Afforestation is widely promoted as a natural climate solution, yet its long-term viability under intensifying climate stress remains uncertain, particularly in drylands where water limitation and climate extremes constrain forest function. We used the Ecosystem Demography model (ED2), constrained by flux tower and inventory data from Yatir Forest (Israel), to assess the resilience of an arid pine ecosystem to 21st-century climate change under historical and two high-emission scenarios (SSP3-7.0, SSP5-8.5) across five global climate models.
Simulations showed initial carbon gains followed by increasing sink instability, as net ecosystem productivity (NEP) became highly variable and frequently switched to carbon source conditions during drought. Drought frequency and severity increased substantially, with even moderate droughts reducing NEP by up to 50% and delaying recovery for several years, accelerating long-term carbon loss. Basal area accumulated in larger DBH classes as existing trees expanded, while limited recruitment led to a size-skewed structure dominated by mature cohorts. Although water-use efficiency initially improved under elevated CO₂, it declined after mid-century as aridity intensified, reflecting emerging ecohydrological constraints on forest function. Model validation revealed systematic overestimation of gross primary productivity (2–3×) and seasonal NEP biases, contributing to uncertainty in future carbon balance projections.
Our findings suggest that dryland afforestation may shift from carbon sink to source sooner than many projections indicate, as demographic bottlenecks, drought legacies, and ecohydrological constraints destabilize carbon balance. This underscores the need for rigorous model evaluation in climatically constrained systems and for integrating demographic thresholds into afforestation planning and climate mitigation strategies.
