Investigation of climate change effect on streamflows: The case of Eastern Black Sea basin

In this study, the effects of climate change on streamflows, under the influence of RCP4.5 and RCP8.5 concentration scenarios, which predict an increase in greenhouse gas concentration in the 5th assessment report of the Intergovernmental Panel on Climate Change (IPCC), GFDL-ESM2M, HadGEM2-ES and MPI-ESM-MR the years 2021-2040 (P1), 2041-2070 (P2) and 2071-2099 (P3) were investigated using the General Circulation Model (GCM). In this study, 17 Stream Observation Stations and 11 Meteorology Stations were used in the Eastern Black Sea Basin, which is the basin that receives the most precipitation in Turkey. Within the scope of the study, machine learning Artificial Neural Networks (ANN), Adaptive Neural Fuzzy Inference System (ANFIS), Multivariate Adaptive Regression Splines (MARS) methods and classical regression method were used for modeling streamflows. The model with the highest prediction performance was determined from the Streamflow Prediction Models (SPM) calibrated with historical data. Future flows were forecasted until 2100, with the highest predicted SPM inputs being daily Temperature and Precipitation from GCMs. Using future streamflow findings, stream water availability under the influence of climate change, streamflows during dry and wet periods, and low (Q95) and high (Q5) streamflows were investigated in the basin. As a result of the study, the period in which the presence of stream water in the basin which will be mostly affected, was determined as P3. It has also been determined that the water originating from the streamflows in the basin will decrease by up to 320.93*106 m³/year in the most pessimistic case, and will increase up to 569.42*106 m³/year in the most optimistic case. It has been determined that the average streamflows in the basin will increase by 80% in the dry period and decrease by 40% in the wet period. According to the determined reference values, it has been determined that low flows will decrease by approximately 25% and high flows will decrease by approximately 35%.