| Meteorological drought is frequently happened in nature, which is mainly affected by rainfall. To explore the drought occurrence regularity and rainfall trend of the semiarid region of northwest Liaoning, meteorological data of 56 years from year 1956 to 2006 based on 21 observation sites in the region is selected, the rainfall space-time distribution characteristics of Chaoyang and other four cities are studied by using time series forecasting related theory and method, A series of models suitable for describing drought and rainfall prediction in northwest Liaoning are developed, which has important theoretical significance and practical application value for guiding of drought-resisting measures scientifically and rationally, reducing disaster losses to guarantee the food security and economic development of the northwest region of Liaoning.Main results of this research are as follows:(1) Single station data are computed based on Z index, precipitation anomaly index and humid index, It shows that Z index in seasonal drought statistics are in good consistent with the actual situation in the northwest Liaoning Province, the regional drought index base on Z index of single station shows that three is 21 years of drought, accounted for 50% of the total year, which is consistent with the practical situation. This indicates that the classification standard of regional drought proposed above is suitable for describing the characteristics of regional meteorological drought in northwest Liaoning.(2) SPI index is adopted to calculate the frequency, intensity and coverage area of the drought in northwest Liaoning. Analysis of space-time evolvement characteristics of results shows that Chaoyang and Jianping station in the southwest of northwest Liaoning have the highest drought frequency compared with other ones. Annual intensity and coverage area of the drought are in a stable fluctuation with little ascending and descending trend. While in the seasonal drought analysis, the drought trend is increasing in spring and summer, and the contrary in autumn and winter.(3) The experimental modal analysis(EMD) combining with threshold auto-regressive model(TAR) is proposed, it is applied in rainfall time series prediction of Fuxin City. The results show that the predicted precision through combination of IMF components composed of EMD decomposition and TAR model is higher than the individual TAR model does.(4) Rainfall forecast of the northwest region of Liaoning is applied by using chaotic time series prediction method, based on the analysis of the advantages and disadvantages of method used, the local linear segmentation method is proposed, it has a relatively lower predictive error in rainfall forecast in cities of northwest Liaoning except very few years.(5) Year of drought occurrence is forecast by using the optimized grey calamity prediction method, it has a good predictive agreement with the actual drought year. Multivariate grey prediction model MGM (1,3) is established to forecast rainfall in Chaoyang based on the grey relative analysis method, it has a high precision in both fitting and forecasting, and the future year of seasonal drought in Chaoyang are predicted by using the grey catastrophe model. |
PDF Full Text Request