| چکیده انگلیسی مقاله |
Introduction: Precipitation, as one of the main components of the water balance, plays an important role in the spatial and temporal distribution of available water and is the most important factor that directly interferes with the hydrological cycle. Due to the lack of up-to-date and long-term precipitation data with appropriate accuracy and the high spatially and temporally variability of this quantity, it is very difficult to monitor it over large areas; also, the cost of establishing a precipitation measuring station, the shortage of stations, the lack of installation of recording devices in undulating areas, point-based measurements and the inability to generalize measurements over large areas, as well as the lack of the desired ability to record torrential and heavy convective rainfall, have always faced researchers in the fields of atmospheric and hydrology with challenges in measuring precipitation; therefore, the use of satellite products is a suitable alternative to obtain precipitation data, especially in areas without statistics and areas with a low density of ground stations. However, satellite products also have numerous errors; For this reason, it is essential to evaluate and verify the accuracy of these products before use. Therefore, in the present study, the precipitation products of the three satellites including CHIRPS, MERRA, and TRMM were evaluated on a monthly scale in the country of Iran.
Materials and Methods: In this study, data from 222 synoptic stations in Iran were received from the National Meteorological Organization on a monthly time scale from January 2005 to December 2019, and precipitation products from CHIRPS, TRMM, and MERRA satellites were downloaded from the NASA website and converted into uniform numerical data after data format was standardized; then, satellite data and data from ground-based synoptic stations were integrated together, and finally, the estimated and observed data were validated to obtain the satellite forecast error rate using statistical indices including Bias, MAE, RMSE, R, and R2, and the accuracy and success rate of the sensors were verified using conformity indices including POD, FAR, and CSI.
Results and Discussion: According to the results, TRMM has shown good performance compared to other satellites with RMSE = 23.84 and R2 = 0.69. Other indicators also indicate the superiority of this satellite compared to other satellites. MERRA satellite with RMSE = 30.57 and R2 = 0.56 has shown poor performance compared to TRMM and stronger performance compared to CHIRPS and is in second place in this respect. CHRIPS satellite also shows poorer performance compared to the other two satellites in almost all indicators. According to this table and the resulting Bias value, all three satellites have underestimated the rainfall compared to the actual value; however, TRMM satellite has less underestimation compared to the other two satellites and has performed better than the other two satellites in this indicator.
Conclusion: The accuracy of each station showed that the data of all three satellites, according to the POD index, have a low and close to zero variation range, and according to the FAR and CSI indices, this difference is around 0.5; so that the largest of them is related to the MERRA satellite products with a variation range of 0.148, which shows that according to these indices, the data of these satellites have a high reliability. Based on the FAR and CSI results, it can be seen that, although, in all stations, the MERRA products had the lowest error and mistake rate in detecting non-rainy days and rainy days with a very slight difference, but, based on the results of this study, it can be said that overall, the TRMM satellite products have appropriate accuracy, detection, and desirable consistency in all assessments. |