The Climate Response Of Rare And Endangered Plant Tsuga Longibracteata Growth

The climate response of rare and endangered plant Tsuga longibracteata growth were studied by tree-ring width index in Tianbaoyan nature reserve of Fujian. The stepwise regression analysis, regression after extracting the principal components and multidimensional time series analysis were used to study the quantity relation between Tsuga longibracteata growth and climatic factors. The key climatic factors that effect tree growth were selected from many climatic factors to establish forecast model ;By analyzing the relationship between the ENSO event and tree growth in recent years, the effects of global climate abnormality on the growth of Tsuga longibracteata was discussed; The possible effects of sun-spot on the growth of Tsuga longibracteata were also studied, and the growth of Tsuga longibracteata by sun-pot and auto-correlated factors were predicted ;In order to study the nonlinear relationship between the sun-pot and Tsuga longibracteata, the artificial neural network was used to establish prediction model, which would provide a analysis method for sun-spot and tree growth. The results indicated as follows:(1) In 18 climate factors , based on stepwise regression analysis and regression after extracting the principal components ,three climate factor were select that effect breast diameter growth of Tsuga longibracteata.They were average ground temperature of year, average value of extreme and the minimum relative humidity of 5, 6, 9 and 10 month, annual temperature range. Among them 15cm average ground temperature of year and annual temperature range had negative effects on breast diameter growth of Tsuga longibracteata, but the average value of extreme and the minimum relative humidity of 5,6,9 and 10 month had positive effects on breast diameter growth of Tsuga longibracteata.Three main climatic factor were used as multidimensional time series model control factor to establish CAR model of breast diameter growth of Tsuga longibracteata.The result showed that there were also relationship to 15cm average ground temperature one year ago and the average value of extreme and the minimum relative humidity of 5,6,9 and 10 month one year ago.(2) The effects of global climate abnormality on the growth of Tsuga longibracteata was discussed by analyzing the data of ENSO event and breast diameter growth of Tsuga longibracteata.The study showed that EI Nino/anti-EI Nino events perhaps had no impact on the breast diameter growth of Tsuga longibracteata in current year. A linear equation was established with Southern Oscillation index and auto-correlated factors of Tsuga longibracteata- ring width index, it was found that the breast diameter growth of Tsuga longibracteata natural forests were significant positive correlated with its auto-correlated factors. The partial correlation coefficient of Southern Oscillation index had not reach (α= 0.10) the significant test. It showed that Southern Oscillation had little effect on the breast diameter growth of Tsuga longibracteata natural forests.(3) The effects of solar radiation and solar activities on global environment and biology were direct and complicated. In this paper, wavelet transform is used. It showed that there were negative correlation between sunspot and the tree-ring width index of Tsuga longibracteata.This negative correlation signed that sunspot had inhibitory effect on the growth of Tsuga longibracteata. Stepwise regression analysis was used to study the quantitative relationship between them. The results showed that the breast diameter growth of Tsuga longibracteata was affected by its ring index of one ,two and eleven years ago and the sunspot of one and six years ago. Among them, ring index of one and two years ago and the sunspot of one and six years ago had positive effects on breast diameter growth of Tsuga longibracteata except the ring index of eleven years ago. The prediction model of breast diameter growth about Tsuga longibracteata was established by stepwise regression analysis with the sunspot and Auto-correlated factors as forecast factors. There was a high simulating precision that reached 90.04%, which improved 5.65 percent point than climatic factors.(4) By stepwise regression analysis, 5 factors that affected relative sunspot number were selected. They were Tsuga longibracteata ring index of two and eleven years ago and the sunspot of one, four and ten years ago. The artificial neural network of sunspot was established with 5 factors. The simulating regression precision of model was 88%, and the predicting precision of model was 86.23%. Based on the artificial neural network model, the sensitivity analysis on each variable of model showed that the effect of sunspot of four years ago on sunspot is even, and the other factors effects of on sunspot were sensitive, which showed that five variables all had significant effects on sunspot.(5) In this paper, Wave-type time series was used to simulated the breast diameter growth of Tsuga longibracteata natural forests. The results showed the simulating regression precision of model was 98.83%, which was the best of this paper. The simulating regression precision of wave-type time series was better than the stepwise regression analysis with climatic factors (simulating regression precision was 84.39%) and than the multidimensional time series model (simulating regression precision was 89.77%) and the forecast model with sunspot and auto-correlated factors (simulating regression precision was 90.04%).