Identifying urban heat island: the Barcelona case

Abstract

There is a large consensus that cities have a special role in the process of climate change. Cities are responsible for 75% of global energy consumption and 80% of GHG emissions, both due to "lifestyle" generated in the last 150 years and changes associated with urbanization process in the era of globalization and urban sprawl. The specialized literature has devoted many efforts to analyze the contribution of urban systems to climate change, occupying the Urban Heat Island (UHI) an important place in studies on urban climate. In this sense, the use of remote sensing technology has allowed detailed mapping of (daytime) land surface temperature (LST) for urban and metropolitan systems. These studies have demonstrated the key role played by vegetation, impervious soil and land uses to explain differences in the spatial distribution of LST. However, the information provided by satellites has important limitations: especially the low resolution of the thermal band of night images. MODIS, for example, provides valuable information on the night LST; however, the spatial resolution of the thermal band is about one km², scale clearly insufficient to identify accurately the spatial structure of the UHI. In the opposite site, Landsat offers a more accepTabla spatial resolution (30 m² / pixel for the visible bands of the electromagnetic spectrum as well as 60 to 100 m² / pixel in the thermal bands), but does not provide information about night soil temperature. In addition, it is at night when the urban heat island becomes more evident. Therefore, to determine the night LST in an appropriate scale (as offered by Landsat) remains a significant challenge in studies aimed at identifying the spatial structure of the UHI. In Metropolitan Area of Barcelona (AMB, 3,200 km² and 4.8 million inhabitants), the (day) highest temperatures are not in the CBD but in areas that are more peripheral specialized in economic activity such as industrial parks, producing a "donut" in the spatial distribution of the LST. Bare soil also shows a (day) high surface temperature. In contrast, sprawled areas have a less pronounced LST. The spatial structure of the LST, however, changed significantly during the night: compact and sprawl areas maintain high levels of heat, facing the agricultural soil, which cools more sharply as also happens in the industrial land. UHI appears therefore overnight. This paper aims to show the spatial patterns of Urban Heat Island in the Metropolitan Area of Barcelona at medium scale (30 m² / pixel). It develops a new methodology aimed at modeling the night temperature at one km² resolution (MODIS) and then extrapolating this methodology to a most accurate scale of 30 m² / pixel (Landsat). The study allows identifying differences in (night) LST according to the distribution of land use, quantity and quality of the vegetation, intensity of urban sprawl, spatial distribution of economic activity and type of urban morphology (continuous vs. scattered urbanization).