TL;DR
This study leverages satellite data to map the urban heat island (UHI) effect across diverse biomes in the continental US, revealing that impervious surface area is a strong predictor of land surface temperature, but ecological context significantly modulates UHI intensity. The findings highlight that urban growth in forested biomes leads to higher summer UHI, potentially increasing cooling energy demands.
What Problem Does This Paper Address?
Urbanization is a rapidly growing global phenomenon with significant environmental implications, including the urban heat island (UHI) effect. The UHI is characterized by higher temperatures in urban areas compared to their rural surroundings, impacting human health, energy consumption, and local weather. While previous studies have established a link between urban development and UHI, a comprehensive understanding of how this effect varies across different ecological settings (biomes) and urban sizes has been limited. This paper aims to bridge this gap by using remote sensing data to quantify the UHI across various biomes in the continental US, exploring the interplay between urban structure (impervious surface area), urban size, and ecological context.
Methods
The researchers employed a multi-faceted approach combining remote sensing data with GIS analysis. They utilized the Landsat-derived impervious surface area (ISA) dataset (NLCD 2001) to quantify the extent of man-made surfaces at a 30m resolution. This data was used to define five distinct urban zones within each city: Urban Core (≥75% ISA), Urban1 (75-50% ISA), Urban2 (50-25% ISA), Suburban (0-5 km buffer), and Rural (45-50 km buffer).
MODIS-Aqua Version 5 8-day composite land surface temperature (LST) data (MOD11A2) and 16-day composite NDVI (Normalized Difference Vegetation Index) at a 1km resolution were used to characterize surface temperature and vegetation density, respectively. These datasets, averaged over three annual cycles (2003-2005), were analyzed within the defined ISA zones.
To account for ecological context, the study stratified its analysis by terrestrial ecoregions defined by Olson et al. (2001), grouping cities into eight biomes. Further subdivision of two major biomes (Temperate Broadleaf and Mixed Forest, and Temperate Grasslands, Savannas and Shrublands) was performed to capture finer ecological distinctions.
Topographic data from the SRTM30 dataset was used to filter out temperature differences due to elevation and shading. Population data from US Census Grids 2000 was used to identify the 38 most populous urban areas in the continental US.
Datasets
- Landsat ETM+ and IKONOS imagery: Used to derive the impervious surface area (ISA) dataset (NLCD 2001) at 30m resolution.
- MODIS-Aqua Version 5:
- MOD11A2 (8-day composite LST): Used for land surface temperature measurements at 1km resolution.
- NDVI (16-day composite): Used to assess vegetation density at 1km resolution.
- US Census Grids 2000: Used for population data to identify the study cities.
- SRTM30 dataset: Used for topographic data (elevation).
- Olson et al. (2001) Terrestrial Ecoregions Map: Used to classify cities by biome.
Key Insights and Ideas
- ISA as a Dominant UHI Driver: Impervious surface area is a strong predictor of land surface temperature, explaining a significant portion of the variance in LST across most biomes. This underscores the direct impact of urban development on surface heating.
- Ecological Context is Crucial: The intensity of the UHI is not solely determined by urban development but is significantly modulated by the surrounding biome. This means the same amount of impervious surface can lead to different UHI magnitudes depending on the ecological setting.
- Forested Biomes Amplify Summer UHI: Cities located in biomes dominated by temperate broadleaf and mixed forests exhibit the largest summer daytime UHI amplitudes. This is attributed to the contrast between heat-absorbing urban surfaces and the cooling effects of dense vegetation (evapotranspiration) in surrounding rural areas.
- Desert Cities Exhibit a "Heat Sink" Phenomenon: In arid and semi-arid desert biomes, a "U-shaped" temperature gradient is observed, where urban cores are warmer, but suburban/rural fringe areas can be cooler than the urban core. This suggests a potential heat sink effect, possibly due to increased vegetation growth in these areas supported by irrigation.
- Seasonal Asymmetry of UHI: The UHI is significantly more pronounced during summer daytime than during winter. This has implications for energy consumption, suggesting higher cooling demands in summer, especially in forested biomes.
- Urban Size Matters: For cities within temperate forest biomes, there is a strong positive correlation between urban area size and UHI amplitude, with larger cities experiencing more intense heat islands.
Key Results
- ISA explains 70% of LST variance across all studied cities.
- The average yearly UHI (urban-rural temperature difference) is 2.9 °C.
- The average summer daytime UHI is 4.3 °C, while the average winter daytime UHI is 1.3 °C.
- Cities in temperate broadleaf and mixed forest biomes show the largest average summer daytime UHI, reaching up to 8 °C.
- For cities in temperate forest biomes, the UHI amplitude is strongly correlated with size, with urban areas larger than 1000 km² experiencing UHI of about 10 °C during summer days.
- The relationship between UHI amplitude and city size in Northeastern temperate forests is log-linear, explained by the equation: $\Delta T_{\text{Urban-Rural}} = 3.48\log(\text{Area}) + 1.75$, accounting for 71% of the variance.
- In desert biomes, the summer daytime UHI can be slightly negative (around -1 °C), indicating a heat sink effect.
Limitations and Future Work
- Spatial Resolution of MODIS LST: While MODIS LST data (1km resolution) is suitable for broad-scale analysis, it may not capture fine-scale UHI variations within urban areas. The study acknowledges that higher spatial resolution data would be beneficial for more detailed investigations.
- Ecological Context Granularity: The study uses broad biome classifications. Future work could benefit from higher spatial resolution biome information to capture ecoregional variations within cities, allowing for a more nuanced understanding of the interplay between urban surfaces and their immediate ecological settings.
- ISA as a Proxy: While ISA is a good indicator of development intensity, it doesn't directly capture other crucial factors like albedo or 3D structure, which also influence surface temperature.
- Generalizability of Desert UHI: The study notes considerable variation in UHI response within desert cities, suggesting that factors beyond broad biome classification might be at play.
The authors suggest that future research should focus on using higher spatial resolution data to investigate the interplay between urban surfaces and their ecological settings at a finer scale. This would enable more detailed city-scale investigations into UHI phenomena.
Summary updated