Non stationary frequency analysis of standart duration annual maximum precipitation in Aegean Region

The construction of water structures is a high-cost and long-lasting, laborious process. Frequency analysis of extreme hydrological events (such as downpours), which should be taken into account during the design and planning of these structures, falls into the priority field of work of water engineers. In recent years, precipitation has spotted to become irregular as a result of global warming, the effect of which has occurred due to increasing climate change (Ministry of Agriculture and Rural Affairs, 2007) Since the trends in precipitation, disrupt the characteristics of stationarity, non stationary containing time-varying parameters should be used in modeling extreme precipitation In this study, modelling of precipitation processes under the non-stationary conditions, as a consequence of the climate change, was carried out by using different probability distribution functions. Within the scope of the study, which was modelling Standard annual maximum precipitation data, trend tests were applied first. Regardless of the test results, Generalized Extreme Value (GEV) distributions with both stationary and non-stationary parameters were fitted to the data received from stations for different periods of time. GEV distribution parameters, goodness of the fit of the distributions and the precipitations expected to occur during various return periods were calculated via the exTremes and ismev packages in the R Studio open source software. As a result of the study, it was found that non-stationary distributions are more suitable for precipitation periods where the trend is observed.