A parameter estimation scheme of the land surface

Abstract A parameter estimation scheme for the land surface model VIC-3L is discussed. VIC-3L parameters, i.e. the variable infiltration curve parameter (B), the baseflow parameters (Dm and Ds) and the depth of the second soil layer (d2), are chosen for calibration by a systematic manual calibration approach. The VIC-3L parameters are calibrated in the first half of the period of record for 12 MOPEX watersheds in France and used to predict the stream-flow for the second half of the record. Comparisons of the simulated results using the a priori parameters with those using calibrated parameters show that the calibrated parameters, when evaluated against the a priori parameter estimates, are able to reduce the model bias and increase the Nash-Sutcliffe efficiency coefficient of the streamflow simulation. The averaged hourly Nash-Sutcliffe model efficiency coefficient for the 12 watersheds increases from 0.463 to 0.623, and that for the calibrated parameters in predicting the hourly streamflow for the validation period is 0.553. A sensitivity analysis on the calibrated parameters mentioned above shows that the variable infiltration curve parameter (B) and the depth of the second soil layer (d2) are more sensi-tive than the other two parameters (Dm and Ds). Therefore, suitable calibration for the parameters of the land surface model VIC-3L, especially for the variable infiltration curve parameter (B) and the depth of the second soil layer (d2) is very important for simulating land surface behaviour in a specific region