Hi! I am using the WinSRM model. I have 6 zones and am processing entire year at a time. I am using the MODIS 8 day LST and snow cover data and NWS precipitation data. But when I simulate the results, I get a somewhat 'straight line' as computed streamflow. The model is not able to simulate seasonal high and low streamflows. My current parameters are:
Cs = 0.4-0.8 (0.4 when T>0, 0.8 when T0, 0.8 when T
Land surface temperature and air temperature are completely different, you have to use air temperature. Please donot follow certain articles published by Indian authors that use MODIS LST for Temperature index models. That's wrong. Wherever you measure Air temperature there is a big difference between land and Air temperature in Cryosphere region. For eg. If Ice temperature will be -5 degrees your Air temperature is usually -15, -20 degrees.
Apart from that if you are working on a broad basin say 15 - 25k Sq. km, these data give comparatively better results on calibrations. Using these data and calibration whatever you change for a basin for 500-5000 Sq.km you ll not get good results. Please be clear about scale of parameters and the model selection. If it's a mesoscale data, please go for mesoscale model that gives regional picture and climate trend! WINSRM will he better at local level parameters, high resolution snow-cover. Calibrations will have very small role only in temperature index model!
Arun Gunasekaran , Thank you for kind reply!
My watershed is ~60000 km2, and elevation ranges from 860 - 3445 m.
If the MODIS LST data is not useful for the mode, can you suggest a better remotely sensed dataset? Otherwise I will use the weather station data. The problem is however, that I don't have any weather station data after 2500 m altitude. I will try to find a weather station.
Otherwise, can you suggest a better model if WiNSRM model is inappropriate for this area?
I appreciate your help!
Arun Gunasekaran , I ran the WinSRM model using 6 different weather stations' air temperature data (Right now I used TAvg). There is no weather station present in highest zone, so I just used a closest weather station for trial. The lapse rate I assumed was 0.6°C/100m. I also changed some other parameters like (X from 1.01-1.05, Y from 0-0.2), but I am still not able to get the good correlation. NSE is still negative. The only difference I observed was that the computed streamflow and volume was lesser than the measured streamflow, in contrast to the opposite I got when using MODIS LST data.
Data used for current simulation:
Air temperature: Weather stations (Tavg °C)
Precipitation: Weather stations (cm)
Snow-cover: MODIS snow cover data (CDC snow fraction)
Cr, Cs, DDF, Tcrit were the same as mentioned in 1st comment.
Can you please tell me what should I try now?
Thank you!
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators