The objective of this study was to explore satellite-based vegetation phenology to characterize ecosystem variability in the northern China steppes. Advanced Very High Resolution Radiometer (AVHRR) maximum Normalized Difference Vegetation Index (NDVI) composite data set from 1982 to 1990 was used to detect the onset of vegetation green-up for the steppe ecosystems. Our findings show a strong linkage between onset of green-up and climatic factors in the region. An Ecosystem Variability Index (EVI) was derived using the coefficient of variation for date of onset of green-up. The EVI showed the meadow steppe/forest ecosystems to be least ecologically variable and the desert steppe to be most variable. Large areas of typical steppe, near the desert steppe ecotone, exhibited similar EVI values to those of the desert steppe, suggesting greater ecosystem variability in this typical steppe area. This suggests that large expanses of typical steppe in this region are undergoing ecosystem changes that could result in lower grassland productivity and eventual changes in land use practices.
There continues to be controversies among scientists whether humans are contributing to land degradation (desertification) in arid and semi-arid regions of the world. One area of considerable interest is the Gobi desert of central Asia, which is one of the largest deserts on Earth. The Gobi Desert is of particular value for addressing this question because it is divided by two countries that employ vastly different land management practices. Land use in China is high intensity and in Mongolia land use is of low intensity.
In this study, climate and satellite remotely sensed data from 1982-1999 were used to investigate interannual variations in the areal extent the Gobi Desert boundary. Our results show substantial year-to-year variations in the size of the Gobi desert which was strongly correlated with annual precipitation (R2 = 0.81, P<0.000). Based on results of logistical analysis of the climatic and remotely sensed data, an actual evapotransporation threshold of 180 mm per year was identified as the factor for discriminating between areas of desert and steppes within the study area. Correlation values between the areal extend of the desert and climate data were highly significant for study areas in both Mongolia and Inner Mongolia, suggesting that at the scale of our study, human activities contributed little to interannual desert boundary fluctuation. Due to data availability constraints, we were only able to examine satellite imagery over an 18- year period which did not include both a wet and dry cycle for this region. Therefore, a more complete understanding of Gobi Desert boundary response to interannual climatic variation will require studies extended over a longer time period.