Study On Regionalization For Water And Soil Conservation And Prediction Of Suspended Sediment Concentration In The Loess Hilly-Gully Region

Water is the fountainhead of life, soil is the essence of survival, water and soil resourcesis the basic material condition for human survival and development, and is the basic resourcesfor economic and social development. With economic and social development, the man-madesoil and water loss become more and more serious, and this problem is paid more and moreattention by the government and public, so the research on soil and water loss problem hasbeen a hot spot.This paper took the comprehensive management of soil and water loss as the main line,and on the basis of this line it conducted a study on the regionalization model for soil andwater conservation and the prediction model for suspended sediment concentration. Theregionalization for soil and water conservation is to construct a rational spatial layout with thelimited water and soil resources for comprehensive management and to maximize theproduction benefits of the limited resources, and this can better regulate the schemes of waterand soil loss management in the region; the most direct response of water and soil loss is thesuspended sediment concentration in the river, in order to explore the variation of suspendedsediment in the river, the suspended sediment prediction model is established to control themanagement degree of soil and water loss.Based on the reference to a large number of papers and materials, the present situationand existing main problem at home and abroad about regionalization for soil and waterconservation and suspended sediment prediction are comprehensively analyzed. Consideringthe external features in loess hilly-gullied region including natural environment, socialeconomic conditions, land utilization and the loss of water and soil resources etc,10regionalization indexes of reclamation rate, forest coverage rate, population density,evaporation and soil erosion area ratio etc are selected from the influential factors which consist agricultural production, land productivity, hydrological and soil erosion etc, andprincipal component analysis method is used to test the influence degree of each index onregionalization for water and soil conservation. Bayes discrimination analysis is used for thefirst time to conduct the regionalization for water and soil conservation, and is successfullyapplied to divide the districts in the loess hilly-gullied areas of Gansu, Ningxia and Qinghai.By comparing its results with the regionalization results in which the region is dividedaccording to the soil erosion types, it explained the rationality of regionalization by Bayesdiscrimination analysis and proposed measures and suggestions for Yongjing Countycomprehensive management. The research object is the data which was collected in1967~1997at Lanzhou hydrological station in Hilly-gullied areas of the Loess Plateau, and selectwater level, annual rainfall, annual runoff and sediment transport rate as independentvariables for suspended sediment concentration, take1967~1992data as original samples and1993~1997data as test samples, and use runs theory to test the interannual variationrandomness of the suspended sediment concentration. BP neural network algorithm is used toforecast the suspended sediment concentration, and particle swarm algorithm, geneticalgorithm and wavelet analysis theory are adopted to improve BP neural network. With thesefour algorithms, the suspended sediment concentration in1993~1997was predicted. Theworst prediction effect of these models is the unimproved BP neural network, the predictionaccuracy of the particle swarm optimization neural network and the genetic neural networkalgorithm is better and has wavelet neural network the highest prediction accuracy in the fouralgorithms. The regionalization for water and soil conservation and the prediction ofsuspended sediment concentration facilitates the management of water and soil loss inHilly-gullied areas of the Loess Plateau.