|
Abstract:
|
Extreme weather events have incredible impacts on agriculture, public health, and natural resources. Quantifying the likelihood of extreme weather events is important to financial institutions, regulatory agencies, and many other stakeholders. Extreme value theory has played a critical role in describing the probability of extreme weather-related outcomes. Typically, extreme value theory for spatial data focuses on the behavior of componentwise maxima over contiguous blocks of time. This approach does not necessarily preserve the spatial and temporal structure of extreme weather since the extreme values may occur at different times. We propose a new method for quantifying the likelihood of extreme weather events that preserves the spatial and temporal structure of weather extremes. Utilizing principles from times series analysis, spatial statistics, and functional data analysis, the methodology provides a novel approach for describing extreme weather events. This methodology will be applied in describing the distribution of extreme weather events in the three decades following 2040 using computer model data from the North American Regional Climate Change Assessment Program.
|
