wrf_alessandro
Member
Dear WRF users,
I am running WRF 4.2.2 to examine the impact of urbanization on surface climate; I apply ndown in a triply-nested domain to downscale ERA5 from 15 km to 1 km. I am using the corine landcover to define the different urban categories, while for the 1 km domain I use the BEP model.
At the moment I am trying four different configurations to determine which options give the best results; in two experiments I convent the corine landcover into the USGS classes then I test different sf_surface_physics schemes (i.e. NOAH and NOAH-MP); in the other two experiments I convert the corine lancover into the MODIFIED_IGBP_MODIS_NOAH categories trying NOAH and NOAH-MP.
Looking at the simulated surface air temperature (please see the attached figure) I noted very large differences (> 2C) between NOAH and NOAH-MP and surprisingly in the experiment with NOAH-MP and MODIFIED_IGBP_MODIS_NOAH the heat island effect is barely visible, while in the other three experiments the temperature over the test city is clearly larger than surrounding area. I wonder why in the IGBP_MODIS experiment with NOAH the temperature follows the city extension while when I use NOAH-MP I do not see the temperature following the urban area? A lower sensitivity of NOAH-MP to buildings would explain this outcome, but results from the USGS experiment with NOAH-MP does not support this hypothesis. Besides, the agreement between MODIFIED_IGBP_MODIS_NOAH and USGS, when NOAH is used, would exclude issues during the conversion of corine landcover.
Unfortunately, I already performed the simulations over d01 and d02 using MODIFIED_IGBP_MODIS_NOAH with NOAH-MP, so I was wondering whether I can switch the sf_surface_physics or the landcover for d03 only once the last ndown in completed?
Any suggestions are very much appreciated.
Alessandro
I am running WRF 4.2.2 to examine the impact of urbanization on surface climate; I apply ndown in a triply-nested domain to downscale ERA5 from 15 km to 1 km. I am using the corine landcover to define the different urban categories, while for the 1 km domain I use the BEP model.
At the moment I am trying four different configurations to determine which options give the best results; in two experiments I convent the corine landcover into the USGS classes then I test different sf_surface_physics schemes (i.e. NOAH and NOAH-MP); in the other two experiments I convert the corine lancover into the MODIFIED_IGBP_MODIS_NOAH categories trying NOAH and NOAH-MP.
Looking at the simulated surface air temperature (please see the attached figure) I noted very large differences (> 2C) between NOAH and NOAH-MP and surprisingly in the experiment with NOAH-MP and MODIFIED_IGBP_MODIS_NOAH the heat island effect is barely visible, while in the other three experiments the temperature over the test city is clearly larger than surrounding area. I wonder why in the IGBP_MODIS experiment with NOAH the temperature follows the city extension while when I use NOAH-MP I do not see the temperature following the urban area? A lower sensitivity of NOAH-MP to buildings would explain this outcome, but results from the USGS experiment with NOAH-MP does not support this hypothesis. Besides, the agreement between MODIFIED_IGBP_MODIS_NOAH and USGS, when NOAH is used, would exclude issues during the conversion of corine landcover.
Unfortunately, I already performed the simulations over d01 and d02 using MODIFIED_IGBP_MODIS_NOAH with NOAH-MP, so I was wondering whether I can switch the sf_surface_physics or the landcover for d03 only once the last ndown in completed?
Any suggestions are very much appreciated.
Alessandro
