Examined Urban Districts
The dynamic simulations of the urban open space and building climate were carried out in two different districts of Erfurt and Dresden (cf.Figure 1).
The city of Erfurt, Thuringia’s capital, is located in the middle of Germany. Erfurt’s urban climate is characterized by its location in a basin which is only open to the north direction. The city lies deeper in the Erfurt Basin (ca. 190m above sea level, ASL), which is almost completely surrounded by mountain ranges (up to 440m ASL). The small river Gera runs from the south-west through the city and is important, together with its green corridor, for the ventilation of Erfurt with fresh and cold air.
The investigated example quarter Oststadt (Germany, 50°58’43’’ N, 11°02’44’’ E) is part of the district Krämpfervorstadt in the east of Erfurt. Buildings with four to five floors and block structures from the 20th century mainly characterize the urban structure. Moderate up to strong thermal heat stress describes the local urban climate for the inhabitants in summer due to the densely built structure and a lack of urban green. The UHI effect is comparatively strong due to the high heat storage capacities and the limited air exchange (Steeneveld et al. 2011). The nocturnal heat load is quite high and leads to a comparatively high minimum air temperature (see Figure 6). In this way, the effects of climate change with an increasing duration and severity of heat episodes will be particularly noticeable in a city district like Erfurt's Oststadt.
Dresden is located in the eastern part of Germany and is assigned, similar to Erfurt, to the climatic zone of the humid moderate climate of the middle latitudes. Although the oceanic weather influence predominates, a stronger continentality can be recorded compared to the western parts of Germany, recognizable for example by the larger annual fluctuation of the air temperature. The location in the thermally favoured Elbe valley with its up to 200m high ridges is decisive for the climate of the urban region. Compared to other metropolitan areas in a basin (e.g., Erfurt), Dresden is relatively well ventilated, due to the strong orographic influence of the open valley on the wind field.
The example quarter in Gorbitz is located in the western part of the Saxon state capital Dresden, Germany (51°02’44’’ N, 13°40’28’’ E). The foundation stone for the largest and at the same time latest prefabricated estate of Dresden was laid in 1981. Mostly, 6-storey large-panel construction type of multi-residential buildings can be found in Gorbitz with numerous public and private green spaces. Gorbitz lies above the Elbe Valley on the outskirts. The local climate benefits from the comparatively high urban green volume as well as from the topography with ventilation corridors at the north-western and south-eastern edge of the quarter and cold air drainage flows due to the slope characteristics of this district. Gorbitz is characterized by favourable climatic conditions with lower heat load especially during night in comparison to Erfurt Oststadt. However, highly sealed areas and locations without shadowing are characterized by strong heat stress for the inhabitants in Gorbitz during daytime as well.
Examined Multi-Residential Buildings
The two districts differ in both urban structure and building types. While the district in Dresden mainly consists of large-panel construction (LPC) buildings, erected between 1980 and 1990, the district in Erfurt mainly consists of older multi-residential buildings of the so-called “Gründerzeithaus” (GZH) architecture from the turn of the 20th century. For investigations with BPS, we build a 3D simulation model of one representative GZH and LPC building (location see Figure 1) to investigate the effect of the different MMS input on the indoor thermal comfort. The comparison of the building views of the chosen GZH building in Figure 2 and LPC building in Figure 3 shows significant differences in their physical structure and architectural design. While the GZH building is characterised by a brick façade, a saddle roof and balconies, the LPC exhibits a pragmatic cubic structure with no decorative façade elements and a flat ventilated roof. The buildings also vary in their component structure. The walls of the GZH consist of bricks while the ceilings consist of wooden beam constructions. In contrast, walls and ceilings of the LPC consist of reinforced concrete. The GZH building was renovated in the year 2002 by a conversion of the saddle roof into two attic dwellings and associated insulation of the roof and the exterior walls in the attic. Other opaque building components in the full storeys remain unchanged, only the windows were replaced by a double-glazing and balcony plants were installed at the backyard oriented façade. In Table 3 of the appendix the building component structures of the LPC and GZH buildings are opposed, highlighting the enhanced thermal insulation standard of the LPC building.
Microscale Meteorological Simulation
In a first series of tests, the urban climate model ENVI-met was used to investigate the influence of the 3D urban structures on the microscale urban climate and to provide the meteorological input data for the BPS (Bruse et al. 1998, Liu et al. 2021). ENVI-met is a software to model the urban microclimate in three spatial dimensions with its diurnal variability (https://www.envi-met.com/ last access: 2022-02-17). A special characteristic is the high spatial resolution of the microscale model down to a few square meters. The effects and interactions between grey, green and blue structural elements with the airflow, as well as the temperature and humidity fields are numerically resolved by ENVI-met. The model has been applied and evaluated in recent studies (Ali-Toudert et al. 2006, Fahmy et al. 2011, Goldberg et al. 2013, Liu et al. 2021).
In the present study, we used ENVI-met V3.1 to simulate the course of a hot and an average summer day in the investigated urban districts Erfurt Oststadt and Dresden Gorbitz. This model version is not up-to-date, but we applied it in this study because of the consistency to former model applications in these areas. Another reason is the cost-free availability and unrestricted licence agreements of this version making it still attractive for usage by application partners in municipal offices.
In order to map the urban structures in the model geometry, satellite and aerial images were used (cf. Figure 1) as well as on-site inspections and further freely available data sources (‘Geoportal Thueringen’: https://www.geoportal-th.de/de-de/, 2021-07-30, ‘Themed city map of Dresden’: https://stadtplan.dresden.de/, 2021-07-30). The structural elements were manually digitized. This results in datasets with realistic spatial arrangement and geometrical properties of buildings, an urban vegetation layer with different properties (kinds of trees, shrubs, grassland), as well as sealed or partly sealed areas (asphalted surfaces, paved or gravel covered areas), and unsealed surfaces. Figure 4 shows the resulting area input files prepared for the ENVI-met simulations with a horizontal resolution of 4x4 m² within the core model domain.
The ENVI-met runs were set up for a rather hot, cloudless summer day (July 15) with very high solar irradiation and the development of an autochthonous weather situation. The model simulations were initialized at 20:00Local Time (due to nearly neutral stratification of the atmosphere). Here, the initial conditions fromTable 4(see Appendix) wereselected. The following 28 hours were simulated to generate a full daily cycle for all meteorological variables.
Especially the daily course of air temperature, solar irradiation (direct and diffuse radiation), wind speed and wind direction was evaluated at different locations (so-called receptor points) every five simulated minutes regarding the vertical profile(number of height levels: 29, central height of the level above ground surface: 1.00, 3.00, 5.12, 7.49, 10.15, 13.13,…, 362.07 m).The total model area was analysed every simulated hour.
The analysis of the ENVI-met simulations showed that the values for solar irradiation are very high for the selected conditions (geographic location and altitude of the cities; air turbidity). Furthermore, the daily amplitudes of the air temperature in ENVI-met are much lower as observed from selected weather situation nearby (see Figure 8). In order to reflect these conditions more realistically, we applied the atmospheric boundary layer model HIRVAC (HIgh Resolution Vegetation Atmosphere Coupler) for a further series of model simulations. HIRVAC is a numerical micro-meteorological model (Mix et al. 1994), which was applied as one-dimensional version to describe the exchange between soil, vegetation and atmosphere with a high vertical resolution. Reynolds-averaged equations of airflow, heat and moisture exchange are solved at about 100 vertical model layers until the top of the model at a height of about 2 km. Additionally to the atmospheric sub-model, a soil model and a vegetation model complete the HIRVAC model core (Ziemann 1998). About the half of the model levels are inside a typical tree layer. Additional terms in the basic equations enables the realistic simulation of the exchange processes between vegetation elements and the surrounding atmosphere. The distance between the model layers is a few centimetres (in comparison to a few metres in ENVI-met) in the near of the surface and increases with height. In this way, it is also possible to use suitable parameterizations for the transition from the laminar to the turbulent boundary layer. HIRVAC was evaluated using measurement values(Goldberg et al. 2001, Baums et al. 2005, Fischer et al. 2008).
Building Performance Simulation
To simulate the evolving room temperatures in both LPC and GZH buildings the BPS software IDA ICE 4.8[1] (EQUA 2018)and TRNSYS 18(Hiller et al. 2015)was applied. To consider the evolving vertical air temperature gradient within the building, the entire LPC and GZH was simulated in the 3D model although only the living room of the first and top floor are discussed in this article in detail (seeFigure 2andFigure 3). Structural building components were implemented in the BPS model aslisted inTable 3(Appendix) to achieve realistic heat storage and heat transmission characteristics for each room to the building exterior as well as to neighbouring rooms.
We listed the applied boundary conditions like shading situation, assumed window ventilation behaviour, internal heat gains, etc. in the Appendix. The validity of the BPS outcomes was tested by calibrating the simulated room temperature time series for several rooms to monitored indoor temperatures of the GZH(Schünemann et al. 2021)building for an entire summer period.
Model chain from microscale meteorological simulation to building performance simulation
To investigate the impact of variation in urban summer climate on indoor overheating in buildings we developed a model chain where the meteorological results of the MMS were used as meteorological input for indoor thermal comfort evaluation by BPS(Schünemann et al. 2020). For this purpose, ENVI-met models of two districts Erfurt and of Dresden were used. To record the meteorological output at the GZH and LPC locations within the district, we defined receptor points at the buildings location within the ENVI-met district model(Schünemann et al. 2020). The resulting meteorological data of outdoor air temperature, wind direction, wind speed, direct and indirect solar irradiation at these receptor points were taken from 10 m height above ground to reflect the outdoor conditions at the height of the focussed top floor flats in the BPS model of the GZH and LPC building. The thermal inertia of buildings requires that BPS are commonly are simulated for several weeks or at least for several days. Thus, we generated a 14 days long precondition phase followed by a 14 days ongoing heat wave by two different ENVI-met simulations: using an average summer day with 3/8 cloudiness and an initial air temperature of 22 °C for the precondition phase and for the heat wave a hot day with clear sky and 28 °C initial temperature at 20:00 Local Time. On one side, the daily cycle of air temperature (and other relevant meteorological quantities) of the summer day was simply duplicated to derive the precondition phase. On the other side, a period was simulated using long-term data (7 days) from ENVI-met simulations, which were extended to 14 days using a program-based method to recreate an amplifying heat wave with increasing maximum and minimum temperatures day by day and under the assumption of a nearly constant daily temperature amplitude. The resulting summer period of 28 days was used as meteorological input for the BPS, where shading of surrounding objects as well as wind and temperature gradient-driven air exchange were taken into account. For better clarity, the whole model chain process is visualised in the scheme inFigure 5(A: ENVI-met), starting with spatially resolved MMS using ENVI-met, the derivation of the 28 days long summer period as input for BPStoanalyse the indoor thermal conditions in the GZH and LPC rooms.
Besides the MMS tool ENVI-met, we applied the 1D coupled vegetation-boundary layer model HIRVAC in a second model chain (see “B: HIRVAC” inFigure 5) to evaluate the impact of different MMS models on indoor thermal comfort in BPS. The initial conditions for the HIRVAC runs were chosen similar to the ENVI-met simulations. In contrast to ENVI-met, in HIRVAC the entire 28 days long summer period (consisting of a 14 days long preconditioning phase (cloudy, average summer days) and a 14 days long heat wave (hot, clear days)) could be simulated directly by the model. HIRVAC was driven by the conditions listed inTable 5of the Appendix.
As meteorological input in BPS for the variant “B: HIRVAC”, we took the outdoor air temperature and solar irradiation from HIRVAC and the wind conditions from ENVI-met because local 3D wind effects by 3D spatial obstacles cannot be reproduced by the 1D model version of HIRVAC.
The third applied model chain variant “C: Mix ENVI-met & HIRVAC” in Figure 5 combines ENVI-met and HIRVAC simulations by using solar irradiation from HIRVAC and outdoor air temperature as well as wind conditions from ENVI-met.
[1] Validation history of IDA ICE: http://www.equaonline.com/iceuser/new_validationreports.html and http://www.equaonline.com/iceuser/new_certificates.html (accessed at 17.02.2022)