Evaluating the capabilities of Logistic Model Tree in predicting the occurrence probability of daily precipitation

Due to the location of Iran in arid and semi-arid regions and the inhomogeneous distribution of precipitation, predicting the occurrence of precipitation is important, therefore, researchers are implementing novel methods to identify and predict this parameter accurately. Thus the purpose of the current study is to investigate the capabilities of Logistic Model Tree (LMT) in predicting the occurrence of daily precipitation at Parsabad station using 1 to 3-day meteorological data. For this purpose, meteorological data for 2004-2016 were collected, and three combined scenarios of meteorological parameters were considered for calibration and validation of the studied method. The results showed that the prediction accuracy of the best-case scenario using the data from 2 days ago was about 79%, however, with the data from 1 and 3 days ago, the daily precipitation was with 80% prediction accuracy. Finally, by investigating the evaluation criteria, scenario 1 with the input parameters of minimum, maximum and average relative humidity (%), temperature (oC), total sunshine hours, and wind speed (m/s) was determined as the most accurate scenario to predict daily precipitation.