Resilience to drought in a dry forest: Insights from demographic rates

- Resilience capacity is critical to assess forest vulnerability to drought.
- Resilience assessments need to consider simultaneously main demographic rates.
- To evaluate resilience forest structure and management must be considered.
- High resilience could allow the persistence of a forest under climate change.
- Compensation between demographic rates can boost forest resilience.

Concern is rising on whether forest function and structure will recover from drought-related impacts, which are expected to increase under global warming. Understanding demographic mechanisms underlying resilience (i.e. capacity of a system or individual to restore its structure and function prior to a disturbance) is critical for properly assessing forest vulnerability to drought. The simultaneous estimation of resilience of the main demographic rates governing tree population dynamics (growth, recruitment and survival) allows for a comprehensive assessment of forest response capacity. We evaluate the resilience of a large Pinus pinaster forest (approx. 7500 ha) in Southern Europe to one of the driest decades of the last 70 years (i.e. 1942-52). As forest structure and management influence demographic rates, their effects were removed prior to calculating resilience values for growth, recruitment and survival. The extremely dry conditions negatively impacted tree growth and recruitment during drought and slightly decreased survival in the decade after drought. Resilience values were mostly high, despite some low values for recruitment or survival in some forest sections, which indicate a general recovery of growth and recruitment and an absence of widespread reductions in survival after drought. A joint analysis of the Demographic Resilience Index (calculated by combining growth, recruitment and survival resilience; DRI) and resilience values of each rate allows to detect demographic compensation effects. High DRI values, even in sections where resilience in recruitment or survival was low, indicate that low resilience values in a given rate were compensated by high resilience in the remaining rates. Recorded resilience could allow the long-term persistence of the studied forest, although increased frequency and intensity of droughts might exceed the critical threshold of system's resilience. Our approach provides a step toward an exhaustive resilience assessment; however, further research should consider potential resilience thresholds arising from more complex non-linear dynamics.