[Dhsvm-users] Calibration and validation with Short term data

[Keith Adams]

I think I read somewhere that calibration should use 20 data values 
(time steps) for each parameter in the model.
That is, for 1 hour time step, 20 hours for each vegetation and soil 
parameter.
Possibly 20 hours for each parameter that your are changing during 
calibration?

I could try to find the reference if you wanted it.

Keith Adams.




Safeeq wrote:


> Aloha DHSVM users

> 

>  

> 

> I am  trying to calibrate and validate DHSVM modeling using 3 yrs of 

> hourly data but kind of confused whether I should go with this short 

> record or not. The other stations which has long term historical climate 

> and precipitation data are located outside of the study area.  Can 

> somebody suggest me whether I should go with short term onsite records 

> or use the historical data from outside? So far my initial states file 

> looks ok since I ran the model for 3 years and then using the final 

> states as initial states after renaming them. I am running on hourly 

> time step with 10 m resolution.

> 

>  

> 

> Hope someone can guide me through

> 

>  

> 

> Safeeq

> 

> 

> ------------------------------------------------------------------------

> 

> _______________________________________________

> Dhsvm-users mailing list

> Dhsvm-users at u.washington.edu

> http://mailman1.u.washington.edu/mailman/listinfo/dhsvm-users