Comparative Model Estimates of Interception Loss in a Coniferous Forest Stand

Measurements of throughfall in a coniferous forest stand were analyzed and compared with model estimates of interception loss using three different types of models: the Nordic HBV model, the AMOR model and a simplified Rutter model. Two years of seasonal data were available. The 90% confidence interval on estimated average throughfall over the area decreased with storm size and approached a constant value of 15% for events larger than 10 mm of rainfall. Average interception loss was 27% during both the 1993 and 1994 seasons. The Rutter model performed slightly better than the other two; however, all models failed to reproduce the very high interception losses following some of the largest storms. The AMOR model, which is a modified version of the British MORECS model, also underestimated the loss for small and medium-sized storms, and it is necessary to include a linked storage in the model. The Nordic HBV model proved satisfactory as compared to the other two more data demanding models; it does, however, require calibration. The Rutter model has the largest potential to account for the special meteorological conditions prevailing during periods of high losses.