Question 1:
Is the calculation of 2-meter relative humidity done in metgrid.exe during WPS pre-processing?
Because when I was checking the driver data of ERA5, I found that the variable of 2-meter relative humidity was not included. However, this variable is a necessary variable for running real.exe later (that is, the files met_ em. d0 * need to include 2-meter relative humidity). Is this what it means? What variables are used to calculate the 2-meter relative humidity?
Question 2:
When I use the CanESM5 model in CMIP6 to drive the WRF model, the data (. nc format) contained in CanESM5 are: ta (high-altitude temperature), hur (relative humidity), ua (zonal wind), va (meridional wind), zg (potential height), ps (surface pressure), psl (sea level pressure), tos (sea surface temperature), tas (2-meter temperature), mrsol (soil moisture), tsl (soil temperature), lsm (land ocean mask).
I don't use the ungrib.exe program in WPS. I have a fortran program that can directly convert. nc file format to an intermediate file.
In my Fortran program, I use the first layer (1000hPa) of relative humidity、zonal wind and meridional wind from the pressure layer data instead of 2-meter relative humidity、10 meter zonal and meridional wind. The processing method for SKINTEMP is to first assign the value of tos (sea surface temperature), and then fill in the missing values of tos (sea surface temperature) on land with tas (2-meter temperature). Are these treatments reasonable? Then I ran long-term (1961-2014) simulations, such as mrol (soil moisture), tsl (soil temperature), tas (2-meter temperature), lsm (land ocean mask). I only produced one month's average data, while the rest were inputted for 6 hours. Is that okay?
Question 3:
Is there any suitable strategy for me to shorten my running time while running long-term simulations (1961-2014). Will there be significant cumulative errors if continuous simulation is conducted?
Last edited: