Things of note regarding these scripts: --SMOKE, I/O API-- The scripts assume the SMOKE executables are located in $SMK_HOME/subsys/smoke24/$SMOKE_EXE where SMK_HOME is whatever directory you unzip the script .zip files to, and SMOKE_EXE is set by the subsys/scripts/platform script described below. (example: SMOKE_EXE = Linux_ia64ifort) Also, required I/O API programs must be here: $SMK_HOME/subsys/ioapi/$SMOKE_EXE Finally, make sure you create the directories $SMK_HOME/intermed and $SMK_HOME/smoke_out before starting. (If you don't, the script will crash and tell you to create these directories.) --meteorology-- Make sure you change the definition of MET_ROOT at the top of the beis and on_moves scripts to point to your copy of the meteorological inputs (MCIP). --script order-- Each point sector has two scripts, a "Onetime" script and an "Annual" script. The "Onetime" script comes first. For the on_moves sectors, the base script goes first, followed by "temp_adj" and then "reports". Run the 36km jobs before running the 12km jobs. You could skip the 36km runs and start with the 12km jobs, but you would have to edit the run settings file (RUNSET) so that the scripts do not skip the non-grid-dependent SMOKE programs (e.g. Smkinven, Spcmat, Temporal). --domain totals-- Included in the script package is an optional QA program called "domain totals". After Mrggrid is run for each day, this program outputs a file containing domain-wide daily totals for each species from the merged model-ready emissions. By default, this program is turned off. To use the domain totals program, follow these instructions: - Compile the following Fortran program ${SMK_HOME}/subsys/smoke24/scripts/run/domain_totals/domaintotal_all_species.F A Makefile is provided in the same directory. - Move the resulting executable (domaintotal_new.x) to the parent directory: ${SMK_HOME}/subsys/smoke24/scripts/run/ - In the runsettings file (.../scripts/cases/${CASE}/run_settings_*.txt), comment out the lines that say "mrggrid, [grid name], domain, ..." --python-- These scripts require python. Here are all the prerequesites from James's MOVES script package: 1) Python__ 2.4.x or newer 2) Numerical python 24.2 This is an older version of the more robust numPy package. It is available at: http://sourceforge.net/projects/numpy/files/Old%20Numeric/ 3) Scientific python This package allows interfacing with netCDF files. It is available at: http://sourcesup.cru.fr/projects/scientific-py/ 4) UNIX-type system with the csh installed 5) netCDF C libraries in your LD_LIBRARY_PATH environment variable These libraries are required by scientific python and available at: http://www.unidata.ucar.edu/downloads/netcdf/netcdf-3_6_3/index.jsp --land use-- Make sure your copies of the biogenic land use files (BELD3_A, BELD3_B, and BELD3_TOT) are in the ${SMK_HOME}/inputs/2005ck_05b/beis/ directory and that the filenames match the run script. See [**Note B3**] in the beis run scripts. The beld3 inputs required to create the land use files are here: ftp://ftp.epa.gov/EmisInventory/beld_3/b3_files/ --gridding surrogates-- The gridding surrogate .zips under ftp://ftp.epa.gov/EmisInventory/2005v4 contain only the newest Canadian surrogates. For the USA and Mexican surrogates, go to ftp://ftp.epa.gov/EmisInventory/2002v3CAP/ancillary_smoke/ and copy the USA* and MEX* files (-NOT- the CAN* files) into the 36US1_surg_revised and 12km_surg_revised directories. --othar_hg-- In order to duplicate our results exactly, the othar_hg sector needs the following files from 2002v3CAP/ancillary_smoke: - The original 36km and 12km Canadian gridding surrogates, installed in ge_dat/36km_surg and ge_dat/12km_surg, respectively. Include the USA and MEX surrogates in this directory as well (included with the 2002v3CAP archive) to avoid a SMOKE crash. See [**Note 1**] in the run script. - amgref_us_can_allmex3_14nov2007_v4.txt, with the following two records manually removed from the file to avoid a script crash: 112000;2201020000;100 113000;2810030000;901 amgref_us_can_allmex3_26may2009_v11.txt from the 2005v4 package can also be used, but the results will be different than our original results. See [**Note 2**] in the run script to change this setting. - amptref_v3_3_11dec2007_v3.txt. To use the temporal cross-reference from the 2005v4 package (which will work but will give results different than our original results), remove the line with [**Note 3**] from the run script. --ptipm NAMEBREAK_DAILY-- For the automatic creation of PTDAY .lst files to work, the variable NAMEBREAK_DAILY may need to be adjusted. See [**Note 1P**] in the ptipm run script. --mercury, chlorine, sea salt-- For the final merge, be sure to install the contents of the .tar and .zip files in EmisInventory/2005v4/mercury_cl2_seasalt according to the following instructions: - hg_land*, hg_ocean*, and hg_volcanic: Install to ${SMK_HOME}/intermed/hgnatural_05b/. - ocean_cl2.zip: Install to ${SMK_HOME}/intermed/2005ck_05b/. - seasalt.zip: This is only required if you intend on converting to CAMx. If you do not intend on converting your emissions to CAMx-ready emissions, remove the line that says seasalt from each of the SECTORLIST files located in this directory. If you do need sea salt emissions, install the contents of this .zip to ${SMK_HOME}/intermed/2005ck_05b/. Then, for the scripts to work for 36km, you must create softlinks in this manner: > ln -s sslt.${YYYYMMDD}.36US1.GONG.cmaq.ncf emis_mole_seasalt_${YYYYMMDD}_36US1_cmaq_cb05_soa_2005ck_05b.ncf This step is not necessary for 12km. --future year dependencies-- The beis, avefire, and five oth* sectors are not rerun for any of the future year cases. Instead, the future year cases use 2005ck_05b emissions from these seven sectors. Therefore, to process a complete set of future year emissions, you must first process these seven sectors from the 2005ck_05b case. The 2012ck_*_05b tagging cases all use the 2012ck_05b base inventories, with one exception (see below). Tagging cases have a separate ancillary_smoke .zip file that contains extra files needed for the tagging runs (e.g. GSTAG files). Note that some sectors in the tagging cases are dependent upon sectors from other cases, including other tagging cases. The SECTORLIST included in each case/grid script package lists which sector cooresponds with which case. Examples: - 2012ck_pmtag_nh3_05b 12EUS1 uses emissions from 2012ck_pmtag_05b for four sectors: afdust, seca_c3, and the two on_moves sectors. There is no 2012ck_pmtag_05b 36US1 case, so 36US1 scripts for these four sectors are included in the 2012ck_pmtag_nh3_05b package. - 2012ck_pmtag_ptipm_05b uses 2012ck_pmtag_nh3_05b ptipm emissions, plus 2012ck_05b emissions for all other sectors. - 2012ck_pmtag_ptnonipm_05b has new ptnonipm emissions (including new inventories - this is the only tagging run with a separate set of inventories), and uses 2012ck_05b emissions for all other sectors. For 2014ck2_catr1_05b, 2014ck2_05b emissions must be created for all sectors except ptipm. --runtimes for tagging runs-- Run times for some of the tagging runs - in particular, the 2012ck_o3tag_05b 12EUS1 run - could be very long, and may require more memory than your system's default per-job allocation. Plan accordingly.