. Chem. Phys. Discuss., 15, 5327-5358, doi:10.5194/acpd-15-5327-2015.<br />
**Zhu, L., D. K. Henze, J. Bash, G. Jeong, K. Cady-Pereira, M. Shephard, M. Luo, F. Paulot, and S. Capps, Global evaluation of ammonia bi-directional exchange, Atmos. Chem. Phys. Discuss., 15, 4823-4877, doi:10.5194/acpd-15-4823-2015.<br />
**Jiang, Z., J. R. Worden, D. B. A. Jones, J.-T. Lin, W. Verstraeten, and D. K. Henze, Constraints on Asian ozone using Aura TES, OMI and Terra MOPITT, Atmos. Chem. Phys., 15, 99-112.<br />
**Lee, C.J., R.V. Martin, Henze D.K., Brauer M., Cohen A., and A. van Donkelaar, Sensitivity of global particulate-matter-related mortality to local precursor emissions, Environ. Sci. Technol., 49(7), 4335–4344, doi:10.1021/acs.est.5b00873, 2015.<br />
**Bousserez, N., D. K. Henze, A. Perkins, K. W. Bowman, M.Lee, J.Liu, D.B.A. Jones, F. Deng, Improved analysis-error covariance matrix for high-dimensional variational inversions: application to source estimation using a 3D atmospheric transport model. Q.J.R. Meteorol. Soc.. doi: 10.1002/qj.2495<br />
<br />
* '''2014'''<br />
**Jiang, Z., D. B. A Jones, H. M. Worden, and D. K. Henze, Sensitivity of inferred regional CO source estimates to the vertical structure in CO as observed by MOPITT, Atmos. Chem. Phys. Discuss., 14, 22939-22984<br />
**Zhu, Q., Q. Zhuang, D. K. Henze, K. Bowman, M. Chen, Y. Liu, Y. He, H. Matsueda, T. Machida, and Y. Sawa, Constraining terrestrial ecosystem CO2 fluxes by integrating models of biogeochemistry and atmospheric transport and data of surface carbon fluxes and atmospheric CO2 concentrations, Atmos. Chem. Phys. Discuss., 14, 22587-22638<br />
**Mao, Y. H., Q. B. Li, D. K. Henze, Z. Jiang, D. B. A. Jones, M. Kopacz, C. He, L. Qi, M. Gao, W.-M. Hao, and K.-N. Liou, Variational estimates of black carbon emissions in the western United States, Atmos. Chem. Phys. Discuss., 14, 21865-21916<br />
**Liu, J., Bowman, K., Lee, M., Henze, D., Bousserez, N., Brix, H., Collatz, G., Menemenlis, D., Ott, L., Pawson, S., Jones, D., Nassar, R.. Carbon monitoring system flux estimation and attribution: impact of ACOS-GOSAT XCO2 sampling on the inference of terrestrial biospheric sources and sinks. Tellus B, North America, 66, may. 2014. Available at: http://www.tellusb.net/index.php/tellusb/article/view/22486<br />
** Lee, H., D. K. Henze, B. Alexander, and L. T., Murray, Investigating the sensitivity of surface-level nitrate seasonality in Antarctica to primary sources using a global model, Atmos. Environ., 89, 757--767, doi:0.1016/j.atmosenv.2014.03.003<br />
**Paulot, F., D. J. Jacob, R. W. Pinder, J. O. Bash, K. Travis, D. K. Henze, Ammonia emissions in the United States, Europe, and China derived by high-resolution inversion of ammonium wet deposition data: Interpretation with a new agricultural emissions inventory (MASAGE_NH3), J. Geophys. Res., 119, 7, 4343--4364, doi:10.1002/2013JD021130.<br />
**Lapina, K., D. K. Henze, J. B. Milford, M. Huang, M. Lin, A. M. Fiore, G. Carmichael, G. G. Pfister, and K. W. Bowman, Assessment of source contributions to seasonal vegetative exposure to ozone in the U.S., J. Geophys. Res., 119, 324-340, doi:10.1002/2013JD020905<br />
**Shen, Z., J. Liu, L. W. Horowitz, D. K. Henze, S. Fan, H. Levy II, D. L. Mauzerall, J. Lin, and S. Tao,.: Analysis of transpacific transport of black carbon during HIPPO-3: implications for black carbon aging, Atmos. Chem. Phys. Discuss., 14, 505-540, doi:10.5194/acpd-14-505-2014<br />
**Wells, K. C., D. B. Millet, K. E. Cady-Pereira, M. W. Shephard, D. K. Henze, N. Bousserez, E. C. Apel, J. de Gouw, C. Warneke, H. B. Singh, Quantifying global terrestrial methanol emissions using observations from the TES satellite sensor, Atmos. Chem. Phys., 14, 2555-2570<br />
<br />
* '''2013'''<br />
**Walker, T. W., D. B. A. Jones, D. K. Henze, Z. Jiang, M. Parrington, F. Paulot, and Y. Rochon, Adjoint sensitivity analysis of North American surface ozone concentrations: Implications for dry deposition, American Geophysical Union, Fall Meeting 2013, abstract id. A33L-03.<br />
**Deng, F., D. B. A. Jones, D. K. Henze, N. Bousserez, K. W. Bowman, J. B. Fisher, R. Nassar, C. O'Dell, D. Wunch, P. O. Wennberg, E. A. Kort, S. C. Wofsy, T. Blumenstock, N. M. Deutscher, D. Griffith, F. Hase, P. Heikkinen, V. Sherlock, K. Strong, R. Sussmann, and T. Warneke, Inferring regional sources and sinks of atmospheric CO2 from GOSAT XCO2 data, Atmos. Chem. Phys. Discuss., 13, 26327-26388<br />
**Meland, B., X. Xu, D. K. Henze, J. Wang, Assessing remote polarimetric measurements sensitivities to aerosol emissions using the GEOS-Chem adjoint model, Atmos. Meas. Tech. Discuss., 6, 5447-5493, doi:10.5194/amtd-6-5447-2013<br />
**Xu, X., J. Wang, D. K. Henze, W. Qu, M. Kopacz, Constraints on Aerosol Sources Using GEOS-Chem Adjoint and MODIS Radiances, and Evaluation with Multi-sensor (OMI, MISR) data, J. Geophys. Res., 118, doi:10.1002/jgrd.50515.<br />
**Paulot, F., D. J. Jacob and D. K. Henze, Sources and processes contributing to nitrogen deposition in biodiversity hotspots worldwide, Environ. Sci. Technol, 47, 3226-3233, doi:10.1021/es3027727.<br />
**Koo, J., Q. Wang, D. K. Henze, I. A. Waitz, S.R.H. Barrett, Spatial sensitivities of human health risk to intercontinental and high-altitude pollution, Atmos. Environ., 71, 140-147.<br />
**Kharol, S., R. V. Martin, S. Philip, S. Vogel, D. K. Henze, D. Chen, Y. Wang, Q. Zhang, C. L. Heald, Persistent Sensitivity of Asian Aerosol to Emissions of Nitrogen Oxides, Geophys. Res. Lett., 40, 1021-1026, doi:10.1002/grl.50234.<br />
**Jiang, Z., D. B. A. Jones, H. M. Worden, M. N. Deeter, D. K. Henze, J. Worden, and K. W. Bowman, Quantifying the impact of model biases in convective transport on inferred CO source estimates using multi-spectral CO retrievals from MOPITT, J. Geophys. Res.,118, doi:10.1029/jgrd.50216<br />
**L. Zhu, D. K. Henze, K. E. Cady-Pereira, M. W. Shephard, M. Luo, R. W. Pinder, J. O. Bash, G. Jeong, Constraining U.S. ammonia emissions using TES remote sensing observations and the GEOS-Chem adjoint model, J. Geophys. Res., 118, doi:10.1002/jgrd.50166<br />
**Koo, J., Q. Wang, D. K. Henze, I. A. Waitz, S.R.H. Barrett, Spatial sensitivities of human health risk to intercontinental and high-altitude pollution, Atmos. Environ., 71, 140-147<br />
<br />
* '''2012'''<br />
**Bowman, K. W., and D. K. Henze, Attribution of direct ozone radiative forcing to spatially-resolved emissions, Geophys. Res. Lett., 39, L22704, doi:10.1029/2012GL053274.<br />
**Henze, D. K., D. T. Shindell, F. Akhtar, R. J. D. Spurr, R. W. Pinder, D. Loughlin, M. Kopacz, K. Singh, and C. Shim, Spatially refined aerosol direct radiative forcing efficiencies, Environ. Sci. Technol., 46, 9511 - 9518, dx.doi.org/10.1021/es301993s.<br />
**Karydis, V. A., S. L. Capps, R. H. Moore, A. Russell, D. K. Henze, and A. Nenes, Using a global aerosol model adjoint to unravel the footprint of spatially-distributed emissions on cloud droplet number and cloud albedo, Geophys. Res. Lett., 39, L24804, doi:10.1029/2012GL053346.<br />
**Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden, The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098.<br />
**Paulot, F., D. K. Henze, and P. O. Wennberg, Impact of the isoprene photochemical cascade on tropical ozone, Atmos. Chem. Phys., 12, 1307-1325.<br />
**Singh, K. and A. Sandu, 2012: Variational chemical data assimilation with approximate adjoints. Computers and Geosciences, 40, 10-18.<br />
**Turner, A., D. K. Henze, R. V. Martin, and A. Hakami, The spatial extent of source influences on modeled column concentrations of short-lived species, Geophys. Res. Lett., 39, L12806, doi:10.1029/2012GL051832.<br />
**Walker, T., D. B. A. Jones, M. Parrington, D. K. Henze, L. T. Murray, J. W. Bottenheim, K. Anlauf, J. R. Worden, K. W. Bowman, C. Shim, K. Singh, M. Kopacz, D. W. Tarasick, J. Davies, P. von der Gathen, and C. C. Carouge, Impacts of midlatitude precursor emissions and local photochemistry on ozone abundances in the Arctic, J. Geophys. Res.,117, D01305 doi:10.1029/2011JD016370.<br />
**Wang, J., X. Xu, D. K. Henze, Q. Ji, S.-C. Tsay, J. Huang, Top-Down Estimate of Dust Emissions through Integration of MODIS and MISR Aerosol Retrievals with the GEOS-Chem adjoint model, Geophys. Res. Lett., 39, L08802.<br />
**Wecht, K. J., D. J. Jacob, S. C. Wofsy, E. A. Kort, J. R. Worden, S. S. Kulawik, D. K. Henze, M. Kopacz, and V. H. Payne, Validation of TES methane with HIPPO aircraft observations: implications for inverse modeling of methane sources, Atmos. Chem. Phys., 12, 1823-1832.<br />
**Singh, K. and A. Sandu (2012). "Variational chemical data assimilation with approximate adjoints." Computers and Geosciences 40: 10-18.<br />
<br />
* '''2011'''<br />
**Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald, Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
**Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald (2011), Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
**Kopacz, M., D. L. Mauzerall, J. Wang, E. M. Leibensperger, D. K. Henze, and K. Singh, Origin and radiative forcing of black carbon transported to the Himalayas and Tibetan Plateau, Atmos. Chem. Phys., 11, 2837-2852.<br />
<br />
* '''2010'''<br />
**Kopacz, M., D. J. Jacob, J. A. Fisher, J. A. Logan, L. Zhang, I. A Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010), Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atoms. Chem. Phys., 10, 855-876.<br />
**Kopacz, M., D.J. Jacob, J.A. Fisher, J. A. Logan, L. Zhang, I. A. Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010): Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atmos. Chem. Phys., 10, 855-876. http://www.atmos-chem-phys.net/10/855/2010/acp-10-855-2010.<br />
**Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden (2012), The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098<br />
**Singh, K., Jardak, M., Sandu, A., Bowman, K., Lee, M., and Jones, D. (2010): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. Discuss., 3, 1783-1827, doi:10.5194/gmdd-3-1783-2010. http://www.geosci-model-dev-discuss.net/3/1783/2010/gmdd-3-1783-2010.html<br />
<br />
* '''2009'''<br />
**Eller, P., K. Singh, A. Sandu, K. Bowman, D. K. Henze, and M. Lee (2009), Implementation and evaluation of an array of chemical solvers in a global chemical transport model, Geosci. Mod. Devel., 2, 185-207.<br />
**Henze, D. K., J. H. Seinfeld and D. T. Shindell, (2009), Inverse modeling and mapping U.S. air quality influences of inorganic PM2.5 precursor emissions with the adjoint of GEOS-Chem, Atoms. Chem. Phys., 9, 5877-5903.<br />
**Kopacz, M., D. J. Jacob, D. K. Henze, C. L. Heald, D. G. Streets, and Q. Zhang (2009), A comparison of analytical and adjoint Bayesian inversion methods for constraining Asian sources of CO using satellite (MOPITT) measurements of CO columns, J. Geophys. Res., 114, D04305, doi:10.1029/2007JD009264.<br />
**Pye, H. O. T., H. Liao, S. Wu, L. J. Mickely, D. J. Jacob, D. K. Henze, and J. H. Seinfeld (2009), Effect of changes in climate and emissions on future sulfate-nitrate-ammonium aerosol levels in the United States, J. Geophys. Res., 114, D01205, doi:10.1029/2008JD010701.<br />
**Zhang, L., D. J. Jacob, M. Kopacz, D. K. Henze, K. Singh, and D. A. Jaffe (2009), Intercontinental source attribution of ozone pollution at western U.S. sites using an adjoint method, Geophys. Res. Lett., 36, L11810, doi:10.1029/2009GL037950.<br />
<br />
* '''2007'''<br />
**Henze, D. K., A. Hakami and J. H. Seinfeld (2007), Development of the adjoint of GEOS-Chem, Atmos. Chem. Phys., 7, 2413-2433.<br />
<br />
=== Conference proceedings === <br />
* Adjoint inversion of CO sources using combined MOPITT, SCIAMACHY and AIRS CO columns, presented by Monika Kopacz at the COSPAR Scientific Assembly, Montreal, July 18, 2008. http://acmg.seas.harvard.edu/presentations/powerpoints/mak2008/COSPAR_MKopacz_July2008.ppt<br />
<br />
* Singh, K., P. Eller, A. Sandu, D. K. Henze, K. Bowman, M. Kopacz, and M. Lee (2009), Towards the construction of a standard geos-chem adjoint model, ACM High Performance Computing Conference.<br />
<br />
* Kopacz, M., Mauzerall, D.L., Leibensperger, E.M., Wang, J., Henze, D.K., Singh, K., Shim, C. Identifying the origin and estimating the radiative forcing of BC in the Himalayas: an analysis using the global GEOS-Chem adjoint model, European Geophysical Union meeting, Vienna, May 4, 2010.<br />
<br />
* Kopacz, M., Jacob, D.J., Fisher, J.A., Logan, J.A., Zhang, L., Megretskaia, I.A., Yantosca, R.M., Singh, K., Henze, D.K., Burrows, J.P., Buchwitz, M., Khlystova, I., McMillan, W.W., Gille, J.C., Edwards, D.P., Eldering, A., Thouret, V., Nedelec, P. Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), European Geophysical Union meeting, Vienna, May 7, 2010.<br />
<br />
* Tang, J., Zhuang, Q. and Xiong, X. (2010), 4D-Var inversion of atmospheric methane fluxes by assimilating SCIAMACHY and AIRS satellite retrievals, AGU, Dec. 18, 2010, http:/web.ics.purdue.edu/~tang16/agu2010_tang.ppt<br />
<br />
* Bousserez, N., R. V. Martin, K. W. Bowman, D.K. Henze, M. Kopacz, K. Singh, C. Shim, C. Wespes, Improving the lightning NOx source using satellite observations: a 4D-var analysis approach, AGU, Dec., 2010, http://myweb.dal.ca/nc689777/AGU_liNOx_poster_final.pdf</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint&diff=22106GEOS-Chem Adjoint2015-07-03T05:24:52Z<p>Hyung-Min Lee: /* Current GEOS-Chem Adjoint Research Projects (please add yours!) */</p>
<hr />
<div><big><big><strong>Adjoint Working Group</strong></big></big><br />
<br />
== Contact information ==<br />
<br />
{| border=1 cellspacing=0 cellpadding=5<br />
<br />
|-<br />
|bgcolor="#CCCCCC"|'''Adjoint Model Scientist'''<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|bgcolor="#CCCCCC"|'''Adjoint Model and Data Assimilation<br>Working Group Co-Chairs'''<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman], [mailto:dbj@atmosp.physics.utoronto.ca Dylan B.A. Jones]<br />
|-<br />
|bgcolor="#CCCCCC"|'''GC adjoint support team'''<br />
|[mailto:yanko.davila@colorado.edu Yanko Davila], [mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez]<br />
|-<br />
|bgcolor="#CCCCCC"|'''Adjoint Working Group email list'''<br />
|<tt>geos-chem-adjoint@seas.harvard.edu</tt><br />
|-<br />
|bgcolor="#CCCCCC"|'''To subscribe to email list'''<br />
|Send email to [mailto:geos-chem-adjoint-join@seas.harvard.edu <tt>geos-chem-adjoint-join@seas.harvard.edu</tt>]<br />
|-<br />
|bgcolor="#CCCCCC"|'''To unsubscribe from email list'''<br />
|Send email to [mailto:geos-chem-adjoint-leave@seas.harvard.edu <tt>geos-chem-adjoint-leave@seas.harvard.edu</tt>]<br />
|}<br />
<br />
== Historical Development ==<br />
Original work on the adjoint of GEOS-Chem v6 began in 2003, focusing on the adjoint of the offline aerosol simulation. By 2005, the adjoint was expanded to include a tagged CO simulation and a full chemistry simulation; an adjoint of GEOS-Chem v7 was also developed in the following years. Each of these branches of the adjoint code were been constructed in a hybrid fashion using a combination of automatic differentiation software ([http://www.autodiff.com/tamc TAMC], [http://people.cs.vt.edu/~asandu/Software/Kpp/ KPP]) and manual coding of both discrete and continuous adjoints. They shared many common elements yet had unique features for different applications. <br />
<br />
During the summer of 2009, the existing branches were merged and updated to bring the adjoint into alignment with the latest release of GEOS-Chem, v8-02-01. This merged adjoint model is now the standard adjoint code into which all further development efforts will be placed.<br />
<br />
== Forward Model Code ==<br />
<br />
The forward model on which the adjoint is based originally corresponded to GEOS-Chem v8-02-01. It was subsequently updated as follows:<br />
<br />
* KPP solver for gas-phase chemistry (as in GCv8-02-03)<br />
* Implement Bond 2007 BC/OC emissions (as in GCv8-02-02)<br />
* Apply bug fixes from GCv8-02-02 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-02#Previous_issues_now_resolved_in_v8-02-02 here]<br />
* Apply bug fixes from GCv8-02-03 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-03#Previous_issues_now_resolved_in_v8-02-03 here]<br />
* Apply bug fixes from GCv8-02-04 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-04#Previous_issues_now_resolved_in_v8-02-04 here]<br />
<br />
All bug fixes and model updates were previous listed at the top of inverse_driver.f. We have now switched to documenting the code development cycle here in the wiki, see the following section. <br />
<br />
== Code Versions, Bug Fixes and Developments == <br />
<br />
=== Current GEOS-Chem adjoint version released ===<br />
*[[GEOS-Chem_Adjoint_v35]] (You will download this version when you check out.)<br />
<br />
=== Previous GEOS-Chem adjoint versions released ===<br />
*[[GEOS-Chem_Adjoint_v34]]<br />
*[[GEOS-Chem_Adjoint_v33]] <br />
*[[GEOS-Chem_Adjoint_v32]]<br />
*[[GEOS-Chem_Adjoint_v31]]<br />
<br />
== Summary of Main Adjoint Code Supported Features ==<br />
<br />
=== Features === <br />
* Meteorological fields<br />
** GEOS-3 <i>needs testing</i><br />
** GEOS-4<br />
** GEOS-5 <br />
* model resolution<br />
** 4 x 5<br />
** 2 x 2.5 <br />
** Nested Asia and NA<br />
* Forward model processes<br />
** convection<br />
** advection<br />
** PBL mixing<br />
** dry deposition<br />
** wet deposition<br />
** strat / trop exchange with LINOZ and new GMI strat chem (v9-01-03)<br />
** NOy up fluxes (now replaced with new GMI strat chem)<br />
** aerosols <br />
*** inorganic aerosol thermodynamics with RPMARES<br />
*** inorganic aerosol thermodynamics with ISORROPIA <i> in progress </i><br />
*** sulfate chemistry <br />
*** BC<br />
*** SOA, Dust, sea salt <i> needs doing </i><br />
*** aerosol surface area feedbacks <i> needs updating </i><br />
*** aerosol optical feedbacks <i> needs doing </i><br />
** emissions<br />
*** all standard emissions included<br />
* Simulation modes<br />
** full chemistry<br />
** tagged CO<br />
** tagged Ox <br />
** CH4<br />
** offline aerosols (for BC and dust only)<br />
** CO2<br />
* Observational Operators<br />
** MOPITT CO column<br />
** SCIAMACHY CO column<br />
** AIRS CO column<br />
** IMPROVE BC<br />
** CASTNet (NH4+) <i> needs updating </i> <br />
** GOME / SCIAMACHY NO2 column <i> needs updating </i> <br />
*** using KNMI retrieval (Henze)<br />
*** using Dalhousie retrieval (Shim)<br />
** SCIAMACHY/OMI NO2<br />
*** using Dalhousie retrieval (Bousserez, Padmanabhan)<br />
** TES NH3 <br />
** TES O3 <br />
** GOSAT CO2 <br />
** MLS O3 and TES CO2 <i> in progress </i><br />
* Control parameters<br />
** Initial Conditions scaling factors (linear or log)<br />
** Emissions scaling factors (linear or log) <br />
*** NH3, primary BC/OC, SO2: anthropogenic, natural, bioburn, biomass, ship<br />
*** NOx: soil, aircraft, anthropogenic, biofuel, bioburn<br />
*** Lightning NOx: injection height, yield <i> in progress </i><br />
*** all other gas-phase tracers: anthropogenic, biofuel, bioburn <br />
* Adjoint sensitivities<br />
** w.r.t. all implemented control parameters<br />
** w.r.t Reaction Rate Parameters <br />
** w.r.t all emissions <br />
** of AQ attainment metrics <i> needs updating </i><br />
** of spatiotemporally averaged species concentrations (e.g., arctic O3) <br />
* Other<br />
** Inverse Hessian approximation <br />
** off-diagonal covariance matrices <i> needs updating </i><br />
** 3D-Var <i> needs updating </i><br />
<br />
Features may be qualified as: <br />
* <i>needs testing</i>: an implemented feature that we haven't fully used yet<br />
* <i>needs updating</i>: a feature developed with a previous branch that has yet to be updated to GEOS-Chem v8 and the merged adjoint<br />
* <i>needs doing</i>: a feature nobody has tackled the adjoint of yet<br />
* <i>in progress</i>: a feature currently under development <br />
* <i>in pipeline</i>: a feature which has been submitted and awaiting integration into the CVS repository<br />
<br />
=== Primary code developers ===<br />
Monika Kopacz, Kumaresh Singh, Changsub Shim, Daven Henze<br />
<br />
=== Adjoint model lead scientist ===<br />
Daven Henze<br />
<br />
<br />
<br />
== Resources ==<br />
<br />
=== User's guide ===<br />
A user's guide v35 is available. [http://adjoint.colorado.edu/~yanko/gcadj_std/GC_adj_man.pdf User's Guide v35] <br />
<br />
Previous version v32-v33 available at [http://adjoint.colorado.edu/~daven/gcadj_std/GC_adj_man.pdf User's Guide v32]<br />
<br />
Quick Introduction to GitLab is available at [http://adjoint.colorado.edu/~yanko/gcadj_std/GitLab_Tutorial.pdf GitLab Tutorial]<br />
<br />
=== Code flowchart ===<br />
Meemong Lee has created a detailed flowchart of the inverse model code structure. http://adjoint.colorado.edu/~daven/gcadj_std/flowchart.pdf<br />
<br />
=== Plotting tools ===<br />
<br />
Some IDL and MATLAB routines for plotting benchmark results. http://adjoint.colorado.edu/~daven/gcadj_std/tools.tar.gz<br />
<br />
=== Background papers and presentations ===<br />
<br />
Several articles and presentations (including a GC adjoint modeling clinic overview from IGC5) providing background information about adjoints. http://adjoint.colorado.edu/~daven/gcadj_std/adj_articles.tar.gz<br />
<br />
== Distribution and Use ==<br />
Code for the adjoint is distributed through GITLAB, a web interface connected to a GIT server located at adjoint.colorado.edu. You can access GITLAB at http://adjoint.colorado.edu:8080 after your account is created. Here is our [[Quick Start Guide.]] <br />
<br />
Even if your office mate has a copy of the code, the best way to obtain the model is to get an account for yourself and download a version from the repository. So please do not copy code directly from others or pass the code along to third parties. This vastly helps with tracking developments and keeping up with model updates. <br />
<br />
Use of the adjoint model code follows standard practice for GEOS-Chem. It is expected that any developments that come of individual applications based on this community model will eventually be given back to the community by incorporation of new developments into the standard adjoint code. New development should be submitted to Daven Henze for inclusion in the standard adjoint model code.<br />
<br />
Using GIT gives the users the ability to change the code and commit their changes without affecting the main repository hosted at adjoint.colorado.edu. Users can work with their modified versions of the code and even create their own tags because GIT acts as a local repository. When ready to submit your update to the community just create a new branch with your modifications. Send an email Daven Henze explaining your contributions and we'll do our best to include them as soon as possible.<br />
<br />
=== Quick guide to GIT === <br />
As of version 34 we started using git versioning system and the GitLab web interface. We prepared [http://adjoint.colorado.edu/~yanko/gcadj_std/GitLab_Tutorial.pdf GitLab Tutorial] to help users get used to the web interface and git. We recommend first taking a look at GIT manual to get a general feel for how this tools works (e.g., [http://git-scm.com/documentation GIT Documentation] or [http://www.kernel.org/pub/software/scm/git/docs/git.html GIT Manual Page]). <br />
<br />
<br />
Useful GIT commands: <br />
<br />
Initial download:<br />
git clone ssh://git@adjoint.colorado.edu:2222/yanko.davila/gcadj_std.git<br />
<br />
Status of project vs the current repository:<br />
git status<br />
<br />
Check difference of files (differences have colors for easy reading)<br />
git diff --color <wildcard> [<wildcard>] <path>/foo_mod.f<br />
<br />
Checkout specific version<br />
git checkout <wildcard><br />
<br />
Replacing a file with the newest version from the repository:<br />
git checkout origin/master -- <path>/foo_mod.f<br />
<br />
Merging changes in a file: ([http://www.kernel.org/pub/software/scm/git/docs/git-merge.html Reference])<br />
git merge -m <wildcard> <br />
<br />
Comitting<br />
git commit -a<br />
<br />
Tagging a version<br />
git tag -a TAGNAME<br />
<br />
Deleting a tag<br />
git tag -d TAGNAME<br />
git push origin :refs/tags/TAGNAME<br />
<br />
Saving changes to repository<br />
git push<br />
<br />
Saving tags to repository<br />
git push --tags<br />
<br />
List the history of a file:<br />
git log -- <path>/foo_mod.f<br />
<br />
Add a file to the repository<br />
git add <file_name><br />
<br />
Delete a file from the repository<br />
git rm <file_name><br />
<br />
Determine current version<br />
git show HEAD [ | grep commit]<br />
<br />
Download remote changes, rewind your local branch, then replays all your changes over the top of your current branch one by one, until you’re all up to date.<br />
git pull --rebase<br />
<br />
There are several wildcards that you can use on git for example:<br/><br />
''"origin/master"'' - Latest version on the repository <br/><br />
''"HEAD"'' - Latest version as of your last download <br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/72773d7b1a7bd757957d2c186acb3dbb107d5323 v33i]"'' - Specific TAG, find all tag names on GitLab<br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/32d5c926ef529c1dfc640c9dec4a925b24db575c 32d5c926e]"'' - Specific COMMIT, find all commit numbers on GitLab <br/><br />
<br />
[[Using_Git_with_GEOS-Chem|Here is the foward model documentation of git.]]<br />
<br />
=== Backward compatibility (CVS) === <br />
<br />
For people using old version of the code we still have active our CVS repository, but note that the latest version on CVS is v33i-patch2. Here you can find our [[Quick Guide to CVS]]<br />
<br />
== Crediting GEOS-Chem adjoint developers ==<br />
<br />
We aim to make distribution of adjoint model code as immediate as possible. A consequence is that many features may not yet be publicly documented. Therefore, giving code developers due credit is of utmost importance. <br />
<br />
Authors of new additions to the standard code should be offered co-authorship on the first round of presentations and publications to come of their development. Features currently falling in this category and their developers are:<br />
* (v35) HTAP Emissions Inventory. Kateryna Lapina, Daven Henze and Yanko Davila, CU Boulder.<br />
* (v35) NEI2008 Emissions Inventory. Katie Travis; Fabien Paulot, Harvard; Hyungmin Lee, and Daven Henze, CU Boulder. <br />
* (v35) Deposition based cost function. Fabien Paulot, Harvard; Daven Henze and Yanko Davila, CU Boulder. <br />
* (v34) Implementation of the sensitivity to reaction rate constants. Developers: Hyungmin Lee, CU Boulder; Thomas ; Fabien Paulot (Harvard); Daven Henze, CU Boulder and Yanko Davila, CU Boulder. <br />
* (v34) ISOROPIA II adjoint. Developer: Shannon Capps, EPA.<br />
* (v34) Off-diagonal covariance error matrices implementation. Developers: Nicolas Bousserez, CU Boulder; Kumaresh; Yanko Davila, CU Boulder. <br />
* (v32) Nested full chemistry adjoint. Developers: Zhe Jiang, University of Toronto; Daven Henze, CU Boulder.<br />
<br />
<br />
Citation of the appropriate [[GEOS-Chem_Adjoint#Journal_Articles|journal articles]] for mature developments is also encouraged, as well as considering aspects of [http://acmg.seas.harvard.edu/geos/geos_credit.html co-authorship for the forward model]. <br />
<br />
Overall, if you have any questions about authorship, even for a conference presentation, please contact Daven Henze.<br />
<br />
== Current GEOS-Chem Adjoint Research Projects (please add yours!) ==<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|- bgcolor="#cccccc"<br />
!User Group <br />
!Description <br />
!Contact Person<br />
|-<br />
|CU Boulder<br />
|Aerosol precursors, CO2, O3; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|CU Boulder<br />
|Inverse modeling/optimization: variational source inversions, posterior error estimates in high-dimension; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Nicolas Bousserez]<br />
|-<br />
|CU Boulder<br />
|Identifying long-range sources of atmospheric pollutants utilizing an adjoint method<br />
|[mailto:alicia.camacho@valpo.edu Alicia Camacho]<br />
|-<br />
|CU Boulder<br />
|Sensitivity of nitrate deposition in Federal Class I areas in the US<br />
|[mailto:hyungmin.lee@colorado.edu Hyung-Min Lee]<br />
|-<br />
|CU Boulder<br />
|Constraints on Aerosol Sources<br />
|[mailto:zhen.qu@colorado.edu Zhen Qu]<br />
|-<br />
|Harvard<br />
|Methane<br />
|Alex Turner, aturner [at] fas.harvard.edu<br />
|-<br />
|Harvard<br />
|Methane<br />
|Bram Maasakkers, maasakkers [at] seas.harvard.edu<br />
|-<br />
|Harvard<br />
|Smoke Emissions in SE Asia<br />
|Patrick Kim, kim68 [at] fas.harvard.edu<br />
|-<br />
|Harvard<br />
|Harvard Emissions Component (HEMCO)<br />
|[mailto:ckeller@seas.harvard.edu Christoph Keller]<br />
|-<br />
|MIT<br />
|Air quality, aircraft emissions and sensitivities<br />
|[mailto:dedoussi@mit.edu Irene Dedoussi]<br />
|-<br />
|Dalhousie University<br />
|Mass balance vs adjoint approaches for top-down constraints on NOx emissions<br />
|Matthew Cooper<br />
|-<br />
|Dalhousie University<br />
|Trends in OC and BC emissions over the US <br />
|Nathaniel Egan-Pimblett<br />
|-<br />
|Dalhousie University<br />
|Sensitivity of global PM2.5-induced mortality to emissions<br />
|[mailto:colin.lee@dal.ca Colin Lee]<br />
|-<br />
|JPL<br />
|Microwave Limb Sounder (MLS) Ozone assimilation and CMS<br />
|[mailto:meemong.lee@jpl.nasa.gov Meemong Lee]<br />
|-<br />
|JPL <br />
| TES ozone assimilation/attribution of ozone radiative forcing<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|University of Edinburgh<br />
|Quantifying the impact of boreal forest fires on tropospheric oxidants over the Atlantic<br />
|[mailto:mark.parrington@ed.ac.uk Mark Parrington]<br />
|-<br />
|Peking University<br />
|Satellite constraints on VOC emissions <br />
|[mailto:tmfu@pku.edu.cn May Fu]<br />
|-<br />
|UEA<br />
|CO2 assimilation<br />
|[mailto:zhaohui.chen@uea.ac.uk Chen]<br />
|-<br />
|Purdue University<br />
|Feedback between terrestrial ecosystem processes and atmospheric CO2 signals<br />
|[mailto:zhuq@purdue.edu Qing Zhu]<br />
|-<br />
|Purdue University<br />
|Feedback between aquatic ecosystem processes and atmospheric CH4 signals<br />
|[mailto:tan80@purdue.edu Zeli Tan]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and reactive nitrogen to precursor emissions<br />
|[mailto:twalker@atmosp.physics.utoronto.ca Thomas Walker]<br />
|-<br />
|JPL<br />
|Adjoint analysis for carbon monoxide and ozone<br />
|[mailto:Zhe.Jiang@jpl.nasa.gov Zhe Jiang]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and CO to precursor emissions<br />
|[mailto:cwhaley@atmosp.physics.utoronto.ca Cynthia Whaley]<br />
|-<br />
|Georgia Tech / CU Boulder<br />
|ISORROPIA adjoint development; NH3 assimilation; cloud droplet sensitivities<br />
|[mailto:shannon.capps@colorado.edu Shannon Capps]<br />
|-<br />
|Peking University<br />
|Source attributions of tropospheric ozone over North China<br />
|[mailto:linjt@pku.edu.cn Jintai Lin]<br />
|-<br />
|University of Wollongong<br />
|Sensitivity of ozone and adjoint analysis of CO over Australasia.<br />
|[mailto:rb864@uowmail.edu.au Rebecca Buchholz]<br />
|-<br />
|University of Toronto<br />
|CO2 assimilation & transport model bias estimation<br />
|[mailto:mkeller@atmosp.physics.utoronto.ca Martin Keller]<br />
|-<br />
|University of Wisconsin<br />
|CO2 assimilation and forecast & Temperature profile retrieval <br />
|[mailto:wenguang.bai@ssec.wisc.edu Wenguang Bai]<br />
|-<br />
|University of Leicester (UK)<br />
|Top-down estimates of Amazon isoprene emissions<br />
|[mailto:mpb14@le.ac.uk Michael Barkley]<br />
|-<br />
|Anyang University<br />
|Aerosol emission modeling in East Asia <br />
|[mailto:koo@anyang.ac.kr Youn Seo Koo]<br />
|-<br />
|Nanjing University<br />
|Inverse modeling of terrestrial ecosystem carbon flux<br />
|[mailto:hengmao.wang@gmail.com Hengmao Wang]<br />
|-<br />
|Tsinghua University<br />
|Nested-gird simulations with the adjoint model<br />
|[mailto:nany12@mails.tsinghua.edu.cn Nan Yang]<br />
|-<br />
|University of Minnesota<br />
|Inverse modeling of VOC sources based on TES and IASI measurements<br />
||[mailto:kcw@umn.edu Kelley Wells], [mailto:dbm@umn.edu Dylan Millet]<br />
|-<br />
|University of Minnesota & CU Boulder<br />
|Inverse modeling of N2O sources<br />
|[mailto:dbm@umn.edu Dylan Millet], [mailto:Daven.Henze@Colorado.EDU Daven Henze]<br />
|-<br />
|Tsinghua University<br />
|Inverse modeling of anthropogenic emissions over East Asia<br />
|[mailto:qiangzhang@tsinghua.edu.cn Qiang Zhang]<br />
|-<br />
|UCLA<br />
|Constrain black carbon emission<br />
|[mailto:qiling@atmos.ucla.edu Ling Qi]<br />
|-<br />
|UCLA<br />
|source attribution of ozone in the western U.S.<br />
|[mailto:gaomei@atmos.ucla.edu Mei Gao]<br />
|-<br />
|University of Toronto<br />
|Methane inverse modeling<br />
|[mailto:stanevich@atmosp.physics.utoronto.ca Ilya Stanevich]<br />
|-<br />
|}<br />
<br />
== Publications ==<br />
<br />
=== Journal Articles ===<br />
<br />
* '''In press or sumbitted'''<br />
**Turner, A. J., D. J. Jacob, K. J. Wecht, J. D. Maasakkers, S. C. Biraud, H. Boesch, K. W. Bowman, N. M. Deutscher, M. K. Dubey, D. W. T. Griffith, F. Hase, A. Kuze, J. Notholt, H. Ohyama, R. Parker, V. H. Payne, R. Sussmann, V. A. Velazco, T. Warneke, P. O. Wennberg, and D. Wunch, Estimating global and North American methane emissions with high spatial resolution using GOSAT satellite data, submitted to Atmos. Chem. Phys.<br />
**Turner, A. J. and D. J. Jacob, Balancing aggregation and smoothing errors in inverse models, submitted to Atmos. Chem. Phys.<br />
**Liu, J., Bowman, K., and Henze, D., Source-receptor relationships of column-average CO2 and implications for the impact of observations on flux inversions. JGR-Atmosphere, in press.<br />
**Whaley, C. H., K. Strong, D. B. A. Jones, T. W. Walker, Z. Jiang, D. K. Henze, M. Cooke, C. A. McLinden, M. Pommier, R. L. Mittermeier, P. F. Fogal, Improvements to our understanding of urban ozone air pollution: Sources of Toronto-area ozone during poor air quality events, submitted.<br />
**Zhao, Y. H., L. Zhang, Y. P. Pan, Y. S. Wang, F. Paulot, and D. K. Henze, Atmospheric nitrogen deposition to the northwestern Pacific: seasonal variation and source attribution, submitted.<br />
**Lapina, K., D. K. Henze, J. B. Milford, C. Cuvelier, and M. Seltzer, Implications of RCP Scenarios for future changes in vegetative exposure to ozone in the Western U.S., Geophys. Res. Let., in press.<br />
**Zhang, L., L. Licheng, Y. Zhao, S. Gong, X. Zhang, D. K. Henze, S. L. Capps, T.-M. Fu, Q. Zhang, Source attribution of particulate matter pollution over North China with the adjoint method, submitted.<br />
**Walker, T. W., D. B. A. Jones, D. K. Henze, Z. Jiang, M. Parrington, F. Paulot, and Y. Rochon, Adjoint sensitivity analysis of North American surface ozone concentrations: Implications for dry deposition, submitted.<br />
**Keller, M, D. B. A. Jones, Z. Jiang, D. K. Henze, and H. Worden, Quantifying model biases in CO emission estimation using weak constraint 4D-Var, submitted.<br />
**Bousserez, N., D. K. Henze, B. Rooney, A. Perkins, K. J. Wecht, A. J. Turner, V. Natraj, J. R. Worden, Constraints on methane emissions in North America from future geostationary remote sensing measurements, submitted.<br />
<br />
* '''2015'''<br />
**Patrick S. Kim, Daniel J. Jacob, Loretta J. Mickley, Shannon N. Koplitz, Miriam E. Marlier, Ruth S. DeFries, Samuel S. Myers, Boon Ning Chewf, Yuhao H. Mao, Sensitivity of population smoke exposure to fire locations in Equatorial Asia, doi:10.1016/j.atmosenv.2014.09.045<br />
**Deng, F., D. B. A. Jones, T. W. Walker, M. Keller, K. W. Bowman, D. K. Henze, R. Nassar, E. A. Kort, S. C. Wofsy, K. A. Walker, A. E. Bourassa, and D. A. Degenstein, Sensitivity analysis of the potential impact of discrepancies in stratosphere-troposphere exchange on inferred sources and sinks of CO2, Atmos. Chem. Phys. Discuss., 15, 10813-10851, doi:10.5194/acpd-15-10813-2015.<br />
**Jiang, Z., D. B. A. Jones, J. R. Worden, H. M. Worden, D. K. Henze, Y. X. Wang, Regional data assimilation of multi-spectral MOPITT observations of CO over North America, Atmos. Chem. Phys. Discuss., 15, 5327-5358, doi:10.5194/acpd-15-5327-2015.<br />
**Zhu, L., D. K. Henze, J. Bash, G. Jeong, K. Cady-Pereira, M. Shephard, M. Luo, F. Paulot, and S. Capps, Global evaluation of ammonia bi-directional exchange, Atmos. Chem. Phys. Discuss., 15, 4823-4877, doi:10.5194/acpd-15-4823-2015.<br />
**Jiang, Z., J. R. Worden, D. B. A. Jones, J.-T. Lin, W. Verstraeten, and D. K. Henze, Constraints on Asian ozone using Aura TES, OMI and Terra MOPITT, Atmos. Chem. Phys., 15, 99-112.<br />
**Lee, C.J., R.V. Martin, Henze D.K., Brauer M., Cohen A., and A. van Donkelaar, Sensitivity of global particulate-matter-related mortality to local precursor emissions, Environ. Sci. Technol., 49(7), 4335–4344, doi:10.1021/acs.est.5b00873, 2015.<br />
**Bousserez, N., D. K. Henze, A. Perkins, K. W. Bowman, M.Lee, J.Liu, D.B.A. Jones, F. Deng, Improved analysis-error covariance matrix for high-dimensional variational inversions: application to source estimation using a 3D atmospheric transport model. Q.J.R. Meteorol. Soc.. doi: 10.1002/qj.2495<br />
<br />
* '''2014'''<br />
**Zhang, L., D. K. Henze, G. A. Grell, G. R. Carmichael, N. Bousserez, Q. Zhang, J. Cao, and Y. Mao, Constraining black carbon aerosol over Southeast Asia using OMI aerosol absorption optical depth and the adjoint of GEOS-Chem, Atmos. Chem. Phys. Discuss. 14, 28385-28452.<br />
**Jiang, Z., D. B. A Jones, H. M. Worden, and D. K. Henze, Sensitivity of inferred regional CO source estimates to the vertical structure in CO as observed by MOPITT, Atmos. Chem. Phys. Discuss., 14, 22939-22984<br />
**Zhu, Q., Q. Zhuang, D. K. Henze, K. Bowman, M. Chen, Y. Liu, Y. He, H. Matsueda, T. Machida, and Y. Sawa, Constraining terrestrial ecosystem CO2 fluxes by integrating models of biogeochemistry and atmospheric transport and data of surface carbon fluxes and atmospheric CO2 concentrations, Atmos. Chem. Phys. Discuss., 14, 22587-22638<br />
**Mao, Y. H., Q. B. Li, D. K. Henze, Z. Jiang, D. B. A. Jones, M. Kopacz, C. He, L. Qi, M. Gao, W.-M. Hao, and K.-N. Liou, Variational estimates of black carbon emissions in the western United States, Atmos. Chem. Phys. Discuss., 14, 21865-21916<br />
**Liu, J., Bowman, K., Lee, M., Henze, D., Bousserez, N., Brix, H., Collatz, G., Menemenlis, D., Ott, L., Pawson, S., Jones, D., Nassar, R.. Carbon monitoring system flux estimation and attribution: impact of ACOS-GOSAT XCO2 sampling on the inference of terrestrial biospheric sources and sinks. Tellus B, North America, 66, may. 2014. Available at: http://www.tellusb.net/index.php/tellusb/article/view/22486<br />
** Lee, H., D. K. Henze, B. Alexander, and L. T., Murray, Investigating the sensitivity of surface-level nitrate seasonality in Antarctica to primary sources using a global model, Atmos. Environ., 89, 757--767, doi:0.1016/j.atmosenv.2014.03.003<br />
**Paulot, F., D. J. Jacob, R. W. Pinder, J. O. Bash, K. Travis, D. K. Henze, Ammonia emissions in the United States, Europe, and China derived by high-resolution inversion of ammonium wet deposition data: Interpretation with a new agricultural emissions inventory (MASAGE_NH3), J. Geophys. Res., 119, 7, 4343--4364, doi:10.1002/2013JD021130.<br />
**Lapina, K., D. K. Henze, J. B. Milford, M. Huang, M. Lin, A. M. Fiore, G. Carmichael, G. G. Pfister, and K. W. Bowman, Assessment of source contributions to seasonal vegetative exposure to ozone in the U.S., J. Geophys. Res., 119, 324-340, doi:10.1002/2013JD020905<br />
**Shen, Z., J. Liu, L. W. Horowitz, D. K. Henze, S. Fan, H. Levy II, D. L. Mauzerall, J. Lin, and S. Tao,.: Analysis of transpacific transport of black carbon during HIPPO-3: implications for black carbon aging, Atmos. Chem. Phys. Discuss., 14, 505-540, doi:10.5194/acpd-14-505-2014<br />
**Wells, K. C., D. B. Millet, K. E. Cady-Pereira, M. W. Shephard, D. K. Henze, N. Bousserez, E. C. Apel, J. de Gouw, C. Warneke, H. B. Singh, Quantifying global terrestrial methanol emissions using observations from the TES satellite sensor, Atmos. Chem. Phys., 14, 2555-2570<br />
<br />
* '''2013'''<br />
**Deng, F., D. B. A. Jones, D. K. Henze, N. Bousserez, K. W. Bowman, J. B. Fisher, R. Nassar, C. O'Dell, D. Wunch, P. O. Wennberg, E. A. Kort, S. C. Wofsy, T. Blumenstock, N. M. Deutscher, D. Griffith, F. Hase, P. Heikkinen, V. Sherlock, K. Strong, R. Sussmann, and T. Warneke, Inferring regional sources and sinks of atmospheric CO2 from GOSAT XCO2 data, Atmos. Chem. Phys. Discuss., 13, 26327-26388<br />
**Meland, B., X. Xu, D. K. Henze, J. Wang, Assessing remote polarimetric measurements sensitivities to aerosol emissions using the GEOS-Chem adjoint model, Atmos. Meas. Tech. Discuss., 6, 5447-5493, doi:10.5194/amtd-6-5447-2013<br />
**Xu, X., J. Wang, D. K. Henze, W. Qu, M. Kopacz, Constraints on Aerosol Sources Using GEOS-Chem Adjoint and MODIS Radiances, and Evaluation with Multi-sensor (OMI, MISR) data, J. Geophys. Res., 118, doi:10.1002/jgrd.50515.<br />
**Paulot, F., D. J. Jacob and D. K. Henze, Sources and processes contributing to nitrogen deposition in biodiversity hotspots worldwide, Environ. Sci. Technol, 47, 3226-3233, doi:10.1021/es3027727.<br />
**Koo, J., Q. Wang, D. K. Henze, I. A. Waitz, S.R.H. Barrett, Spatial sensitivities of human health risk to intercontinental and high-altitude pollution, Atmos. Environ., 71, 140-147.<br />
**Kharol, S., R. V. Martin, S. Philip, S. Vogel, D. K. Henze, D. Chen, Y. Wang, Q. Zhang, C. L. Heald, Persistent Sensitivity of Asian Aerosol to Emissions of Nitrogen Oxides, Geophys. Res. Lett., 40, 1021-1026, doi:10.1002/grl.50234.<br />
**Jiang, Z., D. B. A. Jones, H. M. Worden, M. N. Deeter, D. K. Henze, J. Worden, and K. W. Bowman, Quantifying the impact of model biases in convective transport on inferred CO source estimates using multi-spectral CO retrievals from MOPITT, J. Geophys. Res.,118, doi:10.1029/jgrd.50216<br />
**L. Zhu, D. K. Henze, K. E. Cady-Pereira, M. W. Shephard, M. Luo, R. W. Pinder, J. O. Bash, G. Jeong, Constraining U.S. ammonia emissions using TES remote sensing observations and the GEOS-Chem adjoint model, J. Geophys. Res., 118, doi:10.1002/jgrd.50166<br />
**Koo, J., Q. Wang, D. K. Henze, I. A. Waitz, S.R.H. Barrett, Spatial sensitivities of human health risk to intercontinental and high-altitude pollution, Atmos. Environ., 71, 140-147<br />
<br />
* '''2012'''<br />
**Bowman, K. W., and D. K. Henze, Attribution of direct ozone radiative forcing to spatially-resolved emissions, Geophys. Res. Lett., 39, L22704, doi:10.1029/2012GL053274.<br />
**Henze, D. K., D. T. Shindell, F. Akhtar, R. J. D. Spurr, R. W. Pinder, D. Loughlin, M. Kopacz, K. Singh, and C. Shim, Spatially refined aerosol direct radiative forcing efficiencies, Environ. Sci. Technol., 46, 9511 - 9518, dx.doi.org/10.1021/es301993s.<br />
**Karydis, V. A., S. L. Capps, R. H. Moore, A. Russell, D. K. Henze, and A. Nenes, Using a global aerosol model adjoint to unravel the footprint of spatially-distributed emissions on cloud droplet number and cloud albedo, Geophys. Res. Lett., 39, L24804, doi:10.1029/2012GL053346.<br />
**Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden, The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098.<br />
**Paulot, F., D. K. Henze, and P. O. Wennberg, Impact of the isoprene photochemical cascade on tropical ozone, Atmos. Chem. Phys., 12, 1307-1325.<br />
**Singh, K. and A. Sandu, 2012: Variational chemical data assimilation with approximate adjoints. Computers and Geosciences, 40, 10-18.<br />
**Turner, A., D. K. Henze, R. V. Martin, and A. Hakami, The spatial extent of source influences on modeled column concentrations of short-lived species, Geophys. Res. Lett., 39, L12806, doi:10.1029/2012GL051832.<br />
**Walker, T., D. B. A. Jones, M. Parrington, D. K. Henze, L. T. Murray, J. W. Bottenheim, K. Anlauf, J. R. Worden, K. W. Bowman, C. Shim, K. Singh, M. Kopacz, D. W. Tarasick, J. Davies, P. von der Gathen, and C. C. Carouge, Impacts of midlatitude precursor emissions and local photochemistry on ozone abundances in the Arctic, J. Geophys. Res.,117, D01305 doi:10.1029/2011JD016370.<br />
**Wang, J., X. Xu, D. K. Henze, Q. Ji, S.-C. Tsay, J. Huang, Top-Down Estimate of Dust Emissions through Integration of MODIS and MISR Aerosol Retrievals with the GEOS-Chem adjoint model, Geophys. Res. Lett., 39, L08802.<br />
**Wecht, K. J., D. J. Jacob, S. C. Wofsy, E. A. Kort, J. R. Worden, S. S. Kulawik, D. K. Henze, M. Kopacz, and V. H. Payne, Validation of TES methane with HIPPO aircraft observations: implications for inverse modeling of methane sources, Atmos. Chem. Phys., 12, 1823-1832.<br />
**Singh, K. and A. Sandu (2012). "Variational chemical data assimilation with approximate adjoints." Computers and Geosciences 40: 10-18.<br />
<br />
* '''2011'''<br />
**Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald, Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
**Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald (2011), Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
**Kopacz, M., D. L. Mauzerall, J. Wang, E. M. Leibensperger, D. K. Henze, and K. Singh, Origin and radiative forcing of black carbon transported to the Himalayas and Tibetan Plateau, Atmos. Chem. Phys., 11, 2837-2852.<br />
<br />
* '''2010'''<br />
**Kopacz, M., D. J. Jacob, J. A. Fisher, J. A. Logan, L. Zhang, I. A Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010), Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atoms. Chem. Phys., 10, 855-876.<br />
**Kopacz, M., D.J. Jacob, J.A. Fisher, J. A. Logan, L. Zhang, I. A. Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010): Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atmos. Chem. Phys., 10, 855-876. http://www.atmos-chem-phys.net/10/855/2010/acp-10-855-2010.<br />
**Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden (2012), The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098<br />
**Singh, K., Jardak, M., Sandu, A., Bowman, K., Lee, M., and Jones, D. (2010): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. Discuss., 3, 1783-1827, doi:10.5194/gmdd-3-1783-2010. http://www.geosci-model-dev-discuss.net/3/1783/2010/gmdd-3-1783-2010.html<br />
<br />
* '''2009'''<br />
**Eller, P., K. Singh, A. Sandu, K. Bowman, D. K. Henze, and M. Lee (2009), Implementation and evaluation of an array of chemical solvers in a global chemical transport model, Geosci. Mod. Devel., 2, 185-207.<br />
**Henze, D. K., J. H. Seinfeld and D. T. Shindell, (2009), Inverse modeling and mapping U.S. air quality influences of inorganic PM2.5 precursor emissions with the adjoint of GEOS-Chem, Atoms. Chem. Phys., 9, 5877-5903.<br />
**Kopacz, M., D. J. Jacob, D. K. Henze, C. L. Heald, D. G. Streets, and Q. Zhang (2009), A comparison of analytical and adjoint Bayesian inversion methods for constraining Asian sources of CO using satellite (MOPITT) measurements of CO columns, J. Geophys. Res., 114, D04305, doi:10.1029/2007JD009264.<br />
**Pye, H. O. T., H. Liao, S. Wu, L. J. Mickely, D. J. Jacob, D. K. Henze, and J. H. Seinfeld (2009), Effect of changes in climate and emissions on future sulfate-nitrate-ammonium aerosol levels in the United States, J. Geophys. Res., 114, D01205, doi:10.1029/2008JD010701.<br />
**Zhang, L., D. J. Jacob, M. Kopacz, D. K. Henze, K. Singh, and D. A. Jaffe (2009), Intercontinental source attribution of ozone pollution at western U.S. sites using an adjoint method, Geophys. Res. Lett., 36, L11810, doi:10.1029/2009GL037950.<br />
<br />
* '''2007'''<br />
**Henze, D. K., A. Hakami and J. H. Seinfeld (2007), Development of the adjoint of GEOS-Chem, Atmos. Chem. Phys., 7, 2413-2433.<br />
<br />
=== Conference proceedings === <br />
* Adjoint inversion of CO sources using combined MOPITT, SCIAMACHY and AIRS CO columns, presented by Monika Kopacz at the COSPAR Scientific Assembly, Montreal, July 18, 2008. http://acmg.seas.harvard.edu/presentations/powerpoints/mak2008/COSPAR_MKopacz_July2008.ppt<br />
<br />
* Singh, K., P. Eller, A. Sandu, D. K. Henze, K. Bowman, M. Kopacz, and M. Lee (2009), Towards the construction of a standard geos-chem adjoint model, ACM High Performance Computing Conference.<br />
<br />
* Kopacz, M., Mauzerall, D.L., Leibensperger, E.M., Wang, J., Henze, D.K., Singh, K., Shim, C. Identifying the origin and estimating the radiative forcing of BC in the Himalayas: an analysis using the global GEOS-Chem adjoint model, European Geophysical Union meeting, Vienna, May 4, 2010.<br />
<br />
* Kopacz, M., Jacob, D.J., Fisher, J.A., Logan, J.A., Zhang, L., Megretskaia, I.A., Yantosca, R.M., Singh, K., Henze, D.K., Burrows, J.P., Buchwitz, M., Khlystova, I., McMillan, W.W., Gille, J.C., Edwards, D.P., Eldering, A., Thouret, V., Nedelec, P. Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), European Geophysical Union meeting, Vienna, May 7, 2010.<br />
<br />
* Tang, J., Zhuang, Q. and Xiong, X. (2010), 4D-Var inversion of atmospheric methane fluxes by assimilating SCIAMACHY and AIRS satellite retrievals, AGU, Dec. 18, 2010, http:/web.ics.purdue.edu/~tang16/agu2010_tang.ppt<br />
<br />
* Bousserez, N., R. V. Martin, K. W. Bowman, D.K. Henze, M. Kopacz, K. Singh, C. Shim, C. Wespes, Improving the lightning NOx source using satellite observations: a 4D-var analysis approach, AGU, Dec., 2010, http://myweb.dal.ca/nc689777/AGU_liNOx_poster_final.pdf</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint&diff=13626GEOS-Chem Adjoint2013-07-03T23:08:39Z<p>Hyung-Min Lee: /* Crediting GEOS-Chem adjoint developers */</p>
<hr />
<div><big><big><strong>Adjoint and Data Assimilation Working Group</strong></big></big><br />
<br />
== Contact information ==<br />
<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|-<br />
|'''Adjoint Working Group Co-Chairs'''<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman] and [mailto:dbj@atmosp.physics.utoronto.ca Dylan Jones]<br />
|-<br />
|-<br />
|'''Adjoint Model Scientist'''<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|'''GC adjoint support team'''<br />
|[mailto:yanko.davila@colorado.edu Yanko Davila], [mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez]<br />
|-<br />
|'''Adjoint Working Group email list'''<br />
|<tt>geos-chem-adjoint@seas.harvard.edu</tt><br />
|-<br />
|'''To subscribe to email list'''<br />
|Send email to [mailto:geos-chem-adjoint-join@seas.harvard.edu <tt>geos-chem-adjoint-join@seas.harvard.edu</tt>]<br />
|-<br />
|'''To unsubscribe from email list'''<br />
|Send email to [mailto:geos-chem-adjoint-leave@seas.harvard.edu <tt>geos-chem-adjoint-leave@seas.harvard.edu</tt>]<br />
|}<br />
<br />
== Historical Development ==<br />
Original work on the adjoint of GEOS-Chem v6 began in 2003, focusing on the adjoint of the offline aerosol simulation. By 2005, the adjoint was expanded to include a tagged CO simulation and a full chemistry simulation; an adjoint of GEOS-Chem v7 was also developed in the following years. Each of these branches of the adjoint code were been constructed in a hybrid fashion using a combination of automatic differentiation software ([http://www.autodiff.com/tamc TAMC], [http://people.cs.vt.edu/~asandu/Software/Kpp/ KPP]) and manual coding of both discrete and continuous adjoints. They shared many common elements yet had unique features for different applications. <br />
<br />
During the summer of 2009, the existing branches were merged and updated to bring the adjoint into alignment with the latest release of GEOS-Chem, v8-02-01. This merged adjoint model is now the standard adjoint code into which all further development efforts will be placed.<br />
<br />
== Forward Model Code ==<br />
<br />
The forward model on which the adjoint is based originally corresponded to GEOS-Chem v8-02-01. It was subsequently updated as follows:<br />
<br />
* KPP solver for gas-phase chemistry (as in GCv8-02-03)<br />
* Implement Bond 2007 BC/OC emissions (as in GCv8-02-02)<br />
* Apply bug fixes from GCv8-02-02 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-02#Previous_issues_now_resolved_in_v8-02-02 here]<br />
* Apply bug fixes from GCv8-02-03 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-03#Previous_issues_now_resolved_in_v8-02-03 here]<br />
* Apply bug fixes from GCv8-02-04 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-04#Previous_issues_now_resolved_in_v8-02-04 here]<br />
<br />
All bug fixes and model updates were previous listed at the top of inverse_driver.f. We have now switched to documenting the code development cycle here in the wiki, see the following section. <br />
<br />
== Code Versions, Bug Fixes and Developments == <br />
<br />
=== Current GEOS-Chem adjoint version released ===<br />
*[[GEOS-Chem_Adjoint_v34]] (You will download this version when you check out.)<br />
<br />
=== Previous GEOS-Chem adjoint versions released ===<br />
*[[GEOS-Chem_Adjoint_v33]] <br />
*[[GEOS-Chem_Adjoint_v32]]<br />
*[[GEOS-Chem_Adjoint_v31]]<br />
<br />
== Summary of Main Adjoint Code Supported Features ==<br />
<br />
=== Features === <br />
* Meteorological fields<br />
** GEOS-3 <i>needs testing</i><br />
** GEOS-4<br />
** GEOS-5 <br />
* model resolution<br />
** 4 x 5<br />
** 2 x 2.5 <br />
** Nested Asia and NA<br />
* Forward model processes<br />
** convection<br />
** advection<br />
** PBL mixing<br />
** dry deposition<br />
** wet deposition<br />
** strat / trop exchange with LINOZ and new GMI strat chem (v9-01-03)<br />
** NOy up fluxes (now replaced with new GMI strat chem)<br />
** aerosols <br />
*** inorganic aerosol thermodynamics with RPMARES<br />
*** inorganic aerosol thermodynamics with ISORROPIA <i> in progress </i><br />
*** sulfate chemistry <br />
*** BC<br />
*** SOA, Dust, sea salt <i> needs doing </i><br />
*** aerosol surface area feedbacks <i> needs updating </i><br />
*** aerosol optical feedbacks <i> needs doing </i><br />
** emissions<br />
*** all standard emissions included<br />
* Simulation modes<br />
** full chemistry<br />
** tagged CO<br />
** tagged Ox <br />
** CH4<br />
** offline aerosols (for BC and dust only)<br />
** CO2<br />
* Observational Operators<br />
** MOPITT CO column<br />
** SCIAMACHY CO column<br />
** AIRS CO column<br />
** IMPROVE BC<br />
** CASTNet (NH4+) <i> needs updating </i> <br />
** GOME / SCIAMACHY NO2 column <i> needs updating </i> <br />
*** using KNMI retrieval (Henze)<br />
*** using Dalhousie retrieval (Shim)<br />
** SCIAMACHY/OMI NO2<br />
*** using Dalhousie retrieval (Bousserez, Padmanabhan)<br />
** TES NH3 <br />
** TES O3 <br />
** GOSAT CO2 <br />
** MLS O3 and TES CO2 <i> in progress </i><br />
* Control parameters<br />
** Initial Conditions scaling factors (linear or log)<br />
** Emissions scaling factors (linear or log) <br />
*** NH3, primary BC/OC, SO2: anthropogenic, natural, bioburn, biomass, ship<br />
*** NOx: soil, aircraft, anthropogenic, biofuel, bioburn<br />
*** Lightning NOx: injection height, yield <i> in progress </i><br />
*** all other gas-phase tracers: anthropogenic, biofuel, bioburn <br />
* Adjoint sensitivities<br />
** w.r.t. all implemented control parameters<br />
** w.r.t Reaction Rate Parameters <br />
** w.r.t all emissions <br />
** of AQ attainment metrics <i> needs updating </i><br />
** of spatiotemporally averaged species concentrations (e.g., arctic O3) <br />
* Other<br />
** Inverse Hessian approximation <br />
** off-diagonal covariance matrices <i> needs updating </i><br />
** 3D-Var <i> needs updating </i><br />
<br />
Features may be qualified as: <br />
* <i>needs testing</i>: an implemented feature that we haven't fully used yet<br />
* <i>needs updating</i>: a feature developed with a previous branch that has yet to be updated to GEOS-Chem v8 and the merged adjoint<br />
* <i>needs doing</i>: a feature nobody has tackled the adjoint of yet<br />
* <i>in progress</i>: a feature currently under development <br />
* <i>in pipeline</i>: a feature which has been submitted and awaiting integration into the CVS repository<br />
<br />
=== Primary code developers ===<br />
Monika Kopacz, Kumaresh Singh, Changsub Shim, Daven Henze<br />
<br />
=== Adjoint model lead scientist ===<br />
Daven Henze<br />
<br />
<br />
<br />
== Resources ==<br />
<br />
=== User's guide ===<br />
A user's guide v34 is available. [http://adjoint.colorado.edu/~yanko/gcadj_std/GC_adj_man.pdf User's Guide v34] <br />
<br />
Previous version v32-v33 available at [http://adjoint.colorado.edu/~daven/gcadj_std/GC_adj_man.pdf User's Guide v32]<br />
<br />
=== Code flowchart ===<br />
Meemong Lee has created a detailed flowchart of the inverse model code structure. http://adjoint.colorado.edu/~daven/gcadj_std/flowchart.pdf<br />
<br />
=== Plotting tools ===<br />
<br />
Some IDL and MATLAB routines for plotting benchmark results. http://adjoint.colorado.edu/~daven/gcadj_std/tools.tar.gz<br />
<br />
=== Background papers and presentations ===<br />
<br />
Several articles and presentations (including a GC adjoint modeling clinic overview from IGC5) providing background information about adjoints. http://adjoint.colorado.edu/~daven/gcadj_std/adj_articles.tar.gz<br />
<br />
== Distribution and Use ==<br />
Code for the adjoint is distributed through GITLAB, a web interface connected to a GIT server located at adjoint.colorado.edu. You can access GITLAB at http://adjoint.colorado.edu:8080 after your account is created. Here is our [[Quick Start Guide.]] <br />
<br />
Even if your office mate has a copy of the code, the best way to obtain the model is to get an account for yourself and download a version from the repository. So please do not copy code directly from others or pass the code along to third parties. This vastly helps with tracking developments and keeping up with model updates. <br />
<br />
Use of the adjoint model code follows standard practice for GEOS-Chem. It is expected that any developments that come of individual applications based on this community model will eventually be given back to the community by incorporation of new developments into the standard adjoint code. New development should be submitted to Daven Henze for inclusion in the standard adjoint model code.<br />
<br />
Using GIT gives the users the ability to change the code and commit their changes without affecting the main repository hosted at adjoint.colorado.edu. Users can work with their tuned versions of the code and even create their own tags because GIT acts as a local repository. When ready to submit your update to the comunity just send Daven Henze your [[Using_Git_with_GEOS-Chem#Sharing_your_revisions_with_others_.28and_vice_versa.29|patch]] and we'll take care of the rest.<br />
<br />
=== Quick guide to GIT === <br />
As of version 34 we started using git versioning system. We recommend first taking a look at GIT manual to get a general feel for how this tools works (e.g., [http://git-scm.com/documentation GIT Documentation] or [http://www.kernel.org/pub/software/scm/git/docs/git.html GIT Manual Page]). <br />
<br />
<br />
Useful GIT commands: <br />
<br />
Initial download:<br />
git clone ssh://git@adjoint.colorado.edu:2222/gcadj_std.git<br />
<br />
Status of project vs the current repository:<br />
git status<br />
<br />
Check difference of files (differences have colors for easy reading)<br />
git diff --word-diff=color <wildcard> [<wildcard>] <path>/foo_mod.f<br />
<br />
Checkout specific version<br />
git checkout <wildcard><br />
<br />
Replacing a file with the newest version from the repository:<br />
git checkout origin/master -- <path>/foo_mod.f<br />
<br />
Merging changes in a file: ([http://www.kernel.org/pub/software/scm/git/docs/git-merge.html Reference])<br />
git merge -m <wildcard> <br />
<br />
Comitting<br />
git commit -a<br />
<br />
Tagging a version<br />
git tag -a TAGNAME<br />
<br />
Deleting a tag<br />
git tag -d TAGNAME<br />
git push origin :refs/tags/TAGNAME<br />
<br />
Saving changes to repository<br />
git push<br />
<br />
Saving tags to repository<br />
git push --tags<br />
<br />
List the history of a file:<br />
git log -- <path>/foo_mod.f<br />
<br />
Add a file to the repository<br />
git add <file_name><br />
<br />
Delete a file from the repository<br />
git rm <file_name><br />
<br />
Determine current version<br />
git show HEAD [ | grep commit]<br />
<br />
Download remote changes, rewind your local branch, then replays all your changes over the top of your current branch one by one, until you’re all up to date.<br />
git pull --rebase<br />
<br />
There are several wildcards that you can use on git for example:<br/><br />
''"origin/master"'' - Latest version on the repository <br/><br />
''"HEAD"'' - Latest version as of your last download <br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/72773d7b1a7bd757957d2c186acb3dbb107d5323 v33i]"'' - Specific TAG, find all tag names on GitLab<br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/32d5c926ef529c1dfc640c9dec4a925b24db575c 32d5c926e]"'' - Specific COMMIT, find all commit numbers on GitLab <br/><br />
<br />
[[Using_Git_with_GEOS-Chem|Here is the foward model documentation of git.]]<br />
<br />
=== Backward compatibility (CVS) === <br />
<br />
For people using old version of the code we still have active our CVS repository, but note that the latest version on CVS is v33i-patch2. Here you can find our [[Quick Guide to CVS]]<br />
<br />
== Crediting GEOS-Chem adjoint developers ==<br />
<br />
We aim to make distribution of adjoint model code as immediate as possible. A consequence is that many features may not yet be publicly documented. Therefore, giving code developers due credit is of utmost importance. <br />
<br />
Authors of new additions to the standard code should be offered co-authorship on the first round of presentations and publications to come of their development. Features currently falling in this category and their developers are:<br />
* (v32) MOPITT v5 CO observation operator. Developer: Zhe Jiang, University of Toronto.<br />
* (v32) Dust adjoint. Developer: Xiaoguang (Richard) Xu, University of Nebraska Lincoln.<br />
* (v32) Black carbon offline aerosol adjoint. Developer: Yuhao Mao (UCLA).<br />
* (v32) Nested full chemistry adjoint. Developers: Zhe Jiang, University of Toronto; Daven Henze, CU Boulder.<br />
* (v32) Stratospheric production / loss rate sensitivities. Developer: Hyung-Min Lee, CU Boulder.<br />
* (v32) CH4 adjoint. Developer: Kevin Wecht, Harvard University<br />
* (v29) LIDORT. Developer: Daven Henze, University of Colorado Boulder. Collaborator: Rob Spurr. <br />
* (v28) CO2 adjoint. Developer: Daven Henze, University of Colorado Boulder. Collaborators: Ray Nassar, Kevin Bowman, Dylan Jones.<br />
<br />
Citation of the appropriate [[GEOS-Chem_Adjoint#Journal_Articles|journal articles]] for mature developments is also encouraged, as well as considering aspects of [http://acmg.seas.harvard.edu/geos/geos_credit.html co-authorship for the forward model]. <br />
<br />
Overall, if you have any questions about authorship, even for a conference presentation, please contact Daven Henze.<br />
<br />
== Current GEOS-Chem Adjoint Research Projects (please add yours!) ==<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|- bgcolor="#cccccc"<br />
!User Group <br />
!Description <br />
!Contact Person<br />
|-<br />
|CU Boulder<br />
|Aerosol precursors, CO2, O3; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|CU Boulder<br />
|Inverse modeling/optimization; CO2 fluxes inversion (CMS project); general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Nicolas Bousserez]<br />
|-<br />
|CU Boulder<br />
|Sensitivity of nitrate deposition over Antarctica including stratospheric tracers<br />
|[mailto:hyungmin.lee@colorado.edu Hyung-Min Lee]<br />
|-<br />
|Harvard<br />
|Methane<br />
|Kevin Wecht, wecht [at] fas.harvard.edu<br />
|-<br />
|Harvard<br />
|Smoke Emissions in SE Asia<br />
|Patrick Kim, kim68 [at] fas.harvard.edu<br />
|-<br />
|Harvard<br />
|Methane from GOSAT<br />
|Alex Turner, aturner [at] fas.harvard.edu<br />
|-<br />
|Purdue University<br />
|Methane (SICAMACHY, AIRS and IASI)<br />
|[mailto:tang16@purdue.edu Jinyun Tang]<br />
|-<br />
|MIT<br />
|Aircraft emissions<br />
|[mailto:jaminkoo@mit.edu Jamin Koo]<br />
|-<br />
|MIT<br />
|Air quality, aircraft emissions and sensitivities<br />
|[mailto:bvconsta@mit.edu Bogdan V. Constantin]<br />
|-<br />
|Princeton<br />
|BC sensitivities, general adjoint code development<br />
|Monika Kopacz, mkopacz [at] princeton.edu<br />
|-<br />
|Dalhousie University<br />
|Lightning NOx emissions and impact on tropical ozone using the adjoint<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez (now at CU-Boulder)]<br />
|-<br />
|Dalhousie University<br />
|Surface NOx emissions inversion using SCIAMACHY/OMI NO2 measurements<br />
|Akhila Padmanabhan akhila [at] dal.ca; Nicolas Bousserez [mailto:Nicolas.Bousserez@colorado.edu] (now at CU-Boulder)<br />
|-<br />
|JPL<br />
|Microwave Limb Sounder (MLS) Ozone assimilation<br />
|[mailto:meemong.lee@jpl.nasa.gov Meemong Lee]<br />
|-<br />
|JPL <br />
| TES ozone assimilation/attribution of ozone radiative forcing<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|University of Edinburgh<br />
|Quantifying the impact of boreal forest fires on tropospheric oxidants over the Atlantic<br />
|[mailto:mark.parrington@ed.ac.uk Mark Parrington]<br />
|-<br />
|US EPA<br />
|Integration with economic models for future emission inventory scenario development<br />
|[mailto:akhtar.farhan@epa.gov Farhan Akhtar]<br />
|-<br />
|Peking University<br />
|Satellite constraints on VOC emissions <br />
|[mailto:tmfu@pku.edu.cn May Fu]<br />
|-<br />
|IAP.CAS<br />
|CO2 assimilation<br />
|[mailto:czh@mail.iap.ac.cn Chen]<br />
|-<br />
|Purdue University<br />
|Feedback between terrestrial ecosystem processes and atmospheric co2 signals<br />
|[mailto:zhuq@purdue.edu Qing Zhu]<br />
|-<br />
|Purdue University<br />
|Feedback between aquatic ecosystem processes and atmospheric CH4 signals<br />
|[mailto:tan80@purdue.edu Zeli Tan]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and reactive nitrogen to precursor emissions<br />
|[mailto:twalker@atmosp.physics.utoronto.ca Thomas Walker]<br />
|-<br />
|University of Toronto<br />
|Adjoint analysis for carbon monoxide <br />
|[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and CO to precursor emissions<br />
|[mailto:cwhaley@atmosp.physics.utoronto.ca Cynthia Whaley]<br />
|-<br />
|Georgia Tech / US EPA<br />
|ISORROPIA adjoint development; NH3 assimilation; cloud droplet sensitivities<br />
|[mailto:scapps@gatech.edu Shannon Capps]<br />
|-<br />
|Peking University<br />
|Source attributions of tropospheric ozone over North China<br />
|[mailto:linjt@pku.edu.cn Jintai Lin]<br />
|-<br />
|University of Wollongong<br />
|Sensitivity of ozone and adjoint analysis of CO over Australasia.<br />
|[mailto:rb864@uowmail.edu.au Rebecca Buchholz]<br />
|-<br />
|University of Toronto<br />
|CO2 assimilation & transport model bias estimation<br />
|[mailto:mkeller@atmosp.physics.utoronto.ca Martin Keller]<br />
|-<br />
|University of Wisconsin<br />
|CO2 assimilation and forecast & Temperature profile retrieval <br />
|[mailto:wenguang.bai@ssec.wisc.edu Wenguang Bai]<br />
|-<br />
|Dalhousie University<br />
|Sensitivity of global PM2.5-induced mortality to emissions<br />
|[mailto:colin.lee@dal.ca Colin Lee]<br />
|-<br />
|University of Leicester (UK)<br />
|Top-down estimates of Amazon isoprene emissions<br />
|[mailto:mpb14@le.ac.uk Michael Barkley]<br />
|-<br />
|Anyang University<br />
|Aerosol emission modeling in East Asia <br />
|[mailto:koo@anyang.ac.kr Youn Seo Koo]<br />
|-<br />
|Nanjing University<br />
|Inverse modeling of terrestrial ecosystem carbon flux<br />
|[mailto:hengmao.wang@gmail.com Hengmao Wang]<br />
|-<br />
|Tsinghua University<br />
|Nested-gird simulations with the adjoint model<br />
|[mailto:nany12@mails.tsinghua.edu.cn Nan Yang]<br />
|-<br />
|University of Minnesota<br />
|Inverse modeling of VOC sources based on TES and IASI measurements<br />
||[mailto:kcw@umn.edu Kelley Wells], [mailto:dbm@umn.edu Dylan Millet]<br />
|-<br />
|University of Minnesota & CU Boulder<br />
|Inverse modeling of N2O sources<br />
|[mailto:dbm@umn.edu Dylan Millet], [mailto:Daven.Henze@Colorado.EDU Daven Henze]<br />
|-<br />
|Tsinghua University<br />
|Inverse modeling of anthropogenic emissions over East Asia<br />
|[mailto:qiangzhang@tsinghua.edu.cn Qiang Zhang]<br />
|-<br />
|UCLA<br />
|Constrain black carbon emission<br />
|[mailto:qiling@atmos.ucla.edu Ling Qi]<br />
|}<br />
<br />
== Publications ==<br />
<br />
=== Journal Articles ===<br />
<br />
* '''In press or sumbitted'''<br />
<br />
**Jiang, Z., D. B. A. Jones, H. M. Worden, M. N. Deeter, D. K. Henze, J. Worden, and K. W. Bowman, Quantifying the impact of model biases in convective transport on inferred CO source estimates using multi-spectral CO retrievals from MOPITT, JGR, in press.<br />
**Kharol, S., R. V. Martin, S. Philip, S. Vogel, D. K. Henze, D. Chen, Y. Wang, Q. Zhang, C. L. Heald, Persistent Sensitivity of Asian Aerosol to Emissions of Nitrogen Oxides, GRL, in press.<br />
**Koo, J., Q. Wang, D. K. Henze, I. A. Waitz, S.R.H. Barrett, Spatial sensitivities of human health risk to intercontinental and high-altitude pollution, Atmos. Environ, in press.<br />
**L. Zhu, D. K. Henze, K. E. Cady-Pereira, M. W. Shephard, M. Luo, R. W. Pinder, J. O. Bash, G. Jeong, Constraining U.S. ammonia emissions using TES remote sensing observations and the GEOS-Chem adjoint model, JGR, in press.<br />
**Paulot, F., D. J. Jacob and D. K. Henze, Sources and processes contributing to nitrogen deposition in biodiversity hotspots worldwide, submitted.<br />
**Singh, K., A. Sandu, Variational Chemical Data Assimilation with Approximate Adjoints, submitted. <br />
**Xu, X., J. Wang, D. K. Henze, W. Qu, M. Kopacz, Constraints on Aerosol Sources Using GEOS-Chem Adjoint and MODIS Radiances, and Evaluation with Multi-sensor (OMI, MISR) data, submitted.<br />
<br />
* '''2012'''<br />
**Bowman, K. W., and D. K. Henze, Attribution of direct ozone radiative forcing to spatially-resolved emissions, Geophys. Res. Lett., 39, L22704, doi:10.1029/2012GL053274. GRL Editor's highlight, Science Daily, KGNU radio interview.<br />
**Henze, D. K., D. T. Shindell, F. Akhtar, R. J. D. Spurr, R. W. Pinder, D. Loughlin, M. Kopacz, K. Singh, and C. Shim, Spatially refined aerosol direct radiative forcing efficiencies, Environ. Sci. Technol., 46, 9511 - 9518, dx.doi.org/10.1021/es301993s.<br />
**Karydis, V. A., S. L. Capps, R. H. Moore, A. Russell, D. K. Henze, and A. Nenes, Using a global aerosol model adjoint to unravel the footprint of spatially-distributed emissions on cloud droplet number and cloud albedo, Geophys. Res. Lett., 39, L24804, doi:10.1029/2012GL053346.<br />
**Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden, The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098, online.<br />
**Paulot, F., D. K. Henze, and P. O. Wennberg, Impact of the isoprene photochemical cascade on tropical ozone, Atmos. Chem. Phys., 12, 1307-1325, online.<br />
**Singh, K. and A. Sandu, 2012: Variational chemical data assimilation with approximate adjoints. Computers and Geosciences, 40, 10-18.<br />
**Turner, A., D. K. Henze, R. V. Martin, and A. Hakami, The spatial extent of source influences on modeled column concentrations of short-lived species, Geophys. Res. Lett., 39, L12806, doi:10.1029/2012GL051832.<br />
**Walker, T., D. B. A. Jones, M. Parrington, D. K. Henze, L. T. Murray, J. W. Bottenheim, K. Anlauf, J. R. Worden, K. W. Bowman, C. Shim, K. Singh, M. Kopacz, D. W. Tarasick, J. Davies, P. von der Gathen, and C. C. Carouge, Impacts of midlatitude precursor emissions and local photochemistry on ozone abundances in the Arctic, J. Geophys. Res.,117, D01305 doi:10.1029/2011JD016370.<br />
**Wang, J., X. Xu, D. K. Henze, Q. Ji, S.-C. Tsay, J. Huang, Top-Down Estimate of Dust Emissions through Integration of MODIS and MISR Aerosol Retrievals with the GEOS-Chem adjoint model, Geophys. Res. Lett., 39, L08802.<br />
**Wecht, K. J., D. J. Jacob, S. C. Wofsy, E. A. Kort, J. R. Worden, S. S. Kulawik, D. K. Henze, M. Kopacz, and V. H. Payne, Validation of TES methane with HIPPO aircraft observations: implications for inverse modeling of methane sources, Atmos. Chem. Phys., 12, 1823-1832, online.<br />
<br />
* '''2011'''<br />
**Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald, Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
**Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald (2011), Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
**Kopacz, M., D. L. Mauzerall, J. Wang, E. M. Leibensperger, D. K. Henze, and K. Singh, Origin and radiative forcing of black carbon transported to the Himalayas and Tibetan Plateau, Atmos. Chem. Phys., 11, 2837-2852 online.<br />
<br />
* '''2010'''<br />
**Kopacz, M., D. J. Jacob, J. A. Fisher, J. A. Logan, L. Zhang, I. A Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010), Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atoms. Chem. Phys., 10, 855-876, online.<br />
**Kopacz, M., D.J. Jacob, J.A. Fisher, J. A. Logan, L. Zhang, I. A. Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010): Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atmos. Chem. Phys., 10, 855-876. http://www.atmos-chem-phys.net/10/855/2010/acp-10-855-2010.<br />
**Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden (2012), The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098<br />
**Singh, K., Jardak, M., Sandu, A., Bowman, K., Lee, M., and Jones, D. (2010): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. Discuss., 3, 1783-1827, doi:10.5194/gmdd-3-1783-2010. http://www.geosci-model-dev-discuss.net/3/1783/2010/gmdd-3-1783-2010.html<br />
<br />
* '''2009'''<br />
**Eller, P., K. Singh, A. Sandu, K. Bowman, D. K. Henze, and M. Lee (2009), Implementation and evaluation of an array of chemical solvers in a global chemical transport model, Geosci. Mod. Devel., 2, 185-207. online<br />
**Henze, D. K., J. H. Seinfeld and D. T. Shindell, (2009), Inverse modeling and mapping U.S. air quality influences of inorganic PM2.5 precursor emissions with the adjoint of GEOS-Chem, Atoms. Chem. Phys., 9, 5877-5903, online.<br />
**Kopacz, M., D. J. Jacob, D. K. Henze, C. L. Heald, D. G. Streets, and Q. Zhang (2009), A comparison of analytical and adjoint Bayesian inversion methods for constraining Asian sources of CO using satellite (MOPITT) measurements of CO columns, J. Geophys. Res., 114, D04305, doi:10.1029/2007JD009264.<br />
**Pye, H. O. T., H. Liao, S. Wu, L. J. Mickely, D. J. Jacob, D. K. Henze, and J. H. Seinfeld (2009), Effect of changes in climate and emissions on future sulfate-nitrate-ammonium aerosol levels in the United States, J. Geophys. Res., 114, D01205, doi:10.1029/2008JD010701.<br />
**Zhang, L., D. J. Jacob, M. Kopacz, D. K. Henze, K. Singh, and D. A. Jaffe (2009), Intercontinental source attribution of ozone pollution at western U.S. sites using an adjoint method, Geophys. Res. Lett., 36, L11810, doi:10.1029/2009GL037950.<br />
<br />
* '''2007'''<br />
**Henze, D. K., A. Hakami and J. H. Seinfeld (2007), Development of the adjoint of GEOS-Chem, Atmos. Chem. Phys., 7, 2413-2433.<br />
<br />
=== Conference proceedings === <br />
* Adjoint inversion of CO sources using combined MOPITT, SCIAMACHY and AIRS CO columns, presented by Monika Kopacz at the COSPAR Scientific Assembly, Montreal, July 18, 2008. http://acmg.seas.harvard.edu/presentations/powerpoints/mak2008/COSPAR_MKopacz_July2008.ppt<br />
<br />
* Singh, K., P. Eller, A. Sandu, D. K. Henze, K. Bowman, M. Kopacz, and M. Lee (2009), Towards the construction of a standard geos-chem adjoint model, ACM High Performance Computing Conference.<br />
<br />
* Kopacz, M., Mauzerall, D.L., Leibensperger, E.M., Wang, J., Henze, D.K., Singh, K., Shim, C. Identifying the origin and estimating the radiative forcing of BC in the Himalayas: an analysis using the global GEOS-Chem adjoint model, European Geophysical Union meeting, Vienna, May 4, 2010.<br />
<br />
* Kopacz, M., Jacob, D.J., Fisher, J.A., Logan, J.A., Zhang, L., Megretskaia, I.A., Yantosca, R.M., Singh, K., Henze, D.K., Burrows, J.P., Buchwitz, M., Khlystova, I., McMillan, W.W., Gille, J.C., Edwards, D.P., Eldering, A., Thouret, V., Nedelec, P. Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), European Geophysical Union meeting, Vienna, May 7, 2010.<br />
<br />
* Tang, J., Zhuang, Q. and Xiong, X. (2010), 4D-Var inversion of atmospheric methane fluxes by assimilating SCIAMACHY and AIRS satellite retrievals, AGU, Dec. 18, 2010, http:/web.ics.purdue.edu/~tang16/agu2010_tang.ppt<br />
<br />
* Bousserez, N., R. V. Martin, K. W. Bowman, D.K. Henze, M. Kopacz, K. Singh, C. Shim, C. Wespes, Improving the lightning NOx source using satellite observations: a 4D-var analysis approach, AGU, Dec., 2010, http://myweb.dal.ca/nc689777/AGU_liNOx_poster_final.pdf</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v32&diff=13625GEOS-Chem Adjoint v322013-07-03T23:07:40Z<p>Hyung-Min Lee: /* Add strat production and loss rates as active control parameters (adj32_025) */</p>
<hr />
<div>== Overview ==<br />
Under development. Expected release date 02/15/2012<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v31]]<br />
*Next version: [[GEOS-Chem_Adjoint_v33]]<br />
<br />
== What's new in this version ==<br />
<br />
=== Updates to forward model ===<br />
<br />
----<br />
<br />
==== Update tagged CO simulation (adj32_017) ====<br />
Several updates from Zhe Jiang have been implemented for the tagged-CO implementation. See also changes to the adjoint model, [[GEOS-Chem_Adjoint_v32#Update_tagged_CO_simulation_.28adj32_017.29]]<br />
<br />
===== Update to tagged_co_mod.f =====<br />
<br />
<tt>tagged_co_mod.f</tt> has been updated to distribute emissions through the boundary layer and removing obsolete COSF, TIME_STEPS, GET_CO_CH4, SET_CO_CH4, and GET_OPT_CO.<br />
<br />
Also, F_OF_PBL and ED_CO were declared thread PRIVATE in several parallel do loops. <br />
<br />
<br />
===== Updates to <tt>time_mod.f</tt> for tagged-CO =====<br />
'''''[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]''''' wrote<br />
<br />
:I currently found it perhaps can only affect tagged-CO inversion. Thus, it can not be said as a BUG.<br />
<br />
:For example, if we do our inversion for June 1-15, when the backward running, because June 15 is not the last day of the calendar month, these emissions will not be read. However, we don't need to read them since the model will use the value in the forward run. For the tagged-CO backward run, the CH4 concentration will only be read if June 15 is the last day of the calendar year. Thus, in the original code, the CH4 concentration, and thus gradient on CH4 oxidation, is ZERO.<br />
<br />
IF ( ITS_A_NEW_YEAR() ) THEN<br />
CALL GET_GLOBAL_CH4( GET_YEAR(), .TRUE., A3090S, A0030S, A0030N, A3090N )<br />
ENDIF<br />
<br />
:In my previous code, I saved the CH4 concentration in the forward run and then read them in the backward run. After I modified time_mod.f, I have removed the code for the CH4 saving/reading. Anyway, the effect is trivial. I can solve the problem in the tagged_CO inversion by modifying the previous code.<br />
<br />
To address these issues, subroutines ITS_A_NEW_MONTH and ITS_A_NEW_YEAR have been updated in <tt>time_mod.f</tt>. <br />
<br />
<br />
--[[User:Daven|Daven]] 18:18, 8 January 2012 (EST)<br />
<br />
==== Update several files to v9 code to support full-chem nested simulation (adj32_015) ====<br />
Updates to the following files were made by Zhe Jiang to support the NESTED_NA and NESTED_SD simulation:<br />
<br />
===== Update DEAD dust emissions for nested simulations and MERRA (v9-01-01) =====<br />
When the forward model was updated to support MERRA, the dust emissions got retuned [[MERRA_implementation_details#Dust_emissions]]. But also the emissions for GEOS-5 were affected, and support was added for nested grid. <br />
<br />
The <tt>dust_dead_mod.f</tt> file has been updated to match that in v9-01-01 (dust_dead_mod.f,v 1.3 2010/02/02 16:57:53 bmy). Common blocks CMN_SIZE and CMN_GCTM are reinstated for backwards compatibility.<br />
<br />
===== Updated TOMS O3 =====<br />
<tt>toms_mod.f</tt> was updated to v9-01-02. This uses the newer TOMS inventory TOMS_200906 and allows for rescaling of years up to 2008 (previous version stopped at 2007). <br />
<br />
===== cac_mod.f =====<br />
<tt>cac_mod.f</tt> was updated to v9-01-02.<br />
<br />
===== emissions_mod.f =====<br />
<tt>emissions_mod.f</tt> was updated to support GRID05x0666 for CAC, NEI05 and EMEP emissions following v9-01-02.<br />
<br />
===== file_mod.f =====<br />
<tt>file_mod.f</tt> was updated to include IU_BC_05x06.<br />
<br />
===== CMN_SIZE =====<br />
<tt>CMN_SIZE</tt> was updated to include size of NA and SD nested domain.<br />
<br />
===== commsoil.h =====<br />
<tt>commsoil.h</tt> was updated to include size of NA and SD nested domain.<br />
<br />
===== define.h =====<br />
<tt>define.h</tt> was updated to include the NESTED_SD option.<br />
<br />
===== tpcore_bc_mod.f =====<br />
<tt>tpcore_bc_mod.f</tt> has been updated to include <br />
*OPEN_BC_FILE_05x0666<br />
*SAVE_GLOBAL_TPCORE_BC_05x0666<br />
*DO_WINDOW_TPCORE_BC_05x0666<br />
*CLEAN_WINDOW_TPCORE_BC_05x0666<br />
*READ_WINDOW_TPCORE_BC_05x0666<br />
*ITS_TIME_FOR_BC_05x0666INIT_TPCORE_BC_05x0666<br />
and the BC_05x06 array.<br />
<br />
===== regrid_1x1_mod.f =====<br />
<tt>regrid_1x1_mod</tt> has been updated to support the NESTED_SD domain.<br />
<br />
===== Update lightning_nox_mod.f =====<br />
See [[GEOS-Chem_Adjoint_v32#Updated_lightning_parameterization_and_fix_for_cloud-top-height_algorithm_.28v9-01-01.29]]<br />
<br />
===== acetone_mod.f =====<br />
<tt>acetone_mod.f</tt> has been updated to support the nested domains. It now matches v9-01-02, with the additional update for the adjoint that we do not multiply and divide by ACETONE, thus making OCEAN_SINK_ACET self-adjoint.<br />
<br />
===== input_mod.f =====<br />
<tt>input_mod.f</tt> has been updated in several places to support nested domains. <br />
<br />
===== transport_mod.f =====<br />
<tt>transport_mod.f </tt> has been updated to support the NESTED_SD domain.<br />
<br />
===== co2_mod.f =====<br />
<tt>co2_mod.f</tt> has been updated to support nested domains.<br />
<br />
<br />
--[[User:Daven|Daven]] 15:00, 18 January 2012 (EST)<br />
<br />
<br />
===== nei05 and input_mod.f =====<br />
<tt>nei2005_anthro_mod.f</tt> and <tt>input_mod.f</tt> have been updates, see [[GEOS-Chem_Adjoint_v32#Fixes_for_minor_issues_affecting_nested-grid_simulations_.28v9_01_02e.29]]<br />
<br />
==== Update CH4 simulation (adj32_023) ====<br />
[[GEOS-Chem_Adjoint_v32#Include_adjoint_of_CH4_simulation_.28adj32_023.29]]<br />
<br />
==== Update stratospheric chemistry and prod/loss rates (adj32_025) ====<br />
[[GEOS-Chem_Adjoint_v32#Add_strat_production_and_loss_rates_as_active_control_parameters_.28adj32_025.29]]<br />
<br />
=== Bug fixes in forward model ===<br />
<br />
----<br />
<br />
==== Prevent overflow / underflow in rpmares (adj32_001) ====<br />
<br />
===== Ionic strength below zero =====<br />
On prospero, we often get the message: <tt>Ionic strength below zero...negative concentrations</tt>.<br />
<br />
Despite the setting of negative roots to zero, sometimes the code still ends up with negative CRUTES. Trap for it here to prevent mysterious crashing later <br />
<br />
! Trap for negative CRUTES (dkh, 01/06/12, adj32_001)) <br />
IF ( CRUTES(1) < 0d0 ) THEN<br />
IF ( CRUTES(1) .ne. ABS(CRUTES(1) ) ) CRUTES(1)=1d9<br />
IF ( CRUTES(2) .ne. ABS(CRUTES(2) ) ) CRUTES(2)=1d9<br />
IF ( CRUTES(3) .ne. ABS(CRUTES(3) ) ) CRUTES(3)=1d9<br />
CRUTES(1) = MIN( CRUTES(1), CRUTES(2), CRUTES(3) )<br />
IF ( CRUTES(1) < 0d0 ) THEN<br />
print*, ' CRUTES = ', CRUTES , NR<br />
CALL ERROR_STOP( ' checking for neg val failed',<br />
& ' rpmares_mod.f ' )<br />
ENDIF<br />
ENDIF<br />
<br />
In addition, reinstate the following<br />
<br />
! reinstate (dkh, 02/12/12, adj32_001) <br />
IF (HPLUS <= 0d0) GOTO 100<br />
--[[User:Daven|Daven]] 14:58, 6 January 2012 (EST)<br />
<br />
===== RPMARES: not safe to divide...exit =====<br />
When running with LADJ = F, this error would arise. The reason is not entirely clear, but with the exact same inputs, it sometimes goes away by adding a print statement here or there, so is likely owing to overflow / underflow within RPMARES somewhere. The RPMARES_FORADJ routine is not subject to this instability, so now we move the LADJ flag within RPMARES_FORADJ and always use this routine. <br />
<br />
--[[User:Daven|Daven]] 12:32, 12 February 2012 (EST)<br />
<br />
<br />
===== Compiler level fix =====<br />
'''''[mailto:fabienpaulot@gmail.com Fabien Paulot]''''' wrote<br />
<br />
:Hi Daven,<br />
<br />
:I am sorry to bug you again about that issue but this time I may have a solution. It seems the problem with RPMARES is related with the compilation option.<br />
<br />
:I don't get the error ("not safe to divide") when I use the following option<br />
<br />
:FFLAGS = -cpp -w -O2 -auto -noalign -convert big_endian -fp-model precise and remove the -mp option from the F90 compiler options.<br />
<br />
:(Note switching to O2 is not the reason why it's working but Bob suggested not to use O3).<br />
<br />
<br />
We now implement the -fp-model precise flag and do not use the -mp flag. The latter was originally implemented to allow KPP code to run on 32-bit systems. <br />
<br />
This compiler level fix does address the two issues with rpmares noted above on prospero. But we will leave those patches in the code for now, as they don't hurt, and we'll see how this plays out on other platforms.<br />
<br />
--[[User:Daven|Daven]] 00:34, 1 March 2012 (EST)<br />
<br />
==== Updated lightning parameterization and fix for cloud-top-height algorithm (v9-01-01) ====<br />
Updates to <tt>a6_read_mod.f</tt> and <tt>lightning_nox_mod.f</tt>. Details in Lee's description: http://people.seas.harvard.edu/~ltmurray/LNOx.v9-01-01.Release.Notes.pd<br />
<br />
--[[User:Daven|Daven]] 15:18, 6 January 2012 (EST)<br />
<br />
==== Fix for GEIA emissions scaling factor over Botswana (v9-01-02h) ====<br />
Updates to <tt>scale_anthro_mod.f</tt>. See [[GEOS-Chem_v9-01-02#Fix_for_GEIA_emissions_scaling_factor_over_Botswana]]<br />
<br />
--[[User:Daven|Daven]] 16:16, 6 January 2012 (EST)<br />
<br />
==== TINY parameter in convection_mod.f (v8-03-02) ====<br />
Update to <tt>convection_mod.f</tt> [[GEOS-Chem_v8-03-02#TINY_parameter_in_convection_mod.f]]<br />
<br />
--[[User:Daven|Daven]] 16:16, 6 January 2012 (EST)<br />
<br />
==== Updated to REGRID_05x0666_NESTED for non-standard domains (v?-?-?) ====<br />
'''''[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]''''' wrote<br />
<br />
:I revised the SUBROUTINE REGRID_05x0666_NESTED.<br />
<br />
:The problem in the original code is that the Longitude of LL (Left-Low of nested window) has to be even number and the latitude of LL has to be integer number. It has no influence on our standard NA/CH/EU nested setting.<br />
<br />
:I think it would be safer to fix this problem because we perhaps need to define new nested window in future.<br />
<br />
This subroutine has now been replaced with Zhe's update. Not sure if there is a corresponding update in the v9 forward model yet. <br />
<br />
--[[User:Daven|Daven]] 19:06, 6 January 2012 (EST)<br />
<br />
==== Fix variable declarations in UPBDFLX_NOY (adj32_008) ====<br />
'''''[mailto:twalker@atmosp.physics.utoronto.ca Thomas Walker]''''' noted that having DTDYN, AIRDENS, PNOY declared as REAL*4 can lead to overflow in the adjoint model. The <tt>upbdflx_mod.f</tt> has now been updated to use REAL*8 instead. <br />
<br />
--[[User:Daven|Daven]] 18:00, 8 January 2012 (EST)<br />
<br />
==== NEI05 ID names for non-full chemistry simulation (adj32_015) ====<br />
<br />
Explanation from Zhe...<br />
<br />
IF ( .not. ITS_A_FULLCHEM_SIM() ) THEN !Zhe (16/3/2011)<br />
! Zhe (16/3/2011), (dkh, <br />
SPECIES_ID_SAVE = (/ IDTNOX, IDTCO, IDTSO2, IDTSO4, IDTNH3,<br />
$ IDTACET, IDTALK4, IDTC2H6, IDTC3H8,<br />
$ IDTOCPI, IDTBCPI,<br />
$ IDTALD2, IDTCH2O, IDTPRPE, IDTMEK<br />
$ /) <br />
<br />
IDTNOX = 1<br />
IDTCO = 4<br />
IDTSO2 = 26<br />
IDTSO4 = 27<br />
IDTNH3 = 30<br />
IDTACET = 9<br />
IDTALK4 = 5<br />
IDTC2H6 = 21<br />
IDTC3H8 = 19<br />
IDTOCPI = 35<br />
IDTBCPI = 34<br />
IDTALD2 = 11<br />
IDTCH2O = 20<br />
IDTPRPE = 18<br />
IDTMEK = 10<br />
<br />
ENDIF<br />
<br />
--[[User:Daven|Daven]] 19:46, 16 January 2012 (EST)<br />
<br />
==== Fixes for minor issues affecting nested-grid simulations (v9_01_02e) ====<br />
<br />
Updates to <tt>nei2005_anthro_mod.f</tt> and <tt>input_mod.f</tt>, see [[GEOS-Chem_v9-01-02#Fixes_for_minor_issues_affecting_nested-grid_simulations]]<br />
<br />
==== Fixes for tagged CO EMEP emissions (adj32_019) ====<br />
To allow for tagged-CO simulation to use the EMEP emissions, use hardwired tracer numbers in emep_mod.f This should eventually be migrated into the standard forward model code. <br />
<br />
--[[User:Daven|Daven]] 15:23, 9 February 2012 (EST)<br />
<br />
==== Fix for nested boundary condition (adj32_021) ====<br />
In <tt>tpcore_bc_mod.f</tt>, subroutine DO_WINDOW_TPCORE_BC_05x0666, the OpenMP statements were causing problems, so remove these for now. <br />
<br />
--[[User:Daven|Daven]] 15:23, 9 February 2012 (EST)<br />
<br />
==== Fix for nested boundary condition (adj32_022) ====<br />
Updated to<tt>transport_mod.f</tt> and <tt>diag3.f</tt>. This issue has been corrected in the standard forward model with the following note:<br />
<br />
! Note: the mass flux diagnostic arrays (MASSFLEW, MASSFLNS and MASSFLUP)<br />
! are incremented upside-down (level 1 = top of the atmosphere).<br />
! The levels order is reversed only when written out in diag3.f<br />
! (ccc, 3/8/10),<br />
<br />
--[[User:Daven|Daven]] 15:23, 9 February 2012 (EST)<br />
<br />
--[[User:Daven|Daven]] 19:51, 16 January 2012 (EST)<br />
<br />
=== Updates to adjoint model ===<br />
<br />
<br />
----<br />
<br />
==== Updates to aerosol direct radiative forcing adjoint ====<br />
Several updates in <tt>lidort_mod.f</tt> and <tt>mie_mod.f</tt>:<br />
*From Farhan Akhtar, now include primary OC in the LIDORT radiative forcing calculation. <br />
*Now switch back to using analytic derivatives from mie table rather than FD<br />
<br />
--[[User:Daven|Daven]] 18:50, 8 January 2012 (EST)<br />
<br />
==== Dust adjoint and MODIS AOD observation operator (adj32_011) ====<br />
From Richard Xiaoguang Xu at UNL, the dust adjoint includes sensitivity with respect to dust emissions from the DEAD module. Also included in this update is an observation operator for assimilating AOD observations from MODIS using the UNL retrieval. <br />
<br />
--[[User:Daven|Daven]] 00:23, 13 January 2012 (EST)<br />
<br />
==== Inverse Hessian approximation (adj32_012) ====<br />
The inverse Hessian approximation (e.g., Henze et al., 2009) using the DFP formula is now implemented with the LINVH flag in <tt>input.gcadj</tt>. Instructions for how to perform this calculation are included in the UPDATE_HESSIAN subroutine in the new <tt>inv_hessian_mod.f</tt> file. <br />
<br />
--[[User:Daven|Daven]] 00:23, 13 January 2012 (EST)<br />
<br />
==== Offline aerosol adjoint. (adj32_013) ====<br />
Yuhao Mao and Qinbin Li at UCLA have implemented the offline aerosol simulation adjoint for black carbon. This update also included an observation operator for measurements of these species at IMPROVE sites, IMPROVE_BC_OC_OBS and <tt>improve_bc_mod.f</tt>.<br />
<br />
The offline aerosol adjoint will also work for the recently implemented dust aerosol adjoint (it does not yet include support for sulfate aerosol, sea salt or SOA). For example, you may now run with <tt>input.geos</tt><br />
<br />
%%% TRACER MENU %%% : <br />
Type of simulation : 10<br />
Number of Tracers : 4<br />
Tracer Entries -------> : TR# Name g/mole Tracer Members; () = emitted<br />
Tracer #1 : 1 BCPI 12.0<br />
Tracer #2 : 2 OCPI 12.0<br />
Tracer #3 : 3 BCPO 12.0<br />
Tracer #4 : 4 OCPO 12.0<br />
<br />
and <tt>input.gcadj</tt><br />
<br />
FOR LADJ_EMS :<br />
NNEMS: ems groups implemented : 12<br />
Emission entries ------------> : EMS# ems_name opt SF_DEFAULT REG_PARAM ERROR<br />
Emission #1 : 1 IDADJ_EBCPI_an T 1 1 1<br />
Emission #2 : 2 IDADJ_EBCPO_an T 1 1 1<br />
Emission #3 : 3 IDADJ_EOCPI_an T 1 1 1<br />
Emission #4 : 4 IDADJ_EOCPO_an T 1 1 1<br />
Emission #5 : 5 IDADJ_EBCPI_bf T 1 1 1<br />
Emission #6 : 6 IDADJ_EBCPO_bf T 1 1 1<br />
Emission #7 : 7 IDADJ_EOCPI_bf T 1 1 1<br />
Emission #8 : 8 IDADJ_EOCPO_bf T 1 1 1<br />
Emission #9 : 9 IDADJ_EBCPI_bb T 1 1 1<br />
Emission #10 : 10 IDADJ_EBCPO_bb T 1 1 1<br />
Emission #11 : 11 IDADJ_EOCPI_bb T 1 1 1<br />
Emission #12 : 12 IDADJ_EOCPO_bb T 1 1 1<br />
<br />
--[[User:Daven|Daven]] 19:27, 13 January 2012 (EST)<br />
<br />
==== Augmented list of supported VOC sensitivities (adj32_014) ====<br />
From Kateryna Lapina, sensitivities of anthropogenic, biofuel and biomass burning emissions of all VOCs (ALK4, MEK, ALD2, PRPE, C3H8, CH2O C2H6). To implement, simply list these in <tt>input.gcadj</tt> as:<br />
<br />
Emission #34 : 34 IDADJ_EALK4_an T 1 1 1<br />
Emission #35 : 35 IDADJ_EALK4_bf T 1 1 1<br />
Emission #36 : 36 IDADJ_EALK4_bb T 1 1 1<br />
Emission #37 : 37 IDADJ_EACET_an T 1 1 1<br />
Emission #38 : 38 IDADJ_EACET_bf T 1 1 1<br />
Emission #39 : 39 IDADJ_EACET_bb T 1 1 1<br />
Emission #40 : 40 IDADJ_EMEK_an T 1 1 1<br />
Emission #41 : 41 IDADJ_EMEK_bf T 1 1 1<br />
Emission #42 : 42 IDADJ_EMEK_bb T 1 1 1<br />
Emission #43 : 43 IDADJ_EALD2_an T 1 1 1<br />
Emission #44 : 44 IDADJ_EALD2_bf T 1 1 1<br />
Emission #45 : 45 IDADJ_EALD2_bb T 1 1 1<br />
Emission #46 : 46 IDADJ_EPRPE_an T 1 1 1<br />
Emission #47 : 47 IDADJ_EPRPE_bf T 1 1 1<br />
Emission #48 : 48 IDADJ_EPRPE_bb T 1 1 1<br />
Emission #49 : 49 IDADJ_EC3H8_an T 1 1 1<br />
Emission #50 : 50 IDADJ_EC3H8_bf T 1 1 1<br />
Emission #51 : 51 IDADJ_EC3H8_bb T 1 1 1<br />
Emission #52 : 52 IDADJ_ECH2O_an T 1 1 1<br />
Emission #53 : 53 IDADJ_ECH2O_bf T 1 1 1<br />
Emission #54 : 54 IDADJ_ECH2O_bb T 1 1 1<br />
Emission #55 : 55 IDADJ_EC2H6_an T 1 1 1<br />
Emission #56 : 56 IDADJ_EC2H6_bf T 1 1 1<br />
Emission #57 : 57 IDADJ_EC2H6_bb T 1 1 1<br />
<br />
where in the example above emissions 1-33 were the previously defined standard set.<br />
<br />
--[[User:Daven|Daven]] 19:27, 13 January 2012 (EST)<br />
<br />
==== Remove bin_data.f and adj_bin_data.f ====<br />
<br />
From Zhe Jiang, remove the obsolete files bin_data.f and adj_bin_data.f, and remove their listings in Makefile.ifort.hdf<br />
<br />
--[[User:Daven|Daven]] 20:27, 16 January 2012 (EST)<br />
<br />
==== Several updates to support nested full chemistry adjoint (adj32_015) ====<br />
Several updates from Zhe Jiang support running the adjoint full-chemistry simulation over the NESTED_NA and NESTED_SD domain. The latter is a new domain, like NESTED_NA, but slightly less padding around NA. <br />
<br />
===== Support nested in emissions_adj.f =====<br />
<br />
<tt>emissions_adj.f</tt> has been updated to support nested emissions from the CAC, NEI05 and EMEP inventories.<br />
<br />
===== Cleanup code in adj_arrays_mod.f =====<br />
Remove obsolete variable NOR in <tt>adj_arrays_mod.f</tt><br />
<br />
<br />
--[[User:Daven|Daven]] 15:01, 18 January 2012 (EST)<br />
<br />
===== Do not call GET_BOUNDING_BOX in input_adj_mod.f =====<br />
When calculating IFD and JFD from a user specified latitude and longitude, the call to GET_BOUNDING_BOX causes a crash, so we need to modify this part of the code to work on the nested domain. For now, just re-order the code so that this is only called if USEINDEX is false, and run the nested code with USEINDEX true.<br />
<br />
==== MOPITT v5 observation operator (adj32_016) ====<br />
Zhe Jiang added support for the MOPITT v5 products. <br />
* Add MOPITT_V5_CO_OBS pre-processor flag to define_adj.h <br />
*Updates to <tt>adj_arrays_mod.f</tt>, <tt>mopitt_obs_mod.f </tt>, <tt>HdfVdModule.f90</tt>, <tt>input_adj_mod.f</tt>.<br />
<br />
--[[User:Daven|Daven]] 15:01, 18 January 2012 (EST)<br />
<br />
==== Updates for tagged CO simulation (adj32_017) ====<br />
Several updates from Zhe Jiang have been implemented for the tagged-CO implementation. See also changes to the forward model, [[GEOS-Chem_Adjoint_v32#Updates_for_tagged_CO_simulation__.28adj32_017.29]]<br />
<br />
===== Include CH4 gradient =====<br />
Update <tt>inverse_mod.f</tt> so that oxidation of CH4 is incorporated in the gradient <br />
<br />
===== Cleanup and small fixes to tagged_co_adj_mod.f =====<br />
Remove COSF and TIME_STEPS in <tt>tagged_co_adj_mod.f</tt>.<br />
<br />
===== Cleanup and merge CALC_APRIORI =====<br />
Updates for calculation of the penalty term:<br />
*Now valid for full chemistry and tagged CO simulations<br />
*Add subroutine READ_APERROR<br />
<br />
===== Enforce factor of 0.5 in cost function definition =====<br />
A consequence of the merging of the full-chem and CO routines in CALC_APRIORI is that the definition of the cost function needs to be made consistent. Originally, the CO code omitted the factors of 1/2 in both terms of the cost function. To correct for this, the cost function (and adjoint forcing) has been updated in the following CO observation operators to match the definition in the full-chem simulation:<br />
*airs_co_obs_mod.f<br />
*mopitt_obs_mod.f<br />
*sciabr_co_obs_mod<br />
<br />
--[[User:Daven|Daven]] 15:01, 18 January 2012 (EST)<br />
<br />
==== Include adjoint of CH4 simulation (adj32_023) ====<br />
Updates from Kevin Wecht. Modifications in the following files:<br />
* checkpt_mod.f<br />
* adj_arrays_mod.f<br />
* chemistry_adj_mod.f<br />
* cleanup_adj.f<br />
* define_adj.h<br />
* emissions_adj_mod.f<br />
* geos_chem_adj_mod.f<br />
* input_adj_mod.f<br />
<br />
* a3_read_mod.f<br />
* dao_mod.f<br />
<br />
* geos_chem_mod.f<br />
* global_ch4_mod.f<br />
* global_oh_mod.f<br />
<br />
* CMN_DIAG and diag3.f, diag_mod.f, initialize.f, ndxx_setup.f<br />
* julday_mod.f (fwd model bug fix)<br />
<br />
New files<br />
* netcdf_util_mod.f<br />
* global_ch4_adj_mod.f<br />
* tes_ch4_mod.f<br />
* scia_ch4_mod.f<br />
* leo_ch4_mod.f<br />
* mem_ch4_mod.f<br />
*geocape_ch4_mod.f<br />
<br />
===== Add support for ND19, ND59 and ND60 =====<br />
Modified CMN_DIAG and diag3.f, diag_mod.f, initialize.f, ndxx_setup.f<br />
<br />
===== Add support for OH_LOSS =====<br />
Modified diag_oh_mod.f <br />
<br />
<br />
--[[User:Daven|Daven]] 13:26, 12 February 2012 (EST)<br />
<br />
==== Add population weighted cost function option (adj32_024) ====<br />
Update from Steven Vogel adds a new option for cost function evaluation, LPOP_UGM3, and a new file <tt>population_mod.f</tt>. <br />
<br />
--[[User:Daven|Daven]] 01:10, 14 February 2012 (EST)<br />
<br />
==== Add strat production and loss rates as active control parameters (adj32_025) ====<br />
Hyung-Min Lee has developed the adjoint with respect to production and loss control parameters, LADJ_STRAT, based on the new stratospheric fluxes from Lee Murray [[Stratospheric_chemistry]]. Also fix debug printing for JLOP. <br />
<br />
Updated files include:<br />
*logical_adj_mod.f<br />
*adj_arrays_mod.f<br />
*checkpoint_mod.f<br />
*checkpt_mod.f<br />
*chemdr_adj.f<br />
*geos_chem_adj_mod.f<br />
*input_adj_mod.f<br />
*inverse_driver.f<br />
*inverse_mod.f<br />
*linoz_adj_mod.f (completely replaced)<br />
*gamap_mod.f<br />
*partition_mod.f<br />
*tacer_mod.f<br />
*strat_chem_adj_mod.f (added)<br />
<br />
<br />
fwd:<br />
*logical_mod.f<br />
<br />
modified:<br />
*chemdr.f<br />
*geos_chem_mod.f (remove UPBDFLX_NOY and DO_UPBDFLX)<br />
*input_mod.f<br />
*linoz_mod.f (replace w/v9)<br />
*strat_chem_mod.f (add from v9)<br />
*Makefile.ifort.* (because new strat chem uses netcdf)<br />
<br />
--[[User:Daven|Daven]] 15:22, 14 February 2012 (EST)<br />
<br />
==== Calculate array sizes dynamically in CINSPECT (adj32_026) ====<br />
The arrays VAR_FD and RCONST_FD have been moved to adj_arrays_mod.f and are now dynamically allocated. The following error trap (which would frequently cause problems for new users) is no longer relevant:<br />
<br />
IF ( NCHEM > 1000 ) THEN <br />
CALL ERROR_STOP( 'Need to boost NCHEM max', 'chemistry_mod.f')<br />
ENDIF <br />
<br />
--[[User:Daven|Daven]] 16:22, 23 February 2012 (EST)<br />
<br />
=== Bug fixes in adjoint model ===<br />
<br />
----<br />
<br />
==== Mover error checking for O3 assimilation (adj32_002) ====<br />
'''''[mailto:N.Bousserez@dal.ca Nicolas Bousserez] wrote:'''''<br />
<br />
:the code crashes in "input_adj_mod.f" when doing the checking of observation settings (see line 2027 in my code). It is apparently due to "IDCSPEC_ADJ", which is not allocated/defined yet at the time the program executes those lines. I'm skipping these checks for now then. Note: The code crashes also when using "TES_O3_OBS" instead of "SCIA_DAL_NO2_OBS", so it is not related to what I implemented apparently.<br />
<br />
The following block of code has been moved from <tt>input_adj_mod.f</tt> to subroutine <tt>INIT_CSPEC_ADJ</tt> in <tt>adj_arrays_mod.f</tt>, as it can only be called after the <tt>CSPEC</tt> variables have been initialized. <br />
<br />
#elif defined ( TES_O3_OBS ) <br />
! Since the O3 obs operators will pass adjoints back <br />
! to CSPEC via CSPEC_AFTER_CHEM_ADJ, we need to make sure that <br />
! these species are listed as observed species<br />
FOUND = .FALSE.<br />
DO N = 1, NOBS_CSPEC<br />
<br />
IF ( TRIM( NAMEGAS( IDCSPEC_ADJ(N) ) ) == 'O3' ) THEN<br />
FOUND = .TRUE.<br />
ENDIF<br />
<br />
ENDDO<br />
IF ( .not. FOUND ) THEN<br />
<br />
CALL ERROR_STOP( ' Need to list O3 as observed species',<br />
& ' input_adj_mod ' )<br />
ENDIF<br />
#endif<br />
<br />
--[[User:Daven|Daven]] 16:38, 6 January 2012 (EST)<br />
<br />
==== Retire CSPEC_NO2_ADJ (adj32_003) ====<br />
<br />
'''''[mailto:N.Bousserez@dal.ca Nicolas Bousserez] wrote:'''''<br />
<br />
:In "chemdr_adj.F" I removed this part:<br />
#if defined( SCIA_KNMI_NO2_OBS ) || defined( SCIA_DAL_NO2_OBS )<br />
! Apply forcing from satellite observations<br />
CSPEC_ADJ(:,IDNO2) = CSPEC_ADJ(:,IDNO2) + CSPEC_NO2_ADJ(:)<br />
CSPEC_NO2_ADJ(:) = 0d0<br />
#endif<br />
:because it seems like I'm already updating the adjoint of NO2 in "sciadal_no2_obs_mod.f" through CSPEC_AFTER_CHEM_ADJ. I would like to be sure that it is correct though.<br />
<br />
'''''[mailto:daven.henze@colorado.edu Daven Henze] wrote:'''''<br />
<br />
:Yes, that is correct. I didn't want to take it out incase you had built your obs operator using CSPEC_NO2_ADJ. But since you used the more generic CSPEC_AFTER_CHEM_ADJ, then we can delete this section entirely and use the section higher up in this file ( IF (LCSPEC_OBS) ...) to apply the forcing from your obs operator to CSPEC_ADJ.<br />
<br />
Now use <tt>CSPEC_AFTER_CHEM_ADJ</tt> instead of <tt>CSPEC_NO2_ADJ</tt> in <tt>chemdr_adj.f</tt>.<br />
<br />
--[[User:Daven|Daven]] 18:04, 6 January 2012 (EST)<br />
<br />
==== Fix to NaNs in STT_ADJ everywhere after adjoint of partitioning (adj32_004) ====<br />
The forward gas-phase chemical solver may sometimes need to repeat the internal integration of the reaction rate ODEs within a single grid cell. When this happens, we now only redo the model calculation during the forward model run, not the adjoint, as this can lead to a corrupt CSPEC_ADJ array globally (not sure how that happens, but it does), eventually triggering a check for NaNs after the partitioning adjoint. See updates to <tt>chemistry_mod.f</tt> and <tt>gckpp_adj_Integrator.f90</tt>. Thanks to Richard Xu for helping pinpoint this bug. <br />
<br />
--[[User:Daven|Daven]] 18:03, 6 January 2012 (EST)<br />
<br />
==== Fix to NaNs in sulfate aerosol adjoint tracers (adj32_005) ====<br />
Owing to cancelation of errors, it's possible to seemingly randomly end up with NaNs coming out of <tt>CHEM_SO2_ADJ</tt> . This can happen if L3 ends up equaling 1d0 exactly. We now trap for this and skip in the rate event this happens. <br />
<br />
--[[User:Daven|Daven]] 18:03, 6 January 2012 (EST)<br />
<br />
==== Remove superfluous debugging which had <tt>LAVHRRLAI</tt> set to FALSE ====<br />
Noted by Fabien. Upated in <tt>input_mod.f</tt>.<br />
<br />
--[[User:Daven|Daven]] 18:07, 6 January 2012 (EST)<br />
<br />
==== Fix for reseting deposition adjoints (adj32_006) ====<br />
<br />
'''''[mailto:fabienpaulot@gmail.com Fabien Paulot] wrote:'''''<br />
<br />
:It seems it's "a bug" in the code. To keep track of the dry deposition flux of HNO3 I use dryhno3 (in cspec array). The CSPEC_ADJ(DRYHNO3,:) should be reset to zero at every time step but it is not. That's because you are saving in chk_cspec before it is reset to zero. Easy fix, just force to zero in chemdr_adj before the forcing is applied et voila, it works (the adjoint does not diverge any more after one hour). It should not affect anything as long as there is no forcing on dryhno3 (or any dry dep counters for that matter). <br />
<br />
We add the following to reset the dry dep adjoints in <tt>chemdr_adj.f</tt>:<br />
!$OMP PARALLEL DO<br />
!$OMP+DEFAULT( SHARED )<br />
!$OMP+PRIVATE( N, NK, JJ )<br />
DO N = 1,NUMDEP<br />
NK = NTDEP(N)<br />
IF (NK.NE.0) THEN<br />
JJ = IRM(NPRODLO+1,NK,NCS)<br />
IF (JJ.GT.0) THEN<br />
CSPEC_ADJ(:,JJ) = 0.0D0<br />
ENDIF<br />
ENDIF<br />
ENDDO<br />
!$OMP END PARALLEL DO<br />
<br />
--[[User:Daven|Daven]] 18:34, 6 January 2012 (EST)<br />
<br />
==== Update comments for STT_ADJ unit conversion ====<br />
'''''[mailto:mkeller@atmosp.physics.utoronto.ca Martin Keller] wrote:'''''<br />
<br />
:I'm currently guessing that CONVERT_UNITS(1,etc.) does something like:<br />
<br />
STT: kg -> v/v<br />
<br />
STT_ADJ: J/[v/v] -> J/kg<br />
<br />
:and similarly CONVERT_UNITS(2,etc.)<br />
<br />
STT: v/v -> kg<br />
<br />
STT_ADJ: J/kg -> J/[v/v]<br />
<br />
:From a dimensional analysis this would seem plausible. It would also explain some of the errors I'm currently fighting with. If this is correct, then a short comment should perhaps be added to the code.<br />
<br />
The comments in <tt>geos_chem_adj_mod.f</tt> have been updated surrounding calls to CONVERT_UNITS.<br />
<br />
--[[User:Daven|Daven]] 18:43, 6 January 2012 (EST)<br />
<br />
==== Only write upbd checkpoint files if necessary (adj32_007) ====<br />
Thanks to Farhan Akhtar for pointing this out. We now check to make sure writing these checkpoint files is necessary (i.e., it is an adjoint run, and not just a pseudo obs calculation) in <tt>linoz_mod.f</tt>:<br />
<br />
! Now see if checkpointing is necessary (fa, dkh, 01/06/12, adj32_007) <br />
IF ( LADJ .and. N_CALC > 0 ) THEN<br />
CALL MAKE_UPBDFLX_CHKFILE( NYMD, NHMS, TAU )<br />
ENDIF <br />
<br />
--[[User:Daven|Daven]] 18:54, 6 January 2012 (EST)<br />
<br />
==== Fix for reading TES diagnostic flags from input file (adj32_010) ====<br />
The problem was that if the CO diagnostics (LDCOSAT) were set to false, then the line counting would be off for any subsequent lines read in input.gcadj. The file input_adj_mod.f has been updated to work properly regardless of the value of LDCOSAT. <br />
<br />
<br />
==== Bug fix for unit conversion in aircraft and chemical source terms for CO2 adjoint (adj32_018) ====<br />
The units of the emissions scaling factor adjoints for the aircraft and chemical source terms were incorrect. These sources are in units per volume, not area, so their conversion needed to be<br />
<br />
E_CO2_ADJ = E_CO2_ADJ * BOXVL(I,J,L)<br />
& * DTSRCE / XNUMOL_CO2<br />
<br />
instead of<br />
<br />
!!E_CO2_ADJ = E_CO2_ADJ * A_CM2 * DTSRCE / XNUMOL_CO2<br />
<br />
Also, the ordering of the application of the CO2 emissions scaling factors was made consistent across all sectors such that the ND04 diagnostic is correct (previously the scaling had been applied after the diagnostic for ocean exchange, balanced biosphere, and CO2 surface correction for CO oxidation). <br />
<br />
--[[User:Daven|Daven]] 10:32, 9 February 2012 (EST)<br />
<br />
==== Bug fix for parallel loops in tagged_co_mod.f (adj32_017) ====<br />
<br />
F_OF_PBL and ED_CO were declared thread PRIVATE in several parallel do loops. <br />
<br />
--[[User:Daven|Daven]] 18:59, 8 January 2012 (EST)<br />
<br />
==== Bug fix for adjoints in nested cushions (adj32_020) ====<br />
In <tt>transport_mod.f</tt>, now call DO_WINDOW_TPCORE_BC_ADJ to set the adjoint values to zero within the cushion. <br />
<br />
--[[User:Daven|Daven]] 15:26, 9 February 2012 (EST)<br />
<br />
== Outstanding issues not yet resolved in v32 ==<br />
New features to add to the pipeline for the v33 release:<br />
*Reaction rates / deposition / updated chemistry (Paulot, Harvard)<br />
*ETOH simulations (Millet, UMN)<br />
*NH3 updates (Zhu, Colorado)<br />
*NO2/SO2 obs operators (Bousserez, Dalhousie)<br />
*ANISO (Capps, Georgia Tech)<br />
*Update from strat_chem_201106 to strat_chem_201109 in strat_chem_mod.f ?<br />
<br />
Some bug fixes that need implementing<br />
* Fix bug in Makefile.ifort.hdf, .lidort and .netcdf: need to list define.h in set of include files for strat_chem_mod.f <br />
* Trap for xFD not in trop in geos_chem_mod.f, line 1622: add "IF JLOP(IFD,JFD,LFD) > 0" to trap for LFD not in tropospher<br />
* Fix gckpp_adj_Integrator.f90 <br />
* Increase magnitude of filter in partition_adj.f <br />
* Fix NH3 seasonality -- check with Steve<br />
* Fix line 4737 of lidort_mod.f (Jamin Koo, 04/20/2012)<br />
* Remove M0 and M0_ADJ from adj_arrays_mod.f (not used) hml<br />
<br />
<br />
And some bugs of note with no fix yet<br />
*Can't specify IFD, JFD using latitudes with nested domain<br />
*Can't use NEI emissions with tagged-CO</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v33&diff=13624GEOS-Chem Adjoint v332013-07-03T23:06:25Z<p>Hyung-Min Lee: /* What's new in this version */</p>
<hr />
<div>== Overview ==<br />
<br />
BETA RELEASE<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v32]]<br />
<br />
=== What's new in this version ===<br />
<br />
GEOS-Chem Adjoint v33 contains the following major updates and improvements. <br />
<br />
{| border=1 cellspacing=0 cellpadding=5 <br />
|- bgcolor="#CCCCCC"<br />
!Feature<br />
!Type<br />
!Submitted by<br />
!Committed by<br />
!Version<br />
!Status<br />
|-<br />
|[[#Add_TES_O3_IRK_.28v33a.29|Add TES O3 IRK observation operator]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Kevin Bowman (JPL)<br />
|Daven Henze (CU Boulder)<br />
|v33a<br />
|Approved 13 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_wet_deposition_adjoint_.28v33b.29|Fix wetdep adjoint]]<br />
|Adjoint bug fix<br />
|Fabien Paulot (Harvard University), Daven Henze (CU Boulder)<br />
|Daven Henze (CU Boulder)<br />
|v33b<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_checkpointing_of_aircraft_SO2_emissions_.28v33c.29|Fix checkpointing of aircraft SO2 emissions]]<br />
|Adjoint bug fix<br />
|Daven Henze, Kateryna Lapina (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33c<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_linoz_adjoint_.28v33d.29|Fix linoz adjoint]]<br />
|Adjoint bug fix<br />
|Hyung-Min Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33d<br />
|Approved 07 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Adjoint_stability_update_.28v33e.29| Adjoint stability update]]<br />
|Adjoint bug fix<br />
|Jamin Koo (MIT), Daven Henze (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_SE_Asia_NH3_seasonality_.28v33e.29| Fix SE Asia NH3 seasonality]]<br />
|Forward bug fix<br />
|Shailesh Kharol (Dalhousie), Steven Vogel (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_problem_with_LICS_flag_in_using_SCHEM_.28v33f.29| Fix problem with LICS flag in using SCHEM]]<br />
|Adjoint bug fix<br />
|Hyung-Min Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33f<br />
|Approved 13 July 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Inverse_Hessian_and_pseudo-obs_cost_function_formulation_updates_.28v33g.29| L-BFGS inverse Hessian and pseudo-obs cost function formulation updates]]<br />
|Adjoint update<br />
|Nicolas Bousserez (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33g<br />
|Approved 10 August 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#BC_fix_to_support_offline_simulation_and_AOD_.28v33h.29| BC fix to support offline simulation and AOD]]<br />
|Adjoint update<br />
|Li Zhang (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33h<br />
|Approved 5 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Two_Minor_Bug_Fixes_.28v33h.29| Two minor bug fixes]]<br />
|Adjoint bug fix<br />
|Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33h<br />
|Approved 5 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_Monoterpene_SOA_zero_bug | Biomass Emissions bug fix]]<br />
|Forward bug fix<br />
| Jeffrey Pierce (CSU)<br />
|Yanko Davila (CU Boulder)<br />
|v33h<br />
|Approved 5 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#L-BFGS_inverse_Hessian_Update_.28v33i.29 | L-BFGS inverse Hessian update]]<br />
|Adjoint Update<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33i<br />
|Approved 21 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Pseudo-Obs_Test_Bug_Fix_.28v33i.29 | Pseudo-obs test bug fix]]<br />
|Adjoint bug fix<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33i<br />
|Approved 21 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Version_33i_patches_.28v33i-patch1.29 | Version 33i patch1]]<br />
|Adjoint bug fix<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33i-patch1<br />
|Approved 9 October 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Version_33i_patches_.28v33i-patch2.29| Version 33i patch2]]<br />
|Adjoint bug fix<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33i-patch2<br />
|Approved 22 October 2012<br />
|-bgcolor="#CCFFFF"<br />
|}<br />
<br />
==== Updates to forward model ====<br />
<br />
==== Bug fixes in forward model ====<br />
<br />
===== Fix SE Asia NH3 seasonality (v33e) =====<br />
<br />
The bug fix from the forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Anthropogenic_emissions#Streets_2000 fixing the seasonality of the NH3 emissions over SE Asia] has also been applied.<br />
<br />
1 week benchmark is approved [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_003]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:40, 10 May 2012 (EDT)<br />
<br />
===== Fix Monoterpene SOA zero bug =====<br />
<br />
The bug fix from the forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Setting_up_GEOS-Chem_nested_grid_simulations#Monoterpene_SOA_zero_bug_fix Monoterpene SOA zero bug fix] has been applied.<br />
<br />
--[[User:Yanko|Yanko]] 16:29, 6 September 2012 (EDT)<br />
<br />
<br />
==== Updates to adjoint model ====<br />
<br />
===== Add TES O3 IRK (v33a) =====<br />
<br />
Implemented observation operator for assimilation of TES O3 IRKs. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_001]<br />
<br />
--[[User:Daven|Daven]] 19:03, 13 April 2012 (EDT)<br />
<br />
===== BC fix to support offline simulation and AOD (v33h) =====<br />
- Fix to support offline BC simulation and AOD calculation<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE&pli=1#gid=0 v33h]<br />
<br />
--[[User:Yanko|Yanko]] 16:43, 6 September 2012 (EDT)<br />
<br />
===== L-BFGS inverse Hessian Update (v33i) =====<br />
<br />
Now use the background error instead of the initial L-BFGS inverse Hessian diagonal value where the gradient of the cost function is very small (in this case the inverse Hessian value is not updated by the algorithm).<br />
Calculate and output the Degree of Freedom for Signal.<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE&pli=1#gid=0 v33i]<br />
<br />
--[[User:Yanko|Yanko]] 20:34, 21 September 2012 (EDT)<br />
<br />
==== Bug fixes in adjoint model ====<br />
<br />
===== Fix wet deposition adjoint (v33b) =====<br />
The adjoint was not properly handling evaporation. This fix helps make the adjoint of species such as NH3 better in a few locations where adjoints were diverging from forward model sensitivities by up to 50%. <br />
<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_002 vs 33b_002]<br />
<br />
--[[User:Daven|Daven]] 12:51, 20 April 2012 (EDT)<br />
<br />
===== Fix checkpointing of aircraft SO2 emissions (v33c) =====<br />
<br />
The SO2 aircraft emissions are distributed in the model based on a mapping from the 20 level inventory to the GEOS-Chem levels. This mapping depends upon the air masses in the models. The emissions are read in during the first hour of the month in the forward model, but on the last hour of the month in the adjoint. Thus they were being calculated inconsistently in the adjoint compared to the forward model. We now checkpoint these emissions rather than recalculate them<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33c_001 - 33c_004]<br />
<br />
--[[User:Daven|Daven]] 23:08, 7 May 2012 (EDT)<br />
<br />
===== Fix linoz adjoint (v33d) =====<br />
<br />
While the linoz routine is linear, and should be self-adjoint, it isn't exactly self-adjoint. For example, given initial forcings of STT_ADJ = 0d0 in every grid cell, it can generate non-zero values of STT_ADJ as output, which is incorrect. <br />
<br />
In this update we use the discrete adjoint based upon reformulating the code to be consistent with the strat chem routine for the non-Ox species. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33d_001 - 33c_007]<br />
<br />
--[[User:Daven|Daven]] 23:14, 7 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33d (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
===== Adjoint stability update (v33e) =====<br />
<br />
Several fixes were made to increase the numerical stability of the code<br />
*Change the filter to 1d-10 in partition_adj.f<br />
*Better error trapping in gckpp_adj_Integrator.f90 to prevent NaNs following a failed integration step during the adjoint<br />
<br />
<br />
<br />
A few other small changes to improve performance were included in this update. Overall, the benchmark performance is not impacted by these update, see [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_001]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:26, 10 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33e (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.'''<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
<br />
===== Fix problem with LICS flag in using SCHEM (v33f) =====<br />
<br />
New stratospheric chemistry module (strat_chem_mod.f), was set to loop over the number of tracers (NSTPL) that has read from input.gcadj. It caused a problem when LADJ_STRAT is turned off, LADJ_STRAT is not appeared in input.gcadj (NSTPL = 0), or LADJ = F for spin-up.<br />
<br />
--[[User:Hyung-Min Lee|Hyung-Min Lee]] 19:26, 14 Jul 2012 (EDT)<br />
<br />
===== Inverse Hessian and pseudo-obs cost function formulation updates (v33g) =====<br />
- Possibility to compute the L-BFGS inverse Hessian, which gives the posterior error covariance matrix of the inversion. Here only the diagonal elements (the error variances) of the matrix are computed. Unlike the DFP method, this method allows to compute a posterior errors for large dimensional problems (e.g. high-resolution inversions).<br />
<br />
- The equations for the initial estimate of the inverse Hessian in the DFP algorithm were also updated to allow for quicker convergence. <br />
<br />
- The cost function for the pseudo-observations inversion test has been modified to properly account for the error term specified in the input.gcadj input file (ERROR).<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=1 v33g] <br />
<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 05:34, 11 August 2012 (EDT)<br />
<br />
===== Two Minor Bug Fixes (v33h) =====<br />
- Fix for OpenMP in subroutine EMITHIGH_ADJ. This can correct some errors for the BC emissions adjoints. Without this fix, users may notice some error in FD vs ADJ calculations for shorter simulations.<br />
<br />
- Update subroutine CHEM_SO2_ADJ to avoid underflow.<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE&pli=1#gid=0 v33h]<br />
<br />
--[[User:Yanko|Yanko]] 16:43, 6 September 2012 (EDT)<br />
<br />
===== Pseudo-Obs Test Bug Fix (v33i) =====<br />
- Bug fix for the definition of the observation error in the pseudo-obs inversion test and filter to avoid NaN adjoint value. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE&pli=1#gid=0 v33i]<br />
--[[User:Yanko|Yanko]] 20:34, 21 September 2012 (EDT)<br />
<br />
=====Version 33i patches (v33i-patch1) =====<br />
- Bug fix for too small observation errors when doing pseudo-assimilation experiments<br />
<br />
- Bug Fix for the DFP.<br />
<br />
--[[User:Yanko|Yanko]] 17:42, 1 November 2012 (EDT)<br />
<br />
=====Version 33i patches (v33i-patch2) =====<br />
<br />
- Bug Fix for memory problems caused by the amount of iteration steps kept in memory while doing a Hessian approximation.<br />
<br />
--[[User:Yanko|Yanko]] 17:42, 1 November 2012 (EDT)<br />
<br />
== Outstanding issues not yet resolved in v33 ==<br />
<br />
=== Updates in the pipeline ===<br />
*Reaction rate sensitivities (Paulot, Harvard; Walker, UT)<br />
*Update inverse Hessian approximation (Bousserez, CUB)<br />
*Off-diagonal error covariance matrices (Singh, VT; Liu, JPL)<br />
*Deposition-based cost function and updated chemistry (Paulot, Harvard)<br />
*ETOH simulations (Millet, UMN)<br />
*NH3 updates (Zhu, Colorado)<br />
*NO2/SO2 obs operators (Bousserez, Dalhousie)<br />
*ANISO (Capps, Georgia Tech)</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v33&diff=13623GEOS-Chem Adjoint v332013-07-03T23:05:31Z<p>Hyung-Min Lee: /* Fix problem with LICS flag in using SCHEM (v33f) */</p>
<hr />
<div>== Overview ==<br />
<br />
BETA RELEASE<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v32]]<br />
<br />
=== What's new in this version ===<br />
<br />
GEOS-Chem Adjoint v33 contains the following major updates and improvements. <br />
<br />
{| border=1 cellspacing=0 cellpadding=5 <br />
|- bgcolor="#CCCCCC"<br />
!Feature<br />
!Type<br />
!Submitted by<br />
!Committed by<br />
!Version<br />
!Status<br />
|-<br />
|[[#Add_TES_O3_IRK_.28v33a.29|Add TES O3 IRK observation operator]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Kevin Bowman (JPL)<br />
|Daven Henze (CU Boulder)<br />
|v33a<br />
|Approved 13 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_wet_deposition_adjoint_.28v33b.29|Fix wetdep adjoint]]<br />
|Adjoint bug fix<br />
|Fabien Paulot (Harvard University), Daven Henze (CU Boulder)<br />
|Daven Henze (CU Boulder)<br />
|v33b<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_checkpointing_of_aircraft_SO2_emissions_.28v33c.29|Fix checkpointing of aircraft SO2 emissions]]<br />
|Adjoint bug fix<br />
|Daven Henze, Kateryna Lapina (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33c<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_linoz_adjoint_.28v33d.29|Fix linoz adjoint]]<br />
|Adjoint bug fix<br />
|Hyungmin Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33d<br />
|Approved 07 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Adjoint_stability_update_.28v33e.29| Adjoint stability update]]<br />
|Adjoint bug fix<br />
|Jamin Koo (MIT), Daven Henze (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_SE_Asia_NH3_seasonality_.28v33e.29| Fix SE Asia NH3 seasonality]]<br />
|Forward bug fix<br />
|Shailesh Kharol (Dalhousie), Steven Vogel (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_problem_with_LICS_flag_in_using_SCHEM_.28v33f.29| Fix problem with LICS flag in using SCHEM]]<br />
|Adjoint bug fix<br />
|Hyungmin Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33f<br />
|Approved 13 July 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Inverse_Hessian_and_pseudo-obs_cost_function_formulation_updates_.28v33g.29| L-BFGS inverse Hessian and pseudo-obs cost function formulation updates]]<br />
|Adjoint update<br />
|Nicolas Bousserez (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33g<br />
|Approved 10 August 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#BC_fix_to_support_offline_simulation_and_AOD_.28v33h.29| BC fix to support offline simulation and AOD]]<br />
|Adjoint update<br />
|Li Zhang (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33h<br />
|Approved 5 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Two_Minor_Bug_Fixes_.28v33h.29| Two minor bug fixes]]<br />
|Adjoint bug fix<br />
|Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33h<br />
|Approved 5 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_Monoterpene_SOA_zero_bug | Biomass Emissions bug fix]]<br />
|Forward bug fix<br />
| Jeffrey Pierce (CSU)<br />
|Yanko Davila (CU Boulder)<br />
|v33h<br />
|Approved 5 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#L-BFGS_inverse_Hessian_Update_.28v33i.29 | L-BFGS inverse Hessian update]]<br />
|Adjoint Update<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33i<br />
|Approved 21 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Pseudo-Obs_Test_Bug_Fix_.28v33i.29 | Pseudo-obs test bug fix]]<br />
|Adjoint bug fix<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33i<br />
|Approved 21 September 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Version_33i_patches_.28v33i-patch1.29 | Version 33i patch1]]<br />
|Adjoint bug fix<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33i-patch1<br />
|Approved 9 October 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Version_33i_patches_.28v33i-patch2.29| Version 33i patch2]]<br />
|Adjoint bug fix<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v33i-patch2<br />
|Approved 22 October 2012<br />
|-bgcolor="#CCFFFF"<br />
|}<br />
<br />
==== Updates to forward model ====<br />
<br />
==== Bug fixes in forward model ====<br />
<br />
===== Fix SE Asia NH3 seasonality (v33e) =====<br />
<br />
The bug fix from the forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Anthropogenic_emissions#Streets_2000 fixing the seasonality of the NH3 emissions over SE Asia] has also been applied.<br />
<br />
1 week benchmark is approved [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_003]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:40, 10 May 2012 (EDT)<br />
<br />
===== Fix Monoterpene SOA zero bug =====<br />
<br />
The bug fix from the forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Setting_up_GEOS-Chem_nested_grid_simulations#Monoterpene_SOA_zero_bug_fix Monoterpene SOA zero bug fix] has been applied.<br />
<br />
--[[User:Yanko|Yanko]] 16:29, 6 September 2012 (EDT)<br />
<br />
<br />
==== Updates to adjoint model ====<br />
<br />
===== Add TES O3 IRK (v33a) =====<br />
<br />
Implemented observation operator for assimilation of TES O3 IRKs. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_001]<br />
<br />
--[[User:Daven|Daven]] 19:03, 13 April 2012 (EDT)<br />
<br />
===== BC fix to support offline simulation and AOD (v33h) =====<br />
- Fix to support offline BC simulation and AOD calculation<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE&pli=1#gid=0 v33h]<br />
<br />
--[[User:Yanko|Yanko]] 16:43, 6 September 2012 (EDT)<br />
<br />
===== L-BFGS inverse Hessian Update (v33i) =====<br />
<br />
Now use the background error instead of the initial L-BFGS inverse Hessian diagonal value where the gradient of the cost function is very small (in this case the inverse Hessian value is not updated by the algorithm).<br />
Calculate and output the Degree of Freedom for Signal.<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE&pli=1#gid=0 v33i]<br />
<br />
--[[User:Yanko|Yanko]] 20:34, 21 September 2012 (EDT)<br />
<br />
==== Bug fixes in adjoint model ====<br />
<br />
===== Fix wet deposition adjoint (v33b) =====<br />
The adjoint was not properly handling evaporation. This fix helps make the adjoint of species such as NH3 better in a few locations where adjoints were diverging from forward model sensitivities by up to 50%. <br />
<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_002 vs 33b_002]<br />
<br />
--[[User:Daven|Daven]] 12:51, 20 April 2012 (EDT)<br />
<br />
===== Fix checkpointing of aircraft SO2 emissions (v33c) =====<br />
<br />
The SO2 aircraft emissions are distributed in the model based on a mapping from the 20 level inventory to the GEOS-Chem levels. This mapping depends upon the air masses in the models. The emissions are read in during the first hour of the month in the forward model, but on the last hour of the month in the adjoint. Thus they were being calculated inconsistently in the adjoint compared to the forward model. We now checkpoint these emissions rather than recalculate them<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33c_001 - 33c_004]<br />
<br />
--[[User:Daven|Daven]] 23:08, 7 May 2012 (EDT)<br />
<br />
===== Fix linoz adjoint (v33d) =====<br />
<br />
While the linoz routine is linear, and should be self-adjoint, it isn't exactly self-adjoint. For example, given initial forcings of STT_ADJ = 0d0 in every grid cell, it can generate non-zero values of STT_ADJ as output, which is incorrect. <br />
<br />
In this update we use the discrete adjoint based upon reformulating the code to be consistent with the strat chem routine for the non-Ox species. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33d_001 - 33c_007]<br />
<br />
--[[User:Daven|Daven]] 23:14, 7 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33d (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
===== Adjoint stability update (v33e) =====<br />
<br />
Several fixes were made to increase the numerical stability of the code<br />
*Change the filter to 1d-10 in partition_adj.f<br />
*Better error trapping in gckpp_adj_Integrator.f90 to prevent NaNs following a failed integration step during the adjoint<br />
<br />
<br />
<br />
A few other small changes to improve performance were included in this update. Overall, the benchmark performance is not impacted by these update, see [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_001]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:26, 10 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33e (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.'''<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
<br />
===== Fix problem with LICS flag in using SCHEM (v33f) =====<br />
<br />
New stratospheric chemistry module (strat_chem_mod.f), was set to loop over the number of tracers (NSTPL) that has read from input.gcadj. It caused a problem when LADJ_STRAT is turned off, LADJ_STRAT is not appeared in input.gcadj (NSTPL = 0), or LADJ = F for spin-up.<br />
<br />
--[[User:Hyung-Min Lee|Hyung-Min Lee]] 19:26, 14 Jul 2012 (EDT)<br />
<br />
===== Inverse Hessian and pseudo-obs cost function formulation updates (v33g) =====<br />
- Possibility to compute the L-BFGS inverse Hessian, which gives the posterior error covariance matrix of the inversion. Here only the diagonal elements (the error variances) of the matrix are computed. Unlike the DFP method, this method allows to compute a posterior errors for large dimensional problems (e.g. high-resolution inversions).<br />
<br />
- The equations for the initial estimate of the inverse Hessian in the DFP algorithm were also updated to allow for quicker convergence. <br />
<br />
- The cost function for the pseudo-observations inversion test has been modified to properly account for the error term specified in the input.gcadj input file (ERROR).<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=1 v33g] <br />
<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 05:34, 11 August 2012 (EDT)<br />
<br />
===== Two Minor Bug Fixes (v33h) =====<br />
- Fix for OpenMP in subroutine EMITHIGH_ADJ. This can correct some errors for the BC emissions adjoints. Without this fix, users may notice some error in FD vs ADJ calculations for shorter simulations.<br />
<br />
- Update subroutine CHEM_SO2_ADJ to avoid underflow.<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE&pli=1#gid=0 v33h]<br />
<br />
--[[User:Yanko|Yanko]] 16:43, 6 September 2012 (EDT)<br />
<br />
===== Pseudo-Obs Test Bug Fix (v33i) =====<br />
- Bug fix for the definition of the observation error in the pseudo-obs inversion test and filter to avoid NaN adjoint value. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE&pli=1#gid=0 v33i]<br />
--[[User:Yanko|Yanko]] 20:34, 21 September 2012 (EDT)<br />
<br />
=====Version 33i patches (v33i-patch1) =====<br />
- Bug fix for too small observation errors when doing pseudo-assimilation experiments<br />
<br />
- Bug Fix for the DFP.<br />
<br />
--[[User:Yanko|Yanko]] 17:42, 1 November 2012 (EDT)<br />
<br />
=====Version 33i patches (v33i-patch2) =====<br />
<br />
- Bug Fix for memory problems caused by the amount of iteration steps kept in memory while doing a Hessian approximation.<br />
<br />
--[[User:Yanko|Yanko]] 17:42, 1 November 2012 (EDT)<br />
<br />
== Outstanding issues not yet resolved in v33 ==<br />
<br />
=== Updates in the pipeline ===<br />
*Reaction rate sensitivities (Paulot, Harvard; Walker, UT)<br />
*Update inverse Hessian approximation (Bousserez, CUB)<br />
*Off-diagonal error covariance matrices (Singh, VT; Liu, JPL)<br />
*Deposition-based cost function and updated chemistry (Paulot, Harvard)<br />
*ETOH simulations (Millet, UMN)<br />
*NH3 updates (Zhu, Colorado)<br />
*NO2/SO2 obs operators (Bousserez, Dalhousie)<br />
*ANISO (Capps, Georgia Tech)</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v34&diff=13622GEOS-Chem Adjoint v342013-07-03T23:03:45Z<p>Hyung-Min Lee: /* Updated GMI database */</p>
<hr />
<div>== Overview ==<br />
<br />
BETA RELEASE<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v33]]<br />
<br />
=== What's new in this version ===<br />
<br />
GEOS-Chem Adjoint v34 contains the following major updates and improvements. <br />
<br />
{| border=1 cellspacing=0 cellpadding=5 <br />
|- bgcolor="#CCCCCC"<br />
!Feature<br />
!Type<br />
!Submitted by<br />
!Committed by<br />
!Version<br />
!Status<br />
<br />
|-<br />
|[[#Building_Process_Updated |Building Process Updated]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Add_GFED3_Emissions |Add support for GFED3 Emissions]]<br />
|Forward Model Update<br />
|Melissa Payer (Harvard)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Checkpoint_files_creation_updated |Checkpoint files creation updated]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Population_mod_Updated |population-weighted cost function updated]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Fix_GFED2_Emissions |Add support for GFED2 8day and 3hr Emissions]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#M_Undefined |M undefined]]<br />
|Adjoint Bug Fix<br />
|Daven Henze (CU Boulder), Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Error_compiling_on_ifort_12 |Error compiling on ifort 12]]<br />
|Adjoint Bug Fix<br />
|Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Set_observation_errors_to_1.25_in_pseudo-obs_test |Set observation errors to 1% in pseudo-obs test]]<br />
|Adjoint Bug Fix<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34a<br />
|Approved 11 Dec 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Update_to_Nested_Domains |Update to Nested Domains]]<br />
|Forward Model Bug Fix<br />
|Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34b<br />
|Approved 25 Jan 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Fixed_Makefile_Bug |Fixed Makefile Bug]]<br />
|Adjoint Bug Fix<br />
|Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34b<br />
|Approved 25 Jan 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Added_support_for_ICOADS_ship_emissions |Added support for ICOADS ship emissions]]<br />
|Forward Model Update<br />
|Katerina Lapina (CU Boulder), Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34c<br />
|Approved 5 Feb 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Washout_fix_for_non-aerosol_species |Washout fix for non-aerosol species]]<br />
|Forward Model Bug Fix<br />
|Helen Amos<br />
|Yanko Davila (CU Boulder)<br />
|v34d<br />
|Approved 4 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Add_support_for_RETRO_emissions |Add support for RETRO emissions]]<br />
|Forward Model Update<br />
|Wes Reinhart (U. Minnesota), Dylan Millet (U. Minnesota)<br />
|Yanko Davila (CU Boulder)<br />
|v34e<br />
|Approved 12 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Bug_fixes_for_biofuel_emissions |Bug fixes for biofuel emissions]]<br />
|Forward Model Bug Fix<br />
|Melissa Payer<br />
|Yanko Davila (CU Boulder)<br />
|v34e<br />
|Approved 12 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Fix_bug_in_EDGAR_ship_emissions_code | Fix bug in EDGAR ship emissions code]]<br />
|Forward Model Bug Fix<br />
|Michael Long<br />
|Yanko Davila (CU Boulder)<br />
|v34e<br />
|Approved 12 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Bug_fix_for_ship_NOx_emissions | Bug fix for ship NOx emissions]]<br />
|Forward Model Bug Fix<br />
|Chris Holmes<br />
|Yanko Davila (CU Boulder)<br />
|v34e<br />
|Approved 12 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Added_support_for_off-diagonal_covariance | Added support for off-diagonal covariance]]<br />
|Adjoint Model Update<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34f<br />
|Approved 4 Apr 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Nested_Transport_Fix | Nested Transport Fix]]<br />
|Forward Model Bug Fix<br />
|Lin Zhang (Harvard)<br />
|Yanko Davila (CU Boulder)<br />
|v34g<br />
|Approved 4 Apr 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Bug_fix_for_mass_transport_flux_diagnostics | Bug fix for mass transport flux diagnostics]]<br />
|Forward Model Bug Fix<br />
|Mat Evans<br />
|Yanko Davila (CU Boulder)<br />
|v34h<br />
|Approved 17 Apr 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Bug_fix_in_streets_anthro_mod.F_for_offline_simulations | Bug fix in streets_anthro_mod.F for offline simulations]]<br />
|Forward Model Bug Fix<br />
|Fabien Paulot<br />
|Yanko Davila (CU Boulder)<br />
|v34h<br />
|Approved 17 Apr 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Added_support_for_ISOROPIA_II | Added support for ISOROPIA II ]]<br />
|Adjoint Model Update<br />
|Shannon Capps (EPA), Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34i<br />
|Approved 19 Apr 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
|-<br />
|[[#Diurnal_NOx_and_regional_biofuel_emissions_fix | Diurnal NOx and regional biofuel emissions fix ]]<br />
|Forward Model Bug Fix<br />
|Christoph Keller<br />
|Yanko Davila (CU Boulder)<br />
|v34j<br />
|Approved 14 May 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
|-<br />
|[[#Fix_for_isoprene_nitrates_in_globchem.dat | Fix for isoprene nitrates in globchem.dat ]]<br />
|Forward Model Bug Fix<br />
|Fabien Paulot<br />
|Yanko Davila (CU Boulder)<br />
|v34k<br />
|Approved 15 May 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
-<br />
|[[#Added_support_for_reaction_rate_constants_defined_in_input.gcadj | Added support for reaction rate constants defined in input.gcadj ]]<br />
|Adjoint Model Update<br />
|Hyunmin Lee (CU Boulder), Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34m<br />
|Approved 25 June 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
-<br />
|[[#Updated_GMI_database | Updated GMI database ]]<br />
|Forward Model Update<br />
|Hyunmin Lee (CU Boulder), Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34n<br />
|Approved 3 July 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
|}<br />
<br />
<br />
==== - Updates to forward model ====<br />
<br />
===== Add GFED3 Emissions =====<br />
<br />
Implemented GFDE3 Emissions<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34 tests 005 and 006]<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Added support for ICOADS ship emissions =====<br />
<br />
Added support for ICOADS ship emissions.<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34c test 003]<br />
<br />
--[[User:Yanko|Yanko]] 13:49, 5 February 2013 (EST)<br />
<br />
===== Add support for RETRO emissions =====<br />
<br />
Added support for [[GEOS-Chem_v9-01-02_benchmark_history#v9-01-02k | RETRO emissions.]] <br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34e test 001]<br />
<br />
--[[User:Yanko|Yanko]] 14:14, 14 March 2013 (EDT)<br />
<br />
===== Updated GMI database =====<br />
Now we are using the same GMI data inventory as GEOS-Chem forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Stratospheric_chemistry#New_implementation_of_linearized_stratospheric_chemistry v9-01-03o]. <br />
With the adjoint model, one can calculate the sensitivities with respect to production and loss rates of [http://wiki.seas.harvard.edu/geos-chem/index.php/Stratospheric_chemistry#Original_Species 24 original species].<br />
Note that the GEOS-Chem adjoint does not support Bromine chemistry yet.<br />
<br />
--[[User:Hyung-Min Lee|Hyung-Min Lee]] 19:01, 3 July 2013 (EDT)<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=1 gcadj_std_v34n ]<br />
<br />
--[[User:Yanko|Yanko]] 17:37, 3 July 2013 (EDT)<br />
<br />
==== - Bug fixes in forward model ====<br />
<br />
===== Update to Nested Domains =====<br />
<br />
Improved nested simulations. STRAT_CHEM update for regrid GMI files to nested domain.<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=1 gcadj_std_v34b]<br />
<br />
--[[User:Yanko|Yanko]] 15:53, 25 January 2013 (EST)<br />
<br />
===== Washout fix for non-aerosol species ===== <br />
<br />
The Wet deposition algorithm was underestimating washout of non-aerosol species. [[GEOS-Chem_v9-01-02#Washout_fix_for_non-aerosol_species | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 16:00, 4 March 2013 (EST)<br />
<br />
===== Bug fixes for biofuel emissions ===== <br />
<br />
Fix for double-counting of residential biofuel when using RETRO and fix in biofuel emissions using Streets inventory. [[GEOS-Chem_v9-01-02#Bug_fixes_for_biofuel_emissions | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 19:43, 13 March 2013 (EDT)<br />
<br />
===== Fix bug in EDGAR ship emissions code ===== <br />
<br />
GEOS-Chem default configuration uses the ICOADS ship emissions instead of the EDGAR ship emissions. However, we still pass <tt>EDGAR_CO_SHIP</tt> to routine <tt>COMPUTE_EDGAR_CO</tt> even if it is not allocated. [[Ship_emissions#Potential_bug_in_EDGAR_ship_emissions_code | See forward wiki. ]]<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34e test 003 to 008]<br />
<br />
--[[User:Yanko|Yanko]] 19:43, 13 March 2013 (EDT)<br />
<br />
===== Bug fix for ship NOx emissions ===== <br />
<br />
Ship NOx emissions are supposed to be emitted as (1 mol HNO3 + 10 mol O3) per mol NOx in the inventory. The code previously used molar masses incorrectly so that the ratios were actually 0.73 mol HNO3 and 9.5 mol O3. Thus, this bug fix will slightly increase O3 and HNO3 over the oceans. [[Ship_emissions#Bug_in_ship_NOx_emissions | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 19:43, 13 March 2013 (EDT)<br />
<br />
===== Nested Transport Fix ===== <br />
<br />
Applied a bug fix that affects transport for nested simulations with GEOS-5 met fields. [[Setting_up_GEOS-Chem_nested_grid_simulations#Transport_fix | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 18:59, 4 April 2013 (EDT)<br />
<br />
===== Bug fix for mass transport flux diagnostics ===== <br />
<br />
Applied a bug fix to LD24,LD25and LD25 diagnostics. [[GEOS-Chem_v9-01-03#Bug_fix_for_mass_transport_flux_diagnostics | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 15:27, 17 April 2013 (EDT)<br />
<br />
===== Bug fix in streets_anthro_mod.F for offline simulations ===== <br />
<br />
Applied a bug fix to avoid issues with the offline chemistry where idtnh3 is not equal to 30. [[GEOS-Chem_v9-01-03#Bug_fix_in_streets_anthro_mod.F_for_offline_simulations | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 15:27, 17 April 2013 (EDT)<br />
<br />
===== Diurnal NOx and regional biofuel emissions fix ===== <br />
<br />
Diurnal NOx scale factors should be based on UTC instead of local time. Also we currently double count biofuel emissions over Canada and Mexico if the respective local inventories (CAC and BRAVO) are used. [[GEOS-Chem_v9-02#Bug_fixes_in_biofuel_mod.F_and_emfossil.F | See forward wiki. ]]<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34j test 001]<br />
<br />
--[[User:Yanko|Yanko]] 13:07, 15 May 2013 (EDT)<br />
<br />
===== Fix for isoprene nitrates in globchem.dat ===== <br />
<br />
Fabien Paulot discovered an error in the reaction for isoprene nitrates in globchem.dat for the SOA and standard simulations. [[GEOS-Chem_v8-03-02#Fix_for_isoprene_nitrates_in_globchem.dat | See forward wiki. ]]<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34k test 001]<br />
<br />
--[[User:Yanko|Yanko]] 15:49, 15 May 2013 (EDT)<br />
<br />
==== - Updates to adjoint model ====<br />
<br />
=====Building Process Updated =====<br />
<br />
Building process updated to match forward model. Find more information about changes on the run script on the [http://adjoint.colorado.edu/~yanko/gcadj_std/GC_adj_man.pdf User's Guide Section 2.4 Page 27]<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Checkpoint files creation updated =====<br />
<br />
Now *chk* files are not created at X=0. Checkpoint files are not created at X=2, 3 if performing a finite diference test.<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Population_mod Updated =====<br />
<br />
population_mod.f updated.<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Fix GFED2 Emissions =====<br />
<br />
Fixed an error when using GFED2 8Daily and 3Hr emissions, the error occurred when running backwards because the emissions were called twice.<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34 tests 003 and 004 ]<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Added support for off-diagonal covariance ===== <br />
<br />
- Now possibility to define correlations (off-diagonal terms) in the background (prior) covariance error matrix. User can define the latitude and longitude correlation length in input.gcadj. See Singh et al. (2011): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. ,4, 299-316 for more information. <br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34f test 003 to 005]<br />
<br />
--[[User:Yanko|Yanko]] 17:38, 4 April 2013 (EDT)<br />
<br />
===== Added support for ISOROPIA II ===== <br />
<br />
- The ISORROPIA II package performs aerosol thermodynamical equilibrium. It partitions nitrate (HNO3 and NIT) and ammonia (NH3 and NH4) between the gas and aerosol phases. Inputs to the partitioning routine include temperature and RH. ISORROPIA II has significant benefits over previous implementations of ISORROPIA, especially for partitioning of nitrate at low RH. [[ ISORROPIA_II | Click here more information. ]]<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34i test 003 to 005]<br />
<br />
--[[User:Yanko|Yanko]] 12:19, 22 April 2013 (EDT)<br />
<br />
===== Added support for reaction rate constants defined in input.gcadj =====<br />
<br />
- Reaction rate constants can now be selected as control variables for sensitivity analysis or optimization. The complete set of reaction rate constants, or subsets, can be specified in input.gcadj. <br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34m test 001 to 011]<br />
<br />
==== - Bug fixes in adjoint model ====<br />
<br />
===== M Undefined =====<br />
<br />
M Undefined on line 1199 of tagged_co_adj_mod.f, this only affects tagged_co emissions inversions.<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Error compiling on ifort 12 =====<br />
<br />
Fixed an error that made imposible to compile sciabr_co_obs_mod.f on ifort 12. The error was caused by an openmp pragma. This error is entirely a problem of the compiler and not the code itself, this should me resolved on version 13 of the compiler. Meanwhile we just commented the pragma to avoid problems. <br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Set observation errors to 1% in pseudo-obs test =====<br />
<br />
Setting the observation errors to 1% instead of 100% in the pseudo-obs test.<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=1 gcadj_std_v34a ]<br />
<br />
--[[User:Yanko|Yanko]] 14:08, 18 December 2012 (EST)<br />
<br />
===== Fixed Makefile Bug =====<br />
<br />
Fixed a Makefile bug preventing 'make help' from running at checkout, also added example of IFORT_OPT flag on run script.<br />
<br />
--[[User:Yanko|Yanko]] 15:53, 25 January 2013 (EST)</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v34&diff=13621GEOS-Chem Adjoint v342013-07-03T23:01:48Z<p>Hyung-Min Lee: /* Updated GMI database */</p>
<hr />
<div>== Overview ==<br />
<br />
BETA RELEASE<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v33]]<br />
<br />
=== What's new in this version ===<br />
<br />
GEOS-Chem Adjoint v34 contains the following major updates and improvements. <br />
<br />
{| border=1 cellspacing=0 cellpadding=5 <br />
|- bgcolor="#CCCCCC"<br />
!Feature<br />
!Type<br />
!Submitted by<br />
!Committed by<br />
!Version<br />
!Status<br />
<br />
|-<br />
|[[#Building_Process_Updated |Building Process Updated]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Add_GFED3_Emissions |Add support for GFED3 Emissions]]<br />
|Forward Model Update<br />
|Melissa Payer (Harvard)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Checkpoint_files_creation_updated |Checkpoint files creation updated]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Population_mod_Updated |population-weighted cost function updated]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Fix_GFED2_Emissions |Add support for GFED2 8day and 3hr Emissions]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#M_Undefined |M undefined]]<br />
|Adjoint Bug Fix<br />
|Daven Henze (CU Boulder), Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Error_compiling_on_ifort_12 |Error compiling on ifort 12]]<br />
|Adjoint Bug Fix<br />
|Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34<br />
|Approved 14 Nov 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Set_observation_errors_to_1.25_in_pseudo-obs_test |Set observation errors to 1% in pseudo-obs test]]<br />
|Adjoint Bug Fix<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34a<br />
|Approved 11 Dec 2012<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Update_to_Nested_Domains |Update to Nested Domains]]<br />
|Forward Model Bug Fix<br />
|Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34b<br />
|Approved 25 Jan 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Fixed_Makefile_Bug |Fixed Makefile Bug]]<br />
|Adjoint Bug Fix<br />
|Yanko Davila (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34b<br />
|Approved 25 Jan 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Added_support_for_ICOADS_ship_emissions |Added support for ICOADS ship emissions]]<br />
|Forward Model Update<br />
|Katerina Lapina (CU Boulder), Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34c<br />
|Approved 5 Feb 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Washout_fix_for_non-aerosol_species |Washout fix for non-aerosol species]]<br />
|Forward Model Bug Fix<br />
|Helen Amos<br />
|Yanko Davila (CU Boulder)<br />
|v34d<br />
|Approved 4 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Add_support_for_RETRO_emissions |Add support for RETRO emissions]]<br />
|Forward Model Update<br />
|Wes Reinhart (U. Minnesota), Dylan Millet (U. Minnesota)<br />
|Yanko Davila (CU Boulder)<br />
|v34e<br />
|Approved 12 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Bug_fixes_for_biofuel_emissions |Bug fixes for biofuel emissions]]<br />
|Forward Model Bug Fix<br />
|Melissa Payer<br />
|Yanko Davila (CU Boulder)<br />
|v34e<br />
|Approved 12 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Fix_bug_in_EDGAR_ship_emissions_code | Fix bug in EDGAR ship emissions code]]<br />
|Forward Model Bug Fix<br />
|Michael Long<br />
|Yanko Davila (CU Boulder)<br />
|v34e<br />
|Approved 12 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Bug_fix_for_ship_NOx_emissions | Bug fix for ship NOx emissions]]<br />
|Forward Model Bug Fix<br />
|Chris Holmes<br />
|Yanko Davila (CU Boulder)<br />
|v34e<br />
|Approved 12 Mar 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Added_support_for_off-diagonal_covariance | Added support for off-diagonal covariance]]<br />
|Adjoint Model Update<br />
|Nicolas Bousserez (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34f<br />
|Approved 4 Apr 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Nested_Transport_Fix | Nested Transport Fix]]<br />
|Forward Model Bug Fix<br />
|Lin Zhang (Harvard)<br />
|Yanko Davila (CU Boulder)<br />
|v34g<br />
|Approved 4 Apr 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Bug_fix_for_mass_transport_flux_diagnostics | Bug fix for mass transport flux diagnostics]]<br />
|Forward Model Bug Fix<br />
|Mat Evans<br />
|Yanko Davila (CU Boulder)<br />
|v34h<br />
|Approved 17 Apr 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Bug_fix_in_streets_anthro_mod.F_for_offline_simulations | Bug fix in streets_anthro_mod.F for offline simulations]]<br />
|Forward Model Bug Fix<br />
|Fabien Paulot<br />
|Yanko Davila (CU Boulder)<br />
|v34h<br />
|Approved 17 Apr 2013<br />
|-bgcolor="#CCFFFF"<br />
<br />
|-<br />
|[[#Added_support_for_ISOROPIA_II | Added support for ISOROPIA II ]]<br />
|Adjoint Model Update<br />
|Shannon Capps (EPA), Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34i<br />
|Approved 19 Apr 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
|-<br />
|[[#Diurnal_NOx_and_regional_biofuel_emissions_fix | Diurnal NOx and regional biofuel emissions fix ]]<br />
|Forward Model Bug Fix<br />
|Christoph Keller<br />
|Yanko Davila (CU Boulder)<br />
|v34j<br />
|Approved 14 May 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
|-<br />
|[[#Fix_for_isoprene_nitrates_in_globchem.dat | Fix for isoprene nitrates in globchem.dat ]]<br />
|Forward Model Bug Fix<br />
|Fabien Paulot<br />
|Yanko Davila (CU Boulder)<br />
|v34k<br />
|Approved 15 May 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
-<br />
|[[#Added_support_for_reaction_rate_constants_defined_in_input.gcadj | Added support for reaction rate constants defined in input.gcadj ]]<br />
|Adjoint Model Update<br />
|Hyunmin Lee (CU Boulder), Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34m<br />
|Approved 25 June 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
-<br />
|[[#Updated_GMI_database | Updated GMI database ]]<br />
|Forward Model Update<br />
|Hyunmin Lee (CU Boulder), Daven Henze (CU Boulder)<br />
|Yanko Davila (CU Boulder)<br />
|v34n<br />
|Approved 3 July 2013<br />
|-bgcolor="#CCFFFF<br />
<br />
|}<br />
<br />
<br />
==== - Updates to forward model ====<br />
<br />
===== Add GFED3 Emissions =====<br />
<br />
Implemented GFDE3 Emissions<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34 tests 005 and 006]<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Added support for ICOADS ship emissions =====<br />
<br />
Added support for ICOADS ship emissions.<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34c test 003]<br />
<br />
--[[User:Yanko|Yanko]] 13:49, 5 February 2013 (EST)<br />
<br />
===== Add support for RETRO emissions =====<br />
<br />
Added support for [[GEOS-Chem_v9-01-02_benchmark_history#v9-01-02k | RETRO emissions.]] <br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34e test 001]<br />
<br />
--[[User:Yanko|Yanko]] 14:14, 14 March 2013 (EDT)<br />
<br />
===== Updated GMI database =====<br />
Now we are using the same GMI data inventory as GEOS-Chem forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Stratospheric_chemistry#New_implementation_of_linearized_stratospheric_chemistry v9-03-01o]. <br />
With the adjoint model, one can calculate the sensitivities with respect to production and loss rates of [http://wiki.seas.harvard.edu/geos-chem/index.php/Stratospheric_chemistry#Original_Species 24 original species].<br />
Note that the GEOS-Chem adjoint does not support Bromine chemistry yet.<br />
<br />
--[[User:Hyung-Min Lee|Hyung-Min Lee]] 19:01, 3 July 2013 (EDT)<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=1 gcadj_std_v34n ]<br />
<br />
--[[User:Yanko|Yanko]] 17:37, 3 July 2013 (EDT)<br />
<br />
==== - Bug fixes in forward model ====<br />
<br />
===== Update to Nested Domains =====<br />
<br />
Improved nested simulations. STRAT_CHEM update for regrid GMI files to nested domain.<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=1 gcadj_std_v34b]<br />
<br />
--[[User:Yanko|Yanko]] 15:53, 25 January 2013 (EST)<br />
<br />
===== Washout fix for non-aerosol species ===== <br />
<br />
The Wet deposition algorithm was underestimating washout of non-aerosol species. [[GEOS-Chem_v9-01-02#Washout_fix_for_non-aerosol_species | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 16:00, 4 March 2013 (EST)<br />
<br />
===== Bug fixes for biofuel emissions ===== <br />
<br />
Fix for double-counting of residential biofuel when using RETRO and fix in biofuel emissions using Streets inventory. [[GEOS-Chem_v9-01-02#Bug_fixes_for_biofuel_emissions | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 19:43, 13 March 2013 (EDT)<br />
<br />
===== Fix bug in EDGAR ship emissions code ===== <br />
<br />
GEOS-Chem default configuration uses the ICOADS ship emissions instead of the EDGAR ship emissions. However, we still pass <tt>EDGAR_CO_SHIP</tt> to routine <tt>COMPUTE_EDGAR_CO</tt> even if it is not allocated. [[Ship_emissions#Potential_bug_in_EDGAR_ship_emissions_code | See forward wiki. ]]<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34e test 003 to 008]<br />
<br />
--[[User:Yanko|Yanko]] 19:43, 13 March 2013 (EDT)<br />
<br />
===== Bug fix for ship NOx emissions ===== <br />
<br />
Ship NOx emissions are supposed to be emitted as (1 mol HNO3 + 10 mol O3) per mol NOx in the inventory. The code previously used molar masses incorrectly so that the ratios were actually 0.73 mol HNO3 and 9.5 mol O3. Thus, this bug fix will slightly increase O3 and HNO3 over the oceans. [[Ship_emissions#Bug_in_ship_NOx_emissions | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 19:43, 13 March 2013 (EDT)<br />
<br />
===== Nested Transport Fix ===== <br />
<br />
Applied a bug fix that affects transport for nested simulations with GEOS-5 met fields. [[Setting_up_GEOS-Chem_nested_grid_simulations#Transport_fix | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 18:59, 4 April 2013 (EDT)<br />
<br />
===== Bug fix for mass transport flux diagnostics ===== <br />
<br />
Applied a bug fix to LD24,LD25and LD25 diagnostics. [[GEOS-Chem_v9-01-03#Bug_fix_for_mass_transport_flux_diagnostics | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 15:27, 17 April 2013 (EDT)<br />
<br />
===== Bug fix in streets_anthro_mod.F for offline simulations ===== <br />
<br />
Applied a bug fix to avoid issues with the offline chemistry where idtnh3 is not equal to 30. [[GEOS-Chem_v9-01-03#Bug_fix_in_streets_anthro_mod.F_for_offline_simulations | See forward wiki. ]]<br />
<br />
--[[User:Yanko|Yanko]] 15:27, 17 April 2013 (EDT)<br />
<br />
===== Diurnal NOx and regional biofuel emissions fix ===== <br />
<br />
Diurnal NOx scale factors should be based on UTC instead of local time. Also we currently double count biofuel emissions over Canada and Mexico if the respective local inventories (CAC and BRAVO) are used. [[GEOS-Chem_v9-02#Bug_fixes_in_biofuel_mod.F_and_emfossil.F | See forward wiki. ]]<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34j test 001]<br />
<br />
--[[User:Yanko|Yanko]] 13:07, 15 May 2013 (EDT)<br />
<br />
===== Fix for isoprene nitrates in globchem.dat ===== <br />
<br />
Fabien Paulot discovered an error in the reaction for isoprene nitrates in globchem.dat for the SOA and standard simulations. [[GEOS-Chem_v8-03-02#Fix_for_isoprene_nitrates_in_globchem.dat | See forward wiki. ]]<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34k test 001]<br />
<br />
--[[User:Yanko|Yanko]] 15:49, 15 May 2013 (EDT)<br />
<br />
==== - Updates to adjoint model ====<br />
<br />
=====Building Process Updated =====<br />
<br />
Building process updated to match forward model. Find more information about changes on the run script on the [http://adjoint.colorado.edu/~yanko/gcadj_std/GC_adj_man.pdf User's Guide Section 2.4 Page 27]<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Checkpoint files creation updated =====<br />
<br />
Now *chk* files are not created at X=0. Checkpoint files are not created at X=2, 3 if performing a finite diference test.<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Population_mod Updated =====<br />
<br />
population_mod.f updated.<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Fix GFED2 Emissions =====<br />
<br />
Fixed an error when using GFED2 8Daily and 3Hr emissions, the error occurred when running backwards because the emissions were called twice.<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34 tests 003 and 004 ]<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Added support for off-diagonal covariance ===== <br />
<br />
- Now possibility to define correlations (off-diagonal terms) in the background (prior) covariance error matrix. User can define the latitude and longitude correlation length in input.gcadj. See Singh et al. (2011): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. ,4, 299-316 for more information. <br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34f test 003 to 005]<br />
<br />
--[[User:Yanko|Yanko]] 17:38, 4 April 2013 (EDT)<br />
<br />
===== Added support for ISOROPIA II ===== <br />
<br />
- The ISORROPIA II package performs aerosol thermodynamical equilibrium. It partitions nitrate (HNO3 and NIT) and ammonia (NH3 and NH4) between the gas and aerosol phases. Inputs to the partitioning routine include temperature and RH. ISORROPIA II has significant benefits over previous implementations of ISORROPIA, especially for partitioning of nitrate at low RH. [[ ISORROPIA_II | Click here more information. ]]<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34i test 003 to 005]<br />
<br />
--[[User:Yanko|Yanko]] 12:19, 22 April 2013 (EDT)<br />
<br />
===== Added support for reaction rate constants defined in input.gcadj =====<br />
<br />
- Reaction rate constants can now be selected as control variables for sensitivity analysis or optimization. The complete set of reaction rate constants, or subsets, can be specified in input.gcadj. <br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 gcadj_std_v34m test 001 to 011]<br />
<br />
==== - Bug fixes in adjoint model ====<br />
<br />
===== M Undefined =====<br />
<br />
M Undefined on line 1199 of tagged_co_adj_mod.f, this only affects tagged_co emissions inversions.<br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Error compiling on ifort 12 =====<br />
<br />
Fixed an error that made imposible to compile sciabr_co_obs_mod.f on ifort 12. The error was caused by an openmp pragma. This error is entirely a problem of the compiler and not the code itself, this should me resolved on version 13 of the compiler. Meanwhile we just commented the pragma to avoid problems. <br />
<br />
--[[User:Yanko|Yanko]] 19:06, 14 November 2012 (EST)<br />
<br />
===== Set observation errors to 1% in pseudo-obs test =====<br />
<br />
Setting the observation errors to 1% instead of 100% in the pseudo-obs test.<br />
<br />
See Benchmarks results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=1 gcadj_std_v34a ]<br />
<br />
--[[User:Yanko|Yanko]] 14:08, 18 December 2012 (EST)<br />
<br />
===== Fixed Makefile Bug =====<br />
<br />
Fixed a Makefile bug preventing 'make help' from running at checkout, also added example of IFORT_OPT flag on run script.<br />
<br />
--[[User:Yanko|Yanko]] 15:53, 25 January 2013 (EST)</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint&diff=12632GEOS-Chem Adjoint2013-02-19T23:31:32Z<p>Hyung-Min Lee: /* Current GEOS-Chem Adjoint Research Projects (please add yours!) */</p>
<hr />
<div><big><big><strong>Adjoint and Data Assimilation Working Group</strong></big></big><br />
<br />
== Contact information ==<br />
<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|-<br />
|'''Adjoint Working Group Co-Chairs'''<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman] and [mailto:dbj@atmosp.physics.utoronto.ca Dylan Jones]<br />
|-<br />
|-<br />
|'''Adjoint Model Scientist'''<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|'''Adjoint Co-Model Scientist/GC adjoint support team'''<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez]<br />
|-<br />
|-<br />
|'''GC adjoint support team'''<br />
|[mailto:yanko.davila@colorado.edu Yanko Davila]<br />
|-<br />
|'''Adjoint Working Group email list'''<br />
|<tt>geos-chem-adjoint@seas.harvard.edu</tt><br />
|-<br />
|'''To subscribe to email list'''<br />
|Send email to [mailto:geos-chem-adjoint-join@seas.harvard.edu <tt>geos-chem-adjoint-join@seas.harvard.edu</tt>]<br />
|-<br />
|'''To unsubscribe from email list'''<br />
|Send email to [mailto:geos-chem-adjoint-leave@seas.harvard.edu <tt>geos-chem-adjoint-leave@seas.harvard.edu</tt>]<br />
|}<br />
<br />
== Historical Development ==<br />
Original work on the adjoint of GEOS-Chem v6 began in 2003, focusing on the adjoint of the offline aerosol simulation. By 2005, the adjoint was expanded to include a tagged CO simulation and a full chemistry simulation; an adjoint of GEOS-Chem v7 was also developed in the following years. Each of these branches of the adjoint code were been constructed in a hybrid fashion using a combination of automatic differentiation software ([http://www.autodiff.com/tamc TAMC], [http://people.cs.vt.edu/~asandu/Software/Kpp/ KPP]) and manual coding of both discrete and continuous adjoints. They shared many common elements yet had unique features for different applications. <br />
<br />
During the summer of 2009, the existing branches were merged and updated to bring the adjoint into alignment with the latest release of GEOS-Chem, v8-02-01. This merged adjoint model is now the standard adjoint code into which all further development efforts will be placed.<br />
<br />
== Forward Model Code ==<br />
<br />
The forward model on which the adjoint is based originally corresponded to GEOS-Chem v8-02-01. It was subsequently updated as follows:<br />
<br />
* KPP solver for gas-phase chemistry (as in GCv8-02-03)<br />
* Implement Bond 2007 BC/OC emissions (as in GCv8-02-02)<br />
* Apply bug fixes from GCv8-02-02 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-02#Previous_issues_now_resolved_in_v8-02-02 here]<br />
* Apply bug fixes from GCv8-02-03 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-03#Previous_issues_now_resolved_in_v8-02-03 here]<br />
* Apply bug fixes from GCv8-02-04 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-04#Previous_issues_now_resolved_in_v8-02-04 here]<br />
<br />
All bug fixes and model updates were previous listed at the top of inverse_driver.f. We have now switched to documenting the code development cycle here in the wiki, see the following section. <br />
<br />
== Code Versions, Bug Fixes and Developments == <br />
<br />
=== Current GEOS-Chem adjoint version released ===<br />
[[GEOS-Chem_Adjoint_v34]] (You will download this version when you check out.)<br />
<br />
=== Previous GEOS-Chem adjoint versions released ===<br />
*[[GEOS-Chem_Adjoint_v33]] <br />
*[[GEOS-Chem_Adjoint_v32]]<br />
*[[GEOS-Chem_Adjoint_v31]]<br />
<br />
== Summary of Main Adjoint Code Supported Features ==<br />
<br />
=== Features === <br />
* Meteorological fields<br />
** GEOS-3 <i>needs testing</i><br />
** GEOS-4<br />
** GEOS-5 <br />
* model resolution<br />
** 4 x 5<br />
** 2 x 2.5 <br />
** Nested Asia and NA<br />
* Forward model processes<br />
** convection<br />
** advection<br />
** PBL mixing<br />
** dry deposition<br />
** wet deposition<br />
** strat / trop exchange with LINOZ and new GMI strat chem (v9-01-03)<br />
** NOy up fluxes (now replaced with new GMI strat chem)<br />
** aerosols <br />
*** inorganic aerosol thermodynamics with RPMARES<br />
*** inorganic aerosol thermodynamics with ISORROPIA <i> in progress </i><br />
*** sulfate chemistry <br />
*** BC<br />
*** SOA, Dust, sea salt <i> needs doing </i><br />
*** aerosol surface area feedbacks <i> needs updating </i><br />
*** aerosol optical feedbacks <i> needs doing </i><br />
** emissions<br />
*** all standard emissions included<br />
* Simulation modes<br />
** full chemistry<br />
** tagged CO<br />
** tagged Ox <br />
** CH4<br />
** offline aerosols (for BC and dust only)<br />
** CO2<br />
* Observational Operators<br />
** MOPITT CO column<br />
** SCIAMACHY CO column<br />
** AIRS CO column<br />
** IMPROVE BC<br />
** CASTNet (NH4+) <i> needs updating </i> <br />
** GOME / SCIAMACHY NO2 column <i> needs updating </i> <br />
*** using KNMI retrieval (Henze)<br />
*** using Dalhousie retrieval (Shim)<br />
** SCIAMACHY/OMI NO2<br />
*** using Dalhousie retrieval (Bousserez, Padmanabhan)<br />
** TES NH3 <br />
** TES O3 <br />
** GOSAT CO2 <br />
** MLS O3 and TES CO2 <i> in progress </i><br />
* Control parameters<br />
** Initial Conditions scaling factors (linear or log)<br />
** Emissions scaling factors (linear or log) <br />
*** NH3, primary BC/OC, SO2: anthropogenic, natural, bioburn, biomass, ship<br />
*** NOx: soil, aircraft, anthropogenic, biofuel, bioburn<br />
*** Lightning NOx: injection height, yield <i> in progress </i><br />
*** all other gas-phase tracers: anthropogenic, biofuel, bioburn <br />
* Adjoint sensitivities<br />
** w.r.t. all implemented control parameters<br />
** w.r.t Reaction Rate Parameters <br />
** w.r.t all emissions <br />
** of AQ attainment metrics <i> needs updating </i><br />
** of spatiotemporally averaged species concentrations (e.g., arctic O3) <br />
* Other<br />
** Inverse Hessian approximation <br />
** off-diagonal covariance matrices <i> needs updating </i><br />
** 3D-Var <i> needs updating </i><br />
<br />
Features may be qualified as: <br />
* <i>needs testing</i>: an implemented feature that we haven't fully used yet<br />
* <i>needs updating</i>: a feature developed with a previous branch that has yet to be updated to GEOS-Chem v8 and the merged adjoint<br />
* <i>needs doing</i>: a feature nobody has tackled the adjoint of yet<br />
* <i>in progress</i>: a feature currently under development <br />
* <i>in pipeline</i>: a feature which has been submitted and awaiting integration into the CVS repository<br />
<br />
=== Primary code developers ===<br />
Monika Kopacz, Kumaresh Singh, Changsub Shim, Daven Henze<br />
<br />
=== Adjoint model lead scientist ===<br />
Daven Henze<br />
<br />
<br />
<br />
== Resources ==<br />
<br />
=== User's guide ===<br />
A user's guide v34 is available. [http://adjoint.colorado.edu/~yanko/gcadj_std/GC_adj_man.pdf User's Guide v34] <br />
<br />
Previous version v32-v33 available at [http://adjoint.colorado.edu/~daven/gcadj_std/GC_adj_man.pdf User's Guide v32]<br />
<br />
=== Code flowchart ===<br />
Meemong Lee has created a detailed flowchart of the inverse model code structure. http://adjoint.colorado.edu/~daven/gcadj_std/flowchart.pdf<br />
<br />
=== Plotting tools ===<br />
<br />
Some IDL and MATLAB routines for plotting benchmark results. http://adjoint.colorado.edu/~daven/gcadj_std/tools.tar.gz<br />
<br />
=== Background papers and presentations ===<br />
<br />
Several articles and presentations (including a GC adjoint modeling clinic overview from IGC5) providing background information about adjoints. http://adjoint.colorado.edu/~daven/gcadj_std/adj_articles.tar.gz<br />
<br />
== Distribution and Use ==<br />
Code for the adjoint is distributed through GITLAB, a web interface connected to a GIT server located at adjoint.colorado.edu. You can access GITLAB at http://adjoint.colorado.edu:8080 after your account is created. Here is our [[Quick Start Guide.]] <br />
<br />
Even if your office mate has a copy of the code, the best way to obtain the model is to get an account for yourself and download a version from the repository. So please do not copy code directly from others or pass the code along to third parties. This vastly helps with tracking developments and keeping up with model updates. <br />
<br />
Use of the adjoint model code follows standard practice for GEOS-Chem. It is expected that any developments that come of individual applications based on this community model will eventually be given back to the community by incorporation of new developments into the standard adjoint code. New development should be submitted to Daven Henze for inclusion in the standard adjoint model code.<br />
<br />
Using GIT gives the users the ability to change the code and commit their changes without affecting the main repository hosted at adjoint.colorado.edu. Users can work with their tuned versions of the code and even create their own tags because GIT acts as a local repository. When ready to submit your update to the comunity just send Daven Henze your [[Using_Git_with_GEOS-Chem#Sharing_your_revisions_with_others_.28and_vice_versa.29|patch]] and we'll take care of the rest.<br />
<br />
=== Quick guide to GIT === <br />
As of version 34 we started using git versioning system. We recommend first taking a look at GIT manual to get a general feel for how this tools works (e.g., [http://git-scm.com/documentation GIT Documentation] or [http://www.kernel.org/pub/software/scm/git/docs/git.html GIT Manual Page]). <br />
<br />
<br />
Useful GIT commands: <br />
<br />
Initial download:<br />
git clone ssh://git@adjoint.colorado.edu:2222/gcadj_std.git<br />
<br />
Status of project vs the current repository:<br />
git status<br />
<br />
Check difference of files (differences have colors for easy reading)<br />
git diff --word-diff=color <wildcard> [<wildcard>] <path>/foo_mod.f<br />
<br />
Checkout specific version<br />
git checkout <wildcard><br />
<br />
Replacing a file with the newest version from the repository:<br />
git checkout origin/master -- <path>/foo_mod.f<br />
<br />
Merging changes in a file: ([http://www.kernel.org/pub/software/scm/git/docs/git-merge.html Reference])<br />
git merge -m <wildcard> <br />
<br />
Comitting<br />
git commit -a<br />
<br />
Tagging a version<br />
git tag -a TAGNAME<br />
<br />
Deleting a tag<br />
git tag -d TAGNAME<br />
git push origin :refs/tags/TAGNAME<br />
<br />
Saving changes to repository<br />
git push<br />
<br />
Saving tags to repository<br />
git push --tags<br />
<br />
List the history of a file:<br />
git log -- <path>/foo_mod.f<br />
<br />
Add a file to the repository<br />
git add <file_name><br />
<br />
Delete a file from the repository<br />
git rm <file_name><br />
<br />
Determine current version<br />
git show HEAD [ | grep commit]<br />
<br />
Download remote changes, rewind your local branch, then replays all your changes over the top of your current branch one by one, until you’re all up to date.<br />
git pull --rebase<br />
<br />
There are several wildcards that you can use on git for example:<br/><br />
''"origin/master"'' - Latest version on the repository <br/><br />
''"HEAD"'' - Latest version as of your last download <br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/72773d7b1a7bd757957d2c186acb3dbb107d5323 v33i]"'' - Specific TAG, find all tag names on GitLab<br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/32d5c926ef529c1dfc640c9dec4a925b24db575c 32d5c926e]"'' - Specific COMMIT, find all commit numbers on GitLab <br/><br />
<br />
[[Using_Git_with_GEOS-Chem|Here is the foward model documentation of git.]]<br />
<br />
=== Backward compatibility (CVS) === <br />
<br />
For people using old version of the code we still have active our CVS repository, but note that the latest version on CVS is v33i-patch2. Here you can find our [[Quick Guide to CVS]]<br />
<br />
== Crediting GEOS-Chem adjoint developers ==<br />
<br />
We aim to make distribution of adjoint model code as immediate as possible. A consequence is that many features may not yet be publicly documented. Therefore, giving code developers due credit is of utmost importance. <br />
<br />
Authors of new additions to the standard code should be offered co-authorship on the first round of presentations and publications to come of their development. Features currently falling in this category and their developers are:<br />
* (v32) MOPITT v5 CO observation operator. Developer: Zhe Jiang, University of Toronto.<br />
* (v32) Dust adjoint. Developer: Xiaoguang (Richard) Xu, University of Nebraska Lincoln.<br />
* (v32) Black carbon offline aerosol adjoint. Developer: Yuhao Mao (UCLA).<br />
* (v32) Nested full chemistry adjoint. Developers: Zhe Jiang, University of Toronto; Daven Henze, CU Boulder.<br />
* (v32) Stratospheric production / loss rate sensitivities. Developer: Hyungmin Lee, CU Boulder.<br />
* (v32) CH4 adjoint. Developer: Kevin Wecht, Harvard University<br />
* (v29) LIDORT. Developer: Daven Henze, University of Colorado Boulder. Collaborator: Rob Spurr. <br />
* (v28) CO2 adjoint. Developer: Daven Henze, University of Colorado Boulder. Collaborators: Ray Nassar, Kevin Bowman, Dylan Jones.<br />
<br />
Citation of the appropriate [[GEOS-Chem_Adjoint#Journal_Articles|journal articles]] for mature developments is also encouraged, as well as considering aspects of [http://acmg.seas.harvard.edu/geos/geos_credit.html co-authorship for the forward model]. <br />
<br />
Overall, if you have any questions about authorship, even for a conference presentation, please contact Daven Henze.<br />
<br />
== Current GEOS-Chem Adjoint Research Projects (please add yours!) ==<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|- bgcolor="#cccccc"<br />
!User Group <br />
!Description <br />
!Contact Person<br />
|-<br />
|CU Boulder<br />
|Aerosol precursors, CO2, O3; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|CU Boulder<br />
|Inverse modeling/optimization; CO2 fluxes inversion (CMS project); general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Nicolas Bousserez]<br />
|-<br />
|CU Boulder<br />
|Sensitivity of nitrate deposition over Antarctica including stratospheric tracers<br />
|[mailto:hyungmin.lee@colorado.edu Hyung-Min Lee]<br />
|-<br />
|Harvard<br />
|Methane<br />
|Kevin Wecht, wecht [at] fas.harvard.edu<br />
|-<br />
|Harvard<br />
|Smoke Emissions in SE Asia<br />
|Patrick Kim, kim68 [at] fas.harvard.edu<br />
|-<br />
|Purdue University<br />
|Methane (SICAMACHY, AIRS and IASI)<br />
|[mailto:tang16@purdue.edu Jinyun Tang]<br />
|-<br />
|MIT<br />
|Aircraft emissions<br />
|[mailto:jaminkoo@mit.edu Jamin Koo]<br />
|-<br />
|MIT<br />
|Air quality, aircraft emissions and sensitivities<br />
|[mailto:bvconsta@mit.edu Bogdan V. Constantin]<br />
|-<br />
|Princeton<br />
|BC sensitivities, general adjoint code development<br />
|Monika Kopacz, mkopacz [at] princeton.edu<br />
|-<br />
|Dalhousie University<br />
|Lightning NOx emissions and impact on tropical ozone using the adjoint<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez (now at CU-Boulder)]<br />
|-<br />
|Dalhousie University<br />
|Surface NOx emissions inversion using SCIAMACHY/OMI NO2 measurements<br />
|Akhila Padmanabhan akhila [at] dal.ca; Nicolas Bousserez [mailto:Nicolas.Bousserez@colorado.edu] (now at CU-Boulder)<br />
|-<br />
|JPL<br />
|Microwave Limb Sounder (MLS) Ozone assimilation<br />
|[mailto:meemong.lee@jpl.nasa.gov Meemong Lee]<br />
|-<br />
|JPL <br />
| TES ozone assimilation/attribution of ozone radiative forcing<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|University of Edinburgh<br />
|Quantifying the impact of boreal forest fires on tropospheric oxidants over the Atlantic<br />
|[mailto:mark.parrington@ed.ac.uk Mark Parrington]<br />
|-<br />
|US EPA<br />
|Integration with economic models for future emission inventory scenario development<br />
|[mailto:akhtar.farhan@epa.gov Farhan Akhtar]<br />
|-<br />
|Peking University<br />
|Satellite constraints on VOC emissions <br />
|[mailto:tmfu@pku.edu.cn May Fu]<br />
|-<br />
|CU Boulder<br />
|Spatial extent of source influences on aerosol precursor columns<br />
|[mailto:alexander.turner@colorado.edu Alex Turner]<br />
|-<br />
|IAP.CAS<br />
|CO2 assimilation<br />
|[mailto:czh@mail.iap.ac.cn Chen]<br />
|-<br />
|Purdue University<br />
|Feedback between terrestrial ecosystem processes and atmospheric co2 signals<br />
|[mailto:zhuq@purdue.edu Qing Zhu]<br />
|-<br />
|Purdue University<br />
|Feedback between aquatic ecosystem processes and atmospheric CH4 signals<br />
|[mailto:tan80@purdue.edu Zeli Tan]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and reactive nitrogen to precursor emissions<br />
|[mailto:twalker@atmosp.physics.utoronto.ca Thomas Walker]<br />
|-<br />
|University of Toronto<br />
|Adjoint analysis for carbon monoxide <br />
|[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and CO to precursor emissions<br />
|[mailto:cwhaley@atmosp.physics.utoronto.ca Cynthia Whaley]<br />
|-<br />
|Georgia Tech<br />
|ISORROPIA adjoint development; inorganic aerosol precursors<br />
|[mailto:scapps@gatech.edu Shannon Capps]<br />
|-<br />
|Peking University<br />
|Source attributions of tropospheric ozone over North China<br />
|[mailto:linjt@pku.edu.cn Jintai Lin]<br />
|-<br />
|University of Wollongong<br />
|Sensitivity of ozone and adjoint analysis of CO over Australasia.<br />
|[mailto:rb864@uowmail.edu.au Rebecca Buchholz]<br />
|-<br />
|University of Toronto<br />
|CO2 assimilation & transport model bias estimation<br />
|[mailto:mkeller@atmosp.physics.utoronto.ca Martin Keller]<br />
|-<br />
|University of Wisconsin<br />
|CO2 assimilation and forecast & Temperature profile retrieval <br />
|[mailto:wenguang.bai@ssec.wisc.edu Wenguang Bai]<br />
|-<br />
|Dalhousie University<br />
|Sensitivity of global PM2.5-induced mortality to emissions<br />
|[mailto:colin.lee@dal.ca Colin Lee]<br />
|-<br />
|University of Leicester (UK)<br />
|Top-down estimates of Amazon isoprene emissions<br />
|[mailto:mpb14@le.ac.uk Michael Barkley]<br />
|-<br />
|Anyang University<br />
|Aerosol emission modeling in East Asia <br />
|[mailto:koo@anyang.ac.kr Youn Seo Koo]<br />
|-<br />
|Nanjing University<br />
|Inverse modeling of terrestrial ecosystem carbon flux<br />
|[mailto:hengmao.wang@gmail.com Hengmao Wang]<br />
|-<br />
|Tsinghua University<br />
|Nested-gird simulations with the adjoint model<br />
|[mailto:nany12@mails.tsinghua.edu.cn Nan Yang]<br />
|-<br />
|University of Minnesota<br />
|Inverse modeling of VOC sources based on TES and IASI measurements<br />
|[mailto:dbm@umn.edu Dylan Millet]<br />
|-<br />
|Tsinghua University<br />
|Inverse modeling of anthropogenic emissions over East Asia<br />
|[mailto:qiangzhang@tsinghua.edu.cn Qiang Zhang]<br />
|-<br />
|UCLA<br />
|Constrain black carbon emission<br />
|[mailto:qiling@atmos.ucla.edu Ling Qi]<br />
|}<br />
<br />
== Publications ==<br />
<br />
=== Journal Articles ===<br />
*Wang, J., X. Xu, D. K. Henze, Q. Ji, S.-C. Tsay, J. Huang, Top-Down Estimate of Dust Emissions through Integration of MODIS and MISR Aerosol Retrievals with the GEOS-Chem adjoint model, submitted.<br />
<br />
*Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden (2012), The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098<br />
<br />
*Bowman, K. W., and D. K. Henze, Attribution of direct ozone radiative forcing to spatially-resolved emissions, submitted.<br />
<br />
*Paulot, F., D. K. Henze, and P. O. Wennberg (2012), Impact of the isoprene photochemical cascade on tropical ozone, Atmos. Chem. Phys., 12, 1307-1325, 2012.<br />
<br />
*Henze, D. K., D. T. Shindell, F. Akhtar, R. J. D. Spurr, R. W. Pinder, D. Loughlin, M. Kopacz, K. Singh, and C. Shim, Spatially refined aerosol direct radiative forcing efficiencies, submitted.<br />
<br />
*Turner, A., D. K. Henze, R. V. Martin, and A. Hakami (2012), The spatial extent of source influences on modeled column concentrations of short-lived species, ''Geophys. Res. Lett.'', doi: 10.1029/2012GL051832, in press.<br />
<br />
*Jiang, Z., D. B. A. Jones, H. M. Worden, M. N. Deeter, D. K. Henze, J. Worden, and K. W. Bowman, Quantifying the impact of model biases in convective transport on inferred CO source estimates using multi-spectral CO retrievals from MOPITT, submitted.<br />
<br />
*Wecht, K. J., D. J. Jacob, S. C. Wofsy, E. A. Kort, J. R. Worden, S. S. Kulawik, D. K. Henze, M. Kopacz, and V. H. Payne, Validation of TES methane with HIPPO aircraft observations: implications for inverse modeling of methane sources, Atmos. Chem. Phys. Discuss., 11, 27887-27911.<br />
<br />
*Walker, T., D. B. A. Jones, M. Parrington, D. K. Henze, L. T. Murray, J. W. Bottenheim, K. Anlauf, J. R. Worden, K. W. Bowman, C. Shim, K. Singh, M. Kopacz, D. W. Tarasick, J. Davies, P. von der Gathen, and C. C. Carouge (2012), Impacts of midlatitude precursor emissions and local photochemistry on ozone abundances in the Arctic, J. Geophys. Res., doi:10.1029/2011JD016370.<br />
<br />
*Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald (2011), Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
<br />
*Singh, K., A. Sandu, Variational Chemical Data Assimilation with Approximate Adjoints, submitted. <br />
<br />
*Singh, K., Jardak, M., Sandu, A., Bowman, K., Lee, M., and Jones, D. (2010): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. Discuss., 3, 1783-1827, doi:10.5194/gmdd-3-1783-2010. http://www.geosci-model-dev-discuss.net/3/1783/2010/gmdd-3-1783-2010.html<br />
<br />
*Kopacz, M., D. L. Mauzerall, J. Wang, E. M. Leibensperger, D. K. Henze, and K. Singh (2011), Origin and radiative forcing of black carbon transported to the Himalayas and Tibetan Plateau, Atmos. Chem. Phys., 11, 2837-2852, 2011. http://www.atmos-chem-phys.net/11/2837/2011/acp-11-2837-2011.html<br />
<br />
* Kopacz, M., D.J. Jacob, J.A. Fisher, J. A. Logan, L. Zhang, I. A. Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010): Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atmos. Chem. Phys., 10, 855-876. http://www.atmos-chem-phys.net/10/855/2010/acp-10-855-2010.pdf<br />
<br />
* Kopacz, M., D. J. Jacob, D. K. Henze, C. L. Heald, D. G. Streets, and Q. Zhang (2009), A comparison of analytical and adjoint Bayesian inversion methods for constraining Asian sources of CO using satellite (MOPITT) measurements of CO columns, J. Geophys. Res., doi:0.1029/2007JD009264. http://acmg.seas.harvard.edu/publications/KopaczJGR2009_2007JD009264.pdf<br />
<br />
* Henze, D. K., J. H. Seinfeld and D. T. Shindell (2009), Inverse modeling and mapping U.S. air quality influences of inorganic PM2.5 precursor emissions with the adjoint of GEOS-Chem, Atmos. Chem. Phys., 9, 5877-5903.<br />
<br />
* Zhang, L., D. J. Jacob, M. Kopacz, D. K. Henze, K. Singh, and D. A. Jaffe (2009), Intercontinental source attribution of ozone pollution at western U.S. sites using an adjoint method, Geophys. Res. Lett., 36, L11810, doi:10.1029/2009GL037950<br />
<br />
* Henze, D. K., A. Hakami and J. H. Seinfeld (2007), Development of the adjoint of GEOS-Chem, Atmos. Chem. Phys., 7, 2413-2433.<br />
<br />
=== Conference proceedings === <br />
* Adjoint inversion of CO sources using combined MOPITT, SCIAMACHY and AIRS CO columns, presented by Monika Kopacz at the COSPAR Scientific Assembly, Montreal, July 18, 2008. http://acmg.seas.harvard.edu/presentations/powerpoints/mak2008/COSPAR_MKopacz_July2008.ppt<br />
<br />
* Singh, K., P. Eller, A. Sandu, D. K. Henze, K. Bowman, M. Kopacz, and M. Lee (2009), Towards the construction of a standard geos-chem adjoint model, ACM High Performance Computing Conference.<br />
<br />
* Kopacz, M., Mauzerall, D.L., Leibensperger, E.M., Wang, J., Henze, D.K., Singh, K., Shim, C. Identifying the origin and estimating the radiative forcing of BC in the Himalayas: an analysis using the global GEOS-Chem adjoint model, European Geophysical Union meeting, Vienna, May 4, 2010.<br />
<br />
* Kopacz, M., Jacob, D.J., Fisher, J.A., Logan, J.A., Zhang, L., Megretskaia, I.A., Yantosca, R.M., Singh, K., Henze, D.K., Burrows, J.P., Buchwitz, M., Khlystova, I., McMillan, W.W., Gille, J.C., Edwards, D.P., Eldering, A., Thouret, V., Nedelec, P. Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), European Geophysical Union meeting, Vienna, May 7, 2010.<br />
<br />
* Tang, J., Zhuang, Q. and Xiong, X. (2010), 4D-Var inversion of atmospheric methane fluxes by assimilating SCIAMACHY and AIRS satellite retrievals, AGU, Dec. 18, 2010, http:/web.ics.purdue.edu/~tang16/agu2010_tang.ppt<br />
<br />
* Bousserez, N., R. V. Martin, K. W. Bowman, D.K. Henze, M. Kopacz, K. Singh, C. Shim, C. Wespes, Improving the lightning NOx source using satellite observations: a 4D-var analysis approach, AGU, Dec., 2010, http://myweb.dal.ca/nc689777/AGU_liNOx_poster_final.pdf</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint&diff=12631GEOS-Chem Adjoint2013-02-19T23:29:02Z<p>Hyung-Min Lee: /* Current GEOS-Chem Adjoint Research Projects (please add yours!) */</p>
<hr />
<div><big><big><strong>Adjoint and Data Assimilation Working Group</strong></big></big><br />
<br />
== Contact information ==<br />
<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|-<br />
|'''Adjoint Working Group Co-Chairs'''<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman] and [mailto:dbj@atmosp.physics.utoronto.ca Dylan Jones]<br />
|-<br />
|-<br />
|'''Adjoint Model Scientist'''<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|'''Adjoint Co-Model Scientist/GC adjoint support team'''<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez]<br />
|-<br />
|-<br />
|'''GC adjoint support team'''<br />
|[mailto:yanko.davila@colorado.edu Yanko Davila]<br />
|-<br />
|'''Adjoint Working Group email list'''<br />
|<tt>geos-chem-adjoint@seas.harvard.edu</tt><br />
|-<br />
|'''To subscribe to email list'''<br />
|Send email to [mailto:geos-chem-adjoint-join@seas.harvard.edu <tt>geos-chem-adjoint-join@seas.harvard.edu</tt>]<br />
|-<br />
|'''To unsubscribe from email list'''<br />
|Send email to [mailto:geos-chem-adjoint-leave@seas.harvard.edu <tt>geos-chem-adjoint-leave@seas.harvard.edu</tt>]<br />
|}<br />
<br />
== Historical Development ==<br />
Original work on the adjoint of GEOS-Chem v6 began in 2003, focusing on the adjoint of the offline aerosol simulation. By 2005, the adjoint was expanded to include a tagged CO simulation and a full chemistry simulation; an adjoint of GEOS-Chem v7 was also developed in the following years. Each of these branches of the adjoint code were been constructed in a hybrid fashion using a combination of automatic differentiation software ([http://www.autodiff.com/tamc TAMC], [http://people.cs.vt.edu/~asandu/Software/Kpp/ KPP]) and manual coding of both discrete and continuous adjoints. They shared many common elements yet had unique features for different applications. <br />
<br />
During the summer of 2009, the existing branches were merged and updated to bring the adjoint into alignment with the latest release of GEOS-Chem, v8-02-01. This merged adjoint model is now the standard adjoint code into which all further development efforts will be placed.<br />
<br />
== Forward Model Code ==<br />
<br />
The forward model on which the adjoint is based originally corresponded to GEOS-Chem v8-02-01. It was subsequently updated as follows:<br />
<br />
* KPP solver for gas-phase chemistry (as in GCv8-02-03)<br />
* Implement Bond 2007 BC/OC emissions (as in GCv8-02-02)<br />
* Apply bug fixes from GCv8-02-02 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-02#Previous_issues_now_resolved_in_v8-02-02 here]<br />
* Apply bug fixes from GCv8-02-03 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-03#Previous_issues_now_resolved_in_v8-02-03 here]<br />
* Apply bug fixes from GCv8-02-04 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-04#Previous_issues_now_resolved_in_v8-02-04 here]<br />
<br />
All bug fixes and model updates were previous listed at the top of inverse_driver.f. We have now switched to documenting the code development cycle here in the wiki, see the following section. <br />
<br />
== Code Versions, Bug Fixes and Developments == <br />
<br />
=== Current GEOS-Chem adjoint version released ===<br />
[[GEOS-Chem_Adjoint_v34]] (You will download this version when you check out.)<br />
<br />
=== Previous GEOS-Chem adjoint versions released ===<br />
*[[GEOS-Chem_Adjoint_v33]] <br />
*[[GEOS-Chem_Adjoint_v32]]<br />
*[[GEOS-Chem_Adjoint_v31]]<br />
<br />
== Summary of Main Adjoint Code Supported Features ==<br />
<br />
=== Features === <br />
* Meteorological fields<br />
** GEOS-3 <i>needs testing</i><br />
** GEOS-4<br />
** GEOS-5 <br />
* model resolution<br />
** 4 x 5<br />
** 2 x 2.5 <br />
** Nested Asia and NA<br />
* Forward model processes<br />
** convection<br />
** advection<br />
** PBL mixing<br />
** dry deposition<br />
** wet deposition<br />
** strat / trop exchange with LINOZ and new GMI strat chem (v9-01-03)<br />
** NOy up fluxes (now replaced with new GMI strat chem)<br />
** aerosols <br />
*** inorganic aerosol thermodynamics with RPMARES<br />
*** inorganic aerosol thermodynamics with ISORROPIA <i> in progress </i><br />
*** sulfate chemistry <br />
*** BC<br />
*** SOA, Dust, sea salt <i> needs doing </i><br />
*** aerosol surface area feedbacks <i> needs updating </i><br />
*** aerosol optical feedbacks <i> needs doing </i><br />
** emissions<br />
*** all standard emissions included<br />
* Simulation modes<br />
** full chemistry<br />
** tagged CO<br />
** tagged Ox <br />
** CH4<br />
** offline aerosols (for BC and dust only)<br />
** CO2<br />
* Observational Operators<br />
** MOPITT CO column<br />
** SCIAMACHY CO column<br />
** AIRS CO column<br />
** IMPROVE BC<br />
** CASTNet (NH4+) <i> needs updating </i> <br />
** GOME / SCIAMACHY NO2 column <i> needs updating </i> <br />
*** using KNMI retrieval (Henze)<br />
*** using Dalhousie retrieval (Shim)<br />
** SCIAMACHY/OMI NO2<br />
*** using Dalhousie retrieval (Bousserez, Padmanabhan)<br />
** TES NH3 <br />
** TES O3 <br />
** GOSAT CO2 <br />
** MLS O3 and TES CO2 <i> in progress </i><br />
* Control parameters<br />
** Initial Conditions scaling factors (linear or log)<br />
** Emissions scaling factors (linear or log) <br />
*** NH3, primary BC/OC, SO2: anthropogenic, natural, bioburn, biomass, ship<br />
*** NOx: soil, aircraft, anthropogenic, biofuel, bioburn<br />
*** Lightning NOx: injection height, yield <i> in progress </i><br />
*** all other gas-phase tracers: anthropogenic, biofuel, bioburn <br />
* Adjoint sensitivities<br />
** w.r.t. all implemented control parameters<br />
** w.r.t Reaction Rate Parameters <br />
** w.r.t all emissions <br />
** of AQ attainment metrics <i> needs updating </i><br />
** of spatiotemporally averaged species concentrations (e.g., arctic O3) <br />
* Other<br />
** Inverse Hessian approximation <br />
** off-diagonal covariance matrices <i> needs updating </i><br />
** 3D-Var <i> needs updating </i><br />
<br />
Features may be qualified as: <br />
* <i>needs testing</i>: an implemented feature that we haven't fully used yet<br />
* <i>needs updating</i>: a feature developed with a previous branch that has yet to be updated to GEOS-Chem v8 and the merged adjoint<br />
* <i>needs doing</i>: a feature nobody has tackled the adjoint of yet<br />
* <i>in progress</i>: a feature currently under development <br />
* <i>in pipeline</i>: a feature which has been submitted and awaiting integration into the CVS repository<br />
<br />
=== Primary code developers ===<br />
Monika Kopacz, Kumaresh Singh, Changsub Shim, Daven Henze<br />
<br />
=== Adjoint model lead scientist ===<br />
Daven Henze<br />
<br />
<br />
<br />
== Resources ==<br />
<br />
=== User's guide ===<br />
A user's guide v34 is available. [http://adjoint.colorado.edu/~yanko/gcadj_std/GC_adj_man.pdf User's Guide v34] <br />
<br />
Previous version v32-v33 available at [http://adjoint.colorado.edu/~daven/gcadj_std/GC_adj_man.pdf User's Guide v32]<br />
<br />
=== Code flowchart ===<br />
Meemong Lee has created a detailed flowchart of the inverse model code structure. http://adjoint.colorado.edu/~daven/gcadj_std/flowchart.pdf<br />
<br />
=== Plotting tools ===<br />
<br />
Some IDL and MATLAB routines for plotting benchmark results. http://adjoint.colorado.edu/~daven/gcadj_std/tools.tar.gz<br />
<br />
=== Background papers and presentations ===<br />
<br />
Several articles and presentations (including a GC adjoint modeling clinic overview from IGC5) providing background information about adjoints. http://adjoint.colorado.edu/~daven/gcadj_std/adj_articles.tar.gz<br />
<br />
== Distribution and Use ==<br />
Code for the adjoint is distributed through GITLAB, a web interface connected to a GIT server located at adjoint.colorado.edu. You can access GITLAB at http://adjoint.colorado.edu:8080 after your account is created. Here is our [[Quick Start Guide.]] <br />
<br />
Even if your office mate has a copy of the code, the best way to obtain the model is to get an account for yourself and download a version from the repository. So please do not copy code directly from others or pass the code along to third parties. This vastly helps with tracking developments and keeping up with model updates. <br />
<br />
Use of the adjoint model code follows standard practice for GEOS-Chem. It is expected that any developments that come of individual applications based on this community model will eventually be given back to the community by incorporation of new developments into the standard adjoint code. New development should be submitted to Daven Henze for inclusion in the standard adjoint model code.<br />
<br />
Using GIT gives the users the ability to change the code and commit their changes without affecting the main repository hosted at adjoint.colorado.edu. Users can work with their tuned versions of the code and even create their own tags because GIT acts as a local repository. When ready to submit your update to the comunity just send Daven Henze your [[Using_Git_with_GEOS-Chem#Sharing_your_revisions_with_others_.28and_vice_versa.29|patch]] and we'll take care of the rest.<br />
<br />
=== Quick guide to GIT === <br />
As of version 34 we started using git versioning system. We recommend first taking a look at GIT manual to get a general feel for how this tools works (e.g., [http://git-scm.com/documentation GIT Documentation] or [http://www.kernel.org/pub/software/scm/git/docs/git.html GIT Manual Page]). <br />
<br />
<br />
Useful GIT commands: <br />
<br />
Initial download:<br />
git clone ssh://git@adjoint.colorado.edu:2222/gcadj_std.git<br />
<br />
Status of project vs the current repository:<br />
git status<br />
<br />
Check difference of files (differences have colors for easy reading)<br />
git diff --word-diff=color <wildcard> [<wildcard>] <path>/foo_mod.f<br />
<br />
Checkout specific version<br />
git checkout <wildcard><br />
<br />
Replacing a file with the newest version from the repository:<br />
git checkout origin/master -- <path>/foo_mod.f<br />
<br />
Merging changes in a file: ([http://www.kernel.org/pub/software/scm/git/docs/git-merge.html Reference])<br />
git merge -m <wildcard> <br />
<br />
Comitting<br />
git commit -a<br />
<br />
Tagging a version<br />
git tag -a TAGNAME<br />
<br />
Deleting a tag<br />
git tag -d TAGNAME<br />
git push origin :refs/tags/TAGNAME<br />
<br />
Saving changes to repository<br />
git push<br />
<br />
Saving tags to repository<br />
git push --tags<br />
<br />
List the history of a file:<br />
git log -- <path>/foo_mod.f<br />
<br />
Add a file to the repository<br />
git add <file_name><br />
<br />
Delete a file from the repository<br />
git rm <file_name><br />
<br />
Determine current version<br />
git show HEAD [ | grep commit]<br />
<br />
Download remote changes, rewind your local branch, then replays all your changes over the top of your current branch one by one, until you’re all up to date.<br />
git pull --rebase<br />
<br />
There are several wildcards that you can use on git for example:<br/><br />
''"origin/master"'' - Latest version on the repository <br/><br />
''"HEAD"'' - Latest version as of your last download <br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/72773d7b1a7bd757957d2c186acb3dbb107d5323 v33i]"'' - Specific TAG, find all tag names on GitLab<br/><br />
''"[http://adjoint.colorado.edu:8080/gcadj_std/commits/32d5c926ef529c1dfc640c9dec4a925b24db575c 32d5c926e]"'' - Specific COMMIT, find all commit numbers on GitLab <br/><br />
<br />
[[Using_Git_with_GEOS-Chem|Here is the foward model documentation of git.]]<br />
<br />
=== Backward compatibility (CVS) === <br />
<br />
For people using old version of the code we still have active our CVS repository, but note that the latest version on CVS is v33i-patch2. Here you can find our [[Quick Guide to CVS]]<br />
<br />
== Crediting GEOS-Chem adjoint developers ==<br />
<br />
We aim to make distribution of adjoint model code as immediate as possible. A consequence is that many features may not yet be publicly documented. Therefore, giving code developers due credit is of utmost importance. <br />
<br />
Authors of new additions to the standard code should be offered co-authorship on the first round of presentations and publications to come of their development. Features currently falling in this category and their developers are:<br />
* (v32) MOPITT v5 CO observation operator. Developer: Zhe Jiang, University of Toronto.<br />
* (v32) Dust adjoint. Developer: Xiaoguang (Richard) Xu, University of Nebraska Lincoln.<br />
* (v32) Black carbon offline aerosol adjoint. Developer: Yuhao Mao (UCLA).<br />
* (v32) Nested full chemistry adjoint. Developers: Zhe Jiang, University of Toronto; Daven Henze, CU Boulder.<br />
* (v32) Stratospheric production / loss rate sensitivities. Developer: Hyungmin Lee, CU Boulder.<br />
* (v32) CH4 adjoint. Developer: Kevin Wecht, Harvard University<br />
* (v29) LIDORT. Developer: Daven Henze, University of Colorado Boulder. Collaborator: Rob Spurr. <br />
* (v28) CO2 adjoint. Developer: Daven Henze, University of Colorado Boulder. Collaborators: Ray Nassar, Kevin Bowman, Dylan Jones.<br />
<br />
Citation of the appropriate [[GEOS-Chem_Adjoint#Journal_Articles|journal articles]] for mature developments is also encouraged, as well as considering aspects of [http://acmg.seas.harvard.edu/geos/geos_credit.html co-authorship for the forward model]. <br />
<br />
Overall, if you have any questions about authorship, even for a conference presentation, please contact Daven Henze.<br />
<br />
== Current GEOS-Chem Adjoint Research Projects (please add yours!) ==<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|- bgcolor="#cccccc"<br />
!User Group <br />
!Description <br />
!Contact Person<br />
|-<br />
|CU Boulder<br />
|Aerosol precursors, CO2, O3; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|CU Boulder<br />
|Inverse modeling/optimization; CO2 fluxes inversion (CMS project); general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Nicolas Bousserez]<br />
|-<br />
|Harvard<br />
|Methane<br />
|Kevin Wecht, wecht [at] fas.harvard.edu<br />
|-<br />
|-<br />
|Harvard<br />
|Smoke Emissions in SE Asia<br />
|Patrick Kim, kim68 [at] fas.harvard.edu<br />
|-<br />
|Purdue University<br />
|Methane (SICAMACHY, AIRS and IASI)<br />
|[mailto:tang16@purdue.edu Jinyun Tang]<br />
|-<br />
|MIT<br />
|Aircraft emissions<br />
|[mailto:jaminkoo@mit.edu Jamin Koo]<br />
|-<br />
|MIT<br />
|Air quality, aircraft emissions and sensitivities<br />
|[mailto:bvconsta@mit.edu Bogdan V. Constantin]<br />
|-<br />
|Princeton<br />
|BC sensitivities, general adjoint code development<br />
|Monika Kopacz, mkopacz [at] princeton.edu<br />
|-<br />
|Dalhousie University<br />
|Lightning NOx emissions and impact on tropical ozone using the adjoint<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez (now at CU-Boulder)]<br />
|-<br />
|Dalhousie University<br />
|Surface NOx emissions inversion using SCIAMACHY/OMI NO2 measurements<br />
|Akhila Padmanabhan akhila [at] dal.ca; Nicolas Bousserez [mailto:Nicolas.Bousserez@colorado.edu] (now at CU-Boulder)<br />
|-<br />
|JPL<br />
|Microwave Limb Sounder (MLS) Ozone assimilation<br />
|[mailto:meemong.lee@jpl.nasa.gov Meemong Lee]<br />
|-<br />
|JPL <br />
| TES ozone assimilation/attribution of ozone radiative forcing<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|University of Edinburgh<br />
|Quantifying the impact of boreal forest fires on tropospheric oxidants over the Atlantic<br />
|[mailto:mark.parrington@ed.ac.uk Mark Parrington]<br />
|-<br />
|US EPA<br />
|Integration with economic models for future emission inventory scenario development<br />
|[mailto:akhtar.farhan@epa.gov Farhan Akhtar]<br />
|-<br />
|Peking University<br />
|Satellite constraints on VOC emissions <br />
|[mailto:tmfu@pku.edu.cn May Fu]<br />
|-<br />
|CU Boulder<br />
|Spatial extent of source influences on aerosol precursor columns<br />
|[mailto:alexander.turner@colorado.edu Alex Turner]<br />
|-<br />
|IAP.CAS<br />
|CO2 assimilation<br />
|[mailto:czh@mail.iap.ac.cn Chen]<br />
|-<br />
|Purdue University<br />
|Feedback between terrestrial ecosystem processes and atmospheric co2 signals<br />
|[mailto:zhuq@purdue.edu Qing Zhu]<br />
|-<br />
|Purdue University<br />
|Feedback between aquatic ecosystem processes and atmospheric CH4 signals<br />
|[mailto:tan80@purdue.edu Zeli Tan]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and reactive nitrogen to precursor emissions<br />
|[mailto:twalker@atmosp.physics.utoronto.ca Thomas Walker]<br />
|-<br />
|University of Toronto<br />
|Adjoint analysis for carbon monoxide <br />
|[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and CO to precursor emissions<br />
|[mailto:cwhaley@atmosp.physics.utoronto.ca Cynthia Whaley]<br />
|-<br />
|Georgia Tech<br />
|ISORROPIA adjoint development; inorganic aerosol precursors<br />
|[mailto:scapps@gatech.edu Shannon Capps]<br />
|-<br />
|Peking University<br />
|Source attributions of tropospheric ozone over North China<br />
|[mailto:linjt@pku.edu.cn Jintai Lin]<br />
|-<br />
|University of Wollongong<br />
|Sensitivity of ozone and adjoint analysis of CO over Australasia.<br />
|[mailto:rb864@uowmail.edu.au Rebecca Buchholz]<br />
|-<br />
|University of Toronto<br />
|CO2 assimilation & transport model bias estimation<br />
|[mailto:mkeller@atmosp.physics.utoronto.ca Martin Keller]<br />
|-<br />
|University of Wisconsin<br />
|CO2 assimilation and forecast & Temperature profile retrieval <br />
|[mailto:wenguang.bai@ssec.wisc.edu Wenguang Bai]<br />
|-<br />
|Dalhousie University<br />
|Sensitivity of global PM2.5-induced mortality to emissions<br />
|[mailto:colin.lee@dal.ca Colin Lee]<br />
|-<br />
|University of Leicester (UK)<br />
|Top-down estimates of Amazon isoprene emissions<br />
|[mailto:mpb14@le.ac.uk Michael Barkley]<br />
|-<br />
|Anyang University<br />
|Aerosol emission modeling in East Asia <br />
|[mailto:koo@anyang.ac.kr Youn Seo Koo]<br />
|-<br />
|Nanjing University<br />
|Inverse modeling of terrestrial ecosystem carbon flux<br />
|[mailto:hengmao.wang@gmail.com Hengmao Wang]<br />
|-<br />
|Tsinghua University<br />
|Nested-gird simulations with the adjoint model<br />
|[mailto:nany12@mails.tsinghua.edu.cn Nan Yang]<br />
|-<br />
|University of Minnesota<br />
|Inverse modeling of VOC sources based on TES and IASI measurements<br />
|[mailto:dbm@umn.edu Dylan Millet]<br />
|-<br />
|Tsinghua University<br />
|Inverse modeling of anthropogenic emissions over East Asia<br />
|[mailto:qiangzhang@tsinghua.edu.cn Qiang Zhang]<br />
|-<br />
|UCLA<br />
|Constrain black carbon emission<br />
|[mailto:qiling@atmos.ucla.edu Ling Qi]<br />
|-<br />
|CU Boulder<br />
|Sensitivity of nitrate deposition over Antarctica including stratospheric tracers<br />
|[mailto:hyungmin.lee@colorado.edu Hyung-Min Lee]<br />
|}<br />
<br />
== Publications ==<br />
<br />
=== Journal Articles ===<br />
*Wang, J., X. Xu, D. K. Henze, Q. Ji, S.-C. Tsay, J. Huang, Top-Down Estimate of Dust Emissions through Integration of MODIS and MISR Aerosol Retrievals with the GEOS-Chem adjoint model, submitted.<br />
<br />
*Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden (2012), The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098<br />
<br />
*Bowman, K. W., and D. K. Henze, Attribution of direct ozone radiative forcing to spatially-resolved emissions, submitted.<br />
<br />
*Paulot, F., D. K. Henze, and P. O. Wennberg (2012), Impact of the isoprene photochemical cascade on tropical ozone, Atmos. Chem. Phys., 12, 1307-1325, 2012.<br />
<br />
*Henze, D. K., D. T. Shindell, F. Akhtar, R. J. D. Spurr, R. W. Pinder, D. Loughlin, M. Kopacz, K. Singh, and C. Shim, Spatially refined aerosol direct radiative forcing efficiencies, submitted.<br />
<br />
*Turner, A., D. K. Henze, R. V. Martin, and A. Hakami (2012), The spatial extent of source influences on modeled column concentrations of short-lived species, ''Geophys. Res. Lett.'', doi: 10.1029/2012GL051832, in press.<br />
<br />
*Jiang, Z., D. B. A. Jones, H. M. Worden, M. N. Deeter, D. K. Henze, J. Worden, and K. W. Bowman, Quantifying the impact of model biases in convective transport on inferred CO source estimates using multi-spectral CO retrievals from MOPITT, submitted.<br />
<br />
*Wecht, K. J., D. J. Jacob, S. C. Wofsy, E. A. Kort, J. R. Worden, S. S. Kulawik, D. K. Henze, M. Kopacz, and V. H. Payne, Validation of TES methane with HIPPO aircraft observations: implications for inverse modeling of methane sources, Atmos. Chem. Phys. Discuss., 11, 27887-27911.<br />
<br />
*Walker, T., D. B. A. Jones, M. Parrington, D. K. Henze, L. T. Murray, J. W. Bottenheim, K. Anlauf, J. R. Worden, K. W. Bowman, C. Shim, K. Singh, M. Kopacz, D. W. Tarasick, J. Davies, P. von der Gathen, and C. C. Carouge (2012), Impacts of midlatitude precursor emissions and local photochemistry on ozone abundances in the Arctic, J. Geophys. Res., doi:10.1029/2011JD016370.<br />
<br />
*Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald (2011), Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
<br />
*Singh, K., A. Sandu, Variational Chemical Data Assimilation with Approximate Adjoints, submitted. <br />
<br />
*Singh, K., Jardak, M., Sandu, A., Bowman, K., Lee, M., and Jones, D. (2010): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. Discuss., 3, 1783-1827, doi:10.5194/gmdd-3-1783-2010. http://www.geosci-model-dev-discuss.net/3/1783/2010/gmdd-3-1783-2010.html<br />
<br />
*Kopacz, M., D. L. Mauzerall, J. Wang, E. M. Leibensperger, D. K. Henze, and K. Singh (2011), Origin and radiative forcing of black carbon transported to the Himalayas and Tibetan Plateau, Atmos. Chem. Phys., 11, 2837-2852, 2011. http://www.atmos-chem-phys.net/11/2837/2011/acp-11-2837-2011.html<br />
<br />
* Kopacz, M., D.J. Jacob, J.A. Fisher, J. A. Logan, L. Zhang, I. A. Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010): Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atmos. Chem. Phys., 10, 855-876. http://www.atmos-chem-phys.net/10/855/2010/acp-10-855-2010.pdf<br />
<br />
* Kopacz, M., D. J. Jacob, D. K. Henze, C. L. Heald, D. G. Streets, and Q. Zhang (2009), A comparison of analytical and adjoint Bayesian inversion methods for constraining Asian sources of CO using satellite (MOPITT) measurements of CO columns, J. Geophys. Res., doi:0.1029/2007JD009264. http://acmg.seas.harvard.edu/publications/KopaczJGR2009_2007JD009264.pdf<br />
<br />
* Henze, D. K., J. H. Seinfeld and D. T. Shindell (2009), Inverse modeling and mapping U.S. air quality influences of inorganic PM2.5 precursor emissions with the adjoint of GEOS-Chem, Atmos. Chem. Phys., 9, 5877-5903.<br />
<br />
* Zhang, L., D. J. Jacob, M. Kopacz, D. K. Henze, K. Singh, and D. A. Jaffe (2009), Intercontinental source attribution of ozone pollution at western U.S. sites using an adjoint method, Geophys. Res. Lett., 36, L11810, doi:10.1029/2009GL037950<br />
<br />
* Henze, D. K., A. Hakami and J. H. Seinfeld (2007), Development of the adjoint of GEOS-Chem, Atmos. Chem. Phys., 7, 2413-2433.<br />
<br />
=== Conference proceedings === <br />
* Adjoint inversion of CO sources using combined MOPITT, SCIAMACHY and AIRS CO columns, presented by Monika Kopacz at the COSPAR Scientific Assembly, Montreal, July 18, 2008. http://acmg.seas.harvard.edu/presentations/powerpoints/mak2008/COSPAR_MKopacz_July2008.ppt<br />
<br />
* Singh, K., P. Eller, A. Sandu, D. K. Henze, K. Bowman, M. Kopacz, and M. Lee (2009), Towards the construction of a standard geos-chem adjoint model, ACM High Performance Computing Conference.<br />
<br />
* Kopacz, M., Mauzerall, D.L., Leibensperger, E.M., Wang, J., Henze, D.K., Singh, K., Shim, C. Identifying the origin and estimating the radiative forcing of BC in the Himalayas: an analysis using the global GEOS-Chem adjoint model, European Geophysical Union meeting, Vienna, May 4, 2010.<br />
<br />
* Kopacz, M., Jacob, D.J., Fisher, J.A., Logan, J.A., Zhang, L., Megretskaia, I.A., Yantosca, R.M., Singh, K., Henze, D.K., Burrows, J.P., Buchwitz, M., Khlystova, I., McMillan, W.W., Gille, J.C., Edwards, D.P., Eldering, A., Thouret, V., Nedelec, P. Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), European Geophysical Union meeting, Vienna, May 7, 2010.<br />
<br />
* Tang, J., Zhuang, Q. and Xiong, X. (2010), 4D-Var inversion of atmospheric methane fluxes by assimilating SCIAMACHY and AIRS satellite retrievals, AGU, Dec. 18, 2010, http:/web.ics.purdue.edu/~tang16/agu2010_tang.ppt<br />
<br />
* Bousserez, N., R. V. Martin, K. W. Bowman, D.K. Henze, M. Kopacz, K. Singh, C. Shim, C. Wespes, Improving the lightning NOx source using satellite observations: a 4D-var analysis approach, AGU, Dec., 2010, http://myweb.dal.ca/nc689777/AGU_liNOx_poster_final.pdf</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v33&diff=10900GEOS-Chem Adjoint v332012-07-17T17:20:18Z<p>Hyung-Min Lee: /* What's new in this version */</p>
<hr />
<div>== Overview ==<br />
<br />
BETA RELEASE<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v32]]<br />
<br />
=== What's new in this version ===<br />
<br />
GEOS-Chem Adjoint v33 contains the following major updates and improvements. <br />
<br />
{| border=1 cellspacing=0 cellpadding=5 <br />
|- bgcolor="#CCCCCC"<br />
!Feature<br />
!Type<br />
!Submitted by<br />
!Committed by<br />
!Version<br />
!Status<br />
|-<br />
|[[#Add_TES_O3_IRK_.28v33a.29|Add TES O3 IRK observation operator]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Kevin Bowman (JPL)<br />
|Daven Henze (CU Boulder)<br />
|v33a<br />
|Approved 13 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_wet_deposition_adjoint_.28v33b.29|Fix wetdep adjoint]]<br />
|Adjoint bug fix<br />
|Fabien Paulot (Harvard University), Daven Henze (CU Boulder)<br />
|Daven Henze (CU Boulder)<br />
|v33b<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_checkpointing_of_aircraft_SO2_emissions_.28v33c.29|Fix checkpointing of aircraft SO2 emissions]]<br />
|Adjoint bug fix<br />
|Daven Henze, Kateryna Lapina (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33c<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_linoz_adjoint_.28v33d.29|Fix linoz adjoint]]<br />
|Adjoint bug fix<br />
|Hyungmin Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33d<br />
|Approved 07 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Adjoint_stability_update_.28v33e.29| Adjoint stability update]]<br />
|Adjoint bug fix<br />
|Jamin Koo (MIT), Daven Henze (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_SE_Asia_NH3_seasonality_.28v33e.29| Fix SE Asia NH3 seasonality]]<br />
|Forward bug fix<br />
|Shailesh Kharol (Dalhousie), Steven Vogel (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_problem_with_LICS_flag_in_using_SCHEM_.28v33f.29| Fix problem with LICS flag in using SCHEM]]<br />
|Adjoint bug fix<br />
|Hyungmin Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33f<br />
|Approved 13 July 2012<br />
|-bgcolor="#CCFFFF"<br />
|}<br />
<br />
==== Updates to forward model ====<br />
<br />
==== Bug fixes in forward model ====<br />
<br />
===== Fix SE Asia NH3 seasonality (v33e) =====<br />
<br />
The bug fix from the forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Anthropogenic_emissions#Streets_2000 fixing the seasonality of the NH3 emissions over SE Asia] has also been applied.<br />
<br />
1 week benchmark is approved [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_003]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:40, 10 May 2012 (EDT)<br />
<br />
==== Updates to adjoint model ====<br />
<br />
===== Add TES O3 IRK (v33a) =====<br />
<br />
Implemented observation operator for assimilation of TES O3 IRKs. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_001]<br />
<br />
--[[User:Daven|Daven]] 19:03, 13 April 2012 (EDT)<br />
<br />
==== Bug fixes in adjoint model ====<br />
<br />
===== Fix wet deposition adjoint (v33b) =====<br />
The adjoint was not properly handling evaporation. This fix helps make the adjoint of species such as NH3 better in a few locations where adjoints were diverging from forward model sensitivities by up to 50%. <br />
<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_002 vs 33b_002]<br />
<br />
--[[User:Daven|Daven]] 12:51, 20 April 2012 (EDT)<br />
<br />
===== Fix checkpointing of aircraft SO2 emissions (v33c) =====<br />
<br />
The SO2 aircraft emissions are distributed in the model based on a mapping from the 20 level inventory to the GEOS-Chem levels. This mapping depends upon the air masses in the models. The emissions are read in during the first hour of the month in the forward model, but on the last hour of the month in the adjoint. Thus they were being calculated inconsistently in the adjoint compared to the forward model. We now checkpoint these emissions rather than recalculate them<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33c_001 - 33c_004]<br />
<br />
--[[User:Daven|Daven]] 23:08, 7 May 2012 (EDT)<br />
<br />
===== Fix linoz adjoint (v33d) =====<br />
<br />
While the linoz routine is linear, and should be self-adjoint, it isn't exactly self-adjoint. For example, given initial forcings of STT_ADJ = 0d0 in every grid cell, it can generate non-zero values of STT_ADJ as output, which is incorrect. <br />
<br />
In this update we use the discrete adjoint based upon reformulating the code to be consistent with the strat chem routine for the non-Ox species. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33d_001 - 33c_007]<br />
<br />
--[[User:Daven|Daven]] 23:14, 7 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33d (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
===== Adjoint stability update (v33e) =====<br />
<br />
Several fixes were made to increase the numerical stability of the code<br />
*Change the filter to 1d-10 in partition_adj.f<br />
*Better error trapping in gckpp_adj_Integrator.f90 to prevent NaNs following a failed integration step during the adjoint<br />
<br />
<br />
<br />
A few other small changes to improve performance were included in this update. Overall, the benchmark performance is not impacted by these update, see [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_001]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:26, 10 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33e (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.'''<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
<br />
===== Fix problem with LICS flag in using SCHEM (v33f) =====<br />
<br />
New stratospheric chemistry module (strat_chem_mod.f), was set to loop over the number of tracers (NSTPL) that has read from input.gcadj. It caused a problem when LADJ_STRAT is turned off, LADJ_STRAT is not appeared in input.gcadj (NSTPL = 0), or LADJ = F for spin-up.<br />
<br />
--[[User:Hyung-Min Lee|Hyungmin]] 19:26, 14 Jul 2012 (EDT)<br />
<br />
== Outstanding issues not yet resolved in v33 ==<br />
<br />
=== Updates in the pipeline ===<br />
*Reaction rate sensitivities (Paulot, Harvard; Walker, UT)<br />
*Update inverse Hessian approximation (Bousserez, CUB)<br />
*Off-diagonal error covariance matrices (Singh, VT; Liu, JPL)<br />
*Deposition-based cost function and updated chemistry (Paulot, Harvard)<br />
*ETOH simulations (Millet, UMN)<br />
*NH3 updates (Zhu, Colorado)<br />
*NO2/SO2 obs operators (Bousserez, Dalhousie)<br />
*ANISO (Capps, Georgia Tech)</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v33&diff=10897GEOS-Chem Adjoint v332012-07-14T23:30:25Z<p>Hyung-Min Lee: /* Bug fixes in adjoint model */</p>
<hr />
<div>== Overview ==<br />
<br />
BETA RELEASE<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v32]]<br />
<br />
=== What's new in this version ===<br />
<br />
GEOS-Chem Adjoint v33 contains the following major updates and improvements. <br />
<br />
{| border=1 cellspacing=0 cellpadding=5 <br />
|- bgcolor="#CCCCCC"<br />
!Feature<br />
!Type<br />
!Submitted by<br />
!Committed by<br />
!Version<br />
!Status<br />
|-<br />
|[[#Add_TES_O3_IRK_.28v33a.29|Add TES O3 IRK observation operator]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Kevin Bowman (JPL)<br />
|Daven Henze (CU Boulder)<br />
|v33a<br />
|Approved 13 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_wet_deposition_adjoint_.28v33b.29|Fix wetdep adjoint]]<br />
|Adjoint bug fix<br />
|Fabien Paulot (Harvard University), Daven Henze (CU Boulder)<br />
|Daven Henze (CU Boulder)<br />
|v33b<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_checkpointing_of_aircraft_SO2_emissions_.28v33c.29|Fix checkpointing of aircraft SO2 emissions]]<br />
|Adjoint bug fix<br />
|Daven Henze, Kateryna Lapina (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33c<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_linoz_adjoint_.28v33d.29|Fix linoz adjoint]]<br />
|Adjoint bug fix<br />
|Hyungmin Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33d<br />
|Approved 07 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Adjoint_stability_update_.28v33e.29| Adjoint stability update]]<br />
|Adjoint bug fix<br />
|Jamin Koo (MIT), Daven Henze (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_SE_Asia_NH3_seasonality_.28v33e.29| Fix SE Asia NH3 seasonality]]<br />
|Forward bug fix<br />
|Shailesh Kharol (Dalhousie), Steven Vogel (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_LICS_.28v33e.29| Fix problem with LICS flag in using SCHEM]]<br />
|Adjoint bug fix<br />
|Hyungmin Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33f<br />
|Approved 13 July 2012<br />
|-bgcolor="#CCFFFF"<br />
|}<br />
<br />
==== Updates to forward model ====<br />
<br />
==== Bug fixes in forward model ====<br />
<br />
===== Fix SE Asia NH3 seasonality (v33e) =====<br />
<br />
The bug fix from the forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Anthropogenic_emissions#Streets_2000 fixing the seasonality of the NH3 emissions over SE Asia] has also been applied.<br />
<br />
1 week benchmark is approved [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_003]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:40, 10 May 2012 (EDT)<br />
<br />
==== Updates to adjoint model ====<br />
<br />
===== Add TES O3 IRK (v33a) =====<br />
<br />
Implemented observation operator for assimilation of TES O3 IRKs. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_001]<br />
<br />
--[[User:Daven|Daven]] 19:03, 13 April 2012 (EDT)<br />
<br />
==== Bug fixes in adjoint model ====<br />
<br />
===== Fix wet deposition adjoint (v33b) =====<br />
The adjoint was not properly handling evaporation. This fix helps make the adjoint of species such as NH3 better in a few locations where adjoints were diverging from forward model sensitivities by up to 50%. <br />
<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_002 vs 33b_002]<br />
<br />
--[[User:Daven|Daven]] 12:51, 20 April 2012 (EDT)<br />
<br />
===== Fix checkpointing of aircraft SO2 emissions (v33c) =====<br />
<br />
The SO2 aircraft emissions are distributed in the model based on a mapping from the 20 level inventory to the GEOS-Chem levels. This mapping depends upon the air masses in the models. The emissions are read in during the first hour of the month in the forward model, but on the last hour of the month in the adjoint. Thus they were being calculated inconsistently in the adjoint compared to the forward model. We now checkpoint these emissions rather than recalculate them<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33c_001 - 33c_004]<br />
<br />
--[[User:Daven|Daven]] 23:08, 7 May 2012 (EDT)<br />
<br />
===== Fix linoz adjoint (v33d) =====<br />
<br />
While the linoz routine is linear, and should be self-adjoint, it isn't exactly self-adjoint. For example, given initial forcings of STT_ADJ = 0d0 in every grid cell, it can generate non-zero values of STT_ADJ as output, which is incorrect. <br />
<br />
In this update we use the discrete adjoint based upon reformulating the code to be consistent with the strat chem routine for the non-Ox species. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33d_001 - 33c_007]<br />
<br />
--[[User:Daven|Daven]] 23:14, 7 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33d (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
===== Adjoint stability update (v33e) =====<br />
<br />
Several fixes were made to increase the numerical stability of the code<br />
*Change the filter to 1d-10 in partition_adj.f<br />
*Better error trapping in gckpp_adj_Integrator.f90 to prevent NaNs following a failed integration step during the adjoint<br />
<br />
<br />
<br />
A few other small changes to improve performance were included in this update. Overall, the benchmark performance is not impacted by these update, see [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_001]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:26, 10 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33e (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.'''<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
<br />
===== Fix problem with LICS flag in using SCHEM (v33f) =====<br />
<br />
New stratospheric chemistry module (strat_chem_mod.f), was set to loop over the number of tracers (NSTPL) that has read from input.gcadj. It caused a problem when LADJ_STRAT is turned off, LADJ_STRAT is not appeared in input.gcadj (NSTPL = 0), or LADJ = F for spin-up.<br />
<br />
--[[User:Hyung-Min Lee|Hyungmin]] 19:26, 14 Jul 2012 (EDT)<br />
<br />
== Outstanding issues not yet resolved in v33 ==<br />
<br />
=== Updates in the pipeline ===<br />
*Reaction rate sensitivities (Paulot, Harvard; Walker, UT)<br />
*Update inverse Hessian approximation (Bousserez, CUB)<br />
*Off-diagonal error covariance matrices (Singh, VT; Liu, JPL)<br />
*Deposition-based cost function and updated chemistry (Paulot, Harvard)<br />
*ETOH simulations (Millet, UMN)<br />
*NH3 updates (Zhu, Colorado)<br />
*NO2/SO2 obs operators (Bousserez, Dalhousie)<br />
*ANISO (Capps, Georgia Tech)</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v33&diff=10896GEOS-Chem Adjoint v332012-07-13T21:16:37Z<p>Hyung-Min Lee: /* What's new in this version */</p>
<hr />
<div>== Overview ==<br />
<br />
BETA RELEASE<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v32]]<br />
<br />
=== What's new in this version ===<br />
<br />
GEOS-Chem Adjoint v33 contains the following major updates and improvements. <br />
<br />
{| border=1 cellspacing=0 cellpadding=5 <br />
|- bgcolor="#CCCCCC"<br />
!Feature<br />
!Type<br />
!Submitted by<br />
!Committed by<br />
!Version<br />
!Status<br />
|-<br />
|[[#Add_TES_O3_IRK_.28v33a.29|Add TES O3 IRK observation operator]]<br />
|Adjoint Update<br />
|Daven Henze (CU Boulder), Kevin Bowman (JPL)<br />
|Daven Henze (CU Boulder)<br />
|v33a<br />
|Approved 13 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_wet_deposition_adjoint_.28v33b.29|Fix wetdep adjoint]]<br />
|Adjoint bug fix<br />
|Fabien Paulot (Harvard University), Daven Henze (CU Boulder)<br />
|Daven Henze (CU Boulder)<br />
|v33b<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_checkpointing_of_aircraft_SO2_emissions_.28v33c.29|Fix checkpointing of aircraft SO2 emissions]]<br />
|Adjoint bug fix<br />
|Daven Henze, Kateryna Lapina (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33c<br />
|Approved 20 Apr 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_linoz_adjoint_.28v33d.29|Fix linoz adjoint]]<br />
|Adjoint bug fix<br />
|Hyungmin Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33d<br />
|Approved 07 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Adjoint_stability_update_.28v33e.29| Adjoint stability update]]<br />
|Adjoint bug fix<br />
|Jamin Koo (MIT), Daven Henze (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_SE_Asia_NH3_seasonality_.28v33e.29| Fix SE Asia NH3 seasonality]]<br />
|Forward bug fix<br />
|Shailesh Kharol (Dalhousie), Steven Vogel (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33e<br />
|Approved 09 May 2012<br />
|-bgcolor="#CCFFFF"<br />
|-<br />
|[[#Fix_LICS_.28v33e.29| Fix problem with LICS flag in using SCHEM]]<br />
|Adjoint bug fix<br />
|Hyungmin Lee (CU Boulder)<br />
|Nicolas Bousserez (CU Boulder)<br />
|v33f<br />
|Approved 13 July 2012<br />
|-bgcolor="#CCFFFF"<br />
|}<br />
<br />
==== Updates to forward model ====<br />
<br />
==== Bug fixes in forward model ====<br />
<br />
===== Fix SE Asia NH3 seasonality (v33e) =====<br />
<br />
The bug fix from the forward model [http://wiki.seas.harvard.edu/geos-chem/index.php/Anthropogenic_emissions#Streets_2000 fixing the seasonality of the NH3 emissions over SE Asia] has also been applied.<br />
<br />
1 week benchmark is approved [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_003]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:40, 10 May 2012 (EDT)<br />
<br />
==== Updates to adjoint model ====<br />
<br />
===== Add TES O3 IRK (v33a) =====<br />
<br />
Implemented observation operator for assimilation of TES O3 IRKs. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_001]<br />
<br />
--[[User:Daven|Daven]] 19:03, 13 April 2012 (EDT)<br />
<br />
==== Bug fixes in adjoint model ====<br />
<br />
===== Fix wet deposition adjoint (v33b) =====<br />
The adjoint was not properly handling evaporation. This fix helps make the adjoint of species such as NH3 better in a few locations where adjoints were diverging from forward model sensitivities by up to 50%. <br />
<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33a_002 vs 33b_002]<br />
<br />
--[[User:Daven|Daven]] 12:51, 20 April 2012 (EDT)<br />
<br />
===== Fix checkpointing of aircraft SO2 emissions (v33c) =====<br />
<br />
The SO2 aircraft emissions are distributed in the model based on a mapping from the 20 level inventory to the GEOS-Chem levels. This mapping depends upon the air masses in the models. The emissions are read in during the first hour of the month in the forward model, but on the last hour of the month in the adjoint. Thus they were being calculated inconsistently in the adjoint compared to the forward model. We now checkpoint these emissions rather than recalculate them<br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33c_001 - 33c_004]<br />
<br />
--[[User:Daven|Daven]] 23:08, 7 May 2012 (EDT)<br />
<br />
===== Fix linoz adjoint (v33d) =====<br />
<br />
While the linoz routine is linear, and should be self-adjoint, it isn't exactly self-adjoint. For example, given initial forcings of STT_ADJ = 0d0 in every grid cell, it can generate non-zero values of STT_ADJ as output, which is incorrect. <br />
<br />
In this update we use the discrete adjoint based upon reformulating the code to be consistent with the strat chem routine for the non-Ox species. <br />
<br />
See benchmark results [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33d_001 - 33c_007]<br />
<br />
--[[User:Daven|Daven]] 23:14, 7 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33d (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
===== Adjoint stability update (v33e) =====<br />
<br />
Several fixes were made to increase the numerical stability of the code<br />
*Change the filter to 1d-10 in partition_adj.f<br />
*Better error trapping in gckpp_adj_Integrator.f90 to prevent NaNs following a failed integration step during the adjoint<br />
<br />
<br />
<br />
A few other small changes to improve performance were included in this update. Overall, the benchmark performance is not impacted by these update, see [https://docs.google.com/spreadsheet/ccc?key=0As2MewHKyWpDdDlTeFdPanU2RzlWc0luaFpEQXlyclE#gid=0 33e_001]<br />
<br />
<br />
--[[User:Daven|Daven]] 02:26, 10 May 2012 (EDT)<br />
<br />
'''Important: If you check out version 33e (using the tag option), make sure to also check out the most recent version of linoz_mod.f (cvs checkout linoz_mod.f), since this file had not been updated by mistake.'''<br />
--[[User:Nicolas Bousserez|Nicolas Bousserez]] 22:12, 7 July 2012 (EDT)<br />
<br />
== Outstanding issues not yet resolved in v33 ==<br />
<br />
=== Updates in the pipeline ===<br />
*Reaction rate sensitivities (Paulot, Harvard; Walker, UT)<br />
*Update inverse Hessian approximation (Bousserez, CUB)<br />
*Off-diagonal error covariance matrices (Singh, VT; Liu, JPL)<br />
*Deposition-based cost function and updated chemistry (Paulot, Harvard)<br />
*ETOH simulations (Millet, UMN)<br />
*NH3 updates (Zhu, Colorado)<br />
*NO2/SO2 obs operators (Bousserez, Dalhousie)<br />
*ANISO (Capps, Georgia Tech)</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint&diff=10747GEOS-Chem Adjoint2012-06-20T01:01:59Z<p>Hyung-Min Lee: /* Current GEOS-Chem adjoint versions released */</p>
<hr />
<div><big><big><strong>Adjoint and Data Assimilation Working Group</strong></big></big><br />
<br />
== Contact information ==<br />
<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|-<br />
|'''Adjoint Working Group Co-Chairs'''<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman] and [mailto:dbj@atmosp.physics.utoronto.ca Dylan Jones]<br />
|-<br />
|-<br />
|'''Adjoint Model Scientist'''<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|'''Adjoint Co-Model Scientist/GC adjoint support team'''<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez]<br />
|-<br />
|'''Adjoint Working Group email list'''<br />
|<tt>geos-chem-adjoint@seas.harvard.edu</tt><br />
|-<br />
|'''To subscribe to email list'''<br />
|Send email to [mailto:geos-chem-adjoint-join@seas.harvard.edu <tt>geos-chem-adjoint-join@seas.harvard.edu</tt>]<br />
|-<br />
|'''To unsubscribe from email list'''<br />
|Send email to [mailto:geos-chem-adjoint-leave@seas.harvard.edu <tt>geos-chem-adjoint-leave@seas.harvard.edu</tt>]<br />
|}<br />
<br />
== Historical Development ==<br />
Original work on the adjoint of GEOS-Chem v6 began in 2003, focusing on the adjoint of the offline aerosol simulation. By 2005, the adjoint was expanded to include a tagged CO simulation and a full chemistry simulation; an adjoint of GEOS-Chem v7 was also developed in the following years. Each of these branches of the adjoint code were been constructed in a hybrid fashion using a combination of automatic differentiation software ([http://www.autodiff.com/tamc TAMC], [http://people.cs.vt.edu/~asandu/Software/Kpp/ KPP]) and manual coding of both discrete and continuous adjoints. They shared many common elements yet had unique features for different applications. <br />
<br />
During the summer of 2009, the existing branches were merged and updated to bring the adjoint into alignment with the latest release of GEOS-Chem, v8-02-01. This merged adjoint model is now the standard adjoint code into which all further development efforts will be placed.<br />
<br />
== Forward Model Code ==<br />
<br />
The forward model on which the adjoint is based originally corresponded to GEOS-Chem v8-02-01. It was subsequently updated as follows:<br />
<br />
* KPP solver for gas-phase chemistry (as in GCv8-02-03)<br />
* Implement Bond 2007 BC/OC emissions (as in GCv8-02-02)<br />
* Apply bug fixes from GCv8-02-02 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-02#Previous_issues_now_resolved_in_v8-02-02 here]<br />
* Apply bug fixes from GCv8-02-03 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-03#Previous_issues_now_resolved_in_v8-02-03 here]<br />
* Apply bug fixes from GCv8-02-04 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-04#Previous_issues_now_resolved_in_v8-02-04 here]<br />
<br />
All bug fixes and model updates were previous listed at the top of inverse_driver.f. We have now switched to documenting the code development cycle here in the wiki, see the following section. <br />
<br />
== Code Versions, Bug Fixes and Developments == <br />
<br />
=== Current GEOS-Chem adjoint version released ===<br />
[[GEOS-Chem_Adjoint_v33]] (You will get the latest version on the list when you check out.)<br />
<br />
=== Previous GEOS-Chem adjoint versions released ===<br />
*[[GEOS-Chem_Adjoint_v32]]<br />
*[[GEOS-Chem_Adjoint_v31]]<br />
<br />
== Summary of Main Adjoint Code Supported Features ==<br />
<br />
=== Features === <br />
* Meteorological fields<br />
** GEOS-3 <i>needs testing</i><br />
** GEOS-4<br />
** GEOS-5 <br />
* model resolution<br />
** 4 x 5<br />
** 2 x 2.5 <br />
** Nested Asia and NA<br />
* Forward model processes<br />
** convection<br />
** advection<br />
** PBL mixing<br />
** dry deposition<br />
** wet deposition<br />
** strat / trop exchange with LINOZ and new GMI strat chem (v9-01-03)<br />
** NOy up fluxes (now replaced with new GMI strat chem)<br />
** aerosols <br />
*** inorganic aerosol thermodynamics with RPMARES<br />
*** inorganic aerosol thermodynamics with ISORROPIA <i> in progress </i><br />
*** sulfate chemistry <br />
*** BC<br />
*** SOA, Dust, sea salt <i> needs doing </i><br />
*** aerosol surface area feedbacks <i> needs updating </i><br />
*** aerosol optical feedbacks <i> needs doing </i><br />
** emissions<br />
*** all standard emissions included<br />
* Simulation modes<br />
** full chemistry<br />
** tagged CO<br />
** tagged Ox <br />
** CH4<br />
** offline aerosols (for BC and dust only)<br />
** CO2<br />
* Observational Operators<br />
** MOPITT CO column<br />
** SCIAMACHY CO column<br />
** AIRS CO column<br />
** IMPROVE BC<br />
** CASTNet (NH4+) <i> needs updating </i> <br />
** GOME / SCIAMACHY NO2 column <i> needs updating </i> <br />
*** using KNMI retrieval (Henze)<br />
*** using Dalhousie retrieval (Shim)<br />
** SCIAMACHY/OMI NO2<br />
*** using Dalhousie retrieval (Bousserez, Padmanabhan)<br />
** TES NH3 <br />
** TES O3 <br />
** GOSAT CO2 <br />
** MLS O3 and TES CO2 <i> in progress </i><br />
* Control parameters<br />
** Initial Conditions scaling factors (linear or log)<br />
** Emissions scaling factors (linear or log) <br />
*** NH3, primary BC/OC, SO2: anthropogenic, natural, bioburn, biomass, ship<br />
*** NOx: soil, aircraft, anthropogenic, biofuel, bioburn<br />
*** Lightning NOx: injection height, yield <i> in progress </i><br />
*** all other gas-phase tracers: anthropogenic, biofuel, bioburn <br />
* Adjoint sensitivities<br />
** w.r.t. all implemented control parameters<br />
** w.r.t Reaction Rate Parameters <br />
** w.r.t all emissions <br />
** of AQ attainment metrics <i> needs updating </i><br />
** of spatiotemporally averaged species concentrations (e.g., arctic O3) <br />
* Other<br />
** Inverse Hessian approximation <br />
** off-diagonal covariance matrices <i> needs updating </i><br />
** 3D-Var <i> needs updating </i><br />
<br />
Features may be qualified as: <br />
* <i>needs testing</i>: an implemented feature that we haven't fully used yet<br />
* <i>needs updating</i>: a feature developed with a previous branch that has yet to be updated to GEOS-Chem v8 and the merged adjoint<br />
* <i>needs doing</i>: a feature nobody has tackled the adjoint of yet<br />
* <i>in progress</i>: a feature currently under development <br />
* <i>in pipeline</i>: a feature which has been submitted and awaiting integration into the CVS repository<br />
<br />
=== Primary code developers ===<br />
Monika Kopacz, Kumaresh Singh, Changsub Shim, Daven Henze<br />
<br />
=== Adjoint model lead scientist ===<br />
Daven Henze<br />
<br />
<br />
<br />
== Resources ==<br />
<br />
=== User's guide ===<br />
A user's guide is available. http://adjoint.colorado.edu/%7Edaven/gcadj_std/GC_adj_man.pdf <br />
<br />
=== Code flowchart ===<br />
Meemong Lee has created a detailed flowchart of the inverse model code structure. http://adjoint.colorado.edu/~daven/gcadj_std/flowchart.pdf<br />
<br />
=== Plotting tools ===<br />
<br />
Some IDL and MATLAB routines for plotting benchmark results. http://adjoint.colorado.edu/~daven/gcadj_std/tools.tar.gz<br />
<br />
=== Background papers and presentations ===<br />
<br />
Several articles and presentations (including a GC adjoint modeling clinic overview from IGC5) providing background information about adjoints. http://adjoint.colorado.edu/~daven/gcadj_std/adj_articles.tar.gz<br />
<br />
== Distribution and Use ==<br />
Code for the adjoint is distributed through a CVS server located at adjoint.colorado.edu. Contact Daven Henze to obtain an account on the server.<br />
<br />
Even if your office mate has a copy of the code, the best way to obtain the model is to get a CVS account for yourself and download a version from the repository. So please do not copy code directly from others or pass the code along to third parties. This vastly helps with tracking developments and keeping up with model updates. <br />
<br />
Use of the adjoint model code follows standard practice for GEOS-Chem. It is expected that any developments that come of individual applications based on this community model will eventually be given back to the community by incorporation of new developments into the standard adjoint code. New development should be submitted to Daven Henze for inclusion in the standard adjoint model code.<br />
<br />
=== Quick guide to CVS === <br />
We recommend first taking a look at CVS manual to get a general feel for how this tools works (e.g., http://cvsbook.red-bean.com/cvsbook.html). <br />
<br />
Below are some command commands you may use for developing code and checking the status of code updates.<br />
<br />
Obtain the latest code:<br />
<br />
cvs checkout gcadj_std<br />
<br />
<br />
Generate a list (modified_files.txt) of all the files in your local copy that differ from the current repository code:<br />
<br />
cvs -q status | grep 'Status' | grep -v 'Up-to-date' > modified_files.txt<br />
<br />
<br />
Determine the difference between your local copy of a file and the version that you originally checked out (i.e., see what you changed):<br />
<br />
cvs diff foo_mod.f<br />
<br />
Note: arguments such as a filename are optional. Without listing a specific file, cvs will run the command on all files in the current directory. <br />
<br />
<br />
Determine the difference between your local copy of the code the current version in the repository (i.e., see both what you changed, and what has changed in the repository):<br />
<br />
cvs diff -D "now" foo_mod.f<br />
<br />
<br />
Merge your local file with the current repository version<br />
<br />
cvs update foo_mod.f <br />
<br />
<br />
Obtain a fresh copy of a file in the repository (without merging)<br />
<br />
rm foo_mod.f<br />
cvs update foo_mod.f <br />
<br />
<br />
Difference between the local repository and the tagged version “TAGNAME”<br />
<br />
cvs diff -r TAGNAME (< file name > for specific file)<br />
<br />
<br />
Difference between the two tagged versions “TAGNAME1” and “TAGNAME2” in the repository (shows modified files only):<br />
<br />
cvs diff -N -c -r TAGNAME2 -r TAGNAME2 | grep "Index:" > output.txt<br />
<br />
Please do not use the commit command, which is restricted to the GCadj support team (Daven Henze and Nicolas Bousserez).<br />
<br />
== Crediting GEOS-Chem adjoint developers ==<br />
<br />
We aim to make distribution of adjoint model code as immediate as possible. A consequence is that many features may not yet be publicly documented. Therefore, giving code developers due credit is of utmost importance. <br />
<br />
Authors of new additions to the standard code should be offered co-authorship on the first round of presentations and publications to come of their development. Features currently falling in this category and their developers are:<br />
* (v32) MOPITT v5 CO observation operator. Developer: Zhe Jiang, University of Toronto.<br />
* (v32) Dust adjoint. Developer: Xiaoguang (Richard) Xu, University of Nebraska Lincoln.<br />
* (v32) Black carbon offline aerosol adjoint. Developer: Yuhao Mao (UCLA).<br />
* (v32) Nested full chemistry adjoint. Developers: Zhe Jiang, University of Toronto; Daven Henze, CU Boulder.<br />
* (v32) Stratospheric production / loss rate sensitivities. Developer: Hyungmin Lee, CU Boulder.<br />
* (v32) CH4 adjoint. Developer: Kevin Wecht, Harvard University<br />
* (v29) LIDORT. Developer: Daven Henze, University of Colorado Boulder. Collaborator: Rob Spurr. <br />
* (v28) CO2 adjoint. Developer: Daven Henze, University of Colorado Boulder. Collaborators: Ray Nassar, Kevin Bowman, Dylan Jones.<br />
<br />
Citation of the appropriate [[GEOS-Chem_Adjoint#Journal_Articles|journal articles]] for mature developments is also encouraged, as well as considering aspects of [http://acmg.seas.harvard.edu/geos/geos_credit.html co-authorship for the forward model]. <br />
<br />
Overall, if you have any questions about authorship, even for a conference presentation, please contact Daven Henze.<br />
<br />
== Current GEOS-Chem Adjoint Research Projects (please add yours!) ==<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|- bgcolor="#cccccc"<br />
!User Group <br />
!Description <br />
!Contact Person<br />
|-<br />
|CU Boulder<br />
|Aerosol precursors, CO2, O3; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|CU Boulder<br />
|Inverse modeling/optimization; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Nicolas Bousserez]<br />
|-<br />
|Harvard<br />
|Methane<br />
|Kevin Wecht, wecht [at] fas.harvard.edu<br />
|-<br />
|Purdue University<br />
|Methane (SICAMACHY, AIRS and IASI)<br />
|[mailto:tang16@purdue.edu Jinyun Tang]<br />
|-<br />
|MIT<br />
|Aircraft emissions<br />
|[mailto:jaminkoo@mit.edu Jamin Koo]<br />
|-<br />
|Princeton<br />
|BC sensitivities, general adjoint code development<br />
|Monika Kopacz, mkopacz [at] princeton.edu<br />
|-<br />
|Dalhousie University<br />
|Lightning NOx emissions and impact on tropical ozone using the adjoint<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez (now at CU-Boulder)]<br />
|-<br />
|Dalhousie University<br />
|Surface NOx emissions inversion using SCIAMACHY/OMI NO2 measurements<br />
|Akhila Padmanabhan akhila [at] dal.ca; Nicolas Bousserez [mailto:Nicolas.Bousserez@colorado.edu] (now at CU-Boulder)<br />
|-<br />
|JPL<br />
|Microwave Limb Sounder (MLS) Ozone assimilation<br />
|[mailto:meemong.lee@jpl.nasa.gov Meemong Lee]<br />
|-<br />
|JPL <br />
| TES ozone assimilation/attribution of ozone radiative forcing<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|University of Edinburgh<br />
|Quantifying the impact of boreal forest fires on tropospheric oxidants over the Atlantic<br />
|[mailto:mark.parrington@ed.ac.uk Mark Parrington]<br />
|-<br />
|US EPA<br />
|Integration with economic models for future emission inventory scenario development<br />
|[mailto:akhtar.farhan@epa.gov Farhan Akhtar]<br />
|-<br />
|Peking University<br />
|Satellite constraints on VOC emissions <br />
|[mailto:tmfu@pku.edu.cn May Fu]<br />
|-<br />
|CU Boulder<br />
|Spatial extent of source influences on aerosol precursor columns<br />
|[mailto:alexander.turner@colorado.edu Alex Turner]<br />
|-<br />
|IAP.CAS<br />
|CO2 assimilation<br />
|[mailto:czh@mail.iap.ac.cn Chen]<br />
|-<br />
|Purdue University<br />
|Feedback between terrestrial ecosystem processes and atmospheric co2 signals<br />
|[mailto:zhuq@purdue.edu Qing Zhu]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and reactive nitrogen to precursor emissions<br />
|[mailto:twalker@atmosp.physics.utoronto.ca Thomas Walker]<br />
|-<br />
|University of Toronto<br />
|Adjoint analysis for carbon monoxide <br />
|[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]<br />
|-<br />
|Georgia Tech<br />
|ISORROPIA adjoint development; inorganic aerosol precursors<br />
|[mailto:scapps@gatech.edu Shannon Capps]<br />
|-<br />
|Peking University<br />
|Source attributions of tropospheric ozone over North China<br />
|[mailto:linjt@pku.edu.cn Jintai Lin]<br />
|-<br />
|University of Wollongong<br />
|Sensitivity of ozone and adjoint analysis of CO over Australasia.<br />
|[mailto:rb864@uowmail.edu.au Rebecca Buchholz]<br />
|-<br />
|University of Toronto<br />
|CO2 assimilation & transport model bias estimation<br />
|[mailto:mkeller@atmosp.physics.utoronto.ca Martin Keller]<br />
|-<br />
|Dalhousie University<br />
|Sensitivity of global PM2.5-induced mortality to emissions<br />
|[mailto:colin.lee@dal.ca Colin Lee]<br />
|-<br />
|University of Leicester (UK)<br />
|Top-down estimates of Amazon isoprene emissions<br />
|[mailto:mpb14@le.ac.uk Michael Barkley]<br />
|-<br />
|Anyang University<br />
|Aerosol emission modeling in East Asia <br />
|[mailto:koo@anyang.ac.kr Youn Seo Koo]<br />
|-<br />
|Nanjing University<br />
|Inverse modeling of terrestrial ecosystem carbon flux<br />
|[mailto:hengmao.wang@gmail.com Hengmao Wang]<br />
|-<br />
|Tsinghua University<br />
|Nested-gird simulations with the adjoint model<br />
|[mailto:ny07@mails.tsinghua.edu.cn Nan Yang]<br />
|-<br />
|University of Minnesota<br />
|Inverse modeling of VOC sources based on TES and IASI measurements<br />
|[mailto:dbm@umn.edu Dylan Millet]<br />
|}<br />
<br />
== Publications ==<br />
<br />
=== Journal Articles ===<br />
*Wang, J., X. Xu, D. K. Henze, Q. Ji, S.-C. Tsay, J. Huang, Top-Down Estimate of Dust Emissions through Integration of MODIS and MISR Aerosol Retrievals with the GEOS-Chem adjoint model, submitted.<br />
<br />
*Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden (2012), The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098<br />
<br />
*Bowman, K. W., and D. K. Henze, Attribution of direct ozone radiative forcing to spatially-resolved emissions, submitted.<br />
<br />
*Paulot, F., D. K. Henze, and P. O. Wennberg (2012), Impact of the isoprene photochemical cascade on tropical ozone, Atmos. Chem. Phys., 12, 1307-1325, 2012.<br />
<br />
*Henze, D. K., D. T. Shindell, F. Akhtar, R. J. D. Spurr, R. W. Pinder, D. Loughlin, M. Kopacz, K. Singh, and C. Shim, Spatially refined aerosol direct radiative forcing efficiencies, submitted.<br />
<br />
*Turner, A., D. K. Henze, R. V. Martin, and A. Hakami (2012), The spatial extent of source influences on modeled column concentrations of short-lived species, ''Geophys. Res. Lett.'', doi: 10.1029/2012GL051832, in press.<br />
<br />
*Jiang, Z., D. B. A. Jones, H. M. Worden, M. N. Deeter, D. K. Henze, J. Worden, and K. W. Bowman, Quantifying the impact of model biases in convective transport on inferred CO source estimates using multi-spectral CO retrievals from MOPITT, submitted.<br />
<br />
*Wecht, K. J., D. J. Jacob, S. C. Wofsy, E. A. Kort, J. R. Worden, S. S. Kulawik, D. K. Henze, M. Kopacz, and V. H. Payne, Validation of TES methane with HIPPO aircraft observations: implications for inverse modeling of methane sources, Atmos. Chem. Phys. Discuss., 11, 27887-27911.<br />
<br />
*Walker, T., D. B. A. Jones, M. Parrington, D. K. Henze, L. T. Murray, J. W. Bottenheim, K. Anlauf, J. R. Worden, K. W. Bowman, C. Shim, K. Singh, M. Kopacz, D. W. Tarasick, J. Davies, P. von der Gathen, and C. C. Carouge (2012), Impacts of midlatitude precursor emissions and local photochemistry on ozone abundances in the Arctic, J. Geophys. Res., doi:10.1029/2011JD016370.<br />
<br />
*Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald (2011), Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
<br />
*Singh, K., A. Sandu, Variational Chemical Data Assimilation with Approximate Adjoints, submitted. <br />
<br />
*Singh, K., Jardak, M., Sandu, A., Bowman, K., Lee, M., and Jones, D. (2010): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. Discuss., 3, 1783-1827, doi:10.5194/gmdd-3-1783-2010. http://www.geosci-model-dev-discuss.net/3/1783/2010/gmdd-3-1783-2010.html<br />
<br />
*Kopacz, M., D. L. Mauzerall, J. Wang, E. M. Leibensperger, D. K. Henze, and K. Singh (2011), Origin and radiative forcing of black carbon transported to the Himalayas and Tibetan Plateau, Atmos. Chem. Phys., 11, 2837-2852, 2011. http://www.atmos-chem-phys.net/11/2837/2011/acp-11-2837-2011.html<br />
<br />
* Kopacz, M., D.J. Jacob, J.A. Fisher, J. A. Logan, L. Zhang, I. A. Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010): Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atmos. Chem. Phys., 10, 855-876. http://www.atmos-chem-phys.net/10/855/2010/acp-10-855-2010.pdf<br />
<br />
* Kopacz, M., D. J. Jacob, D. K. Henze, C. L. Heald, D. G. Streets, and Q. Zhang (2009), A comparison of analytical and adjoint Bayesian inversion methods for constraining Asian sources of CO using satellite (MOPITT) measurements of CO columns, J. Geophys. Res., doi:0.1029/2007JD009264. http://acmg.seas.harvard.edu/publications/KopaczJGR2009_2007JD009264.pdf<br />
<br />
* Henze, D. K., J. H. Seinfeld and D. T. Shindell (2009), Inverse modeling and mapping U.S. air quality influences of inorganic PM2.5 precursor emissions with the adjoint of GEOS-Chem, Atmos. Chem. Phys., 9, 5877-5903.<br />
<br />
* Zhang, L., D. J. Jacob, M. Kopacz, D. K. Henze, K. Singh, and D. A. Jaffe (2009), Intercontinental source attribution of ozone pollution at western U.S. sites using an adjoint method, Geophys. Res. Lett., 36, L11810, doi:10.1029/2009GL037950<br />
<br />
* Henze, D. K., A. Hakami and J. H. Seinfeld (2007), Development of the adjoint of GEOS-Chem, Atmos. Chem. Phys., 7, 2413-2433.<br />
<br />
=== Conference proceedings === <br />
* Adjoint inversion of CO sources using combined MOPITT, SCIAMACHY and AIRS CO columns, presented by Monika Kopacz at the COSPAR Scientific Assembly, Montreal, July 18, 2008. http://acmg.seas.harvard.edu/presentations/powerpoints/mak2008/COSPAR_MKopacz_July2008.ppt<br />
<br />
* Singh, K., P. Eller, A. Sandu, D. K. Henze, K. Bowman, M. Kopacz, and M. Lee (2009), Towards the construction of a standard geos-chem adjoint model, ACM High Performance Computing Conference.<br />
<br />
* Kopacz, M., Mauzerall, D.L., Leibensperger, E.M., Wang, J., Henze, D.K., Singh, K., Shim, C. Identifying the origin and estimating the radiative forcing of BC in the Himalayas: an analysis using the global GEOS-Chem adjoint model, European Geophysical Union meeting, Vienna, May 4, 2010.<br />
<br />
* Kopacz, M., Jacob, D.J., Fisher, J.A., Logan, J.A., Zhang, L., Megretskaia, I.A., Yantosca, R.M., Singh, K., Henze, D.K., Burrows, J.P., Buchwitz, M., Khlystova, I., McMillan, W.W., Gille, J.C., Edwards, D.P., Eldering, A., Thouret, V., Nedelec, P. Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), European Geophysical Union meeting, Vienna, May 7, 2010.<br />
<br />
* Tang, J., Zhuang, Q. and Xiong, X. (2010), 4D-Var inversion of atmospheric methane fluxes by assimilating SCIAMACHY and AIRS satellite retrievals, AGU, Dec. 18, 2010, http:/web.ics.purdue.edu/~tang16/agu2010_tang.ppt<br />
<br />
* Bousserez, N., R. V. Martin, K. W. Bowman, D.K. Henze, M. Kopacz, K. Singh, C. Shim, C. Wespes, Improving the lightning NOx source using satellite observations: a 4D-var analysis approach, AGU, Dec., 2010, http://myweb.dal.ca/nc689777/AGU_liNOx_poster_final.pdf</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint&diff=10746GEOS-Chem Adjoint2012-06-20T01:00:52Z<p>Hyung-Min Lee: /* Current GEOS-Chem adjoint versions under development */</p>
<hr />
<div><big><big><strong>Adjoint and Data Assimilation Working Group</strong></big></big><br />
<br />
== Contact information ==<br />
<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|-<br />
|'''Adjoint Working Group Co-Chairs'''<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman] and [mailto:dbj@atmosp.physics.utoronto.ca Dylan Jones]<br />
|-<br />
|-<br />
|'''Adjoint Model Scientist'''<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|'''Adjoint Co-Model Scientist/GC adjoint support team'''<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez]<br />
|-<br />
|'''Adjoint Working Group email list'''<br />
|<tt>geos-chem-adjoint@seas.harvard.edu</tt><br />
|-<br />
|'''To subscribe to email list'''<br />
|Send email to [mailto:geos-chem-adjoint-join@seas.harvard.edu <tt>geos-chem-adjoint-join@seas.harvard.edu</tt>]<br />
|-<br />
|'''To unsubscribe from email list'''<br />
|Send email to [mailto:geos-chem-adjoint-leave@seas.harvard.edu <tt>geos-chem-adjoint-leave@seas.harvard.edu</tt>]<br />
|}<br />
<br />
== Historical Development ==<br />
Original work on the adjoint of GEOS-Chem v6 began in 2003, focusing on the adjoint of the offline aerosol simulation. By 2005, the adjoint was expanded to include a tagged CO simulation and a full chemistry simulation; an adjoint of GEOS-Chem v7 was also developed in the following years. Each of these branches of the adjoint code were been constructed in a hybrid fashion using a combination of automatic differentiation software ([http://www.autodiff.com/tamc TAMC], [http://people.cs.vt.edu/~asandu/Software/Kpp/ KPP]) and manual coding of both discrete and continuous adjoints. They shared many common elements yet had unique features for different applications. <br />
<br />
During the summer of 2009, the existing branches were merged and updated to bring the adjoint into alignment with the latest release of GEOS-Chem, v8-02-01. This merged adjoint model is now the standard adjoint code into which all further development efforts will be placed.<br />
<br />
== Forward Model Code ==<br />
<br />
The forward model on which the adjoint is based originally corresponded to GEOS-Chem v8-02-01. It was subsequently updated as follows:<br />
<br />
* KPP solver for gas-phase chemistry (as in GCv8-02-03)<br />
* Implement Bond 2007 BC/OC emissions (as in GCv8-02-02)<br />
* Apply bug fixes from GCv8-02-02 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-02#Previous_issues_now_resolved_in_v8-02-02 here]<br />
* Apply bug fixes from GCv8-02-03 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-03#Previous_issues_now_resolved_in_v8-02-03 here]<br />
* Apply bug fixes from GCv8-02-04 listed [http://wiki.seas.harvard.edu/geos-chem/index.php/GEOS-Chem_v8-02-04#Previous_issues_now_resolved_in_v8-02-04 here]<br />
<br />
All bug fixes and model updates were previous listed at the top of inverse_driver.f. We have now switched to documenting the code development cycle here in the wiki, see the following section. <br />
<br />
== Code Versions, Bug Fixes and Developments == <br />
<br />
=== Current GEOS-Chem adjoint versions released ===<br />
[[GEOS-Chem_Adjoint_v33]] (You will get the latest version on the list when you check out.)<br />
<br />
=== Previous GEOS-Chem adjoint versions released ===<br />
*[[GEOS-Chem_Adjoint_v32]]<br />
*[[GEOS-Chem_Adjoint_v31]]<br />
<br />
== Summary of Main Adjoint Code Supported Features ==<br />
<br />
=== Features === <br />
* Meteorological fields<br />
** GEOS-3 <i>needs testing</i><br />
** GEOS-4<br />
** GEOS-5 <br />
* model resolution<br />
** 4 x 5<br />
** 2 x 2.5 <br />
** Nested Asia and NA<br />
* Forward model processes<br />
** convection<br />
** advection<br />
** PBL mixing<br />
** dry deposition<br />
** wet deposition<br />
** strat / trop exchange with LINOZ and new GMI strat chem (v9-01-03)<br />
** NOy up fluxes (now replaced with new GMI strat chem)<br />
** aerosols <br />
*** inorganic aerosol thermodynamics with RPMARES<br />
*** inorganic aerosol thermodynamics with ISORROPIA <i> in progress </i><br />
*** sulfate chemistry <br />
*** BC<br />
*** SOA, Dust, sea salt <i> needs doing </i><br />
*** aerosol surface area feedbacks <i> needs updating </i><br />
*** aerosol optical feedbacks <i> needs doing </i><br />
** emissions<br />
*** all standard emissions included<br />
* Simulation modes<br />
** full chemistry<br />
** tagged CO<br />
** tagged Ox <br />
** CH4<br />
** offline aerosols (for BC and dust only)<br />
** CO2<br />
* Observational Operators<br />
** MOPITT CO column<br />
** SCIAMACHY CO column<br />
** AIRS CO column<br />
** IMPROVE BC<br />
** CASTNet (NH4+) <i> needs updating </i> <br />
** GOME / SCIAMACHY NO2 column <i> needs updating </i> <br />
*** using KNMI retrieval (Henze)<br />
*** using Dalhousie retrieval (Shim)<br />
** SCIAMACHY/OMI NO2<br />
*** using Dalhousie retrieval (Bousserez, Padmanabhan)<br />
** TES NH3 <br />
** TES O3 <br />
** GOSAT CO2 <br />
** MLS O3 and TES CO2 <i> in progress </i><br />
* Control parameters<br />
** Initial Conditions scaling factors (linear or log)<br />
** Emissions scaling factors (linear or log) <br />
*** NH3, primary BC/OC, SO2: anthropogenic, natural, bioburn, biomass, ship<br />
*** NOx: soil, aircraft, anthropogenic, biofuel, bioburn<br />
*** Lightning NOx: injection height, yield <i> in progress </i><br />
*** all other gas-phase tracers: anthropogenic, biofuel, bioburn <br />
* Adjoint sensitivities<br />
** w.r.t. all implemented control parameters<br />
** w.r.t Reaction Rate Parameters <br />
** w.r.t all emissions <br />
** of AQ attainment metrics <i> needs updating </i><br />
** of spatiotemporally averaged species concentrations (e.g., arctic O3) <br />
* Other<br />
** Inverse Hessian approximation <br />
** off-diagonal covariance matrices <i> needs updating </i><br />
** 3D-Var <i> needs updating </i><br />
<br />
Features may be qualified as: <br />
* <i>needs testing</i>: an implemented feature that we haven't fully used yet<br />
* <i>needs updating</i>: a feature developed with a previous branch that has yet to be updated to GEOS-Chem v8 and the merged adjoint<br />
* <i>needs doing</i>: a feature nobody has tackled the adjoint of yet<br />
* <i>in progress</i>: a feature currently under development <br />
* <i>in pipeline</i>: a feature which has been submitted and awaiting integration into the CVS repository<br />
<br />
=== Primary code developers ===<br />
Monika Kopacz, Kumaresh Singh, Changsub Shim, Daven Henze<br />
<br />
=== Adjoint model lead scientist ===<br />
Daven Henze<br />
<br />
<br />
<br />
== Resources ==<br />
<br />
=== User's guide ===<br />
A user's guide is available. http://adjoint.colorado.edu/%7Edaven/gcadj_std/GC_adj_man.pdf <br />
<br />
=== Code flowchart ===<br />
Meemong Lee has created a detailed flowchart of the inverse model code structure. http://adjoint.colorado.edu/~daven/gcadj_std/flowchart.pdf<br />
<br />
=== Plotting tools ===<br />
<br />
Some IDL and MATLAB routines for plotting benchmark results. http://adjoint.colorado.edu/~daven/gcadj_std/tools.tar.gz<br />
<br />
=== Background papers and presentations ===<br />
<br />
Several articles and presentations (including a GC adjoint modeling clinic overview from IGC5) providing background information about adjoints. http://adjoint.colorado.edu/~daven/gcadj_std/adj_articles.tar.gz<br />
<br />
== Distribution and Use ==<br />
Code for the adjoint is distributed through a CVS server located at adjoint.colorado.edu. Contact Daven Henze to obtain an account on the server.<br />
<br />
Even if your office mate has a copy of the code, the best way to obtain the model is to get a CVS account for yourself and download a version from the repository. So please do not copy code directly from others or pass the code along to third parties. This vastly helps with tracking developments and keeping up with model updates. <br />
<br />
Use of the adjoint model code follows standard practice for GEOS-Chem. It is expected that any developments that come of individual applications based on this community model will eventually be given back to the community by incorporation of new developments into the standard adjoint code. New development should be submitted to Daven Henze for inclusion in the standard adjoint model code.<br />
<br />
=== Quick guide to CVS === <br />
We recommend first taking a look at CVS manual to get a general feel for how this tools works (e.g., http://cvsbook.red-bean.com/cvsbook.html). <br />
<br />
Below are some command commands you may use for developing code and checking the status of code updates.<br />
<br />
Obtain the latest code:<br />
<br />
cvs checkout gcadj_std<br />
<br />
<br />
Generate a list (modified_files.txt) of all the files in your local copy that differ from the current repository code:<br />
<br />
cvs -q status | grep 'Status' | grep -v 'Up-to-date' > modified_files.txt<br />
<br />
<br />
Determine the difference between your local copy of a file and the version that you originally checked out (i.e., see what you changed):<br />
<br />
cvs diff foo_mod.f<br />
<br />
Note: arguments such as a filename are optional. Without listing a specific file, cvs will run the command on all files in the current directory. <br />
<br />
<br />
Determine the difference between your local copy of the code the current version in the repository (i.e., see both what you changed, and what has changed in the repository):<br />
<br />
cvs diff -D "now" foo_mod.f<br />
<br />
<br />
Merge your local file with the current repository version<br />
<br />
cvs update foo_mod.f <br />
<br />
<br />
Obtain a fresh copy of a file in the repository (without merging)<br />
<br />
rm foo_mod.f<br />
cvs update foo_mod.f <br />
<br />
<br />
Difference between the local repository and the tagged version “TAGNAME”<br />
<br />
cvs diff -r TAGNAME (< file name > for specific file)<br />
<br />
<br />
Difference between the two tagged versions “TAGNAME1” and “TAGNAME2” in the repository (shows modified files only):<br />
<br />
cvs diff -N -c -r TAGNAME2 -r TAGNAME2 | grep "Index:" > output.txt<br />
<br />
Please do not use the commit command, which is restricted to the GCadj support team (Daven Henze and Nicolas Bousserez).<br />
<br />
== Crediting GEOS-Chem adjoint developers ==<br />
<br />
We aim to make distribution of adjoint model code as immediate as possible. A consequence is that many features may not yet be publicly documented. Therefore, giving code developers due credit is of utmost importance. <br />
<br />
Authors of new additions to the standard code should be offered co-authorship on the first round of presentations and publications to come of their development. Features currently falling in this category and their developers are:<br />
* (v32) MOPITT v5 CO observation operator. Developer: Zhe Jiang, University of Toronto.<br />
* (v32) Dust adjoint. Developer: Xiaoguang (Richard) Xu, University of Nebraska Lincoln.<br />
* (v32) Black carbon offline aerosol adjoint. Developer: Yuhao Mao (UCLA).<br />
* (v32) Nested full chemistry adjoint. Developers: Zhe Jiang, University of Toronto; Daven Henze, CU Boulder.<br />
* (v32) Stratospheric production / loss rate sensitivities. Developer: Hyungmin Lee, CU Boulder.<br />
* (v32) CH4 adjoint. Developer: Kevin Wecht, Harvard University<br />
* (v29) LIDORT. Developer: Daven Henze, University of Colorado Boulder. Collaborator: Rob Spurr. <br />
* (v28) CO2 adjoint. Developer: Daven Henze, University of Colorado Boulder. Collaborators: Ray Nassar, Kevin Bowman, Dylan Jones.<br />
<br />
Citation of the appropriate [[GEOS-Chem_Adjoint#Journal_Articles|journal articles]] for mature developments is also encouraged, as well as considering aspects of [http://acmg.seas.harvard.edu/geos/geos_credit.html co-authorship for the forward model]. <br />
<br />
Overall, if you have any questions about authorship, even for a conference presentation, please contact Daven Henze.<br />
<br />
== Current GEOS-Chem Adjoint Research Projects (please add yours!) ==<br />
{| border=1 cellspacing=0 cellpadding=5<br />
|- bgcolor="#cccccc"<br />
!User Group <br />
!Description <br />
!Contact Person<br />
|-<br />
|CU Boulder<br />
|Aerosol precursors, CO2, O3; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Daven Henze]<br />
|-<br />
|-<br />
|CU Boulder<br />
|Inverse modeling/optimization; general adjoint code maintenance<br />
|[mailto:daven.henze@colorado.edu Nicolas Bousserez]<br />
|-<br />
|Harvard<br />
|Methane<br />
|Kevin Wecht, wecht [at] fas.harvard.edu<br />
|-<br />
|Purdue University<br />
|Methane (SICAMACHY, AIRS and IASI)<br />
|[mailto:tang16@purdue.edu Jinyun Tang]<br />
|-<br />
|MIT<br />
|Aircraft emissions<br />
|[mailto:jaminkoo@mit.edu Jamin Koo]<br />
|-<br />
|Princeton<br />
|BC sensitivities, general adjoint code development<br />
|Monika Kopacz, mkopacz [at] princeton.edu<br />
|-<br />
|Dalhousie University<br />
|Lightning NOx emissions and impact on tropical ozone using the adjoint<br />
|[mailto:Nicolas.Bousserez@colorado.edu Nicolas Bousserez (now at CU-Boulder)]<br />
|-<br />
|Dalhousie University<br />
|Surface NOx emissions inversion using SCIAMACHY/OMI NO2 measurements<br />
|Akhila Padmanabhan akhila [at] dal.ca; Nicolas Bousserez [mailto:Nicolas.Bousserez@colorado.edu] (now at CU-Boulder)<br />
|-<br />
|JPL<br />
|Microwave Limb Sounder (MLS) Ozone assimilation<br />
|[mailto:meemong.lee@jpl.nasa.gov Meemong Lee]<br />
|-<br />
|JPL <br />
| TES ozone assimilation/attribution of ozone radiative forcing<br />
|[mailto:kevin.bowman@jpl.nasa.gov Kevin Bowman]<br />
|-<br />
|University of Edinburgh<br />
|Quantifying the impact of boreal forest fires on tropospheric oxidants over the Atlantic<br />
|[mailto:mark.parrington@ed.ac.uk Mark Parrington]<br />
|-<br />
|US EPA<br />
|Integration with economic models for future emission inventory scenario development<br />
|[mailto:akhtar.farhan@epa.gov Farhan Akhtar]<br />
|-<br />
|Peking University<br />
|Satellite constraints on VOC emissions <br />
|[mailto:tmfu@pku.edu.cn May Fu]<br />
|-<br />
|CU Boulder<br />
|Spatial extent of source influences on aerosol precursor columns<br />
|[mailto:alexander.turner@colorado.edu Alex Turner]<br />
|-<br />
|IAP.CAS<br />
|CO2 assimilation<br />
|[mailto:czh@mail.iap.ac.cn Chen]<br />
|-<br />
|Purdue University<br />
|Feedback between terrestrial ecosystem processes and atmospheric co2 signals<br />
|[mailto:zhuq@purdue.edu Qing Zhu]<br />
|-<br />
|University of Toronto<br />
|Sensitivity of ozone and reactive nitrogen to precursor emissions<br />
|[mailto:twalker@atmosp.physics.utoronto.ca Thomas Walker]<br />
|-<br />
|University of Toronto<br />
|Adjoint analysis for carbon monoxide <br />
|[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]<br />
|-<br />
|Georgia Tech<br />
|ISORROPIA adjoint development; inorganic aerosol precursors<br />
|[mailto:scapps@gatech.edu Shannon Capps]<br />
|-<br />
|Peking University<br />
|Source attributions of tropospheric ozone over North China<br />
|[mailto:linjt@pku.edu.cn Jintai Lin]<br />
|-<br />
|University of Wollongong<br />
|Sensitivity of ozone and adjoint analysis of CO over Australasia.<br />
|[mailto:rb864@uowmail.edu.au Rebecca Buchholz]<br />
|-<br />
|University of Toronto<br />
|CO2 assimilation & transport model bias estimation<br />
|[mailto:mkeller@atmosp.physics.utoronto.ca Martin Keller]<br />
|-<br />
|Dalhousie University<br />
|Sensitivity of global PM2.5-induced mortality to emissions<br />
|[mailto:colin.lee@dal.ca Colin Lee]<br />
|-<br />
|University of Leicester (UK)<br />
|Top-down estimates of Amazon isoprene emissions<br />
|[mailto:mpb14@le.ac.uk Michael Barkley]<br />
|-<br />
|Anyang University<br />
|Aerosol emission modeling in East Asia <br />
|[mailto:koo@anyang.ac.kr Youn Seo Koo]<br />
|-<br />
|Nanjing University<br />
|Inverse modeling of terrestrial ecosystem carbon flux<br />
|[mailto:hengmao.wang@gmail.com Hengmao Wang]<br />
|-<br />
|Tsinghua University<br />
|Nested-gird simulations with the adjoint model<br />
|[mailto:ny07@mails.tsinghua.edu.cn Nan Yang]<br />
|-<br />
|University of Minnesota<br />
|Inverse modeling of VOC sources based on TES and IASI measurements<br />
|[mailto:dbm@umn.edu Dylan Millet]<br />
|}<br />
<br />
== Publications ==<br />
<br />
=== Journal Articles ===<br />
*Wang, J., X. Xu, D. K. Henze, Q. Ji, S.-C. Tsay, J. Huang, Top-Down Estimate of Dust Emissions through Integration of MODIS and MISR Aerosol Retrievals with the GEOS-Chem adjoint model, submitted.<br />
<br />
*Parrington, M., P. I. Palmer, D. K. Henze, D. W. Tarasick, E. J. Hyer, R. C. Owen, C. Clerbaux, K. W. Bowman, M. N. Deeter, E. M. Barratt, P.-F. Coheur, D. Hurtmans, M. George, and J. R. Worden (2012), The influence of boreal biomass burning emissions on the distribution of tropospheric ozone over North America and the North Atlantic during 2010, Atmos. Chem. Phys., 12, 2077-2098<br />
<br />
*Bowman, K. W., and D. K. Henze, Attribution of direct ozone radiative forcing to spatially-resolved emissions, submitted.<br />
<br />
*Paulot, F., D. K. Henze, and P. O. Wennberg (2012), Impact of the isoprene photochemical cascade on tropical ozone, Atmos. Chem. Phys., 12, 1307-1325, 2012.<br />
<br />
*Henze, D. K., D. T. Shindell, F. Akhtar, R. J. D. Spurr, R. W. Pinder, D. Loughlin, M. Kopacz, K. Singh, and C. Shim, Spatially refined aerosol direct radiative forcing efficiencies, submitted.<br />
<br />
*Turner, A., D. K. Henze, R. V. Martin, and A. Hakami (2012), The spatial extent of source influences on modeled column concentrations of short-lived species, ''Geophys. Res. Lett.'', doi: 10.1029/2012GL051832, in press.<br />
<br />
*Jiang, Z., D. B. A. Jones, H. M. Worden, M. N. Deeter, D. K. Henze, J. Worden, and K. W. Bowman, Quantifying the impact of model biases in convective transport on inferred CO source estimates using multi-spectral CO retrievals from MOPITT, submitted.<br />
<br />
*Wecht, K. J., D. J. Jacob, S. C. Wofsy, E. A. Kort, J. R. Worden, S. S. Kulawik, D. K. Henze, M. Kopacz, and V. H. Payne, Validation of TES methane with HIPPO aircraft observations: implications for inverse modeling of methane sources, Atmos. Chem. Phys. Discuss., 11, 27887-27911.<br />
<br />
*Walker, T., D. B. A. Jones, M. Parrington, D. K. Henze, L. T. Murray, J. W. Bottenheim, K. Anlauf, J. R. Worden, K. W. Bowman, C. Shim, K. Singh, M. Kopacz, D. W. Tarasick, J. Davies, P. von der Gathen, and C. C. Carouge (2012), Impacts of midlatitude precursor emissions and local photochemistry on ozone abundances in the Arctic, J. Geophys. Res., doi:10.1029/2011JD016370.<br />
<br />
*Jiang, Z., D. B. A. Jones, M. Kopacz, J. Liu, D. K. Henze, and C. Heald (2011), Quantifying the impact of model errors on top-down estimates of carbon monoxide emissions using satellite observations, J. Geophys. Res., 116, D15306, doi:10.1029/2010JD015282.<br />
<br />
*Singh, K., A. Sandu, Variational Chemical Data Assimilation with Approximate Adjoints, submitted. <br />
<br />
*Singh, K., Jardak, M., Sandu, A., Bowman, K., Lee, M., and Jones, D. (2010): Construction of non-diagonal background error covariance matrices for global chemical data assimilation, Geosci. Model Dev. Discuss., 3, 1783-1827, doi:10.5194/gmdd-3-1783-2010. http://www.geosci-model-dev-discuss.net/3/1783/2010/gmdd-3-1783-2010.html<br />
<br />
*Kopacz, M., D. L. Mauzerall, J. Wang, E. M. Leibensperger, D. K. Henze, and K. Singh (2011), Origin and radiative forcing of black carbon transported to the Himalayas and Tibetan Plateau, Atmos. Chem. Phys., 11, 2837-2852, 2011. http://www.atmos-chem-phys.net/11/2837/2011/acp-11-2837-2011.html<br />
<br />
* Kopacz, M., D.J. Jacob, J.A. Fisher, J. A. Logan, L. Zhang, I. A. Megretskaia, R. M. Yantosca, K. Singh, D. K. Henze, J. P. Burrows, M. Buchwitz, I. Khlystova, W. W. McMillan, J. C. Gille, D. P. Edwards, A. Eldering, V. Thouret, and P. Nedelec (2010): Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), Atmos. Chem. Phys., 10, 855-876. http://www.atmos-chem-phys.net/10/855/2010/acp-10-855-2010.pdf<br />
<br />
* Kopacz, M., D. J. Jacob, D. K. Henze, C. L. Heald, D. G. Streets, and Q. Zhang (2009), A comparison of analytical and adjoint Bayesian inversion methods for constraining Asian sources of CO using satellite (MOPITT) measurements of CO columns, J. Geophys. Res., doi:0.1029/2007JD009264. http://acmg.seas.harvard.edu/publications/KopaczJGR2009_2007JD009264.pdf<br />
<br />
* Henze, D. K., J. H. Seinfeld and D. T. Shindell (2009), Inverse modeling and mapping U.S. air quality influences of inorganic PM2.5 precursor emissions with the adjoint of GEOS-Chem, Atmos. Chem. Phys., 9, 5877-5903.<br />
<br />
* Zhang, L., D. J. Jacob, M. Kopacz, D. K. Henze, K. Singh, and D. A. Jaffe (2009), Intercontinental source attribution of ozone pollution at western U.S. sites using an adjoint method, Geophys. Res. Lett., 36, L11810, doi:10.1029/2009GL037950<br />
<br />
* Henze, D. K., A. Hakami and J. H. Seinfeld (2007), Development of the adjoint of GEOS-Chem, Atmos. Chem. Phys., 7, 2413-2433.<br />
<br />
=== Conference proceedings === <br />
* Adjoint inversion of CO sources using combined MOPITT, SCIAMACHY and AIRS CO columns, presented by Monika Kopacz at the COSPAR Scientific Assembly, Montreal, July 18, 2008. http://acmg.seas.harvard.edu/presentations/powerpoints/mak2008/COSPAR_MKopacz_July2008.ppt<br />
<br />
* Singh, K., P. Eller, A. Sandu, D. K. Henze, K. Bowman, M. Kopacz, and M. Lee (2009), Towards the construction of a standard geos-chem adjoint model, ACM High Performance Computing Conference.<br />
<br />
* Kopacz, M., Mauzerall, D.L., Leibensperger, E.M., Wang, J., Henze, D.K., Singh, K., Shim, C. Identifying the origin and estimating the radiative forcing of BC in the Himalayas: an analysis using the global GEOS-Chem adjoint model, European Geophysical Union meeting, Vienna, May 4, 2010.<br />
<br />
* Kopacz, M., Jacob, D.J., Fisher, J.A., Logan, J.A., Zhang, L., Megretskaia, I.A., Yantosca, R.M., Singh, K., Henze, D.K., Burrows, J.P., Buchwitz, M., Khlystova, I., McMillan, W.W., Gille, J.C., Edwards, D.P., Eldering, A., Thouret, V., Nedelec, P. Global estimates of CO sources with high resolution by adjoint inversion of multiple satellite datasets (MOPITT, AIRS, SCIAMACHY, TES), European Geophysical Union meeting, Vienna, May 7, 2010.<br />
<br />
* Tang, J., Zhuang, Q. and Xiong, X. (2010), 4D-Var inversion of atmospheric methane fluxes by assimilating SCIAMACHY and AIRS satellite retrievals, AGU, Dec. 18, 2010, http:/web.ics.purdue.edu/~tang16/agu2010_tang.ppt<br />
<br />
* Bousserez, N., R. V. Martin, K. W. Bowman, D.K. Henze, M. Kopacz, K. Singh, C. Shim, C. Wespes, Improving the lightning NOx source using satellite observations: a 4D-var analysis approach, AGU, Dec., 2010, http://myweb.dal.ca/nc689777/AGU_liNOx_poster_final.pdf</div>Hyung-Min Leehttps://wiki.seas.harvard.edu/geos-chem/index.php?title=GEOS-Chem_Adjoint_v32&diff=9779GEOS-Chem Adjoint v322012-03-06T20:46:43Z<p>Hyung-Min Lee: /* Add strat production and loss rates as active control parameters (adj32_025) */</p>
<hr />
<div>== Overview ==<br />
Under development. Expected release date 02/15/2012<br />
<br />
*Previous version: [[GEOS-Chem_Adjoint_v31]]<br />
*Next version: [[GEOS-Chem_Adjoint_v33]]<br />
<br />
== What's new in this version ==<br />
<br />
=== Updates to forward model ===<br />
<br />
----<br />
<br />
==== Update tagged CO simulation (adj32_017) ====<br />
Several updates from Zhe Jiang have been implemented for the tagged-CO implementation. See also changes to the adjoint model, [[GEOS-Chem_Adjoint_v32#Update_tagged_CO_simulation_.28adj32_017.29]]<br />
<br />
===== Update to tagged_co_mod.f =====<br />
<br />
<tt>tagged_co_mod.f</tt> has been updated to distribute emissions through the boundary layer and removing obsolete COSF, TIME_STEPS, GET_CO_CH4, SET_CO_CH4, and GET_OPT_CO.<br />
<br />
Also, F_OF_PBL and ED_CO were declared thread PRIVATE in several parallel do loops. <br />
<br />
<br />
===== Updates to <tt>time_mod.f</tt> for tagged-CO =====<br />
'''''[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]''''' wrote<br />
<br />
:I currently found it perhaps can only affect tagged-CO inversion. Thus, it can not be said as a BUG.<br />
<br />
:For example, if we do our inversion for June 1-15, when the backward running, because June 15 is not the last day of the calendar month, these emissions will not be read. However, we don't need to read them since the model will use the value in the forward run. For the tagged-CO backward run, the CH4 concentration will only be read if June 15 is the last day of the calendar year. Thus, in the original code, the CH4 concentration, and thus gradient on CH4 oxidation, is ZERO.<br />
<br />
IF ( ITS_A_NEW_YEAR() ) THEN<br />
CALL GET_GLOBAL_CH4( GET_YEAR(), .TRUE., A3090S, A0030S, A0030N, A3090N )<br />
ENDIF<br />
<br />
:In my previous code, I saved the CH4 concentration in the forward run and then read them in the backward run. After I modified time_mod.f, I have removed the code for the CH4 saving/reading. Anyway, the effect is trivial. I can solve the problem in the tagged_CO inversion by modifying the previous code.<br />
<br />
To address these issues, subroutines ITS_A_NEW_MONTH and ITS_A_NEW_YEAR have been updated in <tt>time_mod.f</tt>. <br />
<br />
<br />
--[[User:Daven|Daven]] 18:18, 8 January 2012 (EST)<br />
<br />
==== Update several files to v9 code to support full-chem nested simulation (adj32_015) ====<br />
Updates to the following files were made by Zhe Jiang to support the NESTED_NA and NESTED_SD simulation:<br />
<br />
===== Update DEAD dust emissions for nested simulations and MERRA (v9-01-01) =====<br />
When the forward model was updated to support MERRA, the dust emissions got retuned [[MERRA_implementation_details#Dust_emissions]]. But also the emissions for GEOS-5 were affected, and support was added for nested grid. <br />
<br />
The <tt>dust_dead_mod.f</tt> file has been updated to match that in v9-01-01 (dust_dead_mod.f,v 1.3 2010/02/02 16:57:53 bmy). Common blocks CMN_SIZE and CMN_GCTM are reinstated for backwards compatibility.<br />
<br />
===== Updated TOMS O3 =====<br />
<tt>toms_mod.f</tt> was updated to v9-01-02. This uses the newer TOMS inventory TOMS_200906 and allows for rescaling of years up to 2008 (previous version stopped at 2007). <br />
<br />
===== cac_mod.f =====<br />
<tt>cac_mod.f</tt> was updated to v9-01-02.<br />
<br />
===== emissions_mod.f =====<br />
<tt>emissions_mod.f</tt> was updated to support GRID05x0666 for CAC, NEI05 and EMEP emissions following v9-01-02.<br />
<br />
===== file_mod.f =====<br />
<tt>file_mod.f</tt> was updated to include IU_BC_05x06.<br />
<br />
===== CMN_SIZE =====<br />
<tt>CMN_SIZE</tt> was updated to include size of NA and SD nested domain.<br />
<br />
===== commsoil.h =====<br />
<tt>commsoil.h</tt> was updated to include size of NA and SD nested domain.<br />
<br />
===== define.h =====<br />
<tt>define.h</tt> was updated to include the NESTED_SD option.<br />
<br />
===== tpcore_bc_mod.f =====<br />
<tt>tpcore_bc_mod.f</tt> has been updated to include <br />
*OPEN_BC_FILE_05x0666<br />
*SAVE_GLOBAL_TPCORE_BC_05x0666<br />
*DO_WINDOW_TPCORE_BC_05x0666<br />
*CLEAN_WINDOW_TPCORE_BC_05x0666<br />
*READ_WINDOW_TPCORE_BC_05x0666<br />
*ITS_TIME_FOR_BC_05x0666INIT_TPCORE_BC_05x0666<br />
and the BC_05x06 array.<br />
<br />
===== regrid_1x1_mod.f =====<br />
<tt>regrid_1x1_mod</tt> has been updated to support the NESTED_SD domain.<br />
<br />
===== Update lightning_nox_mod.f =====<br />
See [[GEOS-Chem_Adjoint_v32#Updated_lightning_parameterization_and_fix_for_cloud-top-height_algorithm_.28v9-01-01.29]]<br />
<br />
===== acetone_mod.f =====<br />
<tt>acetone_mod.f</tt> has been updated to support the nested domains. It now matches v9-01-02, with the additional update for the adjoint that we do not multiply and divide by ACETONE, thus making OCEAN_SINK_ACET self-adjoint.<br />
<br />
===== input_mod.f =====<br />
<tt>input_mod.f</tt> has been updated in several places to support nested domains. <br />
<br />
===== transport_mod.f =====<br />
<tt>transport_mod.f </tt> has been updated to support the NESTED_SD domain.<br />
<br />
===== co2_mod.f =====<br />
<tt>co2_mod.f</tt> has been updated to support nested domains.<br />
<br />
<br />
--[[User:Daven|Daven]] 15:00, 18 January 2012 (EST)<br />
<br />
<br />
===== nei05 and input_mod.f =====<br />
<tt>nei2005_anthro_mod.f</tt> and <tt>input_mod.f</tt> have been updates, see [[GEOS-Chem_Adjoint_v32#Fixes_for_minor_issues_affecting_nested-grid_simulations_.28v9_01_02e.29]]<br />
<br />
==== Update CH4 simulation (adj32_023) ====<br />
[[GEOS-Chem_Adjoint_v32#Include_adjoint_of_CH4_simulation_.28adj32_023.29]]<br />
<br />
==== Update stratospheric chemistry and prod/loss rates (adj32_025) ====<br />
[[GEOS-Chem_Adjoint_v32#Add_strat_production_and_loss_rates_as_active_control_parameters_.28adj32_025.29]]<br />
<br />
=== Bug fixes in forward model ===<br />
<br />
----<br />
<br />
==== Prevent overflow / underflow in rpmares (adj32_001) ====<br />
<br />
===== Ionic strength below zero =====<br />
On prospero, we often get the message: <tt>Ionic strength below zero...negative concentrations</tt>.<br />
<br />
Despite the setting of negative roots to zero, sometimes the code still ends up with negative CRUTES. Trap for it here to prevent mysterious crashing later <br />
<br />
! Trap for negative CRUTES (dkh, 01/06/12, adj32_001)) <br />
IF ( CRUTES(1) < 0d0 ) THEN<br />
IF ( CRUTES(1) .ne. ABS(CRUTES(1) ) ) CRUTES(1)=1d9<br />
IF ( CRUTES(2) .ne. ABS(CRUTES(2) ) ) CRUTES(2)=1d9<br />
IF ( CRUTES(3) .ne. ABS(CRUTES(3) ) ) CRUTES(3)=1d9<br />
CRUTES(1) = MIN( CRUTES(1), CRUTES(2), CRUTES(3) )<br />
IF ( CRUTES(1) < 0d0 ) THEN<br />
print*, ' CRUTES = ', CRUTES , NR<br />
CALL ERROR_STOP( ' checking for neg val failed',<br />
& ' rpmares_mod.f ' )<br />
ENDIF<br />
ENDIF<br />
<br />
In addition, reinstate the following<br />
<br />
! reinstate (dkh, 02/12/12, adj32_001) <br />
IF (HPLUS <= 0d0) GOTO 100<br />
--[[User:Daven|Daven]] 14:58, 6 January 2012 (EST)<br />
<br />
===== RPMARES: not safe to divide...exit =====<br />
When running with LADJ = F, this error would arise. The reason is not entirely clear, but with the exact same inputs, it sometimes goes away by adding a print statement here or there, so is likely owing to overflow / underflow within RPMARES somewhere. The RPMARES_FORADJ routine is not subject to this instability, so now we move the LADJ flag within RPMARES_FORADJ and always use this routine. <br />
<br />
--[[User:Daven|Daven]] 12:32, 12 February 2012 (EST)<br />
<br />
<br />
===== Compiler level fix =====<br />
'''''[mailto:fabienpaulot@gmail.com Fabien Paulot]''''' wrote<br />
<br />
:Hi Daven,<br />
<br />
:I am sorry to bug you again about that issue but this time I may have a solution. It seems the problem with RPMARES is related with the compilation option.<br />
<br />
:I don't get the error ("not safe to divide") when I use the following option<br />
<br />
:FFLAGS = -cpp -w -O2 -auto -noalign -convert big_endian -fp-model precise and remove the -mp option from the F90 compiler options.<br />
<br />
:(Note switching to O2 is not the reason why it's working but Bob suggested not to use O3).<br />
<br />
<br />
We now implement the -fp-model precise flag and do not use the -mp flag. The latter was originally implemented to allow KPP code to run on 32-bit systems. <br />
<br />
This compiler level fix does address the two issues with rpmares noted above on prospero. But we will leave those patches in the code for now, as they don't hurt, and we'll see how this plays out on other platforms.<br />
<br />
--[[User:Daven|Daven]] 00:34, 1 March 2012 (EST)<br />
<br />
==== Updated lightning parameterization and fix for cloud-top-height algorithm (v9-01-01) ====<br />
Updates to <tt>a6_read_mod.f</tt> and <tt>lightning_nox_mod.f</tt>. Details in Lee's description: http://people.seas.harvard.edu/~ltmurray/LNOx.v9-01-01.Release.Notes.pd<br />
<br />
--[[User:Daven|Daven]] 15:18, 6 January 2012 (EST)<br />
<br />
==== Fix for GEIA emissions scaling factor over Botswana (v9-01-02h) ====<br />
Updates to <tt>scale_anthro_mod.f</tt>. See [[GEOS-Chem_v9-01-02#Fix_for_GEIA_emissions_scaling_factor_over_Botswana]]<br />
<br />
--[[User:Daven|Daven]] 16:16, 6 January 2012 (EST)<br />
<br />
==== TINY parameter in convection_mod.f (v8-03-02) ====<br />
Update to <tt>convection_mod.f</tt> [[GEOS-Chem_v8-03-02#TINY_parameter_in_convection_mod.f]]<br />
<br />
--[[User:Daven|Daven]] 16:16, 6 January 2012 (EST)<br />
<br />
==== Updated to REGRID_05x0666_NESTED for non-standard domains (v?-?-?) ====<br />
'''''[mailto:zjiang@atmosp.physics.utoronto.ca Zhe Jiang]''''' wrote<br />
<br />
:I revised the SUBROUTINE REGRID_05x0666_NESTED.<br />
<br />
:The problem in the original code is that the Longitude of LL (Left-Low of nested window) has to be even number and the latitude of LL has to be integer number. It has no influence on our standard NA/CH/EU nested setting.<br />
<br />
:I think it would be safer to fix this problem because we perhaps need to define new nested window in future.<br />
<br />
This subroutine has now been replaced with Zhe's update. Not sure if there is a corresponding update in the v9 forward model yet. <br />
<br />
--[[User:Daven|Daven]] 19:06, 6 January 2012 (EST)<br />
<br />
==== Fix variable declarations in UPBDFLX_NOY (adj32_008) ====<br />
'''''[mailto:twalker@atmosp.physics.utoronto.ca Thomas Walker]''''' noted that having DTDYN, AIRDENS, PNOY declared as REAL*4 can lead to overflow in the adjoint model. The <tt>upbdflx_mod.f</tt> has now been updated to use REAL*8 instead. <br />
<br />
--[[User:Daven|Daven]] 18:00, 8 January 2012 (EST)<br />
<br />
==== NEI05 ID names for non-full chemistry simulation (adj32_015) ====<br />
<br />
Explanation from Zhe...<br />
<br />
IF ( .not. ITS_A_FULLCHEM_SIM() ) THEN !Zhe (16/3/2011)<br />
! Zhe (16/3/2011), (dkh, <br />
SPECIES_ID_SAVE = (/ IDTNOX, IDTCO, IDTSO2, IDTSO4, IDTNH3,<br />
$ IDTACET, IDTALK4, IDTC2H6, IDTC3H8,<br />
$ IDTOCPI, IDTBCPI,<br />
$ IDTALD2, IDTCH2O, IDTPRPE, IDTMEK<br />
$ /) <br />
<br />
IDTNOX = 1<br />
IDTCO = 4<br />
IDTSO2 = 26<br />
IDTSO4 = 27<br />
IDTNH3 = 30<br />
IDTACET = 9<br />
IDTALK4 = 5<br />
IDTC2H6 = 21<br />
IDTC3H8 = 19<br />
IDTOCPI = 35<br />
IDTBCPI = 34<br />
IDTALD2 = 11<br />
IDTCH2O = 20<br />
IDTPRPE = 18<br />
IDTMEK = 10<br />
<br />
ENDIF<br />
<br />
--[[User:Daven|Daven]] 19:46, 16 January 2012 (EST)<br />
<br />
==== Fixes for minor issues affecting nested-grid simulations (v9_01_02e) ====<br />
<br />
Updates to <tt>nei2005_anthro_mod.f</tt> and <tt>input_mod.f</tt>, see [[GEOS-Chem_v9-01-02#Fixes_for_minor_issues_affecting_nested-grid_simulations]]<br />
<br />
==== Fixes for tagged CO EMEP emissions (adj32_019) ====<br />
To allow for tagged-CO simulation to use the EMEP emissions, use hardwired tracer numbers in emep_mod.f This should eventually be migrated into the standard forward model code. <br />
<br />
--[[User:Daven|Daven]] 15:23, 9 February 2012 (EST)<br />
<br />
==== Fix for nested boundary condition (adj32_021) ====<br />
In <tt>tpcore_bc_mod.f</tt>, subroutine DO_WINDOW_TPCORE_BC_05x0666, the OpenMP statements were causing problems, so remove these for now. <br />
<br />
--[[User:Daven|Daven]] 15:23, 9 February 2012 (EST)<br />
<br />
==== Fix for nested boundary condition (adj32_022) ====<br />
Updated to<tt>transport_mod.f</tt> and <tt>diag3.f</tt>. This issue has been corrected in the standard forward model with the following note:<br />
<br />
! Note: the mass flux diagnostic arrays (MASSFLEW, MASSFLNS and MASSFLUP)<br />
! are incremented upside-down (level 1 = top of the atmosphere).<br />
! The levels order is reversed only when written out in diag3.f<br />
! (ccc, 3/8/10),<br />
<br />
--[[User:Daven|Daven]] 15:23, 9 February 2012 (EST)<br />
<br />
--[[User:Daven|Daven]] 19:51, 16 January 2012 (EST)<br />
<br />
=== Updates to adjoint model ===<br />
<br />
<br />
----<br />
<br />
==== Updates to aerosol direct radiative forcing adjoint ====<br />
Several updates in <tt>lidort_mod.f</tt> and <tt>mie_mod.f</tt>:<br />
*From Farhan Akhtar, now include primary OC in the LIDORT radiative forcing calculation. <br />
*Now switch back to using analytic derivatives from mie table rather than FD<br />
<br />
--[[User:Daven|Daven]] 18:50, 8 January 2012 (EST)<br />
<br />
==== Dust adjoint and MODIS AOD observation operator (adj32_011) ====<br />
From Richard Xiaoguang Xu at UNL, the dust adjoint includes sensitivity with respect to dust emissions from the DEAD module. Also included in this update is an observation operator for assimilating AOD observations from MODIS using the UNL retrieval. <br />
<br />
--[[User:Daven|Daven]] 00:23, 13 January 2012 (EST)<br />
<br />
==== Inverse Hessian approximation (adj32_012) ====<br />
The inverse Hessian approximation (e.g., Henze et al., 2009) using the DFP formula is now implemented with the LINVH flag in <tt>input.gcadj</tt>. Instructions for how to perform this calculation are included in the UPDATE_HESSIAN subroutine in the new <tt>inv_hessian_mod.f</tt> file. <br />
<br />
--[[User:Daven|Daven]] 00:23, 13 January 2012 (EST)<br />
<br />
==== Offline aerosol adjoint. (adj32_013) ====<br />
Yuhao Mao and Qinbin Li at UCLA have implemented the offline aerosol simulation adjoint for black carbon. This update also included an observation operator for measurements of these species at IMPROVE sites, IMPROVE_BC_OC_OBS and <tt>improve_bc_mod.f</tt>.<br />
<br />
The offline aerosol adjoint will also work for the recently implemented dust aerosol adjoint (it does not yet include support for sulfate aerosol, sea salt or SOA). For example, you may now run with <tt>input.geos</tt><br />
<br />
%%% TRACER MENU %%% : <br />
Type of simulation : 10<br />
Number of Tracers : 4<br />
Tracer Entries -------> : TR# Name g/mole Tracer Members; () = emitted<br />
Tracer #1 : 1 BCPI 12.0<br />
Tracer #2 : 2 OCPI 12.0<br />
Tracer #3 : 3 BCPO 12.0<br />
Tracer #4 : 4 OCPO 12.0<br />
<br />
and <tt>input.gcadj</tt><br />
<br />
FOR LADJ_EMS :<br />
NNEMS: ems groups implemented : 12<br />
Emission entries ------------> : EMS# ems_name opt SF_DEFAULT REG_PARAM ERROR<br />
Emission #1 : 1 IDADJ_EBCPI_an T 1 1 1<br />
Emission #2 : 2 IDADJ_EBCPO_an T 1 1 1<br />
Emission #3 : 3 IDADJ_EOCPI_an T 1 1 1<br />
Emission #4 : 4 IDADJ_EOCPO_an T 1 1 1<br />
Emission #5 : 5 IDADJ_EBCPI_bf T 1 1 1<br />
Emission #6 : 6 IDADJ_EBCPO_bf T 1 1 1<br />
Emission #7 : 7 IDADJ_EOCPI_bf T 1 1 1<br />
Emission #8 : 8 IDADJ_EOCPO_bf T 1 1 1<br />
Emission #9 : 9 IDADJ_EBCPI_bb T 1 1 1<br />
Emission #10 : 10 IDADJ_EBCPO_bb T 1 1 1<br />
Emission #11 : 11 IDADJ_EOCPI_bb T 1 1 1<br />
Emission #12 : 12 IDADJ_EOCPO_bb T 1 1 1<br />
<br />
--[[User:Daven|Daven]] 19:27, 13 January 2012 (EST)<br />
<br />
==== Augmented list of supported VOC sensitivities (adj32_014) ====<br />
From Kateryna Lapina, sensitivities of anthropogenic, biofuel and biomass burning emissions of all VOCs (ALK4, MEK, ALD2, PRPE, C3H8, CH2O C2H6). To implement, simply list these in <tt>input.gcadj</tt> as:<br />
<br />
Emission #34 : 34 IDADJ_EALK4_an T 1 1 1<br />
Emission #35 : 35 IDADJ_EALK4_bf T 1 1 1<br />
Emission #36 : 36 IDADJ_EALK4_bb T 1 1 1<br />
Emission #37 : 37 IDADJ_EACET_an T 1 1 1<br />
Emission #38 : 38 IDADJ_EACET_bf T 1 1 1<br />
Emission #39 : 39 IDADJ_EACET_bb T 1 1 1<br />
Emission #40 : 40 IDADJ_EMEK_an T 1 1 1<br />
Emission #41 : 41 IDADJ_EMEK_bf T 1 1 1<br />
Emission #42 : 42 IDADJ_EMEK_bb T 1 1 1<br />
Emission #43 : 43 IDADJ_EALD2_an T 1 1 1<br />
Emission #44 : 44 IDADJ_EALD2_bf T 1 1 1<br />
Emission #45 : 45 IDADJ_EALD2_bb T 1 1 1<br />
Emission #46 : 46 IDADJ_EPRPE_an T 1 1 1<br />
Emission #47 : 47 IDADJ_EPRPE_bf T 1 1 1<br />
Emission #48 : 48 IDADJ_EPRPE_bb T 1 1 1<br />
Emission #49 : 49 IDADJ_EC3H8_an T 1 1 1<br />
Emission #50 : 50 IDADJ_EC3H8_bf T 1 1 1<br />
Emission #51 : 51 IDADJ_EC3H8_bb T 1 1 1<br />
Emission #52 : 52 IDADJ_ECH2O_an T 1 1 1<br />
Emission #53 : 53 IDADJ_ECH2O_bf T 1 1 1<br />
Emission #54 : 54 IDADJ_ECH2O_bb T 1 1 1<br />
Emission #55 : 55 IDADJ_EC2H6_an T 1 1 1<br />
Emission #56 : 56 IDADJ_EC2H6_bf T 1 1 1<br />
Emission #57 : 57 IDADJ_EC2H6_bb T 1 1 1<br />
<br />
where in the example above emissions 1-33 were the previously defined standard set.<br />
<br />
--[[User:Daven|Daven]] 19:27, 13 January 2012 (EST)<br />
<br />
==== Remove bin_data.f and adj_bin_data.f ====<br />
<br />
From Zhe Jiang, remove the obsolete files bin_data.f and adj_bin_data.f, and remove their listings in Makefile.ifort.hdf<br />
<br />
--[[User:Daven|Daven]] 20:27, 16 January 2012 (EST)<br />
<br />
==== Several updates to support nested full chemistry adjoint (adj32_015) ====<br />
Several updates from Zhe Jiang support running the adjoint full-chemistry simulation over the NESTED_NA and NESTED_SD domain. The latter is a new domain, like NESTED_NA, but slightly less padding around NA. <br />
<br />
===== Support nested in emissions_adj.f =====<br />
<br />
<tt>emissions_adj.f</tt> has been updated to support nested emissions from the CAC, NEI05 and EMEP inventories.<br />
<br />
===== Cleanup code in adj_arrays_mod.f =====<br />
Remove obsolete variable NOR in <tt>adj_arrays_mod.f</tt><br />
<br />
<br />
--[[User:Daven|Daven]] 15:01, 18 January 2012 (EST)<br />
<br />
===== Do not call GET_BOUNDING_BOX in input_adj_mod.f =====<br />
When calculating IFD and JFD from a user specified latitude and longitude, the call to GET_BOUNDING_BOX causes a crash, so we need to modify this part of the code to work on the nested domain. For now, just re-order the code so that this is only called if USEINDEX is false, and run the nested code with USEINDEX true.<br />
<br />
==== MOPITT v5 observation operator (adj32_016) ====<br />
Zhe Jiang added support for the MOPITT v5 products. <br />
* Add MOPITT_V5_CO_OBS pre-processor flag to define_adj.h <br />
*Updates to <tt>adj_arrays_mod.f</tt>, <tt>mopitt_obs_mod.f </tt>, <tt>HdfVdModule.f90</tt>, <tt>input_adj_mod.f</tt>.<br />
<br />
--[[User:Daven|Daven]] 15:01, 18 January 2012 (EST)<br />
<br />
==== Updates for tagged CO simulation (adj32_017) ====<br />
Several updates from Zhe Jiang have been implemented for the tagged-CO implementation. See also changes to the forward model, [[GEOS-Chem_Adjoint_v32#Updates_for_tagged_CO_simulation__.28adj32_017.29]]<br />
<br />
===== Include CH4 gradient =====<br />
Update <tt>inverse_mod.f</tt> so that oxidation of CH4 is incorporated in the gradient <br />
<br />
===== Cleanup and small fixes to tagged_co_adj_mod.f =====<br />
Remove COSF and TIME_STEPS in <tt>tagged_co_adj_mod.f</tt>.<br />
<br />
===== Cleanup and merge CALC_APRIORI =====<br />
Updates for calculation of the penalty term:<br />
*Now valid for full chemistry and tagged CO simulations<br />
*Add subroutine READ_APERROR<br />
<br />
===== Enforce factor of 0.5 in cost function definition =====<br />
A consequence of the merging of the full-chem and CO routines in CALC_APRIORI is that the definition of the cost function needs to be made consistent. Originally, the CO code omitted the factors of 1/2 in both terms of the cost function. To correct for this, the cost function (and adjoint forcing) has been updated in the following CO observation operators to match the definition in the full-chem simulation:<br />
*airs_co_obs_mod.f<br />
*mopitt_obs_mod.f<br />
*sciabr_co_obs_mod<br />
<br />
--[[User:Daven|Daven]] 15:01, 18 January 2012 (EST)<br />
<br />
==== Include adjoint of CH4 simulation (adj32_023) ====<br />
Updates from Kevin Wecht. Modifications in the following files:<br />
* checkpt_mod.f<br />
* adj_arrays_mod.f<br />
* chemistry_adj_mod.f<br />
* cleanup_adj.f<br />
* define_adj.h<br />
* emissions_adj_mod.f<br />
* geos_chem_adj_mod.f<br />
* input_adj_mod.f<br />
<br />
* a3_read_mod.f<br />
* dao_mod.f<br />
<br />
* geos_chem_mod.f<br />
* global_ch4_mod.f<br />
* global_oh_mod.f<br />
<br />
* CMN_DIAG and diag3.f, diag_mod.f, initialize.f, ndxx_setup.f<br />
* julday_mod.f (fwd model bug fix)<br />
<br />
New files<br />
* netcdf_util_mod.f<br />
* global_ch4_adj_mod.f<br />
* tes_ch4_mod.f<br />
* scia_ch4_mod.f<br />
* leo_ch4_mod.f<br />
* mem_ch4_mod.f<br />
*geocape_ch4_mod.f<br />
<br />
===== Add support for ND19, ND59 and ND60 =====<br />
Modified CMN_DIAG and diag3.f, diag_mod.f, initialize.f, ndxx_setup.f<br />
<br />
===== Add support for OH_LOSS =====<br />
Modified diag_oh_mod.f <br />
<br />
<br />
--[[User:Daven|Daven]] 13:26, 12 February 2012 (EST)<br />
<br />
==== Add population weighted cost function option (adj32_024) ====<br />
Update from Steven Vogel adds a new option for cost function evaluation, LPOP_UGM3, and a new file <tt>population_mod.f</tt>. <br />
<br />
--[[User:Daven|Daven]] 01:10, 14 February 2012 (EST)<br />
<br />
==== Add strat production and loss rates as active control parameters (adj32_025) ====<br />
Hyungmin Lee has developed the adjoint with respect to production and loss control parameters, LADJ_STRAT, based on the new stratospheric fluxes from Lee Murray [[Stratospheric_chemistry]]. Also fix debug printing for JLOP. <br />
<br />
Updated files include:<br />
*logical_adj_mod.f<br />
*adj_arrays_mod.f<br />
*checkpoint_mod.f<br />
*checkpt_mod.f<br />
*chemdr_adj.f<br />
*geos_chem_adj_mod.f<br />
*input_adj_mod.f<br />
*inverse_driver.f<br />
*inverse_mod.f<br />
*linoz_adj_mod.f (completely replaced)<br />
*gamap_mod.f<br />
*partition_mod.f<br />
*tacer_mod.f<br />
*strat_chem_adj_mod.f (added)<br />
<br />
<br />
fwd:<br />
*logical_mod.f<br />
<br />
modified:<br />
*chemdr.f<br />
*geos_chem_mod.f (remove UPBDFLX_NOY and DO_UPBDFLX)<br />
*input_mod.f<br />
*linoz_mod.f (replace w/v9)<br />
*strat_chem_mod.f (add from v9)<br />
*Makefile.ifort.* (because new strat chem uses netcdf)<br />
<br />
--[[User:Daven|Daven]] 15:22, 14 February 2012 (EST)<br />
<br />
==== Calculate array sizes dynamically in CINSPECT (adj32_026) ====<br />
The arrays VAR_FD and RCONST_FD have been moved to adj_arrays_mod.f and are now dynamically allocated. The following error trap (which would frequently cause problems for new users) is no longer relevant:<br />
<br />
IF ( NCHEM > 1000 ) THEN <br />
CALL ERROR_STOP( 'Need to boost NCHEM max', 'chemistry_mod.f')<br />
ENDIF <br />
<br />
--[[User:Daven|Daven]] 16:22, 23 February 2012 (EST)<br />
<br />
=== Bug fixes in adjoint model ===<br />
<br />
----<br />
<br />
==== Mover error checking for O3 assimilation (adj32_002) ====<br />
'''''[mailto:N.Bousserez@dal.ca Nicolas Bousserez] wrote:'''''<br />
<br />
:the code crashes in "input_adj_mod.f" when doing the checking of observation settings (see line 2027 in my code). It is apparently due to "IDCSPEC_ADJ", which is not allocated/defined yet at the time the program executes those lines. I'm skipping these checks for now then. Note: The code crashes also when using "TES_O3_OBS" instead of "SCIA_DAL_NO2_OBS", so it is not related to what I implemented apparently.<br />
<br />
The following block of code has been moved from <tt>input_adj_mod.f</tt> to subroutine <tt>INIT_CSPEC_ADJ</tt> in <tt>adj_arrays_mod.f</tt>, as it can only be called after the <tt>CSPEC</tt> variables have been initialized. <br />
<br />
#elif defined ( TES_O3_OBS ) <br />
! Since the O3 obs operators will pass adjoints back <br />
! to CSPEC via CSPEC_AFTER_CHEM_ADJ, we need to make sure that <br />
! these species are listed as observed species<br />
FOUND = .FALSE.<br />
DO N = 1, NOBS_CSPEC<br />
<br />
IF ( TRIM( NAMEGAS( IDCSPEC_ADJ(N) ) ) == 'O3' ) THEN<br />
FOUND = .TRUE.<br />
ENDIF<br />
<br />
ENDDO<br />
IF ( .not. FOUND ) THEN<br />
<br />
CALL ERROR_STOP( ' Need to list O3 as observed species',<br />
& ' input_adj_mod ' )<br />
ENDIF<br />
#endif<br />
<br />
--[[User:Daven|Daven]] 16:38, 6 January 2012 (EST)<br />
<br />
==== Retire CSPEC_NO2_ADJ (adj32_003) ====<br />
<br />
'''''[mailto:N.Bousserez@dal.ca Nicolas Bousserez] wrote:'''''<br />
<br />
:In "chemdr_adj.F" I removed this part:<br />
#if defined( SCIA_KNMI_NO2_OBS ) || defined( SCIA_DAL_NO2_OBS )<br />
! Apply forcing from satellite observations<br />
CSPEC_ADJ(:,IDNO2) = CSPEC_ADJ(:,IDNO2) + CSPEC_NO2_ADJ(:)<br />
CSPEC_NO2_ADJ(:) = 0d0<br />
#endif<br />
:because it seems like I'm already updating the adjoint of NO2 in "sciadal_no2_obs_mod.f" through CSPEC_AFTER_CHEM_ADJ. I would like to be sure that it is correct though.<br />
<br />
'''''[mailto:daven.henze@colorado.edu Daven Henze] wrote:'''''<br />
<br />
:Yes, that is correct. I didn't want to take it out incase you had built your obs operator using CSPEC_NO2_ADJ. But since you used the more generic CSPEC_AFTER_CHEM_ADJ, then we can delete this section entirely and use the section higher up in this file ( IF (LCSPEC_OBS) ...) to apply the forcing from your obs operator to CSPEC_ADJ.<br />
<br />
Now use <tt>CSPEC_AFTER_CHEM_ADJ</tt> instead of <tt>CSPEC_NO2_ADJ</tt> in <tt>chemdr_adj.f</tt>.<br />
<br />
--[[User:Daven|Daven]] 18:04, 6 January 2012 (EST)<br />
<br />
==== Fix to NaNs in STT_ADJ everywhere after adjoint of partitioning (adj32_004) ====<br />
The forward gas-phase chemical solver may sometimes need to repeat the internal integration of the reaction rate ODEs within a single grid cell. When this happens, we now only redo the model calculation during the forward model run, not the adjoint, as this can lead to a corrupt CSPEC_ADJ array globally (not sure how that happens, but it does), eventually triggering a check for NaNs after the partitioning adjoint. See updates to <tt>chemistry_mod.f</tt> and <tt>gckpp_adj_Integrator.f90</tt>. Thanks to Richard Xu for helping pinpoint this bug. <br />
<br />
--[[User:Daven|Daven]] 18:03, 6 January 2012 (EST)<br />
<br />
==== Fix to NaNs in sulfate aerosol adjoint tracers (adj32_005) ====<br />
Owing to cancelation of errors, it's possible to seemingly randomly end up with NaNs coming out of <tt>CHEM_SO2_ADJ</tt> . This can happen if L3 ends up equaling 1d0 exactly. We now trap for this and skip in the rate event this happens. <br />
<br />
--[[User:Daven|Daven]] 18:03, 6 January 2012 (EST)<br />
<br />
==== Remove superfluous debugging which had <tt>LAVHRRLAI</tt> set to FALSE ====<br />
Noted by Fabien. Upated in <tt>input_mod.f</tt>.<br />
<br />
--[[User:Daven|Daven]] 18:07, 6 January 2012 (EST)<br />
<br />
==== Fix for reseting deposition adjoints (adj32_006) ====<br />
<br />
'''''[mailto:fabienpaulot@gmail.com Fabien Paulot] wrote:'''''<br />
<br />
:It seems it's "a bug" in the code. To keep track of the dry deposition flux of HNO3 I use dryhno3 (in cspec array). The CSPEC_ADJ(DRYHNO3,:) should be reset to zero at every time step but it is not. That's because you are saving in chk_cspec before it is reset to zero. Easy fix, just force to zero in chemdr_adj before the forcing is applied et voila, it works (the adjoint does not diverge any more after one hour). It should not affect anything as long as there is no forcing on dryhno3 (or any dry dep counters for that matter). <br />
<br />
We add the following to reset the dry dep adjoints in <tt>chemdr_adj.f</tt>:<br />
!$OMP PARALLEL DO<br />
!$OMP+DEFAULT( SHARED )<br />
!$OMP+PRIVATE( N, NK, JJ )<br />
DO N = 1,NUMDEP<br />
NK = NTDEP(N)<br />
IF (NK.NE.0) THEN<br />
JJ = IRM(NPRODLO+1,NK,NCS)<br />
IF (JJ.GT.0) THEN<br />
CSPEC_ADJ(:,JJ) = 0.0D0<br />
ENDIF<br />
ENDIF<br />
ENDDO<br />
!$OMP END PARALLEL DO<br />
<br />
--[[User:Daven|Daven]] 18:34, 6 January 2012 (EST)<br />
<br />
==== Update comments for STT_ADJ unit conversion ====<br />
'''''[mailto:mkeller@atmosp.physics.utoronto.ca Martin Keller] wrote:'''''<br />
<br />
:I'm currently guessing that CONVERT_UNITS(1,etc.) does something like:<br />
<br />
STT: kg -> v/v<br />
<br />
STT_ADJ: J/[v/v] -> J/kg<br />
<br />
:and similarly CONVERT_UNITS(2,etc.)<br />
<br />
STT: v/v -> kg<br />
<br />
STT_ADJ: J/kg -> J/[v/v]<br />
<br />
:From a dimensional analysis this would seem plausible. It would also explain some of the errors I'm currently fighting with. If this is correct, then a short comment should perhaps be added to the code.<br />
<br />
The comments in <tt>geos_chem_adj_mod.f</tt> have been updated surrounding calls to CONVERT_UNITS.<br />
<br />
--[[User:Daven|Daven]] 18:43, 6 January 2012 (EST)<br />
<br />
==== Only write upbd checkpoint files if necessary (adj32_007) ====<br />
Thanks to Farhan Akhtar for pointing this out. We now check to make sure writing these checkpoint files is necessary (i.e., it is an adjoint run, and not just a pseudo obs calculation) in <tt>linoz_mod.f</tt>:<br />
<br />
! Now see if checkpointing is necessary (fa, dkh, 01/06/12, adj32_007) <br />
IF ( LADJ .and. N_CALC > 0 ) THEN<br />
CALL MAKE_UPBDFLX_CHKFILE( NYMD, NHMS, TAU )<br />
ENDIF <br />
<br />
--[[User:Daven|Daven]] 18:54, 6 January 2012 (EST)<br />
<br />
==== Fix for reading TES diagnostic flags from input file (adj32_010) ====<br />
The problem was that if the CO diagnostics (LDCOSAT) were set to false, then the line counting would be off for any subsequent lines read in input.gcadj. The file input_adj_mod.f has been updated to work properly regardless of the value of LDCOSAT. <br />
<br />
<br />
==== Bug fix for unit conversion in aircraft and chemical source terms for CO2 adjoint (adj32_018) ====<br />
The units of the emissions scaling factor adjoints for the aircraft and chemical source terms were incorrect. These sources are in units per volume, not area, so their conversion needed to be<br />
<br />
E_CO2_ADJ = E_CO2_ADJ * BOXVL(I,J,L)<br />
& * DTSRCE / XNUMOL_CO2<br />
<br />
instead of<br />
<br />
!!E_CO2_ADJ = E_CO2_ADJ * A_CM2 * DTSRCE / XNUMOL_CO2<br />
<br />
Also, the ordering of the application of the CO2 emissions scaling factors was made consistent across all sectors such that the ND04 diagnostic is correct (previously the scaling had been applied after the diagnostic for ocean exchange, balanced biosphere, and CO2 surface correction for CO oxidation). <br />
<br />
--[[User:Daven|Daven]] 10:32, 9 February 2012 (EST)<br />
<br />
==== Bug fix for parallel loops in tagged_co_mod.f (adj32_017) ====<br />
<br />
F_OF_PBL and ED_CO were declared thread PRIVATE in several parallel do loops. <br />
<br />
--[[User:Daven|Daven]] 18:59, 8 January 2012 (EST)<br />
<br />
==== Bug fix for adjoints in nested cushions (adj32_020) ====<br />
In <tt>transport_mod.f</tt>, now call DO_WINDOW_TPCORE_BC_ADJ to set the adjoint values to zero within the cushion. <br />
<br />
--[[User:Daven|Daven]] 15:26, 9 February 2012 (EST)<br />
<br />
== Outstanding issues not yet resolved in v32 ==<br />
This is just a quick list of the features that we're working on putting into v32. As they are integrated and validated, we'll move these up to section 2 and expand the descriptions. <br />
*NH3 updates (Zhu, Colorado)<br />
*O3 IRK (Henze, Colorado)<br />
*NO2/SO2 obs operators (Bousserez, Dalhousie)<br />
*ANISO (Capps, Georgia Tech)<br />
<br />
Things to potentially add to the pipeline for the v33 release:<br />
*Reaction rates / deposition / updated chemistry (Paulot, Harvard)<br />
*ETOH simulations (Millet, UMN)<br />
<br />
And some bugs of note with no fix yet<br />
*Can't specify IFD, JFD using latitudes with nested domain<br />
*Can't use NEI emissions with tagged-CO</div>Hyung-Min Lee