The influence of the urban heat island effect on the energy performance of residential buildings in a city with an oceanic climate during the summer period: Case of Valdivia, Chile

Sustainable Cities and Society 97 (2023) 104766 Available online 3 July 2023 2210-6707/© 2023 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). The influence of the urban heat island effect on the energy performance of residential buildings in a city with an oceanic climate during the summer period: Case of Valdivia, Chile Konstantin Verichev a , Cristian Salazar-Concha b , Carmen Díaz-L´ opez c , Manuel Carpio d , e , * a Institute of Civil Engineering, Universidad Austral de Chile, General Lagos 2086, Valdivia, Chile b Administration Institute, Universidad Austral de Chile, Independencia 631, Valdivia, Chile c Department of Building Construction I, University of Seville, Avenida Reina Mercedes 2, Seville, Spain d Department of Construction Engineering and Management, School of Engineering, Pontificia Universidad Cat´ olica de Chile, Avenida Vicu˜ na Mackenna 4860, Santiago, Chile e Centro Nacional de Excelencia para la Industria de la Madera (CENAMAD), Pontificia Universidad Cat´ olica de Chile, Santiago, Chile ARTICLE INFO Keywords: Urban heat island Cooling degree days Local climate zone Building Energy demand Energy consumption ABSTRACT The intensification of urban growth and climate change are affecting the energy and thermal performance of buildings. In recent years, the issue of building resilience to urban heat island (UHI) conditions has become increasingly relevant. The energy performance of buildings can vary significantly in different areas of the same city, regardless of their size. The aim of this study was to evaluate the energy effects of UHI intensity differences in various local climate zones (LCZs) of Valdivia city, Chile, on a typical residential dwelling. A simplified methodology was used, based on the assessment of cooling degree days variations and heat gains variations inside the studied house through the thermal envelope. Valdivia has a homogeneous urban morphology, and three types of low-rise LCZs prevail in the city (LCZ 3, 6, and 9). The results showed that the average cooling demand for a 66 m 2 residential dwelling during 39 summer days was 158 kWh for LCZ 9, 219 kWh for LCZ 6, and 289 kWh for LCZ 3, compared to the rural demand of 114 kWh. These results suggest that the energy effects of UHI can be significant, and that it is important to study the microclimatic conditions in various LCZs for a correct understanding of the UHI energy effects on different buildings. 1. Introduction The process of urbanisation is directly related to the change and transformation of natural ecosystems. The increase in construction and the development of urban infrastructure lead to a constant growth in the use of different materials that change the balance of radiation, humidity and thermal conditions in the urban ecosystem. The continuous development and growth of the urban environment leads to increased air pollution and greenhouse gas emissions. All these factors cause a difference in air temperature between urban and rural areas. This phenomenon is known as urban heat island (UHI) (Oke, 1973, 1987). The UHI phenomenon has been defined and described at various times and in various major cities around the world, both in the northern hemisphere – Madrid (Yagüe et al., 1991), Seoul (Kim & Baik, 2002, 2005), New York (Gaffin et al., 2008) and London (Kolokotroni & Giridharan, 2008) – and in the southern hemisphere – Santiago de Chile (Sarricolea & Martín-Vide, 2014; Smith & Romero, 2016) Buenos Aires (Bejar´ an & Camilloni, 2003; Camilloni & Barrucand, 2012) and Melbourne (Morris & Simmonds, 2000). Generally, scientific papers present UHI research in large cities, but research in small cities and semi-urban areas has obtained valuable results (Fujibe, 2009). For example, research in small northern hemisphere cities such as Fairbanks, Alaska, in the USA (33,000 inhabitants), showed an absolute maximum between urban and rural area temperature – maximum urban heat island intensity (UHII), in the same measurement period, of about 4.5 ◦C (Magee et al., 1999). Likewise, in the city of Apatity, Kola Peninsula of Russia (58,000 inhabitants) a UHII of 1.9 ◦C was obtained from sensor measurements (Konstantinov et al., 2015). In cities with temperate climates, such as the city of Bilbao in Spain (350,000 inhabitants), the average UHII value for the summer period * Corresponding author at: Department of Construction Engineering and Management, School of Engineering, Pontificia Universidad Cat´ olica de Chile, Avenida Vicu˜ na Mackenna 4860, Santiago, Chile. E-mail address: [email protected] (M. Carpio). Contents lists available at ScienceDirect Sustainable Cities and Society journal homepage: www.elsevier.com/locate/scs https://doi.org/10.1016/j.scs.2023.104766 Received 5 December 2022; Received in revised form 29 June 2023; Accepted 29 June 2023 Sustainable Cities and Society 97 (2023) 104766 2 was 2.54. (±1.00)◦C (Acero et al., 2013). For the city of Christchurch, New Zealand (360,000 inhabitants) located in the temperate climate zone of the southern hemisphere, it was observed for a winter period UHII of up to 5.0 ◦C on some days (Tapper et al., 1981). UHII depends on different factors: density of urbanisation, urban green or blue areas, types of urban paving and also strongly dependent of the weather conditions, all of which affect urban thermal comfort (Carpio et al., 2020). The density of urbanization and paving varies in different parts of cities. This is why the urban classification system developed by Stewart and Oke (2012) is widely used for UHI research. This system defines 17 urban local climate zones (LCZs) (Stewart & Oke, 2012). The effect of UHI on the energy efficiency of buildings can be quantified through heating degree days (HDD) and cooling degree days (CDD) (Verichev et al., 2021b). HDD is the annual or period sum of average daily or hourly temperatures below a specific base temperature, which may vary depending on the type of HDD application task or national building codes. CDD is the annual or period sum of average daily or hourly temperatures above a given base temperature (Christenson et al., 2006; Matzarakis & Balafoutis, 2004; Park et al., 2021). Santamouris analysed the effect of UHI on the energy demand of buildings in different cities over 4 decades (Santamouris, 2014). The results showed that the demand for heating decreases in urban centres by an average of 19%, while the demand for cooling in urban centres increases by an average of 23%. Moreover, these percentages will grow in parallel with future changes in the global climate (IPCC, 2021). Several studies have analysed the effects of UHI on CDD during the summer period. For example, a study on UHI in Lisbon found that CDD with base temperature 23 ◦C (CDD23) for the period from May to September varied from 351 ◦C to 470 ◦C for four virtual stations of MM5 (fifth-generation mesoscale model) (Oxizidis et al., 2008). In London, measurements at 77 locations between August 1999 and 2000 showed a variation of CDD12 from 156 ◦C to 217 ◦C and 165 ◦C to 233 ◦C, respectively (Kolokotroni et al., 2009). A variation of CDD12 from 292 ◦C to 408 ◦C and CDD12 from 308 ◦C to 430 ◦C was reported at 77 stations in London for July and August 2008, respectively (Kolokotroni et al., 2010). At six monitoring stations in Bahrain, the annual variations of CDD18 and CDD24 were in the ranges of 3339 ◦C to 3911 ◦C and 1746 ◦C to 2089 ◦C, respectively (Radhi & Sharples, 2013). In the city of Agrinio, Greece, with 93,000 inhabitants, variations of CDD27 between June and August were 251 ◦C rural and 304 ◦C urban (Vardoulakis et al., 2013). In Rome, between 2015 and 2017 annual CDD20 in five localities presented a variation from 101 ◦C to 260 ◦C (Zinzi et al., 2018). In Seoul, during 2020, at 57 stations between June and September, CDD26 urban average was 85 ◦C and rural average was 30 ◦C (Su et al., 2021). These increases in temperature and CDD have a substantial impact on the behaviour and performance of buildings throughout their life cycle if they are not taken into account at the design stage (Díaz-L´ opez et al., 2021). If buildings are designed without considering this extra thermal load caused by UHI, over a short period of time buildings will not be able to provide the adequate thermal comfort for which they were designed, incurring extra costs in terms of energy consumption, and causing deterioration of their structural components (Díaz-L´ opez et al., 2021). Many researches have shown the increase in cooling demand between the urban centre and the periphery. In Athens in 1996, measurements at 30 automatic stations throughout the city showed that the cooling demand for the month of August varied from 5 kWh/m 2 to 11.5 kWh/m 2 between the city centre and the urban periphery (Santamouris et al., 2001). In London in 2006, a 19% increase in cooling energy was observed for a typical 60 m 2 office in the city centre compared to the periphery (Kolokotroni et al., 2006). In Hong Kong in 2011, a 10% increase in air-conditioning demand was observed for a 95 m 2 residential flat in the city centre compared to the periphery (Chan, 2011). In Tokyo in 2012, a 27.5% increase in cooling energy was identified in nine standard building (Hirano & Fujita, 2012). In Melbourne in 2014, an 11.4% increase in cooling energy consumption was observed for a 1,350 m 2 residential building (Ren et al., 2012). Although the importance of the UHI effect on the energy demand of buildings is widely recognised by the scientific community, the quantitative impacts have not been sufficiently studied (Li et al., 2019). Furthermore, in recent research on the effect of UHI on the energy demand of buildings, most studies focus on large cities with more than one million inhabitants, ignoring the issues that UHI brings to small and medium sized cities that are expanding. Also in recent years, there has been a growing interest in analysing the influence of LCZs on the energy performance of buildings (Yang et al., 2020). For example, a study in Fig. 1. Map of the geographical location of the city of Valdivia (left) and map of the location of the microclimatic sensors (right). K. Verichev et al. Sustainable Cities and Society 97 (2023) 104766 3 India, city of Nagpur, the energy consumption of residential and office buildings in three different LCZs (LCZ 3, LCZ 6 and LCZ 9) was analysed in a tropical savanna climate (Kotharkar et al., 2022). The urban structure of cities can vary from country to country. For example, in China, the urban structure is characterized by a predominance of LCZ 1, LCZ 2, LCZ 4, LCZ 5 in cities such as Wuhan and Shanghai (Yang et al., 2023; Zhang et al., 2022). L´ opez-Guerrero et al. note that only 7% of studies on the thermal and energy effects of UHIs on buildings were conducted in cities with a population of less than 300,000 inhabitants, and most studies (65.2%) were conducted in cities located only in three K¨ oppen climatic zones (L´ opez-Guerrero et al., 2022). Consequently, the thermo-energy interaction of buildings with the urban environment depend on a number of factors, including general climatic conditions, urban structure, and the size of the city amongst other parameters. In recent years, several methodologies have been developed in studying the impact of UHI on the energy efficiency of houses. These methodologies aim to link LCZs with building energy simulation, modify meteorological archives for energy simulation of houses for each zone, and conduct detailed energy simulations. However, these methodologies are complex and require validation in different climatic and urban conditions (Mao et al., 2017; Salvati et al., 2019). Thus, methods of microclimatic monitoring and simplified energy assessment can be useful for understanding the energy effects of UHI on buildings (Kotharkar et al., 2022). Therefore, the aim of this paper is to analyse the UHI phenomenon through a temperature measurement experiment in a small city, in order to obtain quantitative estimations of UHI effects at multiple locations in the city with respect to LCZs, CDD and variation of energy consumption of a typical dwelling over a period of time in summer. The city selected as a case study will be Valdivia, in southern Chile, with 155,000 inhabitants and an oceanic climate. 2. Material and methods This section will begin by providing a general geographical description of the city under study. It will then be followed by a detailed explanation of the microclimate experiment and the methodology employed to determine the LCZs and evaluate the variations in energy consumption for a typical residential building. 2.1. Geographical description of the location of the city of Valdivia The city of Valdivia is located in southern Chile (39◦48 ′ S and 73◦14 ′ W), at the confluence of the Calle-Calle, Valdivia and Cruces rivers. It is situated 15 km from the Pacific Ocean coast (Fig. 1). Valdivia is surrounded by hills with altitudes not exceeding 300 m above sea level. The relief altitudes within the city do not exceed 30 m. The city is primarily characterized by low-rise residential buildings, with no significant industrial presence. In addition, Valdivia has a fairly high proportion of green areas per inhabitant (Rojas et al., 2016). The city’s climate is marine west coast climate (warm and dry summer) with oceanic influence (Cfb), based on modified K¨ oppen–Geiger classification (Sarricolea et al., 2016). It has an average annual rainfall of 1,800 mm with maximum rainfall in May, June and July (Castillo, 2001). The average daily temperature in the summer months in the city does not exceed 21 ◦С, while in winter the average daily temperature does not fall below 5 ◦С. This type of climate and rainfall regime form an ecoregion characterised by evergreen forests, also called Valdivian Forest. The city has areas with wet soils and wetland areas. The daily value of the global horizontal solar radiation in the summer months reaches 7.44 kWh/m 2 /day, and in the winter months it is 1.05 kWh/m 2 /day (Chile, 2019). 2.2. Microclimatic research methods To organise the microclimatic measurement experiment and identify the thermal differences of the air at a level of 2 m, 10 iButton "Maxim Integrated" (Maxim Integrated, 2020) DS1923 temperature recording sensors with a measurement accuracy of 0.5 ◦C were used. These sensors are self-contained and allow data to be recorded and stored over long periods of time. They are protected from sunlight and rain by a solar radiation shield. Sensors were installed uniformly throughout the city (Fig. 1). All sensors were placed at a standardised height of 2 m on a natural underlying surface (grass) and at a considerable distance from buildings, walls and asphalt (more than 3 m). Temperature data collection through the installed sensors was carried out for 39 days starting on 24 January 2017, with a resolution of hourly measurements. The geographical coordinates of the measurement points are presented in Table 1. Each location will be characterised according to its LCZ type. To determine the main synoptic processes in the city of Valdivia during the period of the microclimatic measurements, daily synoptic maps of atmospheric pressure with analysis of frontal systems, provided by the Meteorological Service of the Chilean Navy (Armada de Chile, 2022) and trajectories of air masses (HYSPLIT, 2020), were analysed. To compare results of measurements in the city with rural meteorological data, the record of the Pichoy meteorological station was used as a "rural" reference station. The Pichoy station belongs to the Meteorological Office of Chile and is located 35 km north-east of the city centre (altitude 20 m above sea level) (Chile, 2019). Table 1 Geographical coordinates of microclimatic experiment locations. Point Lat/Lon Alt., m Point Lat/Lon Alt., m 1 39◦47´40.26´´/ 73◦13´23.49´´ 14 6 39◦49´42.39´´/ 73◦14´30.25´´ 17 2 39◦47´58.98´´/ 73◦12´27.10´´ 7 7 39◦50´09.39´´/ 73◦12´48.20´´ 16 3 39◦48´52.82´´/ 73◦15´31.62´´ 15 8 39◦50´28.23´´/ 73◦16´28.05´´ 17 4 39◦49´07.25´´/ 73◦14´44.15´´ 10 9 39◦50´44.37´´/ 73◦15´06.48´´ 17 5 39◦49´05.84´´/ 73◦13´54.15´´ 10 10 39◦50´45.29´´/ 73◦13´´58.04´ 8 Table 2 Values of surface cover properties for LCZs classification. LCZ Name Building surface fraction Impervious surface fraction Pervious surface fraction Height of roughness elements, m LCZ 1 Compact high-rise 40–60 40–60 <10 >25 LCZ 2 Compact midrise 40–70 30–50 <20 10–25 LCZ 3 Compact low-rise 40–70 20–50 <30 3–10 LCZ 4 Open high-rise 20–40 30–40 30–40 >25 LCZ 5 Open midrise 20–40 30–50 20–40 10–25 LCZ 6 Open low-rise 20–40 20–50 30–60 3–10 LCZ 7 Lightweight low-rise 60–90 <20 <30 2–4 LCZ 8 Large low-rise 30–50 40–50 <20 3–10 LCZ 9 Sparsely built 10–20 <20 60–80 3–10 LCZ 10 Heavy industry 20–30 20–40 40–50 5–15 K. Verichev et al. Sustainable Cities and Society 97 (2023) 104766 4 Fig. 2. Example of building surface fraction definition for measurement point 8. Fig. 3. Plan and 3D model of the dwelling under study. K. Verichev et al. Sustainable Cities and Society 97 (2023) 104766 5 To determine the LCZ in which each measurement point is located, the classification of Stewart and Oke (2012) was used (Stewart & Oke, 2012). Table 2 presents the main surface cover parameters characterizing various LCZs according to Stewart and Oke (2012). To determine LCZ, the following information is needed: (i) building surface fraction; (ii) impervious surface fraction; (iii) pervious surface fraction; and (vi) height of roughness elements. The first two parameters were calculated using Google Earth Pro. An area of 40,000 m 2 was defined around each measurement point, and the corresponding surfaces were calculated. Surfaces corresponding to water bodies and surfaces covered with vegetation were also identified, as these are constituent components of pervious surfaces. Satellite images from the period corresponding to the microclimatic experiment (January 2017) were used to identify these surfaces. Fig. 2 shows an example of the definition of the building surface fraction for location 8 of the microclimatic experiment. The spatial resolution of 200 m for the area around the locations of microclimatic measurements was determined within the spatial resolution range defined by Oke (2002) for microclimatic studies (Oke, 2002). 2.3. Estimated energy consumption for cooling of the typical dwelling To quantify the effect of UHI, we used the CDD parameter, which is widely used in the field of building energy (Mourshed, 2012). The “hourly” method is used to calculate the CDD (Verichev & Carpio, 2018): CDD =[∑ n i=1 (Ti−Tb)+]1 24 (1) where, T i is the hourly temperature measured at the ith hour of the day; index (+) means that it is only necessary to calculate positive differences between the base temperature (T b ) and T i . In this study, the standard base temperature of 20 ◦C and the recommended temperature of 25.3 ◦C from the Sustainable Building Standards for dwellings in Chile were used (Government of Chile, 2019). In accordance with Eq. (1), the values of CDD20 and CDD25.3 for the period of the experiment were calculated for each point of microclimatic measurements. To demonstrate the differences in energy demand for cooling in the microclimatological study period of a representative dwelling, a typical dwelling from southern Chile was used (Fig. 3). The dwelling has a floor area of 66.37 m 2 and was used in previous studies (Verichev et al., 2020, 2021a). Table 3 shows the thermal transmittance value of the elements of the thermal envelope of this dwelling and the exchange surface with the outside environment. Considering one type of representative building located in different LCZs of the city is a common practice for assessing energy differences (Assimakopoulos et al., 2007; Zinzi & Carnielo, 2017; Zinzi et al., 2018). This technique allows researchers to move from assessing environmental thermal differences within a city to real building energy differences. However, for more accurate implementation of studies based on energy simulation, it is necessary to have a broad spectrum of meteorological data, which can be difficult and costly to obtain in practice. In this paper, we will conduct an energy assessment based on a methodology for assessing the energy consumption of buildings for cooling, which is widely used in life cycle cost assessment optimization of building thermal insulation (Kurekci, 2016; Las-Heras-Casas et al., 2021; L´ opez-Ochoa et al., 2020; Nematchoua et al., 2015; Sisman et al., 2007; Yuan et al., 2017). Present study will analyse the energy consumption for cooling of the same dwelling but located at different microclimatic measurement points. The energy demand of a house for cooling generally depends on the following factors: (i) heat gains inside the house through the thermal envelope; (ii) solar gains; (iii) gains from human occupancy; (iii) gains from electronic equipment; and (iv) ventilation and infiltration rates amongst others. To assess the differences in energy demand and consumption of a typical house within a studied city, it can be assumed that these differences will depend only on the heat gains inside the house through the thermal envelope. Therefore, it is possible to estimate the energy demand for cooling according to the methodology of the studies – (Kurekci, 2016; Las-Heras-Casas et al., 2021; L´ opez-Ochoa et al., 2020; Nematchoua et al., 2015; Sisman et al., 2007; Yuan et al., 2017). With this methodology, differences in demand related to disparity in thermal conditions of various measurement points in the city can be demonstrated. The cooling energy demand for the building envelope per unit of exchange area per study period, in kWh/m 2 , is calculated using the following equation: EDcool,e=0.024⋅CDD20⋅Ue(2) where, e refers to the roof, walls, floor and openings that form the thermal envelope of the building; and U e is the thermal transmittance of element e of the building envelope, in W/m 2 ⋅K. The cooling energy demand of the building per unit of living space per period, ED cool , respectively, in kWh/m 2 per study period, is calculated by the following equation: EDcool =∑EDcool,e⋅Aexch,e Aliv (3) where, A exch,e is the exchange surface of the building envelope element e , in m 2 , and, A liv , which is the living area of the building. It will also take on the assumption that the functioning of the analysed house will be 24 h a day. To quantify the urban differences in energy consumption of a typical dwelling for cooling, the formula to be used is: ECcool =EDcool ε (4) where, ε is the thermal efficiency or seasonal energy efficiency ratio of the cooling system. For this estimation, the Chilean market split system Table 3 Thermal transmittance of the elements of the thermal envelope of the dwelling under study and exchange surface. Structural elements U, W/m 2 ⋅K Exchange surface, m 2 Roof 0.329 57.28 Walls 0.629 100.81 Floor 0.266 45.94 Doors 1.060 3.28 Windows 3.160 9.86 Table 4 LCZ definition results for the microclimatological experiment points. Location 1 2 3 4 5 6 7 8 9 10 LCZ 9 3 3 3 6 6 6 9 3 6 Building surface fraction,% 10.0 51.5 48.8 63.2 29.0 36.0 35.3 16.5 49.5 37.8 Impervious surface fraction,% 2.8 25.5 21.4 17.2 32.6 20.1 20.8 0 20.4 26.4 Vegetation surface fraction,% 78 7.6 11.1 6.6 16.9 32.7 16.8 49 16.6 31.8 Water surface fraction,% 2.4 0 0 0 3.7 0 0 0 0 0 Height of roughness elements, m 3–10 3–10 3–10 3–10 3–10 3–10 3–10 3–10 3–10 3–10 K. Verichev et al. Sustainable Cities and Society 97 (2023) 104766 6 Fig. 4. Examples of Valdivia LCZs. K. Verichev et al. Sustainable Cities and Society 97 (2023) 104766 7 will be used. – “Samsung© Split Inverter Muro 9000 BTU/hr” with ε =3.21 (https://www.samsung.com/cl/air-conditioners/wall-mo unt/ar09tsfzawk-zs/). The price of electricity for the study period was 0.23 USD/kWh (Chilean Central Bank, 2022). In this study, we will only consider the energy consumption required for cooling the dwelling. This choice is based on the fact that the experiment was conducted during the summer, and Chileans living in the southern part of the country generally exhibit a relatively high tolerance for lower temperatures (P´ erez-Fargallo et al., 2018). 3. Results and discussion This section begins with a presentation of the results of determining the LCZs at the locations of the microclimatic experiment. The meteorological-synoptic analysis results for the microclimatic experiment period are then discussed. Finally, an analysis of the CDD under UHI conditions is conducted, along with a presentation of estimated energy demand and energy consumption data for cooling the dwelling under study. 3.1. Description of LCZs of the studied urban locations Table 4 presents the results of the analysis of satellite images obtained from Google Earth Pro. It was identified that the city of Valdivia does not present major variability in LCZs, only three types of LCZs were identified – 3, 6 and 9: (i) four locations of the microclimatic experiment were located in LCZ 3 (Table 4). The average fraction of building surfaces in these locations was 53.3%, while the average fraction of vegetation surfaces was 10.5%. For example, in Fig. 4, it can be seen location of the measurement point 4 with 63.2% of the building surface fraction; (ii) four locations of the microclimatic experiment were situated in LCZ 6 (Table 4). The average fraction of building surfaces in these locations was 34.5%, while the average fraction of vegetation surfaces was 24.6%. As an illustration, measurement point 5 is positioned within LCZ 6 (Fig. 4) with a building fraction of 29% and an impervious surface fraction of 32.6%, the highest amongst all the locations in the city; (iii) two locations of this microclimatic experiment were situated in LCZ 9 (Table 4). These locations where characterized by a high proportion of vegetated surfaces. For example, measurement point 8 was positioned within LCZ 9 (Fig. 4) with a building fraction of 16.5% and no impervious surfaces. Additionally, it is worth noting that in Table 4, the sum of fractions of identified surfaces for each location is less than 100%. This is because the remaining fraction accounts for pervious surfaces, such as bare soil without vegetation or roads and sidewalks without impervious pavement. It is worth mentioning that all the measurement points in this study were situated in LCZs characterized by buildings with an average height ranging from 3 to 10 m. It is also worth adding that locations 1–3 and 8–10 were located in purely residential urban areas with different building densities. Locations 4–6 were located in mixed residential- Fig. 5. Main meteorological parameters of the rural meteorological station in the period of the microclimatic experiment. K. Verichev et al. Sustainable Cities and Society 97 (2023) 104766 8 downtown urban areas. And location 7 was located in a mixed residential-industrial urban area. 3.2. Meteorological conditions during the period of the microclimatic experiment To analyse the microclimatic results, it is necessary to identify the general synoptic processes that determine the atmospheric conditions during the experiment. Figs. 5 and 6 show the measurement data of the main atmospheric parameters (pressure, temperature and relative humidity) recorded at the rural meteorological station. Using all the synoptic information and data from this station, the synoptic processes and situations during the microclimatic experiment were analysed. Two typical and repetitive synoptic situations were identified during the experiment that determined the weather in the city: •Anticyclonic period: Generally, when the city was located in the south-eastern part of the South Pacific Subtropical Anticyclone in a weak barometric gradient field, surface and atmospheric heating was caused during the day by direct solar radiation and intense cooling of the underlying surface at night in clear conditions. For this reason, on some days, a daily temperature amplitude between 8.9 ◦C at night and up to 25 ◦C at midday could be observed (Fig. 5). •Cyclonic period: When the city was affected by cyclones and frontal systems, cyclonic periods were characterised by daily minimum temperature amplitudes and low atmospheric pressure (Fig. 5). The passage of frontal systems was associated with atmospheric precipitation, for example, on the 26th and 27th days of the experiment, precipitation with an intensity of 45 mm was observed, which also affected the increase in the average daily values of relative humidity on these and subsequent days (Fig. 6). Generally, in anticyclonic conditions, air masses came from the tropical part of the Pacific Ocean, and in the case of cyclonic periods, air masses came from the temperate and Antarctic part of the Pacific Ocean. Oceanic air masses were always observed. This can be seen in Fig. 7, which shows the daily trajectories (−72 h) of the air masses that reached the city of Valdivia at a level of 500 m (lower tropospheric layer) during the study period. It also shows that during the period of the microclimatic experiment, no interventions of continental air masses and periods with heat waves were observed. During the study period, only 11 of the 39 days of the experiment, temperatures exceeded 25 ◦C, a typical summer temperature for this region of the world. 3.3. Analysis of temperature measurements and CDD values during microclimatic experiment After the analysis of the general synoptic situations and LCZs of the studied urban locations, an analysis of the microclimatic measurement data recorded at each of the 10 measurement points in the city was carried out. Fig. 8 shows the daily average values of the temperature at all urban measurement points and at the rural meteorological station. The temporal evolution of the results at the 10 points and at the rural station are synchronous. In general, the temperatures observed at the rural meteorological station are lower than in urban locations. Table 5 presents the results of the daily average urban heat island intensity (UHII) or urban-rural temperature difference during the microclimatic experiment at the 10 urban measurement points. It can be seen that points 8 and 1 have a minimal difference with the data from the rural station of +0.08 ◦C and +0.32 ◦C, respectively, and measurement points 9 and 4 have a larger difference to the rural station at +2.25 ◦C y +2.05 ◦C, respectively. The average value of the UHII in the city of Valdivia is similar to, for example, the city of Bilbao in Spain, which has Fig. 6. Results of relative humidity measurements of the rural meteorological station in the period of the microclimatic experiment. K. Verichev et al. Sustainable Cities and Society 97 (2023) 104766 9 an average daily temperature difference of +0.70 ◦C for the whole summer period (Acero et al., 2013). In general, the average daily temperatures at the urban measurement points exceeded the values of the rural metrological station, although at the urban points located in LCZ 9 colder values of the average daily temperatures were observed (Fig. 9). If the results of the CDD20 calculation at different measurement points in the city during the microclimate experiment are analysed (Table 5), it is noticeable that between points 8 and 9 there is almost double the difference in CDD20, being the two most critical UHI points in the city. It can be concluded that between point 8 and point 9 the same dwelling with the same air-cooling system will consume almost twice as much energy for air cooling during the period of the experiment. This leads to a problem of optimisation in the houses, according to the actual urban microclimatic conditions in terms of thermal insulation and the search for an optimal cooling system. In terms of the energy rating of dwellings, if a company builds typical modular or prefabricated houses with a pre-established energy assessment, they should adjust their design and air-conditioning systems according to the urban microclimatic realities. Another issue is outdoor thermal comfort in different urban areas, which was not analysed in this research, but is currently attracting scientific interest related to UHI and climate change. Furthermore, there is a strong correlation between the UHII and the values of CDD20 and CDD25.3 during the microclimatic experiment. This is shown in Fig. 10, which has a correlation coefficient of R 2 =0.97 for CDD20 and R 2 =0.95 for CDD25.3. In the case of the rural station for the period of the experiment the value of CDD20 was 36.8 ◦C and CDD25.3 of 5.6 ◦C, being lower than in any urban area. If approximation formulas are used to calculate CDD20 and CDD25.3 when the UHII is equal to 0, it is possible to obtain values of 45.6 ◦C and 7.8 ◦C, respectively, which are quite close to the real values recorded at the rural meteorological station (Table 5). The observed differences in CDD between the different measurement points in the same city depend on building surface and impervious surface fractions. Analysis of the percentages obtained for each type of urban cover (Table 4), as well as the CDD values obtained for the period of the microclimatic experiment (Table 5), shows a correlation between CDD20 and the sum of building and impervious surface fractions (Fig. 11). These two types of surfaces were summed because they are more associated with direct anthropogenic interference on the environment. It is worth noting that the minimum CDD corresponds to the minimum sum of building and impervious surface fractions, and the maximum CDD corresponds to the maximum sum. This correlation is quite logical for the defined LCZs with low building height, but high building fraction (e.g., point 4 has 63.2%), because in the summer period in Valdivia there is no significant effect of building shading. With increasing building density and impervious paving, the CDD20 value Fig. 7. Daily trajectories (00.00 UTC) of air masses during the microclimatological experiment period for the city of Valdivia. Based on data from the HYSPLIT (2020). K. Verichev et al.