Hazard assessments are invariably a blend of expert interpretations of geophysical events and statistical descriptions of these events. Analyses of the recurrence rate and magnitude of events, their spatial density and their potential effects are essential components of hazard assessment for nuclear facilities. This chapter explores a robust approach to estimating spatial density using kernel methods and describes new methods of quantifying the uncertainty in these estimations using statistical techniques. Some of the spatial density estimation methods presented in this chapter have been used since the mid 1990s. In addition, new tools are emerging that offer improved understanding of spatial density estimates and their application in hazard assessments. For example, algorithms have been developed for numerical optimization of estimates of spatial density. Smoothed bootstrap techniques provide a mechanism for assessing uncertainty in spatial density, especially where information on past events is sparse. Methods in parallel processing have revolutionized the way we explore models of spatial density, in ways that were not practical even a decade ago. These developments are exceedingly encouraging. Although purely quantitative descriptions of spatial density, by themselves, are unlikely to ever be sufficient for assessment of hazard and risk, these quantitative estimations combined with expert judgment provide a powerful tool for improving these assessments. Thus, recent developments in quantitative density estimation will have a significant impact on the quality of geologic hazard assessments for nuclear facilities.