Different result trends in different correction methods by TRAIN

[yunxiao]

Different result trends in different correction methods by TRAIN

 

Dear David and all,

The attachment is the results of ERA, GACOS, linear, MERRA, MODIS, powerlaw, WRF, models for my study area in Hong Kong, and the results of w, u, u-o, u-do.

I use different atmospheric delay correction methods in the Hong Kong area I studied, but the effects of different correction methods are quite different.

The correct result range obtained by the power law method is -298.5 to 394.6, and the corrected result range of the modis data is -130527 to 78744.2. The results obtained by the two methods are significantly larger than those results of the erai model (-2.7 to 3), the gacos model (-12.3 to 14.3), the wrf model (-4.9 to 4.1), and the merra model (-3.3 to 4.2). So I have a few questions to be asked.

1)        Why are the results of power law methods and modis data so large? Is it because the parameters I set are wrong or why? Please give me some guidance. The spatial bands in powerlaw is set [8000 32000] (True or no?). And I used gdalwrap -overwrite Water_Vapor_Near_Infrared_name_band.hdf -of netCDF MODIS_yyyymmdd.nc and gdal_translate MODIS_yyyymmdd.nc -of GMT MODIS_yyyymmdd.grd to generate modis data in grd format. Is it wrong?

2)        Please help me see if the dem in the grd format is correct, see the attached figure (DEM debug test). I used the construct_dem.sh command to generate the dem in grd format.

3)        I want to use GFS data after 2015. How can I download it?

 

Thanks for any hint in advance.

Cheers,

yunxiao