Hello,
I am running into some abnormal soil moisture values in one set of simulations. I have ran a set of 10 summertime (JJA) simulations for both present-day (pd) and end-of-century scenarios (eoc). The eoc uses the pseudo global warming method. After comparing the average soil moisture over the 10 summers, I found that there were very specific locations where the soil moisture was ~150% greater at eoc than pd (below). D02 is shown, but D01 shows the same pattern.
I then looked at soil type and found that these locations correspond to high percentages of sand (soil category #1) (below).
The set of plots below compare the different scenarios for the uppermost soil layer (line color is by year 2012-2021). Those that have "sand" in the title are points where the change in soil moisture was > 100%. Those with non-sand consists of a set of ~200 points near the largest patch sandy soil in the northeast corner of the domain. You can see that in sandy soil during the pd simulations (top left), all 10 show an immediate drop in soil moisture, but nearby "non-sand" points do not (top right). Additionally, during the eoc scenarios (bottom plots), this did not occur in either sand or non-sand points.
It seems like the soil moisture values seen on the top left are erroneous, but I can't figure out why it would just occur in areas with high amounts of sandy soil. The simulations were performed with the same WPS/WRF install. Both scenarios use the same default soil data set in (soiltype_top_30s) and WRF parameterizations. Also, soil moisture was not modified for the PGW eoc scenario. Other details: WPS version is 4.5 and WRF version is 4.5.1; ERA5 forcing data.
I'm not sure what to modify to correct the issue, so I was hoping someone could help identify model settings that I could check or modify. I've attached my namelist.input. It shows 5 domains, but only 2 were used in the simulations (9- and 3-km resolution). Thank you in any guidance you can provide!
I am running into some abnormal soil moisture values in one set of simulations. I have ran a set of 10 summertime (JJA) simulations for both present-day (pd) and end-of-century scenarios (eoc). The eoc uses the pseudo global warming method. After comparing the average soil moisture over the 10 summers, I found that there were very specific locations where the soil moisture was ~150% greater at eoc than pd (below). D02 is shown, but D01 shows the same pattern.
I then looked at soil type and found that these locations correspond to high percentages of sand (soil category #1) (below).
The set of plots below compare the different scenarios for the uppermost soil layer (line color is by year 2012-2021). Those that have "sand" in the title are points where the change in soil moisture was > 100%. Those with non-sand consists of a set of ~200 points near the largest patch sandy soil in the northeast corner of the domain. You can see that in sandy soil during the pd simulations (top left), all 10 show an immediate drop in soil moisture, but nearby "non-sand" points do not (top right). Additionally, during the eoc scenarios (bottom plots), this did not occur in either sand or non-sand points.
It seems like the soil moisture values seen on the top left are erroneous, but I can't figure out why it would just occur in areas with high amounts of sandy soil. The simulations were performed with the same WPS/WRF install. Both scenarios use the same default soil data set in (soiltype_top_30s) and WRF parameterizations. Also, soil moisture was not modified for the PGW eoc scenario. Other details: WPS version is 4.5 and WRF version is 4.5.1; ERA5 forcing data.
I'm not sure what to modify to correct the issue, so I was hoping someone could help identify model settings that I could check or modify. I've attached my namelist.input. It shows 5 domains, but only 2 were used in the simulations (9- and 3-km resolution). Thank you in any guidance you can provide!