The Analysis Of The Characteristics And Prediction Of The Drought In Leizhou Peninsula

With the status of water resource becoming more and more important, the problems of drought and lack of water are concered by more and more people. The author finds out the problem of drought in Leizhou Peninsula is very serious, so the author does some explorative study on predicting the drought in Leizhou Peninsula, and the dissertation includes eight chapters as follws:Chapter 1: the Introduction. This part is to elicit the task of the study, is to introduct some progress on study of drought, and to introduce the primary content of the study at the same time.Chapter 2: the analysis of the character of hydrology in Leizhou Peninsula. This part is to describe the drought in Leizhou Peninsula in. history with table. From the table, we can see the main characteristics of the drought in Leizhou Peninsula, and we analyze the main influencing factors of the drought.Chapter 3: the analysis of the change-point of the hydrological data. This part is to explain the data, which is used to predict the drought in Leizhou Peninsula, and to do responding statistical analysis of the data. We use hydrological aberrance diagnosis method presented by â…ªA Jun, etc. and the random adjacent classification method presented by HU Yuan-lai, etc. to classify the rainfall of plentiful and low water period in Leizhou Peninsula, which result is that plentiful water period is from May to September and low water period is from October to April next year. As well we establish the model of the collection between aveage year-rainfall and the ten rainfall stations based on changeable structure based genetic algorithm.Chapter 4: to change the original rainfall data into plentiful and low water period rainfall data, and to establish auto-regrssive model. In this chapter, to establish threshold auto-regressive model (TAR) to predicting rainfall of plentiful and low water period.Chapter 5: to establish bilinear time series model (BM) to predicting rainfall of plentiful and low water period.Chapter 6: to establish the changeable structure based genetic algorithm auto-regressive model (CSGA-AR) model to predicting rainfall of plentiful and low water period.Chapter 7: to establish the combined forecasting model using the data from chapter 4 to chapter 6. Those models used above include many parameters and coefficients, the optimization of these parameters and coefficients adopted real coding based genetic algorithm presented by JIN Ju-liang, etc.Chapter 8: summary and expectation of the study. From the approachable result and the predicting result of 1999~2001 of the established model, we can see the predicting data posses some consultative value in practical use. At last, we put forward some suggestions for futher study on drought problem.