In the face of global warming, reliable predictions of species and ecosystem responses to intensifying warm ocean temperature extremes are crucial. Currently, a lack of empirical evidence for how heat stress differentially impacts species across their distribution, limits our ability to calibrate predictions with real-world ecological change. Our study uses three decades of observations in kelp abundance to assess how different populations respond to marine heatwaves (MHWs). Further, we compare these realised observations with predictions based on species distribution modelling and experimental approaches. Our results support a hybrid model of thermal performance: thermal limits differ between populations, but volatility in thermal performance increases toward a species’ warm edge. This results in heightened vulnerability of warm-edge populations to MHWs. Moreover, realised impacts of MHWs occur at much smaller thermal anomalies than predicted by experimental approaches and distribution modelling, highlighting the importance of calibrating experimental and modelling approaches with realised observations. This thermal sensitivity analysis is one step towards the establishment of MHW risk profiles, which is an essential dimension to quantifying species responsiveness and vulnerability to MHWs. Further consideration of exposure and adaptivity will help build a comprehensive risk framework – important information for risk management of MHW’s ecological impacts.