Forest decline, attributed to increased aridity under global climate change, has been observed with rising frequency worldwide. One of the knowledge gaps making its spatially explicit prediction difficult is the identification of the climatic settings that generate a significant change in the forest state. A relatively rare sequence of unfavourable climatic events – a short extreme drought followed by a prolonged moderate drought within one decade – has allowed us to examine how rainfall amount affects forest performance.
Large-scale monitoring, at high spatial and temporal resolutions, is required to study climatic effects on forest performance. Therefore, time-series of spatially interpolated rainfall maps, remote sensing images and tree growth data were used to estimate the environmental settings to which the forests are exposed, and the corresponding forest performance responses. Performance was estimated from Normalized Difference Vegetation Index (NDVI) values obtained from 32 Landsat satellite images for 1994–2011. To widen the study perspective we sampled forest performance along a rainfall gradient (250–750 mm) in the plantedPinus halepensis forests in Israel.
Performance response was not spatially homogeneous. Three response types could be identified along the rainfall gradient: stable performance with low correlation to rainfall pattern in the humid region (>500 mm), moderate performance decline with high correlation to rainfall in the intermediate region (350–500 mm), and steep performance decline with intermediate correlation to rainfall in the arid region (