Seasonal patterns of Mediterranean evergreen woodlands (Montado) are explained by long-term precipitation

• What is the importance of precipitation to phenology and productivity of woodlands?
• Historical precipitation influenced more woodlands vegetation than concurrent one.
• Non-linear relationship between precipitation and seasonal metrics was identified.
• A threshold of vegetation response to annual precipitation was found at c. 650 mm.
• Identifying thresholds is essential to anticipate ecosystem degradation in drylands.

In temperate areas, vegetation seasonality and phenology have been mostly associated with temperature changes both in space and time. In drylands, where water is the most limiting factor, we expect that they strongly respond to water availability. The degree to what that response depends more on precipitation that occurred when vegetation seasonality and phenology were measured, or on the long-term precipitation, is not fully known. We hypothesize that in drylands, long-term precipitation better explains the patterns of seasonality and phenology metrics than concurrent one, due to constrains imposed by ecosystem legacy. We correlated long-term precipitation (30 years normal) and concurrent precipitation (12 years) to several seasonal metrics (MODIS, average of 12 years) measured in a savannah-like system, Mediterranean evergreen woodlands, located in southwest Europe (Portugal). We observed that seasonal metrics of productivity and phenology were more significantly related with long-term precipitation than with concurrent precipitation. Comparing the extremes of our gradient we found that drier areas (c. 496 mm long-term annual precipitation) showed average growth cycles of annual plants 25 days shorter and ended 16 days sooner than more rainy regions (c. 739 mm). Evergreen vegetation productivity was shown to be c. 30% lower in drier areas. Moreover, productivity and phenology metrics were non-linearly related to the long-term precipitation, suggesting both are particularly constrained below 600–650 mm. These results suggest a memory effect in the response of vegetation to climate, most probably associated to legacies on soil characteristics and on plant community. It also indicates the existence of ecosystem response thresholds in vegetation's response to precipitation along ecosystem transitions. Overall, this method can be used to track ecosystem services over space in drylands and for managing ecosystems for both mitigation and adaptation to climate change.