Solar radiation, which is used in hydrological modeling, agricultural, solar energy systems, and climatological studies, is the most important element of the energy reaching the earth. The present study compared, the performance of two empirical equations -Angstrom and Hargreaves-Samani equations- and, three machine learning models -Artificial Neural Networks (ANN), Support Vector Machine (SVM), and Long Short-Term Memory (LSTM)-. Various learning models were developed for the variables used in each empirical equation. In the present study, monthly data of six stations in Turkey, three stations receiving the most solar radiation and three stations receiving the least solar radiation, were used. In terms of the mean squared error (MSE), root mean squared error (RMSE), mean absolute error (MAE), and determination coefficient () values of each model, LSTM was the most successful model, followed by ANN and SVM. The MAE value was 2.65 with the Hargreaves-Samani equation and, decreased to 0.987 with the LSTM model while MAE was 1.24 in the Angstrom equation and decreased to 0.747 with the LSTM model. The study revealed that the deep learning model is more appropriate to use compared to the empirical equations even in cases where there is limited data.

No abstract

In the study, variability in reference evapotranspiration (ET0) from Southeastern Anatolian Project (GAP) area was investigated with linear regression method. For the purpose, seasonal ET0 time series were formed from monthly reference evapotranspiration (ET0). The ET0 data sets of three sites showed a statistically significance decreasing trend while there was upward trend in some seasons of two sites. But, variation in all seasonal ET0 time series of Kilis site was not detected.

The unnatural change in the globe under influence of devastating global warming has been quashing the overall functioning of ecosystem since industrial revolution. Thus, the human-induced disaster caused by proportional increase of greenhouse gases in the atmosphere has affected the normal functioning of hydrologic cycle. Under the undesirable condition, the amount of hydrologic variables began to diverge over time. Hydrologic variable should be homogeneous for the reliability of hydraulic structure while predicting necessary design criteria for its construction. Therefore, the test of whether this requirement is true should be performed in the context of any given hydrologic data’s homogeneity before being passed to the implementation of statistical approaches to the data. The study carried out in Yesilirmak basin was realized on homogeneity of seasonal maximum streamflow data from eight gauging stations operated by The General Directorate of State Hydraulic Works (DSI). Yesilirmak River basin area is approximately 5% of surface area of Turkey. Yesilirmak River is one of the major rivers of Turkey and its long is 519 kilometers. There are three main tributaries of the Yesilirmak River, named as Kelkit, Cekerek and Tersakan. Its water is mostly used for purposes as irrigation, drinking, fisheries and wildlife. The parametric and non-parametric procedures, called as standard normal homogeneity, Pettitt, Buishand range and von Neuman ratio were used for this reason. Statistically significant inhomogeneity with respect to the all of the statistic tests taken into account in the study was detected in the considered streamflow data sequences presented.

Standardized Precipitation Index (SPI) is used to determine dry and humid periods according to the cumulative probability method at different time scales. . In this study, the rainfall data between the years of 1980-2018 belonging to of Kayseri Meteorology Station was simulated by CLIGEN stochastic climatic data generator. SPI indices calculated by using observed and simulated precipitation were evaluated with the statistical methods at the time scales of 3-, 6-, 9- and 12- months. The SPI values of 3-, 6-, 9- and 12- month which are observed and simulated with CLIGEN are close to each other and the performance of the model is very high in calculating the SPI values of these time series. However, as the time period increased, the model's representative ability decreased.