Sensitivity of temperate vegetation to precipitation is higher in steppes than in deserts and forests

Abstract

Recognizing the sensitivity of vegetation to precipitation is fundamental to monitoring and assessing ecosystem health and predicting ecosystem response to climate change. However, there is a lack of detailed, large-scale assessment of the sensitivity of temperate vegetation to climate variability. We analyzed the spatiotemporal trends of NDVI in Inner Mongolia (IM) for the period from 2000 to 2020, evaluated the sensitivity of NDVI to precipitation by constructing a linear mixed-effects model, and assessed the relative importance of an array of biotic and abiotic factors in affecting the vegetation sensitivity to precipitation using random forest models. We found that (1) the NDVI in IM had a significant or slight increase in 80.86% of the region during the studied two decades, though it showed a large inter-annual fluctuation, and the areas with stable or decreasing NDVI were mainly in the desert areas. Precipitation and wind speed were the two major drivers for the NDVI changes. (2) The typical steppes demonstrated the highest sensitivity to precipitation, followed by desert steppes and forest steppes, broad-leaved forests, and deserts and coniferous forests. (3) The major factor affecting the vegetation sensitivity to precipitation varied with vegetation types; it is temperature in the coniferous forests, forest steppes and steppe deserts, but it is the sunshine hours in the deciduous broad-leaved forests, the typical steppes, and the desert steppes. The high sensitivity of the typical steppe vegetation to precipitation is related with its location on climate gradient, it has a high production potential but limited by precipitation under a semiarid climate. Our analysis deepens the understanding of the NDVI sensitivity to precipitation and its influencing factors in different vegetation regions, and is valuable for predicting vegetation changes and developing vegetation management strategies to adapt to climate changes.

Keywords

• Vegetation sensitivity
• Remote sensing
• Trend analysis
• Linear mixed-effects model
• Random forest
• Inner Mongolia