Interaction Analysis And High-resolution Projection Of The Meteorological And Hydrological System In Xiangxi River Basin Under Climate Changes

Hydrological forecasting is an important research content of hydrology research, plays an important supporting role for the sustainable utilization of water resources planning and management. Due to human activities and global climate warming effects, the current water resource changing profoundly,including variations and spatial-temporal evolvement. Meanwhile, the runoff restricted and affected by various factors that exhibit the characteristics of complex, random, multi-dimensional. Especially in recent years, with the China’s large water conservancy project put into operation, construction and operation, hydrological spatial-temporal evolvements are becoming more complex, thus to watershed hydrological modeling and prediction requires higher precised. Therefore, hydrological forecasting plays an important role in the effective operation of water resource planning and management or flood control and disaster reduction system. It is of important practical significance and application value to explore the model which can improve the prediction accuracy. In addition, since the industrial revolution, the rapid development of scientific and social economy has led to the increasing of human intervention in nature. Global climate change has become a common phenomenon in academia and the public. So far, it has been observed that climate change has a serious impact on the hydrological cycle of water environment. So in the context of climate change, it is very important to study the impact of climate change on the future water resources. So it meaningful to provide scientific basis for decision makers. Therefore, the goal of this paper is to apply the recursive clustering analysis regression tree ensemble forecasting model to the hydrological forecasting and statistical scale model to make high-resolution projection. Choosing reasonable forecast method is the precondition to make accurate forecast. The main research contents and innovative achievements of this paper include:(1) Firstly, introduced the basic theories, methods, classification and predictive model of the time series prediction, combined with the data of the Xiangxi river, and a runoff for the main factors affecting the results of sensitivity analysis is cited,using three kinds of methods to make a comparative study and then compared the simulation accuracy and forecast accuracy of the recursive clustering analysis regression tree ensemble model, multiple linear regression analysis model and neural network model. Through this comparative study, preliminarily determine regression tree ensemble model is the best choice.(2) Factorial Designs analysis of the relationship between parameter and dependent variable and the quantitative relation between parameters. This paper used regression tree ensemble model of Xiangxi river Watershed rainfall runoff simulation. Through the factorial designs and model combination, the set of the five parameter of the model was analysised by factorial designs, such as rainfall(P), evapotranspiration(E), High temperature(H), low temperature (L), the water level (f), and analyzes the various factors combination of diameter flow Q output effect. The results of this paper provide a template for the future use of the general model, and provide a theoretical support for the introduction of other hydrological models of the design method.(3) Using the SDSM statistical method to predict the changes of future precipitation and temperature in the basin. The results show that there are some differences in the precipitation, but it can be seen that the next 3 periods (2020s (2010-2039),2050s (2040-2069),2080s (2070-2099)) in the long term there will be an increase in the trend of the precipitation, the predicted results show that:the 2020s period of the temperature will be lower than the base year (1961-2000), the 2050s period is basically the same as the base year, the 2080s period will be higher than the base year, and there is a rising trend in the next 3 period.(4) Simulation in A2, B2 scenarios to getting the change of the runoff of the basin in the future. Then put the data of precipitation and temperature into recursive clustering analysis regression tree ensemble model to have basin runoff simulation in the next three hours.