City module - City cluster and urban heat islands (Europe)
|
As one of the most evident influences of human activities on climate, urbanization has significantly affected the terrestrial ecosystems and can be observed, measured and perceived through diverse indicators. Among them, and considered one of the adverse results of ongoing urbanization is a phenomenon called the urban heat island (UHI) effect, i.e. urban areas experiencing elevated temperatures relative to the surrounding countryside (Oke, 1987). Although UHI studies have been conducted in the last decades in various cities, the understanding of the mechanism of UHI needs to be updated and enhanced due to the diversity of cities which vary in size, climate, economic development, etc.
To assess the UHI intensity (UHII) in Europe, we combined land surface temperature data and land cover data. We used the land surface temperatures MYD11A2 data obtained from the Moderate-Resolution Imaging Spectroradiometer (MODIS) on board the NASA's Aqua Earth Observing Satellite. The satellite scans a particular Earth location twice a day, with overpasses approximately at local times of 1:30 and 13:30. respectively. The MOD11A2 data used here represent averages over 8 days at approximately 1000 m spatial resolution with a temporal coverage from Jan. 2006 to Dec. 2011. 
Figure 2. General process flow for retrieving UHI. To identify city areas, the City Clustering Algorithm (CCA) proposed by Rozenfeld et al. (2008) was applied. It defines a "city" as a maximally connected cluster of populated cells with a population density larger than a cut-off threshold. Since it is very resource intensive and time consuming to gather and update gridded population data, high resolution land cover data derived via remote sensing serves as an optimal alternative used here for urban agglomeration identification. We used CORINE land cover data of the year 2006 at 250m spatial resolution (Bossard et al., 2000), reclassifying land cover data into urban and non-urban. The general process of identifying an urban area is illustrated in Fig. 1. CCA involves a clustering parameter l determining up to which distance urban cells are connected with each other, i.e. urban cells within the distance are assigned to the same cluster. Here we initialized radius burning CCA with 500 m (2 times CORINE spatial resolution) as a threshold distance. Analogously, the surrounding boundary areas of cities can be also identified with a similar method iteratively until the area is approximately the same size as the urban ones. The UHII is therefore defined as the mean temperature difference between the cluster and boundary. 
Figure 3. Cluster and boundary identified for Great London Area Results In general, UHI is positively correlated with city size. The larger a city is, the more pronounced the UHI is. In terms of seasonal variation, the high temperature in summer intensifies the UHI for most cities in temperate climate zones. However, Urban Cool Island or Urban Oasis Effect can be also observed in some Mediterranean cities, where the city represents lower0020mean temperature as the surrounding area. The mechanism of UHI in such cities deserves continued research. 
Figure 4. Nighttime land surface temperature map (captured 06.26.2006- 03.07.2006 Aqua) Interpretation aid and possible limitations It is worthy to mention that the land surface temperature data retrieved from MODIS is not equivalent to the near surface air temperature which normally has a smaller range. However, a variety of studies have shown that the air surface temperature and the land surface temperature are closely related. 
Figure 5. 3D illustration of the UHI in Paris
References
|