Comparison and evaluation between MODIS vegetation indices in Northwest China

The Moderate Resolution Imaging Spectroradiometer (MODIS) VI products provide consistent, spatial, and temporal comparisons of global vegetation conditions that can be used to monitor the Earth's terrestrial photosynthetic activity. In order to compare and evaluate the ability that the MODIS two indices (VI) , the normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) , monitor vegetation over a diverse range of biomes in Northwest China, the MODIS vegetation index products data set were used, which are from NASA LP DAAC(Land Process Distributed Active Archive Center) and have 250m resolution and 16-day composting periods in 2004.Different vegetation types including broadleaf forests, needleleaf forests, meadows, grassland, steppes, scrubs, desert and cultivated vegetation were chosen as representative types, based on Vegetation Map of Northwest China. Two CE-313 radiancements, which is portable instrument and has five filters between 450 and 1650 nm, were used to collect the spectrum data for calculating vegetation indices over different vegetation types in the plants growth season. The levels of vegetation cover were observed at the same spots and the same time when the CE-313 radiancements worked. The results show that the NDVI was higher than the EVI in most part of Northwest China. The difference between the two indices increased from deserts, steppes, cultivated vegetation, meadows to forest. Both NDVI and EVI were well indicated the distribution and the growth of various vegetations in arid and semi-arid area. The NDVI saturated in these high biomass types vegetations, such as broadleaf forests, needleleaf forests, meadows and part of cultivated vegetation, during the plants blooming period. The NDVI did not increase with the growth of these vegetations during the period. The EVI was different. It increased with the growth of vegetation. The NDVI saturation threshold was about 0.8. The length of the period that NDVI saturating was different from 1 to 5 months with the type of vegetation. The length of the period of broadleaf forests was the longest and that of the cultivated vegetation was the shortest. Though the height of high-cold meadows was not higher than forty cm in northeast of Tibetan Plateau, the NDVI of these regions saturated for 2 months during the plants flourished. The data gotten from the CE-313 showed that the NDVI of many kinds vegetation, for example, winter wheat, corn, clover blossom and forb high cold meadows, may saturate during the process of becoming mature. Crossplots of the two Vis showed a curvilinear relationship between them, such that the NDVI always had higher values but appeared to reach an asymptotic maximum value. When the NDVI was about 0.8, the NDVI was almost stop responding with the vegetation density while the EVI was still responding. Both NDVI and EVI have a good linear correlation with the levels of vegetation cover according to the data gotten from the CE-313. The Correlation of EVI was better than that of NDVI especially in the higher levels vegetation cover. The Correlation coefficient of EVI was 0.8112 and the NVDI was 0.6946.