Quantifying the long-term trends of changes in terrestrial vegetation on a large scale is an effective method for detecting the effects of global environmental change. In view of the trend towards overall restoration and local degradation of terrestrial vegetation in China, it is necessary to pay attention to the spatial processes of vegetative restoration or degradation, as well as to clarify the temporal and spatial characteristics of vegetative growth in greater geographical detail. However, traditional linear regression analysis has some drawbacks when describing ecological processes. Combining nonparametric linear regression analysis with high-order nonlinear fitting, the temporal and spatial characteristics of terrestrial vegetative growth in China during 1982-2012 were detected using the third generation of Global Inventory Modeling and Mapping Studies (GIMMS3g) dataset. The results showed that high-order curves could be effective. The region joining Ordos City and Shaanxi Gansu Ningxia on the Loess Plateau may have experienced restoration-degradation-restoration processes of vegetative growth. In the Daloushan Mountains, degradation-restoration processes of vegetative growth may have occurred, and the occurrence of several hidden vegetative growth processes was located in different regions of eastern China. Changes in cultivated vegetation were inconsistent with changes in other vegetation types. In southern China and some high-altitude areas, temperature was the primary driver of vegetative growth on an interannual scale, while in the north, the effect of rainfall was more significant. Nevertheless, the influence of climate on vegetation activity in large urban areas was weak. The trend types of degradation-restoration processes in several regions were inconsistent with the implements of regional land development and protection strategy. Thus, the role of human activity cannot be ignored. In future studies, it will be still necessary to quantify the effects of human management on spatial patterns, develop trend-fitting methods, and explore more refined methods of analyzing the driving forces affecting large-scale changes in vegetative growth.
Keywords: China; GIMMS NDVI3g; Influence factors; Interannual vegetation variation; Nonlinear fitting.