Study the Spatial Urban Heat Variations Through a Data-Driven Approach

The associated effects of heat exposure in cities are not spatially uniform. Accurate detection of spatial urban heat variations requires choosing appropriate proxies and relevant resolution of data to avoid Modifiable Areal Unit Problem (MAUP). Selecting data resolutions congruent with system dynamics is challenging, yet it is central in ascertaining appropriate heat mitigation strategies.

Thus, in this study, through a data-driven approach, we aim to explore what information can be gained about heat exposure through different kinds of spatial data at a range of resolutions. To this purpose, we employ various types of known and theoretically based predictors such as the presence of specific street level features (derived from Google Street View images), demographic-related variables (derived from census statistics), and zoning data (e.g., land use, neighborhood planning units, statistics zones, etc.) to explain air temperature, land surface temperature and mean radiant temperature (MRT) variations within the city of Atlanta. Field measurements of air temperature, satellite imagery, and MRT modeling data of various resolutions are treated as dependent variables in multivariate regression analyses. Our findings demonstrate that street-level features can explain the most variation in land surface temperature and MRT among the groups of variables studied in this paper. Moreover, examining different data resolutions in the regression model show that downgrading data resolutions up to a certain threshold value does not affect street-level features' capability to explain temperature considerably, suggesting the usefulness of lower resolution data in the case of high-resolution data paucity (Figure 1).

The results of this study contribute to utilizing appropriate sets of data and relevant resolution of temperature measurements for representing spatial urban heat variations and consequently devising place-based heat mitigation strategies.