GEOS-5

Quick Fire Emission Dataset (QFED)

Revision as of 11:47, 16 August 2010 by Anton.darmenov (talk | contribs) (→Running QFED on discover)

This page to contains information regarding QFED, mostly for developers.

How to Check Out and Build the Code

These are checkout and build instructions for QFED. We presume that the user has an account and appropriate permissions on progress to check code out, and access to the NCCS machines (e.g., discover) to run code.

The name of the module in the cvs repository is QFED. Currently the name of the development tag is qfed-1_0_b4. To checkout the QFED source into the directory ./qfed/src, for example, use the following command 

% cvs -d $CVSROOT co -d ./qfed/src -r qfed-1_0_b4 QFED

Once the code is checked out set the ESMADIR variable to point to the QFED top directory, e.g.: 

% setenv ESMADIR ${HOME}/projects/qfed

Make sure that you have a valid ${BASEDIR}. QFED code is known to compile against Baselibs-v3.1.7.

To build the code navigate to the source directory where the main GNUMakefile resides: 

% cd ./qfed/src

The available options can be seen by running gmake or gmake esma_help: 

% gmake esma_help

  Standard ESMA targets:
  % make esma_install    (builds and install under ESMADIR)
  % make esma_clean      (removes deliverables: *.[aox], etc)
  % make esma_distclean  (leaves in the same state as cvs co)
  % make esma_doc        (generates PDF, installs under ESMADIR)
  % make esma_help       (this message)
  Environment:
     ESMADIR = .../qfed/src/
     BASEDIR = /opt/Baselibs/v3.1.7/
        ARCH = Linux
        SITE = hostname


To build and install QFED : 

% make esma_install

followed by 

% make PYTHON_INSTALL=python_install install

if you wish to install the python modules and scripts as well.

QFED Vegetation Map

The QFED vegetation classes are Tropical Forest, Extratropical Forest, Savanna and Grassland. The QFED vegetation map is derived from the global IGBP land cover data set by aggregating one or more IGBP classes into a single QFED class. QFED Tropical Forest corresponds to IGBP Evergreen Broadleaf for latitudes (30°S, 30°N), whereas QFED Extratropical Forest includes the rest of the IGBP forest types and IGBP Evergreen Broadleaf for latitudes outside the (30°S, 30°N) zone. The IGBP shrubland and savanna classes are aggregated into the QFED Savanna class. Similarly, the IGBP grassland and cropland classes are aggregated into the QFED Grassland class.

 
QFED vegetation map.
 
GFED vegetation map.


The QFED vegetation map uses the following numerical values for the vegetation classes: 

Tropical Forest = 1
Extratropical Forest = 2
Savanna = 3
Grassland = 4

To generate a QFED vegetation map and save it as a NetCDF file one can use the VegType.py script, which is installed in $(ESMADIR)/$(ARCH)/lib/Python/qfed/. The default options are equivalent to running the script with the following arguments: 

VegType.py --resolution='c' --output='QFED.vegetation_map.x288_y181.2010.nc'

The available options and their description can be seen by running the script with '-h' or '--help' option.

Level 3 QFED Products

There are two Level 3 (Global Gridded) QFED products:

  1. The daily fire radiative power (FRP) derived from MODIS Thermal Anomalies/Fire products (MOD14, MYD14). The name of the product is QFED Level 3a.
  2. The Level 3b product is derived from the Level 3a product. It contains the daily QFED fire emissions (BC, OC, CO, CO2, SO2).

The QFED Level 3a product can be generated after obtaining the necessary MOD14 and/or MYD14 granules using the qfed_l3a.py script. Similarly, Level 3b can be generated using the qfed_l3b.py script. Typical usage of the scripts will require one to provide the year and start and end day of the year. For full list of options run the scripts with the '--help' option, e.g. 

% qfed_l3a.py --help

Running QFED on discover

QFED (tag qfed-1_0_a5) has been successfully build and run on discover using Python-2.5.4, pyhdf-0.8.3, Baselibs-3.2.1 and Intel-11.0.083 compiler. First, lets set some environmental variables: 

setenv QFED ${HOME}/qfed/Linux/                                                   # QFED installation directory

setenv PATH ${QFED}/bin/:${SHARE}/dasilva/bin/:${PATH}                            # QFED scripts and n4zip utility 
setenv PYHDF /discover/nobackup/adarmeno/tmp/sandbox/pyhdf-0.8.3/                 # pyhdf - python interface to the HDF library
setenv PYTHONPATH ${QFED}/lib/Python/: \                                          # Python modules
                  ${PYHDF}/lib/python2.5/site-packages/:${PYTHONPATH}                       

setenv QFED_IGBP '/discover/nobackup/rgovinda/Emissions/Vegetation/GL_IGBP_INPE/' # IGBP-INPE dataset
setenv MODIS_FIRE '/discover/nobackup/dao_ops/intermediate/flk/modis/'            # MODIS Fire products

setenv QFED_L3a ${HOME}/tmp/QFED-test/                                            # QFED Level3a output directory
setenv QFED_L3b ${HOME}/tmp/QFED-test/                                            # QFED Level3b output directory

To generate the QFED FRP (QFED-L3a) files for July 19, 2010 (DOY = 200) using both MODIS Aqua and Terra data, we will execute the qfed_l3a.py script, e.g. 

qfed_l3a.py --mxd14=${MODIS_FIRE} --mxd03=${MODIS_FIRE} --igbp=${QFED_IGBP} --output=${QFED_L3a} 2010 200

Next we will generate the actual emissions files for the same day by running the qfed_l3b.py script 

qfed_l3b --input=${QFED_L3a} --output=${QFED_L3b} --ndays=1 2010 200
Retrieved from "https://geos5.org/wiki/index.php?title=Quick_Fire_Emission_Dataset_(QFED)&oldid=671"