Low probability forecasts are often not taken seriously enough. This is particularly problematic for extreme weather events with potentially severe consequences. A fundamental challenge arises from an inherent probability-uncertainty trade-off: widening the forecast area (lower spatial resolution) increases probabilities but also the uncertainty about where the event may occur and therefore, if the user will be affected—“wind gusts in the state of Baden-Württemberg” is more likely than “wind gusts in Tübingen”. We conducted an online experiment with a quota-representative sample for Germany (N = 1182) to investigate how spatial resolution of probabilistic forecasts influence the perceived risk to be affected. We employed a mixed factorial design in which spatial resolution and forecast probability were varied within subjects and forecast presentation format between subjects in three conditions: i) numeric descriptions only, ii) maps retaining similar spatial patterns across resolutions, and iii) maps showing substantial structural change when aggregated, including reduced zero-probability areas and larger contiguous regions with elevated probabilities.

Spatial resolution affected risk perception differently depending on the format: When forecast information was presented numerically, participants estimated their likelihood to be affected to be lower for forecasts with lower spatial resolution. With maps, this decline did not occur; instead, estimates aligned more closely with forecasted probabilities—independent of spatial resolution. Across all formats, trust and perceived usefulness were highest for high-resolution forecasts and decreased with lower resolution. This suggests that although forecast maps might elude the probability-uncertainty trade-off, lower spatial resolution might still affect how the information is perceived and used.