Evaluation of canopy interception schemes in land surface models

SummaryField measurements of interception loss ratio published in the literature for different locations over the globe are used in this study to evaluate four canopy interception schemes incorporated into the Community Land Model version 3 (CLM3). Those are the CLM3 default scheme, the Shuttleworth scheme, a scheme that considers the sub-grid variability of both precipitation and canopy water storage (WW07 scheme), and a herein revised WW07 scheme (rain type-distinguished WW07) that distinguishes for rain type (convective versus stratiform). With the exception of Shuttleworth scheme that exhibits significant temporal resolution sensitivity in simulating the canopy interception loss, the rest three schemes are validated against observations. The rain type-distinguished WW07 scheme exhibits the best performance in producing the annual rainfall interception loss ratio over the globe with an overall mean bias error (MBE) of −0.007 and root mean square error (RMSE) of 0.121. This scheme also produces a realistic seasonality of the rainfall interception loss ratio in tropical areas, which is verified against field measurements from a site in the Amazon rainforest.