Abstract

Selective harvesting of primary rain forest, followed by regeneration, changes the rate and pattern of canopy re-evaporation which is a key element of the canopy-interception budget. These effects then propagate to modify: (a) local convective rainfall (i.e., part of the direct precipitation budget), (b) soil moisture and thus drought stress, and (c) the flood behaviour of rivers. Given typically sparse raingauge networks, the spatio-temporal complexity of direct and canopy-intercepted precipit-ation can result in highly inaccurate precipitation estimates for flood or water-balance modelling.

Within this study, a dense network of 500 direct and net precipitation gauges were installed within a 10 km2 region covered by a mosaic of primary, lightly-impacted and highly-damaged rain forest. Twelve months of measurements from this intensive network were combined with 10-15 years of gross precipitation records for Eastern Sabah (Malaysian Borneo), and analysed using a combination of time-series and geostatistical modelling.

We have been able to identify: (a) diurnal, seasonal and ENSO patterns within the direct precipitation, (b) the spatial extent of events of these characteristic periods, (c) the impact of spatio-temporal patterns in direct precipitation on components of the canopy-interception budget, and (d) systematic variations in canopy re-evaporation between different characteristic elements of the disturbance mosaic.