| Precipitation is a key variable of the global energy and water cycle and it is a primary driver of hydrological models. The accuracy of precipitation data is essential for flood and drought monitoring and forecasting, as well as water resources management. Based on daily precipitation data of 55 meteorological stations over Circum-Bohai-Sea region during 1998~2014, the suitability are evaluated and compared among 3 satellite-based precipitation products of TRMM 3B42 V7, CMORPH and PERSIANN. And then, their errors are explored for different seasons. Results show that generally, TRMM 3B42 V7 has the highest performance and the lowest is for PERSIANN. However, CMORPH has the highest detection probability of rainfall occurrence. The rainfall is slightly overestimated by 3B42 V7 but underestimated by CMORPH and PERSIANN. Seasonally, the skill scores in summer is higher than in other seasons and in winter, the capturing capability is very lower. The performance of 3 satellite-based precipitation products increases from the daily to monthly scale and larger errors exists in the daily rainfall estimates.As the important component of atmospheric circulation and hydrological cycle, the estimate accuracy of precipitation is crucial in hydrology, meteorology and climate. In this study, wavelet multi-resolution analysis is applied to improve TRMM 3B42 daily rainfall estimation at 55 meteorological station over the Circum-Bohai-Sea Region during 1998~2014. The results show: 1) After correction, the correlation coefficient between 3B42 and in-situ data is remarkly increased and the MAE and RMSE is decreased significantly. 2) The probability of detection and the false alarm ratio of daily precipitation event by 3B42 data are improved through wavelet-based correction. 3) Wavelet multi-resolution analysis effectively enhances the precipitation estimation accuracy of each season, especially in winter. 4) However, 3B42 data after the wavelet-based correction still overestimate daily rainfall. |
PDF Full Text Request