<p>Sinkholes pose a significant hazard in karst regions worldwide. Information on sinkhole occurrence rates is crucial for accurate karst hazard assessments and hazard mitigation planning. However, collecting sinkhole occurrence data is very difficult as sinkholes, especially cover-collapse sinkholes, often occur unexpectedly with little or no prior surface manifestation. In this study, we developed a novel method that detects elevation changes between two successive high-resolution lidar-derived digital elevation models and analyzes these changes to identify cover-collapse sinkholes that occurred over a defined time frame. We introduced an efficient multistep filtering procedure, including an innovative aspect-based spatial analysis, to remove noise and non-sinkhole features from the detected elevation changes. We applied the method to lidar data collected in 2014 and 2023 for Hart County, Kentucky, USA, an area of highly developed karst, and identified 2633 cover-collapse sinkholes that occurred during the 10-year period. This yielded an estimated annual occurrence rate of 0.24 cover-collapse sinkholes per km² for the area. We believe this is the first robust method for quantifying sinkhole occurrence rates at the regional scale. Furthermore, using historical aerial imagery, we collected additional chronological information for a subset of the identified cover-collapse sinkhole occurrences, providing a valuable dataset for evaluating occurrence variation over time and studying cover-collapse sinkhole precursors and triggers. Our method can be broadly applied to many karst regions where multitemporal lidar data are available to estimate sinkhole occurrence rates, greatly improving karst susceptibility and risk assessments for the infrastructure and population of these regions.</p>