ekluzek
commented
6 months ago One of the reasons this work is important is because current mechanistic models don't show the important increase in dust emissions over the 20th Century. See the following talk by @dmleung at 2:01
Hi this is Danny, a PhD student at UCLA. Our group (@jfkok) recently developed a dust emission product using a data assimilation approach that better represents the decadal variability of dust emissions historically (1841-2000). The dust emission inventory will be commonly used by other ESMs (e.g., NorESM) for the upcoming AeroCom model intercomparison project to investigate the historical dust radiative forcings. I also developed the code needed to let CTSM read in the prescribed dust emissions using the streams capability. Erik Kluzek @ekluzek thought it would be nice to introduce an extra capability for CTSM to be able to read in dust emission datasets as an alternative to actively simulating dust emissions. We hope to introduce this capability, with the default being CTSM actively simulating dust emissions and the alternative being CTSM reading in the assimilated dust emissions. Users will be able to choose via a namelist option. Furthermore, the capability is not limited to our emission inventory; users can again compile any emission inventory they want and specify the dataset via the user namelist. The code I developed follows other StreamType.F90 in CTSM and will be able to read in dust emission datasets of any temporal and spatial resolutions.
We hope to introduce both the code and the dataset to be part of CTSM.
-- Brief scientific description for code changes:
(setting the namelist variables and default data file to read in the emission inventory) bld/namelist_files/namelist_definition_ctsm.xml bld/namelist_files/namelist_defaults_ctsm.xml bld/CLMBuildNamelist.pm
(setting the namelist variables for passing the data to the StreamMod and the DUSTMod) src/main/clm_driver.F90 src/main/clm_initializeMod.F90 src/main/controlMod.F90
(advancing the prescribed emission data in time and interpolating the emission data in time and space) src/cpl/share_esmf/DustEmisStreamMod.F90
(The DUSTMod will be able to choose between actively simulating the emissions or receiving the prescribed emissions from the StreamMod) src/biogeochem/DUSTMod.F90
The default dust emission inventory is a monthly dataset of 1.9°x2.5° global emissions for 1841-2000. The data is here: /gpfs/fs1/work/dleung/DustCOMM/AeroCom_DustCOMM_c230308/c230321/cdf5_AeroCom_DustCOMM_reconstr_emis_flux_1.9x2.5_1841-2000_gregorian_c230321.nc However, the code is made such that datasets of any spatial and temporal resolutions could be read in and processed by the streams capability of CTSM.
-- Long term goal: We hope to add the suggested feature to the master branch of CESM.
-- sandbox: /gpfs/fs1/work/dleung/ctsm_dustemis_prs_dev
Thank you, Danny