(RESOLVED) The Method of Geogrid.exe input Landuse Data

[Yu-Tzu]

Dear all,

In the GEOGRID.TABLE, there are just some different interpolated and averaged methods for landuse data input.
But how about the method like dominated or mosaic which described in some papers and articles?

What I am confused is, while we use the interpolated and averaged methods, we will get the value calculated by these numerical methods.
*For example, using "Simple four-point average interpolation" with USGS 24-categories：
→F(x,y) = (a11+a12+a21+a22)/4
→F(x,y) = (1[=urban]+1[=urban]+16[=water bodies]+16[=water bodies])/4
→F(x,y) = 8.5
→F(x,y) = "Shrub Land"
But actually, the value is not exist in the raw landuse data!
So, does it mean we will input the wrong land type in our WRF-model simulation?
Furthermore, whether I turn on the "mosaic" in WRF namelist can correct this problem?


Thank you for your reply!

 

[Ming Chen]

The default option to process category variable is 'nearest_neighbor'
For category variable, geogrid program considers all source pixels that lie within each WRF grid cell, and finds the fraction of the WRF grid cell that is comprised of each category in the source data. Thereby, land use type is correct.