Influence of the Representative Concentration Pathways (RCP) scenarios on the bioclimatic design strategies of the built environment

Sustainable Cities and Society 72 (2021) 103042 Available online 25 May 2021 2210-6707/© 2021 The Author(s). 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/). Influence of the Representative Concentration Pathways (RCP) scenarios on the bioclimatic design strategies of the built environment David Bienvenido-Huertas a , Carlos Rubio-Bellido a , *, David Marín-García b , Jacinto Canivell a a Department of Building Construction II, University of Seville, 41012, Seville, Spain b Department of Graphical Expression and Building Engineering, University of Seville, 41012, Seville, Spain ARTICLE INFO Keywords: Bioclimatic design Thermal comfort Climate change Built environment ABSTRACT Over time, whether through traditional knowledge or the constructive implementation, the relationship of the built environment with the climate conditions of a certain place has been developed. The control of these symbiotic solutions based on the climate-conscious design and their strategic approach have been improved to keep better welfare levels. Due to climate change, however, design strategies could be modified in a context of global warming. This research considers the Representative Concentration Pathways (RCP 2.6, 4.5 and RCP 8.5) to analyse the effectiveness of the design strategies throughout the 21 st century. A total of 6 countries (France, Portugal, Spain, Argentina, Brazil, and Chile) were selected to assess both thermal comfort levels and the need for using HVAC systems in each climate zone and in all future scenarios, so 1,450 cases were studied. The results showed that the less conservative climate change scenarios will affect thermal comfort, thus significantly reducing comfort hours in warm climates. In addition, passive design strategies could be less effective in the future, predominating the use of cooling systems. As a result of this research, future design strategies should be dynamic and permeable for possible scenarios. 1. Introduction The oil crisis of the 1970s made people aware of climate change effects. Over the years new generations have considered climate change as a major challenge for society, particularly in the 21 st century. Earth’s ecosystem is strongly changing: the extinction of species, desertification, sea level rise, and more and more extreme thermal conditions (World Wildlife Fund, 2014). Moreover, the Covid-19 pandemic has shown the effects of future pandemics emerging from the new climate change scenario (Manzanedo & Manning, 2020). The high energy consumption from anthropogenic activities is among the main generators of climate change. These activities include the building sector; in quantified data, it is responsible for both 40 % of energy consumption (European Commission, 2006; European Environment Agency, 2018) and 36 % of greenhouse gas emissions (European Commission, 2002; European Union, 2010). This high building energy consumption is due to deficient energy performance because HVAC systems are used for long periods to keep thermal comfort conditions inside dwellings. Consequently, other social problems are emerging, such as fuel poverty (S´ anchez-Guevara S´ anchez, Mavrogianni, & Neila Gonz´ alez, 2017), so having solutions for high energy consumption could improve some aspects in various scopes, apart from the environmental one. To lessen climate change, international goals should be established as the activities of each region greatly affect all over the world. In this regard, 195 countries were committed in the 2015 Paris Conference of the Parties (COP) to reduce greenhouse gas emissions (Tobin, Schmidt, Tosun, & Burns, 2018). Likewise, decarbonisation goals have been established by the year 2050 through various programs or policies, both at a continental (European Commission, 2011) and state scale (Climate Change Act, 2008, 2008; Ministry of the Environment (Japan) (2017)). Most of these policies establish a demanding goal: the reduction in building energy consumption by 90 % in comparison with the actual values. Thus, designing efficient buildings is crucial. For this purpose, climate characteristics must be known, thus guaranteeing a sustainable built environment (Kim, Gu, & Kim, 2018). In addition, architects and engineers should know the microclimate conditions before and after constructing the building to ensure its sustainability (Stavrakakis et al., 2012). In short, both climate behaviour (Tejero-Gonz´ alez, Andr´ es-Chicote, García-Ib´ a˜ nez, Velasco-G´ omez, & Rey-Martínez, 2016) and the vernacular architecture of the region (Manzano-Agugliaro, Montoya, * Corresponding author. E-mail address: [email protected] (C. Rubio-Bellido). Contents lists available at ScienceDirect Sustainable Cities and Society journal homepage: www.elsevier.com/locate/scs https://doi.org/10.1016/j.scs.2021.103042 Received 3 February 2021; Received in revised form 5 May 2021; Accepted 21 May 2021 Sustainable Cities and Society 72 (2021) 103042 2 Sabio-Ortega, & García-Cruz, 2015) should be known in detail to establish adapted design strategies. These design strategies adapted to climate imply many positive aspects: guaranteeing a great percentage of thermal comfort hours (Gaitani, Mihalakakou, & Santamouris, 2007) and reducing energy consumption (Casquero-Modrego & Go˜ ni-Modrego, 2019), greenhouse gas emissions (Omer, 2008) and economic costs (Li, Yu, Liu, & Li, 2011; Taleb, 2014). Moreover, passive design strategies do not just affect buildings because they could be extrapolated to the urban scope. The most appropriate passive strategies in urban environments located in arid (Daemei, Azmoodeh, Zamani, & Khotbehsara, 2018; Hamdan & de Oliveira, 2019) or humid zones (Lu et al., 2017) have been studied. The use of design strategies adapted to climate have been widely analysed: (i) Cicelsky and Meir (2014) determined that, in hot hyper-arid, the combination of extensive insulation, shade, high performance windows, air tightness and seasonal operation of window shutters ensures appropriate strategies for this type of climate; (ii) Khotbehsara, Daemei, and Malekjahan (2019) studied the energy performance of buildings located in four climates in Iran through the type of roof and determined the most appropriate passive design strategies; (iii) Asadi, Fakhari, and Sendi (2016) determined that bioclimatic techniques in Iran could avoid using air conditioning systems in summer; (iv) in another study conducted in the same region, Riahi Zaniani, Taghipour Ghahfarokhi, Jahangiri, and Alidadi Shamsabadi (2019) verified the great potential of using green roofs as passive strategy to reduce cooling demand; and (v) Yang, Fu, He, He, and Liu (2020) analysed the most appropriate design strategies in buildings located in Turpan (China). The most effective measures were passive solar heating in winter and the use of semi-basements and night ventilation in summer. However, the effectiveness of these strategies in relation to climate change has been scarcely studied. Osman and Sevinc (2019) reported that resilient strategies should be focused on a more active cooling by 2070 in arid regions. However, climate change effects in other regions are unknown. Likewise, the studies analysing the influence of climate change scenarios have not considered the most recent climate change scenarios. Osman and Sevinc (2019) used the Special Report on Emissions Scenarios (SRES) of the Intergovernmental Panel on Climate Change (IPCC) (Intergovernmental Panel on Climate Change, 2007). Nevertheless, the SRES were updated through Representative Concentration Pathways (RCP) scenarios (Intergovernmental Panel on Climate Change, 2014). These RCP scenarios establish four tendencies of the evolution of greenhouse gas emissions throughout the 21 st century: a strict mitigation scenario (RCP2.6), two intermediate scenarios (RCP4.5 and RCP6.0), and a scenario with very high greenhouse gas emissions (RCP8.5). These scenarios have been hardly used to study the evolution of energy performance and thermal comfort and are focused on analysing their impact in terms of energy consumption (Cellura, Guarino, Longo, & Tumminia, 2018; Kikumoto, Ooka, Arima, & Yamanaka, 2015; Zhai & Helman, 2019). Thus, this paper analyses the influence of climate change on the design strategies adapted to climate. Thus, this study offers a new approach to the effectiveness of design strategies by considering RCP scenarios. In this regard, the lack of existing studies on RCP scenarios shows the need to analyze the effectiveness of energy improvements in the building stock. The analysis was mainly based on both the variation of the effectiveness of the design strategies to keep thermal comfort and the need for using HVAC systems. For this purpose, the climate zones existing in 6 countries were analysed: 3 developed countries (France, Portugal, and Spain) and 3 developing countries (Argentina, Brazil, and Chile). These countries were selected because of both the characteristics of their climate zones and their regulation on energy efficiency, very similar among them (Bienvenido-Huertas, Oliveira, Rubio-Bellido, & Marín, 2019). Each country stresses the importance of the envelope properties in building energy efficiency. However, design criteria are not established, so the results presented in this paper could be an opportunity for these countries to develop energy efficiency policies. Moreover, knowing the future effectiveness of design strategies leads to know the effectiveness of the energy policies today adopted (da Guarda et al., 2020). Likewise, the analysis was also based on the use of data mining algorithms, such as multilayer perceptrons and k-means, thus representing a new approach. In this regard, these techniques achieved appropriate results in energy analysis approaches (Fan, Ding, & Liao, 2019; Zhang, Cao, & Romagnoli, 2018) and estimated certain variables related to buildings under the effect of climate change (Li, (Jerry) Yu, Haghighat, & Zhang, 2019; Xia, Han, Zhao, & Liang, 2021). 2. Methodology 2.1. Determination of both design strategies and the percentage of thermal comfort hours The methodology of this paper was based on the analysis of the existing climate data in each climate zone in each country to obtain generic design strategies. For this purpose, the comfort model defined in ASHRAE Standard 55, also known as the PMV (Predicted Mean Vote) model, was used because it is generally valid for an international approach (Bienvenido-Huertas, Rubio-Bellido, P´ erez-Fargallo, & Fig. 1. Limits of the thermal comfort zone and the design strategies considered in the study. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 3 Pulido-Arcas, 2020). This model is an experimentally derived algorithm, which considers dry bulb temperature, humidity, air velocity, metabolic activity, and clothing insulation. All this climate variables can be depicted in a psychrometric chart to generate the 16 design strategies considered in the standard. This paper assessed the hourly data of each location (8760 h a year) within the strategies by using the Climate consultant tool. This tool was developed by Ligget and Milne of the Department of Architecture and Urban Design at the University of California (Los Angeles) (UCLA Energy Design Tools Group, 2018). Several studies have used Climate Consultant to study architectural design strategies, but they are scarce. However, it has been recently more used: (i) Bougiatioti and Oikonomou (2020) used Climate Consultant to determine the possibilities of using passive strategies to improve thermal comfort in cold periods in the eastern Mediterranean; (ii) Khotbehsara et al. (2019) analysed the most appropriate passive design strategies in 4 climate zones in Iran; (iii) Osman and Sevinc (2019) determined the design strategies adapted in this research work for the current and future climate of Jartum (Sudan); (iv) Awad and Abd-Rabo (2020) studied the influence of the design strategies on thermal comfort in buildings in the United Arab Emirates; and (v) Gaber et al. (2020) concluded that the most appropriate design strategies for Alexandria are based on solar radiation and the air movement. For this purpose, they were based on the strategies determined through Climate Consultant. Likewise, Climate Consultant is of great interest to analyse certain climate variables (e.g., outdoor temperature or radiation), as some studies have shown: Cao, Bui, and Kjøniksen (2019), Costanzo and Donn (2017), Shafiee, Faizi, Yazdanfar, and Khanmohammadi (2020), and Daemei, Azmoodeh, Zamani, and Khotbehsara (2018). In this study, Climate Consultant was used to determine both the most appropriate design strategy for the climate and the variation of the percentage of thermal comfort annual hours and of the percentage of hours when the use of HVAC systems was required. For this purpose, a psychometric diagram represents the increases in the thermal comfort zones by applying design strategies (DS). In addition, the thermal comfort model used was the model defined in ASHRAE 55-2017 (the predicted mean vote model) (ASHRAE, 2017). This model considered the dry-bulb temperature, relative humidity, air speed, and the metabolic activity. Moreover, the predicted mean vote model is determined Fig. 2. Thermal zones in the 6 countries analysed in the study. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 4 by estimating the clothing used in winter and summer, i.e., it is characterised by defining two thermal comfort zones according to occupants’ clothing insulation: zone of 0.5 clo (corresponding to the use of air conditioning systems) and zone of 1.0 clo (corresponding to the use of heating systems) (Fig. 1). It also considers that users are more comfortable with dryer air when temperatures are slightly higher. The use of this thermal comfort model allows both users’ thermal comfort needs to be known and clothing to be changed according to the season (Kim, Tartarini, Parkinson, Cooper, & de Dear, 2019; Oh, Haberl, & Baltazar, 2020; Ryu, Hong, Seo, & Seo, 2017). Through 11 design strategies, tolerances can be applied to each heating or cooling zone (Fig. 1). Design strategies between 2 and 7 increase the summer thermal comfort zone, and design strategies between 8 and 11 increase the winter thermal comfort zone. 2.2. Climate zones and the obtaining of climate data This study analysed the climate zones in Argentina, Brazil, Chile, France, Portugal, and Spain because of the typical characteristics of the climate zones in each country, defined in their effective construction codes (Fig. 2). These zones encompass from very cold regions to very warm regions. In Argentina, the IRAM 11603 standard (Instituto Argentino de Normalizaci´ on y Certificaci´ on, 2012) divides the climate of the country into 6 main zones: from the hottest (zone I) to the coldest (zone VI). In zones I-IV, there are 4 subcategories which better classify the various microclimates of the country: a (zones with thermal amplitudes greater than 14 ◦C), b (zones with thermal amplitudes lower than 14 ◦C), c (transition zones from zones with greater thermal amplitudes to other with lower thermal amplitudes), and d (coastal zones with low amplitudes throughout the year). The climate zones in Brazil are included in NBR 15220 (Associaç˜ ao Brasileira de Normas T´ ecnicas, 2005), which divides the country into 8 zones according to the positions of the hourly records in the psychometric diagram. In Chile, one of the existing and recent climate classifications is included in the NCh 1079 standard (Instituto Nacional de Normalizaci´ on (Chile) (2008)). This standard conducts a bioclimatic classification of the country in 9 zones, clearly influenced by the latitude and altitude of the zones in the country. In the European countries, climate classifications are based on the combination of two indicators which classify climate according to the type of winter and summer. In France, the climate classification is included in the RT2012 standard (Republique Française, 2010). This classification divides the country into 8 zones obtained by combining 3 winter zones (H1, H2, and H3) and 4 summer zones (a, b, c, and d). In Portugal, the climate classification is included in the decree-law 80/2006 (Minist´ erio das Obras Públicas, 2006). Finally, the climate classification in Spain is included in the royal decree 314/2006 (The Government of Spain, 2006) and in its update with the royal decree 732/2019 (The Government of Spain, 2019). The classification distinguishes 5 winter zones (A, B, C, D, and E) and 4 summer zones (1, 2, 3, and 4) according to the degree days. Winter zones go from the less severe (zone A) to the most severe (zone E), and summer zones go from the less hot (zone 1) to the hottest (zone 4). The climate data of each zone were obtained through METEONORM, a software composed of climate data from 8,325 weather stations divided throughout the planet and widely used (Bellia, Pedace, & Fragliasso, 2015; Hatwaambo, Jain, Perers, & Karlsson, 2009; Kameni et al., 2019), in both the current scenario and climate change scenarios. The coordinates of the selected locations in each zone are indicated in Table 1. The future climate data were obtained with 3 RCP scenarios: RCP 2.6 (low), RCP 4.5 (intermediate), and RCP 8.5 (high). Each scenario considers different evolution tendencies of the greenhouse gas emissions throughout the 21 st century: RCP 2.6 is the scenario closer to the fulfilment of the decarbonisation goals, and RCP 8.5 is the most unfavourable, with an increase in the global mean temperature between 2.6 and 4.8 ◦C. Climate data were obtained for each scenario in each decade of the 21 st century after the performance of this study (i.e., 2030, 2040, 2050, 2060, 2070, 2080, 2090, and 2100). A total of 25 climate data were obtained by each zone (1 from the current scenario, and 24 from future scenarios), resulting in 58 climate zones, so the climate data analysed in this study were 1,450. Each climate data was analysed with Climate Consultant, determining the percentage of thermal comfort hours and the needs for using HVAC systems with and without design strategies (Fig. 3). Thermal comfort assessment and the need for HVAC systems without design strategies are understood when Table 1 Coordinates of the selected climate data in each zone. Zone Latitude Longitude Altitude Zone Latitude Longitude Altitude AR-Ia −29.467 −60.217 30 FR-H1A 48.857 2.351 62 AR-Ib −27.954 −58.809 60 FR-H1B 47.902 1.904 105 AR-IIa −28.568 −66.801 859 FR-H1C 45.187 5.726 215 AR-IIb −31.392 −58.017 45 FR-H2A 47.655 −2.762 13 AR-IIIa −32.750 −60.733 30 FR-H2B 47.084 2.396 152 AR-IIIb −34.600 −58.382 0 FR-H2C 44.350 2.574 590 AR-IVa −22.745 −65.897 3787 FR-H2D 44.735 4.599 305 AR-IVb −37.894 −66.760 30 FR-H3 43.837 4.360 54 AR-IVc −39.524 −69.280 473 PO-I1V1 37.328 −8.600 28 AR-IVd −37.846 −58.256 143 PO-I1V2 38.524 −8.893 58 AR-V −43.081 −68.307 120 PO-I1V3 38.573 −7.907 244 AR-VI −47.412 −70.877 1440 PO-I2V1 41.149 −8.611 73 BR-Z1 −29.168 −51.179 535 PO-I2V2 41.533 −8.417 123 BR-Z2 −25.095 −50.162 850 PO-I2V3 39.823 −7.493 269 BR-Z3 −27.593 −48.553 150 PO-I3V1 40.536 −7.268 816 BR-Z4 −15.794 −47.883 1111 PO-I3V2 41.300 −7.740 607 BR-Z5 −23.961 −46.334 90 PO-I3V3 41.083 −7.867 864 BR-Z6 −16.679 −49.254 765 SP-A3 36.517 −6.283 0 BR-Z7 −8.087 −42.052 363 SP-A4 36.833 −2.450 0 BR-Z8 −1.456 −48.504 16 SP-B3 39.567 2.650 26 CH-An −22.911 −68.200 2569 SP-B4 37.383 −5.983 17 CH-CI −33.450 −70.667 486 SP-C1 43.367 −8.383 0 CH-CL −33.046 −71.616 13 SP-C2 41.383 2.177 0 CH-ND −22.462 −68.927 2341 SP-C3 37.178 −3.601 711 CH-NL −23.646 −70.398 178 SP-C4 38.880 −6.975 196 CH-NVT −30.017 −70.700 835 SP-D1 43.012 −7.557 393 CH-SE −47.267 −72.550 910 SP-D2 40.965 −5.664 806 CH-SI −38.529 −72.435 229 SP-D3 40.419 −3.692 681 CH-SL −36.833 −73.050 49 SP-E1 40.654 −4.696 1121 D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 5 thermal comfort was assessed only for zones of 0.5 and 1.0 clo. As for the building assessment with design strategies, the increases in the thermal comfort limits were considered in the psychometric diagram with the optimal strategies determined by Climate Consultant. 2.3. Data mining This study applied data mining approaches. The first approach was based on searching similarities among the climate zones analysed. For this purpose, cluster analyses were performed. The algorithm k-means was used like in other similar climate analysis studies (Bienvenido-- Huertas, S´ anchez-García, Rubio-Bellido, & Pulido-Arcas, 2021). This algorithm consists of grouping a set of individuals into k-groups. The algorithm begins with a random selection of k individuals from the dataset. Through the analysis of the Euclidean distances to the centroids, an iterative process is carried out until obtaining the optimal group (Hartigan & Wong, 1979). The value of k is therefore a key parameter in the grouping process. The elbow method is used to select the appropriate k-value. Through this method, the representation of the total within-cluster sum of squares allows the optimal value of k to be detected. This study used this method to determine the most appropriate input k-value. Likewise, the silhouette index was also used to define in detail the initial estimate of k obtained with the elbow method. The silhouette index (s(i)) is obtained through Eq. (1) and represents the similarity of each instance with the centroid to which it has been assigned. Thus, the analysis is individually performed in each dataset instance, and the average value represents the global level of the group. Fig. 3. Example of the analysis of the design strategies in the B4 thermal zone (Spain). D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 6 The values obtained with this index oscillate between -1 and 1. Positive values indicate that the instance is in the correct group, while negative values indicate the opposite. The variables considered in the cluster analysis were analysed with 3 dimensions according to the energy needs of each location: percentage of heating hours, percentage of cooling hours, and percentage of dehumidification hours. The analysis was carried out with the climate data of the current scenario and with those of the RCP 2.6, 4.5, and 8.5 scenarios in 2100. s(i) = b(i) − a(i) max{a(i),b(i) } (1) where a(i)is the average distance between instance (i) and the remaining points of the same cluster; and b(i)is the minimum average distance between the instance (i) and the remaining clusters. Likewise, neural networks were used in the study. Their use was associated with the possibility of estimating the energy needs of each location through their coordinates. Although representative locations were selected for each climate zone, the availability of prediction tools allows energy needs to be estimated in other locations. This responds to the possible differences among the locations in the same climate zone (Bienvenido-Huertas et al., 2021). The neural network model used was the multilayer perceptron. This algorithm is characterized by its universal approximation capabilities (Barron, 1993; Cybenko, 1989; Hornik, Stinchcombe, & White, 1989), and is one of the most used algorithms in data mining approaches (Pino-Mejías, P´ erez-Fargallo, Rubio-Bellido, & Pulido-Arcas, 2018; Pino-Mejías, P´ erez-Fargallo, Rubio-Bellido, & Pulido-Arcas, 2017). The architecture of multilayer perceptrons is made up of an input layer, an intermediate layer, and an output layer. The independent variables are in the input layer, and the dependent variable in the output layer. Generally, models that only estimate a single variable are designed (Pino-Mejías et al., 2017, 2018). The multilayer perceptrons analysed are included in Table 2. Moreover, the models were trained by backpropagation (Rumelhart, Hinton, & Williams, 1986; Wang, 1994; Werbos, 1974) and using the Broyden-Fletcher-Goldfarb-Shanno algorithm (Fletcher, 1980). The quality of the estimates was statistically analysed with the coefficient of determination (R 2 ), the mean absolute error (MAE), and the root mean square error (RMSE): R2= ⎛ ⎜ ⎜ ⎝ 1−∑ n i=1 (ai−pi)2 ∑ n i=1 (ai−ai)2 ⎞ ⎟ ⎟ ⎠ (2) MAE =∑ n i=1 |ai−pi| n(3) RMSE = ⎛ ⎜ ⎜ ⎝ ∑ n i=1 (ai−pi)2 n ⎞ ⎟ ⎟ ⎠ 1/2 (4) where pi is the predicted value, ai is the actual value, and n is the number of instances in the dataset. 2.4. Limitations of the study Some limitations should be considered when analysing the results. On the one hand, the thermal comfort model is not based on an adaptive behaviour pattern. The reason is the own limitations of the tool used to determine the design strategies. An adaptive use on the part of users would vary the thermal comfort limits established in ASHRAE 55-2017 in zones of 0.5 and 1 clo, so the percentage of hours to use HVAC systems could vary. On the other hand, this study does not consider the effect of possible urban heat islands (UHI). These phenomena cause that the air temperature in cities is higher than in the surroundings (Santamouris & Asimakopoulos, 1996). This aspect contributes to a greater cooling energy demand in buildings located within an UHI (Hassid et al., 2000; Salvati, Coch Roura, & Cecere, 2017), thus implying a greater risk for population (S´ anchez-Guevara S´ anchez, Nú˜ nez Peir´ o, Taylor, Mavrogianni, & Neila Gonz´ alez, 2019). 3. Results and discussion The analysis was first focused on the evolution of the thermal comfort hours. Table 3 shows the percentage values of the thermal comfort hours in the current scenario, Fig. 4 shows the percentage values obtained throughout the 21 st century with the three RCP scenarios, and Figs. A1–A6 show the time series. In the values obtained in the current scenario, appropriate design strategies adapted to climate significantly increased the percentage of thermal comfort hours in comparison with the buildings without those design strategies. The use of these measures implied an average increase of 41.48 %, with maximum increase values of 82 % (CH-ND). The effectiveness of the design strategies was reduced only in BR-Z7 and BR-Z8 because of the many hourly values with high temperatures and absolute humidity, which demanded the use of air conditioning systems. Nonetheless, the use of design strategies adapted to climate significantly increased the thermal comfort hours and reduced the use of HVAC systems (thus reducing its energy consumption). However, the effectiveness of the design strategies adapted to climate could varied throughout the 21 st century. Two aspects were crucial to know the variability presented by the percentage of thermal comfort hours: the RCP scenario and the type of climate zone. Thus, RCP 2.6 was characterised by having percentage variations close to zero. In this regard, the average annual values of the variations with RCP 2.6 oscillated between -0.38 and 0.50 % in buildings without design strategies, and between -0.66 and 0.37 % in buildings with design strategies. These percentage values caused that the tendencies of the time series of RCP 2.6 were almost horizontal with similar values in the current scenario and in 2100, with an average deviation of 0.40 %. Moreover, the effectiveness of the design strategies did not vary with RCP 2.6 in all the climate zones, obtaining very similar values to those obtained in the current scenario: an average increase value of 41.00 % in the thermal comfort hours and the same tendency detected in the climate zones, with the worst values in BR-Z7 and BR-Z8. This was due to the characteristics of RCP 2.6, which considers that the environmental policies are successful to reduce climate change effects. However, the influence of climate change on the percentage of thermal comfort hours was detected in the other two scenarios. In RCP 4.5, the climate characteristics of each region could vary the percentages of thermal comfort hours. Thus, cold climate zones, such as those in Chile, increased the percentage of thermal comfort hours between 1.00 Table 2 Input and output variable approaches analyzed for multilayer perceptrons. Type Output variable Input variable Without design strategies Percentage of heating hours Latitude, longitude, altitude, year, scenario Percentage of cooling hours Percentage of dehumidification hours With design strategies Percentage of heating hours Latitude, longitude, altitude, year, scenario Percentage of cooling hours Percentage of dehumidification hours D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 7 Table 3 Annual percentage of thermal comfort hours according to the design of the building in the current scenario. Zone Percentage of thermal comfort hours (without design strategies) [%] Percentage of thermal comfort hours (with design strategies) [%] Zone Percentage of thermal comfort hours (without design strategies) [%] Percentage of thermal comfort hours (with design strategies) [%] AR-Ia 14 49 FRH1A 10 46 AR-Ib 13 44 FRH1B 10 42 AR-IIa 23 68 FRH1C 11 46 AR-IIb 18 59 FRH2A 7 46 AR-IIIa 16 55 FRH2B 10 43 AR-IIIb 18 60 FRH2C 9 43 AR-IVa 2 45 FRH2D 14 52 AR-IVb 16 60 FR-H3 16 56 AR-IVc 15 56 POI1V1 20 68 AR-IVd 12 54 POI1V2 19 69 AR-V 7 40 POI1V3 17 63 AR-VI 4 39 POI2V1 12 63 BR-Z1 22 64 POI2V2 11 61 BR-Z2 18 68 POI2V3 18 61 BR-Z3 11 46 POI3V1 11 52 BR-Z4 34 73 POI3V2 12 52 BR-Z5 16 49 POI3V3 10 49 BR-Z6 30 66 SP-A3 20 64 BR-Z7 5 10 SP-A4 20 67 BR-Z8 0 0 SP-B3 10 47 CH-An 7 80 SP-B4 21 66 CH-CI 17 63 SP-C1 10 64 CH-CL 21 66 SP-C2 14 50 CH-ND 7 89 SP-C3 17 62 CH-NL 13 89 SP-C4 19 64 CHNVT 4 57 SP-D1 7 48 CH-SE 6 44 SP-D2 12 48 CH-SI 9 42 SP-D3 19 62 CH-SL 8 52 SP-E1 13 50 D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 8 Fig. 4. Evolution of the percentage of thermal comfort hours according to both the design of the building and the climate change scenario. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 9 and 16.00 % in the building without design strategies, and between 2.83 and 10.90 % in the building with design strategies, except CH-NL in which the percentage of thermal comfort hours was slightly reduced by 4.89 %. Likewise, in the cold climate zones in the other countries (ARIVa, AR-IVb, AR-IVc, AR-V, AR-VI, FR-H1A, FR-H1B, FR-H2A, SP-D1, SP-D2, SP-D3, and SP-E1), the percentage of thermal comfort hours increased between 0.10 and 4.00 % in the building with design strategies, and between 0.38 and 6.18 % in the building with design strategies. Fig. 5. Evolution of the percentage of hours with the need for using HVAC systems in Argentina according to both the building design and the climate change scenario. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 16 appropriate. Moreover, the coefficient of determination was always greater than 90 %, with values in some cases close to 95 %. Likewise, the percentage of error ranged between 2.99 and 6.11 % in MAE, and between 4.02 and 7.74 % in RMSE. Thus, the estimated values were close to the actual values, so multilayer perceptrons could be used by architects, technicians, and policy makers to have a more extensive knowledge of the energy requirements of any location. This would avoid the need for a detailed analysis of each climate characteristic. It is worth stressing that the approach of the building with design strategies was determined based on the existing climate data in the current scenario, thus reproducing the behaviour of future buildings designed by architects and engineers. However, architects and engineers Table 5 Centroids and number of zones associated with each cluster. Scenario Cluster Centroid Number of zones Percentage of cooling hours [%] Percentage of heating hours [%] Percentage of dehumidification hours [%] Current 1 63.68 0.00 33.83 2 2 8.27 72.54 1.73 12 3 12.86 58.28 10.67 12 4 15.89 36.10 28.24 6 5 1.64 87.46 1.86 26 RCP 2.6 (2100) 1 73.20 0.00 25.50 2 2 9.36 71.54 2.65 17 3 15.69 54.70 12.82 12 4 21.30 30.22 31.65 6 5 1.46 85.73 2.60 21 RCP 4.5 (2100) 1 91.23 0.00 8.69 2 2 7.52 70.41 5.99 23 3 21.43 52.17 10.20 15 4 28.59 25.03 33.15 8 5 1.02 86.94 1.40 10 RCP 8.5 (2100) 1 99.44 0.00 0.56 2 2 12.88 59.71 10.65 26 3 31.36 42.54 11.67 15 4 43.54 15.89 31.73 8 5 2.12 83.05 0.91 7 Fig. 12. Point clouds between the actual values and the values predicted by the multilayer perceptrons. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 17 Fig. 13. Box-plots with the saving in heating and cooling hours between the performance of the building with design strategies of the current scenario in the future and the optimisation of the design strategies adapted to future years. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 18 could use future climate scenarios to determine the most appropriate design based on future climate variations. For this reason, the variation of the hours using HVAC systems was analysed in buildings with design strategies adapted to the climate change scenario. Fig. 13 shows the distributions of the saving obtained in heating and cooling hours (the saving in dehumidification hours was not included as they were not obtained with the design strategies adapted to climate change). The optimisation of design strategies affected more zones to save cooling hours than heating hours. The number of heating zones with possibility of reducing heating hours was 20, and the number of zones with possibility of reducing cooling hours was 34. Nevertheless, the quartile values of the saving in heating hours were greater than in cooling. This corresponded to the tendency with greater effectiveness of the design strategies to reduce the heating energy demand instead of cooling energy demand. Regarding the possible variation of the design strategies, Fig. 14 shows the annual percentage of application of the design strategies adapted to each design, i.e., the percentage value to apply design strategies in the decades of the 21 st century. As indicated in Section 2, the design strategies increased the limits associated with the thermal comfort zones of both 0.5 and 1.0 clo from ASHRAE 55-2017. Thus, the use of these design strategies covered more thermal comfort hours than in the case of not having these strategies. The selection of the strategies was optimized to use as less number as possible of strategies with the greatest percentage of thermal comfort hours. The design strategies varied according to the type of climate, with the common aspect of the prevalence of solar protection (DS-1). Thus, the design strategies based Fig. 14. Percentage of years to apply the design strategies throughout the 21 st century, adapting them to the climate variations of each RCP scenario in the climate zones in Argentina. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 19 Fig. A1. Evolution of the percentage of thermal comfort hours in Argentina according to both the design of the building and the climate change scenario. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 20 on both the gains of internal loads (DS-8) and the solar gain in high thermal mass elements (DS-10) were the most appropriate in the climate zones in Argentina. These design strategies are like those obtained in Chile. However, the need for using design strategies adapted to cooling in both countries was different. In Argentina, it was appropriate to use design strategies to reduce the cooling load, such as the use of high thermal mass elements with night dissipation (DS-3) or night ventilation (DS-6), with the latter being the most appropriate strategy in all the decades of the 21 st century; in Chile, the percentage of using these strategies was low, and it was obtained only in the combinations of the most severe RCP scenarios at the end of the 21 st century (e.g., RCP 8.5 in 2090 or in 2100). These results followed the same tendency as in Rubio-Bellido, Pulido-Arcas, and Ureta-Gragera (2015) in the cities of Santiago de Chile and Concepci´ on with the A2 climate change scenario. Similarly, the same heating design strategies (DS-8 and DS-10) were more used in Brazil, although the percentage of heating hours were not so high in the climate zones in the country. Likewise, natural ventilation was strongly important in most of the country. The only existing zone in the country where the design strategies were not effective was BR-Z8 because of the high outdoor temperatures. In the European countries considered in the research, the use of heating (DS-8, DS-9, and DS-10) and cooling design strategies (DS-3, DS-5, and DS-6) were appropriate for the various climates in each country. Nonetheless, the effectiveness of the design strategies adapted to heating is expected to be reduced with RCP 8.5. Thus, these results showed that climate change will vary the effectiveness of passive design strategies in terms of thermal comfort hours. However, the climate zone and the climate change scenario strongly influenced this variation. Thus, achieving international low-carbon goals (RCP 2.6) could guarantee the effectiveness of passive building strategies. However, if other greenhouse gas emissions scenarios emerge, design strategies would be less effective in most of the climate zones due to the double effect produced by the loss of heating energy demand, together with the low effectiveness of passive cooling design strategies due to the increase in outdoor temperatures. Nonetheless, the effects produced by climate change were not negative in all regions. In the cold climate zones analysed in this research, such as the zones in Chile or Argentina, climate change contributed to thermal comfort Fig. A2. Evolution of the percentage of thermal comfort hours in Brazil according to both the design of the building and the climate change scenario. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 21 conditions and decreased heating energy demand. Nonetheless, this was an isolated aspect as climate change reduced thermal comfort hours in a larger number of zones. Thus, it is imperative to include the use of passive building design strategies with another measure, such as the use of effective HVAC systems (de Rubeis, Falasca, Curci, Paoletti, & Ambrosini, 2020), self-production (Fratean & Dobra, 2018; Lobaccaro, Croce, Vettorato, & Carlucci, 2018; Wu & Skye, 2018) or users’ behaviour (Bienvenido-Huertas, S´ anchez-García, Rubio-Bellido, & Oliveira, 2020; Huchuk, O’brien, & Sanner, 2020). The decarbonisation goals planned for the building sector would be achieved more quickly by contributing to reach the category of nearly zero energy consumption buildings (Attia et al., 2017). In addition, a better response of the building stock would be guarantee in view of the future climate evolution throughout the 21 st century (Ciancio et al., 2020; Zhai & Helman, 2019) 4. Conclusions This study analyses the potential of applying passive buildings design strategies in view of the climate change effect. For this purpose, the climate zones in 3 developed countries (France, Portugal, and Spain) and in 3 developing countries (Argentina, Brazil, and Chile) included in the regulations on building energy efficiency were used. In each zone, the climate data of both the current scenario and the RCP scenarios (RCP 2.6, RCP 4.5, and RCP 8.5) were analysed in each decade of the 21 st century. A total of 1,450 climate files were analysed, and the following conclusions were drawn: Fig. A3. Evolution of the percentage of thermal comfort hours in Chile according to both the design of the building and the climate change scenario. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 22 - Thermal comfort in indoor spaces presented various tendencies according to the type of climate and scenario. Thus, RCP 2.6 (i.e., the scenario in which the environmental policies are successful) did not vary thermal comfort hours at the end of the 21 st century. However, RCP 4.5 and 8.5 generated two effects according to the climate: the cold climate zones were favoured by the climate change effect as the number of thermal comfort hours increased, and these hours were reduced in the other zones. - The use of passive design strategies improved users’ thermal comfort in comparison with buildings without these design strategies. In this regard, the buildings with these design strategies obtained an average increase of 41 % in the percentage of thermal comfort hours. Nonetheless, RCP 4.5 and 8.5 could reduce the effectiveness of these strategies to keep a number of thermal comfort hours similar to that of the current scenario. - The percentage of hours with the need for using HVAC systems was also influenced by the RCP scenarios. RCP 2.6 again obtained almost the same result at the end of the century in comparison with the current scenario, with variations between 1 and 6% in the percentage of hours. However, RCP 4.5 and 8.5 considerably reduced the percentage of heating hours, whereas the percentage of cooling hours increased. This greater prevalence of cooling demand in the future implied an effectiveness loss of the passive design strategies as their capacity is worse to guarantee users’ thermal comfort in the hours with a high temperature and humidity. - Design strategies implied the need for considering dynamic and resilient building designs due to the variability presented by the most appropriate design strategies for each zone according to both the year and the RCP scenario. Thus, the use of dynamic building designs would guarantee a greater adaptation in view of the energy demand variations. In this regard, the use of design strategies based on the gains with internal loads and with high thermal mass elements could be interesting in the short term to reduce the heating demand existing in most zones, although the use of strategies to reduce cooling energy demand (e.g. the use of high thermal mass elements with night dissipation) could be more important in the medium and long term. However, this does mean that these strategies are not used in the current scenario; it depends on the climate characteristics of Fig. A4. Evolution of the percentage of thermal comfort hours in France according to both the design of the building and the climate change scenario. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 23 the building. Moreover, the temperature rise estimates could be low in comparison with the actual ones as the effectiveness of the passive strategies could be removed. In this regard, Z8 in Brazil could be an example of how the effectiveness of the strategies could be if a more unfavourable climate than that of RCP 8.5 took place as energy demands could not be reduced with passive strategies. - The use of data mining algorithms could be helpful to understand the influence of climate change on design strategies and to establish policies. Through the cluster analysis, 5 similarity groups were established between the zones with different centroids. These groups have their own characteristics that differentiate them from the others and allow appropriate design measures to be established; however, the evolution of the climate throughout the 21 st century will vary the similarities among locations. Therefore, both the classification of the zones in the current scenario and the changes expected in the future should be considered. Moreover, multilayer perceptrons are an appropriate tool to estimate the percentage variables of the need for using HVAC systems through the coordinates of the location analysed. Therefore, architects and engineers could use multilayer perceptrons to know the energy requirements of any location without the need for a detailed study. To conclude, the results of this study are of interest for architects and engineers to have a knowledge framework to design efficient buildings in the regions analysed. Thus, there would be a quicker decarbonisation of the building stock by constructing buildings adapted to the climate and not just as per the current climate but considering the future evolution. Nonetheless, the limitations to guarantee high percentages of Fig. A5. Evolution of the percentage of thermal comfort hours in Portugal according to both the design of the building and the climate change scenario. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 24 thermal comfort hours in the current scenario, together with the effectiveness loss throughout the 21 st century, force to combine passive strategies with other measures, such as effective systems or users’ appropriate operational pattern. Moreover, these results are of interest for developing policies and regulations on building energy efficiency. The use of the climate zones included in the regulations of each country guarantees a greater traceability on the part of policy designers to establish effective design criteria. In this regard, this paper shows the progressive effectiveness loss presented by the passive heating design strategies, and passive cooling design strategies could slightly reduce Fig. A6. Evolution of the percentage of thermal comfort hours in Spain according to both the design of the building and the climate change scenario. D. Bienvenido-Huertas et al. Sustainable Cities and Society 72 (2021) 103042 25 cooling demands. This paper also contributes to understand that standards should be adapted by considering future scenarios because buildings’ useful life is long (between 50 and 100 years), so the standards could be energetically out-of-date, thus implying high economic investments. Likewise, the groups obtained with the cluster analyses in the various zones could be a starting point to establish interregional energy efficiency policies and regulations. Thus, these results could facilitate synergies between technicians and politicians from different countries regarding the most appropriate design strategies in each cluster. This is an essential aspect because zones from different countries and continents were grouped together. However, future steps are required to increase knowledge among the various climate zones. In this regard, the results of this study are based on the existing climate analysis in each zone. Future studies should increase the analysis by using specific case studies in each region. For this purpose, the case study should be appropriately selected due to the expected variability in the operational patterns of each country. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A References Asadi, S., Fakhari, M., & Sendi, M. (2016). A study on the thermal behavior of traditional residential buildings: Rasoulian house case study. Journal of Building Engineering, 7, 334–342. https://doi.org/10.1016/j.jobe.2016.07.012 American Society of Heating Refrigerating and Air Conditioning Engineers (ASHRAE). (2017). ASHRAE standard 55-2017 thermal environmental conditions for human occupancy. ASHRAE Inc.. Ed. Associaç˜ ao Brasileira de Normas T´ ecnicas. (2005). NBR 15220: Desempenho T´ ermico de edificaç˜ oes. Attia, S., Eleftheriou, P., Xeni, F., Morlot, R., M´ en´ ezo, C., Kostopoulos, V., et al. (2017). Overview and future challenges of nearly zero energy buildings (nZEB) design in Southern Europe. Energy and Buildings, 155(2017), 439–458. https://doi.org/ 10.1016/j.enbuild.2017.09.043 Awad, J., & Abd-Rabo, L. (2020). Daylight and energy performance optimization in hot – Arid regions: Application and adaptation guide for designers in the UAE. Procedia Manufacturing, 44, 237–244. https://doi.org/10.1016/j.promfg.2020.02.227 Barron, A. R. (1993). Universal approximation bounds for superpositions of a sigmoidal function. IEEE Transactions on Information Theory, 39(3), 930–945. Bellia, L., Pedace, A., & Fragliasso, F. (2015). The role of weather data files in Climatebased Daylight Modeling. Solar Energy, 112, 169–182. https://doi.org/10.1016/j. solener.2014.11.033 Bienvenido-Huertas, D., Oliveira, M., Rubio-Bellido, C., & Marín, D. (2019). A comparative analysis of the international regulation of thermal properties in building envelope. Sustainability, 11, 5574. https://doi.org/10.3390/su11205574 Bienvenido-Huertas, D., Rubio-Bellido, C., P´ erez-Fargallo, A., & Pulido-Arcas, J. A. (2020). Energy saving potential in current and future world built environments based on the adaptive comfort approach. Journal of Cleaner Production, 249(119306). https://doi.org/10.1016/j.jclepro.2019.119306 Bienvenido-Huertas, D., S´ anchez-García, D., Rubio-Bellido, C., & Oliveira, M. J. (2020). Influence of adaptive energy saving techniques on office buildings located in cities of the Iberian Peninsula. Sustainable Cities and Society, 53(November 2019), 101944. https://doi.org/10.1016/j.scs.2019.101944 Bienvenido-Huertas, D., S´ anchez-García, D., Rubio-Bellido, C., & Pulido-Arcas, J. A. (2021). Analysing the inequitable energy framework for the implementation of nearly zero energy buildings (nZEB) in Spain. Journal of Building Engineering, 35 (102011). https://doi.org/10.1016/j.jobe.2020.102011 Bougiatioti, F., & Oikonomou, A. (2020). Architectural characteristics and environmental performance of byzantine houses and streets. Building and Environment, 170. https:// doi.org/10.1016/j.buildenv.2019.106605 Cao, V. D., Bui, T. Q., & Kjøniksen, A. L. (2019). Thermal analysis of multi-layer walls containing geopolymer concrete and phase change materials for building applications. Energy, 186. https://doi.org/10.1016/j.energy.2019.07.122 Casquero-Modrego, N., & Go˜ ni-Modrego, M. (2019). Energy retrofit of an existing affordable building envelope in Spain, case study. Sustainable Cities and Society, 44, 395–405. https://doi.org/10.1016/j.scs.2018.09.034 Cellura, M., Guarino, F., Longo, S., & Tumminia, G. (2018). Climate change and the building sector: Modelling and energy implications to an office building in southern Europe. Energy for Sustainable Development, 45(2018), 46–65. https://doi.org/ 10.1016/j.esd.2018.05.001 Ciancio, V., Salata, F., Falasca, S., Curci, G., Golasi, I., & de Wilde, P. (2020). Energy demands of buildings in the framework of climate change: An investigation across Europe. Sustainable Cities and Society. , Article 154244. https://doi.org/10.1016/j. scs.2020.102213 Cicelsky, A., & Meir, I. A. (2014). Parametric analysis of environmentally responsive strategies for building envelopes specific for hot hyperarid regions. Sustainable Cities and Society, 13, 279–302. https://doi.org/10.1016/j.scs.2014.02.003 Climate Change Act. (2008). Climate change act 2008. Costanzo, V., & Donn, M. (2017). Thermal and visual comfort assessment of natural ventilated office buildings in Europe and North America. Energy and Buildings, 140, 210–223. https://doi.org/10.1016/j.enbuild.2017.02.003 Cybenko, G. (1989). Approximation by superpositions of a sigmoidal function. Mathematics of Control, Signals and Systems, 2(4), 303–314. da Guarda, E. L. A., Domingos, R. M. A., Jorge, S. H. M., Durante, L. C., Sanches, J. C. M., Le˜ ao, M., et al. (2020). The influence of climate change on renewable energy systems designed to achieve zero energy buildings in the present: A case study in the Brazilian Savannah. Sustainable Cities and Society, 52(July 2019). https://doi.org/ 10.1016/j.scs.2019.101843 Daemei, A. B., Azmoodeh, M., Zamani, Z., & Khotbehsara, E. M. (2018a). Experimental and simulation studies on the thermal behavior of vertical greenery system for temperature mitigation in urban spaces. Journal of Building Engineering, 20, 277–284. https://doi.org/10.1016/j.jobe.2018.07.024 Daemei, A. B., Azmoodeh, M., Zamani, Z., & Khotbehsara, E. M. (2018b). Experimental and simulation studies on the thermal behavior of vertical greenery system for temperature mitigation in urban spaces. Journal of Building Engineering, 20, 277–284. https://doi.org/10.1016/j.jobe.2018.07.024 de Rubeis, T., Falasca, S., Curci, G., Paoletti, D., & Ambrosini, D. (2020). Sensitivity of heating performance of an energy self-sufficient building to climate zone, climate change and HVAC system solutions. Sustainable Cities and Society, 61(June), 102300. https://doi.org/10.1016/j.scs.2020.102300 European Commission. (2002). Directive 2002/91/EC of the European Parliament and of the council of 16 December 2002 on the energy performance of buildings. Official Journal Of The European Union, 65–71. https://doi.org/10.1039/ap9842100196 European Commission. (2006). Action plan for energy efficiency: Realising the potential. European Commission. (2011). A Roadmap for moving to a competitive low carbon economy in 2050 (pp. 1–15). European Environment Agency. (2018). Final energy consumption by sector and fuel (2016). http://www.eea.europa.eu/data-and-maps/indicators/final-energy-cons umption-by-sector-9/assessment-1. European Union. (2010). Directive 2010/31/EU of the european parliament and of the council of 19 may 2010 on the energy performance of buildings. Official Journal of The European Union, 153. Fan, C., Ding, Y., & Liao, Y. (2019). Analysis of hourly cooling load prediction accuracy with data-mining approaches on different training time scales. Sustainable Cities and Society, 51. https://doi.org/10.1016/j.scs.2019.101717 Fletcher, R. (1980). Practical methods of optimization. John Wiley & Sons. Fratean, A., & Dobra, P. (2018). Control strategies for decreasing energy costs and increasing self-consumption in nearly zero-energy buildings. Sustainable Cities and Society, 39, 459–475. https://doi.org/10.1016/j.scs.2018.03.019 Gaber, N., Ibrahim, A. B., Rashad, A., Wahba, E., El-Sayad, Z., & Bakr, A. F. (2020). Improving pedestrian micro-climate in urban canyons: City Center of Alexandria, Egypt. Urban Climate, 34. https://doi.org/10.1016/j.uclim.2020.100670 Gaitani, N., Mihalakakou, G., & Santamouris, M. (2007). On the use of bioclimatic architecture principles in order to improve thermal comfort conditions in outdoor spaces. Building and Environment, 42(1), 317–324. https://doi.org/10.1016/j. buildenv.2005.08.018 Hamdan, D. M. A., & de Oliveira, F. L. (2019). The impact of urban design elements on microclimate in hot arid climatic conditions: Al Ain City, UAE. Energy and Buildings, 200, 86–103. https://doi.org/10.1016/j.enbuild.2019.07.028 Hartigan, J. A., & Wong, M. A. (1979). Algorithm AS 136: A k-means clustering algorithm. Journal of the Royal Statistical Society. Series C (Applied Statistics), 28(1), 100–108. Hassid, S., Santamouris, M., Papanikolaou, N., Linardi, A., Klitsikas, N., Georgakis, C., et al. (2000). The effect of the Athens heat island on air conditioning load. Energy and Buildings, 32. www.elsevier.comrlocaterenbuild. Hatwaambo, S., Jain, P. C., Perers, B., & Karlsson, B. (2009). Projected beam irradiation at low latitudes using Meteonorm database. Renewable Energy, 34(5), 1394–1398. https://doi.org/10.1016/j.renene.2008.09.011 Hornik, K., Stinchcombe, M., & White, H. (1989). Multilayer feedforward networks are universal approximators. Neural Networks, 2(5), 359–366. https://doi.org/10.1016/ 0893-6080(89)90020-8 Huchuk, B., O’brien, W., & Sanner, S. (2020). Exploring smart thermostat users’ schedule override behaviors and the energy consequences. Science and Technology for the Built Environment. https://doi.org/10.1080/23744731.2020.1814668 Instituto Argentino de Normalizaci´ on y Certificaci´ on. (2012). IRAM 11603: Acondicionamiento t´ ermico de edificios. Clasificaci´ on bioambiental de la República Argentina. Instituto Nacional de Normalizaci´ on (Chile). (2008). NCh 1079:2008. Arquitectura y construcci´ on – Zonificaci´ on clim´ atica habitacional para Chile y recomendaciones para dise˜ no arquitect´ onico. Intergovernmental Panel on Climate Change. (2007). Climate change 2007: The physical science basis. Contribution of working group I to the fourth assessment report of the intergovernmental panel on climate change. D. Bienvenido-Huertas et al.