The Relationship Between Climate Change And Epidemic Of Rubber Tree Powdery Mildew In Hainan Island

Global climate change is a fact; the occurrence of plant diseases has a close relationship with the climate. Using the disease index of powdery mildew of rubber in 1962-2003 in Hainan State Farms and the daily meteorological data in 1961-2003 in Haikou, Dongfang, Danzhou, Qiongzhong, Qionghai, Sanya and Lingshui, this paper firstly carries out fitting to the disease index adopting time series analysis method, analyzes the climate changing features in Hainan province, studies the relationship between disease index and meteorological factors, screens the meteorological factors that are relevant to disease index, and in the end builds the disease forecast model based on meteorological factors employing stepwise regression method. The major results are as follows:The powdery mildew disease indexes for rubber trees of Hainnan State Farm in 1962 to 2003 have been forecasted with the methods of exponential smoothing, autoregression analysis, moving average analysis, and autoregression moving average analysis, and the research results of these four methods have been compared as well. It has turned out that all these four methods can predict the rubber tree powdery mildew occurrence trend well, with the autoregression moving average analysis method having the best forecasting result.By studying the meteorological factors variation trend of Hainan between 1962-2003, it has been found that in the latest 40 years the annual average temperature, the average highest temperature and the average lowest temperature all tend to rise significantly by 0.03℃,0.021 ℃ and 0.036℃ respectively, while the annual sunshine duration tends to decrease sharply by 5.17 hours on average annually and the precipitation has not significant variation.The correlation coefficients between anomalies of meteorological factors and anomalies of disease indexes have been analyzed on the ten-day, monthly, quarterly and annual basis. Through related analyses, key factors and key durations affecting the rubber tree powdery mildew disease indexes have been screened. On this base, the disease indexes forecast models based on the ten-day, monthly, winter, spring and annual key metrological factors have been set up respectively with the stepwise regression analysis method. By comparing the precision of models, it has been found that the model based on the ten-day key meteorological factors has the best fitting effect, with the correlation coefficient of 0.7985 and the root-mean-square error of 8.4148, and that it are followed by the model based on the monthly meteorological factors,with the correlation coefficient of 0.7241 and the root-mean-square error of 9.7420.