Non-linear decision thresholds and adaptation lock-in under climate change: Evidence from a professional decision-making context

Effective climate adaptation depends on how climate-related information, experience, and expectations are translated into decisions across climate-sensitive sectors. Adaptation decision-making is often implicitly assumed to respond linearly to increasing risk, yet growing evidence indicates the presence of thresholds and lock-in effects. Using survey data from a climate-exposed professional sector as an empirical test case, we apply Bayesian inference and explainable machine-learning methods to uncover non-linear adaptation decision dynamics.

Across respondents, the predicted probability of adaptation advocacy is moderate (0.60, 95% CI: 0.52–0.68), establishing a baseline against which critical decision thresholds are detected. While beliefs in local climate impacts are generally strong, beliefs in personally experienced impacts are weaker, highlighting potential gaps between abstract climate information and experiential knowledge. Model-based predictions reveal that adaptation advocacy increases with expected impacts up to critical thresholds but declines beyond them, reflecting risk- and opportunity-associated tipping behaviour rather than monotonic responses.

These non-linear patterns are partly contingent on both the expected impacts of climate change, arising from the interaction of information, personal experience, and prior knowledge, and reported access to adaptation measures considered effective. Incorporating anticipated risk and opportunity alongside perceived actionability alters predicted decision regimes, producing conditional lock-in, where adaptation behaviour persists at a given level but may fail to increase even as risk rises. This demonstrates that climate information alone is not universally effective; rather, it must be tailored to the needs, experiences, and capacities of recipients, as the timing and magnitude of tipping behaviour, shaped by perceived impacts and available options determines whether information translates into meaningful adaptive action.

Although the empirical analysis focuses on one professional group, the identified decision thresholds and lock-in mechanisms are likely relevant across climate-sensitive sectors. These findings have implications for designing climate services that support actionable, context-sensitive adaptation under uncertainty, bridging the gap between information provision and adaptive decision-making.