Impact Of Climate Change On Vegetation In The Northwest China

We used Atmosphere-Ocean General Circulation Models(GCMs) and Generalized Linear Model(GLM) to predict Normalized Difference Vegetation Index(NDVI) in Northwest China in 2085. The variables used were precipitation and temperature. NDVI data were collected from MODerate-resolution Imaging Spectroradiometer(MODIS). The years in which analyses were performed were 2001, 2004, 2007, 2010 and 2013. Temperature and precipitation data were mainly collected from GCMs predictions. Using MAGGIC/SCENGEN 5.3 software, we ranked 20 CMIP3 GCMs and selected the best 2 or 3 GCMs and the best 14 GCMs. NDVI data were upscaled to 2.5° × 2.5° grid. The main results were as follows.(1) The main characters of the climate in Northwest China in the future were that the increases of temperature were large and the changes of precipitation were small. For Shaanxi Province, the ranges of the 14 GCMs predicted precipitation changes were relatively larger and the ranges of predicted temperature changes were relatively smaller compared with other parts of Northwest China.(2) The regionally averaged precipitation changes predicted by 14 GCMs under A1 FI scenario in 2085 were that, in spring, the minimum is-0.49 mm·day-1, the median was 0.41 mm·day-1 and the maximum was 1.26 mm·day-1; in summer, the minimum was-1.04 mm·day-1, the median was 0.12 mm·day-1, and the maximum was 1.52 mm·day-1; in autumn, the minimum was-0.52 mm·day-1, the median was 0.24 mm·day-1, and the maximum was 0.98 mm·day-1; in winter, the minimum was-0.15 mm·day-1, the median was 0.20 mm·day-1, and the maximum was 0.78 mm·day-1. The mean temperature changes of the 14 GCMs predictions were that, in spring, the minimum was 2.65°C, the median was 5.21°C and the maximum was 8.10°C; in summer, the minimum was 4.70°C, the median was 5.60°C and the maximum was 6.5°C; in autumn, the minimum was 5.67°C, the median was 6.04°C and the maximum was 6.42°C; in winter, the minimum was 4.02°C, the median was 6.06°C, and the maximum was 7.91°C.(3) The trends of the changes of NDVI in each season between 2001, 2004, 2007, 2010 and 2013 were non-linear. In spring, NDVI showed an increasing-decreasing-increasing-and-decreasing trend, all changes significant. In summer, NDVI increased for the first three periods significantly and had an insignificant increase in the fourth period. In autumn, NDVI significantly increased for the first two periods, then increased insignificantly and finally decreased significantly. In winter, NDVI increased significantly at first, decreased significantly, decreased insignificantly and increased significantly at last. Comparing the two averaged NDVI in the period of 2001-2007 and the period of 2007-2013, we found that the NDVI in spring decreased significantly, the NDVIs in summer and in winter increased significantly but the NDVI in autumn increased insignificantly.(4) The GLM using temperature and precipitation as variables explained 62% deviance. The GLM using precipitation as variable was the best model. Using the best ranked GCMs and GLM model with the above two variables, NDVI was predicted to increase in each season in 2085. Using the lower limits of the 14 GCMs predictions of temperature changes and precipitation changes, the results was that in spring, the NDVI decreased insignificantly; in summer, NDVI increased significantly; in autumn, NDVI increased insignificantly and in winter, NDVI decreased insignificantly. The upper limits of the 14 GCMs predictions showed that NDVI increased significantly in all seasons. The results generated from precipitation model and the lower limits of 14 GCMs predictions was that NDVI decreased in all seasons. The results from the 14 GCMs confidence interval lower limits were much reliable according to the current vegetation changes in Northwest China.