Abstract

Global Circulation Models (GCMs) are increasingly used to assess possible impacts of tropical land-use change on evapotranspiration and the feedback to precipitation. Evaluation of these scenarios requires an understanding of (i) the natural cyclicity and trends in the driving processes (e.g., precipitation, air temperature and net radiation), and (ii) the land-use change impacts that have been observed within relevant landscape-scale studies.

This study examines the relative impact of several types of forest cover change on water balance and evapotranspiration in South East Asia, against (a) the natural dynamics and stochastic elements inherent within time-series of precipitation and evapotranspiration, and (b) uncertainty in our ability to characterise these dynamics. The very limited number of studies addressing water balance changes as a result of tropical forestry were examined using QAA and other statistical procedures. The Dynamic Harmonic Regression (DHR) model, recently developed at Lancaster, was then used to characterise the time-series dynamics. This state-space stochastic model, was used to detect and estimate cyclic components related to monsoon cycles, short- and longer-term El Niño Southern Oscillation cycles, cycles resulting from Walker cell migration, as well as longer-term drifts in the precipitation and evaporation data from studies undertaken across SE Asia. These results demonstrated (i) some of the features of precipitation and evapotranspiration time-series for SE Asia that GCM simulations might be evaluated against, and (ii) the magnitude of land-use change impacts that such GCMs might be expected to forecast