Development of a Statistical Downscaling Approach for Projecting Future River Runoff from GCM Outputs

C.M. Liu, S.H. Lin, S.S. Lin, T.H. Yen, and J.T. Kuo (Taiwan)


Statistical downscaling; GCM; river runoff; climate change.


In this study, a statistical approach is adopted to project the temperature and precipitation field of the 21st century under SRES A2, B2 over Taiwan (~ 150 x 400 km2 ) based on the global circulation model (GCM) outputs.. Multivariate regression is used to build the relationship between GCM outputs and local observed climate data, and, thereafter, to project the future climate. In the meantime, artificial neural network has been applied to estimate the river runoff based on the projected climatic parameters. Results of projected change of river runoff are then evaluated.

Important Links:

Go Back