| |
The ultimate goal of this study is to assess future water vulnerability over Florida, based on the change in precipitation and evapotranspiration estimated using the most advanced Global Climate Model (GCM) ensemble. We evaluated the skills of CMIP5 (Climate Model Inter-comparison project, phase 5) climate models in reproducing retrospective climatology over the state of Florida for the key climate variables important from the hydrological and agricultural perspectives (i.e., precipitation (Precp), maximum and minimum temperature (Tmax and Tmin), wind speed (Ws), and relative humidity (Rhs)). The biases of raw CMIP5 were estimated using three different grid-based observational datasets as references. Based on the accuracy of various predictors such as mean climatology, temporal variability, extreme frequency, etc., the GCMs were ranked for each of the different reference datasets, climate variables, and predictors. The variation of the ranks was examined and rank-based GCM weights were assigned. The weights were then used to develop future ensembles (for 4 different RCP gas-emission scenarios) for the annual cycle of monthly mean and variance of precipitation and reference evapotranspiration (ETo). Finally the differences between the retrospective and future ensembles were investigated to assess future climate change impacts on water vulnerability using simple indices (e.g., ETo/Precp, drought index, and standardized Precp index). The uncertainties of the assessment were quantified by the spread range of ensembles and a reliability factor for the GCMs estimated using a measure of model biases and convergence criterion. |