The spatial resolution of Global Climate Models (GCMs) is too coarse to resolve regional scale effects and to be used directly in local impact studies. Statistical downscaling techniques offer an alternative to improve regional or local estimates of variables from GCM outputs. In this study, an Adaptive Neuro-Fuzzy Inference System (ANFIS) using Fuzzy C-Means (FCM) clustering is presented and assessed, to reconstruct the climate observed in Iran.\nBoth Automated Statistical Downscaling (ASD) and ANFIS models, as these two models are evaluated and inter-compared, are calibrated using National Center for Environmental Prediction (NCEP) reanalysis data, before using CGCM3 predictors. The comparison is performed over the period of ASD and ANFIS calibration (1961–1975) and over the validation period (1976–1990), using (NCEP) predictors. The criteria for results comparison are: computing the amount of model explained variance (R2), Root Mean Square Error (RMSE) for the estimated statistics and climatic indices.\nMaximum, minimum and mean temperature are generated for the periods 2011–2040, 2041–2070 and 2071–2100 and compared to the 1961–1990 period. Results of the comparison indicates the increase of temperature for each station.

Login to Download PDF

Please login with your Paper ID and password to access the full PDF.

🔑 Login to Download