Temperature variations across socioeconomic status and built environment in the state of Illinois and the Netherlands

Abstract

Urban heat islands cause heat stress to urban residents and their severity is increasing due to both climate change and ever-increasing urbanization. Investigation of the issue has a storied past, but the methods used have all shared serious limitations. The goal of this study was to document and analyze correlations between surface air temperature and the average incomes of residents or the population density of areas affected. The study hypothesized that a positive correlation would be found between surface air temperature and population density, while a negative correlation would be found between surface air temperature and average income. The study’s literature review gives readers an easily comprehensible primer for this topic, covering both the nature of UHI and its methods of investigation. The issue of environmental justice and vulnerability to heat amongst certain populations is also explored as these ideas heavily informed the study’s purpose and hypothesis. This study used crowdsourced point measurements of the surface air temperature taken via NETATMO weather-recording devices in private homes. This data was used to investigate UHI and surface temperatures in general in the state of Illinois, the city of Chicago, the continental Netherlands, and the city of Amsterdam. This data was collected at four different times over the course of August 20, 2021. The point data was then used to create continuous field temperature maps of each study area, extrapolating temperature values between the points for which data was recorded. Average temperatures were then found for each administrative unit (county, municipality, or neighborhood) at each of the four times in each study area, resulting in four air surface temperature maps for each study area that track values throughout the day. Average income per household and population density data were collected for administrative units in all study areas and then compared with the surface air temperature maps to find trends, patterns, and possible correlations. Geographically weighted regression (GWR) and bivariate analysis (BA) were used to analyze the resulting map data in order to better understand how temperature values varied with distance between data points and across all four study areas. The results are strongest and most clear for the Netherlands, where data points are the most robust and widespread. Overall, the GWR results indicate a consistent and moderately strong correlation between surface air temperature and both average income values and population density. The BA results suggest that the correlation is negative for average income values and positive for population density. The majority of the results at each step of the investigative process are shown using maps created in ArcGIS Pro.