Transforming typical hourly simulation weather data files to represent urban locations by using a 3D urban unit representation with micro-climate simulations

Leonidas Bourikas; Patrick James; AbuBakr Bahaj; Mark Jentsch; Tianfeng Shen; David Chow; Jo Darkwa

doi:10.1186/s40984-016-0020-4

Transforming typical hourly simulation weather data files to represent urban locations by using a 3D urban unit representation with micro-climate simulations

Bourikas, Leonidas; James, Patrick; Bahaj, AbuBakr; Jentsch, Mark; Shen, Tianfeng; Chow, David; Darkwa, Jo 2016-07-08 00:00:00 Urban and building energy simulation models are usually driven by typical meteorological year (TMY) weather data often in a TMY2 or EPW format. However, the locations where these historical datasets were collected (usually airports) generally do not represent the local, site specific micro-climates that cities develop. In this paper, a humid sub-tropical climate context has been considered. An idealised “urban unit model” of 250 m radius is being presented as a method of adapting commonly available weather data files to the local micro-climate. This idealised “urban unit model” is based on the main thermal and morphological characteristics of nine sites with residential/institutional (university) use in Hangzhou, China. The area of the urban unit was determined by the region of influence on the air temperature signal at the centre of the unit. Air temperature and relative humidity were monitored and the characteristics of the surroundings assessed (eg green-space, blue-space, built form). The “urban unit model” was then implemented into micro-climatic simulations using a Computational Fluid Dynamics – Surface Energy Balance analysis tool (ENVI-met, Version 4). The “urban unit model” approach used here in the simulations delivered results with performance evaluation indices comparable to previously published work (for air temperature; RMSE <1, index of agreement d > 0.9). The micro-climatic simulation results were then used to adapt the air temperature and relative humidity of the TMY file for Hangzhou to represent the local, site specific morphology under three different weather forcing cases, (ie cloudy/rainy weather (Group 1), clear sky, average weather conditions (Group 2) and clear sky, hot weather (Group 3)). Following model validation, two scenarios (domestic and non-domestic building use) were developed to assess building heating and cooling loads against the business as usual case of using typical meteorological year data files. The final “urban weather projections” obtained from the simulations with the “urban unit model” were used to compare the degree days amongst the reference TMY file, the TMY file with a bulk UHI offset and the TMY file adapted for the site-specific micro-climate (TMY-UWP). The comparison shows that Heating Degree Days (HDD) of the TMY file (1598 days) decreased by 6 % in the “TMY + UHI” case and 13 % in the “TMY-UWP” case showing that the local specific micro-climate is attributed with an additional 7 % (ie from 6 to 13 %) reduction in relation to the bulk UHI effect in the city. The Cooling (Continued on next page) * Correspondence: L.Bourikas@soton.ac.uk Energy & Climate Change Division, Sustainable Energy Research Group (SERG), Faculty of Engineering and the Environment, University of Southampton, Southampton SO17 1BJ, UK Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Bourikas et al. Future Cities and Environment (2016) 2:7 Page 2 of 23 (Continued from previous page) Degree Days (CDD) from the “TMY + UHI” file are 17 % more than the reference TMY (207 days) and the use of the “TMY-UWP” file results to an additional 14 % increase in comparison with the “TMY + UHI” file (ie from 17 to 31 %). This difference between the TMY-UWP and the TMY + UHI files is a reflection of the thermal characteristics of the specific urban morphology of the studied sites compared to the wider city. A dynamic thermal simulation tool (TRNSYS) was used to calculate the heating and cooling load demand change in a domestic and a non-domestic building scenario. The heating and cooling loads calculated with the adapted TMY-UWP file show that in both scenarios there is an increase by approximately 20 % of the cooling load and a 20 % decrease of the heating load. If typical COP values for a reversible air-conditioning system are 2.0 for heating and 3.5 for cooling then the total electricity consumption estimated with the use of the “urbanised” TMY-UWP file will be decreased by 11 % in comparison with the “business as usual” (ie reference TMY) case. Overall, it was found that the proposed method is appropriate for urban and building energy performance simulations in humid sub-tropical climate cities such as Hangzhou, addressing some of the shortfalls of current simulation weather data sets such as the TMY. Keywords: Idealised urban unit model, Micro-climate simulations, Urban weather projections, Cities Introduction locations within the city develop local, site specific Engineering and urban design practise often relies on micro-climates that will not always be represented by thermal simulation modelling in order to achieve regula- these datasets [6]. tory compliance and support decision making on urban Over recent decades many cities have expanded rapidly and building design, as well as sizing of building energy through higher building densities, larger building heights systems that directly effects energy consumption. Urban and growth towards suburban and rural areas. This expan- and building energy simulation models are usually sion results in changes to the balance of the urban energy driven with hourly weather datasets for a ‘typical’ year budget [7] which consists of the radiation, sensible, latent [1]. In humid sub-tropical climates where air-conditioning and anthropogenic heat fluxes and the roughness caused is the primary concern the year format is that of an “aver- by the built environment [7]. These features of the built age year” (eg TMY2 or EPW). Common typical weather environment’s energy and mass equilibrium can cause at- year formats have the form of a synthetic year which is mospheric forcing which alters the local weather condi- compiled from multiple years to be representative of the tions [8] and contributes to the development of distinctive climatic conditions for the site of interest [2]. local micro-climates specific to a site’s characteristics and Heating, Ventilation and Air Conditioning (HVAC) morphology. The micro-climatic factor with the most no- systems are often oversized in order to reduce design ticeable variation regarding changes in the energy budget risk by ensuring that the system will cope with loads is the air temperature [9]. The urban air temperature (T) above the average year’s design day estimations [3]. This difference to the regional non-urban air temperature (eg affects the energy consumption of the building, potentially airport sites) is defined as the urban heat island (UHI) and the thermal comfort of the users and the indoor air qual- the magnitude of this difference is known as the urban ity, whilst at the same time increasing the cost of oper- heat island intensity (UHII) [10]. ation and possibly maintenance for building services [4]. Field observations and experimental measurements The main concern in the case of building energy sys- have been used by various researchers in order to create tems’ oversizing is the way that energy use is estimated, algorithms for the adaptation of weather data to local more specifically the heating, cooling, dehumidification specific morphologies [11–13]. In Hong Kong, urban and mechanical ventilation loads. At the moment, in many heat island intensity observations were used in order to countries, weather data files with a typical meteorological adapt a weather data file and use it in cooling load simu- year (TMY2 or EPW) format are used for building regula- lations [14]. In a field study, Wong et al. [15] pointed tion compliance calculations [5]. However, compliance to out that density, the ratio of building height to building the regulations does not necessarily ensure occupants’ sat- footprint area and especially the green surface area isfaction or represent the real operational conditions of (vegetation) can substantially influence the development the building services. One important consideration for the of the micro-climate. Summertime temperature observa- use of the typical year weather time series in energy con- tions in London suggest additional correlations of UHI sumption predictions and simulations is the representa- intensity with the distance from a thermal hotspot and tiveness ofthelocationwhere thesourcedatahave wind direction [16]. been collected. Many of these locations are, due to histor- More recently, a model was introduced for the estima- ical data availability, airports near large cities. However, tion of the UHI intensity in urban centres and the Bourikas et al. Future Cities and Environment (2016) 2:7 Page 3 of 23 modification of weather data files in the UK [17]. This for the prediction of monthly changes due to the UHI model is based on the statistical analysis of temperature effect in air temperature, RH and wind speed [25]. The data from UK based weather stations from the period Town Energy Balance and an integrated building energy 1961 to 2006. The urban thermal centres were defined model have been coupled with other schemes (ie rural by assigning thresholds to predefined urban fraction pa- station, vertical diffusion and urban boundary layer rameters for each UK grid cell and its surroundings [17]. model) and they have been used for the adaptation of The urban fraction limits were calibrated with visual in- rural air temperature and relative humidity hourly values spection of the fit of calculated ranges for the urban to the urban heat island effect [26]. The so-called Urban fraction parameters to real UK cities [17]. Therefore, in Weather Generator requires a large number of initial order to apply this methodology in places other than the parameters that vary from construction elements pro- UK, this method should be repeated for similar or perties and characteristics (eg albedo and initial equivalent local land cover types and temperature data temperature) to the urban (eg façade to site ratio) and spatial resolution under the same baseline assumptions reference rural site morphology [27]. The land cover of urban fraction limits. analysis for the creation of the simplified “urban unit In a numerical weather prediction approach, the model” can provide most of the required parameters for Urban Canyon model [18] estimates the heat fluxes and the initialisation of the Urban Weather Generator. The the air temperature distribution in an urban street can- use of the 3D “urban unit model” in numerical simula- yon configuration. It introduces a zonal model that di- tion modelling was chosen instead as it can provide an vides the air volume between the buildings into cells approximation of the vertical development of air where the heat and mass balance can be estimated for temperature and relative humidity in the roughness sub- each cell [18]. The results have been used to introduce layer and it can capture the majority of the transient the climatic severity index (CSI) that can assess the im- physical processes at street level. pact of the street canyon morphology on the heating The aim of this study is to provide a comprehensive and cooling demands of locally based buildings [18]. methodology that allows for an integration of urban In a very interesting development, meso-scale weather micro-climate conditions into standard weather datasets forecasting and numerical weather prediction-CFD such as the TMY. In order to achieve this, this paper in- models were coupled in order to predict the micro-scale troduces an idealised “urban unit model” (on a 250 m weather development in wind farms and urban environ- radius) that represents the main thermal and morpho- ments [19, 20]. In these coupling schemes, the numerical logical characteristics of urban sites at street level in the weather prediction models were downscaled in order to neighbourhood scale. This model, which was produced provide the initial and boundary conditions for the using statistical land cover and urban morphology ana- micro-scale models [21]. Despite the benefits from coup- lysis, can be used with simulations as a method of adapt- ling meso- to micro-scale models, the differences in the ing commonly available weather data files to a local horizontal and vertical scales and the time resolution be- specific micro-climate. This methodology is used to tween these models pose a challenge to their widespread adapt air temperature and relative humidity (RH) from successful application. The grid size differences and sub- the TMY file to the effects of local site specific morph- scale grid implications regarding the physical processes ology on urban weather development. The “urban wea- at play are often a threat to the analysis of the results ther projections” resulting from this adaptation were and their correct physical interpretation [22]. In then used for comparing the degree days amongst the addition, numerical models are computationally de- reference TMY file, the TMY file offset for a bulk hourly manding in terms of both power and time. Therefore, UHI intensity (local) and the TMY adapted for the urban their use in micro-scale modelling is usually restricted to weather projections (local + micro). Finally, the TMY file small domains and relatively simple geometries [23]. generated by this methodology was used for the dynamic A more common approach is the coupling of numer- thermal simulation of heating and cooling loads in a do- ical weather prediction models with analytical micro- mestic and a non-domestic building scenario. The po- scale models. The analytical urban canopy models esti- tential improvements in the estimation of building mate the energy balance development and momentum energy consumption were assessed against the business transfer in relation to the urban canopy’s morphology as usual case of applying typical meteorological year data and produce single area-averaged values in order to be files in places with a humid sub-tropical climate such as used as forcing in the scale (typically meso-scale) that the case study city of Hangzhou in China. the weather prediction model resolves [24]. Ren et al. [25] have created a “morphed” TMY file for the city of Methodology Melbourne using a regional weather forecasting model An important issue for the successful application of any coupled with an urban canopy parameterisation scheme existing micro-climatic model are the input data Bourikas et al. Future Cities and Environment (2016) 2:7 Page 4 of 23 prerequisites. Detailed information on the surface mate- Hangzhou (30°15'N 120°10'E) in Zhejiang Province, rials’ properties, morphological and other modelling pa- China [30, 31]. The sensors were installed on lampposts rameters is not however, always readily available. There at a level 3 to 5 m above ground. They logged air is a need for models to estimate the weather conditions temperature at 11-Bit (0.0625 °C) resolution and relative within urban areas as a function of time and urban humidity at 12-Bit (0.04 %) resolution [32]. The manu- morphology [28, 29]. The “urban unit model” introduced facturer stated air temperature accuracy is +/-0.5 °C and above has, therefore, been designed to be as general as the RH accuracy is +/- 5 % RH [32]. For calibration pur- possible in order to facilitate widespread use in building poses, air temperature (°C) and RH (%) readings within thermal and urban design simulations. Ideally, a visual 1 min intervals were compared to the readings of two evaluation (or an automated GIS platform) would be separate thermocouples in an environmental chamber at used to decide the urban class for the site of interest. -10, 0, 10 and 40 °C [33]. All sensors operated within the This would be enough to enable offsetting of the refer- reported accuracy margin and fitted well to the thermo- ence (TMY or real time non-urban) hourly air couple measurements [33]. temperature (T) and relative humidity (RH) for the se- According to Oke [34] air temperature and relative lected urban class for different seasonal weather forcing. humidity observations in the surface layer (3x above A small size neighbourhood was selected as the scale average building height) can be expected to be represen- of interest. An “urban unit” has been defined as an area tative of an area ranging from 100 m to several hundred forming disk (with a radius of 250 m) around the centre meters in a direction upwind and around each sensor. It of a neighbourhood where a temperature and relative is to be expected that most of the sensors were collect- humidity sensor was located. Air temperature and rela- ing measurements in the roughness sub-layer and not in tive humidity were monitored at 26 urban sites in the surface layer [35] (Table 1). The location of the Table 1 Siting of selected sensors in Hangzhou and representation of the size of the “urban unit” (Top Right) Siting of the sensor (blue bullet point) Circle of influence Sensor on a lamppost close to a tall building wall Sensor on a lamppost next to a road Bourikas et al. Future Cities and Environment (2016) 2:7 Page 5 of 23 measurement sites was carefully selected to have as lower than the group average air temperature and their homogeneous characteristics as possible for a large city vegetated surface area is large with an early peak within and the sensors were installed a reasonable distance the first 100 m from the sensor (Fig. 2). Interestingly, away from sites of noticeable surface type change [30]. sensor 12 (dark cyan; -1.0 °C) which also shows a nega- The urban unit’s size, ie the radius of the disk around tive temperature departure from the group mean has a the “centre of a neighbourhood”, has been determined small vegetated area approximately 150 m away from by assessing the vegetation cover’s influence on air the sensor, comparable with those at locations with temperature. For this, twelve urban sites were selected higher than the group average temperature. Its negative across the city centre of Hangzhou, North of the Qiantang trend can, however, probably be explained with the steep River (Fig. 1). rise in vegetated surface area at the annular area from Each sensor was considered to be the centre of con- 150 to 200 m, showing that the influence of vegetation centric circles at radii of 10, 25, 50, 100, 150, 200, 300, remains strong at this distance (Fig. 2). 400 and 500 m. The footprint of the vegetated surface Further evidence comes from the comparison between was estimated and apportioned to these annular areas. sensor 1 (black, +0.5 °C) and sensors 2, 5 and 9 (pink, The analysis was carried out for 14 weeks of hourly data red, gold; +0.2 °C). The regression trend lines indicate collected during summer 2013. The air temperature that the location where sensor 1 resides is warmer than (Tair) in Fig. 2 is the average air temperature departure the locations of sensors 2, 5 and 9 despite the larger veg- from the mean temperature of the 12 sensors in this etated surface within a 0 to 100 m radial distance from group for this 14 week period. The vertical shaded refer- the sensor. In the case of sensors 2, 5 and 9 the vege- ence line marks the 250 m radius from the centre (ie the tated surface area peaks occur later at distances from sensor). The legend shows the mean summer air 100 to 150 m showing the persistent impact of the temperature departure from the group’s mean and the green-space area. In addition, site 5 (red; +0.2 °C) has a goodness of fit of the non-linear regression line in similar vegetated surface area to site 1 (black; +0.5 °C) parentheses. with the only difference being a delay of the peak, that is It is to be expected that the influence of green-space is seen at distances 50 to 100 m farther out. Sensor 11 larger when closer to the sensor and that it diminishes if (dark blue, -0.2 °C) has a lower air temperature than the it is more towards to the outer annular areas. This is group average but the green-space percentage peak oc- shown in Fig. 2. The strong influence of vegetated areas curs after the 250 m radius border. However, in this case close to the sensor is evident from the first peak of a the low air temperature is mainly attributed to the site’s vegetated surface area in comparison to the general proximity to a large wetland (ie Xixi). Based on these re- trend in the group of sensors. The locations of sensors 3 sults a representative circular “urban unit” has been de- (dark green; -1.4 °C) and 10 (purple; -0.4 °C) have a fined with a 250 m radius. It is expected that the total Fig. 1 Location of the 12 sites for the urban unit size assessment (Left) and a typical fisheye image used for estimating the Sky View Factor in Hangzhou (Right). (Note: The Mantou Mountain, National Principle Weather station’s location (reference, typical meteorological year file source) is marked as NP. The colour scheme of the bullet points in the map is consistent with Fig. 2 Bourikas et al. Future Cities and Environment (2016) 2:7 Page 6 of 23 Fig. 2 Regression lines of the percentage vegetated area in each annular area on the distance from the centre (ie the sensor) area of about 200,000 m (250 m radius) surrounding Step 1. A simplified model for urban micro-climatic the sensor will be the representative part of the source simulations area for the air temperature and relative humidity signal. The generation of the “urban unit model” is based on The size of the proposed urban unit also agrees with the following steps: other authors’ studies on the windward distance from a point of roughness or thermal change (~200–500 m) (1)The land cover analysis starts with the inspection and the internal boundary layer extent in zones within and selection of an aerial image of the sensor’s local climate classification schemes (r ~ 200–500 m) [36]. location in Google Earth [37]. The image with the Building on the urban unit as defined above, this study best orthographic projection and quality is selected adopted a combination between a local urban classification (ideally, a clear top plan view image). It is then scheme and the simulation of the local specific weather processed with SketchUp [38], a computer aided development for urban unit layouts in the case study re- design software, using the urban unit described gion with a residential/institutional (university, college) above with the data logger at the centre of the use. Micro-climatic simulation modelling was used to as- 250 m circular area (Fig. 4). sess the influence of the neighbourhoods’ morphology on (2)A selection of metadata (eg high resolution images the local, street level air temperature. The main advantage taken on site) is used in combination with the aerial of using such a surface classification scheme is that it pro- view to draw polygons for the vegetation (green vides generic input data for the simulation model [36] and colour), water (blue colour) and built-up (black that the simulation results can be attributed back to typical colour) surfaces (Fig. 4). The residual is designated urban morphology characteristics. as other impermeable surface (white colour) and There are three key parts to the methodology for gen- includes street and pavement surfaces. A set of erating the “urban weather projections” that result from morphological parameters is then calculated for each the combined urban classification and simulation model- site including: the mean building height H, the ling (Fig. 3): (Step 1) The creation of the idealised “urban roughness length z , the height to width aspect ratio, unit model” for the sites of interest (in this study 9 sites); the frontal area ratio λ , the building surface fraction (Step 2) The normalisation of the reference weather data F , the impervious surface fraction I and the r r with the local monthly UHI patterns for different wea- pervious surface fraction P . ther forcings and (Step 3); The adaptation of the UHI (3)Each urban site is classified according to the land adjusted hourly air temperature and RH data to account cover analysis into a “Local Climate Zone” following for the effect of the site specific generic morphology at an urban classification scheme developed by Stewart street level at the neighbourhood scale. and Oke [36]. Each zone (ie thermally homogenous These three components of the overall methodology region of uniform surface characteristics) in the are described in the following: scheme exhibits a distinctive diurnal temperature Bourikas et al. Future Cities and Environment (2016) 2:7 Page 7 of 23 Fig. 3 Methodology for adapting TMY files to include the effect of the local site specific morphology in cities development profile at sensor height (~1.5 to 3 m) (4)The urban morphology of these sites is further at the local scale [36]. The resulting Local Climate analysed for 5 annular areas. The annular rings’ Zones (LCZs) describe 17 generic environments periphery has been defined at radii of 50, 100, 150, consisting of 10 zones for built-up (eg open 200 and 250 m. Urban morphological parameters high-rise) and 7 for non-urban land cover types and descriptive statistics are then calculated for all (eg scattered trees) [36]. Each zone is represented by a the annular areas (ie 0–50 m (red), 50–100 m set of ten morphological parameters and a descriptive (orange), 100–150 m (blue), 150–200 m (green), definition of the typical location and use of the urban 200–250 m (outer) as shown in Table 2, for an sites classified into a zone. Nine urban sites out of a example of site 2 in Fig. 5). total of 26 investigated areas in Hangzhou were (5)The generic, idealised “urban unit model” is classified as “Local Climate Zone 5” (LCZ5) (Fig. 5) constructed to have a similar planar area ratio and which denotes midrise buildings at a medium density mean weighted (footprint) building height to the arrangement [36]. This study focuses on these nine nine studied sites. The individual surface energy sites classified as “LCZ5”. balances are represented in the model by the Fig. 4 Digital elevation models and land cover have been built upon the aerial image (Left). Four main land surface types have been identified in the five concentric annular areas with the sensor located at the centre of a disk of 250 m radius (Right) Bourikas et al. Future Cities and Environment (2016) 2:7 Page 8 of 23 Fig. 5 The nine studied urban units classified into LCZ5 (locations, marked with blue bullet points) in Hangzhou (Left). (Note: The Mantou Mountain, National Principle Weather station’s location (reference, typical meteorological year file source) is marked as NP). An example of the land cover analysis is shown for Site 2 (Right; radius 250 m). Buildings are marked with black and vegetation with green pervious, impervious and building footprint surface generic “urban unit model” comprises of square based area ratios. The morphology characteristics of each boxes (ie blocks) with a non-uniform height in a stag- annular area are based on the median value gered irregular array (3 in Table 4). observations from the statistical analysis for the nine The staggered block array (3 in Table 4) has a north– LCZ5 urban units (Table 3). The median was south orientation. Each block has a base equal to the preferred over the mean because it is not affected by computational grid cells’ horizontal dimensions. For ex- extremely low or high values and the calculated ample, the minimum building footprint area in this distributions were rather skewed than normal. study was 64 m because the computational grid cells had horizontal dimensions of 8 m (x) x 8 m (y). Regard- In idealised models building geometry is usually ing the vertical grid dimension (z), any buildings and substituted with arrays of cubes. Common methods use vegetation in the urban unit model had a minimum cubes in staggered or aligned arrays (eg [35], [39–42]). height equal to the height of the first vertical grid cell (ie Cubes in regular arrays are spaced in repeated intervals 0.50 m). Each block can represent a building (black), equal at all directions to the cube’s edge length (ie aspect vegetated surface (green, grass or tree) or water surface ratio = 1) (1 in Table 4). In this study however, the (blue, zero height). The residual space between the Table 2 Analysis of the morphological parameters for each annular area for a site with a LCZ5 classification (site 2 in Fig. 5) Site 2. R-r (m) P I F z σ =H r r r 0 H 0–50 (red) 0.05 0.64 0.31 1.5 0.52 50–100 (orange) 0.11 0.62 0.27 1.3 0.45 100–150 (blue) 0.17 0.56 0.27 1.2 1.08 150–200 (green) 0.16 0.60 0.24 1.7 0.99 200–250 (outer) 0.19 0.56 0.25 2.1 0.87 P pervious surface fraction, I impervious surface fraction, F building surface fraction, z roughness length, σ =H standard deviation of the building height r r r 0 adjusted for the area weighted average height Bourikas et al. Future Cities and Environment (2016) 2:7 Page 9 of 23 th th Table 3 The median and the range (in brackets) of the key morphological parameters (10 to 90 centile) for the nine LCZ5 urban units shown in Fig. 5 Annulus R-r (m) H (m) P (%) I (%) λ (%) F (%) d (m) z (m) r r f r 0 0–50 20 (15–22) 7 (0–15) 72 (50–85) 17 (7–25) 21 (7–30) 6.7 (3.6–11.1) 1.8 (1.3–4) 50–100 18 (13–25) 11 (0–17) 61 (54–70) 15 (6–20) 27 (8–40) 8.9 (3.6–12.9) 1.4 (0.5–3) 100–150 20 (13–24) 15 (0–25) 65 (56–70) 13 (6–20) 23 (4–30) 7.6 (2.0–10.6) 1.8 (0.5–3) 150–200 17 (12–20) 15 (5–25) 60 (50–66) 14 (10–17) 24 (19–28) 8.6 (5.5–9.7) 1.2 (0.7–2) 200–250 19 (16–24) 15 (12–25) 56 (54–66) 13 (10–20) 23 (16–27) 8.3 (5.8–11.3) 1.8 (0.7–3) (R-r refers to the inner and outer radius of the annuli borders) H building footprint-area-weighted average height, P pervious surface fraction, I impervious surface fraction, λ frontal area ratio, F building surface fraction, d r r f r zero plane displacement height, z roughness length blocks represents the impervious surface (grey, eg roads, steps for the normalisation of the reference weather paved areas) (Fig. 4, right). The distance between the data are as follows: building blocks in each annular area is random and the number of the blocks representing buildings and vegeta- (1)((A) in Fig. 6) - The global horizontal solar radiation tion was defined by the estimated F and P ratios re- (GHR) from the typical meteorological year (TMY) r r spectively. The distribution of the blocks in each annular file (584570_CSWD) was used for creating three th area is similar in all notional quarter annuli (ie 1/4 of distinctive groups in order to account for different the total annular area). The changes to the packing weather forcing in each month (Fig. 6). Group 1 density and distance between the blocks produced a ran- represents days with overcast sky conditions and domly dispersed layout that is expected to better fit the rain events. Group 2 represents days with clear skies high spatial inhomogeneity of real cities than a regular and average or “highly likely” weather conditions for staggered cube array. the month, while Group 3 represents rather warm/ th th hot days with a clear sky (Fig. 6). The 25 and 75 st rd Step 2. Normalisation of the reference weather data with centiles (1 and 3 quartiles) of the GHR values the local UHI patterns (local scale) were used as the cut off points for the categorisation The reference weather station (official, NP in Fig. 5) for of the hourly TMY data into the groups. Group 1 Hangzhou is located at Mantou Mountain (30.23 N, for a given month contains all the days in that 120.17 E, at an elevation of 42 m). This weather sta- month that have hourly GHR values equal to or less th tion is the source of the typical meteorological year than the 25 centile GHR value for the respective (TMY) files for Hangzhou and it reports data with hours (lowest 25 % of GHR values). Similarly, Group Hangzhou international airport’s reference code 2 contains all the days in a month with hourly GHR (ZSHC) and the code 584570 in the World Meteoro- values within the interquartile range of the GHR logical Organisation’s (WMO) weather station list. The data for the respective hours (50 % of GHR values). Table 4 Common urban morphology representations and the idealised “urban unit model” in this study (3) Idealised models used to represent urban morphology (1) Uniform height and aspect ratio (2) Variable height and aspect ratio (3) Variable height, aspect ratio and shape of blocks plus water (blue) and vegetation (green) surfaces. Models (1) and (2) in table were adapted from [42] Bourikas et al. Future Cities and Environment (2016) 2:7 Page 10 of 23 Fig. 6 Methodology flowchart of the model’s validation and the generation of the adapted –“urbanised” weather dataset Therefore, Group 2 is expected to be the most (3)((C) in Fig. 6) - The data collected for each day from representative, “highly likely” weather forcing the sensors of the nine studied sites (LCZ5) scenario. Group 3 contains days that had most of highlighted in Fig. 5 as well as 10 additional the hourly GHR values in the upper quartile (highest (“sample”) sites in Hangzhou (Fig. 7) were allocated 25 % of GHR values). Here, the definition of “most to the weather forcing groups as defined in (2) ((B) of the hourly values” in this context relates to days in Fig. 6). Further to this, the observations from the with less than three hours with GHR values that do 10 “sample” sites were used to create a generic hourly not fit into the specific group and where these hours UHI pattern which relates to the difference of the are not between 12:00 and 16:00 h. sample sites’ hourly average observations to the Following the grouping of the days, descriptive reference weather station data for each month and statistics were calculated for the daily mean air weather forcing group. The hourly average temperature, temperature range, daily mean RH and observations from the nine studied sites (LCZ5) were maximum temperature for all the days in each then used to validate this method. Here we are group. The remaining days from the TMY file that comparing real measurements of the 9 LCZ5 sites did not fit into any category were then distributed with the simulation results of the “urban unit model” into either group 1, 2 or 3 according to their forced with measured data from the reference (TMY matches of daily mean air temperature, temperature source) weather station offset by the generic UHI range, daily mean RH and maximum temperature. effect as measured in the 10 “sample” sites. (2)((B) in Fig. 6) - The descriptive statistics (ie mean T, (4)The normalisation of the TMY reference weather T range, mean RH, max T) of (1) were also calculated data with the generic UHI group patterns (T and RH for observations from the reference weather station offsets) was based on a simple offset of the hourly (TMY source) for the period from December 2012 to mean temperature and RH values for each weather December 2013 (Fig. 6). The results were then forcing group (Eq. 1). compared with the descriptive statistics ((A) in Fig. 6) of the weather forcing groups from the TMY data file as determined in (1). Following this, the individual days TRðÞ H urb ¼ TRðÞ H ref Group;hr Group;hr þ UHIðÞ RH offset ð1Þ of the 2012–2013 observations from the reference Group;hr weather station dataset were distributed into the monthly weather forcing groups according to the four Where T(RH)urb is the air temperature (RH) after the criteria (ie mean T, T range, mean RH, max T from adjustment to the hourly UHI (RH offset) pattern, TMY data analysis) in descending order of weighting. T(RH)ref is the reference air temperature (RH) from the Bourikas et al. Future Cities and Environment (2016) 2:7 Page 11 of 23 Fig. 7 Location of the 10 “sample” sites (purple bullet points) used for the assessment of generic UHI patterns in Hangzhou, China in relation to the 9 studied (LCZ5, blue bullet points) sites TMY weather file and UHI (RH offset) is the positive or from the simulations express the weather change at street negative air temperature (RH) offset due to the urban level in relation to the baseline-reference weather (eg wea- heat island effect for each weather forcing group and ther at airport sites or non-urban sites) caused by the ef- hour respectively. fect of the site specific morphology in the city. These projections have the format of an additional hourly offset Step 3. Adaptation of the “localised” TMY data to to the “localised” dataset for each weather forcing group. include the effects of the site specific morphology ENVI-met is a three dimensional non-hydrostatic nu- (local scale + morphology = micro scale) merical micro-climatic model that couples an atmos- The idealised “urban unit model” (250 m radius) intro- pheric, a soil and a one-dimensional (1-D) vegetation duced above was implemented into micro-climatic simu- model and the surface energy balance. The atmospheric lations using a computational fluid dynamics – surface model is based on incompressible Reynolds averaged energy balance analysis tool (ENVI-met Version 4). In this Navier Stokes (RANS) equations [43]. Wind speed and final stage of data processing the hourly “localised” data direction remain constant during the simulation. The generated according to methodology section (2) was used effect of the surrounding urban environment on the to initialise and force the hourly weather conditions in the turbulent kinetic energy and the turbulent energy dissi- simulation. The “urban weather projections” resulting pation rate were modelled using cyclic (periodic) lateral Bourikas et al. Future Cities and Environment (2016) 2:7 Page 12 of 23 and outflow boundary conditions (ie turbulence from Two scenarios were created for the assessment of the last grid cells at outflow boundary are copied to the first solution’s sensitivity to the horizontal grid dimensions; grid cell at the inflow boundary) [43]. Air temperature one for winter that represents cold clear sky conditions (at 2 m above ground) and relative humidity at the in- and one for summer that represents hot weather with a flow boundary were forced hourly with the “localised” clear sky in Hangzhou. The results from the case studies air temperature and relative humidity TMY data for in both scenarios showed that the air temperature differ- 24 h. The turbulence field was updated every 10 min; ences between the cases are less than 0.5 °C and for RH solar radiation was modelled with a dynamic time step the difference was in the range of 2–3 % RH units. (ie shorter when solar radiation is near its peak (1 s) and The sensitivity to the vertical grid cell dimension was longer during morning and afternoon (2 s)); the internal not assessed because it was considered important to a) temperature of buildings (free running) is calculated ac- have a solution at the height between 2.5 and 5 m above cording to the heat transfer through walls and roofs, the ground (ie middle of grid cells, solution at 2.80, 3.44, where all walls and roofs have the same thermal trans- 4.20 and 5.09 m) where the observations have been col- mittance and albedo. The spin-up period was set to 4 h lected and b) have at least 10 grid cells in the lower (starting at 20:00 on the day before the simulated day). 20 m of the domain and an expansion ratio below 20 % The computational domain in ENVI-met comprises an for the cells above the 20 m threshold. equidistant grid that can be compressed or stretched in In addition, two scenarios have been investigated for the vertical (z, height) dimension by using an expansion assessing the sensitivity of the model to the distribution ratio but there is no option for the local refinement of and the amount of green-space in the model. The first the horizontal computational grid. The starting grid cell scenario (Scenario 1) compared the air temperature at height, ie the height for the cell in contact with the 3.5 m height above ground (T ) in Case 1, where the 3.5m ground surface, was set to 0.5 m and the grid remained vegetation was distributed according to the statistical re- equidistant below the height of 2.5 m with a grid cell sults from the land surface analysis, with Case 2, where spacing equal to dz = 0.5 m. The combination of a 0.5 m all the vegetation surface area was moved to the centre starting grid cell height with an 18 % grid height expan- of the urban unit and Case 3 where the vegetated area sion ratio above 2.5 m resulted in a vertical grid with 16 was moved towards the outer annuli (Fig. 8). The pervi- grid cells at the lower part of the domain (ie the lower ous surface area ratio (ie 0.15) remained the same for 20 m within the roughness sub-layer). The urban unit’s the urban unit model. 250 m radius resulted in 3D computational grids of 72 × Scenario 2 compared the air temperature (T ) from 3.5m 72 × 28 grid cells with a horizontal resolution of 8 m. Case 1 with 5 additional Cases (4 to 8) that had increas- A sensitivity analysis of the simulation results showed ing ratios of pervious surface area that was distributed that an increase of the horizontal resolution from 8 to evenly (same percentage) in each annular area (Fig. 9). 6 m and 3 m (Table 5) delivered no significant change in Each case had 5 % RH points more vegetated surface the model output. The coarsening ratio was not constant area (ie in the form of grass) than the previous one up because of limitations set by the computational domain to a maximum of P =0.4 which represents the upper size and the fixed computational grid (ie the modelled limit for the “Local Climate Zone 5” classification. geometry should fit to an integer number of grid cells). The air temperature development in the idealised In addition, it was not possible to assess the grid sensitiv- “urban unit” for both scenarios was simulated with ity to computational grid dimensions below 3 m × 3 m ENVI-met (Version 4) for August 10, 2013 which repre- due to the simulation domain size limitations (ie 250 × sented a sunny hot day in Hangzhou. The sky was clear. 250 grid cells maximum). That is because the urban unit The previous four days had been dry with 41 °C max- has a diameter of 500 m and a number of grid cells in imum air temperature and similar weather conditions as proximity to the domain borders must remain empty. the day of the simulations. The analysis was conducted Grid cell resolutions coarser than 8 m were not assessed for 24 h from 00:00 China Standard Time (CST) to because they were deemed too low for the purposes of this 23:00 CST. study. Table 5 Case studies for the assessment of the solution’s sensitivity to grid resolution Case Horizontal grid Vertical grid resolution (dz) [m] Computational domain dimensions Coarsening ratio r i,i +1 resolution (dx,dy) [m] (x,y,z) [grid cells] (+no of nesting grids) Case 1 (3, 3) First 5 grid cells’ height: 0.5 m, from 2.25 m to 207 × 207 × 28 (+6) n.a. the top of the 3D domain: dz = 1.18 × dz n n-1 Case 2 (6, 6) Same as Case 1 104 × 104 × 28 (+5) r = 2.0 1,2 Case 3 (8, 8) Same as Case 1 72 × 72 × 28 (+4) r = 1.3 2,3 Bourikas et al. Future Cities and Environment (2016) 2:7 Page 13 of 23 Fig. 8 Distribution of the vegetated surface area in the urban unit for the cases considered in Scenario 1 of the green-space sensitivity analysis In both scenarios (green-space amount and distribu- successive hot, dry summer days, a reduction in soil tion) the modelled temperature varied less than 0.5 °C water content will negate, to a large extent, the cooling between the cases showing that the solution is not sensi- benefits of the added vegetation. tive to the distribution of the green-space in the model In the Chinese building regulations Hangzhou is clas- and the model performs as expected regarding the dif- sified into the Hot Summer - Cold Winter climate zone ferences in the amount of vegetation. [44]. In this zone, residential apartments are typically Specifically, the assessment of the impact of the vege- mixed mode with split air-conditioning and natural ven- tation’s location on the air temperature development in tilation [45]. For the purpose of assessing the potential the urban canopy revealed that the proximity to “green” improvements in the estimation of building energy con- – vegetated space can decrease the urban heat island in- sumption against the “business as usual” case of using tensity during night-time and the maximum day-time air TMY data files the heating and cooling degree days were temperature. The marginal difference between the cases calculated and compared amongst the reference TMY with a central allocation of the vegetated surface area file, the TMY file with a bulk hourly UHI offset and the and those were the vegetation was positioned at the “urbanised” TMY file after its adaptation to the “urban outer border of the urban unit is an indication that the weather projections”. The degree days have been calcu- distance to a vegetated area is not enough to alone pro- lated according to the data from the 584570_CSWD duce large cooling benefits during the day and attenuate TMY file for Hangzhou at a base temperature of 18 °C the night-time urban heat island intensity. In Scenario 2 [45] for heating and 26 °C [45] for cooling. The “urban (amount of green-space), an increase to the urban unit’s unit model” methodology was validated for each of the 4 permeable surface area showed a small decrease in the seasons and 3 different weather forcing conditions average air temperature across the urban unit. The case against the hourly average air temperature and RH ob- with the largest vegetated surface area had the lowest servations from the 9 studied sites on a given day repre- daily air temperatures. A shift was noted in the air sentative of the weather forcing conditions. The main temperature distribution towards a higher occurrence parameters for the “urban unit model” validation are frequency of temperatures at the cooler end. The differ- shown in Appendix 1. Looking at the dates given in ences between the cases were more evident in the aver- Appendix 1 the “urban unit model” validation simula- age surface temperatures. The results suggest that high tions were forced with the hourly weather data from the percentages of vegetated space can reduce the surface reference weather station (TMY source) overlaid with temperatures within the cities. There were, however, also the representative urban heat island effect of the weather strong indications that in places with a humid sub- forcing group as calculated from the 10 “sample” sites’ tropical climate such as Hangzhou, in the case of observations (and not the observed UHI during the Fig. 9 The different percentages of vegetated surface area in the “urban unit” for the cases in Scenario 2 of the green-space sensitivity analysis. Top plan view of the computational domain (Right) for the cases with P = 0.2 and P = 0.4 r r Bourikas et al. Future Cities and Environment (2016) 2:7 Page 14 of 23 simulated day) (see also Figs. 6 and 7). The hourly wea- Results and discussion of the “urban unit model” ther forcing for the “urban weather projections” simula- validation tions was undertaken with the hourly average air This evaluation of the model’s performance (Figs. 10, 11, temperature and RH for the respective weather forcing 12 and 13) showed that urban micro-climatic simula- group in the TMY file, adjusted by the representative tions using the idealised “urban unit model” captures to bulk hourly UHI intensity (same as in the validation within 1 °C the main characteristics of the diurnal air case) for the weather forcing groups. temperature development in all seasons. Fig. 10 Comparison of the observed (black line) and the modelled (red line) air temperature (left) and RH (right) at 3.5 m above ground for the 3 days in January 2013 representative for the weather forcing groups. (Time given in China Standard Time – CST: GMT + 8) Bourikas et al. Future Cities and Environment (2016) 2:7 Page 15 of 23 Fig. 11 Comparison of the observed (black line) and the modelled (red line) air temperature (left) and RH (right) at 3.5 m above ground for the 3 days in May 2013 representative for the weather forcing groups. (Time given in China Standard Time – CST: GMT + 8) If the model output is a perfect prediction we would have to be identical. The ENVI-met model would have expect the observed temperatures (black line; average of to perfectly forecast the average temperature/RH devel- hourly observations from the nine “LCZ5” sites, see opment at street level. Figs. 10, 11, 12 and 13) to be identical to the modelled In the weather forcing group 1 winter scenario (January temperatures (red line in Figs. 10, 11, 12 and 13). In this 13, 2013; Fig. 10 (top)) the modelled air temperature at case the UHI effect experienced by the 9 LCZ5 sites and 3.5 m above ground (Air Temp. [ C]; red line) is a very the 10 “sample” weather stations within the city would good fit to the observed temperature (black line; RMSE: Bourikas et al. Future Cities and Environment (2016) 2:7 Page 16 of 23 Fig. 12 Comparison of the observed (black line) and the modelled (red line) air temperature (left) and RH (right) at 3.5 m above ground for the 3 days in June 2013 representative for the weather forcing groups. (Time given in China Standard Time – CST: GMT + 8) 0.4, MAPE: 5 %, Table 6). The simulation results predict in the urban weather conditions’ prediction when com- the night-time UHI better than the reference weather sta- pared to the reference weather station. In the case of wea- tion (TMY source station) observations (circles in Figs. 10, ther forcing group 2 the simulated air temperature is a 11, 12 and 13). The RH results (red line in Fig. 10 (right)) good fit to the observations for most hours of the day are representative of the observed RH diurnal trend. The (Fig. 10 (middle)). The air temperature is overestimated simulated RH values are low compared to the average of early in the morning but the high values potentially better the hourly observations but still represent an improvement represent the urban conditions than the measured data Bourikas et al. Future Cities and Environment (2016) 2:7 Page 17 of 23 Fig. 13 Comparison of the observed (black line) and the modelled (red line) air temperature (left) and RH (right) at 3.5 m above ground for the 3 days in October 2013 representative for the weather forcing groups. (Time given in China Standard Time – CST: GMT + 8) from the reference weather station. In the weather forcing an inaccurate representation of thermal mass and heat group 3 case, winter daytime air temperatures are storage in the model [46] and the modelling of thermal underestimated with the model failing to predict the diffusivity by ENVI-met [47]. temperature peak around 13:00 h. Nevertheless, the The simulated RH in January for both the weather for- simulation results fit the observed data relatively well at cing group 2 and group 3 cases (Fig. 10), again replicates night-time, in the early morning and afternoon. The the daily observed trend and the RH predicted levels are failure to accurately predict the peak could be a result of comparable with the reference weather station observations Bourikas et al. Future Cities and Environment (2016) 2:7 Page 18 of 23 Table 6 Model performance indices determined for the winter and autumn simulations with the urban unit model Model performance indices T (RH) - winter (January) T (RH) - autumn (October) Group 1 Group 2 Group 3 Group 1 Group 2 Group 3 Mean Squared Error (MSE) 0.13 (56.1) 0.55 (116.1) 0.70 (258.5) 0.50 (61.6) 0.31 (261.2) 0.39 (287.1) Mean Bias Error (MBE) −0.14 (-7.4) 0.12 (-10) 0.16 (-15.4) −0.42 (-7.5) −0.17 (-15.9) −0.02 (-15.3) MSE systematic 0.11 (55.8) 0.49 (112.5) 0.40 (246.3) 0.33 (56.7) 0.22 (255.4) 0.23 (280.9) Root Mean Squared Error (RMSE) 0.36 (7.50) 0.74 (10.75) 0.84 (16.05) 0.70 (7.85) 0.56 (16.16) 0.63 (16.95) RMSE systematic 0.33 (7.47) 0.70 (10.61) 0.64 (15.70) 0.58 (7.53) 0.47 (15.98) 0.48 (16.76) MSE unsystematic 0.02 (0.73) 0.05 (3.05) 0.30 (11.17) 0.17 (4.89) 0.09 (5.85) 0.16 (6.26) RMSE unsystematic 0.15 (0.85) 0.23 (1.75) 0.55 (3.34) 0.41 (2.21) 0.31 (2.42) 0.40 (2.50) Mean Absolute Percentage Error (MAPE) 5 % (8 %) 21 % (12 %) 11 % (18 %) 3 % (9 %) 2 % (21 %) 2 % (21 %) Index of agreement d 0.92 (0.13) 0.97 (0.78) 0.98 (0.72) 0.95 (0.68) 0.99 (0.48) 0.99 (0.77) (especially when considering the 5 % units RH sensor ac- to the urban observations showing the existence of a ra- curacy) for the largest part of the day. ther small urban heat island effect in Hangzhou during The model underestimates the RH during early morn- spring. The RH in the urban unit model was again ing before 08:00 o’clock indicating a possible discrepancy underestimated with the error being acceptable (RMSE: between the modelled vegetation properties (ie amount 7, MAPE: 7 %, Table 7) in the case that represents over- and type of trees, grass) and reality. However, this differ- cast sky conditions (group 1, Fig. 11 (top)) but signifi- ence is not expected to have a significant effect on the cant (RMSE: 15, MAPE: 20 %, Table 7) in the weather model’s application because 1) the predicted RH values forcing group 3 case (ie clear sky, hot weather). Overall, are relatively close to the reference weather station ob- in the spring scenario the night-time urban heat island servations and 2) the largest discrepancy is early in the intensity was overestimated across all the weather for- morning and late at night in winter when typically dehu- cing groups. However, in May night-time the air midification is not an option when split AC units oper- temperature is still low and the air-conditioning demand ate in heating mode (or auxiliary heating sources are if any is expected to be minimal. In the weather forcing used instead). The satisfactory prediction of the expected group 3 case (Fig. 11 (bottom)) the high air temperature urban heat island during the night (ΔT )in at noon suggests that indoor temperatures are highly Case – reference all three cases is a further indication that this level of in- likely to exceed the comfort band threshold of 27 C accuracy is not detrimental to the overall function of the [48] creating a demand for cooling. The simulated air model. temperature peaks are a good fit to the urban observa- In the spring scenario (May 2013, Fig. 11) the air tions and in most cases they represent the urban wea- temperature was overestimated during night and early ther development better than the reference weather morning. The simulated air temperature in the after- station measurements. noon was representative of the observed air temperature In the summer scenario (June, Fig. 12), the air across all three weather forcing groups. The reference temperature predictions from all three weather forcing weather station observations were consistently very close groups fit the observations from the studied LCZ5 sites Table 7 Model performance indices determined for the spring and summer simulations with the urban unit model Model performance indices T (RH) - spring (May) T (RH) - summer (June) Group 1 Group 2 Group 3 Group 1 Group 2 Group 3 Mean Squared Error (MSE) 0.08 (48.5) 0.31 (91.3) 0.90 (226.5) 0.15 (77.2) 0.52 (65.2) 0.67 (88.3) Mean Bias Error (MBE) 0.15 (-6.8) 0.31 (-9.2) 0.84 (-14.4) −0.11 (-8.4) 0.23 (-6.9) 0.75 (-8.9) MSE systematic 0.04 (46) 0.11 (89.8) 0.80 (219.4) 0.04 (73.5) 0.08 (53.6) 0.56 (84.0) Root Mean Squared Error (RMSE) 0.29 (6.9) 0.56 (9.6) 0.95 (15.1) 0.38 (8.8) 0.72 (8.1) 0.82 (9.4) RMSE systematic 0.19 (6.78) 0.33 (9.47) 0.89 (14.81) 0.20 (8.56) 0.27 (7.32) 0.75 (9.17) MSE unsystematic 0.05 (1.96) 0.20 (2.45) 0.10 (8.37) 0.11 (3.71) 0.45 (11.60) 0.11 (4.73) RMSE unsystematic 0.22 (1.40) 0.45 (1.57) 0.32 (2.89) 0.33 (1.93) 0.67 (3.41) 0.33 (2.17) Mean Absolute Percentage Error (MAPE) 1 % (7 %) 2 % (11 %) 3 % (20 %) 2 % (9 %) 2 % (9 %) 2 % (13 %) Index of agreement d 0.78 (0.17) 0.97 (0.82) 0.98 (0.73) 0.94 (0.29) 0.98 (0.89) 0.98 (0.85) Bourikas et al. Future Cities and Environment (2016) 2:7 Page 19 of 23 reasonably well (RMSE: 0.4 to 0.8, MAPE: 2 %, Table 7). uncertainties involved in the initialisation of the model In June 2013, in the weather forcing group 3 case, the and the simulation itself. Furthermore, the model per- air temperature is over predicted at noon and early formance evaluation indices’ scores for air temperature morning. This recurring pattern in the simulation of hot (Tables 6 and 7) are comparable to other published work weather conditions in spring and summer can be the re- [46, 49]. The index of agreement d takes values in the sult of underrepresentation of the permeable surface and range {0, 1} with a value of 1 indicating a perfect match soil water content in the model [47]. The RH simulation between the model prediction and the observations results support this argument with the error of the simu- [46, 50]. The index of agreement scores low in the wea- lated RH in morning being in the range of 15 RH per- ther forcing group 1 results for all seasons. However the centage units. Mean Absolute Percentage Error of the air temperature Overall, the RH is underestimated during large parts was consistently less than 10 % in almost all cases and the of the day but the simulated RH levels are comparably systematic component of the root mean square error close to the observations with a RMSE around 10 RH (RMSEs) was lower than 0.5 C in most cases and always units and MAPE of 10 % (Table 7). lower than 1 C. The systematic component of the error In October (autumn, Fig.13) the model simulates the (ie RMSEs), which represents the error attributed to the urban heat island development during the night with the simulation and the error integrated into the initialisation simulated air temperature being close to the observa- estimates, should approach 0 [46]. The unsystematic com- tions from the LCZ5 sites. The modelled air temperature ponent should approach the value of RMSE [50]. The rela- development is more realistic for the urban environment tively low RMSE values in conjunction with the fact that than the measured reference temperature. At noon and the model results were closer to the urban observations under hot weather conditions (weather forcing group 3), than the reference measured values show that the model’s the model fails to accurately predict the peak but the performance is acceptable. simulated temperature is still an improvement in com- The accuracy of the RH prediction is most important parison with the reference air temperature observa- during the cooling season when dehumidification is re- tions. In general, domestic buildings in October are quired. In Hangzhou and other similar cities with a expected to be free-running (no heating or cooling). humid sub-tropical climate the full cooling season is ex- There is a small demand for cooling in September and pected to last from June to September [31, 51]. The for heating from mid-November onwards. The valid- validation showed that in June and for the summer sce- ation of the model showed that cooling demand in au- nario the error is below 10 RH percentage units. How- tumn is likely to be under estimated at noon and in the ever, the measured RH from the reference weather early afternoon but its prediction is largely improved at station fits better to the urban observations than the night when the air temperature simulation results fit simulated RH. Furthermore there is only a marginal RH the urban observations very well (RMSE: 0.6 to 0.7, difference between the urban and the reference sites. MAPE: ~2 %, Table 6). The simulated RH follows the Therefore, the urban weather projections were used to observed diurnal RH trend but in weather forcing adapt the air temperature only in the TMY file and not groups 2 and 3 the error is large with the RMSE equal the relative humidity. to 16 and 17 RH % units respectively and the MAPE in A full scale error analysis was not undertaken due to the range of 20 %. lack of available data/input parameters notably the This discrepancy might be a result of weather events that hourly global solar radiation and soil properties. The val- ENVI-met cannot simulate such as mist, haze and rain. idation has been restricted to days that are typical for th Specifically on the 12 of October there was mist until the weather forcing groups’ conditions. The main 08:00 in the morning when it dispersed and at the same sources of systematic error are the input initialisation time the model’s prediction accuracy increased. In the parameters, in particular hourly global solar radiation. evening the large error in the RH prediction can be attrib- Other sources of systematic error include the initial uted to the modelled soil water content and the position of boundary conditions, building and vegetation properties. some of the LCZ5 sites being close to large water bodies. Unsystematic errors might be a result of the thermal dif- In late autumn the RH plays little role as there is no dehu- fusivity [47] and turbulence modelling, the modelling of midification load associated with heating. In addition, the evapotranspiration and the total heat advection to the largest RH error in October was noticed early in the morn- atmosphere. ing and during the evening when the temperature is ex- pected to be within the comfort band of the thermostatic Results and discussion of the TMY file for set points and there will be no cooling requirement. Hangzhou adapted with the urban unit model Overall, the accuracy of the model can be viewed as Following the “urban unit model” validation the TMY satisfactory in relation to its purpose, considering the file for Hangzhou (584570_CSWD) was adapted to Bourikas et al. Future Cities and Environment (2016) 2:7 Page 20 of 23 include the “urban weather projections” (UWP) ac- the simulation results with the use of the TMY-UWP cording to the methods outlined above. A comparison file. amongst the reference TMY, the TMY overlaid with the bulk “city” UHI effect (TMY + UHI) and the micro- Conclusions scale “urban weather projections” (TMY-UWP) is The validation of the “urban unit model” and the results shown in Fig. 14. In the “TMY + UHI” HDD have de- of its implementation into micro-climatic simulations creased by 6 % from 1598 days in the reference TMY show that there is a potential for the simplification of and in the “TMY-UWP” file by 13 % showing that the urban site modelling and for the wider application of the local specific micro-climate attributes an additional 7 % method introduced in this paper as a tool for adapting heating load reduction to the bulk UHI effect of the typical meteorological weather data files to represent the city. This difference between the “TMY-UWP” and the neighbourhood scale of cities with a humid sub-tropical “TMY + UHI” files is a reflection of the thermal charac- climate. teristics of the specific urban morphology of the stud- Overall, the comparison amongst the reference (ie ied sites compared to the wider city. CDD in the “TMY 1598 HDD), a bulk “city” UHI effect and the adapted for +UHI” file are 17 % more than the reference TMY file the “urban weather projections” TMY files indicates that (207 days). The “TMY-UWP” file shows an additional there was a 6 % decrease in HDD that is attributed to 14 % increase in comparison with the “TMY + UHI” the bulk “city” UHI effect and an additional 7 % (ie from file. 6 to 13 %) that can be directly attributed to the local The impact of the change in HDD and CDD was specific urban morphology of the 9 LCZ5 sites. The total assessed in terms of heating and cooling loads for a increase of CDD from the reference (ie 207 CDD) to the domestic and a non-domestic building scenario in “urban weather projections” TMY files is in the range of Hangzhou. For this purpose, the “urbanised” (TMY- 30 %. This assessment indicates that cooling loads can UWP) and the reference TMY files were used within a be significantly underestimated in the business as usual dynamic thermal simulation tool (TRNSYS Version case of using the reference TMY file and that the urban 17.1) to model the annual heating and cooling load for effect on air temperature should not be neglected. both scenarios. The calculated heating and cooling loads The heating and cooling demand has been further for both the domestic and the non-domestic building calculated for a domestic and a non-domestic build- case show that the energy performance simulations with ing case in Hangzhou. It has been observed that in the reference TMY file result in an approximate 20 % both cases there is an approximate 20 % increase of over prediction of the heating load and 20 % under pre- the cooling load and a 20 % decrease of the heating diction of the cooling load demand in comparison with load. If typical COP values for a reversible air- Fig. 14 Comparison of the degree days in the reference TMY (TMY reference, gray; 1598 HDD, 207 CDD) with the TMY overlaid with the bulk “city” UHI effect (TMY + UHI, orange; 1495 HDD, 243 CDD) and the micro-scale “urban weather projections” (TMY-UWP, red; 1383 HDD, 279 CDD). T is the temperature threshold used for the calculation of the HDD and CDD. The area with the light blue background marks the base cooling period Bourikas et al. Future Cities and Environment (2016) 2:7 Page 21 of 23 conditioning system are taken as 2.0 for heating and Table 9 Main input parameters for the simulations with the urban unit model in June 2013 3.5for coolingthenthe total electricity consumption estimated with the use of the “urban weather projec- Input parameter Jun 14, Jun 23, Jun 18, Source 2013 2013 2013 tions” TMY file will be decreased by 11 % in com- Weather forcing Group Group 1 Group 2 Group 3 parison with the “business as usual” (ie reference Specific humidity 2500 m 8.9 12 11.3 [52] TMY) case.Thisisan interesting result showingthe (750 mbar) gr w/kg dry air impact highly efficient heat pumps and air- Prevalent wind direction 45 90 225 [53] conditioning systems can have on the electricity con- (N = 0 clockwise) sumption of cities with a humid sub-tropical climate. Wind speed 10 m ab. gr. m/s 3.1 2.8 3.1 [53] However, this assumes a cooling set-point of 26 °C. Roughness length z 0.1 0.1 0.1 [36] If a lower set-point is used the predicted energy sav- ings will be lost. Mean wall albedo 0.23 0.23 0.23 [49] In the majority of cases it would be difficult to justify Mean roof albedo 0.50 0.50 0.50 the additional level of analysis described here to develop Wall heat transmittance 1.4 1.4 1.4 −2 −1 the localised weather data file over the generic bulk W. m .K “city” UHI file. The method is shown to deliver an en- Roof heat transmittance 0.9 0.9 0.9 −2 −1 hancement which in large developments could be justi- W. m .K fied. Single, smaller/residential buildings can take the Underground soil temperature 294.3 299 300 [52] bulk UHI correction approach. This methodology would (Upper-Middle-Deep layer) K 292.7 294 293 also benefit initial stages of urban planning and could in- 290.6 291 291 form decisions on the use and the urban form of existing Underground soil humidity 38 % 39 % 38 % [52] and new developments in the city. (Upper-Middle-Deep layer) 38 % 37 % 38 % Future work is looking to expand this methodology to different urban morphologies and to transfer it to differ- 37 % 36 % 37 % ent climates. Appendix 1 Table 8 Main input parameters for the simulations with the Table 10 Main input parameters for the simulations with the urban unit model in January 2013 urban unit model in May 2013 Input parameter Jan 13, Jan 12, Jan 24, Source Input parameter May 17, May 10, May 22, Source 2013 2013 2013 2013 2013 2013 Weather forcing Group Group 1 Group 2 Group 3 Weather forcing Group Group 1 Group 2 Group 3 Specific humidity 2500 m 3.3 2.0 1.1 [52] Specific humidity 2500 m 8.3 7.2 3.9 [52] (750 mbar) gr w/kg dry air (750 mbar) gr w/kg dry air Prevalent wind direction 22 225 225 [53] Prevalent wind direction (N = 0 0 110 0 [53] (N = 0 clockwise) clockwise) Wind speed 10 m ab. gr. m/s 3.2 2.1 2.5 [53] Wind speed 10 m ab. gr. m/s 1.6 2.8 2.8 [53] Roughness length z 0.1 0.1 0.1 [36] Roughness length z 0.1 0.1 0.1 [36] 0 0 Mean wall albedo 0.23 0.23 0.23 [49] Mean wall albedo 0.23 0.23 0.23 [49] Mean roof albedo 0.50 0.50 0.50 Mean roof albedo 0.50 0.50 0.50 Wall heat transmittance 1.4 1.4 1.4 Wall heat transmittance 1.4 1.4 1.4 −2 −1 −2 −1 W. m .K W. m .K Roof heat transmittance 0.9 0.9 0.9 Roof heat transmittance 0.9 0.9 0.9 −2 −1 −2 −1 W. m .K W. m .K Underground soil temperature 278.4 277.3 280.6 [52] Underground soil temperature 292.7 293.2 296.0 [52] (Upper-Middle-Deep layer) K (Upper-Middle-Deep layer) K 279.6 279.5 281.4 290.3 288.9 291.0 282.6 283.0 282.4 288.0 287.2 288.0 Underground soil humidity 38 % 36 % 34 % [52] Underground soil humidity 35 % 37 % 33 % [52] (Upper-Middle-Deep layer) (Upper-Middle-Deep layer) 38 % 36 % 35 % 35 % 37 % 34 % 36 % 35 % 35 % 34 % 35 % 34 % Bourikas et al. Future Cities and Environment (2016) 2:7 Page 22 of 23 Table 11 Main input parameters for the simulations with the Author details Energy & Climate Change Division, Sustainable Energy Research Group urban unit model in October 2013 (SERG), Faculty of Engineering and the Environment, University of Input parameter October October October Source 2 Southampton, Southampton SO17 1BJ, UK. Urban Energy Systems, Faculty 18, 2013 05, 2013 12, 2013 3 of Civil Engineering, Bauhaus-Universität Weimar, Weimar, Germany. Centre for Sustainable Energy Technologies (CSET), University of Nottingham Weather forcing Group Group 1 Group 2 Group 3 Ningbo, Ningbo, People’s Republic of China. School of Architecture, Faculty Specific humidity 2500 m 5.3 0.6 0.8 [52] of Humanities and Social Sciences, University of Liverpool, Liverpool, UK. (750 mbar) gr w/kg dry air Faculty of Engineering, University of Nottingham, Nottingham, UK. Prevalent wind direction 0 225 200 [53] Received: 2 February 2016 Accepted: 30 June 2016 (N = 0 clockwise) Wind speed 10 m ab. gr. m/s 3.0 2.5 2.2 [53] Roughness length z 0.1 0.1 0.1 [36] References Mean wall albedo 0.23 0.23 0.23 [49] 1. Crawley DB (1998) Which weather data should you use for energy simulations of commercial buildings? In: ASHRAE Transactions. ASHRAE, Mean roof albedo 0.50 0.50 0.50 Atlanta, USA, pp 498–515 Wall heat transmittance 1.4 1.4 1.4 2. Hacker J, Capon R, Mylona A (2009) Use of climate change scenarios for −2 −1 W. m .K building simulation: the CIBSE future weather years. The Chartered Institution of Building Services Engineers, London, UK Roof heat transmittance 0.9 0.9 0.9 −2 −1 3. Thomas PC, Moller S (2006) HVAC system size - getting it right. In: Clients W. m .K Driving Innovation: Moving Ideas into Practice. Cooperative Research Centre Underground soil temperature 291.0 291.3 292.5 [52] for Construction Innovation, Brisbane, Australia (Upper-Middle-Deep layer) K 4. Burdick A (2011) Strategy Guideline: Accurate heating and cooling load 293.1 292.8 291.3 calculations. Oak Ridge, USA. 294.1 294.0 293.1 5. US Department of Energy (2016) EnergyPlus - Weather Data Sources. Available via https://energyplus.net/weather/sources. Accessed Underground soil humidity 33 % 32 % 29 % [52] 19 June 2016 (Upper-Middle-Deep layer) 34 % 33 % 32 % 6. Mylona A (2012) The use of UKCP09 to produce weather files for building simulation. Build Serv Eng Res Technol 33(1):51–62 34 % 33 % 32 % 7. Oke TR (1982) The energetic basis of the urban heat island. Q J R Meteorol Soc 108(455):1–24 8. Collier CG (2006) The impact of urban areas on weather. Q J R Meteorol Soc 132(614):1–25 Abbreviations 9. Taha H, Konopacki S, Gabersek S (1999) Impacts of Large-Scale Surface CDD, cooling degree day; CDH, Ccooling degree hour; CST, China standard Modifications on Meteorological Conditions and Energy Use: A 10-Region time; GHR, global horizontal radiation; GMT, Greenwich mean time; HDD, Modeling Study. Theor Appl Climatol 62(3):175–185 heating degree Dday; HDH, heating degree hour; HVAC, heating, ventilation 10. Memon RA, Leung DYC, Liu CH (2009) An investigation of urban heat island and air conditioning; LCZ, local climate zone; MAPE, mean absolute intensity (UHII) as an indicator of urban heating. Atmos Res 94(3):491–500 percentage error; MBE, mean bias error; MSE, mean squared error; RANS, 11. Jusuf ST, Wong NH (2009) Development of empirical models for an Reynolds averaged Navier stokes; RH, relative humidity; RMSE, root mean estate level air temperature prediction in Singapore. Paper presented at squared error; TKE, turbulent kinetic energy; TMY, typical meteorological year; 7th International Conference on Urban Climate. June 29 - July 3, UHI, urban heat island; UWP, urban weather projections Yokohama, Japan. 12. Kolokotroni M, Davies M, Croxford B, Bhuiyan S, Mavrogianni A (2010) A Acknowledgement validated methodology for the prediction of heating and cooling energy This work is part of the activities of the University of Southampton’s Energy demand for buildings within the Urban Heat Island: Case-study of London. and Climate Change Division and the Sustainable Energy Research Group Sol Energy 84(12):2246–2255 (www.energy.soton.ac.uk) on cities and infrastructure. It is partly supported 13. Kolokotroni M, Ren X, Davies M, Mavrogianni A (2012) London's urban heat by the EPSRC Grant EP/J017698/1, “Transforming the Engineering of Cities to island: Impact on current and future energy consumption in office Deliver Societal and Planetary Wellbeing” and EP/K012347/1, “International buildings. Energy Buildings 47:302–311 Centre for Infrastructure Futures (ICIF)”. The installation work of the sensors’ 14. Chan ALS (2011) Developing a modified typical meteorological year network in Hangzhou and Ningbo is supported by the Ningbo Natural weather file for Hong Kong taking into account the urban heat island Science Foundation (No. 2012A610173) and the Ningbo Housing and Urban- effect. Build Environ 46(12):2434–2441 Rural Development Committee (No. 201206). 15. Wong NH, Jusuf ST, Tan CL (2011) Integrated urban microclimate assessment method as a sustainable urban development and urban design Authors’ contributions tool. Landsc Urban Plan 100:386–389 LB developed the methodology, the “urban unit model” and managed the 16. Watkins R, Palmer J, Kolokotroni M, Littlefair P (2002) The London heat work in this study. PABJ provided guidance, reviewed and supervised all island: results from summertime monitoring. Build Serv Eng Res Technol stages of this study. ABSB contributed to the management of this study and 23(2):97–106 reviewed the final outcome. MFJ contributed to the development of the 17. Kershaw T, Sanderson M, Coley D, Eames M (2010) Estimation of the urban initial idea for this study and reviewed parts of the work and the final heat island for UK climate change projections. Build Serv Eng Res Technol outcome. TS installed and managed the sensors network in China, collected 31(3):251–263 the observations and contributed to their analysis. DHCC contributed to the 18. De La Flor FS, Doḿ ınguez SA (2004) Modelling microclimate in urban development of the sensors network in China and supervised the data environments and assessing its influence on the performance of analysis. JD contributed to the management of the sensors network and surrounding buildings. Energy Buildings 36(5):403–413 reviewed the data analysis and parts of this study. All authors read and 19. Chen F, Kusaka H, Bornstein R, Ching J, Grimmond CSB, Grossman-Clarke S, approved the final manuscript. Loridan T, Manning KW, Martilli A, Miao S, Sailor D, Salamanca FP, Taha H, Tewari M, Wang X, Wyszogrodzki AA, Zhang C (2011) The integrated Competing interests WRF/urban modelling system: development, evaluation, and applications to The authors declare that they have no competing interests. urban environmental problems. Int J Climatol 31(2):273–288 Bourikas et al. Future Cities and Environment (2016) 2:7 Page 23 of 23 20. Yamada T, Koike K (2011) Downscaling mesoscale meteorological models 43. Bruse M, Fleer H (1998) Simulating surface–plant–air interactions inside for computational wind engineering applications. J Wind Eng Ind Aerodyn urban environments with a three dimensional numerical model. Environ 99(4):199–216 Model Softw 13(3–4):373–384 21. Tewari M, Kusaka H, Chen F, Coirier WJ, Kim S, Wyszogrodzki AA, Warner TT 44. Wang H et al (2014) Feasibility and optimization of aerogel glazing system (2010) Impact of coupling a microscale computational fluid dynamics for building energy efficiency in different climates. Int J Low-Carbon model with a mesoscale model on urban scale contaminant transport and Technol 0:1–8 dispersion. Atmos Res 96(4):656–664 45. Wang X, Altan H, Kang J (2015) Parametric study on the performance of green residential buildings in China. Frontiers of Archit Res 4(1):56–67 22. Schlünzen KH (2010) Joint modelling of obstacle induced and mesoscale 46. Middel A et al (2014) Impact of urban form and design on mid-afternoon changes-current limits and challenges. Paper presented at The Fifth microclimate in Phoenix Local Climate Zones. Landsc Urban Plan 122:16–28 International Symposium on Computational Wind Engineering (CWE2010). 47. Maggiotto G, Buccolieri R, Santo MA, Leo LS, Di Sabatino S (2014) Validation May 23-27, 2010, Chapel Hill, North Carolina, USA. of temperature-perturbation and CFD-based modelling for the prediction of 23. Baklanov A, Martilli A, Grimmond CSB, Mahura A, Ching J, Calmet I, Clark P, the thermal urban environment: the Lecce (IT) case study. Environ Model Esau I, Dandou A, Zilitinkevich S, Best MJ, Mestayer P, Santiago JL, Tombrou Softw 60:69–83 M, Petersen C, Porson A, Salamanca F, Amstrup B (2010) Hierarchy of Urban 48. Nikolou S (2011) Low carbon city living in Guangzhou, China. Unpublished, Canopy Parameterisations for different scale models. MEGAPOLI Project MSc Thesis. University of Southampton, Southampton, UK Scientific Report 10-04. Danish Meteorological Institute, DMI, Copenhagen, 49. Yang X et al (2013) Evaluation of a microclimate model for predicting the Available via http://megapoli.dmi.dk/publ/MEGAPOLI_sr10-04.pdf thermal behavior of different ground surfaces. Build Environ 60:93–104 24. Martilli A, Santiago J (2007) CFD simulation of airflow over a regular array of 50. Willmott CJ (1982) Some comments on the evaluation of model cubes. Part II: analysis of spatial average properties. Bound-Layer Meteorol performance. Bull Am Meteorol Soc 63(11):1309–1313 122(3):635–654 51. Hu T, Yoshino H, Jiang Z (2013) Analysis on urban residential energy 25. Ren Z, Wang X, Chen D, Wang C, Thatcher M (2014) Constructing weather consumption of Hot Summer & Cold Winter Zone in China. Sustainable data for building simulation considering urban heat island. Build Serv Eng Cities Society 6:85–91 Res Technol 35(1):69–82 52. National Centers for Environmental Prediction/National Weather 26. Bueno B, Norford L, Hidalgo J, Pigeon G (2012) The urban weather Service/NOAA/U.S. Department of Commerce. 2000, updated daily. generator. J Build Perform Simul 6(4):269–281 NCEP FNL Operational Model Global Tropospheric Analyses, continuing 27. Massachusetts Institute of Technology (2015) Urban heat island effect from July 1999. Research Data Archive at the National Center for modelling software: Urban Weather Generator version 3.0.0 MIT, Available Atmospheric Research, Computational and Information Systems via http://urbanmicroclimate.scripts.mit.edu/uwg.php. Accessed 26 Feb 2016 Laboratory. dx.doi.org/10.5065/D6M043C6. Accessed 26 Jan 2016. 28. Arnfield AJ (2003) Two decades of urban climate research: a review of 53. The Weather Underground (2014) Hangzhou weather data from Mantou turbulence, exchanges of energy and water, and the urban heat island. Int J mountain's National Principle WMO-listed weather station. Climatol 23(1):1–26 www.wunderground.com. Accessed 26 Jan 2016 29. Bouyer J, Inard C, Musy M (2011) Microclimatic coupling as a solution to improve building energy simulation in an urban context. Energy Buildings 43(7):1549–1559 30. Bourikas L et al (2013) Addressing the challenge of interpreting microclimatic weather data from urban sites. J Power Energy Eng 1:7–15 31. Shen T et al (2014) Impact of Urban Heat Island on Building Cooling Energy Consumption in Hangzhou. Abstracts of the 13th International Conference on Sustainable Energy Technologies, Geneva, Switzerland 32. Maxim Integrated (2013) iButton Temperature/Humidity logger with 8 kb data logger memory, Available via www.maximintegrated.com/products/ ibutton/data-logging/ 33. Shen T et al. (2013) Generating a modified weather data file for urban building design and sustainable urban planning accounting for the Urban Heat Island (UHI) effect. In: Abstracts of the 12th International Conference on Sustainable Energy Technologies, Hong Kong Polytechnic University, 26-29 August 2013. 34. Oke TR (2006) Initial guidance to obtain representative meteorological observations at urban sites. In: Instruments and observing methods Report No81, World Meteorological Organization. Available via www.wmo.int/ pages/prog/www/IMOP/publications/IOM-81/IOM-81-UrbanMetObs.pdf. Accessed 26 Jan 2016. 35. Cheng H, Castro IP (2002) Near wall flow over urban-like roughness. Bound-Layer Meteorol 104(2):229–259 36. Stewart ID, Oke TR (2012) Local Climate Zones for Urban Temperature Studies. Bull Am Meteorol Soc 93(12):1879–1900 37. Google Earth (2013) Satellite Images of Hangzhou and Ningbo 38. Trimble Navigation Limited (2013) SketchUp 3D CAD software, Available via Submit your manuscript to a www.sketchup.com/ journal and beneﬁ t from: 39. Xie Z, Castro IP (2006) Large-eddy simulation for urban micro-meteorology. J Hydrodynamics, Ser B 18(3, Supplement):259–264 7 Convenient online submission 40. Santiago J, Martilli A, Martín F (2007) CFD simulation of airflow over a 7 Rigorous peer review regular array of cubes. Part I: Three-dimensional simulation of the flow and validation with wind-tunnel measurements. Bound-Layer Meteorol 7 Immediate publication on acceptance 122(3):609–634 7 Open access: articles freely available online 41. Kanda M, Moriizumi T (2009) Momentum and Heat Transfer over Urban-like 7 High visibility within the ﬁ eld Surfaces. Bound-Layer Meteorol 131(3):385–401 7 Retaining the copyright to your article 42. Millward-Hopkins JT et al (2013) Aerodynamic Parameters of a UK City Derived from Morphological Data. Bound-Layer Meteorol 146(3):447–468 Submit your next manuscript at 7 springeropen.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Future Cities and Environment Springer Journals http://www.deepdyve.com/lp/springer-journals/transforming-typical-hourly-simulation-weather-data-files-to-represent-uwvZd0IYGy

Loading next page...

References (62)

L Bourikas (2013)
Addressing the challenge of interpreting microclimatic weather data from urban sites
J Power Energy Eng, 1
Urban heat island effect modelling software: Urban Weather Generator version 3.0.0 MIT
(2013)
SketchUp 3D CAD software
T. Hu, H. Yoshino, Zhongtian Jiang (2013)
Analysis on urban residential energy consumption of Hot Summer & Cold Winter Zone in China
Sustainable Cities and Society, 6
M. Bruse, H. Fleer (1998)
Simulating surface–plant–air interactions inside urban environments with a three dimensional numerical model
Environmental Modelling and Software, 13
(2011)
Low carbon city living in Guangzhou, China. Unpublished, MSc Thesis
P. Thomas, S. Moller (2006)
HVAC system size - getting it right
Bruno Bueno, L. Norford, J. Hidalgo, G. Pigeon (2013)
The urban weather generator
Journal of Building Performance Simulation, 6
EnergyPlus -Weather Data Sources
Fei Chen, H. Kusaka, R. Bornstein, J. Ching, C. Grimmond, 4. Susanne, Grossman-Clarke, T. Loridan, Kevin Manning, A. Martilli, Shiguang Miao, D. Sailor, Francisco Salamanca, H. Taha, M. Tewari, Xuemei Wang, A. Wyszogrodzki, Chao-Lin Zhang (2010)
The Integrated WRF/Urban Modeling System: Development, Evaluation, and Applications to Urban Environmental Problems
A. Middel, Kathrin Häb, A. Brazel, C. Martin, S. Guhathakurta (2014)
Impact of urban form and design on mid-afternoon microclimate in Phoenix Local Climate Zones
Landscape and Urban Planning, 122
(2013)
Generating a modified weather data file for urban building design and sustainable urban planning accounting for the Urban Heat Island (UHI) effect
(2013)
Satellite Images of Hangzhou and Ningbo
Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory. dx
Z. Ren, X. Wang, D. Chen, C. Wang, M. Thatcher (2014)
Constructing weather data for building simulation considering urban heat island
Building Services Engineering Research & Technology, 35
(2013)
iButton Temperature/Humidity logger with 8 kb data logger memory
DB Crawley (1998)
ASHRAE Transactions
K. Schlünzen, D. Grawe, S. Bohnenstengel, I. Schlüter, R. Koppmann (2011)
Joint modelling of obstacle induced and mesoscale changes—Current limits and challenges
Journal of Wind Engineering and Industrial Aerodynamics, 99
M. Kanda, T. Moriizumi (2009)
Momentum and Heat Transfer over Urban-like Surfaces
Boundary-Layer Meteorology, 131
S. Jusuf, W. Hien (2009)
Development of empirical models for an estate level air temperature prediction in Singapore
Xi Wang, H. Altan, Jian Kang (2015)
Parametric study on the performance of green residential buildings in China
Collection of Frontiers of Architectural Research, 4
Maxim Integrated (2013) iButton Temperature/Humidity logger with 8 kb data logger memory
T. Kershaw, M. Sanderson, D. Coley, M. Eames (2010)
Estimation of the urban heat island for UK climate change projections
Building Services Engineering Research & Technology, 31
J. Hacker, R. Capon, A. Mylona (2009)
Use of Climate Change Scenarios for Building Simulation: the CIBSE Future Weather Years
A Baklanov, A Martilli, CSB Grimmond, A Mahura, J Ching, I Calmet, P Clark, I Esau, A Dandou, S Zilitinkevich, MJ Best, P Mestayer, JL Santiago, M Tombrou, C Petersen, A Porson, F Salamanca, B Amstrup (2010)
Hierarchy of Urban Canopy Parameterisations for different scale models. MEGAPOLI Project Scientific Report 10-04
J. Santiago, A. Martilli, F. Martín (2007)
CFD simulation of airflow over a regular array of cubes. Part I: Three-dimensional simulation of the flow and validation with wind-tunnel measurements
Boundary-Layer Meteorology, 122
M. Tewari, H. Kusaka, Fei Chen, W. Coirier, Sura Kim, A. Wyszogrodzki, T. Warner (2010)
Impact of coupling a microscale computational fluid dynamics model with a mesoscale model on urban scale contaminant transport and dispersion
Atmospheric Research, 96
R. Memon, D. Leung, Chun-Ho Liu (2009)
An investigation of urban heat island intensity (UHII) as an indicator of urban heating
Atmospheric Research, 94
H. Taha, S. Konopacki, S. Gabersek (1998)
Impacts of Large-Scale Surface Modifications on Meteorological Conditions and Energy Use: A 10-Region Modeling Study
Theoretical and Applied Climatology, 62
Iain Stewart, T. Oke (2012)
Local Climate Zones for Urban Temperature Studies
Bulletin of the American Meteorological Society, 93
Initial guidance to obtain representative meteorological observations at urban sites
M. Kolokotroni, X. Ren, M. Davies, A. Mavrogianni (2012)
London's urban heat island: Impact on current and future energy consumption in office buildings
Energy and Buildings, 47
A. Chan (2011)
Developing a modified typical meteorological year weather file for Hong Kong taking into account the urban heat island effect
Building and Environment, 46
R. Watkins, J. Palmer, M. Kolokotroni, P. Littlefair (2002)
The London Heat Island: results from summertime monitoring
Building Services Engineering Research and Technology, 23
Huan Wang, Huijun Wu, Yunfei Ding, Jing-xi Feng, Shengwei Wang (2015)
Feasibility and optimization of aerogel glazing system for building energy efficiency in different climates
International Journal of Low-carbon Technologies, 10
A. Arnfield (2003)
Two decades of urban climate research: a review of turbulence, exchanges of energy and water, and the urban heat island
International Journal of Climatology, 23
J. Bouyer, C. Inard, M. Musy (2011)
Microclimatic coupling as a solution to improve building energy simulation in an urban context
Energy and Buildings, 43
A. Burdick (2011)
Strategy Guideline: Accurate Heating and Cooling Load Calculations
Xiaoshan Yang, Lihua Zhao, M. Bruse, Q. Meng (2013)
Evaluation of a microclimate model for predicting the thermal behavior of different ground surfaces
Building and Environment, 60
N. Wong, S. Jusuf, C. Tan (2011)
Integrated urban microclimate assessment method as a sustainable urban development and urban design tool
Landscape and Urban Planning, 100
T Shen (2014)
Impact of Urban Heat Island on Building Cooling Energy Consumption in Hangzhou
A. Martilli, J. Santiago (2007)
CFD simulation of airflow over a regular array of cubes. Part II: analysis of spatial average properties
Boundary-Layer Meteorology, 122
(2010)
Hierarchy of Urban Canopy Parameterisations for different scale models
T. Oke (1982)
The energetic basis of the urban heat island
Quarterly Journal of the Royal Meteorological Society, 108
X Wang, H Altan, J Kang (2015)
Parametric study on the performance of green residential buildings in China
Frontiers of Archit Res, 4
M. Kolokotroni, M. Davies, B. Croxford, S. Bhuiyan, A. Mavrogianni (2010)
A validated methodology for the prediction of heating and cooling energy demand for buildings within the Urban Heat Island: Case-study of London
Solar Energy, 84
J. Millward-Hopkins, A. Tomlin, L. Ma, D. Ingham, M. Pourkashanian (2013)
Aerodynamic Parameters of a UK City Derived from Morphological Data
Boundary-Layer Meteorology, 146
PC Thomas, S Moller (2006)
Clients Driving Innovation: Moving Ideas into Practice
Hangzhou weather data from Mantou mountain's National Principle WMO-listed weather station
Zheng-Tong Xie, I. Castro (2006)
Large-eddy simulation for urban micro-meteorology
Journal of Hydrodynamics, 18
(2011)
Low carbon city living in Guangzhou, China
C. Collier (2006)
The impact of urban areas on weather
Quarterly Journal of the Royal Meteorological Society, 132
(2000)
National Centers for Environmental Prediction/National Weather Service/NOAA/U.S. Department of Commerce
L. Bourikas, Tianfeng Shen, P. James, D. Chow, M. Jentsch, J. Darkwa, A. Bahaj (2013)
Addressing the challenge of interpreting microclimatic weather data collected from urban sites
Journal of Power and Energy Engineering, 01
Hong Cheng, I. Castro (2002)
Near Wall Flow over Urban-like Roughness
Boundary-Layer Meteorology, 104
C. Willmott (1982)
Some Comments on the Evaluation of Model Performance
Bulletin of the American Meteorological Society, 63
D. Crawley (1998)
Which Weather Data Should You Use for Energy Simulations of Commercial Buildings
F. Flor, Servando Domínguez (2004)
Modelling microclimate in urban environments and assessing its influence on the performance of surrounding buildings
Energy and Buildings, 36
G. Maggiotto, R. Buccolieri, M. Santo, L. Leo, S. Sabatino (2014)
Validation of temperature-perturbation and CFD-based modelling for the prediction of the thermal urban environment: the Lecce (IT) case study
Environ. Model. Softw., 60
A. Mylona (2012)
The use of UKCP09 to produce weather files for building simulation
Building Services Engineering Research & Technology, 33
(2006)
Initial guidance to obtain representative meteorological observations at urban sites. In: Instruments and observing methods Report No81
Tetsuji Yamada, K. Koike (2011)
Downscaling mesoscale meteorological models for computational wind engineering applications
Journal of Wind Engineering and Industrial Aerodynamics, 99

Publisher: Springer Journals
Copyright: Copyright © 2016 by The Author(s)
Subject: Energy; Energy Efficiency (incl. Buildings); Renewable and Green Energy; Energy Technology; Landscape/Regional and Urban Planning
eISSN: 2363-9075
DOI: 10.1186/s40984-016-0020-4
Publisher site: See Article on Publisher Site

Abstract

Urban and building energy simulation models are usually driven by typical meteorological year (TMY) weather data often in a TMY2 or EPW format. However, the locations where these historical datasets were collected (usually airports) generally do not represent the local, site specific micro-climates that cities develop. In this paper, a humid sub-tropical climate context has been considered. An idealised “urban unit model” of 250 m radius is being presented as a method of adapting commonly available weather data files to the local micro-climate. This idealised “urban unit model” is based on the main thermal and morphological characteristics of nine sites with residential/institutional (university) use in Hangzhou, China. The area of the urban unit was determined by the region of influence on the air temperature signal at the centre of the unit. Air temperature and relative humidity were monitored and the characteristics of the surroundings assessed (eg green-space, blue-space, built form). The “urban unit model” was then implemented into micro-climatic simulations using a Computational Fluid Dynamics – Surface Energy Balance analysis tool (ENVI-met, Version 4). The “urban unit model” approach used here in the simulations delivered results with performance evaluation indices comparable to previously published work (for air temperature; RMSE <1, index of agreement d > 0.9). The micro-climatic simulation results were then used to adapt the air temperature and relative humidity of the TMY file for Hangzhou to represent the local, site specific morphology under three different weather forcing cases, (ie cloudy/rainy weather (Group 1), clear sky, average weather conditions (Group 2) and clear sky, hot weather (Group 3)). Following model validation, two scenarios (domestic and non-domestic building use) were developed to assess building heating and cooling loads against the business as usual case of using typical meteorological year data files. The final “urban weather projections” obtained from the simulations with the “urban unit model” were used to compare the degree days amongst the reference TMY file, the TMY file with a bulk UHI offset and the TMY file adapted for the site-specific micro-climate (TMY-UWP). The comparison shows that Heating Degree Days (HDD) of the TMY file (1598 days) decreased by 6 % in the “TMY + UHI” case and 13 % in the “TMY-UWP” case showing that the local specific micro-climate is attributed with an additional 7 % (ie from 6 to 13 %) reduction in relation to the bulk UHI effect in the city. The Cooling (Continued on next page) * Correspondence: L.Bourikas@soton.ac.uk Energy & Climate Change Division, Sustainable Energy Research Group (SERG), Faculty of Engineering and the Environment, University of Southampton, Southampton SO17 1BJ, UK Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Bourikas et al. Future Cities and Environment (2016) 2:7 Page 2 of 23 (Continued from previous page) Degree Days (CDD) from the “TMY + UHI” file are 17 % more than the reference TMY (207 days) and the use of the “TMY-UWP” file results to an additional 14 % increase in comparison with the “TMY + UHI” file (ie from 17 to 31 %). This difference between the TMY-UWP and the TMY + UHI files is a reflection of the thermal characteristics of the specific urban morphology of the studied sites compared to the wider city. A dynamic thermal simulation tool (TRNSYS) was used to calculate the heating and cooling load demand change in a domestic and a non-domestic building scenario. The heating and cooling loads calculated with the adapted TMY-UWP file show that in both scenarios there is an increase by approximately 20 % of the cooling load and a 20 % decrease of the heating load. If typical COP values for a reversible air-conditioning system are 2.0 for heating and 3.5 for cooling then the total electricity consumption estimated with the use of the “urbanised” TMY-UWP file will be decreased by 11 % in comparison with the “business as usual” (ie reference TMY) case. Overall, it was found that the proposed method is appropriate for urban and building energy performance simulations in humid sub-tropical climate cities such as Hangzhou, addressing some of the shortfalls of current simulation weather data sets such as the TMY. Keywords: Idealised urban unit model, Micro-climate simulations, Urban weather projections, Cities Introduction locations within the city develop local, site specific Engineering and urban design practise often relies on micro-climates that will not always be represented by thermal simulation modelling in order to achieve regula- these datasets [6]. tory compliance and support decision making on urban Over recent decades many cities have expanded rapidly and building design, as well as sizing of building energy through higher building densities, larger building heights systems that directly effects energy consumption. Urban and growth towards suburban and rural areas. This expan- and building energy simulation models are usually sion results in changes to the balance of the urban energy driven with hourly weather datasets for a ‘typical’ year budget [7] which consists of the radiation, sensible, latent [1]. In humid sub-tropical climates where air-conditioning and anthropogenic heat fluxes and the roughness caused is the primary concern the year format is that of an “aver- by the built environment [7]. These features of the built age year” (eg TMY2 or EPW). Common typical weather environment’s energy and mass equilibrium can cause at- year formats have the form of a synthetic year which is mospheric forcing which alters the local weather condi- compiled from multiple years to be representative of the tions [8] and contributes to the development of distinctive climatic conditions for the site of interest [2]. local micro-climates specific to a site’s characteristics and Heating, Ventilation and Air Conditioning (HVAC) morphology. The micro-climatic factor with the most no- systems are often oversized in order to reduce design ticeable variation regarding changes in the energy budget risk by ensuring that the system will cope with loads is the air temperature [9]. The urban air temperature (T) above the average year’s design day estimations [3]. This difference to the regional non-urban air temperature (eg affects the energy consumption of the building, potentially airport sites) is defined as the urban heat island (UHI) and the thermal comfort of the users and the indoor air qual- the magnitude of this difference is known as the urban ity, whilst at the same time increasing the cost of oper- heat island intensity (UHII) [10]. ation and possibly maintenance for building services [4]. Field observations and experimental measurements The main concern in the case of building energy sys- have been used by various researchers in order to create tems’ oversizing is the way that energy use is estimated, algorithms for the adaptation of weather data to local more specifically the heating, cooling, dehumidification specific morphologies [11–13]. In Hong Kong, urban and mechanical ventilation loads. At the moment, in many heat island intensity observations were used in order to countries, weather data files with a typical meteorological adapt a weather data file and use it in cooling load simu- year (TMY2 or EPW) format are used for building regula- lations [14]. In a field study, Wong et al. [15] pointed tion compliance calculations [5]. However, compliance to out that density, the ratio of building height to building the regulations does not necessarily ensure occupants’ sat- footprint area and especially the green surface area isfaction or represent the real operational conditions of (vegetation) can substantially influence the development the building services. One important consideration for the of the micro-climate. Summertime temperature observa- use of the typical year weather time series in energy con- tions in London suggest additional correlations of UHI sumption predictions and simulations is the representa- intensity with the distance from a thermal hotspot and tiveness ofthelocationwhere thesourcedatahave wind direction [16]. been collected. Many of these locations are, due to histor- More recently, a model was introduced for the estima- ical data availability, airports near large cities. However, tion of the UHI intensity in urban centres and the Bourikas et al. Future Cities and Environment (2016) 2:7 Page 3 of 23 modification of weather data files in the UK [17]. This for the prediction of monthly changes due to the UHI model is based on the statistical analysis of temperature effect in air temperature, RH and wind speed [25]. The data from UK based weather stations from the period Town Energy Balance and an integrated building energy 1961 to 2006. The urban thermal centres were defined model have been coupled with other schemes (ie rural by assigning thresholds to predefined urban fraction pa- station, vertical diffusion and urban boundary layer rameters for each UK grid cell and its surroundings [17]. model) and they have been used for the adaptation of The urban fraction limits were calibrated with visual in- rural air temperature and relative humidity hourly values spection of the fit of calculated ranges for the urban to the urban heat island effect [26]. The so-called Urban fraction parameters to real UK cities [17]. Therefore, in Weather Generator requires a large number of initial order to apply this methodology in places other than the parameters that vary from construction elements pro- UK, this method should be repeated for similar or perties and characteristics (eg albedo and initial equivalent local land cover types and temperature data temperature) to the urban (eg façade to site ratio) and spatial resolution under the same baseline assumptions reference rural site morphology [27]. The land cover of urban fraction limits. analysis for the creation of the simplified “urban unit In a numerical weather prediction approach, the model” can provide most of the required parameters for Urban Canyon model [18] estimates the heat fluxes and the initialisation of the Urban Weather Generator. The the air temperature distribution in an urban street can- use of the 3D “urban unit model” in numerical simula- yon configuration. It introduces a zonal model that di- tion modelling was chosen instead as it can provide an vides the air volume between the buildings into cells approximation of the vertical development of air where the heat and mass balance can be estimated for temperature and relative humidity in the roughness sub- each cell [18]. The results have been used to introduce layer and it can capture the majority of the transient the climatic severity index (CSI) that can assess the im- physical processes at street level. pact of the street canyon morphology on the heating The aim of this study is to provide a comprehensive and cooling demands of locally based buildings [18]. methodology that allows for an integration of urban In a very interesting development, meso-scale weather micro-climate conditions into standard weather datasets forecasting and numerical weather prediction-CFD such as the TMY. In order to achieve this, this paper in- models were coupled in order to predict the micro-scale troduces an idealised “urban unit model” (on a 250 m weather development in wind farms and urban environ- radius) that represents the main thermal and morpho- ments [19, 20]. In these coupling schemes, the numerical logical characteristics of urban sites at street level in the weather prediction models were downscaled in order to neighbourhood scale. This model, which was produced provide the initial and boundary conditions for the using statistical land cover and urban morphology ana- micro-scale models [21]. Despite the benefits from coup- lysis, can be used with simulations as a method of adapt- ling meso- to micro-scale models, the differences in the ing commonly available weather data files to a local horizontal and vertical scales and the time resolution be- specific micro-climate. This methodology is used to tween these models pose a challenge to their widespread adapt air temperature and relative humidity (RH) from successful application. The grid size differences and sub- the TMY file to the effects of local site specific morph- scale grid implications regarding the physical processes ology on urban weather development. The “urban wea- at play are often a threat to the analysis of the results ther projections” resulting from this adaptation were and their correct physical interpretation [22]. In then used for comparing the degree days amongst the addition, numerical models are computationally de- reference TMY file, the TMY file offset for a bulk hourly manding in terms of both power and time. Therefore, UHI intensity (local) and the TMY adapted for the urban their use in micro-scale modelling is usually restricted to weather projections (local + micro). Finally, the TMY file small domains and relatively simple geometries [23]. generated by this methodology was used for the dynamic A more common approach is the coupling of numer- thermal simulation of heating and cooling loads in a do- ical weather prediction models with analytical micro- mestic and a non-domestic building scenario. The po- scale models. The analytical urban canopy models esti- tential improvements in the estimation of building mate the energy balance development and momentum energy consumption were assessed against the business transfer in relation to the urban canopy’s morphology as usual case of applying typical meteorological year data and produce single area-averaged values in order to be files in places with a humid sub-tropical climate such as used as forcing in the scale (typically meso-scale) that the case study city of Hangzhou in China. the weather prediction model resolves [24]. Ren et al. [25] have created a “morphed” TMY file for the city of Methodology Melbourne using a regional weather forecasting model An important issue for the successful application of any coupled with an urban canopy parameterisation scheme existing micro-climatic model are the input data Bourikas et al. Future Cities and Environment (2016) 2:7 Page 4 of 23 prerequisites. Detailed information on the surface mate- Hangzhou (30°15'N 120°10'E) in Zhejiang Province, rials’ properties, morphological and other modelling pa- China [30, 31]. The sensors were installed on lampposts rameters is not however, always readily available. There at a level 3 to 5 m above ground. They logged air is a need for models to estimate the weather conditions temperature at 11-Bit (0.0625 °C) resolution and relative within urban areas as a function of time and urban humidity at 12-Bit (0.04 %) resolution [32]. The manu- morphology [28, 29]. The “urban unit model” introduced facturer stated air temperature accuracy is +/-0.5 °C and above has, therefore, been designed to be as general as the RH accuracy is +/- 5 % RH [32]. For calibration pur- possible in order to facilitate widespread use in building poses, air temperature (°C) and RH (%) readings within thermal and urban design simulations. Ideally, a visual 1 min intervals were compared to the readings of two evaluation (or an automated GIS platform) would be separate thermocouples in an environmental chamber at used to decide the urban class for the site of interest. -10, 0, 10 and 40 °C [33]. All sensors operated within the This would be enough to enable offsetting of the refer- reported accuracy margin and fitted well to the thermo- ence (TMY or real time non-urban) hourly air couple measurements [33]. temperature (T) and relative humidity (RH) for the se- According to Oke [34] air temperature and relative lected urban class for different seasonal weather forcing. humidity observations in the surface layer (3x above A small size neighbourhood was selected as the scale average building height) can be expected to be represen- of interest. An “urban unit” has been defined as an area tative of an area ranging from 100 m to several hundred forming disk (with a radius of 250 m) around the centre meters in a direction upwind and around each sensor. It of a neighbourhood where a temperature and relative is to be expected that most of the sensors were collect- humidity sensor was located. Air temperature and rela- ing measurements in the roughness sub-layer and not in tive humidity were monitored at 26 urban sites in the surface layer [35] (Table 1). The location of the Table 1 Siting of selected sensors in Hangzhou and representation of the size of the “urban unit” (Top Right) Siting of the sensor (blue bullet point) Circle of influence Sensor on a lamppost close to a tall building wall Sensor on a lamppost next to a road Bourikas et al. Future Cities and Environment (2016) 2:7 Page 5 of 23 measurement sites was carefully selected to have as lower than the group average air temperature and their homogeneous characteristics as possible for a large city vegetated surface area is large with an early peak within and the sensors were installed a reasonable distance the first 100 m from the sensor (Fig. 2). Interestingly, away from sites of noticeable surface type change [30]. sensor 12 (dark cyan; -1.0 °C) which also shows a nega- The urban unit’s size, ie the radius of the disk around tive temperature departure from the group mean has a the “centre of a neighbourhood”, has been determined small vegetated area approximately 150 m away from by assessing the vegetation cover’s influence on air the sensor, comparable with those at locations with temperature. For this, twelve urban sites were selected higher than the group average temperature. Its negative across the city centre of Hangzhou, North of the Qiantang trend can, however, probably be explained with the steep River (Fig. 1). rise in vegetated surface area at the annular area from Each sensor was considered to be the centre of con- 150 to 200 m, showing that the influence of vegetation centric circles at radii of 10, 25, 50, 100, 150, 200, 300, remains strong at this distance (Fig. 2). 400 and 500 m. The footprint of the vegetated surface Further evidence comes from the comparison between was estimated and apportioned to these annular areas. sensor 1 (black, +0.5 °C) and sensors 2, 5 and 9 (pink, The analysis was carried out for 14 weeks of hourly data red, gold; +0.2 °C). The regression trend lines indicate collected during summer 2013. The air temperature that the location where sensor 1 resides is warmer than (Tair) in Fig. 2 is the average air temperature departure the locations of sensors 2, 5 and 9 despite the larger veg- from the mean temperature of the 12 sensors in this etated surface within a 0 to 100 m radial distance from group for this 14 week period. The vertical shaded refer- the sensor. In the case of sensors 2, 5 and 9 the vege- ence line marks the 250 m radius from the centre (ie the tated surface area peaks occur later at distances from sensor). The legend shows the mean summer air 100 to 150 m showing the persistent impact of the temperature departure from the group’s mean and the green-space area. In addition, site 5 (red; +0.2 °C) has a goodness of fit of the non-linear regression line in similar vegetated surface area to site 1 (black; +0.5 °C) parentheses. with the only difference being a delay of the peak, that is It is to be expected that the influence of green-space is seen at distances 50 to 100 m farther out. Sensor 11 larger when closer to the sensor and that it diminishes if (dark blue, -0.2 °C) has a lower air temperature than the it is more towards to the outer annular areas. This is group average but the green-space percentage peak oc- shown in Fig. 2. The strong influence of vegetated areas curs after the 250 m radius border. However, in this case close to the sensor is evident from the first peak of a the low air temperature is mainly attributed to the site’s vegetated surface area in comparison to the general proximity to a large wetland (ie Xixi). Based on these re- trend in the group of sensors. The locations of sensors 3 sults a representative circular “urban unit” has been de- (dark green; -1.4 °C) and 10 (purple; -0.4 °C) have a fined with a 250 m radius. It is expected that the total Fig. 1 Location of the 12 sites for the urban unit size assessment (Left) and a typical fisheye image used for estimating the Sky View Factor in Hangzhou (Right). (Note: The Mantou Mountain, National Principle Weather station’s location (reference, typical meteorological year file source) is marked as NP. The colour scheme of the bullet points in the map is consistent with Fig. 2 Bourikas et al. Future Cities and Environment (2016) 2:7 Page 6 of 23 Fig. 2 Regression lines of the percentage vegetated area in each annular area on the distance from the centre (ie the sensor) area of about 200,000 m (250 m radius) surrounding Step 1. A simplified model for urban micro-climatic the sensor will be the representative part of the source simulations area for the air temperature and relative humidity signal. The generation of the “urban unit model” is based on The size of the proposed urban unit also agrees with the following steps: other authors’ studies on the windward distance from a point of roughness or thermal change (~200–500 m) (1)The land cover analysis starts with the inspection and the internal boundary layer extent in zones within and selection of an aerial image of the sensor’s local climate classification schemes (r ~ 200–500 m) [36]. location in Google Earth [37]. The image with the Building on the urban unit as defined above, this study best orthographic projection and quality is selected adopted a combination between a local urban classification (ideally, a clear top plan view image). It is then scheme and the simulation of the local specific weather processed with SketchUp [38], a computer aided development for urban unit layouts in the case study re- design software, using the urban unit described gion with a residential/institutional (university, college) above with the data logger at the centre of the use. Micro-climatic simulation modelling was used to as- 250 m circular area (Fig. 4). sess the influence of the neighbourhoods’ morphology on (2)A selection of metadata (eg high resolution images the local, street level air temperature. The main advantage taken on site) is used in combination with the aerial of using such a surface classification scheme is that it pro- view to draw polygons for the vegetation (green vides generic input data for the simulation model [36] and colour), water (blue colour) and built-up (black that the simulation results can be attributed back to typical colour) surfaces (Fig. 4). The residual is designated urban morphology characteristics. as other impermeable surface (white colour) and There are three key parts to the methodology for gen- includes street and pavement surfaces. A set of erating the “urban weather projections” that result from morphological parameters is then calculated for each the combined urban classification and simulation model- site including: the mean building height H, the ling (Fig. 3): (Step 1) The creation of the idealised “urban roughness length z , the height to width aspect ratio, unit model” for the sites of interest (in this study 9 sites); the frontal area ratio λ , the building surface fraction (Step 2) The normalisation of the reference weather data F , the impervious surface fraction I and the r r with the local monthly UHI patterns for different wea- pervious surface fraction P . ther forcings and (Step 3); The adaptation of the UHI (3)Each urban site is classified according to the land adjusted hourly air temperature and RH data to account cover analysis into a “Local Climate Zone” following for the effect of the site specific generic morphology at an urban classification scheme developed by Stewart street level at the neighbourhood scale. and Oke [36]. Each zone (ie thermally homogenous These three components of the overall methodology region of uniform surface characteristics) in the are described in the following: scheme exhibits a distinctive diurnal temperature Bourikas et al. Future Cities and Environment (2016) 2:7 Page 7 of 23 Fig. 3 Methodology for adapting TMY files to include the effect of the local site specific morphology in cities development profile at sensor height (~1.5 to 3 m) (4)The urban morphology of these sites is further at the local scale [36]. The resulting Local Climate analysed for 5 annular areas. The annular rings’ Zones (LCZs) describe 17 generic environments periphery has been defined at radii of 50, 100, 150, consisting of 10 zones for built-up (eg open 200 and 250 m. Urban morphological parameters high-rise) and 7 for non-urban land cover types and descriptive statistics are then calculated for all (eg scattered trees) [36]. Each zone is represented by a the annular areas (ie 0–50 m (red), 50–100 m set of ten morphological parameters and a descriptive (orange), 100–150 m (blue), 150–200 m (green), definition of the typical location and use of the urban 200–250 m (outer) as shown in Table 2, for an sites classified into a zone. Nine urban sites out of a example of site 2 in Fig. 5). total of 26 investigated areas in Hangzhou were (5)The generic, idealised “urban unit model” is classified as “Local Climate Zone 5” (LCZ5) (Fig. 5) constructed to have a similar planar area ratio and which denotes midrise buildings at a medium density mean weighted (footprint) building height to the arrangement [36]. This study focuses on these nine nine studied sites. The individual surface energy sites classified as “LCZ5”. balances are represented in the model by the Fig. 4 Digital elevation models and land cover have been built upon the aerial image (Left). Four main land surface types have been identified in the five concentric annular areas with the sensor located at the centre of a disk of 250 m radius (Right) Bourikas et al. Future Cities and Environment (2016) 2:7 Page 8 of 23 Fig. 5 The nine studied urban units classified into LCZ5 (locations, marked with blue bullet points) in Hangzhou (Left). (Note: The Mantou Mountain, National Principle Weather station’s location (reference, typical meteorological year file source) is marked as NP). An example of the land cover analysis is shown for Site 2 (Right; radius 250 m). Buildings are marked with black and vegetation with green pervious, impervious and building footprint surface generic “urban unit model” comprises of square based area ratios. The morphology characteristics of each boxes (ie blocks) with a non-uniform height in a stag- annular area are based on the median value gered irregular array (3 in Table 4). observations from the statistical analysis for the nine The staggered block array (3 in Table 4) has a north– LCZ5 urban units (Table 3). The median was south orientation. Each block has a base equal to the preferred over the mean because it is not affected by computational grid cells’ horizontal dimensions. For ex- extremely low or high values and the calculated ample, the minimum building footprint area in this distributions were rather skewed than normal. study was 64 m because the computational grid cells had horizontal dimensions of 8 m (x) x 8 m (y). Regard- In idealised models building geometry is usually ing the vertical grid dimension (z), any buildings and substituted with arrays of cubes. Common methods use vegetation in the urban unit model had a minimum cubes in staggered or aligned arrays (eg [35], [39–42]). height equal to the height of the first vertical grid cell (ie Cubes in regular arrays are spaced in repeated intervals 0.50 m). Each block can represent a building (black), equal at all directions to the cube’s edge length (ie aspect vegetated surface (green, grass or tree) or water surface ratio = 1) (1 in Table 4). In this study however, the (blue, zero height). The residual space between the Table 2 Analysis of the morphological parameters for each annular area for a site with a LCZ5 classification (site 2 in Fig. 5) Site 2. R-r (m) P I F z σ =H r r r 0 H 0–50 (red) 0.05 0.64 0.31 1.5 0.52 50–100 (orange) 0.11 0.62 0.27 1.3 0.45 100–150 (blue) 0.17 0.56 0.27 1.2 1.08 150–200 (green) 0.16 0.60 0.24 1.7 0.99 200–250 (outer) 0.19 0.56 0.25 2.1 0.87 P pervious surface fraction, I impervious surface fraction, F building surface fraction, z roughness length, σ =H standard deviation of the building height r r r 0 adjusted for the area weighted average height Bourikas et al. Future Cities and Environment (2016) 2:7 Page 9 of 23 th th Table 3 The median and the range (in brackets) of the key morphological parameters (10 to 90 centile) for the nine LCZ5 urban units shown in Fig. 5 Annulus R-r (m) H (m) P (%) I (%) λ (%) F (%) d (m) z (m) r r f r 0 0–50 20 (15–22) 7 (0–15) 72 (50–85) 17 (7–25) 21 (7–30) 6.7 (3.6–11.1) 1.8 (1.3–4) 50–100 18 (13–25) 11 (0–17) 61 (54–70) 15 (6–20) 27 (8–40) 8.9 (3.6–12.9) 1.4 (0.5–3) 100–150 20 (13–24) 15 (0–25) 65 (56–70) 13 (6–20) 23 (4–30) 7.6 (2.0–10.6) 1.8 (0.5–3) 150–200 17 (12–20) 15 (5–25) 60 (50–66) 14 (10–17) 24 (19–28) 8.6 (5.5–9.7) 1.2 (0.7–2) 200–250 19 (16–24) 15 (12–25) 56 (54–66) 13 (10–20) 23 (16–27) 8.3 (5.8–11.3) 1.8 (0.7–3) (R-r refers to the inner and outer radius of the annuli borders) H building footprint-area-weighted average height, P pervious surface fraction, I impervious surface fraction, λ frontal area ratio, F building surface fraction, d r r f r zero plane displacement height, z roughness length blocks represents the impervious surface (grey, eg roads, steps for the normalisation of the reference weather paved areas) (Fig. 4, right). The distance between the data are as follows: building blocks in each annular area is random and the number of the blocks representing buildings and vegeta- (1)((A) in Fig. 6) - The global horizontal solar radiation tion was defined by the estimated F and P ratios re- (GHR) from the typical meteorological year (TMY) r r spectively. The distribution of the blocks in each annular file (584570_CSWD) was used for creating three th area is similar in all notional quarter annuli (ie 1/4 of distinctive groups in order to account for different the total annular area). The changes to the packing weather forcing in each month (Fig. 6). Group 1 density and distance between the blocks produced a ran- represents days with overcast sky conditions and domly dispersed layout that is expected to better fit the rain events. Group 2 represents days with clear skies high spatial inhomogeneity of real cities than a regular and average or “highly likely” weather conditions for staggered cube array. the month, while Group 3 represents rather warm/ th th hot days with a clear sky (Fig. 6). The 25 and 75 st rd Step 2. Normalisation of the reference weather data with centiles (1 and 3 quartiles) of the GHR values the local UHI patterns (local scale) were used as the cut off points for the categorisation The reference weather station (official, NP in Fig. 5) for of the hourly TMY data into the groups. Group 1 Hangzhou is located at Mantou Mountain (30.23 N, for a given month contains all the days in that 120.17 E, at an elevation of 42 m). This weather sta- month that have hourly GHR values equal to or less th tion is the source of the typical meteorological year than the 25 centile GHR value for the respective (TMY) files for Hangzhou and it reports data with hours (lowest 25 % of GHR values). Similarly, Group Hangzhou international airport’s reference code 2 contains all the days in a month with hourly GHR (ZSHC) and the code 584570 in the World Meteoro- values within the interquartile range of the GHR logical Organisation’s (WMO) weather station list. The data for the respective hours (50 % of GHR values). Table 4 Common urban morphology representations and the idealised “urban unit model” in this study (3) Idealised models used to represent urban morphology (1) Uniform height and aspect ratio (2) Variable height and aspect ratio (3) Variable height, aspect ratio and shape of blocks plus water (blue) and vegetation (green) surfaces. Models (1) and (2) in table were adapted from [42] Bourikas et al. Future Cities and Environment (2016) 2:7 Page 10 of 23 Fig. 6 Methodology flowchart of the model’s validation and the generation of the adapted –“urbanised” weather dataset Therefore, Group 2 is expected to be the most (3)((C) in Fig. 6) - The data collected for each day from representative, “highly likely” weather forcing the sensors of the nine studied sites (LCZ5) scenario. Group 3 contains days that had most of highlighted in Fig. 5 as well as 10 additional the hourly GHR values in the upper quartile (highest (“sample”) sites in Hangzhou (Fig. 7) were allocated 25 % of GHR values). Here, the definition of “most to the weather forcing groups as defined in (2) ((B) of the hourly values” in this context relates to days in Fig. 6). Further to this, the observations from the with less than three hours with GHR values that do 10 “sample” sites were used to create a generic hourly not fit into the specific group and where these hours UHI pattern which relates to the difference of the are not between 12:00 and 16:00 h. sample sites’ hourly average observations to the Following the grouping of the days, descriptive reference weather station data for each month and statistics were calculated for the daily mean air weather forcing group. The hourly average temperature, temperature range, daily mean RH and observations from the nine studied sites (LCZ5) were maximum temperature for all the days in each then used to validate this method. Here we are group. The remaining days from the TMY file that comparing real measurements of the 9 LCZ5 sites did not fit into any category were then distributed with the simulation results of the “urban unit model” into either group 1, 2 or 3 according to their forced with measured data from the reference (TMY matches of daily mean air temperature, temperature source) weather station offset by the generic UHI range, daily mean RH and maximum temperature. effect as measured in the 10 “sample” sites. (2)((B) in Fig. 6) - The descriptive statistics (ie mean T, (4)The normalisation of the TMY reference weather T range, mean RH, max T) of (1) were also calculated data with the generic UHI group patterns (T and RH for observations from the reference weather station offsets) was based on a simple offset of the hourly (TMY source) for the period from December 2012 to mean temperature and RH values for each weather December 2013 (Fig. 6). The results were then forcing group (Eq. 1). compared with the descriptive statistics ((A) in Fig. 6) of the weather forcing groups from the TMY data file as determined in (1). Following this, the individual days TRðÞ H urb ¼ TRðÞ H ref Group;hr Group;hr þ UHIðÞ RH offset ð1Þ of the 2012–2013 observations from the reference Group;hr weather station dataset were distributed into the monthly weather forcing groups according to the four Where T(RH)urb is the air temperature (RH) after the criteria (ie mean T, T range, mean RH, max T from adjustment to the hourly UHI (RH offset) pattern, TMY data analysis) in descending order of weighting. T(RH)ref is the reference air temperature (RH) from the Bourikas et al. Future Cities and Environment (2016) 2:7 Page 11 of 23 Fig. 7 Location of the 10 “sample” sites (purple bullet points) used for the assessment of generic UHI patterns in Hangzhou, China in relation to the 9 studied (LCZ5, blue bullet points) sites TMY weather file and UHI (RH offset) is the positive or from the simulations express the weather change at street negative air temperature (RH) offset due to the urban level in relation to the baseline-reference weather (eg wea- heat island effect for each weather forcing group and ther at airport sites or non-urban sites) caused by the ef- hour respectively. fect of the site specific morphology in the city. These projections have the format of an additional hourly offset Step 3. Adaptation of the “localised” TMY data to to the “localised” dataset for each weather forcing group. include the effects of the site specific morphology ENVI-met is a three dimensional non-hydrostatic nu- (local scale + morphology = micro scale) merical micro-climatic model that couples an atmos- The idealised “urban unit model” (250 m radius) intro- pheric, a soil and a one-dimensional (1-D) vegetation duced above was implemented into micro-climatic simu- model and the surface energy balance. The atmospheric lations using a computational fluid dynamics – surface model is based on incompressible Reynolds averaged energy balance analysis tool (ENVI-met Version 4). In this Navier Stokes (RANS) equations [43]. Wind speed and final stage of data processing the hourly “localised” data direction remain constant during the simulation. The generated according to methodology section (2) was used effect of the surrounding urban environment on the to initialise and force the hourly weather conditions in the turbulent kinetic energy and the turbulent energy dissi- simulation. The “urban weather projections” resulting pation rate were modelled using cyclic (periodic) lateral Bourikas et al. Future Cities and Environment (2016) 2:7 Page 12 of 23 and outflow boundary conditions (ie turbulence from Two scenarios were created for the assessment of the last grid cells at outflow boundary are copied to the first solution’s sensitivity to the horizontal grid dimensions; grid cell at the inflow boundary) [43]. Air temperature one for winter that represents cold clear sky conditions (at 2 m above ground) and relative humidity at the in- and one for summer that represents hot weather with a flow boundary were forced hourly with the “localised” clear sky in Hangzhou. The results from the case studies air temperature and relative humidity TMY data for in both scenarios showed that the air temperature differ- 24 h. The turbulence field was updated every 10 min; ences between the cases are less than 0.5 °C and for RH solar radiation was modelled with a dynamic time step the difference was in the range of 2–3 % RH units. (ie shorter when solar radiation is near its peak (1 s) and The sensitivity to the vertical grid cell dimension was longer during morning and afternoon (2 s)); the internal not assessed because it was considered important to a) temperature of buildings (free running) is calculated ac- have a solution at the height between 2.5 and 5 m above cording to the heat transfer through walls and roofs, the ground (ie middle of grid cells, solution at 2.80, 3.44, where all walls and roofs have the same thermal trans- 4.20 and 5.09 m) where the observations have been col- mittance and albedo. The spin-up period was set to 4 h lected and b) have at least 10 grid cells in the lower (starting at 20:00 on the day before the simulated day). 20 m of the domain and an expansion ratio below 20 % The computational domain in ENVI-met comprises an for the cells above the 20 m threshold. equidistant grid that can be compressed or stretched in In addition, two scenarios have been investigated for the vertical (z, height) dimension by using an expansion assessing the sensitivity of the model to the distribution ratio but there is no option for the local refinement of and the amount of green-space in the model. The first the horizontal computational grid. The starting grid cell scenario (Scenario 1) compared the air temperature at height, ie the height for the cell in contact with the 3.5 m height above ground (T ) in Case 1, where the 3.5m ground surface, was set to 0.5 m and the grid remained vegetation was distributed according to the statistical re- equidistant below the height of 2.5 m with a grid cell sults from the land surface analysis, with Case 2, where spacing equal to dz = 0.5 m. The combination of a 0.5 m all the vegetation surface area was moved to the centre starting grid cell height with an 18 % grid height expan- of the urban unit and Case 3 where the vegetated area sion ratio above 2.5 m resulted in a vertical grid with 16 was moved towards the outer annuli (Fig. 8). The pervi- grid cells at the lower part of the domain (ie the lower ous surface area ratio (ie 0.15) remained the same for 20 m within the roughness sub-layer). The urban unit’s the urban unit model. 250 m radius resulted in 3D computational grids of 72 × Scenario 2 compared the air temperature (T ) from 3.5m 72 × 28 grid cells with a horizontal resolution of 8 m. Case 1 with 5 additional Cases (4 to 8) that had increas- A sensitivity analysis of the simulation results showed ing ratios of pervious surface area that was distributed that an increase of the horizontal resolution from 8 to evenly (same percentage) in each annular area (Fig. 9). 6 m and 3 m (Table 5) delivered no significant change in Each case had 5 % RH points more vegetated surface the model output. The coarsening ratio was not constant area (ie in the form of grass) than the previous one up because of limitations set by the computational domain to a maximum of P =0.4 which represents the upper size and the fixed computational grid (ie the modelled limit for the “Local Climate Zone 5” classification. geometry should fit to an integer number of grid cells). The air temperature development in the idealised In addition, it was not possible to assess the grid sensitiv- “urban unit” for both scenarios was simulated with ity to computational grid dimensions below 3 m × 3 m ENVI-met (Version 4) for August 10, 2013 which repre- due to the simulation domain size limitations (ie 250 × sented a sunny hot day in Hangzhou. The sky was clear. 250 grid cells maximum). That is because the urban unit The previous four days had been dry with 41 °C max- has a diameter of 500 m and a number of grid cells in imum air temperature and similar weather conditions as proximity to the domain borders must remain empty. the day of the simulations. The analysis was conducted Grid cell resolutions coarser than 8 m were not assessed for 24 h from 00:00 China Standard Time (CST) to because they were deemed too low for the purposes of this 23:00 CST. study. Table 5 Case studies for the assessment of the solution’s sensitivity to grid resolution Case Horizontal grid Vertical grid resolution (dz) [m] Computational domain dimensions Coarsening ratio r i,i +1 resolution (dx,dy) [m] (x,y,z) [grid cells] (+no of nesting grids) Case 1 (3, 3) First 5 grid cells’ height: 0.5 m, from 2.25 m to 207 × 207 × 28 (+6) n.a. the top of the 3D domain: dz = 1.18 × dz n n-1 Case 2 (6, 6) Same as Case 1 104 × 104 × 28 (+5) r = 2.0 1,2 Case 3 (8, 8) Same as Case 1 72 × 72 × 28 (+4) r = 1.3 2,3 Bourikas et al. Future Cities and Environment (2016) 2:7 Page 13 of 23 Fig. 8 Distribution of the vegetated surface area in the urban unit for the cases considered in Scenario 1 of the green-space sensitivity analysis In both scenarios (green-space amount and distribu- successive hot, dry summer days, a reduction in soil tion) the modelled temperature varied less than 0.5 °C water content will negate, to a large extent, the cooling between the cases showing that the solution is not sensi- benefits of the added vegetation. tive to the distribution of the green-space in the model In the Chinese building regulations Hangzhou is clas- and the model performs as expected regarding the dif- sified into the Hot Summer - Cold Winter climate zone ferences in the amount of vegetation. [44]. In this zone, residential apartments are typically Specifically, the assessment of the impact of the vege- mixed mode with split air-conditioning and natural ven- tation’s location on the air temperature development in tilation [45]. For the purpose of assessing the potential the urban canopy revealed that the proximity to “green” improvements in the estimation of building energy con- – vegetated space can decrease the urban heat island in- sumption against the “business as usual” case of using tensity during night-time and the maximum day-time air TMY data files the heating and cooling degree days were temperature. The marginal difference between the cases calculated and compared amongst the reference TMY with a central allocation of the vegetated surface area file, the TMY file with a bulk hourly UHI offset and the and those were the vegetation was positioned at the “urbanised” TMY file after its adaptation to the “urban outer border of the urban unit is an indication that the weather projections”. The degree days have been calcu- distance to a vegetated area is not enough to alone pro- lated according to the data from the 584570_CSWD duce large cooling benefits during the day and attenuate TMY file for Hangzhou at a base temperature of 18 °C the night-time urban heat island intensity. In Scenario 2 [45] for heating and 26 °C [45] for cooling. The “urban (amount of green-space), an increase to the urban unit’s unit model” methodology was validated for each of the 4 permeable surface area showed a small decrease in the seasons and 3 different weather forcing conditions average air temperature across the urban unit. The case against the hourly average air temperature and RH ob- with the largest vegetated surface area had the lowest servations from the 9 studied sites on a given day repre- daily air temperatures. A shift was noted in the air sentative of the weather forcing conditions. The main temperature distribution towards a higher occurrence parameters for the “urban unit model” validation are frequency of temperatures at the cooler end. The differ- shown in Appendix 1. Looking at the dates given in ences between the cases were more evident in the aver- Appendix 1 the “urban unit model” validation simula- age surface temperatures. The results suggest that high tions were forced with the hourly weather data from the percentages of vegetated space can reduce the surface reference weather station (TMY source) overlaid with temperatures within the cities. There were, however, also the representative urban heat island effect of the weather strong indications that in places with a humid sub- forcing group as calculated from the 10 “sample” sites’ tropical climate such as Hangzhou, in the case of observations (and not the observed UHI during the Fig. 9 The different percentages of vegetated surface area in the “urban unit” for the cases in Scenario 2 of the green-space sensitivity analysis. Top plan view of the computational domain (Right) for the cases with P = 0.2 and P = 0.4 r r Bourikas et al. Future Cities and Environment (2016) 2:7 Page 14 of 23 simulated day) (see also Figs. 6 and 7). The hourly wea- Results and discussion of the “urban unit model” ther forcing for the “urban weather projections” simula- validation tions was undertaken with the hourly average air This evaluation of the model’s performance (Figs. 10, 11, temperature and RH for the respective weather forcing 12 and 13) showed that urban micro-climatic simula- group in the TMY file, adjusted by the representative tions using the idealised “urban unit model” captures to bulk hourly UHI intensity (same as in the validation within 1 °C the main characteristics of the diurnal air case) for the weather forcing groups. temperature development in all seasons. Fig. 10 Comparison of the observed (black line) and the modelled (red line) air temperature (left) and RH (right) at 3.5 m above ground for the 3 days in January 2013 representative for the weather forcing groups. (Time given in China Standard Time – CST: GMT + 8) Bourikas et al. Future Cities and Environment (2016) 2:7 Page 15 of 23 Fig. 11 Comparison of the observed (black line) and the modelled (red line) air temperature (left) and RH (right) at 3.5 m above ground for the 3 days in May 2013 representative for the weather forcing groups. (Time given in China Standard Time – CST: GMT + 8) If the model output is a perfect prediction we would have to be identical. The ENVI-met model would have expect the observed temperatures (black line; average of to perfectly forecast the average temperature/RH devel- hourly observations from the nine “LCZ5” sites, see opment at street level. Figs. 10, 11, 12 and 13) to be identical to the modelled In the weather forcing group 1 winter scenario (January temperatures (red line in Figs. 10, 11, 12 and 13). In this 13, 2013; Fig. 10 (top)) the modelled air temperature at case the UHI effect experienced by the 9 LCZ5 sites and 3.5 m above ground (Air Temp. [ C]; red line) is a very the 10 “sample” weather stations within the city would good fit to the observed temperature (black line; RMSE: Bourikas et al. Future Cities and Environment (2016) 2:7 Page 16 of 23 Fig. 12 Comparison of the observed (black line) and the modelled (red line) air temperature (left) and RH (right) at 3.5 m above ground for the 3 days in June 2013 representative for the weather forcing groups. (Time given in China Standard Time – CST: GMT + 8) 0.4, MAPE: 5 %, Table 6). The simulation results predict in the urban weather conditions’ prediction when com- the night-time UHI better than the reference weather sta- pared to the reference weather station. In the case of wea- tion (TMY source station) observations (circles in Figs. 10, ther forcing group 2 the simulated air temperature is a 11, 12 and 13). The RH results (red line in Fig. 10 (right)) good fit to the observations for most hours of the day are representative of the observed RH diurnal trend. The (Fig. 10 (middle)). The air temperature is overestimated simulated RH values are low compared to the average of early in the morning but the high values potentially better the hourly observations but still represent an improvement represent the urban conditions than the measured data Bourikas et al. Future Cities and Environment (2016) 2:7 Page 17 of 23 Fig. 13 Comparison of the observed (black line) and the modelled (red line) air temperature (left) and RH (right) at 3.5 m above ground for the 3 days in October 2013 representative for the weather forcing groups. (Time given in China Standard Time – CST: GMT + 8) from the reference weather station. In the weather forcing an inaccurate representation of thermal mass and heat group 3 case, winter daytime air temperatures are storage in the model [46] and the modelling of thermal underestimated with the model failing to predict the diffusivity by ENVI-met [47]. temperature peak around 13:00 h. Nevertheless, the The simulated RH in January for both the weather for- simulation results fit the observed data relatively well at cing group 2 and group 3 cases (Fig. 10), again replicates night-time, in the early morning and afternoon. The the daily observed trend and the RH predicted levels are failure to accurately predict the peak could be a result of comparable with the reference weather station observations Bourikas et al. Future Cities and Environment (2016) 2:7 Page 18 of 23 Table 6 Model performance indices determined for the winter and autumn simulations with the urban unit model Model performance indices T (RH) - winter (January) T (RH) - autumn (October) Group 1 Group 2 Group 3 Group 1 Group 2 Group 3 Mean Squared Error (MSE) 0.13 (56.1) 0.55 (116.1) 0.70 (258.5) 0.50 (61.6) 0.31 (261.2) 0.39 (287.1) Mean Bias Error (MBE) −0.14 (-7.4) 0.12 (-10) 0.16 (-15.4) −0.42 (-7.5) −0.17 (-15.9) −0.02 (-15.3) MSE systematic 0.11 (55.8) 0.49 (112.5) 0.40 (246.3) 0.33 (56.7) 0.22 (255.4) 0.23 (280.9) Root Mean Squared Error (RMSE) 0.36 (7.50) 0.74 (10.75) 0.84 (16.05) 0.70 (7.85) 0.56 (16.16) 0.63 (16.95) RMSE systematic 0.33 (7.47) 0.70 (10.61) 0.64 (15.70) 0.58 (7.53) 0.47 (15.98) 0.48 (16.76) MSE unsystematic 0.02 (0.73) 0.05 (3.05) 0.30 (11.17) 0.17 (4.89) 0.09 (5.85) 0.16 (6.26) RMSE unsystematic 0.15 (0.85) 0.23 (1.75) 0.55 (3.34) 0.41 (2.21) 0.31 (2.42) 0.40 (2.50) Mean Absolute Percentage Error (MAPE) 5 % (8 %) 21 % (12 %) 11 % (18 %) 3 % (9 %) 2 % (21 %) 2 % (21 %) Index of agreement d 0.92 (0.13) 0.97 (0.78) 0.98 (0.72) 0.95 (0.68) 0.99 (0.48) 0.99 (0.77) (especially when considering the 5 % units RH sensor ac- to the urban observations showing the existence of a ra- curacy) for the largest part of the day. ther small urban heat island effect in Hangzhou during The model underestimates the RH during early morn- spring. The RH in the urban unit model was again ing before 08:00 o’clock indicating a possible discrepancy underestimated with the error being acceptable (RMSE: between the modelled vegetation properties (ie amount 7, MAPE: 7 %, Table 7) in the case that represents over- and type of trees, grass) and reality. However, this differ- cast sky conditions (group 1, Fig. 11 (top)) but signifi- ence is not expected to have a significant effect on the cant (RMSE: 15, MAPE: 20 %, Table 7) in the weather model’s application because 1) the predicted RH values forcing group 3 case (ie clear sky, hot weather). Overall, are relatively close to the reference weather station ob- in the spring scenario the night-time urban heat island servations and 2) the largest discrepancy is early in the intensity was overestimated across all the weather for- morning and late at night in winter when typically dehu- cing groups. However, in May night-time the air midification is not an option when split AC units oper- temperature is still low and the air-conditioning demand ate in heating mode (or auxiliary heating sources are if any is expected to be minimal. In the weather forcing used instead). The satisfactory prediction of the expected group 3 case (Fig. 11 (bottom)) the high air temperature urban heat island during the night (ΔT )in at noon suggests that indoor temperatures are highly Case – reference all three cases is a further indication that this level of in- likely to exceed the comfort band threshold of 27 C accuracy is not detrimental to the overall function of the [48] creating a demand for cooling. The simulated air model. temperature peaks are a good fit to the urban observa- In the spring scenario (May 2013, Fig. 11) the air tions and in most cases they represent the urban wea- temperature was overestimated during night and early ther development better than the reference weather morning. The simulated air temperature in the after- station measurements. noon was representative of the observed air temperature In the summer scenario (June, Fig. 12), the air across all three weather forcing groups. The reference temperature predictions from all three weather forcing weather station observations were consistently very close groups fit the observations from the studied LCZ5 sites Table 7 Model performance indices determined for the spring and summer simulations with the urban unit model Model performance indices T (RH) - spring (May) T (RH) - summer (June) Group 1 Group 2 Group 3 Group 1 Group 2 Group 3 Mean Squared Error (MSE) 0.08 (48.5) 0.31 (91.3) 0.90 (226.5) 0.15 (77.2) 0.52 (65.2) 0.67 (88.3) Mean Bias Error (MBE) 0.15 (-6.8) 0.31 (-9.2) 0.84 (-14.4) −0.11 (-8.4) 0.23 (-6.9) 0.75 (-8.9) MSE systematic 0.04 (46) 0.11 (89.8) 0.80 (219.4) 0.04 (73.5) 0.08 (53.6) 0.56 (84.0) Root Mean Squared Error (RMSE) 0.29 (6.9) 0.56 (9.6) 0.95 (15.1) 0.38 (8.8) 0.72 (8.1) 0.82 (9.4) RMSE systematic 0.19 (6.78) 0.33 (9.47) 0.89 (14.81) 0.20 (8.56) 0.27 (7.32) 0.75 (9.17) MSE unsystematic 0.05 (1.96) 0.20 (2.45) 0.10 (8.37) 0.11 (3.71) 0.45 (11.60) 0.11 (4.73) RMSE unsystematic 0.22 (1.40) 0.45 (1.57) 0.32 (2.89) 0.33 (1.93) 0.67 (3.41) 0.33 (2.17) Mean Absolute Percentage Error (MAPE) 1 % (7 %) 2 % (11 %) 3 % (20 %) 2 % (9 %) 2 % (9 %) 2 % (13 %) Index of agreement d 0.78 (0.17) 0.97 (0.82) 0.98 (0.73) 0.94 (0.29) 0.98 (0.89) 0.98 (0.85) Bourikas et al. Future Cities and Environment (2016) 2:7 Page 19 of 23 reasonably well (RMSE: 0.4 to 0.8, MAPE: 2 %, Table 7). uncertainties involved in the initialisation of the model In June 2013, in the weather forcing group 3 case, the and the simulation itself. Furthermore, the model per- air temperature is over predicted at noon and early formance evaluation indices’ scores for air temperature morning. This recurring pattern in the simulation of hot (Tables 6 and 7) are comparable to other published work weather conditions in spring and summer can be the re- [46, 49]. The index of agreement d takes values in the sult of underrepresentation of the permeable surface and range {0, 1} with a value of 1 indicating a perfect match soil water content in the model [47]. The RH simulation between the model prediction and the observations results support this argument with the error of the simu- [46, 50]. The index of agreement scores low in the wea- lated RH in morning being in the range of 15 RH per- ther forcing group 1 results for all seasons. However the centage units. Mean Absolute Percentage Error of the air temperature Overall, the RH is underestimated during large parts was consistently less than 10 % in almost all cases and the of the day but the simulated RH levels are comparably systematic component of the root mean square error close to the observations with a RMSE around 10 RH (RMSEs) was lower than 0.5 C in most cases and always units and MAPE of 10 % (Table 7). lower than 1 C. The systematic component of the error In October (autumn, Fig.13) the model simulates the (ie RMSEs), which represents the error attributed to the urban heat island development during the night with the simulation and the error integrated into the initialisation simulated air temperature being close to the observa- estimates, should approach 0 [46]. The unsystematic com- tions from the LCZ5 sites. The modelled air temperature ponent should approach the value of RMSE [50]. The rela- development is more realistic for the urban environment tively low RMSE values in conjunction with the fact that than the measured reference temperature. At noon and the model results were closer to the urban observations under hot weather conditions (weather forcing group 3), than the reference measured values show that the model’s the model fails to accurately predict the peak but the performance is acceptable. simulated temperature is still an improvement in com- The accuracy of the RH prediction is most important parison with the reference air temperature observa- during the cooling season when dehumidification is re- tions. In general, domestic buildings in October are quired. In Hangzhou and other similar cities with a expected to be free-running (no heating or cooling). humid sub-tropical climate the full cooling season is ex- There is a small demand for cooling in September and pected to last from June to September [31, 51]. The for heating from mid-November onwards. The valid- validation showed that in June and for the summer sce- ation of the model showed that cooling demand in au- nario the error is below 10 RH percentage units. How- tumn is likely to be under estimated at noon and in the ever, the measured RH from the reference weather early afternoon but its prediction is largely improved at station fits better to the urban observations than the night when the air temperature simulation results fit simulated RH. Furthermore there is only a marginal RH the urban observations very well (RMSE: 0.6 to 0.7, difference between the urban and the reference sites. MAPE: ~2 %, Table 6). The simulated RH follows the Therefore, the urban weather projections were used to observed diurnal RH trend but in weather forcing adapt the air temperature only in the TMY file and not groups 2 and 3 the error is large with the RMSE equal the relative humidity. to 16 and 17 RH % units respectively and the MAPE in A full scale error analysis was not undertaken due to the range of 20 %. lack of available data/input parameters notably the This discrepancy might be a result of weather events that hourly global solar radiation and soil properties. The val- ENVI-met cannot simulate such as mist, haze and rain. idation has been restricted to days that are typical for th Specifically on the 12 of October there was mist until the weather forcing groups’ conditions. The main 08:00 in the morning when it dispersed and at the same sources of systematic error are the input initialisation time the model’s prediction accuracy increased. In the parameters, in particular hourly global solar radiation. evening the large error in the RH prediction can be attrib- Other sources of systematic error include the initial uted to the modelled soil water content and the position of boundary conditions, building and vegetation properties. some of the LCZ5 sites being close to large water bodies. Unsystematic errors might be a result of the thermal dif- In late autumn the RH plays little role as there is no dehu- fusivity [47] and turbulence modelling, the modelling of midification load associated with heating. In addition, the evapotranspiration and the total heat advection to the largest RH error in October was noticed early in the morn- atmosphere. ing and during the evening when the temperature is ex- pected to be within the comfort band of the thermostatic Results and discussion of the TMY file for set points and there will be no cooling requirement. Hangzhou adapted with the urban unit model Overall, the accuracy of the model can be viewed as Following the “urban unit model” validation the TMY satisfactory in relation to its purpose, considering the file for Hangzhou (584570_CSWD) was adapted to Bourikas et al. Future Cities and Environment (2016) 2:7 Page 20 of 23 include the “urban weather projections” (UWP) ac- the simulation results with the use of the TMY-UWP cording to the methods outlined above. A comparison file. amongst the reference TMY, the TMY overlaid with the bulk “city” UHI effect (TMY + UHI) and the micro- Conclusions scale “urban weather projections” (TMY-UWP) is The validation of the “urban unit model” and the results shown in Fig. 14. In the “TMY + UHI” HDD have de- of its implementation into micro-climatic simulations creased by 6 % from 1598 days in the reference TMY show that there is a potential for the simplification of and in the “TMY-UWP” file by 13 % showing that the urban site modelling and for the wider application of the local specific micro-climate attributes an additional 7 % method introduced in this paper as a tool for adapting heating load reduction to the bulk UHI effect of the typical meteorological weather data files to represent the city. This difference between the “TMY-UWP” and the neighbourhood scale of cities with a humid sub-tropical “TMY + UHI” files is a reflection of the thermal charac- climate. teristics of the specific urban morphology of the stud- Overall, the comparison amongst the reference (ie ied sites compared to the wider city. CDD in the “TMY 1598 HDD), a bulk “city” UHI effect and the adapted for +UHI” file are 17 % more than the reference TMY file the “urban weather projections” TMY files indicates that (207 days). The “TMY-UWP” file shows an additional there was a 6 % decrease in HDD that is attributed to 14 % increase in comparison with the “TMY + UHI” the bulk “city” UHI effect and an additional 7 % (ie from file. 6 to 13 %) that can be directly attributed to the local The impact of the change in HDD and CDD was specific urban morphology of the 9 LCZ5 sites. The total assessed in terms of heating and cooling loads for a increase of CDD from the reference (ie 207 CDD) to the domestic and a non-domestic building scenario in “urban weather projections” TMY files is in the range of Hangzhou. For this purpose, the “urbanised” (TMY- 30 %. This assessment indicates that cooling loads can UWP) and the reference TMY files were used within a be significantly underestimated in the business as usual dynamic thermal simulation tool (TRNSYS Version case of using the reference TMY file and that the urban 17.1) to model the annual heating and cooling load for effect on air temperature should not be neglected. both scenarios. The calculated heating and cooling loads The heating and cooling demand has been further for both the domestic and the non-domestic building calculated for a domestic and a non-domestic build- case show that the energy performance simulations with ing case in Hangzhou. It has been observed that in the reference TMY file result in an approximate 20 % both cases there is an approximate 20 % increase of over prediction of the heating load and 20 % under pre- the cooling load and a 20 % decrease of the heating diction of the cooling load demand in comparison with load. If typical COP values for a reversible air- Fig. 14 Comparison of the degree days in the reference TMY (TMY reference, gray; 1598 HDD, 207 CDD) with the TMY overlaid with the bulk “city” UHI effect (TMY + UHI, orange; 1495 HDD, 243 CDD) and the micro-scale “urban weather projections” (TMY-UWP, red; 1383 HDD, 279 CDD). T is the temperature threshold used for the calculation of the HDD and CDD. The area with the light blue background marks the base cooling period Bourikas et al. Future Cities and Environment (2016) 2:7 Page 21 of 23 conditioning system are taken as 2.0 for heating and Table 9 Main input parameters for the simulations with the urban unit model in June 2013 3.5for coolingthenthe total electricity consumption estimated with the use of the “urban weather projec- Input parameter Jun 14, Jun 23, Jun 18, Source 2013 2013 2013 tions” TMY file will be decreased by 11 % in com- Weather forcing Group Group 1 Group 2 Group 3 parison with the “business as usual” (ie reference Specific humidity 2500 m 8.9 12 11.3 [52] TMY) case.Thisisan interesting result showingthe (750 mbar) gr w/kg dry air impact highly efficient heat pumps and air- Prevalent wind direction 45 90 225 [53] conditioning systems can have on the electricity con- (N = 0 clockwise) sumption of cities with a humid sub-tropical climate. Wind speed 10 m ab. gr. m/s 3.1 2.8 3.1 [53] However, this assumes a cooling set-point of 26 °C. Roughness length z 0.1 0.1 0.1 [36] If a lower set-point is used the predicted energy sav- ings will be lost. Mean wall albedo 0.23 0.23 0.23 [49] In the majority of cases it would be difficult to justify Mean roof albedo 0.50 0.50 0.50 the additional level of analysis described here to develop Wall heat transmittance 1.4 1.4 1.4 −2 −1 the localised weather data file over the generic bulk W. m .K “city” UHI file. The method is shown to deliver an en- Roof heat transmittance 0.9 0.9 0.9 −2 −1 hancement which in large developments could be justi- W. m .K fied. Single, smaller/residential buildings can take the Underground soil temperature 294.3 299 300 [52] bulk UHI correction approach. This methodology would (Upper-Middle-Deep layer) K 292.7 294 293 also benefit initial stages of urban planning and could in- 290.6 291 291 form decisions on the use and the urban form of existing Underground soil humidity 38 % 39 % 38 % [52] and new developments in the city. (Upper-Middle-Deep layer) 38 % 37 % 38 % Future work is looking to expand this methodology to different urban morphologies and to transfer it to differ- 37 % 36 % 37 % ent climates. Appendix 1 Table 8 Main input parameters for the simulations with the Table 10 Main input parameters for the simulations with the urban unit model in January 2013 urban unit model in May 2013 Input parameter Jan 13, Jan 12, Jan 24, Source Input parameter May 17, May 10, May 22, Source 2013 2013 2013 2013 2013 2013 Weather forcing Group Group 1 Group 2 Group 3 Weather forcing Group Group 1 Group 2 Group 3 Specific humidity 2500 m 3.3 2.0 1.1 [52] Specific humidity 2500 m 8.3 7.2 3.9 [52] (750 mbar) gr w/kg dry air (750 mbar) gr w/kg dry air Prevalent wind direction 22 225 225 [53] Prevalent wind direction (N = 0 0 110 0 [53] (N = 0 clockwise) clockwise) Wind speed 10 m ab. gr. m/s 3.2 2.1 2.5 [53] Wind speed 10 m ab. gr. m/s 1.6 2.8 2.8 [53] Roughness length z 0.1 0.1 0.1 [36] Roughness length z 0.1 0.1 0.1 [36] 0 0 Mean wall albedo 0.23 0.23 0.23 [49] Mean wall albedo 0.23 0.23 0.23 [49] Mean roof albedo 0.50 0.50 0.50 Mean roof albedo 0.50 0.50 0.50 Wall heat transmittance 1.4 1.4 1.4 Wall heat transmittance 1.4 1.4 1.4 −2 −1 −2 −1 W. m .K W. m .K Roof heat transmittance 0.9 0.9 0.9 Roof heat transmittance 0.9 0.9 0.9 −2 −1 −2 −1 W. m .K W. m .K Underground soil temperature 278.4 277.3 280.6 [52] Underground soil temperature 292.7 293.2 296.0 [52] (Upper-Middle-Deep layer) K (Upper-Middle-Deep layer) K 279.6 279.5 281.4 290.3 288.9 291.0 282.6 283.0 282.4 288.0 287.2 288.0 Underground soil humidity 38 % 36 % 34 % [52] Underground soil humidity 35 % 37 % 33 % [52] (Upper-Middle-Deep layer) (Upper-Middle-Deep layer) 38 % 36 % 35 % 35 % 37 % 34 % 36 % 35 % 35 % 34 % 35 % 34 % Bourikas et al. Future Cities and Environment (2016) 2:7 Page 22 of 23 Table 11 Main input parameters for the simulations with the Author details Energy & Climate Change Division, Sustainable Energy Research Group urban unit model in October 2013 (SERG), Faculty of Engineering and the Environment, University of Input parameter October October October Source 2 Southampton, Southampton SO17 1BJ, UK. Urban Energy Systems, Faculty 18, 2013 05, 2013 12, 2013 3 of Civil Engineering, Bauhaus-Universität Weimar, Weimar, Germany. Centre for Sustainable Energy Technologies (CSET), University of Nottingham Weather forcing Group Group 1 Group 2 Group 3 Ningbo, Ningbo, People’s Republic of China. School of Architecture, Faculty Specific humidity 2500 m 5.3 0.6 0.8 [52] of Humanities and Social Sciences, University of Liverpool, Liverpool, UK. (750 mbar) gr w/kg dry air Faculty of Engineering, University of Nottingham, Nottingham, UK. Prevalent wind direction 0 225 200 [53] Received: 2 February 2016 Accepted: 30 June 2016 (N = 0 clockwise) Wind speed 10 m ab. gr. m/s 3.0 2.5 2.2 [53] Roughness length z 0.1 0.1 0.1 [36] References Mean wall albedo 0.23 0.23 0.23 [49] 1. Crawley DB (1998) Which weather data should you use for energy simulations of commercial buildings? In: ASHRAE Transactions. ASHRAE, Mean roof albedo 0.50 0.50 0.50 Atlanta, USA, pp 498–515 Wall heat transmittance 1.4 1.4 1.4 2. Hacker J, Capon R, Mylona A (2009) Use of climate change scenarios for −2 −1 W. m .K building simulation: the CIBSE future weather years. The Chartered Institution of Building Services Engineers, London, UK Roof heat transmittance 0.9 0.9 0.9 −2 −1 3. Thomas PC, Moller S (2006) HVAC system size - getting it right. In: Clients W. m .K Driving Innovation: Moving Ideas into Practice. Cooperative Research Centre Underground soil temperature 291.0 291.3 292.5 [52] for Construction Innovation, Brisbane, Australia (Upper-Middle-Deep layer) K 4. Burdick A (2011) Strategy Guideline: Accurate heating and cooling load 293.1 292.8 291.3 calculations. Oak Ridge, USA. 294.1 294.0 293.1 5. US Department of Energy (2016) EnergyPlus - Weather Data Sources. Available via https://energyplus.net/weather/sources. Accessed Underground soil humidity 33 % 32 % 29 % [52] 19 June 2016 (Upper-Middle-Deep layer) 34 % 33 % 32 % 6. Mylona A (2012) The use of UKCP09 to produce weather files for building simulation. Build Serv Eng Res Technol 33(1):51–62 34 % 33 % 32 % 7. Oke TR (1982) The energetic basis of the urban heat island. Q J R Meteorol Soc 108(455):1–24 8. Collier CG (2006) The impact of urban areas on weather. Q J R Meteorol Soc 132(614):1–25 Abbreviations 9. Taha H, Konopacki S, Gabersek S (1999) Impacts of Large-Scale Surface CDD, cooling degree day; CDH, Ccooling degree hour; CST, China standard Modifications on Meteorological Conditions and Energy Use: A 10-Region time; GHR, global horizontal radiation; GMT, Greenwich mean time; HDD, Modeling Study. Theor Appl Climatol 62(3):175–185 heating degree Dday; HDH, heating degree hour; HVAC, heating, ventilation 10. Memon RA, Leung DYC, Liu CH (2009) An investigation of urban heat island and air conditioning; LCZ, local climate zone; MAPE, mean absolute intensity (UHII) as an indicator of urban heating. Atmos Res 94(3):491–500 percentage error; MBE, mean bias error; MSE, mean squared error; RANS, 11. Jusuf ST, Wong NH (2009) Development of empirical models for an Reynolds averaged Navier stokes; RH, relative humidity; RMSE, root mean estate level air temperature prediction in Singapore. Paper presented at squared error; TKE, turbulent kinetic energy; TMY, typical meteorological year; 7th International Conference on Urban Climate. June 29 - July 3, UHI, urban heat island; UWP, urban weather projections Yokohama, Japan. 12. Kolokotroni M, Davies M, Croxford B, Bhuiyan S, Mavrogianni A (2010) A Acknowledgement validated methodology for the prediction of heating and cooling energy This work is part of the activities of the University of Southampton’s Energy demand for buildings within the Urban Heat Island: Case-study of London. and Climate Change Division and the Sustainable Energy Research Group Sol Energy 84(12):2246–2255 (www.energy.soton.ac.uk) on cities and infrastructure. It is partly supported 13. Kolokotroni M, Ren X, Davies M, Mavrogianni A (2012) London's urban heat by the EPSRC Grant EP/J017698/1, “Transforming the Engineering of Cities to island: Impact on current and future energy consumption in office Deliver Societal and Planetary Wellbeing” and EP/K012347/1, “International buildings. Energy Buildings 47:302–311 Centre for Infrastructure Futures (ICIF)”. The installation work of the sensors’ 14. Chan ALS (2011) Developing a modified typical meteorological year network in Hangzhou and Ningbo is supported by the Ningbo Natural weather file for Hong Kong taking into account the urban heat island Science Foundation (No. 2012A610173) and the Ningbo Housing and Urban- effect. Build Environ 46(12):2434–2441 Rural Development Committee (No. 201206). 15. Wong NH, Jusuf ST, Tan CL (2011) Integrated urban microclimate assessment method as a sustainable urban development and urban design Authors’ contributions tool. Landsc Urban Plan 100:386–389 LB developed the methodology, the “urban unit model” and managed the 16. Watkins R, Palmer J, Kolokotroni M, Littlefair P (2002) The London heat work in this study. PABJ provided guidance, reviewed and supervised all island: results from summertime monitoring. Build Serv Eng Res Technol stages of this study. ABSB contributed to the management of this study and 23(2):97–106 reviewed the final outcome. MFJ contributed to the development of the 17. Kershaw T, Sanderson M, Coley D, Eames M (2010) Estimation of the urban initial idea for this study and reviewed parts of the work and the final heat island for UK climate change projections. Build Serv Eng Res Technol outcome. TS installed and managed the sensors network in China, collected 31(3):251–263 the observations and contributed to their analysis. DHCC contributed to the 18. De La Flor FS, Doḿ ınguez SA (2004) Modelling microclimate in urban development of the sensors network in China and supervised the data environments and assessing its influence on the performance of analysis. JD contributed to the management of the sensors network and surrounding buildings. Energy Buildings 36(5):403–413 reviewed the data analysis and parts of this study. All authors read and 19. Chen F, Kusaka H, Bornstein R, Ching J, Grimmond CSB, Grossman-Clarke S, approved the final manuscript. Loridan T, Manning KW, Martilli A, Miao S, Sailor D, Salamanca FP, Taha H, Tewari M, Wang X, Wyszogrodzki AA, Zhang C (2011) The integrated Competing interests WRF/urban modelling system: development, evaluation, and applications to The authors declare that they have no competing interests. urban environmental problems. Int J Climatol 31(2):273–288 Bourikas et al. Future Cities and Environment (2016) 2:7 Page 23 of 23 20. Yamada T, Koike K (2011) Downscaling mesoscale meteorological models 43. Bruse M, Fleer H (1998) Simulating surface–plant–air interactions inside for computational wind engineering applications. J Wind Eng Ind Aerodyn urban environments with a three dimensional numerical model. Environ 99(4):199–216 Model Softw 13(3–4):373–384 21. Tewari M, Kusaka H, Chen F, Coirier WJ, Kim S, Wyszogrodzki AA, Warner TT 44. Wang H et al (2014) Feasibility and optimization of aerogel glazing system (2010) Impact of coupling a microscale computational fluid dynamics for building energy efficiency in different climates. Int J Low-Carbon model with a mesoscale model on urban scale contaminant transport and Technol 0:1–8 dispersion. Atmos Res 96(4):656–664 45. Wang X, Altan H, Kang J (2015) Parametric study on the performance of green residential buildings in China. Frontiers of Archit Res 4(1):56–67 22. Schlünzen KH (2010) Joint modelling of obstacle induced and mesoscale 46. Middel A et al (2014) Impact of urban form and design on mid-afternoon changes-current limits and challenges. Paper presented at The Fifth microclimate in Phoenix Local Climate Zones. Landsc Urban Plan 122:16–28 International Symposium on Computational Wind Engineering (CWE2010). 47. Maggiotto G, Buccolieri R, Santo MA, Leo LS, Di Sabatino S (2014) Validation May 23-27, 2010, Chapel Hill, North Carolina, USA. of temperature-perturbation and CFD-based modelling for the prediction of 23. Baklanov A, Martilli A, Grimmond CSB, Mahura A, Ching J, Calmet I, Clark P, the thermal urban environment: the Lecce (IT) case study. Environ Model Esau I, Dandou A, Zilitinkevich S, Best MJ, Mestayer P, Santiago JL, Tombrou Softw 60:69–83 M, Petersen C, Porson A, Salamanca F, Amstrup B (2010) Hierarchy of Urban 48. Nikolou S (2011) Low carbon city living in Guangzhou, China. Unpublished, Canopy Parameterisations for different scale models. MEGAPOLI Project MSc Thesis. University of Southampton, Southampton, UK Scientific Report 10-04. Danish Meteorological Institute, DMI, Copenhagen, 49. Yang X et al (2013) Evaluation of a microclimate model for predicting the Available via http://megapoli.dmi.dk/publ/MEGAPOLI_sr10-04.pdf thermal behavior of different ground surfaces. Build Environ 60:93–104 24. Martilli A, Santiago J (2007) CFD simulation of airflow over a regular array of 50. Willmott CJ (1982) Some comments on the evaluation of model cubes. Part II: analysis of spatial average properties. Bound-Layer Meteorol performance. Bull Am Meteorol Soc 63(11):1309–1313 122(3):635–654 51. Hu T, Yoshino H, Jiang Z (2013) Analysis on urban residential energy 25. Ren Z, Wang X, Chen D, Wang C, Thatcher M (2014) Constructing weather consumption of Hot Summer & Cold Winter Zone in China. Sustainable data for building simulation considering urban heat island. Build Serv Eng Cities Society 6:85–91 Res Technol 35(1):69–82 52. National Centers for Environmental Prediction/National Weather 26. Bueno B, Norford L, Hidalgo J, Pigeon G (2012) The urban weather Service/NOAA/U.S. Department of Commerce. 2000, updated daily. generator. J Build Perform Simul 6(4):269–281 NCEP FNL Operational Model Global Tropospheric Analyses, continuing 27. Massachusetts Institute of Technology (2015) Urban heat island effect from July 1999. Research Data Archive at the National Center for modelling software: Urban Weather Generator version 3.0.0 MIT, Available Atmospheric Research, Computational and Information Systems via http://urbanmicroclimate.scripts.mit.edu/uwg.php. Accessed 26 Feb 2016 Laboratory. dx.doi.org/10.5065/D6M043C6. Accessed 26 Jan 2016. 28. Arnfield AJ (2003) Two decades of urban climate research: a review of 53. The Weather Underground (2014) Hangzhou weather data from Mantou turbulence, exchanges of energy and water, and the urban heat island. Int J mountain's National Principle WMO-listed weather station. Climatol 23(1):1–26 www.wunderground.com. Accessed 26 Jan 2016 29. Bouyer J, Inard C, Musy M (2011) Microclimatic coupling as a solution to improve building energy simulation in an urban context. Energy Buildings 43(7):1549–1559 30. Bourikas L et al (2013) Addressing the challenge of interpreting microclimatic weather data from urban sites. J Power Energy Eng 1:7–15 31. Shen T et al (2014) Impact of Urban Heat Island on Building Cooling Energy Consumption in Hangzhou. Abstracts of the 13th International Conference on Sustainable Energy Technologies, Geneva, Switzerland 32. Maxim Integrated (2013) iButton Temperature/Humidity logger with 8 kb data logger memory, Available via www.maximintegrated.com/products/ ibutton/data-logging/ 33. Shen T et al. (2013) Generating a modified weather data file for urban building design and sustainable urban planning accounting for the Urban Heat Island (UHI) effect. In: Abstracts of the 12th International Conference on Sustainable Energy Technologies, Hong Kong Polytechnic University, 26-29 August 2013. 34. Oke TR (2006) Initial guidance to obtain representative meteorological observations at urban sites. In: Instruments and observing methods Report No81, World Meteorological Organization. Available via www.wmo.int/ pages/prog/www/IMOP/publications/IOM-81/IOM-81-UrbanMetObs.pdf. Accessed 26 Jan 2016. 35. Cheng H, Castro IP (2002) Near wall flow over urban-like roughness. Bound-Layer Meteorol 104(2):229–259 36. Stewart ID, Oke TR (2012) Local Climate Zones for Urban Temperature Studies. Bull Am Meteorol Soc 93(12):1879–1900 37. Google Earth (2013) Satellite Images of Hangzhou and Ningbo 38. Trimble Navigation Limited (2013) SketchUp 3D CAD software, Available via Submit your manuscript to a www.sketchup.com/ journal and beneﬁ t from: 39. Xie Z, Castro IP (2006) Large-eddy simulation for urban micro-meteorology. J Hydrodynamics, Ser B 18(3, Supplement):259–264 7 Convenient online submission 40. Santiago J, Martilli A, Martín F (2007) CFD simulation of airflow over a 7 Rigorous peer review regular array of cubes. Part I: Three-dimensional simulation of the flow and validation with wind-tunnel measurements. Bound-Layer Meteorol 7 Immediate publication on acceptance 122(3):609–634 7 Open access: articles freely available online 41. Kanda M, Moriizumi T (2009) Momentum and Heat Transfer over Urban-like 7 High visibility within the ﬁ eld Surfaces. Bound-Layer Meteorol 131(3):385–401 7 Retaining the copyright to your article 42. Millward-Hopkins JT et al (2013) Aerodynamic Parameters of a UK City Derived from Morphological Data. Bound-Layer Meteorol 146(3):447–468 Submit your next manuscript at 7 springeropen.com

Journal

Future Cities and Environment – Springer Journals

Published: Jul 8, 2016

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Transforming typical hourly simulation weather data files to represent urban locations by using a 3D urban unit representation with micro-climate simulations

Transforming typical hourly simulation weather data files to represent urban locations by using a 3D urban unit representation with micro-climate simulations

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Transforming typical hourly simulation weather data files to represent urban locations by using a 3D urban unit representation with micro-climate simulations

Transforming typical hourly simulation weather data files to represent urban locations by using a 3D urban unit representation with micro-climate simulations

References (62)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies