A significant advancement in meteorological science has been presented in a preprint titled "The resolution of risk: covariance-informed spatiotemporal clustering improves the detection of hazardous weather events" on EGUsphere, an Earth System Science Open Archive platform. This research introduces a novel framework that employs covariance-informed spatiotemporal clustering to enhance the accuracy of detecting hazardous weather phenomena, such as heat waves and severe storms. The methodology addresses a critical gap in current practices, where the native resolution of multi-dimensional datasets, like those from ECMWF Reanalysis version 5 (ERA5), often fails to equivalently resolve fluctuations across space and time, leading to biased clustering. The study proposes a framework to quantify the relationship between space and time using space-time separable covariance modeling. By empirically deriving an equivalent spatial resolution for a given temporal resolution, the method ensures unbiased clustering across three dimensions. When applied to detect heat waves and severe storms across the Southeastern US using ERA5 data from 1940 to 2023, the covariance-informed resolution demonstrably improved the recall of heat waves, increasing it from 0.92 to 0.94 when compared against the NOAA Storm Events Database from 2019 to 2023. This indicates a more effective reconstruction of historical weather events. The research highlights the importance of data normalization prior to weather event reconstruction and represents a crucial step toward more precise identification and evaluation of the scale and variability of extreme weather, with vital implications for global disaster preparedness and climate change adaptation strategies, particularly relevant for vulnerable regions like India.