Towards a compound-event-oriented climate model evaluation: a decomposition of the underlying biases in multivariate fire and heat stress hazards

Abstract. Climate models' outputs are affected by biases that need to be detected and adjusted model climate impacts. Many hazards climate-related impacts associated with the interaction between multiple drivers, i.e. compound events. So far typically assessed based on hazard of interest, it is unclear how much a potential bias in dependence drivers contributes overall interact. Here, copula theory, we develop multivariate bias-assessment framework, which allows for disentangling indicators terms underlying univariate their statistical dependence. Based this dissect fire heat stress suite global models considering two simplified indicators: wet-bulb globe temperature (WBGT) Chandler burning index (CBI). Both indices solely rely relative humidity. The spatial pattern well represented models. However, substantial exist representation extreme conditions, especially CBI (spatial average absolute bias: 21 ?C) due driven humidity (20 ?C). Biases WBGT (1.1 small compared (1.9 (1.4 ?C), as compensate each other. In many regions, also related (0.85 important WBGT, indicates well-designed physically adjustment procedures should considered depend drivers. proposed compound-event-oriented evaluation easily applicable other types. Furthermore, can contribute improved present future risk assessments through increasing our understanding biases' sources simulation