Using SPOT-VGT NDVI as a successive ecological indicator for understanding the environmental implications in the Tarim River Basin, China

Ni-Bin Chang
2010 Journal of Applied Remote Sensing  
The resilience and vulnerability of terrestrial ecosystem in the Tarim River Basin, Xinjiang is critical in sustainable development of the northwest region in China. To learn more about causes of the ecosystem evolution in this wide region, vegetation dynamics can be a surrogate indicator of environmental responses and human perturbations. This paper aims to use the inter-annual and intra-annual coefficient of variation (CoV) derived by the SPOT-VGT Normalized Difference Vegetation Index (NDVI)
more » ... as an integrated measure of vegetation dynamics to address the environmental implications in response to climate change. To finally pin down the vegetation dynamics, the intra-annual CoV based on monthly NDVI values and the inter-annual CoV based on seasonally accumulated NDVI values were respectively calculated. Such vegetation dynamics can then be associated with precipitation patterns extracted from the Tropical Rainfall Measuring Mission (TRMM) data and irrigation efforts reflecting the cross-linkages between human society and natural systems. Such a remote sensing analysis enables us to explore the complex vegetation dynamics in terms of distribution and evolution of the collective features of heterogeneity over local soil characteristics, climate change impacts, and anthropogenic activities at differing space and time scales. Findings clearly indicate that the vegetation changes had an obvious trend in some high mountainous areas as a result of climate change whereas the vegetation changes in fluvial plains reflected the increasing evidence of human perturbations due to anthropogenic activities. Some possible environmental implications were finally elaborated from those crosslinkages between economic development and resources depletion in the context of sustainable development. Terms of Use: and natural systems to improve the understanding necessary to respond to global climate changes. Vegetation changes, as an important component of terrestrial ecosystems, may become such an ecological index via coupling the phenological effects and the anthropogenic impacts on a long-term basis [4] . It is evident that recent climatic changes have already affected species physiology, distribution, and phenology [5] . Causal attribution of recent biological trends to climate changes is complicated by non-climatic influences that dominate local, short-term biological changes [6] . This phenomenon is especially relevant in arid regions which cover about 45% of Earth's land surface; yet their ecosystem dynamics have been overlooked for a long time partly because arid ecosystems seem to have low rates of biological activities as the result of sparse biota [7] . However, Rotenberg and Yakir indicated that the dryland Yatir Forest in Israel can take up carbon at rates similar to those of pine forests in continental Europe [8] . Overall, arid ecosystems are very sensitive to a variety of physical, chemical and biological degradation processes [9] . These types of ecosystems are vulnerable due to poor water availability, and the spatial-temporal dynamics of vegetation are therefore largely impacted [10] . The Tarim River Basin, which is the biggest endorheic basin in Central Asia, can be considered as one of the least water-endowed regions in the world. There are a variety of vegetation types in the Tarim River Basin, caused primarily by regional, geologic, topographic, anthropogenic and climatic differences. Studies showed that climate conditions favor more river runoff during the past decade [11, 12] . But the water scarcity in the Tarim River has been exacerbated due to the growing water demand driven by population increase and agricultural development in the last decade [13, 14] . This makes the terrestrial vegetation dynamics of particular significance in terms of plant phenology, soil degradation and organic dynamics [15] . Improvements in the understanding of these factors can certainly provide better environmental management [16] . Remote sensing has obvious advantage in large-scale vegetation cover monitoring [17] . Multitemporal change detection with the aid of remote sensing data can be carried out by simple overlay of classified maps [18] . Vegetation indices have been developed to qualitatively and quantitatively assess vegetation cover using spectral measurements [19] . There are over thirty five vegetation indices that may be applied in different studies [17, 19, 20 ]. Yet each type of vegetation index has its own strength and weakness. For instance, the Ratio Vegetation Index (RVI) is the ratio of the near-infrared band divided by the red band that is popular in agricultural monitoring [19, 20] . The Near Infrared (NIR) increases with increasing canopy, whereas the red decreases with decreasing canopy. Hence, the RVI can be defined as NIR/Red (Landsat: Band 4/Band 3) such that brighter pixels are correlated with more biomass. This index, however, does not perform well when the vegetation cover is less than 50%; but it is the best choice at the dense vegetation cover [21] . Chlorophyll in actively growing vegetation is a strong absorber of red radiation, and the cell-wall structure of healthy leaves strongly reflects NIR radiation. Therefore, greater photosynthetic activity will result in lower reflectance in the red band and higher reflectance in the NIR. By combining these two spectral regions into a normalized ratio, the sensitivity to photosynthetic activity of vegetation may be enhanced. This ratio is the well-known Normalized Difference Vegetation Index (NDVI), which takes advantage of the spectral response difference of the chlorophyll-loaded vegetal tissues between the red and infra-red channels. The higher the NDVI values, the larger the green vegetation densities are present. The NDVI had been successfully utilized to monitor and assess vegetation covers from regional to global scale [20, 23] . In particular, the NDVI may reduce the effect of sensor degradation by normalizing the spectral bands [22] . It is generally agreed that the NDVI is sensitive to low dense vegetation such as observed in semi-arid areas [24, 25] . Many sensors have the geometric and spectral characteristics associated with different bands to monitor vegetation (e.g., AVHRR, Landsat-TM, SPOT-VGT and MODIS), which
doi:10.1117/1.3518454 fatcat:uovegc3cnzgixjjz3e3alhmwwa