Episodic air pollution events that occur because of wildfires, dust storms and industrial incidents can expose populations to particulate matter (PM) concentrations in the thousands of µg m-3. Such events have increased in frequency and duration over recent years, with this trend predicted to continue in the short to medium term because of climate warming. The human health cost of episodic PM events can be significant, and inflammatory responses are measurable even after only a few hours of exposure. Consequently, advice for the protection of public health should be available as quickly as possible, yet the shortest averaging period for which PM exposure guideline values (GVs) are available is 24-hours. To address this problem, we have developed a novel approach, based on Receiver Operating Characteristic (ROC) statistical analysis, that derives 1-hour threshold concentrations that have a probabilistic relationship with 24-hour GVs. The ROC analysis was carried out on PM10 and PM2.5 monitoring data from across the US for the period 2014 to 2019. Validation of the model against US Air Quality Index (AQI) 24-hour breakpoint concentrations for PM showed that the maximum-observed 1-hour PM concentration in any rolling 24-hour averaging period is an excellent predictor of exceedances of 24-hour GVs.