Abstract Forest restoration has become a global priority due to its potential benefits (carbon storage, biodiversity, etc.), but its successful delivery depends on decisions made by people, especially landholders. Despite this, our theory-based understanding of what drives landholders’ decisions to create woodland types important to achieve ecosystem restoration at scale, such as mixed native woodlands which have the highest potential for environmental benefits, remains limited. Building on Diffusion of Innovations Theory, we survey landholders in and around the Cairngorms National Park in Scotland to quantify the effects of woodland creation initiatives, individual landholders’, and contextual characteristics on decisions to create new mixed native woodland using a Bayesian Bernoulli distributed logistic regression. We find that landholders are more likely to have engaged in mixed native woodland creation if they believe it is compatible with their current land use practices; or if they anticipate increases to wildlife habitat or soil quality. Landholders that create mixed native woodland are more likely to perceive interventions to control herbivore damage to new woodland as difficult to understand. Our results suggest that, in contexts with established targets and economic incentives for woodland creation, governments could more effectively promote the uptake of mixed native woodland creation by: emphasising woodlands’ compatibility with other land uses such as agriculture (e.g. agroforestry) or some types of hunting; maintaining or improving the flexibility of woodland creation actions; facilitating access to easy-to-understand evidence-based comparisons of the potential benefits of different woodland types; and supporting landholders’ knowledge needs, especially for herbivore control tasks like fencing. Overall, we show that to achieve ecosystem restoration and its multiple benefits at scale, programme designs need to be more sensitive to landholders’ perspectives whilst incorporating safeguards against negative outcomes.
Bellot et al. (Fri,) studied this question.