90 likes | 217 Vues
This document outlines the essential aspects of the SMOKE (Sholy Modeling of Emission Known Emission) Case Beta Testing, including the configuration and customization of the ASSIGNS file. It explains how this file controls naming conventions for output files and directories based on specific parameters like SCTOR, CASE, and GRID. Caution is advised when modifying the ASSIGNS file due to the complexities of rerunning Jobs. The document also describes script organization, log analysis, and annual reporting, ensuring effective emissions modeling and analysis.
E N D
Background and Customizing SMOKE CaseBeta Testing February 11, 2010
ASSIGNS file • generic ASSIGNS file • works for all Jobs within a Case • works for all Cases • all specific parameters and inputs are specified in Job wrapper script • controls naming convention for files/directories • many of the names derived from SECTOR, CASE and GRID • intermed • reports • logs • model-ready outputs • recommend not modifying this file • if you want to change the naming conventions or the directory hierarchy, be VERY careful modifying the ASSIGNS file. The rerunning of Jobs and the similarities b/w Cases can make naming, overwriting files, etc. quite tricky
script directories under the scripts folder (/data4/smoke_emf/subsys26/smoke26/scripts): • cases: case specific Job directories • Uses Case abbreviation • Job wrapper scripts • Job logs • EMF command client logs • emf: EMF enabled scripts • semi-generic (area annual, point annual, point daily, etc) • controls running of SMOKE executables (smkinven, temporal, etc) • run: helper scripts • smk_run • helper scripts (set_days, qa_run, emf_cleanup) • annual_report: creates annual report by state/pollutant • log_analyzer: creates summaries of SMOKE logs
georegions abbreviation used for file and directory naming Sector: Emissions modeling sector(e.g., ptipm, ptnonipm) processing processing IOAPI grid name matches against GRIDDESC file
run_settings file • Used for controlling which parts of SMOKE are not run: • not a dataset (not managed by EMF), just a configuration file in the Job wrapper dir • can turn off parts of SMOKE by sector, grid, and date • esp. useful when rerunning large Jobs due to errors or running nested grids • By convention, the run_settings file is called: • /data4/smoke_emf/subsys26/smoke26/scripts/cases/$CASE/run_settings_$CASE.txt # Sector, Grid, Environment Variable Name, Part, Start Date, End Date, Value all , 12EUS1, smkinven , PART1, 0, 0, N all , 12EUS1, spcmat , PART1, 0, 0, N all , 12EUS1, temporal , PART2, 0, 0, N ptfire, 36US1, smkreport, PART1, 0, 0, N ptfire, 12EUS1, smkreport, PART1, 0, 0, N
sectorlist • Used by sectormerge • can merge sector specific model-ready emissions from multiple cases • controls naming • controls representative dates and holidays • controls which sectors are merged into final model-ready files (inline sectors) • recommend you create a new sectorlist per case • Used by annual_report
Post-processing: annual reports • calculates annual totals from smkmerge daily reports • uses representative dates (SECTORLIST)) • Python script runs automatically (registers output with EMF) • sector specific • converts moles -> tons • aggregate species (NOx, PM2.5)
Post-processing: log analyzer • Summarizes the warnings and errors in SMOKE logs: • set of known messages (identifies known and unknown warnings/errors) • assigns priority to each message (1 usually means Error, stop the script) • has multiple report levels (level 1 – summary, level 3 – most specific) • Python script runs automatically (registers output with EMF)