Climatic performance of urban textures: Analysis tools for a Mediterranean urban context

Highlights

• Urban morphology has the highest impact on UHI intensity in Mediterranean climate.

• UHI intensity can be estimated based on three descriptors of urban morphology.

• Climate performance of urban textures varies in summer time and winter time.

• Compactness and vertical density are key morphological features for UHI assessment.

Abstract

Urban heat island effect is almost always neglected in building energy simulations, due to difficulties in obtaining site-specific climate data with a district-scale resolution. This study aims at filling this gap for the Mediterranean urban context, presenting a set of tools to estimate the climatic performance of urban fabric at the local scale.

The results are based on climatic analysis conducted in Rome (Italy) and Barcelona (Spain) with the Urban Weather Generator (UWG) model, validated using temperature measurements taken in urban meteorological stations. Parametric analysis of the UHI intensity was performed considering five key variables: urban morphology, vegetation cover, anthropogenic heat from buildings, anthropogenic heat from traffic and albedo. The results show that the variability of urban morphology has the major impact on urban temperature. Two robust relationships between three morphology descriptors of urban fabric and UHI intensity were established applying multiple regression analysis. Such relationships indicate that both the horizontal and the vertical density of buildings play a major role on the temperature increase in urban areas.

Easy-to-use graphical tools have been provided to compare the climate performance of different urban textures and to estimate the average UHI intensity variability in Mediterranean cities.

Introduction

Heat vulnerability in urban areas is one of the most concerning impacts of climate change in hot climates [1]. The increase in frequency of strong heat waves observed in Europe and the Mediterranean region since 2000 has caused severe impacts on population's health and comfort [2], [3], [4]. Mediterranean urban environments are particularly prone to temperature increase because of some inherent characteristics, such as lack of humidity, scarce vegetation, high building density and presence of anthropogenic heat sources, which are responsible of the so-called ‘Urban Heat Island’ (UHI) effect. In this context, the combination of heat waves and UHI causes a significant increase of buildings cooling demand and electricity consumption [5], [6], [7], which already accounts for the biggest share of the carbon emission related to the building stock [8]. For this reason, the impact of UHI intensity on building energy consumption in urban environments should be carefully considered to effectively meet the target of CO₂ emission reduction set for the next decades [9].

The increase of temperature in urban areas is due to anthropogenic sources of heat, such as the buildings' cooling and heating systems, and to mechanical and thermal phenomena determined by the interaction of urban surfaces with the lower layers of atmosphere. The impact of built environments on air temperature increase has been computed coupling mesoscale meteorological models with building energy models [10], [11], [12], [13]; these studies showed that the waste heat from buildings’ systems and the albedo of surfaces play a fundamental role on the increase of urban temperature. On the other way round, the increase of urban temperature affects the energy performance of buildings in urban areas. Recently, new methodologies for predicting buildings energy loads at a district and an urban scale have been proposed [14], [15], [16], [17], [18], [19]. Most of these take into consideration the impact of microclimatic modifications produced by built environments, such as the UHI effect. The results indicate that the energy performance of buildings is influenced by the urban form and building density, which modify both urban temperature and indoor solar gains [20], [21], [22], [23], [24], [25].

The influence of urban geometry on UHI intensity has been investigated in several experimental and numerical studies [26], [27], [28], [29], [30]. However, the findings of these kind of studies are not of help to obtain urban climate data with a local scale resolution, because they are normally based on a sole geometrical parameter, namely the canyon “aspect ratio”—the ratio of the buildings height (H) to the width (W) of the road—or the canyon “Sky View Factor” (SVF). These dimensionless parameters alone are not fully adequate to characterize the three-dimensional features of different urban morphologies, because to same values of SVF or H/W ratios, different urban configurations may exist (Fig. 1). Other studies on urban morphology demonstrated in fact that more than one metric is needed to understand the urban typologies associated to a certain density [25], [31], [32]. The dimensions and the shape of buildings are fundamental variables to understand the interaction between atmosphere and urban surfaces that determines the local urban climate; therefore, a more suitable set of morphology descriptors should be identified to quantitatively assess the climate performance of urban textures.

This work aims at filling this research gap for the Mediterranean context, introducing a set of analytical and graphical tools capable of predicting the climatic performance of different urban textures based on relevant morphological descriptors. This is aimed at providing easy-to-use analysis tools for designers and planners to perform urban climate analyses at the local scale and to adjust weather files for building energy modelling so as to include the UHI effect in urban areas. To this aim, the objective of this study is threefold. First, to validate the climate model ‘Urban Weather Generator’ in the Mediterranean context. Second, to identify quantitative relationships between a set of descriptors of urban morphology and the estimated UHI intensity during winter season and summer season at the local scale. Third, to provide graphical tools to compare easily the climatic performance of different urban textures in a Mediterranean urban context.

The UHI and its impact on building energy performance have been widely investigated in previous studies [33], [34], [35]. In the Mediterranean zone, the maximum UHI intensity has been found to vary between 2°C and 10°C or even more [34], [36]. Many experimental and numerical studies have been carried out in Rome and Barcelona, two of the largest urban areas of the Mediterranean basin. The results showed a variation of the average maximum UHI intensity between 2°C and 5°C [37], [38], [39], [40], [41], determining an increase up to 57% of the cooling demand of residential buildings [37], [42], [43]. The impact on non-domestic buildings (e.g. university) is instead around 10% in this climate [44], [45]. The negative impact of UHI intensity on cooling loads and cooling potential of natural ventilation has been highlighted also in colder climates, such as in London, UK, or in Basel, Switzerland [6], [46], [47].

Other studies on the UHI highlighted that the spatial distribution of urban temperature is not uniform, the latter being dependent mainly on the density of buildings [48], [49]. This fact can be explained by considering three typical phenomena that characterize densely built environments: (1) solar radiation absorption is increased because of multiple reflections between the surfaces of the urban canyons [50], (2) turbulent sensible heat transfer out of the canyon is reduced due to the building proximity that decreases wind speed [51], [52] and (3) long-wave radiation loss from within the canyon is reduced due to the screening by the flanking buildings [28], [46].

An analytical relationship between urban geometry and UHI intensity was firstly presented by Oke [53], who found a logarithmic relationship between the increase of UHI intensity and the increase of the canyon aspect ratio. Many other empirical relationships have also been identified for different cities and climates, as presented in the reviews by Unger [50], [54], [55], [56]; the results are however quite inconsistent and hardly comparable, due to the different climates, urban characteristics and observational periods.

Oke also proposed a practical methodology to link urban climatology analysis to urban planning, introducing the concept of ‘Urban Climate Zones’ (UCZ) [57], successively evolved in ‘Local Climate Zones’ (LCZ) [49]. The LCZ method divides the urban fabric into homogeneous categories in terms of built types, land cover types and land use, with the aim of predicting intra-urban temperature patterns and UHI intensities. Built types are classified in ten categories, comprehensive of compact, open and sparse developments with different building height (high-rise, midrise, low-rise). Recently, effective methodologies to produce LCZ maps using data from remote sensing images and GIS databases have also been presented [58], [59], [60]. However, the classification of built types in the LCZ method has limitations regarding the possibility to cover the many building configurations typically observed in real urban environments [61].

To estimate the UHI intensity at the district scale, many urban energy balance models (UEB) [62], [63] and computational fluid dynamic models (CFD) have also been proposed. UEBs simulate the urban energy fluxes at the local scale based on the building-air volume balance defined by Oke [64]. They use incoming shortwave and long wave radiative fluxes, air temperature, specific humidity, wind components and anthropogenic heat flux as forcing conditions and model the outgoing radiative fluxes, turbulent sensible heat flux, turbulent latent heat flux and net heat storage flux of a given urban system. In these models, urban morphology is described using different sets of parameters, varying from model to model [62].

CFD models can be much more accurate than UEBs, because they solve the governing equation of fluid motion with high spatial resolution [26]. Many studies on the impact of urban morphology on urban microclimate have been developed with CFDs [65], [66], [67], [68]. Procedures for coupling building energy models with CFD models or urban canopy parametrizations have also been proposed to account for the reciprocal interaction between indoors and outdoors in urban contexts [12], [26], [69], [70], [71], [72].

UEB and CFD models thus represent an important progress for climate analysis in urban areas. Nevertheless, they have had poor application in urban planning and building design. CFD tools and coupled models can provide high-resolutions microclimate data for short times, but presuppose expert knowledge and suffer from very long calculation times, which limit their adoption especially for analysis at the district scale [73]. UEB models are less sophisticated but they also need a basic knowledge of urban climate physics to be used properly; also, their application is limited by the fact that the output data—i.e. urban energy fluxes or surface temperatures—are not directly usable to perform building energy analysis. In fact, this lack of communication between tools, disciplines and expertise remains the main limitation for an integrated assessment of the building energy performance in urban environment.

A step forward was done with the development of the ‘Urban Weather Generator’ (UWG) [74]. UWG is based on the UEB model ‘Town Energy Balance’ [63], [75], including a detailed building energy model [76], [77] to account for the reciprocal interaction between buildings energy performance and urban climate (i.e. heat transfer through building fabric and waste heat from heating, ventilation and air conditioning systems). UWG uses hourly data measured at rural meteorological stations as forcing conditions to calculate urban energy fluxes and to generate urban weather files that capture the UHI intensity for a reference urban area. The model consists of four calculation components: the “Rural Station Model" (RSM), the "Vertical Diffusion Model" (VDM), the "Urban Boundary-layer model" (UBL) and the "Urban Canopy and Building Energy Model” (UC-BEM)”. The model workflow is as follow: the input weather data are used by the RSM to calculate the rural sensible heat flux, which is used by the VDM to calculate the vertical profile of air temperature above the rural weather station. These data, along with the urban sensible heat flux calculated by the UC-BEM, are used by the UBL to calculate air temperature above the urban canopy layer. The UC-BEM calculates urban sensible heat flux and urban canyon air temperature and humidity from radiation and precipitation data, air velocity and humidity measured at the weather station and from the air temperature calculated by the UBL. The equations and interrelations between models are described in detail in the relevant publication [74]. The main advantages of this model with respect to others are: (1) short computation time, (2) use of hourly weather files to force the calculation, which allows for an assessment of the hourly variability of UHI intensity over a year and (3) direct usability of the output weather files to run building energy performance simulations with EnergyPlus. To perform simulations with UWG, two files are needed: a rural weather file and an XML file which describes all the features of the urban area that affect the phenomenon, urban morphology included [74]. The first version of the model (V 1.0) has been validated in Boston, Basel and Toulouse, showing an average root mean squared error (RMSE) of about 1 K [74], [78]. Updated versions of the model have recently been validated in Singapore and Abu Dhabi [79], [80].

UWG appears to be particularly suitable for comparative analysis of the climatic performance of the urban textures as it considers a morphology parameterization based on three independent descriptors: the “site coverage ratio”, the “average building height” in the area and the “façade to site ratio”. This set of parameters provides quantitative details on building density and qualitative information on the urban typology (this point is discussed in detail in Section 4). Results from this model could thus be used for quantitative comparisons of the climatic performance of different urban textures as well as to derive qualitative insights into the relationship between urban morphology and climate performance at the local scale.