Modelling rainfall interception by a lowland tropical rain forest in northeastern Puerto Rico

Recent surveys of tropical forest water use suggest that rainfall interception by the canopy is largest in wet maritime locations. To investigate the underlying processes at one such location-the Luquillo Experimental Forest in eastern Puerto Rico-66 days of detailed throughfall and above-canopy climatic data were collected in 1996 and analysed using the Rutter and Gash models of rainfall interception. Throughfall occurred on 80% of the days distributed over 80 rainfall events. Measured interception loss was 50% of gross precipitarion. When Penman-Monteith based estimates for the wet canopy evaporation rate (0.11 mm h(-1) on average) and a canopy storage of 1.15 mm were used, both models severely underestimated measured interception loss. A detailed analysis of four storms using the Rutter model showed that optimizing the model for the wet canopy evaporation component yielded much better results than increasing the canopy storage capacity. However, the Rutter model failed to properly estimate throughfall amounts during an exceptionally large event. The analytical model, on the other hand, was capable of representing interception during the extreme event, but once again optimizing wet canopy evaporation rates produced a much better fit than optimizing the canopy storage capacity. As such, the present results support the idea that it is primarily a high rate of evaporation from a wet canopy that is responsible for the observed high interception losses. (C) 1999 Elsevier Science B.V. All rights reserved.