big differences between using option "modis_15s_lake" and "modis_30s_lake"

[JasonXu]

Dear all,

I am about to setup a simulation over China using the WRF / WPS version 4.1. I plan to use new LANDUSEF 'modis_15s_lake' to replace 'modis_30s_lake' but found big differences between them.

To run geogrid I use the following input data:
modis_landuse_20class_15s_with_lakes,
modis_landuse_20class_30s_with_lakes

My namelist.wps is the following:
geog_data_res = 'modis_15s_lake+30s+default',
geog_data_res = 'modis_30s_lake+30s+default'

I attach a figure showing the landuse and lu_index variables from two outputs geo_em.d01.nc.
Are the differences reasonable? which option is better for me?

I will be happy to provide additional information.
Thank you in advance and best regards,

Jason

 

[Ming Chen]

Jason,
I looked at the two plots and I agree that they are quite different. However, these plots are just for fraction of a specific type. Can you upload the plots for lu_index for me to take a look?
We are aware of the differences between the two datasets. Based on our experience in North America, the 15s dataset is more organized and better represents the 'real' lands type, while the 30s dataset shows noisy structure.

 

[JasonXu]

Hi Ming,
Thank you for your help.
The attachments are two plots for "LU_INDEX".

Best wishes
Jason

 

[Ming Chen]

Hi Jason,
The two plots show that the land use type ranges between 1-24. MODIS landuse is within 1-21. Please confirm you did run with MODIS data.