Search

Abstract Remote sensing based vegetation index provides a practical method for the monitoring of vegetation dynamics at regional and global scales. Here, using a long-term remotely sensed normalized difference vegetation index (NDVI) dataset, we quantified the vegetation changes in the Loess Plateau (LP) over the last three decades (1982–2015), which includes the period before the Chinese"Grain for Green Program"(GGP) was launched (1982–1999), and the period after the GGP (1999–2015). The correlations between the NDVI and four climate related variables, i.e., precipitation, temperature, root-soil moisture (RSM), and a drought proxy-standardized evapotranspiration deficit index (SEDI), were also examined. The results indicated that, (i) the GGP strongly changed the vegetation in the LP. The growing-season mean NDVI (GSM-NDVI) and the annual mean NDVI (AM-NDVI) decreased slightly before the GGP launched in 1999, with slopes of −3.38 × 10−3 and − 8.00 × 10−4 year−1, respectively. However, they showed slight and significant (p < 0.05) increases after the GGP, with slopes of 4.75 × 10−3 and 2.32 × 10−3 year−1, respectively. (ii) Climate change (i.e., warming and drying) resulted in adverse effects on vegetation in the LP during the period before the GGP. However, the observed changes (i.e., wetting and reduced drought) exerted positive effects on the vegetation during the period after the GGP. (iii) Inter-annual variations of spatially averaged NDVI over the LP were primarily determined by RSM rather than any other climate related variables. In the southeastern LP, the inter-annual variation of GSM-NDVI was mainly determined by precipitation and SEDI, while the inter-annual variation of AM-NDVI was mainly caused by SEDI and RSM. Inter-annual variations of both GSM-NDVI and AM-NDVI were mainly determined by SEDI and RSM in the northwestern LP, and by temperature in the southwestern LP. Graphical abstract Unlabelled Image Highlights • The Grain for Green Program (GGP) greatly changed vegetation in the Loess Plateau. • Climate changes exerted adverse effects on vegetation in the LP before the GGP. • Wetting and reduced drought positively affected vegetation in the LP after the GGP. • The inter-annual variation of NDVI was primarily determined by soil moisture. [ABSTRACT FROM AUTHOR]

As the most frequent and longest-lasting climate-related disasters, droughts are complex processes caused by hydrological imbalance, and have significant effects on ecosystem function and socioeconomic development. However, most conventional satellite-based drought indices and combined drought indices are mechanically constrained by the fact that they were developed and evaluated for a specific climate or critical ecological region, especially in China. A comprehensive drought index should be developed for extensive application in different climate-hydrology regions. To address this problem, an optimised principal component analysis method (PCA-RF) was proposed to support drought assessment across the transitional area between the humid and arid regions in central China, including the Loess Plateau (LP) and Qinling Mountains (QL). This method eliminated the causal relationships of non-linear changes for multi-aspect parameters such as land surface temperature, precipitation and soil moisture, and was used in developing a new index named "Meteorological drought index by PCA-RF methods (PMDI)". The results showed that PMDI had higher correlations and lower root mean square error than other drought indices (established using the empirical-weights method) with the 1-month standardized precipitation evapotranspiration index (SPEI-1). That means PMDI performed well in meteorological drought monitoring. Additionally, the normalized difference vegetation index (NDVI) and solar-induced chlorophyll fluorescence (SIF) were utilised to evaluate PMDI effectiveness in detecting vegetation response to drought, which represents agricultural drought. The results indicated that correlations between PMDI and vegetation indices (VIs) were stronger when severe or extreme droughts occurred, and that their relationships were affected by the high spatiotemporal heterogeneity in vegetation change over the LP and QL. Overall, PMDI was shown here to be a very suitable index for monitoring meteorological drought (short-term) and has effective action in capturing the response of vegetation signals to agriculture drought events (long-term). Given the excellent performance of PMDI in diverse climate-hydrology conditions of our study regions, we concluded that the PMDI has considerable potential as a drought monitoring methodology on a larger spatial scale. • A new drought index was developed (PMDI), which integrated multiple sources of information describing different aspects. • PMDI performed well in different climate-hydrology regions for short- and long-term drought monitoring. • PMDI has reasonably good potential applicability in drought monitoring on a larger spatial scale. [ABSTRACT FROM AUTHOR]