Climate change has profound impacts on forest ecosystem dynamics and could lead to the emergence of novel ecosystems via changes in species composition, forest structure, and potentially a complete loss of tree cover. Disturbances fundamentally shape those dynamics: the prevailing disturbance regime determines the inherent variability of a system, and its climate-mediated change could accelerate forest transformation. We used the individual-based forest landscape and disturbance model iLand to investigate the resilience of three protected temperate forest landscapes on three continents – selected to represent a gradient from low to high disturbance activity – to changing climate and disturbance regimes. In scenarios of sustained strong global warming, natural disturbances increased across all landscapes regardless of projected changes in precipitation (up to a seven-fold increase in disturbance rate over the 180-year simulation period). Landscapes with historically high disturbance activity had a higher chance of remaining resilient in the future, retaining their structure and composition within the range of variability inherent to the system. However, the risk of regime shift and forest loss was also highest in these systems, suggesting forests may be vulnerable to abrupt change beyond a threshold of increasing disturbance activity. Resilience generally decreased with increasing severity of climate change. Novelty in tree species composition was more common than novelty in forest structure, especially under dry climate scenarios. Forests close to the upper tree line experienced high novelty in structure across all three study systems. Our results highlight common patterns and processes of forest change, while also underlining the diverse and context-specific responses of temperate forest landscapes to climate change. Understanding past and future disturbance regimes can anticipate the magnitude and direction of forest change. Yet, even across a broad gradient of disturbance activity, we conclude that climate change mitigation is the most effective means of maintaining forest resilience.

Funding provided by: European Research CouncilROR ID: https://ror.org/0472cxd90Award Number: 101001905 Funding provided by: Japan Advanced Institute of Science and TechnologyROR ID: https://ror.org/03frj4r98Award Number: JPMXD0722678534 Funding provided by: Japan Science and Technology AgencyROR ID: https://ror.org/00097mb19Award Number: JPMJRS23I4 Funding provided by: Japan Science and Technology AgencyROR ID: https://ror.org/00097mb19Award Number: JPMJRX21I4 Funding provided by: Japan Society for the Promotion of ScienceROR ID: https://ror.org/00hhkn466Award Number: 22KK0102 Funding provided by: National Park ServiceROR ID: https://ror.org/044zqqy65Award Number: Funding provided by: University of Wisconsin–MadisonROR ID: https://ror.org/01y2jtd41Award Number: Funding provided by: Japan Society for the Promotion of ScienceROR ID: https://ror.org/00hhkn466Award Number:

The dataset was simulated using the individual-based forest landscape and disturbance model iLand (Model page).