ANTHROPOGENIC EMISSIONS

Development of anthropogenic emissions for use with WRF/Chem

 

At this time there is no single tool that will construct an anthropogenic emissions data set for any domain that the user selects. This places the requirement upon the user to construct the anthropogenic emissions data set for their particular domain and desired chemistry. Therefore, one must know a priori the preferred domain location and chemistry options that will be used in the simulation. The provided "raw" anthropogenic emissions data set (see below) can be used if the domain is located over the 48 contiguous states of the United States, but there are some limitations. If your simulation domain falls outside the regioned covered by the provided emissions data, then one can use the subsequent conversion routine after building your own 3-D data set. The next section shows the suggested methodology for constructing your own anthropogenic emissions data set.

Anthropogenic emissions methodology for WRF/Chem

 

The methodology for constructing your own anthropogenic emissions data set is:

• Start with raw emissions data (data could come from a variety of data sources)
• A 4-km emissions data set (area and point source) is available for the U.S. (see below)
• Specify, or make a table listing, the speciation for the desired chemical mechanism (see following section)
• Provided U.S. routines assume RADM2 chemical mechanism
• Prepared the 3-D (or 2-D) anthropogenic emissions data set
• Account for rise of emissions from stack, biomass burning, etc.
• Output data in proper format - user can change format to match their needs in module_input_chem_data.F
• Convert the emissions data to a WRF netCDF data file
• Data format matches that used in conversion routine (see module_input_chem_data.F)
• Map data (extracted from the header in file wrfinput_d01) is needed for some plotting routines to function properly

Brief description of how to construct an anthropogenic inventory

Start with a list of known chemical species that are emitted in the domain of interest. These species will need to be translated into a list of chemical species that are used by the numerical model. Often this will result in either lumping several emitted chemical species into one simulated species, or one emitted species being partitioned into fractions of several simulated species. For example, the following anthropogenic speciation list has in the columns the following information:

• the names of the emitted species,
• the names of the emitted species used in the numerical model,
• the fractional amount of the emitted species used in the model under the emissions name,
• the molecular weight (not used),
• the technical name of the emitted species.

Next a program needs to be constructed that will read your emissions data, convert the species using the constructed table and finally map it onto your simulation domain. Note that elevation data for your domain may be needed if the emissions occur above the surface. Provided on the WRF/Chem ftp site is a program that can be used with the US. anthropogenic emissions inventory - emiss_v03.F - which can be used as an example. The product of the program is a binary three-dimensional emissions data file for the simulation domain. The emissions are updated every hour. If the binary format of the output data is not functional for your needs, the output format can be of modified to fit the format for your needs. However, the WRF/Chem subroutine med_read_bin_chem_emiss (mediation (layer) read binary chemistry emissions) inside module module_input_chem_data.F will also need to be modified so that the converted raw emissions data can be read by the convert_emiss program.

 

Brief Description of the available U.S. anthropogenic inventory

 

First, this inventory is meant as an 'interim' inventory until updates are made available. This inventory is based upon the U.S. EPA's 1999 National Emissions Inventory (NEI-99, version 3) released November 2003, and the 4km horizontal resolution spatial surrogates released by EPA in September of 2003. Revisions to the EPA-99v3 inventory as of March 5, 2004 are also included. Canadian area and mobile source emissions are also included south of ~52 deg. N latitude. However, in this inventory Canadian point sources are not included due to issues related to non-disclosure rules within Canada, but are expected to be included in the updated inventory. Emissions estimates for Mexico from the BRAVO study (north of ~24 deg N latitude) are also included. The emissions are representative of a typical summer day (average of weekday and weekend days), as derived by temporal allocation factors specific to each SCC (source classification code) provided by the EPA. All biogenic sources of VOC and NOx, and all fire-related emissions have been removed from the U.S., Canadian and Mexican source files in this inventory.

 

This invetory is designed for regional scale photochemical models of North America that require emissions data for NOx, VOC, CO, SO2, NH3, PM2.5 and PM10. Forty-one speciated VOC compounds, and 5 PM2.5 aerosol species, along with the 7 primary species are divided into 24 average hourly emissions. The large number of VOC species is designed for flexibility in assigning VOC to various photochemical mechansism codes. The 4 km horizontal data spacing of the area emissions obviously allows for relatively high spatial resolution, and less uncertainty when compiling a specific model inventory with coarser horizontal grid spacing.

 

Gridded area emissions are based on EPA's spatial surrogate file designed for the NEI-99 inventory.  Point emissions are given in terms of latitude and longitude location, with stack parameter information included to allow for plume-rise calculations.

 

The intent of this inventory is to provide:

1.      a data-set that various model groups can use as a reference for air quality model development and research associated with field and model studies.

2.      the necessary information, and in a convenient format, that will allow any model group to incorporate this inventory into their specific emissions processing techniques.

3.      continental scale coverage so that models of various spatial scales, from regional to global, can reference a single data-set for North American emissions.

4.      VOC emissions with sufficiently detailed speciation to allow for emissions calculations within many lumped photochemical mechanisms.

Important Caveats and Qualifiers:

1.      Canadian point emissions are not available in this inventory due to issues associated with the nondisclosure rules within Canada. These sources will be included in a future release.

2.      All vegetative biogenic sources of NOx and VOC have been removed from the inventories of the three countries.

3.      All sources associated with wildfires and prescribed burning have been removed from the inventories.

4.      This inventory is only a recompilation of data available from public web-sites and servers at EPA (e.g., http://map.ngdc.noaa.gov/website/al/emissions/viewer.htm) .  It is not meant to be a substitute for more detailed emissions processing (e.g. the Sparse Matrix Operational Kernel Emissions (SMOKE) model), but rather an alternative for those university and institutional research groups that do not have access, or the need for the level of detail, that more advanced emissions processing programs offer.

Specific details about the WRF/Chem emissions inventory

Two Unix tar files, em99v3_file1.tar and em99v3_file2, contain the NEI-99 version 3 emissions inventory developed for the ICARTT/NEAQS-2004

regional forecast model intercomparison study.

The tar file em99v3_file1.tar contains the following 3 subdirectories:

·       point (directory of point type emission files)

·       grid_loc (latitudes and longtitudes of grid centers and corners)

·       test_progs (2 fortran programs that read point and area files)

The tar file em99v3_file2.tar contains the following subdirectory:

·        area4k  (directory of area type emission files)

All files are ascii text files, and all emission files in the area4k/ and point/ subdirectories were compressed using gzip.

All the information you need to understand the structure, units and species contained in the inventory can be found in the two FORTRAN programs in test_progs/ (area_read.f and point_read.f). 
These programs can be compiled with FORTRAN 90, or 77 and run in the test_progs/ directory. A listing of totals and the maximum daily average, and its location for each emitted species is written to a file (area_read.output or point_read.output). The output generated 
from our SUN V880, using 32 bit accuracy, is supplied for comparison in the output files dated 032404. The area_read.output file also gives the latitude and longitudes of the corner points. The program and output also lists the units of each emission. The sums and 
maximums of both the daily averages, as well as for the average hourly emissions between 19:00 and 20:00 UTC (the emission files in the HR20/ subdirectories) are printed in the output files.

The area emissions are specific to a particular Lambert conformal projection on a 1332 (west-to-east) by 1008 (south-to-north) grid, coincident with the spatial surrogate file. (Details of the projection can be found in the LAMBC subroutine in the area_read.f program).

The point emissions are not gridded, but require the text file - Relpnt_info.txt in the point/ subdirectory as an index for all of the point emission files. The Relpnt_info.txt file contains 
103249 records giving various FIP codes, state and EPA identifiers, latitude, longitude, and 5 stack parameters for each distinct point emission. Each point emission file contain 103249 records that correspond sequentially to the records in the Relpnt_info.txt file. 
See the point_read.output file and the point_read.f subroutine for units of the stack parameters in the Relpnt_info.txt file. The emission type (IRTYP variable) is the emissions type in the EPA's point identification file.

               1 = fugitive emission

               2 = vertical stack

               3 = horizontal stack

               4 = goose-necked

               5 = fitted with rain cap

6 = downward facing

Only stack flow parameters (temperature, velocity ..etc) are given when the emission type is not 1 or 6. It is expected that the users of this inventory will use the information in the indexing file to do their own plume-rise calculations.

The 2 subdirectories: point/ and area4k/ contain 24 subdirectories (HR01/, HR02/ .... HR24/) corresponding to UTC hours 0-1, 1-2, ..... 23-24 for the 41 VOC, 5 PM2.5, and 7 primary emissions. 
There is also a dayav/ subdirectory that gives the diurnal average of the emissions over the 24 hours. (The dayav/ subdirectories include speciated VOC within a subdirectory VOCspec/ and the speciated PM2.5 within a subdirectory PM25spec/, while the HRxx/ subdirectories 
include all emissions within one subdirectory).

Units of all emissions are either in short (English) tons per day, or short tons per hr for all of the primary and speciated PM2.5 emissions. Speciated VOC are in units of moles/hr under all 
cases. The primary emissions are by weight of compound, except NOx which is defined as weight of NO2 molecules. Area emissions are rates per grid cell (4km by 4km in the specific map projection), point emissions are in rates per individual point source.

The speciated VOC are listed below, the first 30 compounds correspond to the SAPRC-99 lumped photochemical mechanism (Carter, 1999). The last eleven species are specific VOC compounds separated 
out of the SAPRAC-99 mechanism, chosen either to provide compatability with the RADM2 and RACM mechanisms, or to provide for comparisons to specific VOC compound measurements. VOC speciation was done by applying EPA's SPECIATE 3.2 software to each SCC specific emission, and 
lumping or separating each of the 760 possible NMHC emissions according to the SAPRC-99 lumped species assignments (Carter, 1999). The recommended, SCC specific conversions of VOC to total organic gas were also applied to the raw VOC 
emissions to derive the emissions of less reactive species (e.g. Methane, Ethane, and Acetone).

1) Methane

2) Alkane1 ;k(OH) < 500 /ppm/min; (primarily Ethane)

3) Alkane2 ;500 < k(OH) < 2500 /ppm/min (excluding C3H8,C2H2,organic acids)

4) Alkane3 ;2500 < k(OH) < 5000 /ppm/min (excluding butanes)

5) Alkane4 ;5000 < k(OH) < 10000 /ppm/min (excluding pentanes)

6) Alkane5 ; k(OH) > 10000 /ppm/min

7) Ethylene

8) Olefin1 ; k(OH) < 70000 /ppm/min (excluding propylene)

9) Olefin2 ; k(OH) > 70000 /ppm/min (excluding dienes and styrene)

10) Isoprene

11) Sum of Terpenes

12) Aromatic 1 ; k(OH) < 20000 /ppm/min (excluding benzene and toluene)

13) Aromatic 2 ; k(OH) > 20000 /ppm/min (excluding xylenes)

14) Formaldehyde

15) Acetaldehyde

16) Higher Aldehydes

17) Benzaldehyde

18) Acetone

19) Methyl-Ethyl ketone

20) PRD2 (see SAPRC-99)

21) Methanol

22) Glyoxal

23) Methylglyoxal

24) Biacetyl

25) Phenols

26) Cresols

27) Methacrolein

28) Methylvinyl ketone

29) IPRD (see SAPRC-99)

30) Unreactive

31) Propylene

32) Acetylene

33) Benzene

34) Butanes

35) Pentanes

36) Toluene

37) Xylenes

38) Propane

39) Dienes

40) Styrene

41) Organic Acids

The Speciation of PM2.5 was done similar to that of the VOC, using EPA recommended SCC specific PM2.5 profiles. The list of PM2.5 species is:

1) PMFINE - unspeciated primary PM2.5

2) PSO4    - PM2.5 sulfate

3) PNO3    - PM2.5 nitrate

4) POA     - PM2.5 organic aerosol

5) PEC     - PM2.5 elemental carbon