Akademik Veri Yönetim Sistemi
FRESENIUS ENVIRONMENTAL BULLETIN, cilt.28, sa.3, ss.2022-2034, 2019 (SCI-Expanded)
The scarcity of water and major water demands cause that droughts give rise to important economic, social and environmental results. For this reason, drought periods must be determined beforehand and must be controlled. In this study, stochastic models were developed in order to estimate monthly flow rates of the Karasu River. The best-fitting model was chosen as Autoregressive Movement Average (ARMA(2,1)) Model. By using this model, 100 synthetic series having the same length with the historical series were produced. The drought durations and severities were obtained by subjecting the historical and synthetic series to run analyses for median threshold level (q=0.5). By using these droughts, possible future drought and climate scenarios that might occur in the future may be predicted, precautions may be taken, risks may be foreseen and preparations may be made. In addition, to determine the distribution that best fits the drought characteristics, the Kolmogorov-Smirnov, Anderson-Darling and Chi-Square distribution fitness tests were applied. According to these tests, the majority of the drought durations fit the Gamma distribution, the majority of the drought severities fit the Generalized Extreme Value (GEV) distribution. The determination of the best fit is important in that it facilitates the hydrologic computations and analyses.