Thermal representation as a decision variable in heat-adaptive tourism opportunity screening: evidence from a Madrid pilot

Extreme heat is a time- and place-sensitive management problem for urban tourism, yet broad climate-suitability assessment operates at a coarse scale, while downstream thermal routing and heat-adjusted accessibility presuppose the candidate set, leaving open the upstream question of which tourism opportunities should remain feasible candidates at a given hour. This study develops a constraint-first, uncertainty-aware architecture for screening urban tourism opportunities under heat and tests how much the choice of thermal representation matters. On a documented extreme-heat day in Madrid, across 27 tourism assets at three times of day, feasibility was assessed two ways: with a simple operational proxy combining ambient air-temperature hazard thresholds and nearby tree-presence information, and with physically based SOLWEIG/UTCI modelling. Changing the thermal representation reclassified one third of outdoor asset-time observations (33.3%, 14 of 42), in both directions. Relative to a conventional nearest-open baseline, constraint-first screening changed the feasible-alternative set in 7 of 8 decision scenarios, the nearest-open option failed screening in 3 of 8, and one scenario returned an explicit no-defensible-alternative outcome when no candidate qualified. The analysis evaluates decision sensitivity to thermal-method choice rather than the accuracy of either method against ground truth. The findings indicate that thermal representation is a substantive modelling choice in heat-adaptive tourism decision support, and that a constraint-first architecture can expose that sensitivity while keeping uncertainty, evidence, and exclusion logic explicit and complementing downstream routing and accessibility methods. Results are limited to a single Madrid pilot without direct field validation of Tmrt/UTCI or observed tourist behaviour.