Predicting the imediate and legacy impact of the 2022 UK extreme summer on forest productivity.

Climate change is increasing the frequency, intensity and duration of drought events. Globally, the observed area in drought was on average 27% for the period 2018-2022, up 74% compared to the period 1981-2017 (S. H. Gebrechorkos et. al. 2025). This global amplification of droughts has contributed to widespread declines in forest productivity and increased drought-induced mortality.

In line with global trends, the United Kingdom has experienced increasingly dry and hot summers, with record temperatures in excess of 40 °C measured in the summer of 2022. As this trend continues under climate change, predicting the immediate and legacy impact of droughts on forests is increasingly important for forecasting ecosystem functioning and resilience, and for constraining how carbon uptake modulates carbon-climate feedbacks.

Land surface models, including the UK land surface model (LSM) JULES (the Joint UK Land Environment Simulator), tend to overestimate the direct effects of water stress on gross primary productivity (GPP) and latent energy (LE) fluxes, whilst lacking a mechanistic representation of post-drought legacy effects. In this study, we implement two stomatal optimisation models into the JULES improving predictions of GPP and LE under drought conditions. We extend the hydraulic components of these models to capture the legacy impact of drought-induced hydraulic conductance loss. The first approach treats conductance loss as instantaneous, depending solely on the historic maximum water stress under drought. The second approach treats conductance loss as cumulative, depending on both the magnitude and length of the drought event.

Focussing on the 2022 UK drought at the Alice Holt eddy covariance site in Southern England, we find that our models predict reductions in GPP and LE of -6% and -20% respectively when comparing 2022 to non-drought years. Flux tower observations indicate a -20[+1,-11]% reduction in LE. While the instantaneous and cumulative hydraulic legacy models predict permanent conductance losses of 11.56[+0.02,-0.3]% and 7[±1]%, respectively, neither captures a notable decline in GPP or LE in the year following the 2022 drought.

We then asked what change in drought extremes would be required to induce hydraulic failure in this UK Oak Woodland. To test this, we repeated the simulated experiment,  intensifying the drought by removing spring rainfall (broadly consistent with spring 2025) and progressively reducing 2022 rainfall, first by half and then to a quarter of its original amount. The instantaneous and cumulative hydraulic legacy models, when applied to the half (quarter) rainfall, predict permanent conductance losses of 16.6[+0.1,-0.4]% (34[+0,-2]%) and 39[+2,-0]% (65.2[+0.6,-0.2]%) respectively. The larger permanent conductance losses under the increased drought conditions were sufficient to induce significant reductions in GPP and LE in the year following the drought.

Our results present an important advance in our ability to forecast the long-term impact of drought on tree productivity and resilience within LSMs. By mechanistically capturing both immediate and legacy hydraulic responses, these predictions provide a robust evidence base for decision-making related to forest management, the resilience of restoration plantings, and the role of forests in achieving net-zero emission strategies.