Temperatures and heating energy in New Zealand houses from a nationally representative study—HEEP

Energy and Buildings 39 (2007) 770–782 www.elsevier.com/locate/enbuild

Temperatures and heating energy in New Zealand houses from a nationally representative study—HEEP L.J. French *, M.J. Camilleri, N.P. Isaacs, A.R. Pollard BRANZ Ltd., Private Bag 50 908, Porirua City, New Zealand

Abstract The household energy end-use project (HEEP) has collected energy and temperature data from a randomly selected, nationally representative sample of about 400 houses throughout New Zealand. This database has been used to explore the drivers of indoor temperatures and heating energy. Initial analysis of the winter living room temperatures shows that heating type, climate and house age are the key drivers. On average, houses heated by solid fuel are the warmest, with houses heated by portable LPG and electric heaters the coldest. Over the three winter months, living rooms are below 20 8C for 83% of the time—and the living room is typically the warmest room. Central heating is in only 5% of houses. Solid fuel is the dominant heating fuel in houses. The lack of air conditioning means that summer temperatures are affected by passive influences (e.g. house design, construction). Summer temperatures are strongly influenced by the house age and the local climate—together these variables explain 69% of the variation in daytime (9 a.m. to 5 p.m.) living room temperatures. In both summer and winter newer (post-1978) houses are warmer—this is beneficial in winter, but the high temperatures in summer are potentially uncomfortable. # 2007 Elsevier B.V. All rights reserved. Keywords: Summer temperatures; Winter temperatures; New Zealand; Residential temperatures; House temperatures; Residential heating; House heating; Heating energy

1. Introduction The household energy end-use project (HEEP) has collected temperatures from living rooms and bedrooms in 397 houses, a statistically representative sample of New Zealand. Annual reports have provided preliminary results through the project, e.g. Isaacs et al. [1]. Both energy usage and temperatures were monitored for a year in each house. Temperature data and survey data collected on the house and occupants have been used in this paper to explore the winter and summer temperatures in houses. Full monitoring of random houses commenced in 1999 and was completed in 2005. During installation information was collected on the house, appliances and the hot water system, as well as the occupants [2]. Other comfort influences such as relative humidity, air movement, surface temperatures and occupant perception of comfort were not measured. HEEP recorded air temperatures at two different heights in the living room and one height in the master bedroom on a

* Corresponding author. Tel.: +64 4 237 1176. E-mail address: [email protected] (L.J. French). 0378-7788/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.enbuild.2007.02.004

10 min basis. Sensors are placed out of direct sunlight and away from heating sources. Monitoring of each house is carried out for a period of approximately 1 year. Each month data was downloaded from each logger, checked thoroughly to remove any inaccuracies, and then stored in a central database. Temperature measurements were carried out with two types of loggers (Fig. 1): a purpose-designed BRANZ temperature logger with an uncertainty of 0.2 8C; and the ‘Tinytag’ with an uncertainty of 0.3 8C. At the start of each monitoring period, the BRANZ temperature loggers were calibrated in a thermal environment with direct traceability to the New Zealand standard temperature reference (see [3]). Nicol and Humphreys [4] and Brager and de Dear [5] suggest that an acceptable internal temperature depends on the external temperature as occupants adapt to the external climate (adaptive temperatures). There is no New Zealand research on how the ambient temperature affects their perception of comfort inside houses or offices. Given the constraints of the data, for the purpose of analysis it has been necessary to select a suitable comfort range. Camilleri [6] and Jaques [7] used 20–25 8C for New Zealand summer conditions, while Donn and Thomas [8] used 16–26 8C to deal with winter and summer. For the purposes of this paper, the

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

771

Figs. 2 and 3 [11] show the North and South Island of New Zealand—the South Island sits below and slightly to the west of the North Island. The black squares on the maps are the areas where HEEP has monitored houses, and the climate zones defined by the NZ Building Code are also shown. 1.2. New Zealand building code

Fig. 1. HEEP temperature loggers.

range of 20–25 8C has been used to represent comfort temperatures. There are a number of other factors that affect people’s comfort including humidity, wind speed, clothing levels and activity levels. It is therefore possible that the occupants are comfortable outside of 20–25 8C or uncomfortable within the range.

New Zealand has required thermal insulation in new houses since 1 April 1978. These modest requirements were increased slightly for houses in the central North Island and all the South Island (Zone 3) in 2000. The other two climate zones cover the remaining area of the North Island (Zones 1 and 2). See Figs. 2 and 3 for the climate zones in New Zealand. Table 1 sets out thermal resistance requirements for common combinations of roof, wall and floor [12,13]. These thermal insulation requirements apply only to new houses. Older houses are not required to upgrade to the current standard, but in some cases roof and floor insulation has been voluntarily installed. A typical house in New Zealand is timber framed with weatherboard cladding, a timber or concrete floor (new houses are typically concrete) and a long-run steel roof. Houses are typically stand-alone with one or two levels. The construction of houses is influenced by New Zealand being an earthquake zone.

1.1. New Zealand’s climate 2. Conditioning the New Zealand home New Zealand is long and narrow, approximately 1600 km in length, with a land area of 270,000 km2, ranging from latitude 378S to 468S [9]. The winter season in New Zealand (a Southern Hemisphere country) is during the months of June, July and August. The summer months are December, January and February. The majority of homes are in a coastal climate, but the central areas of both islands are more continental. The far south is cooler than the far north. For example, the daily mean winter temperature in Invercargill (in the far south) is 6.2 8C compared to 11.9 8C in Kaikohe (in the far north). The mean summer ambient daily temperature in Kaikohe is 18.8 8C, but in Invercargill only 13.3 8C—a difference of 5.5 8C. The annual range of monthly mean temperature (difference between the mean temperature of the warmest and coldest months) is relatively small. In the top of the North Island (Northland) and in western districts of both islands it is about 8 8C, while for the remainder of the North Island and east coast districts of the South Island it is 9–10 8C. Further inland, the annual range can exceed 11 8C, reaching a maximum of 14 8C in Central Otago [10].

New Zealand houses are heated for about half the year, with little heating or cooling in the summer months. Therefore the heating schedule has an influence on temperatures in winter, while the ambient temperature and climate have the largest influence during summer. Only 4% of residential houses in the HEEP sample have air conditioners or reverse cycle heat pumps. Although the use of heat pumps is growing quickly, they currently do not play a major role in summer temperature control or energy use. For more on heating in New Zealand houses see French et al. [14] and Isaacs et al. [15]. 2.1. Winter temperatures Only 5% of New Zealand houses have central heating [16], with most houses only heating one or two rooms. Occupants tend to turn a heater on when they arrive, and off when they leave, or when the room is considered to be warm enough. As a result, temperatures that would be considered comfortable elsewhere in the temperate world are often not achieved. Occupant

Table 1 NZ building code thermal performance requirements 1978 to current Year commenced

1978 2000

Standard

NZS 4218P:1978 NZS 4218:1996

Coverage

New Zealand Zones 1 and 2 Zone 3

R-values (m2 8C/W) Ceiling

Wall

Floor

1.9 1.9 2.5

1.5 1.5 1.9

0.9 1.3 1.3

772

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

Fig. 3. South Island climate zones and HEEP locations (Reprinted with permission from Standards New Zealand).

Fig. 2. North Island climate zones and HEEP locations (Reprinted with permission from Standards New Zealand).

comments such as ‘‘we do quite well with heating, considering the house’’ or ‘‘we just watch TV in bed if it is too cold’’ are not uncommon. The most commonly heated room is the living room which is heated in the evenings in 90% of houses during weekdays and in 87% of houses during the weekends. Only in 6% of houses is there no heating of the living room. Conversely in 50% of houses the bedroom is never heated, and 68% of houses do not heat utility areas (laundry, bathroom, corridor, etc.). The HEEP monitored data was used to determine the length of the heating season for each house. The average length of the heating season ranges from 8.6 months in the cooler far south and 5.5 in the warmer north. Figs. 4 and 5 show the length of the heating—note that 12 houses (approximately 4% of the sample) heat the entire year. In general, these tend to be in the cooler parts of the country (Central North Island and South Island). Conversely, 10 houses (3%) do not appear to heat at all, but these tend to be in the warmer locations (Auckland and north), the 10 houses that do not heat are excluded in the following graphs and tables. As space heating is most common in the evening, for the purpose of the following analysis the ‘winter evening’ (between

5 and 11 p.m. from June to August inclusive) is used as the baseline. Unless otherwise specified, the temperatures reported are for the living room (the part of the house most commonly heated). Fig. 6 shows the distribution of living room mean evening temperatures over the winter months. The mean and median temperature is 17.9 8C. The maximum mean is 23.8 8C and the minimum mean temperature is just 10 8C. Table 2 gives the mean winter temperatures for the four different periods during the day for the living room, bedroom

Fig. 4. Heating months—start and finish.

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

773

Fig. 5. Length of heating season.

and outside ambient temperature. During the day the bedroom is only 2.2 8C warmer than outside and the living room is not much better, averaging 3.8 8C warmer. The mornings are the coldest time inside the average New Zealand house, although the coldest time outside is overnight. The evenings are the warmest, not surprisingly as this is the most common heating time. The bedrooms on average always seem to be slightly lower than the living rooms—at the most there is a difference of 3.8 8C which occurs during the evening. This is most likely caused by heating occurring in the living room and, typically, very little or no heating in the bedrooms. Table 2 also can be used to explore the temperature changes in each space over the day. The mean living room temperature increases from the morning to day to evening, reflecting the use of morning heating, the availability of daytime solar energy and then the evening heating. The temperature drops from the evening to the night, again not surprisingly as only 18% of houses heat the living room at night (11 p.m. to 7 a.m.). The average time the peak temperature is reached in all houses is 5.50 p.m., and there is little regional variation. Only 15% of houses heat the bedroom during the night, but when coupled with the small heat gains from the occupants and appliances the bedroom temperatures are closer to the living room temperatures overnight and during the morning. During the day the temperature difference between the two rooms is 1.6 8C. New Zealand houses are considered low thermal mass, and this is shown in the HEEP sample with most of the houses being light timber construction. There are only two houses in the HEEP sample that have concrete block construction for walls.

Fig. 6. Distribution of mean winter evening living room temperatures.

The effects of heat storage are shown by the delayed change in temperature in the house when the ambient is changing—this is most noticeable during the day to evening period. Internal heat gains during this period are also expected to be high, e.g. meals being cooked, lights being turned on, occupants home from work/school, etc. 2.2. Summer temperatures Fig. 7 shows the distribution of living room mean day (9 a.m. to 5 p.m.) temperatures over the summer months for all HEEP houses throughout New Zealand. Eighty-five percent of the houses have a mean living room daytime temperature between 20 and 25 8C, while less than 1% are over 25 8C and just over 14% are under 20 8C. The average mean daytime living room summer temperature is 21.8 8C, the maximum mean temperature is 25.9 8C and the lowest mean temperature is 16.3 8C. Fig. 8 shows the distribution of the proportion of time between 9 a.m. and 5 p.m. that the living room temperatures are under 20 8C, between 20 and 25 8C, and over 25 8C.

Table 2 Mean winter temperatures: living room, bedroom and ambient Room

Living room Bedroom Ambient

Mean winter temperatures (8C) Morning 7–9 a.m.

Day 9 a.m. to 5 p.m.

Evening 5–11 p.m.

Night 11 p.m. to 7 a.m.

13.5 12.6 7.8

15.8 14.2 12.0

17.8 15.0 9.4

14.8 13.6 7.6

Fig. 7. Mean living room temperatures.

774

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

Over all the houses, the majority (80%) spend less than 25% of the summer day (2 h per day) at temperatures over 25 8C. The majority of the houses are between 20 and 25 8C for the majority of the time. As we have not collected data on the occupants’ perception of comfort or other climatic factors such as air changes per hour, humidity and clothing levels, it is not possible to definitively conclude that these are comfortable temperatures. However, these would be considered comfortable based on the overseas definitions. Table 3 gives the mean summer temperatures for four different periods of the day for the ambient external temperature, the bedroom and the living room. Table 3 shows on average the bedroom is always slightly cooler than the living room. Analysis of the HEEP houses has found that New Zealand houses have randomly orientated windows (on average about 25% of the total glazing is in each compass direction), with living rooms also being randomly orientated. This may explain the small temperature difference between living rooms and bedrooms, as neither can be guaranteed to benefit from the sun. Even though most New Zealand houses are of lightweight construction (timber frame with external cladding), in the absence of air conditioning Table 3 shows there is a time lag in the ambient temperature affecting the internal temperature. The internal spaces have a 4 8C mean temperature range (from 19.2 to 23.1 8C) compared to the 5.6 8C ambient temperature range (from 14.5 to 20.1 8C). Houses with high levels of thermal mass (e.g. concrete or double wall brick) would be expected to have a lower temperature range [8]. However, this could not be confirmed as there are only two such houses in the sample, and with the occupants having an effect on the temperatures in the house it is not considered appropriate to examine them separately. 3. Influences on winter temperatures Occupants heat their houses in winter, so the heating schedule plays a big part in the internal temperatures. The climate, the fuel and the heater type used in the house are also important for achieving higher indoor temperatures, as is the age of the house and the level of thermal insulation. Other factors such as household income have been examined, and they show no significant correlation between income and mean living room temperatures [15]. Further analysis using the equivalised income (taking into account household size) also found no significant correlation between income and the mean living room temperature [1]. Fig. 8. Time spent at given temperature ranges in summer months. Table 3 Mean summer temperature during time periods

Nearly four out of five houses (78%) spend more than half of the day between 20 and 25 8C. Of the other houses (22%), over half of them (13%) spend more than half the day below 20 8C. However, 1% spend over 50% (4 h per day) of the summer day above 25 8C. On the basis of the selected comfort range 1% can be considered to be at uncomfortably high temperatures for over half the day.

Mean temperatures for all houses

Living room (8C) Bedroom (8C) Ambient (8C)

Morning 7–9 a.m.

Day 9 a.m. to 5 p.m.

Evening 5–11 p.m.

Night midnight to 7 a.m.

19.2 19.1 15.8

21.8 21.2 20.1

23.1 22.6 17.9

20.3 20.1 14.5

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

775

Fig. 9. Mean winter evening temperatures by regional council.

3.1. Climate Fig. 9 shows the mean evening living room temperatures and the ambient evening temperature by region from north to south. The houses are grouped by Regional Council, or groups of these Councils when there are small numbers of monitored houses in their regions. Fig. 9 shows there are opposing trends in the indoor and outdoor winter evening temperatures from north to south, although the indoor trend is not straightforward. There are statistically significant differences between the regions, but these are not only related to the climate. Northland (sub-tropical climate) has a lower median evening temperature than the Otago/Southland area, which in turn has a heating season of over 8 months and a lower mean ambient temperature, as shown by the black diamonds on Fig. 9. Houses in the north are heated for a much shorter time than those in the south. They also generally Table 4 Heating energy by regional council—north to south Regional council

Northland Auckland BOP Waikato Gisborne/Hawkes Bay Taranaki/Manawatu–Wanganui Wellington Tasman/Nelson/Marlborough Canterbury Otago/Southland

Net heating energy (kWh)

Sample count

Mean

S.E.

745 2637 2882 4105 2657 4840 2204 2866 2859 4867

153 273 515 590 470 2202 324 578 383 607

28 96 24 46 24 7 34 17 28 27

have less efficient (e.g. open fires) and less powerful heaters. Table 4 and Fig. 10 show that heating energy varies by location, with the houses in the colder climates not surprisingly using more energy. The colder climates also have a wider range in energy use over the houses, as shown by the spread in Fig. 10. The net heating energy includes electricity, natural gas, LPG and solid fuel. The mean space heating energy use is approximately 2800 kWh for 6 of the 10 regional councils. The differences can be explained for: Northland by its warmer climate; the small number (7) of houses monitored in the Taranaki/Manawatu– Wanganui region; Waikato by its colder winter continental climate; and Otago/Southland as the coldest region in the country. 3.2. Heating fuel and heater type The type of heating system is a major driver of the mean living room evening temperature. Often the appliance selection of the original house owner or builder determines the heating fuel, which then becomes a fixed feature for many years. Table 5 shows for each heating fuel type the percentage of time the average winter evening living room spends below 16 8C, in the range of 16–20 8C and above 20 8C. The heating system may be unit heaters (e.g. unflued cabinet LPG heater) or whole-house central heating (e.g. natural gas ducted-air central heating). Table 5 shows that houses heated by solid fuel burners are the warmest for the longest, with 36% of the time above 20 8C. Houses heated by free-standing, portable LPG heaters or portable electric heaters are the coldest, being above 20 8C only 13% or 16% of the time, respectively.

776

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

Fig. 10. Heating energy by regional council—north to south.

The cost of providing heat varies depending on the fuel and appliance:  a plug-in electric heater (capital cost about NZ$75 (US$50)1) costs 15.4c (US10c) per kWh of heat delivered, while a heat pump (capital cost about NZ$2000 (US$1200)) delivers warmth at 5.1 (US3c) per kWh. Electricity ‘lines charges’ add 93c (US60c) per day;  natural gas delivers heat at about 8.6c (US6c) per kWh, to which is added a daily ‘pipe’ charge of 112.3c (US72c);  a reasonably efficient wood burner delivers heat at about 10c (US6c) per kWh. These prices are for Wellington and, apart from the high daily charge for gas, are about average for the country. The wood burner is based on 71% efficiency, heat pump 300%, night store and plug-in heater 100%, and flued natural gas 80% [17]. Although the costs of the different fuels may be relevant, the ‘size’ of the heater appears to be of greater importance. Solid fuel burners are capable of producing large amounts of heat output, although they can be difficult to control. Under ideal conditions, typical solid fuel burner heat output can range from 4 to 25 kW. HEEP houses ran their solid fuel burners between 3 and 5 kW which, even though at the lower end of possible output, would explain the high numbers of solid fuel heated houses spending time above 20 8C. The highest living room winter temperature measured in a HEEP house was 42 8C – which is warmer than any temperature reached during summer – and this house was heated by a solid fuel burner. 1

Capital costs from www.lvmartin.co.nz.

Just under one in five houses (18.5%) reached maximum temperatures above 30 8C in winter—and 81% of these had enclosed solid fuel burners. Almost half the houses (44.5%) reached maximum winter evening temperatures above 25 8C. Fig. 11 illustrates that houses heated with gas or solid fuel are significantly warmer than electric and LPG-heated houses. Note that ‘Natural Gas’ includes reticulated gas and the large home gas (45 kg LPG) cylinders. The LPG in the figures and tables represents the LPG portable heaters, generally with a 9 kg gas bottle. There is a link between the heater type and the winter temperatures. Houses in locations that do not require much heating are more likely to have less powerful heaters (e.g. open fires, electric) compared with houses in the colder climates that are more likely to have solid fuel burners. Gas central heating is limited to North Island cities that have reticulated gas or access to large 45 kg LPG bottles in the rest of the country. Only 14% of New Zealand houses use reticulated natural gas for space heating [18]. Table 6 shows the main heater type with the average winter evening living room temperature. Living rooms heated by open solid fuel burners are coolest, with an average temperature of 16 8C, while those heated by enclosed solid fuel burners are the Table 5 Living room winter evening temperature distribution Heater fuel

<16 8C (%)

S.E. (%)

16–20 8C (%)

S.E. (%)

>20 8C (%)

S.E. (%)

Count

LPG Electricity Natural gas Solid fuel All houses NA

34 33 22 23 28 34

3 3 5 2

53 51 51 41 47 46

3 2 4 2

13 16 27 36 25 19

2 2 5 2

54 103 35 151 328 39

4

3

4

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

Fig. 11. Mean living room winter evening temperature by heating fuel.

warmest, with an average temperature of 18.8 8C. The main heater is occupant reported, but generally it is the heater that is used for the most hours per week and most often is in the living room. Fig. 12 shows the distribution of total household space heating energy by main heater fuel. Table 6 lists heating energy and temperature by the main heater type. Note the heating energy is that used by all heaters in the house, e.g. the main heater may be portable LPG, but the total ‘net energy’ also includes any electric or other heaters. Heater energy use was not able to be extracted for all the monitored houses, and in some cases a small sample with high variability results in a wide distribution as indicated by the S.E. Table 6 shows that in general the use of heaters with higher outputs, and hence higher energy use, leads to higher temperatures. The small sample size for heat pumps and central heating systems limits the conclusions that can be drawn for these types.

777

Fig. 12. Heating energy by main heater fuel.

statistical significance ( p-value 0.00004). This is without considering retrofitting of thermal insulation, the heating fuel, region or occupants’ heating patterns. The housing stock in the Otago/Southland area is older with only 11% of houses being post-1978, and hence insulated from new (see Table 1). The national average is 25% of houses built post-1978. The older housing stock, along with climate, could help explain the low winter temperatures for some of the houses in Otago/Southland (see Fig. 10 and Table 4). Changes to house design and construction may also have led to higher winter temperatures, and these are discussed later in this paper. 4. Influences on summer indoor temperatures The main drivers of summer living room daytime temperature have been found to be the climate and the house age. 4.1. Climate and regional differences

3.3. House age There is a strong relationship between house age and the living room winter evening temperature. Fig. 13 shows a steady decrease in temperature with increased age of houses, i.e. older houses tend to be colder. There is an average rate of fall 0.20  0.05 8C per decade. This result has a very high

Fig. 14 shows the distribution of the mean daytime (9 a.m. to 5 p.m.) temperatures over the summer months for HEEP houses throughout New Zealand. The graph is ordered from the north to the south (left to right); this shows how the warmer climate in the north affects the interior temperature compared to the colder southern climate.

Table 6 Winter living room evening temperatures and heating energy by heater type Main heater type

Mean evening temperature (8C)

S.E. (8C)

Sample count

Net energy (kWh)

S.E. (kWh)

Sample count

Open solid fuel Electric LPG Fixed electric Heat pump Gas Gas central Solid or liquid central Enclosed solid fuel

16.0 16.9 17.0 17.8 18.0 18.1 18.3 18.5 18.8

0.6 0.3 0.2 0.3 0.4 0.5 0.6 0.7 0.2

11 83 54 18 4 28 8 2 142

973 2175 1640 3563 2057 3836 6727 2908 3803

270 241 217 747 1941 550 1936 NA 324

11 76 41 14 3 23 7 1 118

778

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

Fig. 13. Winter evening living room temperatures by year built.

Fig. 14 also gives the mean ambient daytime temperatures— small black squares. It is clear that the warmer the climate, the warmer the living room temperature. For example, the median living room daytime temperature in Northland is 22.5 8C (ambient average 20.7 8C) compared to 19.5 8C (ambient average 16 8C) in Otago/Southland a 3 8C difference in the living room and 4.7 8C in the ambient. The mean of the daytime living room temperatures shown in Fig. 14 ranges from about 20 8C to about 23 8C, apart from the Otago/Southland region where it is 16 8C. Analysis of the data shows that for each increase of 1 8C for the mean external temperature, the mean daytime house

temperature increases by 0.8 8C. The average external temperature is calculated using NIWA CLIDB [19] temperatures for the year the house was monitored. Using climate the ambient temperature alone accounts for 68% of the variance (r2 = 0.68, p-value = 0.0000) in the living room day time temperature. 4.2. House age Fig. 15 shows that newer houses are warmer in summer than older houses. This difference is statistically significant ( pvalue = 0.0000). Note that the ‘Decade House Built’ is the

Fig. 14. Mean summer living room daytime temperatures by regional council.

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

779

Fig. 15. Summer temperatures by house age.

reported decade of original construction, and that many older houses have been significantly modified, including new heating systems and thermal insulation. The mean summer living room temperatures show a trend of increasing by 0.25 8C per decade. This gives a difference of 2.5 8C between houses built at the beginning and the end of the 20th century. The dotted lines in Fig. 15 at 20 8C and 25 8C and mark the selected comfort range, although comfort does depend on more than just temperature and occupants in these houses may be comfortable outside of this range. Apart from the pre-1910 houses (few of which will be in original condition), the mean temperatures for all house ages are within this range. Houses built from 1990 onwards all have a mean daytime living room summer temperature of above 20 8C, but the average temperature in this group is close to 23 8C with extreme means above 25 8C. It is only houses built in the last three decades that have mean day temperatures above 25 8C. The house age and the average external temperature account for 69% (r2 = 0.69) of the summer day temperature variations ( p-value = 0). Using these two variables (house age and external mean temperature) for other times of the day (e.g. morning, evening and night) explains 60–69% of the variation, increasing to 74% for a 24 h mean temperature. The house age alone, without the average external temperature, explains 14% of the variation in daytime living room temperatures. Separate testing has found that the house age and climate are not dependent variables. One issue not explored here, but of concern, is the possible impact of higher summer temperatures due to either climate variability or climate change. As the newer houses tend already to be warmer than the older houses, their adaptation mechanisms to increased temperatures are potentially more problematic. Heat pumps, which can also be used as air

conditioners, are becoming more and more popular, with one supplier reporting increases in sales of up to 35% a year [20]. If they are used to reduce high summer temperatures, this will have undesired impacts on the electricity system and future greenhouse gas emissions. 5. Why are newer houses warmer? This paper has shown that newer houses are warmer in both winter and summer. There are a number of possible reasons such as:  higher thermal insulation: since 1978 all new houses have had to be insulated;  airtightness: newer houses are less ‘leaky’ due to tighter construction and materials;  increased glazing area: a trend to increased use of glass resulting in greater solar gains but higher winter heat losses;  possibly better orientation of windows for passive solar heating (although no clear indication of this can be found in the HEEP sample);  lower ceiling heights leading to lower room volumes;  reduced or no eaves: due to architectural trends these can result in increased solar gains. Using the HEEP sample, some of these options were explored to examine their impact on temperatures. 5.1. Impact of thermal insulation—winter Houses built after 1 April 1978 (see Table 1) were required to include a minimum level of insulation, but the retrofitting of thermal insulation was not required in older houses. As seen in Table 7, there is a 1.0 8C difference in living room evening

780

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

Table 7 Winter temperatures by assumed insulation level House age group

Average winter evening living room (8C)

S.E. (8C)

Sample count

Bedroom overnight (8C)

S.E. (8C)

Sample count

Pre-1978 Post-1978

17.6 18.6

0.1 0.2

265 99

13.2 14.5

0.1 0.2

243 95

temperatures between pre- and post-1978 houses. This pattern is still seen when the houses are separated by region. This pattern can also be seen in overnight bedroom temperatures, although bedrooms are seldom heated. 5.2. Impact of thermal insulation—summer The actual thermal performance of house components (roof, wall, floor, windows) was not measured. It can, however, be assumed that post-1978 houses are likely to have a higher thermal performance than pre-1978 houses. The difference between pre- and post-1978 houses is significant ( pvalue = 0.0004) for the summer daytime temperatures. Although only 5% of the variation in the temperatures is explained from pre- and post-1978, when including climate, 50% of the variation ( p = 0.000) in daytime living room temperatures is explained. This is less than the 69% explained by house age and climate, which suggests that there is more than just the difference in the levels of thermal insulation in preand post-1978 houses that affect the summer living room daytime temperatures. Orme and Parker [21] found that highly insulated UK buildings had a high chance of summertime over-heating if measures are not taken to control the solar and internal gains. Orme and Parker suggested measures including night cooling, external shading, ventilation, reducing internal gains and providing thermal mass. Although they suggested and explored (by modelling) methods of reducing the solar and internal gains, they could not conclude that insulation was the cause of the high temperatures in summer. 6. Discussion New Zealand is considered by its inhabitants as a temperate country, requiring only limited winter heating and even less summer cooling. This paper has explored some of the temperature and energy data collected as part of the HEEP project. It has attempted to quantify the importance of changed NZ Building Code thermal insulation requirements, and identify other important drivers of temperature and energy use. 6.1. Summer As few New Zealand houses are cooled (air conditioned) during the summer, this represents a large sample of naturally ventilated houses, with the ventilation controlled by the occupants’ use of windows and doors. Four percent of HEEP houses had the ability to cool by an air conditioner or a reverse

cycle heat pump. Unfortunately HEEP did not monitor the occupants’ use of windows and doors during the project, as it would be interesting to see the effect of the changing ventilation rates in the house as well as the occupant’s response to temperature. The majority of houses (80%) spend less than one quarter (25%—2 h) of the summer daytime with living room temperatures over 25 8C. Most living rooms are between 20 and 25 8C for the majority of the time. As there is no analysis of ‘comfort’ temperatures for New Zealand, it can only be concluded that, based on overseas norms, these would appear to be comfortable. On average, bedroom temperatures are lower than living room temperatures—by as little as 0.1 8C in the morning (7– 9 a.m.) and as much as 0.6 8C during the day (9 a.m. to 5 p.m.). Inside temperatures have a smaller temperature range than the ambient, showing the temperature stabilising benefit of even low thermal mass construction. As few New Zealand houses have high thermal mass, and with only two in the HEEP sample, it was not possible to see the effect of a high thermal mass building in the New Zealand climate. The house age (represented by decade of construction) and the local climate (the average external temperature over summer) have the largest impact on the summer daytime living room temperatures. Together they explain 69% of the variation in mean summer living room day temperatures. The newer houses are showing a trend of being warmer during the summer with the mean living room temperature increasing by 0.25 8C per decade. Houses built today are 2.5 8C warmer than those built 100 years ago in the same climate. Selected reasons for newer houses being warmer have been explored. The influence of occupant self-reported house airtightness has been found to be marginal, as has the presence of thermal insulation. No obvious relationship has been found between large areas of solar gain (west, north and east facing glazing) and high temperatures. Occupant influence also looks to be significant, but has not been quantified. Preliminary thermal modelling work shows that houses behave differently without occupant influences, e.g. opening and closing windows. Additional simulation models are being explored. Although climate change is not a focus of this paper, it should be noted that the local climate clearly influences the interior temperature. New houses are already warmer than older houses, so a 2–3 8C temperature rise, possibly due to climate change, could make many of the newer houses uncomfortably warm, unless adaptation takes place. There is the danger that the occupants of these newer houses could become reliant on air conditioning, with the resulting higher energy use forming a

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

positive feedback loop into the mechanism of climate change. This is clearly an undesirable result. 6.2. Winter The average New Zealand house has a mean living room temperature below that recommended for comfort and in some cases health. The average winter evening temperature is 17.9 8C, while the mean range is from 10 to 23.8 8C. Only 5% of New Zealand houses have central heating systems. In the other houses, the tendency is to zone heat, with the most common room heated being the living room and the most common time of heating being the winter evening. Solid fuel burners heat the houses well but with little control—they can reach very high temperatures. Houses heated by open fires (solid fuel) and portable electric heaters are the coolest, with mean living room evening temperatures of 16 and 16.9 8C, respectively. Houses heated by enclosed solid fuel burners are the warmest, with a mean living room evening temperature of 18.8 8C. The loggers used to measure the temperature in these houses measure both a combination of air temperature and radiant temperature, although the loggers are placed away from heat sources where possible. It is possible that occupants are placing themselves very close to radiant heat sources. In that case they will be warmer (at least one side of them!) than the temperature our loggers are measuring and we are reporting. Newer houses are warmer during winter than older houses; reasons for this may include higher levels of thermal insulation, increased airtightness, and more efficient or powerful heating systems. Comparing pre- and post-1978 houses, the winter evening living room temperatures in the newer houses are on average 1 8C warmer—1978 is when the first mandatory regulations were introduced for insulation in houses. This temperature difference increases to 1.3 8C in the bedrooms, which seldom have formal heating appliances—the main heating sources are human bodies, televisions, clock radios and pets. 7. Conclusions The work presented in this paper is based on a comprehensive, national sample of houses throughout the main populated climate zones of New Zealand. Although the apparent benefits of thermal insulation on winter indoor temperatures were expected, the results for summer temperatures were not expected. This has opened a door for additional work to explore the other potential drivers of increased indoor temperatures, which may have significance in dealing with adaptation to climate change. The influence of the occupants on the temperatures, and the effect of ventilation, shading, amount of glazing and construction of the house are all areas which will be investigated further. Work has also focused on daytime temperature, and further examination of evening and night temperatures will be done. Of particular importance is the increasing use of winter ‘heat pumps’ which hold the potential to be used as summer ‘air conditioners’. Fifty four thousand

781

single phase air conditioners (most of which are heat pumps) were sold in New Zealand during 2005. Data for 2006 is incomplete at this time but sales were over 70,000 [22]. These in turn will have an unknown impact on the national electricity system, with further possible implications for climate change. It is not known whether the comparatively low winter temperatures are meeting New Zealand households’ comfort expectations. International comfort research would tend to suggest that these are below acceptable comfort temperatures, but additional comfort research is required to quantify any potential future energy demands. In particular, if central heating systems achieve a higher market penetration, they may create an unsustainable load on the national electricity system. Opportunities for fuel shifting may also need to be explored to ensure that future winter heating needs can be met. Acknowledgements We would like to acknowledge the support and assistance of the various HEEP field staff, contractors and house occupants. The research and data collection has financial support from a number of organisations—Foundation for Research, Science and Technology, Building Research and the Building Research Levy, ECCA, Transpower and a series of other Government and industry supporters. References [1] N.P. Isaacs, M. Camilleri, L. French, A. Pollard, K. Saville-Smith, R. Fraser, P. Rossouw, Energy Use in New Zealand Households: Report on the Year 9 Analysis for the Household Energy End-use Project (HEEP), BRANZ Ltd. Study Report 141, Judgeford, Porirua, 2005. [2] N.P. Isaacs, L. Amitrano, M. Camilleri, A. Pollard, A. Stoecklein, Energy Use in New Zealand Households: Report on the Year 6 Analysis for the Household Energy End-use Project (HEEP), BRANZ Ltd. Study Report 115, Judgeford, Porirua, 2002. [3] A.R. Pollard, Investigating the Characterisation of Temperatures within New Zealand Buildings. Masters Thesis, Massey University, Palmerston North, 2001. [4] J.F. Nicol, M.A. Humphreys, Adaptive thermal comfort and sustainable thermal standards for buildings, Energy and Buildings 34 (6) (2002) 563– 572. [5] G.S. Brager, R.J. de Dear, Thermal adaptation in the built environment: a literature review, Energy and Buildings 27 (1) (1998) 83–96. [6] M.T. Camilleri, Implications of Climate Change for the Construction Sector: Houses, BRANZ Ltd. Study Report 94, Judgeford, Porirua, 2000. [7] R.A. Jaques, Summertime Overheating in New Zealand Houses—Influences, Risks and Mitigation Strategies, BRANZ Ltd. Study Report 89, Judgeford, Porirua, 2000. [8] M. Donn, G. Thomas, Designing Comfortable Homes, Cement and Concrete Association of NZ and EECA, Wellington, 2001. [9] Statistics NZ, Profile of New Zealand (www.stats.govt.nz/quick-facts/ default.htm accessed 9 January 2006), 2005. [10] Meteorological Service of New Zealand Ltd. (MetService), The Climate of New Zealand (www.metservice.co.nz/default/index.php?alias=learningcentre accessed 9 February 2006). [11] Standards New Zealand, SNZ PAS 4244:2003—Insulation of Lightweight-Framed and Solid-Timber Houses, Wellington, New Zealand, 2003. [12] N.P. Isaacs, Thermal Efficiency in New Zealand Buildings: A Historical Overview, Centre for the Building Performance Research, Victoria University of Wellington, 1993.

782

L.J. French et al. / Energy and Buildings 39 (2007) 770–782

[13] N.P. Isaacs, Performance based building energy efficiency code, Proceedings of the Global Building Model in the Next Millennium Conference, Melbourne April 12–15 1999 (BRANZ Conference Paper 63: 108–118), 1999. [14] L.J. French, M.T. Camilleri, N.P. Isaacs, A.R. Pollard, Winter Temperatures in New Zealand Houses, Paper presented to 2006 Windsor Conference: Comfort and Energy Use in Buildings—Getting Them Right, April 2006, 27–30 [15] N.P. Isaacs, L. Amitrano, M. Camilleri, L. French, A. Pollard, K. SavilleSmith, R. Fraser, P. Rossouw, Energy Use in New Zealand Households: Report on the Year 8 Analysis for the Household Energy End-use Project (HEEP), BRANZ Ltd. Study Report 133, Judgeford, Porirua, 2004. [16] Statistics NZ, Household Economic Survey (Year ended 30 June 2004). Wellington: Statistics NZ. (www.statistics.govt.nz accessed 9 January 2006), 2004.

[17] Consumer magazine, Shivering Timbers 447, 2005, 38–40. [18] Statistics NZ, 2001 Census of Population and Dwellings—Housing, Statistics New Zealand, Wellington, New Zealand, 2002. [19] A.C. Penny, Climate Database (CLIDB) User’s Manual 5th ed., NIWA, Wellington (www.niwascience.co.nz/services/clidb/ accessed 9 January 2006). [20] G. Ninness, Demand Running Hot for Cool Comfort, Sunday Star Times, 12 February 2006, page D5, 2006. [21] M. Orme, J. Palmer, Control of Overheating in Future Housing—Design Guidance for Low Energy Strategies, Faber Maunsell Ltd., Birmingham UK (copy obtained electronically www.fabermaunsell.com/media/ 4171.pdf). 2003. [22] Energy Efficiency and Conservation Authority (EECA), Single Phase Air Conditioner Sales, Received 3 October 2006, Products Team, EECA, Wellington.