Extreme temperature return level mapping for northwest Turkey.

Risk management and design strategies for extreme temperatures are generally based on a 50-year return level in many national and international legislations. Recent studies have reported significant increasing trends in extreme temperature characteristics all over the world including Turkey. For such an analysis, return level mapping of climatic variables under non-stationarity conditions is a new topic. For this reason, the goal of this study was to introduce a mapping methodology for a 50-year return level of the extreme temperatures for northwest Turkey by considering the non-stationarity conditions. Generalized Extreme Value distribution was fitted for this purpose to de-trended extreme temperature time series. Multiple regression models were developed accordingly based on the physiographic parameters to interpolate point-based return levels over the study area. Observed annual maximum and minimum temperatures as well as the 50-year return levels indicated a high spatial variability over the region. According to the results of the study, Continentality Index, elevation, and latitude are determined to be the appropriate predictors for obtaining annual minimum temperature return level maps while the first two parameters are optimal for a statistical model of the maximum temperatures. Return level extreme temperature maps were produced with 2.5 × 2.5-km grids. Under climate change and its variability, it is of utmost importance for any method to consider the non-stationarity in calculations when modeling the changing conditions. The maps that are produced can be a tool for adapting to intensifying heat waves under climate change and its variability while the applied methodology can be transferred to any other region. [ABSTRACT FROM AUTHOR]