The Stepwise Interpolation Approach Of Precipitation Spatial Analysis In The Upper-Middle Reaches Of The Yangtze River

Using successive regression and GIS technique, the regression equations of the precipitation and the topographical data in the upper-middle reaches of Yangtze River are established based on the 10 years daily, mensal and annual precipitation of 613 meteorological stations, from 1992 to 2001. The topographical data include the elevation, aspect and slope form DEM of 4km resolution. The effect of topography on precipitation can be extract from the precipitation.Secondly, considering the incontinuity of the precipitation distributing, we use improved the Stepwise Interpolation Approach, SIA, to combine with the multi-variable successive regression equations. SIA obtains the result accordant with fact, and considerably improves the accuracy of the spatial interpolation. We gain the distributing instance as a whole by the spatial analyse of the daily, mensal and annual precipitation.Then, we adopt mean error(ME), mean absolute error(MAE), mean relative error(MRE), and the root mean squared interpolation error(RMSIE) and the correlative coefficient(CC) as the determinant standard of interpolation effect. The results show that the mensal precipitation effect of SIA is the best, and the daily precipitation effect is worse reversely.And then, we take the interpolation result of January 2000 for example, which has a good interpolation effect, to analyse the effect of the SIA error. We also adopt mean error, mean absolute error, mean relative error, and the root mean squared interpolation error and the correlative coefficient as the determinant standard under different interpolation steps. Whereafter, we can conclude that the SIA calculation with 5 steps considerably improve the interpolation accuracy, the SIA calculation with 10 steps improve the interpolation accuracy ulteriorly, the SIA iterative results after 50 steps identically converge at the observed precipitation, and the interpolation effect of January is better..Also, by contrasting the daily, mensal and annual precipitation interpolation errorrespectively between SIA and the other methods, we know that the interpolation effect of SIA is better than others, and its accuracy is preferable in the mass.Fllowing our results, we can induce that the SIA embodies the precipitation changes adapt to the topographty and the incontinuity of the precipitation distributing. It can also improve the interpolation accuracy largely. The more steps we interpolate, the better effect we get. Besides, the resolution of 4km can not only image the spatial asymmetry of the precipitation by topographical influence, but also reduce the quantity of calculation. In order to ensure the preferable precision and the simple calculation, we use the resolution of 4km and the SIA calculation with 5 steps.