A. E. Madrid García, J. M. Angulo Ibánez
Extreme values and risk assessment are crucial for spatiotemporal analysis in fields such as environmental science, geophysics, and engineering. Hotspots are commonly identified as regions where a magnitude exceeds a critical threshold. This work presents a framework for assessing spatiotemporal hotspot dynamics based on risk measures. In contrast to traditional approaches, hotspots are defined as regions where risk measures, derived from threshold exceedance indicators, exceed a predefined risk level. The methodology incorporates conditional simulation from the spatiotemporal random field model and a local analysis using sliding windows to assess exceedance set properties and derive empirical risk measurements. Two strategies for threshold selection are examined, highlighting their significance and impact on risk estimation and hotspot mapping. Furthermore, entropy-based measures provide insights into the complexity of hotspot dynamics and the temporal evolution of risk.
Keywords: hotspot, risk mapping, risk measure, spatiotemporal process, threshold exceedance indicator
Scheduled
Spatio-Temporal Statistics I
June 11, 2025 3:30 PM
MR 3