| In recent years,due to the impact of climate change and human activities,the precipitation characteristics of spatial and temporal distribution in the Dalihe River Basin have changed greatly.The rain gage network is the most direct source of precipitation data.Reasonable network density and spatial distribution of rain gages are very important,reflecting the spatial and temporal variation of precipitation in the basin.Pumping station,typical rainstorm analysis method and so on are the most common domestic methods in rain gage network installing,and these methods have single objective function and show no connection between gages.In addition,rain gage network distribution in the Yellow River basin presents a low overall density,while the density is high in part of the basin.Dali river basin in the loess plateau was studied as the research object,basing on the information entropy theory.The maximal information minimum redundancy criterion(MIMR)was used to analyze the relationship and the law of information transmission between the rain gages,as well as the variation characteristics of precipitation at different time scales and different spatial scales.The research results can provide a theoretical basis for optimizing the distribution of rain gage network in different spatial scales of the Loess Plateau.The main work and conclusions are as follows:(1)The temporal and spatial variation characteristics of precipitation is analyzed,as well as the relationship between precipitation information entropy and precipitation in Chabagou River Basin and Dalihe River Basin.Based on the data of 13 rain gages in the Chabagou River Basin and 29 rain gages in the Dalihe River Basin,the spatial and temporal analysis of interannual precipitation was carried out,and the precipitation information entropy was calculated by resampling the precipitation sequence and discretization processing.Then the relationship between precipitation information entropy and precipitation was studied.The results show that the annual precipitation in Chabagou River Basin does not change significantly,and the precipitation concentration ratio is 0.89.The spatial distribution of precipitation is generally higher in the southwest while lower in the north.However,the annual rainfall in the Dalihe River Basin does not change significantly,and the precipitation concentration ratio is 0.88.The precipitation distribution tends to be higher in the east and west of basin,and lower in the middle.After resampling the precipitation series,the precipitation information entropy of the same station increases with the increase of time scale and space scale.(2)Based on the variation law of the joint entropy threshold of rainfall stations at different time scales in the Chabagou and Dali river basins,an optimization plan for the rainfall station network in the watershed at different time scales is proposed.The results show that the changes of joint entropy,mutual information,and redundant information of rainfall series at different time scales are generally the same.As the number of iterations increases,joint information entropy tends to be stable,mutual information first increases and then decreases,and redundant information becomes more and more;In the selection of the threshold,MIMR will capture the sites with heavier rainfall as the central site.The larger the time scale,the smaller the number of sites required to reach the threshold.The number of optimal rainfall stations on the daily,weekly,and 15-day scales in the Chabagou watershed are 10,7,and 4 respectively,and the corresponding rainfall station network density is 20.4km~2 per gage,29.14km~2 per gage,51km~2 per gage.The number of optimal rainfall stations on the daily,weekly,and 15-day scales in the Dali River Basin is 23,10,and 4,respectively,and the corresponding rainfall station network density is 170km~2 per gage,391km~2 per gage,and 977.5km~2 per gage.(3)Analyzed the impact of different spatial scales of the watershed on the rainfall station network.Based on this,the optimization plan of the rainfall station network of the watershed under different spatial scales was evaluated and the performance test of the optimized station network is carried out by using heavy rain,weekly and 15-day accumulated rainfall.The conclusion shows that after the spatial scale changes,Chabagou Basin is a sub-basin of the Dali River Basin,and the information carried by rainfall stations outside the sub-basin will affect the layout of the optimal rainfall stations in the sub-basin.The area rainfall of the optimized station network at the daily,weekly,and 15-day scales in the Chabagou watershed has a strong correlation with the area rainfall of the original station network(r=0.998,0.987,0.975),and the rainfall deviation(RBS=0.8%,0.1%,-0.3%)is small and has high accuracy(NS=0.999,0.999,0.998).The three optimized station networks all meet the requirement of rainfall station network density(200 km~2per gage).The area rainfall of the optimized daily and weekly station network in the Dali River Basin has a strong correlation with the area rainfall of the original station network(r=0.999,0.992),the rainfall deviation(RBS=0.1%,0.8%)is small,and the accuracy is high(NS=0.999,0.999),the 15-day-scale optimized station network will overestimate the basin area rainfall(RBS=4%),and only the daily-scale optimized station network meets the requirements of station network density.The Chabagou watershed can use a daily-scale optimized station network as a new rainfall station network,which can provide 98.5%of rainfall information in the basin,and the station network density is 20.4km~2 per gage;the Dali River Basin can use a daily-scale optimized station network as the new rainfall The station network can provide 98.2%rainfall information in the basin,and the rainfall station density is 170km~2 per gage.The daily-scale optimized station network in the Chabagou watershed and the Dali River Basin can capture the rainstorm centers in their respective basins and reflect the spatial distribution of rainfall in the original station network.The new rainfall station network composed of the Chabagou and Dali river basins has maximizes overall information,maximizes information transfer capacity and minimizes redundant information and can reflect the optimized rainfall distribution of the previous station network.The research results can provide a theoretical basis for the optimal layout of rainfall station networks in different spatial scale basins in the future. |
PDF Full Text Request