Evaluation of Daily, Decade and Monthly Data Satellite Images to Estimate of Precipitation Using Google Earth engine in Khuzestan Province

Since synoptic metrological stations have non-uniformed scattering pattern in Iran and on the other hand precipitation determination and forecasting is essential for irrigation planning, a method precisely determine precipitation of agricultural lands in farm level has great importance. This study was carried out in Google Earth Engine Code programming environment using GPM, TRMM and CHIRPS satellite data which is daily, decade and monthly, respectively in Ahwaz and Izeh metrological stations for calibration and 9 meteorological stations for validation during 2015-2016 and 2017 - 2018 Growing season. Results showed that monthly interval could obtain better accuracy with R2 of 0.99 and NRMSE = 0.36, respectively. The validation results of the rest 9 meteoroidal station indicated that precipitation prediction had 51% and 3.1 mm error and under estimation on average, respectively. The efficiency was reasonable and F-Test showed no significant difference between observed and prediction samples. The standard error value was 14.2 mm which is a significant error and need to work on updated better functions. It can be concluded that this method can be a useful tool for monthly precipitation prediction of areas with no climatic data if integrated with Kriging, co-Kriging and Inverse Distance Weighted (IDW) geostatistical models for interpolation.