Environmental Engineering Reference
In-Depth Information
outside the US). There are two steps (run by two script files) in processing those
new land use data sets. The first step is to pre-process downloaded NLCD data
sets to get rid of spatial overlaps among data sets. The second step is to compute
the gridded land cover fraction of each land use category for user-defined modeling
grids from pre-processed NLCD data sets and MODIS data. The program outputs
gridded fractional coverage of each land category stored in a WRF-ready NetCDF
file.
Figure 1
shows the 30 m NLCD (top) and the WRF GLCC land cover data
(bottom) for deciduous forest processed for 1, 4, and 12 km resolution grids. For the
higher resolution grid domains (1 and 4 km), gridded NLCD data have much finer
gradation and better representation of deciduous forest distribution. Gridded GLCC
land use data tend to underestimate or overestimate percent of deciduous forest
coverage.
NC 1km Grid Domain
TX 4km Grid Domain
East US 12km Grid Domain
2001 NLCD and MODIS Deciduous Forest
GLCC Deciduous Forest
Percent
0 - 10
10 - 25
25 - 50
50 - 75
75 - 100
Fig. 1.
Gridded land cover - deciduous forest at different grid scales
Vegetation and surface related parameter tables for the Pleim-Xiu land surface
model (PX SLM) were modified based on NLCD and MODIS land cover
categories. Then, we ran the WRF and CMAQ models for August 2006 on the
continental US 12 km grid domain using both USGS GLCC and the new NLCD/
MODIS land cover data sets. We used WRF - ARW v3.0 with analysis nudging
winds for all levels, and T, and q
v
above PBL and indirect soil moisture and
temperature nudging in PX LSM. CMAQ v4.7 with CB05 gas-phase chemistry,
AE5 modal aerosols, M3dry dry deposition w/ preliminary bidirectional NH3, and
ACM2 were used for this study. 2006 emissions and Biogenic Emissions Landuse
Database, version 3 (BELD3) data were emission inputs to CMAQ.
