Assessing the relationship between urban structure and the urban heat island: Improving local climate zone mapping and estimating zone and site-scale effects on air temperature
Abstract
Besides temperature difference between `urban' and `rural' areas, the heterogeneity of urban structures (vegetation, impervious surface) causes the magnitude of urban heat island (UHI) vary substantially within the same city. Local climate zone (LCZ) provides rich information about urban biophysical properties, strongly tracks the spatial variability of UHI, and is easy to map using freely-available Landsat 8 data and the Database and Access Portal Tools (WUBDAPT) process. Yet, overall accuracy for WUDAPT based map is only ~60% and the extent to which the LCZ construct usefully partitions temperatures strongly depends on time of day.
In the present study, we evaluate the extent to which higher-resolution mapping methods better describe Washington, DC's urban environment (e.g., buildings, tree canopy) and characterizes spatial variability in air temperature. We employ object-based image analysis (OBIA) technique based on lidar and vegetation abundance data segmented using the `superpixels' algorithm programmed in R for LCZ mapping and assess zone classification accuracy compared to manual zone delineation. We further validate our high-resolution zones based on analysis of variance in day/night near-surface air temperature collected in 2018 and 2020 using car and bicycle traverse in the hottest part of the day and in pre-dawn hours. Subsequently, we estimate the effect of urban structure at sub-zone spatial scales using day/night temperature based on the hypothesis that for some LCZ types temperature tends to vary within LCZ more than among LCZ types at times of high heat and unstable boundary layer conditions. Preliminary work suggests that using high-resolution 2D/3D data and OBIA approach can substantially improve LCZ classification accuracy. 10-class LCZ maps were produced with 64% overall accuracy for WUDAPT and 81% using OBIA. Variance tests on day/night temperature indicate that air temperature differences among LCZ types is strongly significant at night, in accordance with the literature. However, our data suggest tree canopy cover at sub-zone scale is a more important determinant of daytime temperature. The findings of this study can inform UHI mitigation through improved zoning and sub-LCZ rational landscape planning, including targeted tree planting.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2020
- Bibcode:
- 2020AGUFMGC1020003G
- Keywords:
-
- 3307 Boundary layer processes;
- ATMOSPHERIC PROCESSES;
- 0414 Biogeochemical cycles;
- processes;
- and modeling;
- BIOGEOSCIENCES;
- 1630 Impacts of global change;
- GLOBAL CHANGE;
- 1631 Land/atmosphere interactions;
- GLOBAL CHANGE