Trend shifts in satellite-derived vegetation growth in Central Eurasia, 1982-2013.

Abstract

Central Eurasian vegetation is critical for the regional ecological security and the global carbon cycle. However, climatic impacts on vegetation growth in Central Eurasia are uncertain. The reason for this uncertainty lies in the fact that the response of vegetation to climate change showed nonlinearity, seasonality and differences among plant functional types. Based on remotely sensed vegetation index and in-situ meteorological data for the years 1982-2013, in conjunction with the latest land cover type product, we analyzed how vegetation growth trend varied across different seasons and evaluated vegetation response to climate variables at regional, biome and pixel scales. We found a persistent increase in the growing season NDVI over Central Eurasia during 1982-1994, whereas this greening trend has stalled since the mid-1990s in response to increased water deficit. The stalled trend in the growing season NDVI was largely attributed by summer and autumn NDVI changes. Enhanced spring vegetation growth after 2002 was caused by rapid spring warming. The response of vegetation to climatic factors varied in different seasons. Precipitation was the main climate driver for the growing season and summer vegetation growth. Changes in temperature and precipitation during winter and spring controlled the spring vegetation growth. Autumn vegetation growth was mainly dependent on the vegetation growth in summer. We found diverse responses of different vegetation types to climate drivers in Central Eurasia. Forests were more responsive to temperature than to precipitation. Grassland and desert vegetation responded more strongly to precipitation than to temperature in summer but more strongly to temperature than to precipitation in spring. In addition, the growth of desert vegetation was more dependent on winter precipitation than that of grasslands. This study has important implications for improving the performance of terrestrial ecosystem models to predict future vegetation response to climate change.