| Under the influence of human activities and natural factors, the global climate anomaly is likely to persist, resulting in a significant impact and losses to the sustainable development of social economy and people’s life and property. Especially in recent years, a variety of flood disasters caused by extreme precipitation caused a lot of impact on people’s production and living. With the development of satellite remote sensing data, they make up for the areas lacking of precipitation or the shortage of rainfall information, providing a good opportunity for hydrological modeling and forecasting in ungauged basins. And at the same time, with the development of remote sensing precipitation products increasingly rich, the monitoring capability of the extreme rainfall events has become an important topic. Given that, to provide a reference for the effective use of these two kinds of data and lay a solid foundation for more accurately using the satellite data on hydrological modelling and forecasting, this study used China ground meteorological evaluation data to test of two kinds of precipitation products (TRMM (Tropical Rainfall Measuring Mission) and the FY-2D) widely applied both at home and abroad, and evaluated these two kinds of precipitation products in two typical regions of North and South China (Xiangjiang basin and Heihe basin), further testing the monitoring ability in the extreme precipitation events. The major research findings are presented as follows:(1)The study area was in Xiangjiang basin, Heihe basin and its surrounding, using fuzzy comprehensive evaluation method, the correlation coefficient method and the scattering points gradient method, and focus on the meteorological stations precipitation data, TRMM and FY-2D precipitation data to test these two kinds of remote sensing data in daily, monthly scales. On the whole, the two kinds of data with high precision showed a good applicability in the study areas. The quality of data in monthly scale is better than that in daily. The correlation coefficient of monthly precipitation data reaches 0.85 or more, while the daily correlation coefficient value is low, and the average is about 0.4 or lower. Comparing the application of these two kinds of remote sensing data, they application in the Xiangjiang river basin is better. The changing trends of MRE, Bias, K in all meteorological sites are the same, only a few sites having higher MRE, Bias and correlation coefficient at the same time.(2)Daily data were tested by using the fuzzy evaluation method. In the Xiangjiang River Basin the forecasting accuracy of TRMM rainfall data is in medium level, and the accuracy of rainfall from good to bad, followed by, light rain, heavy rain, rain, moderate rain. The Heihe basin located in arid area is typically arid and few rain area, and yearly precipitation is far less than the years of evaporation. During the study period, the Heihe basin is hardly heavy rainy time, and most of the time it’s not raining. Therefore, the best accuracy of TRMM data in the basin is light rain level.(3)The data of elevation, slope and other statistical indicators in the study area were used to do multiple regression analysis to get the influence on the precision of data. The evaluation of the effects of altitude were on the level of 0.42 or more in Xiangjiang basin; the worst accuracy of data were located in the area whose elevation was about 2500km; the effect of slope gradient on the accuracy of TRMM data showed that the increase of slope, the change trend of the precision was getting worse. The evaluation of the effects of altitude were on the level of 0.74 or more in Heihe basin; the worst accuracy of data were located in the area whose elevation was about 1500km; the effect of slope gradient on the accuracy of TRMM data was more complicated that it still had the same change with Xiangjiang basin.(4)Taking the heavy rain events in July 21,2012 in Beijing area as an example, the whole space evolution process of that torrential rain event was monitored and analyzed by the remote sensing. The results showed that the correlation coefficients between these two kinds of satellite remote sensing data and the observed precipitation data were more than 0.91, the error was less than 10%, and these two kinds of data were consistent with the observed rainfall data. But the TRMM and the FY-2D precipitation data were underestimated by 14.42% and 19.86%, the FY-2D was more conservative than the TRMM data. For monitoring the local scale precipitation, TRMM precipitation data and FY-2D had a larger intervals than the data of ground stations, but they also had a good application prospect of monitoring in the national and regional scope.(5) Taking the typhoon "Haikui" in August,2012 as an example, the whole space evolution process of that heavy rainfall caused by typhoon was monitored and analyzed by using the TRMM, the FY-2D and the data from the ground sites. The results showed that the above two kinds of remote sensing precipitation products can seize the position of heavy rainfall accurately, and can be used to distinguish the spiral structure of typhoon, but for the maximum precipitation, especially the TRMM data still underestimated the precipitation. Comparing these two kinds of remote sensing rainfall products, FY-2D were more similar to the data of sites and had better accuracy than TRMM, duing to the addition of the ground observation data. Comparing to the three, the variation of average precipitation along lime series affected by typhoon in the rainfall center area were similar. These two satellite precipitation products Exhibit a phenomenon that the rainfall was underestimated in high strength, overestimated in small strength. Thus, there were still many problems in the satellite precipitation data for the monitoring of extreme precipitation events. Because on the one hand satellite retrieval algorithm reflecting the precipitation mechanism for strong convection rainstorm was not exactly accurate, on the other hand this rare and severe weather phenomenon was accidental, instantaneous changes status and unpredictable. |
PDF Full Text Request