The interpretation and credibility analysis of the precipitation data from the shared forecasts database

The precipitation is a very important input in the traditional hydrological model and at the same time caused the uncertainty of the land surface model. Due to the heterogeneity existed in time and space, plus the simplexes of traditional methods and the insufficient of precipitation measurement technology, the observed precipitation data restricts the applications of the distributed model in simulation and forecasting to a certain extent. With the recent development of measurements and network technology, the meteorological forecasting has become a friendly-accessed, real-time system which can obtain an extended coverage of the land resources. However, the credibility, accuracy and practicability of the interpreted precipitation data needs to be further observed. Through the image processing, the resources were interpreted into the precipitation data which were compared with the measured rainfall values. Then based on the behavior of the hydrological-forecast resources, the method of credence test was selected, and a reliability analysis was conducted from the aspects of point and side. This paper proves that the method of obtaining point precipitation through the network access has a comparatively low precision, but the simulation of precipitation with measured data in temporal and spatial distribution is well established, which can be used as a supplement to the measured precipitation. So the hydrological-forecast resources mentioned above can be applied to the traditional hydrological model, which will greatly impact the study of precipitation in the temporal and spatial distribution, and improves the precision of the hydrological-forecasting.