ABSTRACT
This study aims to evaluate the satellite- and model-based precipitation products against a relatively dense station network of Iran Meteorological Organization (IRIMO) from January 2015 to June 2018 at daily timescale over Iran. The evaluations are carried out based on continuous and categorical statistical metrics. First, we employed daily satellite soil moisture (SM) observations from the Advanced Microwave Scanning Radiometer 2 (AMSR2) for estimating rainfall through the use of the relatively new SM2RAIN approach (SM2R-AMSR2). Next, we validated rainfall estimations from satellite/reanalysis products including version 05 daily Global Precipitation Measurement (GPM) Integrated Multi-Satellite Retrievals for GPM (IMERG) Early (IMERG-DE), Late (IMERG-DL), and Final Daily (IMERG-DF) Runs and the European Centre for Medium-Range Weather Forecasts’ ReAnalysis Interim (ERA-Interim), ECMWF ReAnalysis 5th Generation (ERA5) and SM2R-AMSR2. Based on the probability density function (PDF), all satellite and model products capture more (less) precipitation events than the reference data set for precipitation under (upper) 5.0 mm day−1 threshold. Regarding the mean absolute error (MAE) and root-mean-square error (RMSE) quantities, the ERA5 product has the lowest error relative to other products. Also, the threat score (TS) indicates that this product has the highest accuracy. ERA-Interim and IMERG-DF had roughly similar performances. The results reveal the overall superiority of ERA5 data over other products as it has the highest correlation with observation and the best discrimination between occurrence and non-occurrence of the precipitation events. Also, the results of spatial correlations show that ERAs have a higher agreement with in situ observation when compare to satellite products, over the north to northwest of Iran, where rain is formed mainly by synoptic-scale systems.
Acknowledgment
We are grateful to the scientists in NASA, JAXA, and ECMWF which provide IMERG, AMSR2, and ERA data, respectively, and make them accessible online. Special thanks to the IRIMO for providing the reference data set. We acknowledge the NCL team for their guidance in developing some of the NCL scripts used in this study. The authors are also grateful to Ms Naseem Golestani and Dr Mehdi Rastgoo for their contribution. We also appreciate the contributions of reviewers and the editorial staff of International Journal of Remote Sensing (IJRS), all of whom have contributed to the improvement of the manuscript.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. Kavir Desert.
2. Lut Desert.