ASSESSING NATIONAL PARKS FOR EMERGING CLIMATE TRENDS USING SPACE-TIME PATTERN MINING

Abstract

The need for climate risk assessment is growing in both the private and public sectors. However, conducting a spatially focused physical climate risk assessment can be challenging, as climate data is often large and multidimensional. This project aims to explore whether US national parks are exposed to emerging changes in climate by analyzing historical temperature and precipitation data to identify patterns in spatial clustering over time. Historical precipitation and temperature time series data by county across the contiguous US was extracted at 10-year intervals between 1900 and 2020 for the months of June and December and used to generate space-time cubes. A hot spot analysis was conducted across the cubes leveraging the Getis-Ord Gi* and Mann-Kendall statistics, and 16 classes of hot and cold spot patterns were created across the datasets, both for values and anomalies from the 1-month mean in the 1901-2000 base period. An analysis of total US national parks area coverage by space-time patterns shows that 6.4% was exposed to historical cold spot patterns for June precipitation values, 12.3% was exposed to consecutive hot spot patterns for December precipitation anomalies, and 6.5% and 3.2% was exposed to new hot spot patterns for June and December precipitation anomalies, respectively. The results of this study suggest some emerging precipitation patterns appear to occur in areas where national parks are situated. Understanding changes in climate patterns is important, especially in areas that are designated for conservation, as over time, these factors can have an influence on ecology and biodiversity.