Long-term monitoring of NDVI changes by remote sensing to assess the vulnerability of threatened plants

Abstract

Little is known about how much continuous landscape transformation might affect the most vulnerable elements of biodiversity. In this study, we quantified changes in the normalized difference vegetation index (NDVI) over the past 35 years across locations with threatened plants and in Natura 2000 (N2000) protected areas, in an environmentally heterogeneous region of Southern Europe. First, we estimated the intensity and duration of NDVI gains and losses based on Landsat time series using the LandTrendr algorithm in Google Earth Engine. Then, we tested: 1) whether populations of threatened plants were located in more stable sites than non-threatened plants (i.e., lower NDVI changes); 2) whether NDVI changes around populations of threatened plants differed across habitats and inside/outside N2000 areas, and 3) whether lower NDVI changes occurred in N2000 areas than unprotected areas, thereby indicating their effectiveness at preserving biodiversity. Threatened plants tended to be concentrated in sites with less change irrespective of the habitat where they occurred and their location within protected areas. Occurrence in stable sites also reduced the risk associated with small-sized populations. N2000 areas were in line with the overall greening trend but they experienced less loss events than the unprotected areas, thereby supporting their role in slowing down human-induced land cover changes. Our approach demonstrates how long-term remote sensing monitoring can help to assess the effects of both slow processes and drastic landscape transformation events on priority plants in a comprehensive and rapid manner. This method can identify hidden patterns in extensive regions and guide effective conservation management.