Spatiotemporal changes of normalized difference vegetation index (NDVI) and response to climate extremes and ecological restoration in the Loess Plateau, China

Extreme drought, precipitation, and other extreme climatic events often have impacts on vegetation. Based on meteorological data from 52 stations in the Loess Plateau (LP) and a satellite-derived normalized difference vegetation index (NDVI) from the third-generation Global Inventory Modeling and Mapping Studies (GIMMS3g) dataset, this study investigated the relationship between vegetation change and climatic extremes from 1982 to 2013. Our results showed that the vegetation coverage increased significantly, with a linear rate of 0.025/10a (P < 0.001) from 1982 to 2013. As for the spatial distribution, NDVI revealed an increasing trend from the northwest to the southeast, with about 61.79% of the LP exhibiting a significant increasing trend (P < 0.05). Some temperature extreme indices, including TMAXmean, TMINmean, TN90p, TNx, TX90p, and TXx, increased significantly at rates of 0.77 mm/10a, 0.52 A degrees C/10a, 0.62 A degrees C/10a, 0.80 A degrees C/10a, 5.16 days/10a, and 0.65 A degrees C/10a, respectively. On the other hand, other extreme temperature indices including TX10p and TN10p decreased significantly at rates of -2.77 days/10a and 4.57 days/10a (P < 0.01), respectively. Correlation analysis showed that only TMINmean had a significant relationship with NDVI at the yearly time scale (P < 0.05). At the monthly time scale, vegetation coverage and different vegetation types responded significantly positively to precipitation and temperature extremes (TMAXmean, TMINmean, TNx, TNn, TXn, and TXx) (P < 0.01). All of the precipitation extremes and temperature extremes exhibited significant positive relationships with NDVI during the spring and autumn (P < 0.01). However, the relationship between NDVI and RX1day, TMAXmean, TXn, and TXx was insignificant in summer. Vegetation exhibited a significant negative relationship with precipitation extremes in winter (P < 0.05). In terms of human activity, our results indicate a strong correlation between the cumulative afforestation area and NDVI in Yan'an and Yulin during 1998-2013, r = 0.859 and 0.85, n = 16, P < 0.001.