---

A methodology is presented for the stochastic generation of daily rainfall which accounts for changes to the climatic inputs. The focus of the study is an example catchment in Iran. The methodology addresses the inability of GCMs to provide suitable future scenarios for the time and space scales required for a water resource impact assessment for a small catchment. One stochastic model for rainfall (Neyman-Scott Rectangular Pulses, NSRP, model) is used to generate daily rainfall sequences and then validated using historic records. For present climate conditions, the NSRP model is fitted to observed rainfall statistics. GCM outputs are then downscaled using regressions between atmospheric circulation indices (ACIs) and rainfall statistics. The relationships are then used to predict the rainfall statistics for future conditions using GCM outputs. In this respect, climate change impacts are studied and assessed in this paper. Generated rainfall scenario can then be used as inputs to a rainfall-runoff model in order to generate daily streamflow data which is not investigated here.