NDVI Spatio-temporal Patterns and Climatic Controls Over Northern Patagonia

In this study, we identify the main modes of variability of the Normalized Difference Vegetation Index (NDVI) and their relationships with precipitation and temperature variations across northern Patagonia (36°–45° S). In this approach, we combined a recently developed high-resolution gridded dataset (20 × 20 km) for temperature and precipitation with a re-scaled NDVI grid to spatially match the climate database. Climate–vegetation relationships were analyzed taking into account a wide range of temporal variations (intra- to inter-annual) of both climate and NDVI. An Empirical Orthogonal Function analysis performed on NDVI delimits four regions that are spatially consistent with previous vegetation classifications for northern Patagonia. In addition, these coherent NDVI regions show similarities with the spatial precipitation patterns and the temporal evolution of precipitation over the common period 2001–2010. Both NDVI and precipitation show evident annual cycles over the Mediterranean climatic region in northwestern Patagonia. These annual cycles decrease in amplitude toward the eastern arid rangelands, and to the south on the evergreen all-year-round rainforests. Significant positive relationships between monthly precipitation and NDVI are recorded in the dry temperate rangelands in northeastern Patagonia. In contrast, direct associations between monthly NDVI and precipitation were absent in the Central Patagonia cold grasslands, where seasonal interactions between precipitation, temperature and NDVI appear to be more relevant. Relationships between NDVI and temperature are generally weaker east of the Andes, but significantly positive in late winter/spring over the temperate forests in western North Patagonia. Our results indicate that climate–NDVI relationships in northern Patagonia are biome specifics with the occurrence of temporal lags and precipitation–temperature interactions in the responses of vegetation to climate at some ecosystems.