Prediction of precipitation in Golestan dam watershed using climate signals

Abstract

Global and regional scale climate teleconnection signals, including sea level pressure (SLP) and sea surface temperature (SST), are the main factors influencing the earth’s climate oscillations and are among the most important indices used to predict climatic variables. In this research, the effect of teleconnection signals on monthly maximum 1-day precipitation is examined using artificial neural network (ANN) and 40 years of rainfall data for the Madarsoo watershed located at the upstream of Golestan dam in Northern Iran. The Pearson correlation coefficient was used to determine the correlation between monthly maximum 1-day precipitation and climate signals with different lags. Different ANN models with various combinations of inputs, i.e., correlated SLP and SST with different lags, were then used for forecasting precipitation. Results revealed acceptable performance of ANN in forecasting monthly maximum 1-day precipitation using SST/SLP datasets. For instance, the performance indices including root mean square error (RMSE), correlation (R), and Nash-Sutcliffe (CNS) coefficients for monthly maximum 1-day precipitation of Tangrah rain gauge in August were found to be 6.12, 0.95, and 0.945 mm, respectively, for the test period.