Research On Water Resources Allocation Of Heihe River Basin Based On Runoff Prediction

Heihei River is the second largest inland river in China,located in the northwest arid and semi-arid region.Many big and medium reservoirs were built in this basin,Since the 1960s,people lived in the middle of Heihe River took a lot of surface water from channel and excessively withdrew groundwater with the rapid development of society and economic,causing many environmental problems such as the disappearance of lakes and the ecological degradation in the downstream.This situation severely influences the sustainable development of society and economic and the stabilization and health of ecological environment.Therefore,Combined with Heihe River Plan of the National Natural Science Foundation of China,studying water resources allocation problem of Heihe River basin not only provides scientific decision-making basis for alleviating water resources contradiction in the basin,but also provides reference for water allocation practices in other river basins,which has significant theoretical significance and applicable value.Based on the data and materials between 1957 and 2014,the climate change and runoff variation of Heihe River basin were analyzed.The mid-long term runoff prediction methods were studied subsequently.Then,the scheme set and evaluation index system were built for water resources allocation,respectively.Moreover,the groundwater balance model and the joint allocation model of surface water and groundwater were established and combined together.The optimized problems of water resources allocation were solved through the shared particle swarm optimization algorithm.The water resources allocation schemes and measures were calculated and evaluated.In addition,the water allocation laws were analyzed under the optimized schemes,and the influences of hydrological regime on water resources allocation were also discussed.Finally,the water resources allocation rules were established and tested.The main conclusions of this dissertation were drawn as follow.(1)The precipitation significantly increased at the upstream and midstream of Heihe River,but did not show a significant trend at the downstream.The annual average air temperature obviously grew in the whole basin,and the middle 1990s was a significant turning point.The annual water surface evaporation had a marked increasing tendency at the upstream,but had no apparent trends at the midstream and downstream.Sunspot number and global carbon dioxide concentration had some relationships with the climate change of Heihe River basin.(2)Based on moving window and correlation analysis,this study presented the moving window correlation analysis method,which could effectively eliminate noise interference and accurately captured periodic components of hydrologic time series.Besides,the slope-intercept diagnosis method was also put forward for effectively analyzing variation components of time series based on the slopes and intercepts of piecewise trend lines.(3)The annual runoff series of the upstream of Heihe River had significant 22-year and 6-year periods,while the annual runoff series of Liyuanhe River,an important tributary of Heihe River,had significant 17-year and 19-year periods.Subject to the influences of increasing precipitation and growing air temperature,the annual runoff of the upstream of Heihe River had a remarkable increase trend since the late 1990s.The average ten-day flow and water depth had good logarithmic function relationships at the sections of midstream of Heihe River,and the average ten-day flows had good linear function relationships at two adjacent sections of the downstream of Heihe River.(4)Among the multivariable linear regression model,the artificial neural network,the support vector machine,the gray prediction model and the deterministic components superposition method,the last method was recommended for forecasting the annual runoff of Heihe River basin.In addition,the support vector machine was suitable for predicting the average ten-day flow when dividing wet and dry periods.(5)According to the scenarios of 2020s and 2030s,the water resources allocation scheme set of Heihe River basin was set from the perspective of water supply and water demand aspects.The water resources allocation evaluation index system was also built under the framework of evaluation criteria.Based on the generalized water resources allocation system of Heihe River basin,and the midstream groundwater balance model and the joint allocation model of surface water groundwater were established.Moreover,the shared particle swarm algorithm was realized at a multi-core workstation,which could effectively solved optimization problems of water resources allocation.(6)Based on the non-negative matrix factorization principle,the non-negative matrix factorization method was proposed for evaluating water resources allocation schemes,which was relatively objective and reliable for decision-making.According to the equilibrium optimization principle of system,an equilibrium optimization analysis method was proposed for effectively evaluating equilibrium optimization degrees of water resources allocation schemes.The best water resources allocation schemes of 2020s and 2030s were number 5 and number 11,respectively.Moreover,the measure combining water supply discount with reservoir regulation was suitable for the water resources allocation in Heihe River basin.(7)Under the recommended schemes of different planning years,the surface water withdrawal and the surface water income had good quadric function relationships at the midstream.The relation curves between annual runoff of Zhengyixia gorge and Yingluoxia gorge were above the curve of '97' water distribution scheme.Considering the irrigation,ecology and hydropower generation,this study established reasonable water resources allocation rules based on mid-long term runoff prediction,including the operation functions of Huangzangsi reservoir.