This paper studies the thermal conditions of indoor spaces in a naturally ventilated middle-class residence in a Mediterranean climate and proposes suitable energy-efficient measures for improving comfort. Attainment of indoor thermal comfort is a key performance parameter of good housing. This includes protection from both excessive heat and excessive cold. ASHRAE (the American Society for Heating Refrigeration and Airconditioning Engineers) defines thermal comfort as “that condition of mind which expresses satisfaction with the thermal environment and is assessed by subjective evaluation” (ASHRAE, 2004). Although thermal comfort is a subjective notion, it has significant implications for occupant health and wellbeing. The World Health Organization (WHO) guidelines for health and housing highlight excessive heat and excessive cold as predisposing factors for poor health outcomes (Organization, 2018). The WHO guidelines target both existing housing stock and new housing (Organization, 2018). Excessive indoor cold increases the risk of respiratory and cardiovascular conditions and hypothermia, particularly in temperatures below 18°C [(Jevons et al., 2016); (Ryti et al., 2016)]. Excessive indoor heat impacts mortality and morbidity (Organization, 2018). This risk is exacerbated in summer and during heat waves, especially for the elderly and other vulnerable individuals [(Ormandy & Ezratty, 2016); (Ormandy & Ezratty, 2012)].

On the other hand, thermal comfort is a key driver of energy consumption and greenhouse gas emissions (Kaptan, 2019). Buildings consume more than 30 percent of global energy, 90 percent of which is for space heating and cooling. Buildings are also responsible for 40 percent of carbon dioxide emissions (Source Title, 2020). Buildings provide a significant opportunity to save energy through low-energy design for new buildings and refurbishment to improve the performance of existing housing stock (Source Title, n.d.). Hence, buildings are pivotal to attaining the Paris Agreement target of keeping global temperatures to 1.5°C below pre-industrial levels.

The range for indoor thermal comfort is generally taken to be between 20°C and 25°C (Lechner, 2014). However, with an adaptive approach, 90% of people will find indoor conditions ranging from 18.4°C to 32.9°C acceptable (see the theoretical background below). Changing the indoor thermostat temperature settings can yield huge energy savings [(Zhou et al., 2018); (Stefanović & Gordić, 2016)]. For instance, lowering the thermostat setpoint by 1°C can result in 10% heating energy savings (Hoyt et al., 2005).

An empirical study was undertaken in Cape Town, a city in a temperate climate. Regions with temperate climates, according to (Cohen & Small, 1998), have most of the world's population. Such regions arguably offer great potential for saving building operational energy and mitigating climate change. Specifically, Cape Town has a Mediterranean climate. The Mediterranean climate occurs between latitudes 30° North and 45° South on the western coasts of Africa, Australia, Europe, South America and North America (Service, 2010). The mild Mediterranean climate type offers great potential to save energy for space heating and cooling (Sánchez-García et al., 2018). But it is concerning that even in these temperate conditions, energy-wasteful approaches to thermal comfort are widespread. For example, in the mild Mediterranean region of Andalusia in Spain, 57.4% to 84% of existing homes have air-conditioning (Sánchez-García et al., 2018), a situation which must be avoided by Middle-income countries as they build new houses.

On the Koppen Geiger classification system, Mediterranean climates can be designated as either Csa or Csb. These climates have hot-dry summers. The letter “C” means that this is a Temperate climate type with an average temperature above 0°C but below 18°C in the coolest months. The letter “s” represents dry summers with less than 30mm of precipitation per summer month (summer being April to September in the case of the Northern Hemisphere and October to March in the case of the Southern Hemisphere). The third letter, "b", indicates the average temperature in the warmest month is below 22°C. Cape Town has a Csb climate [(Beck et al., 2018); (Conradie, 2012)], which is the cooler Mediterranean climate type.

Examples of other cities in the world with Csb climate type include: Concepción (Chile), Pasto (Colombia), San Carlos de Bariloche (Argentina), Nakuru (Kenya), San Cristóbal de la Laguna (Spain), Salamanca (Spain), Porto (Portugal), Guarda (Portugal), Rieti (Italy), Ohrid (North Macedonia), Bilecik (Turkey), Kütahya, (Turkey), San Francisco (United States), Santa Barbara ((United States), Seattle, (United States), Victoria, (Canada), Albany (Australia), Devonport (Australia) and Mount Gambier (Australia) (Kudacity, 2019).

Globally, the total building area is growing by 2.5% (IEA) per annum and building stock is set to double by 2050 (Source Title, 2020). Like Cape Town, many of the cities with Csb climates are located in Upper Middle-Income Countries (such as Colombia, North Macedonia and Turkey). These are emerging economies with a high demand for new building stock, which presents an immense opportunity for climate-smart housing design (Source Title, 2020).

In Cape Town, the highest mean monthly temperature is 20.9°C in January/February, and the lowest mean monthly temperature is 14.1°C in July/August (see 3.2 below). Surprisingly, in this mild Cape Town climate, buildings consume 78 percent of mostly coal-generated electricity and contribute 50 percent to carbon emissions (Cape Town, 2015). This is higher than the global average. In one study, 47 percent of Cape Town residents indicated that they use an electric heater for keeping warm in cold weather (S.E.A., 2015). Space heating accounts for 9 percent of residential energy consumption (Cape Town, 2015). Over 80 percent of mid-income respondents identified electric heaters as the most frequently used heat source (S.E.A., 2015). Air conditioning accounts for 6 percent of residential electricity use and 36 percent of commercial electricity use (Cape Town, 2015). The high levels of space heating and air-conditioning are indicative of poor indoor thermal conditions. Indigent households are subject to worse indoor conditions because they live in substandard formal and informal housing (Naicker et al., 2017).

In 2011, the South African Bureau of Standards (SABS) enacted new national building regulations for energy efficiency in buildings. The regulations, SANS 10400-XA:2011 (South African National Standard 10400 Part XA), made it obligatory to use low energy methods for thermal comfort for all new buildings in South Africa (S.A.B.S., 2011). The deemed-to-satisfy design methods of the regulations specify envelope and passive design measures for attaining indoor comfort for various climatic regions in the country. However, literature review does not reveal any systematic climatic-data-based studies of thermal conditions in South African middle-class buildings constructed prior to enaction of the regulations. Klunne (2003) aimed to identify baseline thermal comfort conditions in low-cost housing, but the research based its findings on qualitative interview data only. (Ebhojie, 2019-10) also used interviews to study the energy efficiency of low-cost housing. Some thermal comfort studies have been undertaken using measured temperatures in low-cost housing in Johannesburg and its surrounds [(Mabuya & Scholes, 2020); (Naicker et al., 2017)]. What is common about all these studies is that they focus on low-cost housing and are in South African locations with climates other than the Csb Mediterranean type. This leaves a gap in studying pre-regulation indoor thermal conditions of relatively better constructed middle-class buildings in South Africa, in Cape Town specifically and other Csb climate locations more generally.

Internationally, many adaptive comfort studies have been done for naturally ventilated buildings in climates different from the cool Csb Mediterranean type. For example, (Pathirana et al., 2019) studied adaptive comfort for naturally ventilated buildings in the tropics, while (Soflaei et al., 2020) used an adaptive approach to investigate optimal layouts for courtyards in subtropical desert climate settings. Some adaptive comfort studies have also been undertaken in Csb climates. In the Mediterranean climate city of Seville, (Sánchez-García et al., 2018) observed that building energy efficiency regulations do not yet address existing housing stock and argued that this is a missed opportunity to save energy and reduce carbon emissions in mild climates. (Rubio-Bellido et al., 2017) undertook adaptive comfort studies in a Csb climate using empirically measured temperatures, but this was for only 10% of the year. (Pérez-Fargallo et al., 2018) measured outdoor and indoor temperatures over 9 months in their adaptive study on low-income housing in Chile. But they came up with the surprising recommendation that, to save energy, lowest temperature setpoints should be legislated downwards to 13°C because, in their findings, poorer people could tolerate significantly low temperatures. This proposed setpoint is a whole 5°C below the WHO-recommended lower limit for healthy housing, and yet the authors do not explain why they ignore the WHO recommendation. In their study, (Sánchez-García et al., 2018) and (Sánchez-García et al., 2019) measured temperatures in a northwest-facing and a southwest-facing room in the Mediterranean climate of Seville in Spain. However, their study was limited to 21 days in winter, 30 days in summer and 31 days in spring (a total 81 days) instead of the full year. Furthermore, their objective was to validate a model for long-term future carbon-emission scenario forecasting in the context of climate change.

To the author’s knowledge, the current paper reports results of the first systematic adaptive comfort study of indoor thermal comfort of naturally ventilated spaces in a middle-class residence in a Csb Mediterranean climate using empirically measured temperatures over an entire year at one-hour time steps. Hence, this study makes an important contribution to a growing body of evidence for energy-efficient housing in mild Csb Mediterranean climates.

The study focuses on middle-class housing stock in Cape Town constructed before 2011. An adaptive comfort approach (see 1.1 below) is used to address the research question: For a selected middle-class house in Cape Town, constructed prior to the enaction of the national regulations for energy efficiency and operating without artificial space heating and cooling, what do actual measured temperatures reveal about indoor thermal conditions? Answering this question provides insights into the thermal conditions of a representative building type in Cape Town. The results are important for several reasons. First, by understanding thermal conditions, inferences can be made about potential implications for comfort and health in such housing. Second, they will provide a basis for decisions around suitable energy-efficient and passive solutions for thermal comfort in Cape Town and in similar Csb climates internationally. Finally, they will provide a baseline against which the efficacy of the national building regulations may be gauged in future thermal comfort studies. The results, therefore, contribute to an evidence basis for academia, policy and practice around existing housing stock in Cape Town and other cities in the world with Csb Mediterranean climate.

The following definitions are important in discussing the adaptive comfort approach:

Thermal comfort, as originally defined by ASHRAE, was based on the Predicted Mean Vote / Predicted Percentage of Dissatisfied (PMV/PPD) for artificially conditioned buildings only (Schiavon et al., 2014). Developed by (Fanger, 1972), the PMV model became the basis for ASHRAE comfort standard 55-92. The adaptive comfort hypothesis arises from dissatisfaction with the static PMV model in accounting for thermal comfort in naturally ventilated buildings. (Humphreys, 1975) argued that building occupants were comfortable in indoor thermal conditions covering a range of up to 13K. He attributed this to adaptive factors. These findings were corroborated by (Humphreys, 1978) and (Auliciems, 1981), especially for naturally ventilated buildings. The studies also revealed a strong correlation between indoor thermal neutralities in naturally ventilated buildings and outdoor climate. In a meta-analysis of field studies, (Brager & Dear, 1998) surmised that adaptive factors were behind the observed differences in thermal comfort between air-conditioned and naturally ventilated buildings. Arguing that the PMV approach is not suitable for naturally ventilated buildings, (Dear & Brager, 1998) developed the Adaptive Comfort model for naturally ventilated buildings for when outdoor temperatures are between 10°C and 33.5°C. They found that, as outdoor temperatures fluctuated, occupants of naturally ventilated buildings stayed comfortable within a wider range of indoor thermal conditions. They discovered that neutral thermal conditions in buildings that utilise heating, ventilation and air conditioning (HVAC) technologies are confined to between 21°C and 25°C compared to between 20°C and 27°C in naturally ventilated buildings. Behavioural adjustments (removing/adding clothing, opening/closing windows, switching on a fan, and cultural ones such as a siesta), physiological adjustments (both within an individual’s lifetime and intergenerationally), and psychological adjustments (real/perceived control over indoor thermal conditions, habituation, and future expectations) enable people in naturally ventilated buildings to adjust to a wide range of indoor thermal conditions. They highlighted behavioural and psychological factors as the most significant for adaptive comfort. The more building occupants perceive the availability of options to personally adjust the indoor environment in response to changes in ambient outdoor conditions, the greater the band of adaptive comfort.

Adaptive comfort is dynamically responsive to outdoor temperature variations. It is a function of the effective temperature (ET) (see definition below). For instance, in summer, people will be comfortable in indoor conditions they would consider too warm in winter. And inhabitants of cold climates would accept indoor temperatures that people in warm climates may find too cold.

In (ASHRAE, 2004), the Adaptive Comfort model was incorporated as valid for naturally conditioned buildings under the following conditions:

For any outdoor condition, it should be noted that there is a range of acceptable temperatures below and above the neutral temperatures within which people can adapt to stay thermally comfortable. Adaptive comfort-neutral temperatures and operative temperature ranges for any given location can be presented on the chart with ET as the abscissa and operative temperature as the ordinate (see (Schiavon et al., 2014)). Adaptive comfort is defined for 90% and 80% acceptability. In Table 1 below, the upper and lower limits for the adaptive comfort standard are identified using a web-based adaptive comfort visualisation tool developed by (Tartarini et al., 2020).

Table 1 above shows the range of outdoor effective temperatures (ET) for which the adaptive comfort approach to thermal comfort is valid. The table also shows the minimum and maximum possible neutral temperatures in the adaptive standard. Furthermore, the table shows the limits of adaptive comfort for 80% and 90% acceptability that correspond to the respective lower and upper ET limits.

(Givoni, 1992) proposed passive design methods for naturally ventilated buildings. Many of these methods, such as comfort ventilation, night purge (opening and closing windows) and adjustable solar shading, extend the behavioural and psychological capacity for occupants to adapt to a range of indoor thermal conditions.

By far, the most important scholars on adaptive comfort are de Dear and Brager. Having originally introduced the model in 1998, they set the theoretical premises of adaptive comfort and have authored prolifically on adaptive comfort and indoor climate. Starting by limiting adaptive comfort to natural ventilation, they have lately also extended the approach to air-conditioned spaces [(Dear & Brager, 1998); (Brager & Dear, 2001); (Dear, 2011-11); (Dear & Schiller Brager, 2001); and (Dear & Brager, 2020)].

Compared to steady state approaches to thermal comfort, adaptive comfort as originally conceived by (Dear & Brager, 1998), extends the range of acceptable thermal conditions by increasing the maximum temperature and reducing the minimum temperature for human indoor comfort (see Table 1). Significantly, recent localised empirical studies suggest that the adaptive comfort range can extend the thermal comfort range more than originally proposed. For instance, in Lima, income housing residents experienced low temperatures of 13°C as comfortable (Oraiopoulos et al., 2023-06). While residents in India were found to still be comfortable at a temperature of 35°C (Rawal et al., 2022). The key conclusion is that focused adaptive comfort studies can potentially result in increases in localised comfort ranges, energy and cost savings and climate change mitigation.

Only introduced in 1998, adaptive comfort as a topic has become so widely published that papers solely focusing on literature reviews of publications on the topic have now come to the fore ((Yao et al., 2022); (Dear et al., 2013) and (Karyono et al., 2020)).

A thermal comfort study was undertaken in a two-storey free-standing middle-class residence in Muizenberg, a Cape Town coastal location. (Zizzamia et al., 2019) define a middle-class household in South Africa as having a mean monthly per capita expenditure of 4536.20 ZAR (about 298 USD). Only about 20 percent of South Africans are stable in the middle-class while 60 percent are in transient or chronic poverty (ibid). The per capita household expenditure in the particular house is more than 50 percent higher than the above per capita expenditure. The middle-class residence used in this study is therefore taken to be a representative of mainstream acceptable construction practice in Cape Town. Hence, the residence is considered better in terms of construction and thermal quality than the formal and informal housing for the urban poor [see (Naicker et al., 2017)]. However, because the middle-class residence was constructed in 2006 and therefore predates the SANS10400XA regulations, no specific passive design measures were applied. There are also no active heating and cooling interventions in the building. The building thus provides a suitable setting to study indoor thermal conditions in a house representative of pre-regulation mainstream residential construction practice in Cape Town.

Within the house, two adjacent bedrooms on the upper floor were chosen (see Table 2 below). The two bedrooms are similar, and each has two external walls. They share the length of a southwest-oriented external wall. But the wall containing the window is oriented differently for each room. In the one room, this wall is oriented northwest (Room A), while in the other, it is oriented southeast (Room B). The main characteristics of the chosen bedrooms are described in Table 2 below.

Empirical full-year hourly data logger readings for indoor dry bulb temperature, relative humidity and dew point are captured as supplementary data for each room in the files (Room_A_Data*.xslx and Room_B_data*.xlsx) [see (Sanya, 2024)]. The data is a valuable research outcome as, to the author’s knowledge, it is the only set of publicly available full-year indoor thermal data loggings for residential (envelope-dominated rooms) in Cape Town and probably in any Csb climate type. The data may be used by any researcher in future studies. From this data, the hourly indoor dry bulb temperature loggings for each room are extracted for the adaptive comfort study and discussed further in the results below.

From the LB adaptive comfort simulations, as prevailing outdoor temperature changes, hourly values (totalling 8760 hours for the entire year) are generated for indoor target temperature and upper and lower adaptive comfort temperature limits. The target temperature values are the same for both Room A and Room B. However, the actual hourly indoor temperatures as logged in each room show some differences. The Ms-Excel file Room_A+B_Results*.xlsx [see (Sanya, 2024)] presents the above information in a spreadsheet with seven columns headed respectively as

For any given hour, when the logged dry bulb temperature in the room falls within the lower and upper bounds of adaptive comfort, indoor thermal conditions for that hour are acceptable. If the logged dry bulb temperature falls below the lower bound of adaptive comfort in that hour, indoor conditions are too cold. Where the logged bulb temperature falls above the upper limit in a particular hour, indoor conditions are too warm.

The seven-column results spreadsheet is formatted in MS Excel to allow the results to be sorted (in ascending/ descending order) and/or filtered (greater/ below any chosen value).

Sorting the results by prevailing temperature reveals that the highest effective outdoor temperature (ET) occurs in January and February (Cape Town summer) at 20.9°C. The lowest ET is 12.4°C, occurring in June and July (winter). Therefore, the temperature range of the prevailing outdoor conditions in Cape Town is within the Adaptive Comfort range of applicability (10°C to 33.5°C) highlighted in the theoretical background above. It follows that for such buildings, passive design and behavioural and psychological measures ought to be sufficient to maintain indoor comfort throughout the year without the need for artificial space conditioning.

The upper and lower limits of adaptive comfort are also captured in the spreadsheet. Sorting by the upper temperature bound shows temperatures ranging from 24.1°C to 29°C. While sorting by the lower temperature bound gives a range from 19.1°C to 21.8°C. Hence, the entire adaptive comfort zone for the rooms ranges from 19.1°C to 29°C. This study’s results indicate that in Cape Town and similar Mediterranean locations, the upper comfort range can be extended by 4°C above the 25°C used in steady methods. Considering that a 1°C change can result in 10% heating energy savings (Hoyt et al., 2005), the energy, cost and climate mitigation implications of this potential adjustment are very significant.

For each room, the LB simulation generated an adaptive chart for 90% acceptability (see Figure 2a and Figure 2b below). On the adaptive chart, the prevailing outdoor temperature is on the X-axis while the indoor operative temperature is on the Y-axis (see definitions in Table 4 above).

The lower diagonal line defines hourly lower temperature bounds of comfort. The upper diagonal line temperatures define hourly upper bounds of comfort. Between these two is a diagonal line connecting the hourly target temperatures. On the chart, each hourly indoor dry bulb temperatures as logged on the data logger is also plotted against the respect ive prevailing outdoor temperature – with a collection of plots represented in a colour gradation (blue – fewer hours in this region of the chart; red – most hours in this region of the chart; and others in the colours in between).

Examining Figure 2a shows that when prevailing outdoor conditions approach the maximum of 20.9°C, indoor dry bulb temperatures as measured in Room A are between 23°C – 25°C for most hours of the year and hence are within the comfortable thermal range. This also corresponds to the most frequent condition of adaptive comfort in Room A. However, the chart also indicates that when prevailing outdoor temperatures fall to 13°C and lower, Room A becomes excessively cold for a significant number of hours. Very few hours in the Room have thermal conditions above the upper the bound of comfort.

On the other hand, the adaptive chart in Figure 2b reveals that most temperatures measured in Room B are below the lower comfort bound for most of the year. Even when the prevailing outdoor temperatures approach the 20.9°C maximum mean monthly, about half the hourly indoor temperatures in the room remain below the lower bounds of adaptive comfort. Only a few summer temperatures in Room B are within the band of adaptive comfort. When the prevailing outdoor temperature falls to 13°C and below, all hourly indoor temperatures (covering a substantial number of hours in the year) are below the lower adaptive comfort boundary. The chart shows that indoor conditions in Room B fall below the lower comfort bound for most hours of the year. And none of the hourly temperatures measured in Room B exceed the adaptive comfort upper boundary at any time in the year.

Another important outcome of the simulation is the “percentage comfortable” – hours of the year when the room is within bounds of adaptive comfort. At 55 percent, the hours of the year for which the logged indoor operative temperatures are within the bounds of comfort in Room A (northwest orientation – towards the Equator) are more than double those of Room B (southeast orientation – away from the Equator) at 25 percent. The implication is that in Cape Town, orientation has a very significant impact on indoor thermal comfort. It can be expected that a room oriented ideally to true north (towards the Equator) would perform even better than a room oriented to the south (opposite the Equator). Therefore, in Cape Town and similar Mediterrenean climates, the main living spaces and rooms should, where possible, be orientated towards the Equator side (anywhere from 45^o west of true north to 45° east of true north) while south-facing spaces should be largely be limited to auxillary functions like storerooms, bathrooms and garages.

Considering that the upper bound of comfort in Room A is only exceeded for a few hours and not at all in Room B, it is concluded that, ignoring the influence of direct solar radiation into an indoor space, overheating is not a comfort concern for an envelope dominated naturally ventilated building in Cape Town and similar Mediterranean regions. The design implication here is that sun shading of the openings is sufficient to eliminate indoor space overheating in naturally ventilated envelope-dominated buildings in Cape Town. Such shading (whether fixed or movable) needs to be designed in such a way as to allow for sun radiation penetration to warm spaces when needed for passive heating.

Many of the recorded operative temperatures in Room A and Room B fall below the adaptive comfort boundary during both the cold and warm parts of the year (i.e. temperatures are too cold). The LB simulation results show that the measured operative temperatures in Room B are too cold for 75 percent (6570 hours) of the year. For Room A, the comparable figure of 45 percent (3942 hours) is lower but still significant. This again underscores the importance of good orientation in maintaining thermal comfort in Cape Town. But the most important conclusion here is that because of predominance of excessive indoor cold throughout the year, design measures for conserving internal heat gains (e.g. insulation, elimination of excessive air leakage, low envelope/volume ratio, reduction of glazed fenestration area, use of special glazing) and passive solar heating are recommended. Such measures are necessary regardless of room orientation.

Another outcome from LB is hourly visualisation of the thermal conditions (blue – too cold, yellow comfortable, red - too hot) for each hour of each month of the year (see Figures 3a and 3b below). The visualisations allow a more granular analysis of the above results.

Figure 3a (see above) shows that indoor thermal conditions in Room A are comfortable for most of the year. From January to February, Room A is comfortable for 24 hours on most days. Indoor cold conditions are limited during this period. From March to the end of May, indoor conditions are still predominantly comfortable, but excessive cold creeps in, especially between 6 am and 12 noon on most days. The first few weeks of June are also mostly comfortable, but the later weeks of the month become predominantly cold. Excessive cold is the predominant condition throughout July and up to the first week of September. It is noteworthy that the graph shows swathes of comfortable hours occurring indoors, even in this winter period. From the second week of September, indoor conditions in Room A are mostly comfortable till the end of the year. But the latter half of October and some parts of November still show significant periods with cold indoor conditions especially between 6 pm and 12 noon (this can be attributed to the north-west orientation of the room receiving the sun later in the day and generally higher position of the sun as the December summer solstice approaches). There is indoor overheating for only a few hours in the entire year in Room A. Figure 3b (see above) shows that the most comfortable month in Room B is January. But by the second week of February and throughout the autumn, conditions in Room B are mostly too cold.

Starting in the last week of May and lasting throughout the winter and up to the first week of September, all temperatures in Room B are too cold. In the spring, starting in the second week of September and up to December, Room B attains some indoor thermal comfort, but, even then, conditions remain thermally cold for most of the hours, especially from 6 am to 12 noon. Long periods in October and November are also too cold in the room. January has the highest number of hourly temperatures within the comfort band. Significantly, Room B indoor temperatures are too cold even for large parts of the mild Cape Town summer.

In both Rooms A and Room B, the period 6 am and 12 noon stands out as having the most frequent occurrence of cold for most parts of the year. This period should receive focus in the design process through a responsive approach to the circadian occupancy schedule of the house (for example whether most residents at work during this time) and looking into solutions like designing for rapid morning response for passive solar heating – specifically by utilising the easterly and north-easterly sun in Cape Town [also see (Sanya, n.d.)]

To test this paper’s findings in relation to other Csb climate locations around the world, a sensitivity analysis was undertaken using the adaptive comfort simulation model developed in the study. Four cities are chosen for the sensitivity analysis based on two criteria. First, each city must have full-year hourly EPW climate data available for use in running the simulation. Second, for each city, the daily outdoor average temperatures must be within the range of 10°C to 33.5°C for which adaptive comfort is a valid method. In the southern hemisphere, Augusta (Australia) and Santiago (Chile) are chosen for the sensitivity analysis. In the northern hemisphere, Porto (Portugal) and San Francisco (United States of America) are chosen.

Replacing Cape Town outdoor temperature values with the respective daily means for the above cities gives the sensitivity analysis. Because of the inversion of seasons in the opposite hemispheres, an LB component is used to synchronise LB outdoor temperature with this study’s measured indoor temperatures where necessary. A corollary of this is that, for both hemispheres, it becomes more appropriate to describe Room A as Equator-facing and Room B as facing away from the Equator. The sensitivity analysis results are captured at the end of Figure 4 and should only be referenced as indicative but not absolute.

The sensitivity analysis reveals that Csb climate locations elsewhere give similar results to the Cape Town findings. Orientation remains a major factor for indoor comfort. That is, in all locations, the Equator-facing room is comfortable most of the time; there is no trend of overheating, but excessive cold remains a key issue. The room facing away from the Equator has a significant reduction in comfort, more especially so in the southern hemisphere compared to the northern hemisphere; excessive cold remains as the most significant indoor thermal condition in the year; overheating is the least of concerns. Therefore, the sensitivity analysis suggests that for Csb climate locations the world over, orientation towards the Equator, conservation of internal heat gains and passive heating are a sine quo non for energy efficiency.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sector/ individuals.

Not applicable.

The authors declare there is no conflict.