The Study On Source Areas Of The Immigration And Forecasting Of Immigration Amount Of Brown Planthopper And White-backed Planthopper In Qijiang,Chongqing

The brown planthopper(Nilaparvata lugens (Stal)) and the white backed planthopper (Sogatella furcifera (Horvath)) which belong to Homoptera. Delphacidae, are migration insects with characteristics of globalization, migration, outburst,and devastating damage, etc. They threatened the safety of rice production seriously in recent years. Compared to the eastern region of China, the studies on rice planthoppers migratory and forecasting research are relatively weak in the southwest China. Chongqing is located in the southwest of China with complicated terrains. The valleys respectively formed by Dalou and Wuling mountains in south and southeast Chongqing provides a very favorable terrain conditions for the rice planthoppers to immigrate. Qijiang is an important passage for rice planthoppers to immigrate from Dalou mountains’valley. Given the complexity of Qijiang’s location, obvious contradiction of the development tendency of hybrid rice in China and weakness of the research in rice planthopper. it is significant for accurate warning system and comprehensive control of Chongqing or even the southwest China’s rice region to clarify the distribution of rice planthopper insect source in Qijiang and establish an effective forecasting model. In this essay, the HYSPLIT(Hybrid Single-Particle Lagrangian Integrated Trajectory), WRF(mesoscale weather forecasting model) and GRADS (Grid Analysis and Display System) were used to analysis the insect source and the meteorological background of the migratotry peak days. The Markov model. stepwise regression method, and the BP artificial neural network model were utilized as well to do a forecasting research on WBPH and BPH. The main results were as follows:1 The insect sources of BPH were mainly distributed at the junction of Yunnan, Guizhou and Guangxi as well the southern Yunnan and southwestern Guizhou.Then it was mainly in the south of Yunnan in May. the southwestern parts of Guangxi in June and the south of Yunnan. the southwest of Guizhou and the junction of Guizhou,Yunan and Guangxi in July. Wind shear and rainfall were the main meteorological factors to force the BPH to land in Qijiang.2 The WBPH insect sources of Qijiang mainly came from the juction zone among northwest of Guangxi, southwest of Guizhou and Eastern Yunnan in the first hand, the central Yunnan and the southeast of Guizhou in the second hand.. During high immigration years, the WBPH insects were mainly distributed at the junction of Yunnan, Guizhou and Guangxiin as well as the central part of Yunnan and the west of Guizhou. During shoulder immigration years, the insects were mainly distributed at the junction of Yunnan. Guizhou and Guangxi as well as few regions of southwest Guizhou. During slack immigration years, the insect sources were mainly at the junction of Yunnan, Guizhou and Guangxi and northwest areas of Guizhou. The range of insect sources distribution was the widest in the high immigration years, and the second was the shoulder immigration years, at the same time it was relative concentrated in slack immigration years.3 Using the WRF model to analysis the meteorological background during WBPH and BPH’s common immigration peak days in Qijiang from 23th to 25th July 2007, it proved the mass migration was the result of the collective effect of many factors such as the wind direction shear, the vertical movement of the atmosphere, and the rainfall. It didn’t closely related to the sinking of airflow.4 Markov models were respectively constructed to predict the occurrence tendency of the WBPH and BPH light-trap catches in the main immigrating period. The historical accordance were 100% and 96.15% and the predictive accuracy were 75% and 87.5% respectively.5 WBPH immigration amount predictive model was constructed based on the forecasting mothods of stepwise regression and BP artificial neural network (the 74 atmospheric circulation indexes as predictors). The stepwise regression models could make predictions 2 to 8 month ahead, in which the historical returns and the rate for prediction accuracy of the 9-factors predictive model could both reach 100%. The 10-factors BP artificial neural network model could make prediction a month ahead. The historical returns and predictive accuracy rate were both 100%. which narrow the wrong relatively compared to the stepwise regression model.