Nonstationary frequency analysis of low flows under climate change

Within the scope of the study, the trends of low river flows obtained from flow observation stations were investigated and climatic variables that could cause changes in low flows were tried to be determined. By using the determined climate variables as explanatory variables, non-stationary frequency analysis studies were carried out for low flows in historical and future periods. There are basically 6 stages in achieving this goal. The first one is the determination of the flow observation stations to be used within the scope of the study and predicting the missing observations at these stations. The second is to obtain 7-day low flow data of missing data and to check the homogeneity of the obtained data. The third stage of the study is to investigate the trends of homogeneous 7-day low flow data. The fourth stage is the detailed investigation of climatic factors that cause changes in low flows. It is aimed to evaluate the possible effects of climate variables such as temperature, precipitation and large-scale climate factors (North Atlantic Oscillation, Southern Oscillation, Pacific Decadal Oscillation and Western Mediterranean Oscillation) on low flow changes. In the fifth stage, non-stationary probability distribution models were developed by using the climate variables that have a high correlation with the low flows of each station as the explanatory variable. For this purpose, the Generalized Additive Models for Location Scale and Shape (GAMLSS) approach has been used. In the last stage, low flow predictions that may occur in 20 and 50 year return periods for future periods (2030-2060) and (2070-2100) under historical and climate scenarios were evaluated by using non-stationary models whose mathematical expressions were obtained.