Vegetation response to climate conditions based on NDVI simulations using stepwise cluster analysis for the Three-River Headwaters region of China

Abstract

Located in the hinterland of the Qinghai-Tibet Plateau, the Three-River Headwaters region (THR) features unique eco-environmental conditions and fragile ecosystems, which are very vulnerable to climate change. To investigate the impacts of varying climate conditions, the Normalized Difference Vegetation Index (NDVI) was employed as an indicator to reflect the response of the vegetation dynamics. A series of pixel based VEgetation Dynamics Stepwise-cluster Prediction models (VEDSP) were proposed to establish the relations between NDVI and climate conditions through using the data series of remotely sensed precipitation and temperature. The obtained simulation results for training and testing showed very good agreements with the monthly NDVI observations. Rather than air temperature, the precipitation was identified as the critical climatic factor to result in various NDVI values, especially the 2-month consecutive average precipitation. The developed VEDSP models were further applied to predict the temporal and spatial distributions of NDVI values for five future years (2020, 2040, 2060, 2080 and 2100) according to climate projections of Euro-Mediterranean Center on Climate Change Climate Model (CMCC-CM) under the RCP4.5 scenario. The projected changes of NDVI indicated a slightly significant positive trend in annual average NDVI values, while the monthly peak values of NDVI for the entire THR would decrease by 5.34% in the future relative to the historical averages of the time period from 2000 to 2013. Distinct effects of precipitation and temperature on the response of NDVI were further demonstrated. Findings from this study would be used to help analyze the ecological effects of climate change and enhance the understanding of ecological changes in the future.