Generating regional climate change scenarios with nearest neighbor algorithm in Western Black Sea basin, Turkey

The specific objectives of this thesis are two-fold: 1) demonstrate the ability of the stochastic weather generator to simulate regional monthly maximum and minimum temperature and precipitation data based on the historic records in north western part of Turkey, 2) develop an ensemble of climate scenarios for assessment of watershed vulnerability to extreme events such as floods and droughts. K nearest neighbor weather generator is utilized for this study. A set of statistics is computed for each month and summarized with box-plot to evaluate the model results. These results are based on codes written in R programming language. Since the simulations reproduce the historic statistics reasonably well, the same method is used to produce alternative climate data sets under two different climate change scenarios. These climate scenarios are: 1) increasing precipitation, and 2) increasing temperature. While ?increasing the precipitation? scenario is useful for assessing the risk of floods events, ?increasing the temperature? scenario is useful for the analysis of droughts risk in the basin.