Transfer function based downscaling methods for assessing future rainfall using GCM output

General Circulation Models(GCMS) are tools designed to simulate time series of climate variables globally, accounting for effects of greenhouse gases in the atmosphere. They are good at continental and hemispherical spatial scales and incorporate a large proportion of the complexity of the global systems. They are however unable to represent local subgrid scale features and dynamics. Moreover, accuracy of GCMs is very poor in simulating hydrologic variables such as precipitation, evapotranspiration, etc. These limitations of the GCMs restrict the direct use of their output in hydrology. Downscaling, in the context of hydrology, is a method to project the hydrologic variables (e.g., rainfall and streamflow) at a smaller scale (both spatially and temporally) based on large scale climatological variables (e.g., mean sea level pressure) simulated by a GCM. Transfer function based statistical downscaling methods gain popularity among researchers because of their simplicity. In the present study linear regression, Artificial Neural Network and Support Vector Machine are used as transfer functions in downscaling. The results obtained from all these three methods are compared. The proposed methodology is demonstrated with the projection of monsoon rainfall in meteorological subdivision of North-East India.