Stand-alone BDSNP Soil NO module
User’s Guide
Version 1.0
Rui Zhang, Ben Lash, Quazi Rasool and Daniel Cohan
Department of Civil & Environmental Engineering
Rice University
Houston, TX USA
Contact: cohan@rice.edu
August 2015
1
Table of Contents
1. Introduction ......................................................................................................................................... 5
1.1 How to use this document ....................................................................................................... 5
1.2 BDSNP module overview ......................................................................................................... 5
1.3
2.
BDSNP software and operating system prerequisites ............................................ 7
Software installation .................................................................................................................... 7
2.1 Introduction .................................................................................................................................. 7
2.2 Program structure ...................................................................................................................... 8
2.3 Installing IOAPI and NetCDF libraries ................................................................................ 9
2.4 Build stand-alone BDSNP executable............................................................................... 10
3.
Test BDSNP benchmark case ................................................................................................. 10
3.1 Run benchmark case on 12km CONUS domain............................................................ 10
3.2 Key I/O environment variables in BDSNP module ..................................................... 11
3.2.1 SOILBIOME: soil biome type map ............................................................................. 11
3.2.2 SOILCLIMATE: arid/non-arid climate map ........................................................... 13
The input file $SOILCLIMATE provides the arid and non-arid region
classification over simulation domain, which will be used in the BDSNP module
to determine the soil moisture scaling factor 𝒈𝜽 for soil NO emissions. For the
modeling grid classified as ‘arid’ region, the maximum moisture scaling factor
is corresponding with the water-filled pore space 𝜽 value equal to 0.2; while for
the ‘non-arid’ modeling grid, the maximum moisture scaling factor
corresponds with 𝜃 = 0.3 (Hudman et al., 2012). ......................................................... 13
3.2.3 NFERT: fertilizer map .................................................................................................... 14
3.2.4 NDEP: archive nitrogen deposition file................................................................... 15
3.2.5 GRIDDEC: domain description ................................................................................... 16
3.2.6 METCRO2D: meteorology file ..................................................................................... 17
3.2.7 SOILINSTATE/SOILOUT: restart file ........................................................................ 18
3.2.8 BDSNPOUT: soil NO emission output file ............................................................... 18
4.
5.
Create specific case run using BDSNP module ................................................................ 19
4.1
Regridding static input files ........................................................................................... 20
4.2
Prepare meteorology file by MCIP ............................................................................... 21
BDSNP module further development ................................................................................. 22
5.1 Change emission factor for soil biome ............................................................................ 22
5.2 Replace fertilizer file with EPIC outputs ......................................................................... 22
2
Reference ................................................................................................................................................ 27
Appendix ................................................................................................................................................. 28
Appendix A. Sample setting script of ‘setcase.csh’ ............................................................. 28
Appendix B. Sample run script for BDSNP soil NO benchmark case ........................... 29
Appendix C. Mapping table used to create the soil biome map based on NLCD40
MODIS land cover categories ...................................................................................................... 31
Appendix D. Spatial map of five climate zone over CMAQ CONUS 12km domain
based on Köppen-Geiger climate classification ................................................................... 32
Appendix E. Sample of soil biome emission factor setting include file ...................... 33
3
List of Tables
Table 1 List of key I/O environment variables in BDSNP model .................................... 24
Table 2 List of 24 soil biome emission factor from Steinkamp and Lawrence (2010) .... 25
Table 3 List of output variables from SOILINSTATE (restart file) ................................. 26
List of Figures
Figure 1 Flow chart of inline/offline BDSNP module ........................................................ 6
Figure 2 Soil biome map over continental US (4km resolution) ...................................... 13
Figure 3 Arid (blue) and non-arid (red) region over Continental US (12km resolution) . 14
Figure 4 Daily variation of total N from fertilizer over continual US from Potter et al.
(2011) ................................................................................................................................ 15
Figure 5 Soil N reservoir from deposition from benchmark case over continental US
(resolution 12km) .............................................................................................................. 16
Figure 6 Spatial distribution of daily mean soil NO emission rate from benchmark case
........................................................................................................................................... 19
Figure 7 Regridding biome map from 4km CONUS files to 12km Texas domain .......... 21
Figure 8 Comparison of spatial pattern of fertilizer N input from Potter et al. (2010) and
EPIC simulation ................................................................................................................ 23
4
1. Introduction
1.1 How to use this document
This guide instructs potential users to build and run the stand-alone BDSNP module
to simulate NO emissions from soil for climate and air quality applications. This
document is written in a relatively ad hoc order so that the reader can choose
individual sections for reference based on their own purpose (e.g. background
knowledge, software installation, real case implantation, future model update).
However, the first time user is recommended to read through Section 1 first.
Comments on how to improve the documentation are encouraged.
1.2 BDSNP module overview
The Berkeley-Dalhousie Soil NOx Parameterization (BDSNP) previously has been
implemented in a global chemical transport model (GEOS-Chem) to update the soil
nitrogen oxide emission estimates (Hudman et al., 2010, 2012). Compared to the
soil NO emission module (Yienger and Levy 1995; YL95) currently coded in biogenic
emission models such as BEIS3 and MEGAN, the BDSNP scheme has a more
continuous soil temperature/moisture dependence function and simulates higher
magnitude pulses after dry soil is wetted by precipitation.
Based on the BDSNP scheme, soil NO emissions at each modeling grid and timestep
are determined by the base emission varied by the biome type and the available soil
nitrogen (either form the fertilizer implementation during the growing season or
the nitrogen dry/wet deposition from the atmosphere), the soil temperature and
moisture response function, the soil pulsing after precipitation, and the canopy
reduction due to the resistance. The algorithm for canopy reduction was proposed
by Wang et al. (1998) in GEOS-Chem to consider a more physical basis based on
wind speed, turbulence, canopy structure, deposition constants and other physical
variables.
𝑁𝑂𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝑓𝑙𝑢𝑥 = 𝐴(𝑏𝑖𝑜𝑚𝑒, 𝑠𝑜𝑖𝑙 𝑛𝑖𝑡𝑟𝑜𝑔𝑒𝑛) × 𝑓(𝑇) × 𝑔(𝜃) × 𝑝𝑙𝑢𝑠𝑒(𝑑𝑟𝑦 𝑝𝑒𝑟𝑖𝑜𝑑) ×
𝐶𝑎𝑛𝑜𝑝𝑦 𝑅𝑒𝑑𝑢𝑐𝑡𝑖𝑜𝑛 (𝐿𝐴𝐼, 𝑏𝑖𝑜𝑚𝑒, 𝑚𝑒𝑡)
Benjamin Lash at Rice University (Lash, 2015) successfully implemented the BDSNP
code from GEOS-Chem into the Community Multiscale Air Quality (CMAQ) modeling
system and produced the online soil NO emission estimates using the soil
temperature/moisture information from the Pleim-Xiu land surface model (Pleim
and Xiu, 2003). However, the inline BDSNP module at CMAQ is computationally
expensive and is not cross-platform applicable. For example, BDSNP has not been
implemented in the CAMx air quality model. Therefore, an offline (stand-alone)
version of BDSNP module (stand-alone BDSNP) is developed by Dr. Rui Zhang.
From the software engineering point view, Figure 1 provides the flow chart of the
BDSNP scheme implementation with the option to run inline or offline with the air
5
quality model. Static input files such as arid/non-arid climate zone, soil biome type
(must be consistent with the type of the soil emission factors given by Steinkamp
and Lawrence (2011)) and global fertilizer pool from Potter et al. (2010) are needed
to determine the soil base emission value at each modeling grid. Soil moisture and
temperature as well as some meteorological variables such as radiation, wind and
air pressure are needed (from weather models) to drive the BDSNP to representing
the occurrence of pulsing and the process of canopy reduction. The major difference
with the option of inline and offline BDSNP model is the approach to deal with the
available nitrogen pool from the soil. For the inline BDSNP model (the dash line), the
dry or wet deposition process considered in air quality model will continuously
update the available nitrogen from the atmosphere to the ground; for the standalone BDSNP model (the solid line), only the generic daily nitrogen pool from
deposition process are used. Since soil nitrogen pool needs a long time to build up to
reach the quasi-steady state in the model, a new series of SOILINSTATE files are
needed in the stand-alone BDSNP model to provide the generic daily variation of N
deposition. Here, we performed a full year CMAQ simulation over the 12km
continental U.S. modeling domain in 2005 and recorded their daily total soil N
reservoir from deposition for further usage. Simulating soil NO emissions for a
region outside North America or for a period with sharply different N deposition
rates from 2005 will require development of a new set of SOILINSTATE files.
However, it should be noted that in our in-line CMAQ simulations of soil NO, the soil
N reservoir from deposition is responsible for only about 10 percent of emissions.
Figure 1: Flow chart of inline/offline BDSNP module
6
The inline BDSNP module with CMAQ initially used the soil biome map directly regridded from global atmospheric chemistry model GEOS-Chem, which is too coarse
for regional model implementation. The biome map is related with the land use/
land cover (LU/LC) classification and Köppen-Geiger climate zone definition (Kottek
et al., 2006) and will determine the base soil NO emission strength. A new soil biome
spatial map based on 4-km CONUS 40-category 2006 NLCD-MODIS land use
classification (NLCD40) and climate zone definition is created to replace the GEOSChem biome map to better represent the up-to-date LU/LC change with finer
details. The required mapping the 40 categories to the categories for which
emission rates were reported by Steinkamp and Lawrence (2011).
1.3 BDSNP software and operating system prerequisites
It should be noted that running this version of stand-alone BDSNP model requires
the working knowledge of Unix operating system and FORTRAN programming
language as prerequisites (e.g, how to unzip/untar files, install libraries, link files,
and implement basic commands within Unix, and understand basic Shell script to
invoke FORTRAN executables). If you need to regrid the BDSNP input files (e.g. 4-km
soil biome map and 12-km deposition archive) to a preferred simulation domain,
then you further may need the knowledge/access to Python computer language and
ESRI ArcMAP software or know how to use open-source software such as the spatial
allocator developed by UNC (www.ie.unc.edu/cempd/projects/mims/spatial/).
The system and software requirements for installing and running stand-alone
BDSNP v.1.0 are as follows:
 Linux/Unix operation system
 Shell scripting language
 FORTRAN 90 compiler (e.g. PGI)
 NetCDF v3.6.0 or greater
 IOAPI v3.1
 MCIP v3.6 or greater
The program and provided test case has been successfully tested on a GNU/Linux
x86_64 machine with PGI compiler.
2. Software installation
2.1 Introduction
The stand-alone BDSNP module software installation is fairly straightforward on the
supported Linux platforms. After unzipping the programing the downloading
required libraries and packages, you can refer to the simple steps given below to
install and compile the program. Installing NetCDF and IOAPI libraries is needed
first to compile to BDSNP executables.
7
2.2 Program structure
The first step is to download the provided program compressed file
“BDSNP_SA_V1.0.tar.gz” and untar/unzip the file on designated working directory:
>$tar –xzf BDSNP_SA_V1.0.tar.gz
The folder structure of the BDSNP module is shown as the below tree list:
$/work/dsc1/rz10/BDSNP_SA_V1.0
.
|-bin/
|-biome_ncld2006_4km/
|-inputs/
|---DEP
|---FERT
|---MAP
|---MET
|-lib/
|---IOAPI
|---NETCDF
|-outputs/
|-README
|-setcase.csh
|-src/
|-testdata/
|-work/
File ‘README’ documents the summary of program and basic instruction on how to run
this model. Folder ‘bin’ contained the executables compiled from the source code. Folder
‘biome_ncld2006_4km’ contains the 4km CONUS biome map. Folder ‘inputs’ is the
place to put all the required input files for BDSNP module, which include four subfolders, namely ‘DEP’, ‘FERT’, ‘MAP’, and ‘MET’. More details about the inputs
preparation are given in Section 4. Folder ‘lib’ contains the needed libraries, i.e. NetCDF
and IOAPI. Folder ‘outputs’ is the place for model outputs. Cshell script ‘setcase.csh’ is
the overall control script to setting the folder environment variables. Folder ‘src’ contains
the raw FORTRAN code. Folder ‘testdata’ contain the benchmark inputs and outputs for
the user test. Folder ‘work’ contains the run scripts.
Before installing libraries and compiling the program executables, you should first
change the script ‘setcase.csh’ by setting your current working directory as the
$BDSNPHOME (the sample script of the ‘setcase.csh’ is attached in Appendix A):
$setenv BDSNPHOME [your-current-working-path]
8
After changing the script, you need to source it to invoke the folder environment
variables:
>$source ./setcase.csh
2.3 Installing IOAPI and NetCDF libraries
The NetCDF v3.6.0 or greater can be downloaded at:
http://www.unidata.ucar.edu/downloads/netcdf/index.jsp
The IOAPI v3.1 can be downloaded at:
http://www.cmascenter.org/
Registration may be needed before downloading these packages and corresponding
user’s guides. Users are highly recommended to refer to their user’s guide for
installing the packages.
For NetCDF, there are two ways to install it on Linux/Unix. One is to install the prebuild binary package with the compatible version for your operation system (e.g.
Linux2_x86_64pg_gcc_nomp). The other way is to install it from the tar.gz file. Once
download the compressed file, untar it and go to the netcdf directory and type the
following commands:
>$./configure --prefix=[path-to-netcdf-install]
>$make check install
More detail instruction on the installation of NetCDF can be found at:
http://www.unidata.ucar.edu/software/netcdf/docs/netcdf--‐install/
Once the NetCDF is installed successfully, you need to set the environment variables
to the location where you install it:
>$setenv NETCDF [path-to-netcdf-install]
>$setenv NETCDF_LIB ${NETCDF}/lib
>$PATH ${NETCDF}/bin:${PATH}
>$LD_LIBRARY_PATH ${NETCDF_LIB}:${LD_LIBRARY_PATH}
Then, you need to copy the compiled library ‘libnetcdf.a’ and ‘libnetcdff.a’ to the
folder $BDSNPLIB/NETCDF.
The similar steps are needed to install ioapi on your working system if not have. Be
caution to same complier and compile flags should be used for compiling NetCDF
and IOAPI. The detail instruction on the installation of IOAPI can be found at:
http://www.baronams.com/products/ioapi/AVAIL.html
Again, the corresponding environment variables should be set after successful
install the IOAPI:
>$setenv IOAPI
[path-to-ioapi-install]
9
>$setenv IOAPI_LIB ${IOAPI}/lib
>$PATH ${IOAPI}/bin:${PATH}
>$LD_LIBRARY_PATH ${IOAPI_LIB}:${LD_LIBRARY_PATH}
Also copy the compiled library ‘libioapi.a’ to the folder $ BDSNPLIB/IOAPI.
2.4 Build stand-alone BDSNP executable
Go the directory $BDSNPSRC and change the ‘Makefile’ file by setting the new lib
and include paths as the installed NetCDF and IOAPI directories:
>$ source $BDSNPHOME/setcase.csh
>$ cd $BDSNPSRC
>$ vi Makefile
Change the lines in ‘Makefile’ as following as save the change:
LIB
= -L${BDSNPLIB}/IOAPI –lioapi \
-L${BDSNPLIB}/NETCDF –lnetcdff –lnetcdf
INCLUDE = -I${NETCDF}/include –I${IOAPI}/ioapi/fixed_src
PROGRAM=${BDSNPEXE}/NOX_BDSNP.exe
>$make >& log.bulid_BDSNP &
If success, you can find the executable ‘NOX_BDSNP.exe’ ready to use at the path
$BDSNPEXE.
3. Test BDSNP benchmark case
A benchmark case with the 12km continental U.S. (CONUS) domain setting is
provided for the comparison of outputs by first time users. The inputs and outputs
data are provided at $BDSNPHOME/testdata. The user can utilize those data to test
the functionality and understand the key input/output (I/O) environment variables
setting of the BDSNP module.
3.1 Run benchmark case on 12km CONUS domain
Go the $BDSNPWORK and you can find a sample run script to run the 12km CONUS
benchmark case on the simulation date June 27, 2011 (Julian date 2011178). The
contents of the demo script are attached as Appendix B. For the benchmark run, no
changes are needed for the run script but make sure the setting paths of the input
files in the sample script are at the correct location. Type the following commands to
prepare the input files for the benchmark run:
>$source $BDSNPHOME/setcase.csh
>$cd $BDSNPWORK
>$cp $BDSNPHOME/testcase/BIOME_benchmark.nc $BDSNPINP/MAP
10
>$cp $BDSNPHOME/testcase/SOILCLIAMTE_benchmark.nc $BDSNPINP/MAP
>$cp $BDSNPHOME/testcase/NDEP_2005_CMAQ_benchmark.nc $BDSNPINP/DEP
>$cp $BDSNPHOME/testcase/SOILFERT_2005_CMAQ_benchmark.nc
$BDSNPINP/FERT
>$cp $BDSNPHOME/testcase/METCRO2D_2011178_benchmark.nc
$BDSNPINP/MET
>$cp $BDSNPHOME/testcase/SOILINSTATE_2011177_benchmark.nc
$BDSNPOUT/
After putting the benchmark input files in the correct location, you can invoke the
run script by typing:
>$./run.BDSNP_soil_NO_benchmark.csh &
If the program run successfully, you can find two output files under the directory
$BDSNPOUT, with the name ‘B3GTS_NO_BDSNP_2011178_benchmark.nc’ and
‘SOILINSTATE_2011178_benchmark.nc’ respectively. Those two output files can be
compared with the reference outputs located at $BDSNPHOME/testdata directory
(‘B3GTS_NO_BDSNP_benchmark.nc’ and ‘SOILINSTATE_benchmark.nc’) using
visualization tools such as VERDI (http://www.cmascenter.org/verdi)
or IDL (www.exelisvis.com/ProductsServices/IDL.aspx).
3.2 Key I/O environment variables in BDSNP module
Understanding the input/output (I/O) environment variables defined in BDSNP
module is crucial for the real case implementation and further model development if
needed. There are seven key input environment variables and two key output
environment variables in current version of soil NO BDSNP module. Table 1
summarizes their names and corresponding functionalities. Each I/O environment
variable is generally elaborated in sequence for following subsections.
3.2.1 SOILBIOME: soil biome type map
Based on several field studies for soil NO flux (SNOX) measurement (Stehfest and
Bouwman, 2006), different biome type has different NO emission factors under
different soil conditions (wet or dry) over different sampling geographic locations
(or climate zones). Steinkamp and Lawrence (2011) complied those worldwide
emission factors (a dataset consisting of 112 articles with 583 field measurements
of SNOx covering the period from 1976 to 2010) and regrouped them into 24 soil
biome type base on MODIS land cover category as well as Köppen climate zone
classifications (Kottek et al., 2006) and reported them into the mean values
(arithmetic or geometric). Those soil NO emission factors (wet) implemented in the
current version of stand-alone BDSNP module are listed in Table 2.
11
The folder ‘$BDSNPHOME/biome_ncld2006_4km’ provides the 4-km resolution soil
biome map product over the continental U.S. This fine soil biome map is generated
from the 30-arc-second (approximately 1 kilometer) NCLD2006 dataset with 40
category land cover/land use classifications. A mapping algorithm table (see
Appendix C) was created to connect the land use category to soil biome type (Table
2) based on best available knowledge. For the categories with identical names, such
as ‘evergreen needleleaf forest’, ‘deciduous needleleaf forest’, ‘mixed forest’,
‘savannas’ and ‘grassland’, the mapping is direct. Categories in NLCD40 which are
subsets of the corresponding biome category are consolidated into one category by
addition. For example, ‘permanent snow and ice’ and ‘perennial ice-snow’ in
NLCD40 are combined to form ‘snow and ice’; ‘developed open space’, ‘developed
low intensity’, ‘developed medium intensity’, and ’developed high intensity’ are
added to form ‘urban and built-up lands’. For the categories appearing only in
NLCD40, the mapping algorithm is determined by referring to the CMAQ mapping
scheme, which is documented in each of the CSQY_DATA_* files under the MECHS/
directory at the CMAQ source code release. One of the examples is to map ‘lichens’
and ‘moss’ in NLCD40 to the category ‘grassland’ in soil biome. Furthermore, a
model resolution compatible Köppen climate zone classification (Kottek et al., 2006)
is needed to allocate different emission factor for the same biome type (e.g.
‘grassland’) at different locations. There are five climate zone classifications, namely
A: equatorial, B: arid, C: warm temperature, D: snow, E: polar. A 12-km CONUS
model resolution climate zone classification map (see Appendix D) is also created
using the Spatial Allocator based on the county level text file climate zone
documentation as the surrogate
(http://koeppen-geiger.vu-wien.ac.at/data/KoeppenGeiger.UScounty.txt).
Figure 2 provides the spatial pattern of the 24 soil biome map in 4km resolution
over CONUS domain. The corresponding 1-24 category type names are provided in
Table 2. Only regridding is needed for specific modeling domain implementation
due to the fine texture details (see section 4.1 for detailed procedures).
12
Figure 2 Soil biome map over continental US (4km resolution)
3.2.2 SOILCLIMATE: arid/non-arid climate map
The input file $SOILCLIMATE provides the arid and non-arid region classification over
the simulation domain, which will be used in the BDSNP module to determine the soil
moisture scaling factor 𝒈(𝜽) for soil NO emissions. For the modeling grid classified as
‘arid’ region, the maximum moisture scaling factor corresponds with the water-filled
pore space (𝜽) value equal to 0.2; while for the ‘non-arid’ modeling grid, the maximum
moisture scaling factor corresponds with 𝜽 = 𝟎. 𝟑 (Hudman et al., 2012).
The 12km resolution CONUS arid/non-arid climate map was also provided here (file
name ‘SOILCLIMATE_benchmark.nc’ which is located in the folder ‘$BDSNPHOME/
testdata/’). The users can regrid this static field into the designed modeling domain.
Figure 3 show the distribution of arid/non-arid climate map over continental US
with the blue color corresponding with the arid region while red color
corresponding with the non-arid region.
13
Figure 3 Arid (blue) and non-arid (red) region over Continental US (12km
resolution)
3.2.3 NFERT: fertilizer map
In the BDSNP stand-alone model, the environment variable $NFERT provides the
path of the file with the daily nitrogen input from fertilizer implementation in
addition to the base emission from soil biome. The benchmark case fertilizer
information given here (file name ‘SOILFERT_2006_benchmark.nc’ located in the
folder ‘$BDSNPHOME/testdata/’) is the regridded product from the global GEOSChem BDSNP implementation (Hudman et al. 2012) to 12km resolution CONUS
domain. Hudman et al. (2012) used a global fertilizer database from Potter et al.
(2010) and assumed 37% of fertilizer and manure N is available (1.8 Tg N yr-1) for
potential emission. Figure 4 provides the day-by-day variation of total fertilizer
implementation over CONUS during a year and shows the typical cycle between
growing season and non-growing season.
14
Figure 4 Daily variation of total N from fertilizer over continental US from Potter et
al. (2011)
Without updating, any parameterization of soil NO based on fertilizer will be out of
date within a decade. The integration of the BDSNP and the Environmental Policy
Integrated Climate (EPIC) fertilizer model (Cooter et al. 2012) is one possible stateof-the-science approach for enabling dynamic representation of year-specific
fertilizer fields (see section 5.2).
3.2.4 NDEP: archive nitrogen deposition file
For the stand-alone module, the nitrogen available from wet or dry deposition
cannot be calculated instantaneously through the air quality model but needs to be
determined from offline, pre-computed files. Ben Lash (2015) implemented the
BDSNP module in CMAQ and conducted the whole year 2005 simulation over 12km
CONUS domain using the emission and meteorology files provided by EPA. The daily
available soil N reservoir from deposition was recorded in the restart file
$SOILINSTATE (see section 3.2.7). The path of environment variable $NEDP should
be set as the path of previous day’s $SOILINSTATE archived deposition file. The
whole year 2005 restart files have been archived as the generic deposition fields.
For real case usage, the users need to regrid the deposition fields into the designed
simulation domains and assume the total amount of soil N from deposition at the
same geographic locations over the selected period is comparable with the CMAQ
simulation results in year 2005 during corresponding dates, or scale as desired.
15
Figure 5 provides an example of the spatial distribution of available nitrogen from
deposition
in
the
benchmark
case
on
a
single
day
(file
‘NDEP_2005_CMAQ_benchmark.nc’ located at the folder ‘$BDSNPHOME/testdata/’).
Figure 5 Soil N reservoir from deposition from benchmark case over
continental US (resolution 12km)
3.2.5 GRIDDEC: domain description
The environment variable $GRIDDESC gives the paths of text input files for the grid
description on the configuration of designed simulation domain. The format of
$GRIDESC is the same as CMAQ setting in the IO/API library.
For the benchmark case application (file ‘GRIDDESC_EPA_CONUS_12km’ located at
folder $BDSNPRUN), the domain configuration is set as following:
''
'12CONUS'
2
33.000
45.000
-97.000
-97.000
''
'12US2'
16
40.000
'12CONUS' -2412000.000 -1620000.000
12000.000
12000.000 396 246 1
''
where in line 2, ‘12CONUS’ is the logic name in I/O API library for the domain
projection configuration. In line 3, ‘2’ is the $GDTYP3D in I/O API library to
determine the projection type, which in here is Lambert Conformal Conic. ’33.000’
and ’45.000’ are corresponding with the $P_ALP3D and $P_BET3D in I/O API library
to determine the 1st and 2nd true latitude. The first value ‘-97.000’ in line 3 is the
$P_GAM3D parameter in I/O API library to determine the central meridian
longitude (with the West in negative value). The second value ‘-97.000’ in line 3 is
the $XCENT3D in I/O API library, which represents the center longitude.
Correspondingly, ’40.000’ is the $YCENT3D in I/O API library for the center latitude.
The value ‘12US’ in line 5 is the logic name $GDNAM3D for the simulation domain
name in I/O API library. In line 6, the values ‘-2412000.000’ and ‘-1620000.000’ are
corresponding to the $XORIG3D and $YORIG3D in I/O API library to determine the
distance in meters (m) from the center point of domain the lower left grid point at x
(column) and y (row) directions. The values ’12000.000’ and ‘12000.000’ are
corresponding to the $XCELL3D and $YCELL3D in I/O API library to set the domain
resolution at each modeling gird in meters. The values ‘396’, 246’, ‘1’ are
corresponding to the $NCOLS3D, $NROWS3D, $NTHIK3D in I/O API library to set
the 3D modeling grid number in the x (column), y (row), and z (layer) dimension
respectively.
3.2.6 METCRO2D: meteorology file
Environment variable $METCRO2D provides the path of meteorological variables
needed for the stand-alone BDSNP module. The meteorology file is in NetCDF
format and is normally run by the meteorological model (e.g., WRF) and MCIP (see
section 4.2). A total of nine meteorological variables are needed in the BDSNP
module:
(1) Surface temperature (in K) at 2-m, used to calculate canopy reduction
factor: 𝐶𝑎𝑛𝑜𝑝𝑦 𝑅𝑒𝑑𝑢𝑐𝑡𝑖𝑜𝑛 (𝐿𝐴𝐼, 𝑏𝑖𝑜𝑚𝑒, 𝑚𝑒𝑡)
(2) Ground reaching solar radiation (in W/m2), used to calculate canopy
reduction factor
(3) Surface pressure (in Pa), used to calculate canopy reduction factor
(4) Soil moisture (in m3/m3), used to calculate the dry period and pulsing
strength 𝑝𝑢𝑙𝑠𝑒(𝑑𝑟𝑦 𝑝𝑒𝑟𝑖𝑜𝑑) and scale the base emission by soil moisture
response function: 𝑔(𝜃)
(5) Soil temperature (in K), used to scale the base emission by soil temperature
response function: 𝑓(𝑇)
(6) Surface snow cover, used to calculate canopy reduction factor
(7) Wind speed (in m/s) at 10m, used to calculate canopy reduction factor
(8) Cloud fraction in modeling grid, used to calculate canopy reduction factor
17
(9) Leaf area index, used to calculate canopy reduction factor
It should be noticed that, in this version of the model, the Pleim-Xiu land surface
model (2003) should be selected when running the meteorological model (e.g.,
WRF) and MCIP to provide the diagnosed soil temperature and moisture fields.
3.2.7 SOILINSTATE/SOILOUT: restart file
Environment variables $SOILINSTATE and $SOILOUT are identical and both point to
the path of the restart file generated by the stand-alone BDSNP module (see
Appendix B for the demo run script setting). It contains the previous day’s soil
moisture, pulsing factor, and dry period information for continuous run as well as
several diagnostic variables for further model verification and analysis. Table 3 lists
all four restart variables and nine diagnostic variables and their corresponding
attributes. The output variables provides the breakdown of each process in standalone BDSNP to impact the final soil NO emission estimates, namely the biome base
emission (A_DIAG), fertilizer implement amount (AFERT_DIAFG), soil nitrogen pool
from deposition (NDEPRES), pulse factor (PFACTOR) and the canopy reduction
factor (CRFAVG). For each modeling grid, higher biome base emission plus higher
fertilizer implementation and higher available N from deposition will result in
higher potential to produce soil NO. The pulse factor depends on soil moisture and
the antecedent dry period. The scaling factor CRFAVG represents the loss of NO to
the plant canopy before entering the broader atmosphere.
One of the analyses that can be done with the diagnostic output variables is to
calculate the fraction of NO emissions resulting from soil biome base emissions
(BIOME), fertilizer application (FERT), and N deposition (NDEP). These fractions
sum to 1.
The following formulas can be used to calculate those fractions:
𝐹𝐵𝐼𝑂𝑀𝐸 = 𝐴_𝐷𝐼𝐴𝐺/(𝐴_𝐷𝐼𝐴𝐺 + 𝐴𝐹𝐸𝑅𝑇_𝐷𝐼𝐴𝐺)
𝐹𝐹𝐸𝑅𝑇 =
𝑁𝑅𝐸𝑆_𝐹𝐸𝑅𝑇_𝐷𝐼𝐴𝐺
× 𝐴𝐹𝐸𝑅𝑇_𝐷𝐼𝐴𝐺/(𝐴_𝐷𝐼𝐴𝐺 + 𝐴𝐹𝐸𝑅𝑇_𝐷𝐼𝐴𝐺)
(𝑁𝐷𝐸𝑃𝐸𝑆 + 𝑁𝑅𝐸𝑆_𝐹𝐸𝑅𝑇_𝐷𝐼𝐴𝐺)
𝐹𝑁𝐷𝐸𝑃 =
𝑁𝐷𝐸𝑃𝐸𝑆
× 𝐴𝐹𝐸𝑅𝑇_𝐷𝐼𝐴𝐺/(𝐴_𝐷𝐼𝐴𝐺 + 𝐴𝐹𝐸𝑅𝑇_𝐷𝐼𝐴𝐺)
(𝑁𝐷𝐸𝑃𝐸𝑆 + 𝑁𝑅𝐸𝑆_𝐹𝐸𝑅𝑇_𝐷𝐼𝐴𝐺)
3.2.8 BDSNPOUT: soil NO emission output file
The environment variable $BDSNPOUT is used to set the output path and file name
for the hourly soil NO estimation from BDSNP module. The user can choose the
output unit for the soil NO emission rate either in terms of mass (Ng/s) or moles
(mol/s). Figure 6 shows the spatial distribution of daily mean soil NO emission rate
18
from the benchmark case. For the simulation day, the emission hot spots correlated
well with locations that have either biome types with high base emission factors(e.g.
evergreen broadleaf forest around Mississippi), high fertilizer application rates (e.g.
cropland region near Iowa and Kansas), or high N deposition (e.g. central
California).
Figure 6 Spatial distribution of daily mean soil NO emission rate from benchmark
case
4. Create specific case run using BDSNP module
To run a specific case using the BDSNP module, the user needs to prepare the entire
model domain consistent input files listed in Table 1, which include soil biome map,
arid/non-arid climate region map, archived daily N deposition fields, domain
configuration description file, meteorological outputs in NetCDF format from MCIP
as well as a restart file. After preparing all the needed input files, the user needs to
modify the run script file (see Appendix B for demo) to accommodate the changes
for those key I/O environment variable settings and make sure all the inputs are
located at the correct paths ($BDSNPINP). Then, the user can invoke the model by
typing:
>$ source $BDSNPHOME/setcase.csh
>$ cd $BDSNPRUN
>$./run.BDSNP_soil_NO_[your_case].csh &
19
If run successfully, the outputs from the stand-alone BDSNP module can be found at
$BDSNPOUT and the log files for the simulation can be found at $BDSNPLOG.
The following two subsections gives some details on how to generate the static
input files by regridding the provided data in this package to the user defined
domains (especially for the soil biome map and arid/non-arid climate classification
) and how to prepare the needed meteorological variables through MCIP.
4.1 Regridding static input files
Different methods can be used to mapping the static files from one domain
configuration to another. For commercial user has the ArcGIS installed, it can be
easily achieved by using the toolbox in ArcGIS and Python program if need. For
detail instruction, please assort to corresponding tutorial materials.
Here we describe how to use the open-source software as EPA spatial allocator
(www.ie.unc.edu/cempd/projects/mims/sptial) to do the regridding. No matter
what regridding tools are chosen to use, two principles are highly recommended to
ensure optimal results: (1) keep a consistent projection type from the old to new
domain. In the folder $BDSNPHOME/biome_ncld2006_4km, the 4km resolution
CONUS soil biome map NetCDF file uses the Lambert conformal projection
configuration; (2) If possible, always map from finer resolution to coarser resolution
to keep the detailed texture.
The detailed instructions on how to download, install and operate EPA spatial
allocator can be found at the link:
http://www.ie.unc.edu/cempd/projects/mims/spatial/Spatial_Allocator_UG.pdf.
We use the sample case to regrid the 4km CONUS biome map to 12km TCEQ SIP
modeling domain to demonstrate the procedures:
Step 1. After downloading the EPA spatial allocator, follow the READ allocated at
$SA_HOME to set the system by modifying the script $SA_HOME/bin/sa_setup.csh
Step 2. Prepare all the needed libraries for the spatial allocator and recompile the
executables if needed. The needed libraries are: PROJ4 for raster tool, NetCDF for C
and C++, HDF 4 and HDF5, GEOS, GDAL for raster spatial allocator, and ANN.
Step 3. Prepare the grid description files for the old and new domains. For the
sample case, the GRIDDESC.txt file is like following:
''
'NCLD2006'
2
33.000
45.000
-97.000
-97.000
40.000
'TCEQ3dom'
2
33.000
45.000
-97.000
-97.000
40.000
''
'tx_12km'
'TCEQ3dom' -984000.000 -1632000.000 12000.000 12000.000 149 110 1
20
‘CONUS_4km’
‘NCLD2006’ -2412000.000 -1620000.000 4000.000 4000.000 1416 936 1
''
The meaning of each value in the GRIDDESC can be found in section 3.2.5.
Step 4. Go to the folder $SA_HOME/scripts to modify the sample run script
‘alloc_ioapi_to_new_mapproj.csh’ to update the corresponding environment
variables, especially to make sure the grid name and projection name of input and
output (‘INPUT(OUTPUT)_GRID_NAME’ and ‘INPUT(OUTPUT)_FILE_MAP_PRJN’)
should be indetial to the settings in GRIDDESC.txt in Step 3. When selecting the
methods to allocate the data, be sure to set variable ALLOC_MODE_FILE as
‘ALL_AVERGE’ instead of ‘ALL_AGGREGATE’ since either for the mapping of biome
or air/non-aird climate region, no need to aggregate the values from old domain to
new domain.
Step 5. Invoke the run script and it will call the executable ‘alloctor.exe’ located at
$SA_HOME/bin to do the domain resampling and mapping. The regridding process
calls the library from GDAL for raster spatial allocator to do the transformation. The
regrid static files located at $SA_HOME/output.
Figure 7 provides the comparison of the 4km CONUS biome map to the TCEQ SIP
Texas 12km domain.
Figure 7 Regridding biome map from 4km CONUS files to 12km Texas domain
4.2 Prepare meteorology file by MCIP
The Meteorology-Chemistry Interface Processor (MCIP) links meteorological models
such as MM5 and WRF with the Chemical Transport Model (CTM) of the Models-3
Community Multiscale Air Quality (CMAQ) modeling system to provide a complete
set of meteorological data needed for air quality simulations. Detailed instructions
on
MCIP
download,
install
and
operation
can
be
found
at
http://www.cmascenter.org/. When setting up the MCIP runs, you need to make
21
sure the domain configuration is identical to other static input files (e.g.
GRIDDEC.txt in section 4.1) and the weather model use the Pleim-Xiu land surface
model (2003). The BDSNP stand-alone module needed meteorological variables (see
Table 1) can be found at the surface diagnostic outputs with the logic name
METCRO2D.
5. BDSNP module further development
5.1 Change emission factor for soil biome
The default emission factors for soil biomes used in the current version of module
release are the North American geometric mean values reported by Steinkamp and
Lawrence (2010) (see ‘EF3’ in Table 2). An external text include file named
‘EF_BIOME.EXT’ located at the folder ${BDSNPSRC} (see Appendix E) is prepared
here for the ease of sensitivity study by simply changing the corresponding value for
biome category without changing the source code and recompile the executables.
5.2 Replace fertilizer file with EPIC outputs
The default daily fertilizer map is generated from the global database reported by
Potter et al. (2010), which is relatively coarse resolution and the out-of-date
information (national-level fertilizer data from 1994 to 2001). One of the
alternatives to provide the fertilizer map is to use the Fertilizer Emission Scenario
Tool for CMAQ (FEST-C, http://www.epa.gov/amad/Tools/festC.html) to consider
the different cropland farm management system scenarios and to drive the
Environmental Policy Integrated Climate model (EPIC) simulations. Figure 8
provides the comparison of default fertilizer map (see section 3.2.3, from file
$BDSNPHOME/testdate/SOILFERT_2006_benchmark.nc) and N fertilizer map from
12km EPIC outputs (considering the weighted sum of fertilizer application rate for
ammonia ‘T1_NH3’, nitrate ‘T1_NO3’ and organic nitrogen ‘T1_ON’) over continental
US on July 3. Clearly, it has the similar pattern as Potter N inputs but represents
more details and more up-to-date information.
22
Figure 8 Comparison of spatial pattern of fertilizer N input from Potter et al. (2010)
and EPIC simulation
23
Table 1 List of key I/O environment variables in BDSNP model
Environment
Variable
SOILBIOME
SOILCLIMATE
NFERT
Description
Note
[Input file]: soil biome map for model domain, 24 category
based on NLCD2006 and Köppen climate zone
[Input file]: arid and non-arid region in model domain
[Input file]: fertilizer map for model domain
*can be replaced by
EPIC results, see section
4.2
*current use the 2005
CMAQ results on 12km
CONUS domains
NDEP
[Input file]: nitrogen deposition from previous CMAQ
simulations
GRIDDESC
[Input file]: set the domain size and projection information
METCRO2D
[Input file]: meteorology fields in NetCDF format from
MCIP for model domain, which include the parameters:
(1) TMEP2-surface temperature
(2) RGRND-solar radiation
(3) PRSFC-surface pressure
(4) SOIM1-soil moisture
(5) SOIT1-soil temperature
(6) SNOCOV-snow cover
(7) WSPD10-10m wind speed
(8) CFRAC-cloud fraction
(9) LAI-leaf area index
*soil condition including
moisture and temperature
must be diagnosed by
choosing
Pleim-Xiu
LSM model (2003) in
WRF runs
SOILINSTATE
[Input file]: restart file provides the pulsing factor, dry
period, soil moisture and deposition for continuous run as
well as other model diagnostic parameters
[Output file]: hourly soil NO emission flux in ‘g/s’ or
‘mol/s’ simulated from BDSNP module
*see details in Table 3
BDSNPOUT
SOILOUT
[Output file]: same format as ‘SOILINSTATE’
24
Table 2 List of 24 soil biome emission factor from Steinkamp and Lawrence (2010)
ID
MODIS
land cover
Köppen
main
climate(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Water
Permanent wetland
Snow and ice
Barren
Unclassified
Barren
Closed shrubland
Open shrubland
Open shrubland
Grassland
Savannah
Savannah
Grassland
Woody savannah
Mixed forest
Evergr. broadl. forest
Dec. broadl. forest
Dec. needle. forest
Evergr. Needle. forest
Dec. broadl. forest
Evergr. broadl. forest
Cropland
Urban and build-up lands
Cropland/nat. veg. mosaic
---D,E
-A,B,C
-A,B,C
D,E
D,E
D,E
A,B,C
A,B,C
--C,D,E
C,D,E
--A,B
A,B
----
EF1
(world
geometric
mean)
0
0
0
0
0
0.06
0.09
0.09
0.01
0.84
0.84
0.24
0.42
0.62
0.03
0.36
0.36
0.35
1.66
0.08
0.44
0.57
0.57
0.57
EF2
(world
arithmetic
mean)
0
0
0
0
0
0.06
0.21
0.21
0.01
1.05
1.05
0.97
1.78
0.74
0.14
0.95
0.95
0.95
4.60
0.13
1.14
3.13
3.13
3.14
EF3
(North
American)
0
0
0
0
0
0.06
0.05
0.09
0.01
0.62
0.84
0.24
0.37
0.62
0.00
0.36
0.61
0.35
1.66
0.08
0.44
0.33
0.57
0.57
(1). A-equatorial, B-arid, C-warm temperature, D-snow, E-polar (see Appendix D for spatial map)
25
Table 3 List of output variables from SOILINSTATE (restart file)
Output Variable
Unit
Variable Description
Notes
PFACTOR
--
NO emission current pulse factor
DRYPERIOD
hr
Length of the dry period
NDEPRES
Ng/m2
Soil N reservoir from deposition
SOILMPREV
m3/m3
Soil moisture for previous time
step
THETA_DIAG
--
Moisture WFPS with value [0 1]
*used for
continuous
BDSNP run
*used for
continuous
BDSNP run
*used for
continuous
BDSNP run
*diagnostic para.
WET_TERM_DIAG --
Moisture scale factor
*diagnostic para.
TEMP_DIAG
K
*diagnostic para.
A_DIAG
--
NRES_FERT_DIAG --
Temperature in last simulation
hour
Base emission from soil biome
type
NRES fertilizer only
AFERT_DIAG
--
Fertilizer emission factor
*diagnostic para.
NDEPRATE_DIAG
--
Daily average N deposition rate
CRFAVG
--
PLUSEAVG
--
Daily average canopy reduction
factor
Daily average pulse factor
*used for
continuous
BDSNP run
*diagnostic para.
26
*diagnostic para.
*diagnostic para.
*diagnostic para.
Reference
Cooter, E. J., Bash, J.O., Benson, V., and Ran, L. (2012) Linking agricultural crop
management and air quality models for regional to national-scale nitrogen
assessments. Biogeosciences, 9, no. 10: 4023-4035.
Hudman, R. C., Russell, A.R., Valin, L.C., and Cohen, R.C. (2010) Interannual
variability in soil nitric oxide emissions over the United States as viewed from
space. Atmospheric Chemistry and Physics, 10, 9943-9952.
Hudman, R. C., Moore, N. E., Mebust, A. K., Martin, R. V., Russell, A. R., Valin, L. C.,
and Cohen, R. C. (2012) Steps towards a mechanistic model of global soil nitric
oxide emissions: implementation and space based-constraints, Atmospheric
Chemistry and Physics, 12, 7779-7795, doi:10.5194/acp-12-7779-2012.
Kottek, M., Grieser, J., Beck, C., Rudolf, B., and Rubel, F. (2006) World Map of the
Köppen-Geiger climate classification updated. Meteorologische Zeitshrift, 15,
259-263. DOI: 10.1127/0941-2948/2006/0130.
Lash, Ben (2014). Implementing an improved soil NOx parametrization in the
Community Multiscale Air Quality model: implications for air pollution, Master
thesis, Department of Civil and Environmental Engineering, Rice University.
Pleim, J. E and A. Xiu (2003). Development of a land surface model. Part II: data
assimilation. Journal of Applied Meteorology, 42(12), 1811-1822.
Potter, P., Navin, R., Elena M. B., and Simon D. D. (2010) Characterizing the spatial
patterns of global fertilizer application and manure production. Earth Interactions,
14, no. 2: 1-22.
Stehfest, E. and Bouwman, L. (2006) N2O and NO emission from agricultural fields and
soils under natural vegetation: summarizing available measurement data and
modeling of global annual emissions. Nutrient Cycling in Agroecosystems, 74(3),
207-228. doi: 10.1007/s10705-006-9000-7.
Steinkamp, J. and Lawrence, M. G. (2011) Improvement and evaluation of simulated
global biogenic soil NO emissions in an AC-GCM. Atmospheric Chemistry and
Physics, 11(12), 6063–6082. doi:10.5194/acp-11-6063-2011.
Wang, Y., Jacob, D. J., and Logan, J. A. (1998) Global simulation of tropospheric O3NOx-hydrocarbon chemistry: 1. Model formulation. Journal of Geophysical
Research-Atmospheres, 103, 10713–10725.
Yienger, J. J. and Levy, H. (1995) Empirical-model of global soil-biogenic NOx
emissions. Journal of Geophysical Research-Atmospheres, 100(D6), 1144711464. doi: 10.1029/95jd00370.
27
Appendix
Appendix A. Sample setting script of ‘setcase.csh’
#
setenv BDSNPHOME /work/dsc1/rz10/BDSNP_SA_V1.0
setenv BDSNPSRC $BDSNPHOME/src
setenv BDSNPLIB $BDSNPHOME/lib
setenv BDSNPEXE $BDSNPHOME/bin
setenv BDSNPRUN $BDSNPHOME/work
setenv BDSNPINP $BDSNPHOME/input
setenv BDSNPOUT $BDSNPHOME/output
setenv BDSNPLOG $BDSNPHOME/work/logdir
if ( ! -e $BDSNPINP ) then
mkdir -p $BDSNPINP/MAP
mkdir -p $BDSNPINP/MET
mkdir -p $BDSNPINP/FERT
mkdir -p $BDSNPINP/DEP
endif
if ( ! -e $BDSNPLOG ) mkdir -p $BDSNPLOG
28
Appendix B. Sample run script for BDSNP soil NO benchmark case
#!/bin/csh
#
# Stand-alone BDSNP Soil NO module V1.0
#
## !!! CHANGE setcase.csh to make $BDSNPHOME as your install directory
## Source the case folder
source $cwd/../setcase.csh
################## USER DEFINE ###################
## Grid definition
setenv GRIDDEC $BDSNPWORK/GRIDDESC
setenv GDNAM3D 12US2
## Episode choose in Julian date
set STJD = 201178
set ENJD = 201178
## Soil biome and arid/nonarid map data directory
setenv MAPDIR $BDSNPINP/MAP
## MCIP input directory
setenv METDIR $BDSNPINP/MET
## Fertilizer input directory
setenv FERTDIR $BDSNPINP/FERT
## Nitrogen deposition input directory
setenv DEPDIR $BDSNPINP/DEP
## NFERT file
setenv NFERT $FERTDIR/SOILFERT_2005_benchmark.nc
## NDEP file
setenv NDEP $DEPDIR/NDEP_2005_CMAQ_benchmark.nc
## SOILBIOME file
setenv SOILBIOM $MAPDIR/BIOME_benchmark.nc
## SOILCLIMATE file
setenv SOILCLIMATE $MAPDIR/SOILCLIMATE_benchmark.nc
## Program directory
setenv EXE $BDSNPEXE/NOX_BDSNP.exe
29
## Choose whether start cold run
setenv COLDSTART N
################# End USER DEFINE ################
## Run BDSNP
set CJD = $STJD
#loop between the episode period
while ( $CJD <= $ENJD )
if ( $COLDSTART == 'Y' ) then
## METCRO2D
setenv METCRO2D $METDIR/METCRO2D_${CJD}_benchmark.nc
## NO output from BDSNP module
setenv BDSNPOUT $BDSNPOUT/B3GTS_NO_BDSNP_${CJD}_benchmark.nc
## diagnostic and continue file
setenv SOILOUT $BDSNPOUT/SOILINSTATE_${CJD}_benchmark.nc
if ( -e $BDSNPOUT ) rm -f $BDSNPOUT
time $EXEC >& $BDSNPLOG/log.run.BDSNP.$CJD.txt
else # continous run
set PJD = `./yj_pre $CJD`
## METCRO2D
setenv METCRO2D $METDIR/METCRO2D_${CJD}_benchmark.nc
## SOILIN continuous file
set SOILIN = $BDSNPOUT/SOILINSTATE_${PJD}.nc
## NO output from BDSNP module
setenv BDSNPOUT $BDSNPOUT/B3GTS_NO_BDSNP_${CJD}_benchmark.nc
## diagnostic and continue file
setenv SOILOUT $BDSNPOUT/SOILINSTATE_${CJD}.nc
if ( -e $BDSNPOUT ) rm -f $BDSNPOUT
time $EXEC >& $BDSNPLOG/log.run.BDSNP.$CJD.txt
@ CJD++
end # End while CJD
30
Appendix C. Mapping table used to create the soil biome map based on
NLCD40 MODIS land cover categories
ID
NLCD40 MODIS CATEGORY (40)
ID
1 Evergreen Needleleaf Forest
SOIL BIOME CATEGORY (24)
19 Evergr. Needel. Foresst
2 Evergreen Broadleaf Forest
16 and 21 Evergr. Broadl. Forest
3 Deciduous Needleleaf Forest
18 Dec. Needel. Forest
4 Deciduous Broadleaf Forest
17 and 20 Dec. Broadl. Forest
5 Mixed Forests
15 Mixed Forest
6 Closed Shrublands
7 Closed shurb
7 Open Shrublands
8 and 9 Open shrubland
8 Woody Savannas
14 Woody savannah
9 Savannas
11 and 12 Savannah
10 Grasslands
10 and 13 Grassland
11 Permanent Wetlands
2 Permanent Wetland
12 Croplands
22 Cropland
13 Urban and Built Up
23 Urban and build-up lands
14 Cropland-Natural Vegetation Mosaic
24 Cropland/nat. veg. mosaic
15 Permanent Snow and Ice
3 Snow and ice
16 Barren or Sparsely Vegetated
6 Barren
17 IGBP Water
1 Water
18 Unclassified
1 Water
19 Fill value
1 Water
20 Open Water
1 Water
21 Perennial Ice-Snow
3 Snow and ice
22 Developed Open Space
23 Urban and build-up lands
23 Developed Low Intensity
23 Urban and build-up lands
24 Developed Medium Intensity
23 Urban and build-up lands
25 Developed High Intensity
23 Urban and build-up lands
26 Barren Land (Rock-Sand-Clay)
24 Cropland/nat. veg. mosaic
27 Unconsolidated Shore
24 Cropland/nat. veg. mosaic
28 Deciduous Forest
16 and 21 Evergr. Broadl. Forest
29 Evergreen Forest
19 Evergr. Needel. Foresst
30 Mixed Forest
15 Mixed Forest
31 Dwarf Scrub
8 and 9 Open shrubland
32 Shrub-Scrub
8 and 9 Open shrubland
33 Grassland-Herbaceous
10 and 13 Grassland
34 Sedge-Herbaceous
14 Woody savannah
35 Lichens
10 and 13 Grassland
36 Moss
10 and 13 Grassland
37 Pasture-Hay
24 Cropland/nat. veg. mosaic
38 Cultivated Crops
22 Cropland
39 Woody Wetlands
2 Permanent Wetland
40 Emergent Herbaceous Wetlands
2 Permanent Wetland
31
Appendix D. Spatial map of five climate zone over CMAQ CONUS
12km domain based on Köppen-Geiger climate classification
32
Appendix E. Sample of soil biome emission factor setting include file
${BDSNPSRC}/EF_BIOME.EXT
SPC,CATEGORY,EF
W,1,0.00
PW,2,0.00
SI,3,0.00
B_DE,4,0.00
U,5,0.00
B_ABC,6,0.06
CS,7,0.09
OS_ABC,8,0.09
OS_DE,9,0.01
G_DE,10,0.84
S_DE,11,0.84
S_ABC,12,0.24
G_ABC,13,0.42
WS,14,0.62
MF,15,0.03
EBF_CDE,16,0.36
DBF_CDE,17,0.36
DNF_AB,18,0.35
ENF,19,1.66
DBF_AB,20,0.08
EBF_AB,21,0.44
C,22,0.57
UBL,23,0.57
CVM,24,0.57
33