Skip to content

Latest commit

 

History

History
138 lines (105 loc) · 11.1 KB

CMAQ_UG_appendixC_spatial_data.md

File metadata and controls

138 lines (105 loc) · 11.1 KB
Raw

<< Previous Appendix - Home - Next Appendix >>

Appendix C: Processing Spatial Data for CMAQ Inputs

C.1 Geospatial Data

Air quality modeling requires many spatial data to generate anthropogenic, biogenic, fire, sea salt, dust, and NH3 emissions. In addition, land surface characteristics such as land cover types with vegetation leaf area index (LAI) and fraction, albedo, and soil types are required in modeling the exchanges of heat, moisture, and momentum between the land and atmosphere and dry deposition of trace chemicals (e.g. O3 and NH3). It is important to use a consistent coordinate system for all the geospatial data used in emission, meteorology, and air quality modeling. Most of the geospatial data required for the Sparse Matrix Operator Kernel Emissions (SMOKE)/Weather Research and Forecasting (WRF)/CMAQ modeling can be generated using the Spatial Allocator (SA) that includes three components developed for specific applications: Vector, Raster, and Surrogate Tools.

In using the spatial data, it is important to know the datum, which is a spheroidal (either spherical or ellipsoidal) surface that represents the surface of the earth, and the projection, which is a mathematic transformation that converts a location on the datum to the location on a flat plane. The following sections briefly describe the appropriate datum and projections to use with the CMAQ system and the methods for generating the needed spatial data in the correct form.

C.2 Geodetic datum

A geodetic datum is a coordinate system used to define a location on the Earth. There are many datums used in spatial datasets depending on what geographic regions they are and how the Earth’s surface is approximated as a spheroid. Most of U.S. geospatial data are defined in North American Datum of 1983 (NAD83) and the global data sets are often defined in World Geodetic System 1984 (WGS84).

WRF datasets are in WGS84. All latitude-longitude geographic data sets used in a CMAQ simulation, such as emissions shapefiles, land use or biogenic data files, and the ocean file, should be in WGS84 so that they are spatially aligned with the WRF files. For simulations over North America, NAD83 is only slightly different from the WGS84 datum. As a result, NAD83 can be used for North America domains without introducing spatial misalignment issues in the model datasets.

C.3 Spatial Data Projection

CCTM can use any of the four map projections defined for WRF. The four map projection coordinate systems are regular latitude-longitude geographic, Lambert conformal conic, Mercator, and Polar stereographic. However, users should note that several of the PREP and POST tools that are part of the CMAQ system do not currently support the Mercator projection. These include ICON, BCON, sitecmp, sitecmp_dailyo3, bldoverlay, hr2day and writesite.

It is important to know that in projecting spatial data that is in WGS84 to the CMAQ projection or projecting CMAQ data to another map projection, users SHOULD NOT do any datum transformation. This is consistent with the WRF preprocessing system (WPS). Datum transformation will result in geographic location shifting.

The CMAQ domain projection is defined through the PROJ coordinate transformation software library using a spherical surface with an earth radius of 6370000 m to match the WRF domain projection definition. Once an input dataset is in WGS84 the following examples can be used to define the projection transformation needed to match the WRF data:

Lambert Conformal Conic: "+proj=lcc +a=6370000.0 +b=6370000.0 +lat_1=33 +lat_2=45 +lat_0=40 +lon_0=-97"

Polar stereographic: "+proj=stere +a=6370000.0 +b=6370000.0 +lat_ts=33 +lat_0=90 +lon_0=-97 +k_0=1.0"

Mercator: "+proj=merc +a=6370000.0 +b=6370000.0 +lat_ts=33 +lon_0=0"

Geographic: "+proj=latlong +a=6370000.0 +b=6370000.0"

C.4 Spatial Data Generation

Emission spatial allocation surrogates are required for generating anthropogenic emissions by SMOKE to spatially allocate county-based emission inventories to model grid cells. Emission surrogates can be based on population, roads, airports, railroads, and land use spatial data sets. The SA Vector and Surrogate Tools can be used to generate all needed emission surrogates for SMOKE.

Biogenic emissions require land use input including different tree species. There are two ways to compute the required input for the domain covering the continental U.S. (CONUS).

  1. The original method—re-grid Biogenic Emissions Landuse Database, Version 3 (BELD3) using a SA Vector allocation tool. The BELD3 data is generated from the early 1990s AVHRR land cover data and FIA tree species at the county level.
  2. The second method—use the SA Raster BELD4 land cover generation tool to generate model domain land use data with tree species. Then, a provided utility is used to convert the generated land cover data into an I/O API format for CMAQ input. The limitation for this tool is that the early 1990s county-level FIA tree species table is still used in allocating FIA tree species (this is also the case for the 1st approach).

Fire emissions require fire location, burned areas, and detailed fuel load information. Fire locations are available via satellite detections from the Hazard Mapping System (HMS) or ground level reports from the National Fire and Aviation Management web application. Burn Area estimates can be obtained from GIS based sources such as the Geospatial Multi-Agency Coordination (GeoMac) website or the U.S. National Historical Fire Perimeters Data Basin Dataset. Fuel loading is estimated using a geospatial dataset such as the US Forest Service Fuel Characteristic Classification System (FCCS). All these information sources can be used to estimate fire emissions. An example of a tool that can be used to generate fire emissions is the BlueSky modeling framework. BlueSky modularly links a variety of independent models of fire information, fuel loading, fire consumption, fire emissions, and smoke dispersion. Using these tools and estimating fire emissions can be quite complex so datasets of fire emissions have been created for the community. Examples of such datasets are the Fire INventory from the National Center for Atmospheric Research (FINN) or the Global Fire Emissions Database (GFED).

Sea spray emissions require open ocean and surf zone (50m) buffer fractions for the modeling grid cells in an I/O API file. For most of North American domain, a SA Vector allocation tool can be used to generate the surf zone and open ocean file from a polygon shapefile with land, surf zone buffer, and open ocean in SA data directory. For areas outside U.S., users have to generate a surf zone polygon shapefile with has the same attribute as the file in the SA to use the tool. See the CMAQ Tutorial on creating an ocean file for step by step instructions on creating this CMAQ input file. Chapter 6 has additional information on sea spray module in CMAQ.

DMS and halocarbon emissions are calculated in-line and require the presence of DMS and CHLO in the ocean file. A Python note book can be used to add DMS and CHLO to an existing ocean file. See the CMAQ Tutorial on creating an ocean file for step by step instructions on creating an ocean file and for adding DMS and CHLO to the ocean file.

NH3 emissions from agricultural lands can be estimated using the CMAQ bi-directional NH3 model. The input for the CMAQ bi-directional NH3 model is generated by the Fertilizer Emission Scenario Tool for CMAQ (FEST-C) system. FEST-C contains three main components: Java interface, Environmental Policy Integrated Climate (EPIC) model, and SA Raster Tools. The interface guides users through generating required land user and crop data and EPIC input files and simulating EPIC, and extracting EPIC output for CMAQ. The generated BELD4 land use data by FEST-C needs to be converted into an I/O API format using a utility program in FEST-C for CMAQ input. Note that the BELD4 data used for FEST-C is generated by the 2nd approach described above in Biogenic emission generation approaches.

Land use and land cover data for surface flux modeling in meteorology and air quality can be generated using WPS or the SA Raster Tools. It is important to use consistent land use data in both meteorology and air quality modeling. For the U.S., WPS contains re-gridded 9-arc second (around 250 m resolution) 2011 NLCD land cover, imperviousness, and canopy data while 2011 MODIS land cover is used for areas outside the U.S. In addition, users can use the land use re-gridding tool in the SA Raster Tools system to generate land cover data for any domain directly using NLCD (at 30 m resolution) or/and MODIS land cover data (at 1 km or 500 m resolution). Users can use a provided R utility in SA to update their geogrid land cover data using the more accurate land cover data generating using SA.

<< Previous Appendix - Home - Next Appendix >>
CMAQ User's Guide (c) 2022