Despite the climate threat's severity, global policy responses remain anemic. One political challenge has been the temporal mismatch between short-term climate policy costs and long-term climate policy benefits. Will this policymaking obstacle weaken as the impacts of climate change begin to realize? Here we analyze the impact of a climate-related hazard on public support for costly climate reforms. Using a natural experiment based on randomness in the timing of California wildfires we link, for the first time, threat exposure to realized political behavior rather than self-reported attitudes or behavioral intentions. We find that census block groups within 15 km of a wildfire have approximately 4 to 6 percentage points higher support for costly pro-climate ballot measures. The effects are stronger for block groups closest to wildfires, dropping by approximately 1.7 percentage points for every 10km of distance. Moreover, the effect is concentrated among census block groups with a large or medium concentrations of Democratic voters; by contrast, voters in Republican-dominated census block groups are largely unresponsive to wildfires. Our results suggest that experienced climate threats may only enhance willingness-to-act in areas where the public already holds pro-climate identities.