UNDERSTANDING URBAN HEAT ISLANDS

Introduction

Cities influence most of the basic weather variables which have an effect on humans, such as temperature, humidity, wind, precipitation, or sunshine. Modifications to temperature have received greatest attention because of practical implications for human comfort, morbidity, mortality, air pollution, ecology, or energy use with its potential link to the generation of greenhouse gases. It is therefore unsurprising that the urban heat island (UHI) phenomenon, which characterizes sustained urban warmth compared to its rural surroundings, has established itself as an iconic feature of the urban climate. A seemingly easy to define, understand, and measure phenomenon, this most studied of urban climate effects, however, presents intricate nuances and complexities which are often overlooked by researchers and practitioners not immersed in the field when applying urban climate knowledge in the climate sensitive design of buildings and the planning of cities. The UHI has particularly important implications for cities located in hot (sub)tropical climates where it is generally non-beneficial and mitigation of such local anthropogenic warming with all its undesirable consequences is most pressing.

Concepts

Urban climate scales and vertical structure of urban atmosphere

Understanding urban climate scales is fundamental for characterizing, observing, and modeling UH Is. The atmosphere contains the full spectrum of atmospheric scales forming a continuum from the smallest to the largest. Urban phenomena are part of this complete spectrum and scales range from the microscale (tens of meters, e.g. related to individual urban facets such as walls, roofs, or leaves) to the mesoscale (tens of kilometers, e.g. entire city or urban forest).

A number of distinct layers can be identified within the urban boundary layer (UBL) which refers to that part of the planetary boundary layer (1’BL) which is directly affected by the presence of a city (Figure 11.1). The UBL forms as an internal boundary layer at the rural—urban upwind edge and is advected downstream by the regional wind as a layer of urban-modified air or urban plume. From the bottom upwards the urban canopy layer (UCL) exists below mean height of roughness elements (buildings, trees, etc.). Here all atmospheric variables are spatially highly variable and appear to be specific to a unique location. UCL effects reach beyond roof level in a highly

Figure 11.1 Idealized vertical structure of the urban atmosphere over (a) an urban region at the scale of the whole city (mesoscale), (b) a land-use zone (local scale), and (c) a street canyon (microscale). Gray shaded areas (thick line following surface in (c)) show ‘locations' of the three UH I types corresponding to each scale (see Table 11.1)

Source: Modified after Oke (2006)

turbulent roughness sublayer (RSL) which extends to two to five times the height of roughness elements. In this layer the flow responds to individual roughness elements and individual plumes of air exist. Flow and scalar fields are therefore three-dimensional. The inertial sublayer (ISL) separates the RSL from the mixing layer (ML) and represents a layer where contributions from individual surface elements are uniformly mixed (or blended). Together the RSL and ISL comprise the surface layer (SL) which makes up about 10 percent of the UBL. Here the effects of the city, such as surface heating and cooling, are most intense. The properties of the SL are mixed into the overlying ML by mesoscale processes. Urban climate impacts extend beyond the borders of a city through the advection of atmospheric properties in the form of an urban ‘plume’.

Estimating the urban effect

Much of urban climate research focuses on determining the urban contribution to a specific climate phenomenon. This is particularly true in the case of the UHI for which it is necessary to separate urban from non-urban effects. Lowry (1977) provides a helpful framework to assess the magnitude of the urban effect which was subsequently modified by Oke et al. (2017). A measured variable is assumed to be composed of (1) its background value (for a flat-plane and representing the regional climate) and the respective departures from its background value due to effects from (2) landscape or local climate (e.g. from relief, local water bodies) and (3) human activities (including urban effects). For a station located in an urban area the ensemble of values measured defines the actual urban climate with (3) as a major additive term representing the urban effect which needs to be isolated. (3) is only insignificant if the station is located in a natural area (i.e. not affected by human intervention) and outside of the urban influence.

In most cases determination of the ‘true’ urban effect is not possible because observations from prior to the urban settlement do not exist. In this case Lowry (1977) suggests four surrogate approximations, which are (i) urban—rural differences stratified according to synoptic conditions, (ii) upwind—downwind differences (sometimes used to assess precipitation processes), (iii) urban—rural ratios, and (iv) weekday—weekend differences (measuring the particular weekly anthropogenic cycle which is uniquely related to urbanization). The majority of UHI studies are based on urban—rural differences, i.e. the comparison between ‘urban’ and ‘rural’ locations to estimate the urban contribution. As noted above, however, it cannot be assumed that the rural observations taken in the vicinity of an urban area represent pre-urban conditions and/or are free of human influences, i.e. (3) is zero, because ‘rural’ may also include managed landscapes such as farms and forests. This places a particular emphasis on the selection and proper description of the reference site during the planning, design, and analysis of a study to ensure that it is free from urban influences. Alternatively the reference site may serve a particular research objective which does not necessarily include comparison with a non-urban surface. A classification to help overcome some of these issues and facilitate the transferability of UCL results from one city to another is introduced in the next section.

Local climate zones and definition of UHI

Urban landscape typing according to the important urban climate controls, namely, fabric, land cover, structure and metabolism, provides a useful approach to classify discrete areas of cities with distinct climate modifications. The Local Climate Zone (LCZ) scheme is such a classification of urban and rural sites for temperature studies (Stewart and Oke 2012). The descriptors used to characterize each zone define the degree of (im)permeability, roughness, thermal properties, and anthropogenic heat. LCZs of the built series are classified according to two built densities (compact and open), three classes of building heights (low-, mid-, and high-rise), and four classes which reflect two types of industrial as well as densely built low-rise and sparsely built areas, respectively. Seven types comprise the natural series reflecting different types and densities of vegetation cover, bare surfaces, and water bodies.

The LCZ scheme has particular relevance for UHI studies and is probably a better way to classify urban neighborhoods than the traditionally used land-use classes (e.g. residential, commercial, or industrial) which is not very meaningful in climatic terms. Stewart and Oke (2012) propose to define the UHI magnitude as an LCZ temperature difference, not an ‘urban—rural’ difference. Since the classification covers both built and natural ecosystems it allows for a more refined description of the ‘rural’ reference or background site. Defining UHIs across LCZs is therefore less prone to confusion. It allows for more meaningful results that can be more easily compared across geographic regions. This is because both site location and respective temperature differences use the same unambiguous framework that reflects the actual ability of the surface to modify the local climate.

UHI genesis: urban-rural energy balance differences

Urban structures, materials, land cover, and human activity result in surface energy fluxes which greatly differ in a city compared to those of surrounding rural areas. The surface fluxes are

Figure 11.2 Typical urban—rural surface energy balance differences exemplified by ensemble mean suburban—rural differences observed in Greater Sacramento, 1991 for (a) storage, (b) latent heat, and (c) sensible heat fluxes. An open low-rise suburban and a wet rural grassland site are used for the comparison

Note: differences are positive when the flux is larger at the suburban site

Source: Modified after Oke et al. (2017)

responsible for the atmospheric state and the UHI at the surface and in the atmosphere must therefore be the result of urban/rural surface energy balance (SEB) differences. In an urban area the SEB balances net radiation flux density and anthropogenic heat flux with sensible heat flux density, latent heat flux density and the conduction (storage) of heat into the ground. Also included is the net horizontal flux due to advection. The presence of a city alters all individual SEB components.

Because of generally lower latent heat fluxes in the city as a result of lack of moist vegetation, heat during daytime is preferentially channeled into sensible forms (sensible and storage heat fluxes) which results in a warming of the environment (Figure 11.2). In particular the storage heat flux is significant and often considerably larger in an urban area than its rural surroundings. Key characteristics that influence the size of this flux are the surface materials and the urban structure. The mass of the building fabric presents a large reservoir for heat storage because urban surface materials have good ability to accept, conduct, and diffuse heat. Sensible heat is usually transported into the building volume in the morning and by mid-to-late afternoon it is transferred back to the surface and released into the atmosphere. This helps to maintain a positive sensible flux in cities in the evening and at night contributing to the heat island in the air. The anthropogenic heat flux component is a relatively small fraction of the net radiation input over lesser built-up areas (e.g. low-rise residential neighborhoods) but can be significant in city centers where possibilities for large heat emission from buildings and traffic are amplified.

Differences in urban and rural SEB components generate differences in urban and rural cooling and warming rates at the surface, in the substrate, and in the air generating the four main UHI types as discussed below.