Rainfall interception in a forest stand estimated from grassland meteorological data

Meteorological data collected over grassland were used in the Rutter and analytical interception models to predict interception losses for a Scots pine stand in Thetford Forest during 84 weeks in 1975 and 1976. Rutter and analytical model estimates using grassland data overestimated the observed losses by 37% and 25%, respectively; estimates using data collected over the forest overestimated by 8% and 5%, respectively. Systematic differences in the meteorological parameters controlling the evaporation rate from wet forest and wet grass are the cause of the overestimation of interception losses when grassland data are used in the models. Mean air temperature over the wet forest was 0.75°C cooler, mean wet-bulb depression was 0.25°C smaller, and mean vapour pressure deficit was 0.4 mbar smaller than the mean values observed over wet grass in 116 hr when ⩾ 0.5 mm of rain fell at both observation sites. No significant differences between sites were found when both the grass and the forest canopy were dry. Regression equations for forest-site parameters on grassed-site parameters were used to adjust the grassed-site data for use in the models. Analytical model estimates using adjusted data were closely comparable to both observed losses and model estimates using above-forest data. Rutter model estimates using adjusted data were significantly different from the observed losses for 1976 and for the combined 1975–1976 data, but were comparable with the observed losses and model estimates using above-forest data in 1975.