SPATIAL DISTRIBUTION AND ITS SEASONALITY OF SATELLITE-DERIVED VEGETATION INDEX (NDVI) AND CLIMATE

The Normalized Difference Vegetation Index (NDVI) distribution and its seasonal cycle were investigated in relation to temperature and precipitation over Siberia and its surrounding regions. The analyses used 5-year (1987–1991) monthly means. The monthly mean NDVI was calculated from the third-generation monthly Global Vegetation Index (GVI) product; monthly temperature and precipitation at 611 stations were calculated from Global Daily Summary (GDS) data. The 611 stations were classified by cluster analysis into 10 classes based on the NDVI seasonal cycle (March–October). The geographical distribution characteristics of the NDVI cycle were described using temperature, precipitation and Olson’s land-cover type. In northern regions, where tundra vegetation prevails and temperatures and precipitation are low, the amplitude of the NDVI seasonal cycle is small. In southern regions, where temperatures are high and there is little precipitation, the seasonal amplitude of the NDVI is small because of the arid land type. Forested regions were split into six classes, each characterized by large amplitudes in the NDVI seasonal cycle. The phenological characteristics of the forest classes were noted. For example, a forest-class localized near Lake Baikal shows higher NDVI values, even with the presence of snow cover in March, compared with other regions. This high NDVI value suggests that the exposed green canopy of the coniferous forest can be observed even when snow is present. In addition, the NDVI peaks at stations near 60°N, where the maximum monthly temperature is around 18°C. This result suggests that the optimum temperature-precipitation environment coincides to the area in Siberia where the maximum monthly temperature is 18°C. Copyright © 2001 Royal Meteorological Society.