Climate-driven variability in vegetation greenness over Portugal

The role of climatic variability on vegetation greenness over mainland Portugal (2000-2014) was assessed using the Normalized Difference Vegetation Index (NDVI) as a measure of vegetation greenness. The gridded Standardized Precipitation Evapotranspiration Index (SPEI), gridded maximum (TX) and minimum (TN) temperatures, defined on a monthly basis, were used to assess climatic variability. Different vegetation classes were studied separately: grasslands, holm oak and cork oak, shrubland, deciduous forests and other evergreen forests. Type-specific models were developed to represent NDVI variability, using lagged monthly anomalies (up to 6 mo) of SPEI, TX and TN as exploratory variables. For southern Portugal, occupied mainly by grassland and low density woodland, vegetation greenness is very sensitive to precipitation. This sensitivity is particularly clear in February-March and September. Conversely, in the northwestern regions, where shrubland and evergreen/deciduous forests prevail, vegetation greenness is much less sensitive to precipitation seasonality. After cross-validation, the type specific vegetation models explained 50-88% of the observed NDVI variability (relative errors of 3-7%). Models showed that SPEI significantly correlates with vegetation greenness at 3 and 6 mo timescales. At the same timescales, anomalously low TX and, to a much lower extent, anomalously high TN tend to favour vegetation greenness. It is thus possible to predict vegetation greenness in Portugal, up to 3 mo in advance and for different vegetation types, with some accuracy. A thorough understanding of the relationships between vegetation greenness and precipitation variability may promote a better management of forest/agroforestry systems, water resources, ecosystems and landscapes, particularly under changing climates.