Enhancing Social Vulnerability Index Methods-A Simulated Case Study on Wyalusing-Boston Rail Line

Abstract: LNG imports are vital, their role is limited due to the high liquefaction costs, transportation, and regasification. However, given the constraints on pipeline infrastructure and the public resistance to new pipeline projects, LNG transportation by rail is increasingly considered a viable alternative. The extensive U.S. pipeline network faces significant challenges in expanding to meet growing demand. Building new pipelines requires substantial financial investment and involves lengthy permitting processes. These factors make rapid expansion of pipeline capacity difficult. In New England, these challenges are particularly acute. The region's dense population, social and environmental concerns complicate pipeline construction. The existing pipeline infrastructure often fails to meet demand during peak periods, leading to higher energy prices and supply uncertainties. Transporting LNG by rail offers a promising solution to New England's supply challenges. Rail transport provides flexibility, scalability, and the ability to reach areas not served by pipelines. However, it also brings its own set of challenges and regulatory hurdles out of which social vulnerability is the prominent one. The Social Vulnerability Index (SVI) assesses a community's resilience to external stresses on human health, including natural or human-caused disasters, using four key domains: socioeconomic status, household composition and disability, minority status and language, and housing and transportation. Utilizing data from the 2000 U.S. Census and the 2010 to 2014 American Community Survey (ACS) at the census tract level, the SVI captures demographic homogeneity and is commonly used in policy and planning. We identified the density of Class 2 hazardous incidents as a proxy of LNG form FRA data to conduct simulation to identify impact areas to select affected census tracts. A total of 195 census tracts were extracted where reconstructed SVI tool was applied with 15 census variables from 4 aforementioned categories and three new variables – number of fire stations, hospitals and building footprint within 5-15 minutes driving distance. The result of the optimized SVI method differed from original scores. Also, the multiple linear regression indicates that the Multiple R-squared and adjusted R-squared values for the old SVI calculation method are 1. Theoretically, this suggests a perfect fit for predicting the dependent variable using the 15 independent variables (from 4 theme areas). However, such a scenario is highly unusual in practice. Often, an R-squared of 1 signals an overfitted model. In contrast, the new SVI calculation method incorporating three additional variables yields a multiple R-squared of 0.646 and an adjusted R-squared of 0.634. This suggests a more realistic and robust prediction model. Therefore, we recommend including these new parameters as themes to enhance SVI calculation for such type of transportation and infrastructure development planning to accurate capture the real-world situation.