- Email: [email protected]
Reliability analysis of rainwater tanks: A comparison between South-East and Central Melbourne
Resources, Conservation and Recycling 66 (2012) 1–7
Contents lists available at SciVerse ScienceDirect
Resources, Conservation and Recycling journal homepage: www.elsevier.com/locate/resconrec
Reliability analysis of rainwater tanks: A comparison between South-East and Central Melbourne Monzur Alam Imteaz a,∗ , Ataur Rahman b , Amimul Ahsan c a b c
Faculty of Engineering and Industrial Sciences, Swinburne University of Technology, Melbourne, Australia School of Computing, Engineering and Mathematics, University of Western Sydney, NSW, Australia Green Engineering & Sustainable Technology Lab, Institute of Advanced Technology, University Putra Malaysia (UPM), Malaysia
a r t i c l e
i n f o
Article history: Received 10 November 2011 Received in revised form 8 March 2012 Accepted 25 May 2012 Keywords: Rainwater tank Daily water balance Climatic conditions Reliability Threshold tank size
a b s t r a c t A spreadsheet based daily water balance model is used for the performance analysis and design optimisation of rainwater tanks at two different regions of Melbourne, South-East and Central Melbourne. These two different regions of Melbourne are characterised by notable different topography and rainfall characteristics. From historical rainfall data, three representative years (dry, average and wet) are selected. Reliability is defined as percentage of days in a year when rainwater tank is able to supply the intended partial demand for a particular condition. For the three climatic conditions, a number of reliability charts are produced for domestic rainwater tanks in relations to tank volume, roof area, number of people in a house (i.e. water demand) and percentage of total water demand to be satisfied by the harvested rainwater. It is found that for a relatively small roof size (100 m2 ), 100% reliability cannot be achieved even with a very large tank (10,000 L). Reliability becomes independent of tank size for tank sizes larger than 5000–7000 L depending on the location. This is defined as threshold tank size, relationships with threshold tank sizes and annual rainfall amounts are then established for both the locations. It is found that for a particular annual rainfall amount, threshold tank size for South-East Melbourne is higher than that of Central Melbourne. However, for a relatively large roof size (200 m2 ), approximately 90% reliability can be achieved with a tank size of 10,000 L and with further increase in tank size, a 100% reliability is achievable, except in a dry year. Furthermore, it is found that reliabilities for South-East Melbourne maintain consistent higher values as compared to Central Melbourne. © 2012 Elsevier B.V. All rights reserved.
1. Introduction With increasing population and changing climate regime, water supply systems in many cities of the world are under stress. To tackle this problem, water authorities are adopting several measures including demand management and identifying alternative water sources such as stormwater harvesting, greywater and wastewater reuse and desalination. Among all the alternative water sources, stormwater harvesting perhaps has received the highest level of attention. In Australia, government authorities have been promoting stormwater harvesting through campaigns, as well as offering incentives and grants to promote water saving ideas and innovations. Among all the stormwater harvesting options, rainwater tanks have been most widely studied. Fewkes (1999) conducted studies on residential rainwater tanks in the United Kingdom, producing a series of dimensionless design curves which allows estimation of
∗ Corresponding author. E-mail address: [email protected] (M.A. Imteaz). 0921-3449/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.resconrec.2012.05.009
the rainwater tank size required to obtain a desired performance measure given the roof area and water demand patterns. Vaes and Berlamont (2001) developed a model to determine the effectiveness of rainwater tanks and stormwater runoff using long term historical rainfall data. Coombes and Kuczera (2003) found that for an individual building with a 150 m2 roof area and 1–5 kL tank in Sydney can yield 10–58% mains water savings (depending on the number of people using the building). According to Coombes and Kuczera (2003), depending on roof area and number of occupants, rainwater tank use can result in mains water annual savings of 18–55 kL for 1 kL sized tanks and 25–144 kL for 10 kL sized tanks. In Sweden, Villarreal and Dixon (2005) investigated water savings potential of stormwater harvesting systems from roof areas. Villarreal and Dixon (2005) noted that a mains water saving of 30% can be achieved using a 40 m3 sized tank (toilet and washing machine use only). Ghisi et al. (2007, 2009) investigated the water savings potential from rainwater harvesting systems in Brazil and found that average potential for potable water savings to be 12–79% per year for the cities analysed. Muthukumaran et al. (2011) found that use of rainwater inside a home in regional Victoria in Australia can save up to 40% of potable water use. Farreny et al. (2011)
2
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
examined the quantity and quality of rainwater harvesting in Spain and found that sloping smooth roofs may harvest up to about 50% more rainwater than flat rough roofs. Mun and Han (2012) developed a design and evaluation method for a rainwater harvesting system on the basis of water balance equation and found that a design based on sensitivity analysis and proper management of a rainwater harvesting system should be emphasized to improve the operation efficiency. Imteaz et al. (2012) presented rainwater savings potentials and reliability of rainwater tanks in dry years for Southwest Nigeria. Coombes (2007) conducted studies on the modelling of the rainwater tanks and the opportunities for effective retention storage using the PURRS (Probabilistic Urban Rainwater and Wastewater Reuse Simulator) water balance model. Following over a decade of research into the quality of rainwater collected from roofs, Coombes (2007) identified the potential for rainwater to be utilised far more extensively than many government regulators do recommend. Despite positive outcome from many studies, there remains a general community reluctance to adopt stormwater harvesting including rainwater tanks on a wider scale. Part of the reason for this reluctance can be attributed to lack of information about the effectiveness of a stormwater harvesting system and the optimum storage size required to satisfy the performance requirements under the specific site conditions (Imteaz et al., 2011b). A proper in-depth understanding of the effectiveness of any proposed onsite stormwater harvesting system is often lacking. The predicted change in rainfall patterns in Australia as a result of global warming adds further complexity to planning adequate rainwater harvesting schemes. Furthermore, many studies have used mean annual rainfall data or generated rainfall data in modelling rainwater harvesting system. In an area of highly inter-annual rainfall variability, analysis considering long term mean annual rainfall may not be useful. Eroksuz and Rahman (2010) investigated the water savings potential of rainwater tanks in multi-storied residential buildings for three cities in eastern Australia. They concluded that rainwater tank of appropriate size in a multi-storied building can provide significant water savings even in dry years. They also developed equations for predicting annual rainwater savings potentials for those cities. Khastagir and Jayasuriya (2010) analysed reliability of rainwater tanks and calculated the reliability using a daily water balance model. They produced contours of optimum tank sizes for surrounding areas of Melbourne, considering the historical daily rainfall, the demand for rainwater and the roof area for a supply reliability of 90%. All of these analyses were based on historical daily rainfall data, making an average of cumulative historical savings and other model variables. Through such analysis of averaged variables/parameters rainwater tank users may not get an adequate insight of the expected realistic situation(s) in regards to variability and outcomes. With the impacts of climate change, such ranges of realistic outcomes are expected to be widen further. Furthermore, design charts (contours) presented by Khastagir and Jayasuriya (2010) are based on an expected reliability of 90%. In reality, different user may opt for different reliability. Imteaz et al. (2011a) developed a daily water balance model for the optimisation of rainwater tank size. This model considered daily rainfall, losses due to leakage, spillage and evaporation, roof area, tank volume, rainwater demand, overflow losses and tank top-up requirements in case of shortage. Using the developed model, they presented a range of reliability charts for a central Melbourne location in relations to tank volume, roof area, total water demand and rainwater demand for three different climatic conditions (driest, average and wettest years). However, as the driest and wettest years are very rare events, this paper presents a more realistic representative dry and wet
years. Through statistical analysis of historical annual rainfall data, dry year is defined as having an annual rainfall value closer to 10th percentile value of historical annual rainfall value, and wet year is defined as having an annual rainfall value closer to 90th percentile value. Reliability charts as a function of tank size are produced for dry, average and wet years under different demand scenarios. Furthermore, unlike traditional analysis where a single average rainfall for a large city is widely adopted, this study assesses the impact of spatial variability of rainfall on the reliability of rainwater tanks for Melbourne, the second largest city of Australia. Melbourne has considerable rainfall gradient where rainfall shows remarkable spatial variability. 2. Methodology A spreadsheet based daily water balance model was developed considering daily rainfall, contributing catchment (roof) area, losses due to leakage, spillage and evaporation, storage (tank) volume and water uses. In this model, the primary input value was the daily rainfall amount for three different years (dry, average and wet years). Historical daily rainfall data was collected from Bureau of Meteorology, Australia. For central Melbourne and South-East Melbourne, the selected rainfall stations were located at Scotch College and Scoresby Research Institute, respectively. These two stations are located 19 km apart. The daily runoff volume was calculated from daily rainfall amount by multiplying the rainfall amount with the contributing roof area and deducting the losses. For this study, from the produced runoff, a 10% deduction was applied to account for several losses (leakage, spilling and evaporation). Generated runoff was diverted to the connected available storage tank. Available storage capacity was compared with the accumulated daily runoff. If the accumulated runoff was bigger than available storage volume, excess water (overflow) was deducted from the accumulated runoff. Amount of water use(s) is deducted from the daily accumulated/stored runoff amount, if sufficient amount of water is available in the storage. In a situation, when sufficient amount of water is not available in the storage, model assumes that the remaining water demand is met from the town water supply. The model calculates daily stormwater use, daily water storage in the tank, daily overflow and daily town water use. In addition, model calculates accumulated annual rainwater use, accumulated annual overflow and accumulated annual town water use. The overall procedure can be mathematically described as below: Cumulative water storage equation: St = Vt + St−1 − D
(1)
St = 0, for St < 0
(2)
St = C, for St > C
(3)
where St is the cumulative water stored in the rainwater tank (L) after the end of tth day; VT is the harvested rainwater (L) on the tth day; St−1 is the storage in the tank (L) at the beginning of tth day; D is the daily rainwater demand (L); and C is the capacity of rainwater tank (L). Townwater use equation: TW = D − St , for St < D
(4)
where TW is the townwater use on tth day (L). Overflow equation: OF = St−C , for St > C
(5)
where OF is the overflow on tth day (L). Reliability is calculated with the equation: Re =
N−U × 100 N
(6)
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7 Table 1 Physical features of the study locations. South-East Melbourne
37◦ 50 2.2 S 145◦ 1 45.2 E 18 m Flat
37◦ 53 55.3 S 145◦ 13 49.5 E 77 m Hilly
Dry year 60% demand
Table 2 Historical rainfall characteristics of the study locations. Central Melbourne Mean (mm) Median (mm) Standard deviation Maximum (mm) Minimum (mm)
South-East Melbourne
657 648 144 914 387
Dry year 80% demand
80
Reliability %
Latitude Longitude Average elevation Topography
2 People, Roof: 100m2
100
Central Melbourne
3
60
Average year 80% demand
40
Average year 60% demand
20
Wet year 80% demand
0
Wet year 60% demand
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Tank Size (L)
862 847 160 1237 513
Fig. 1. Reliability curves for a two-people household with a roof size of 100 m2 for South-East Melbourne.
where Re is the reliability of the tank to be able to supply intended demand (%); U is the number of days in a year the tank was unable to meet the demand; and N is the total number of days in a particular year. 3. Data Physical features of the two selected regions are presented in Table 1, which shows South-East Melbourne is located at higher elevation than the Central Melbourne. For central Melbourne and South-East Melbourne daily rainfall data of 41 years (969–2009) and 62 years (1948–2009) were available, respectively. Table 2 summarises the important rainfall statistics for the two study locations, which shows that South-East Melbourne has much higher rainfall than Central Melbourne. Through statistical analysis of total annual rainfall data, three separate years were selected as dry year, average year and wet year. Selected years and corresponding annual rainfall values are shown in Table 3. Developed daily water balance model was simulated with the daily rainfall data for the above-mentioned years for several different tank sizes (1000–10,000 L) and percentage of total water demand to be fulfilled by rainwater tank. To assess the effect of roof area, two roof areas were considered: 100 m2 and 200 m2 . In regards to number of occupants (i.e. total water demand), two scenarios were examined: 2 people and 4 people. An average nonpotable water demand of 185 L/person/day was considered based on an earlier study by Imteaz et al. (2011b). In regards to per cent of total water demand to be fulfilled by rainwater tank, two demand scenarios (60% and 80%) were examined. Several relationship curves are presented to show the effects of different parameters on reliability of rainwater tanks under different climatic conditions.
of 100 m2 for south-eastern Melbourne. Fig. 2 shows the reliability curves as a function of different tank sizes under same climatic, roof size and demand conditions for central Melbourne. From both the figures it is clear that 100% reliability is not achievable with a roof area of 100 m2 , which is quite small to collect an adequate amount of rainfall. In this case, the collected rainfall will be utilised too quickly and the tank will remain empty until the arrival of a new rainfall event. For larger tanks, if the roof area does not increase, reliability would not increase as in that case the tank will remain partially or full empty most of the time as the roof area is too small to catch enough water and/or demand is too high to utilise the water too quickly. With the current conditions, maximum achievable reliability is almost same for both the Central and South-East Melbourne. However, with higher water demand, maximum reliability (in a wet year) for South-East Melbourne (∼75%) is higher than the maximum reliability of Central Melbourne (∼60%). Also, in a dry year, maximum reliabilities in South-East Melbourne (36% and 56%) is much higher than the maximum reliabilities in Central Melbourne (22% and 38%). For South-East Melbourne, it is found that reliabilities cannot be increased further with the increase in tank size for a tank size larger than 5000 L. Whereas, for the same conditions in Central Melbourne, reliabilities can be increased with the increase in tank size for a tank size of up to 7000 L. However, in a dry year in Central Melbourne, reliability remains the same for a tank size larger than 3000 L. In such situations, reliability may be increased with the increase in roof size or decrease in demand. To visualize distinction between reliabilities in Central and South-East Melbourne, Fig. 3a–c shows reliability chart for a twopeople household connected with a roof area of 100 m2 under different demands (80% and 60%) for a dry year, average year and wet year respectively. From the figures, it is found that for both 2 People, Roof: 100m2
4. Results
100
Table 3 Selected rainfalls and corresponding years for the dry, average and wet years. Central Melbourne
Dry year (10 percentile) Average year (50 percentile) Wet year (90 percentile)
Dry year 80% demand
80
Reliability %
Fig. 1 shows the reliability curves as a function of different tank sizes under different demand scenarios (60% and 80%) for a two-people household (i.e. total non-potable water demand of 370 L/day) for a dry year and wet year connected with a roof size
Dry year 60% demand
60
Average year 80% demand Average year 60% demand
40
Wet year 80% demand
20
Wet year 60% demand
South-East Melbourne
Rainfall (mm)
Year
Rainfall (mm)
Year
465 654 858
2007 1983 1974
588 835 1081
1967 1983 1992
0 1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank Size (L) Fig. 2. Reliability curves for a two-people household with a roof size of 100 m2 for Central Melbourne.
4
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
Dry Year, 2 people, Roof: 100m2 60
Reliability (%)
50 40 30 20
Southeast - 80% demand Southeast - 60% demand
10
Central - 80% demand Central - 60% demand
0 1000
2000
3000
4000
5000
6000
7000
8000
900010000
Tank size (L) Average Year, 2 people, Roof: 100m2
90 80
Reliability (%)
70 60 50 40 Southeast - 80% demand
30
Southeast - 60% demand
20
Central - 80% demand 10 0
Central - 60% demand 1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Wet Year, 2 people, Roof: 100m2 100 90
Reliability (%)
80 70 60 50 Southeast - 80% demand
40
Southeast - 60% demand
30
Central - 80% demand
20
Central - 60% demand
10 0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Fig. 3. Comparison of reliabilities for a two-people household connected with a roof area of 100 m2 within south-east and central Melbourne.
the dry year and average year, reliabilities are very similar for 60% demand in Central Melbourne and 80% demand in South-East Melbourne. This reflects the fact that South-East Melbourne gets much higher rainfall than the Central Melbourne. Also, it is found that in a dry year, reliabilities remain same for a tank size larger than 5000 L. However, for an average year and dry year, this threshold tank size is 6000 L and 8000 L, respectively. With the same conditions having four-people household scenario, reliabilities decrease significantly as shown in Fig. 4a–c. From the figures, it is found that maximum reliabilities which can be achieved (for a 60% demand in South-East Melbourne) are 17%, 32% and 50% in a dry year, average year and wet year, respectively. Whereas, for a two-people household scenario, maximum reliabilities of 57%, 78% and 88% can be achieved (for a 60% demand in South-East Melbourne) in
a dry year, average year and wet year, respectively. These figures reveal a drastic change in maximum achievable reliabilities due to change in number of people in a household. Furthermore, it is found that threshold tank sizes (maximum tank size above which reliabilities remain same) are 4000 L, 5000 L and 7000 L in a dry year, average year and wet year respectively. In all the above mentioned scenarios, it can be concluded that reliabilities are limited by either climate condition or roof size. Fig. 5 shows relationships between threshold tank sizes and annual total rainfall amounts. From this figure it is evident that a linear relationship exists with the annual rainfall amount and threshold tank size, i.e. with the increase in annual rainfall, threshold tank size also increases linearly. Also, within these relationship graphs, there are consistent gaps between two-people demand and four-people demand
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
5
Dry Year, 4 people, Roof: 100m2
20 18
Reliability (%)
16 14 12 10 8 Southeast - 80% demand
6
Southeast - 60% demand
4
Central - 80% demand
2
Central - 60% demand
0 1000
2000
3000
4000
5000
6000
7000
8000
900010000
Tank size (L) Average Year, 4 people, Roof: 100m2 35
Reliability (%)
30 25 20 15 Southeast - 80% demand 10
Southeast - 60% demand Central - 80% demand
5
Central - 60% demand 0 1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Wet Year, 4 people, Roof: 100m2 60
Reliability (%)
50 40 30 20
Southeast - 80% demand Southeast - 60% demand
10
Central - 80% demand Central - 60% demand
0 1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Fig. 4. Comparison of reliabilities for a four-people household connected with a roof area of 100 m2 within south-east and central Melbourne.
scenarios. From these plots, threshold tank size for an intermediate demand and for any other annual rainfall amount can be estimated. Also, it is found that lines for threshold tank sizes for Central Melbourne is shifted (and slightly tilted) upward compared to the lines for South-East Melbourne. This reveals that for a same annual rainfall, threshold tank size for South-East Melbourne would be higher than that of entral Melbourne. As the roof area has a significant influence on the expected reliabilities, same model simulations were performed with a larger roof area (200 m2 ) for a four-people household scenario. Fig. 6a–c shows reliability charts under different demands (80% and 60%) for a dry year, average year and wet year, respectively. From these figures, it is found that reliabilities keep on increasing even with a tank size of up to 10,000 L, except for a dry year in Central Melbourne. For Central Melbourne, in a dry year threshold tank size is 5000 L. Also, it is found that for a four-people household scenario, maximum reliabilities of 57%, 78% and 88% can be achieved (for a 60% demand in South-East Melbourne) in a dry year, average
Fig. 5. Relationships between threshold tank sizes and annual total rainfall amounts.
6
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
Dry Year, 4 people, Roof: 200m2 60
Reliability (%)
50 40 30 20
Southeast - 80% demand Southeast - 60% demand
10
Central - 80% demand Central - 60% demand
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Average Year, 4 people, Roof: 200m2 80
Reliability (%)
60
40 Southeast - 80% demand Southeast - 60% demand
20
Central - 80% demand Central - 60% demand 0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Wet Year, 4 people, Roof: 200m2 100
Reliability (%)
80
60
Southeast - 80% demand
40
Southeast - 60% demand Central - 80% demand
20
Central - 60% demand 0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Fig. 6. Comparison of reliabilities for a four-people household connected with a roof area of 200 m2 within south-east and central Melbourne.
year and wet year respectively. These maximum achievable reliabilities are exactly same for a scenario with a half roof size (100 m2 ) and half the demand (2 people). However, the patterns are not same; for a 2 people-100 m2 roof size scenario, reliabilities become independent of the tank size for a tank size larger than threshold tank size. Whereas, for a 4 people, 200 m2 roof size scenario in most cases, reliabilities keep on increasing even with a tank size of 10,000 L, which indicates that in some cases a reliability of 100% may be achieved with a tank size larger than 10,000 L. It should be noted here that reliability is not the only guiding factor in selecting a tank size as all the houses in a big city like Melbourne have townwater supply and hence factors such as tank
space and cost would also influence the selection of a particular tank size. Reliability becomes more important when townwater supply is not available and when a city experiences long drought and struggles to cope with limited water availability and high water demand. 5. Summary and conclusions This paper investigates reliability of rainwater tanks at two different regions (Central and South-East) of Melbourne, under different scenarios i.e. climatic conditions, roof areas, tank volumes, household water demands and portion of total demand to be fulfilled by rainwater. It is found that for a relatively small roof
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
size (100 m2 ), 100% reliability cannot be achieved even with a very large tank (10,000 L). Reliability becomes independent of tank size for tank sizes larger than 5000–7000 L depending on geographical location within Melbourne. This is defined as threshold tank size. With the increased rainwater demand and same smaller tank size, reliability drops down to significant low levels (about 32%). Also, it is found that reliabilities for South-East Melbourne are much higher than reliabilities for Central Melbourne and South-East Melbourne maintains consistent higher reliability values for most of the cases (except in a wet year having higher demands). However, for a relatively large roof size (200 m2 ), approximately 90% reliability can be achieved with a tank size of 10,000 L and it is clear that with further increase in tank size, a 100% reliability is achievable, except in a dry year. Similar to the case with smaller roof size, it is found that reliabilities for South-East Melbourne maintains consistent higher values compared to the reliabilities for Central Melbourne. In general, to achieve a 100% reliability seems to be difficult, because of the fact that a high amount of non-potable water demand (185 L/person/day) has been assumed for this study. In reality, in many parts of the world, the non-potable water demand would be much lower. To identify a threshold tank size for a particular region, which might be the ultimate interest of the end users, relationships among threshold tank sizes and annual rainfall amounts are produced. A linear relationship exists with the annual rainfall amount and threshold tank size, i.e. with the increase in annual rainfall, threshold tank size also increases linearly. Also, within these relationship graphs, there are consistent gaps between two-people demand and four-people demand scenarios. From these fitted lines, threshold tank size for an intermediate demand and for any other annual rainfall amount can be estimated. Also, it is found that the fitted lines for threshold tank sizes for Central Melbourne is shifted upward compared to that of the South-East Melbourne. This reveals that for the same annual rainfall, threshold tank size for South-East Melbourne would be higher than that of Central Melbourne. The considered rainfall stations (which are located 19 km apart) produced significantly different results due to the difference in topographic conditions and rainfall variability. The results reveal that rainwater tank reliability can vary significantly within a large city like Melbourne and emphasises the need to change the design practice of considering a single annual rainfall value for the purpose of rainwater tank sizing.
7
The results outlined in this study would vary with geographical locations i.e. with different climatic conditions or in general with different rainfall intensities and pattern. In reality, there are numerous optimal solutions with different combinations of storage volumes, roof sizes, rainwater demand and number of people in the household. A comprehensive decision support tool would be able to produce reliability, cumulative water savings, cumulative overflow losses and cumulative townwater supply used for any geographical location. References Coombes P. Rainwater research 2007. www.rainwaterharvesting.com.au [viewed 13.03.10]. Coombes P, Kuczera G. Analysis of the performance of rainwater tanks in Australian capital cities. In: 28th International Hydrology and Water Resources Symposium; 2003. Eroksuz E, Rahman A. Rainwater tanks in multi-unit buildings: a case study for three Australian cities. Resources, Conservation and Recycling 2010;54: 1449–52. Farreny R, Morales-Pinzon T, Guisasola A, Taya C, Rieradevall J, Gabarrell X. Roof selection for rainwater harvesting: quantity and quality assessments in Spain. Water Research 2011;45:3245–54. Fewkes A. The use of rainwater for WC flushing: the field testing of a collection system. Building and Environment 1999;34(6):765–72. Ghisi E, Bressan DL, Martini M. Rainwater tank capacity and potential for potable water savings by using rainwater in the residential sector of southeastern Brazil. Building and Environment 2007;42:1654–66. Ghisi E, Tavares DF, Rocha VL. Rainwater harvesting in petrol stations in Brasilia: potential for potable water savings and investment feasibility analysis. Resources, Conservation and Recycling 2009;54:79–85. Imteaz MA, Adeboye O, Rayburg S, Shanableh A. Rainwater harvesting potential for southwest Nigeria using daily water balance model. Resources, Conservation and Recycling 2012;62:51–5. Imteaz MA, Ahsan A, Naser J, Rahman A. Reliability analysis of rainwater tanks in Melbourne using daily water balance model. Resources, Conservation and Recycling 2011a;56:80–6. Imteaz MA, Shanableh A, Rahman A, Ahsan A. Optimisation of rainwater tank design from large roofs: a case study in Melbourne, Australia. Resources, Conservation and Recycling 2011b;55:1022–9. Khastagir A, Jayasuriya N. Optimal sizing of rain water tanks for domestic water conservation. Journal of Hydrology 2010;381:181–8. Mun JS, Han MY. Design and operational parameters of a rooftop rainwater harvesting system: definition, sensitivity and verification. Journal of Environmental Management 2012;93:147–53. Muthukumaran S, Baskaran K, Sexton N. Quantification of potable water savings by residential water conservation and reuse – a case study. Resources, Conservation and Recycling 2011;55:945–52. Vaes G, Berlamont J. The effect of rainwater storage tank on design storms. Urban Water 2001;3:303–7. Villarreal EL, Dixon A. Analysis of a rainwater collection system for domestic water supply in Ringdansen, Norrkoping, Sweden. Building and Environment 2005;40:1174–84.
Contents lists available at SciVerse ScienceDirect
Resources, Conservation and Recycling journal homepage: www.elsevier.com/locate/resconrec
Reliability analysis of rainwater tanks: A comparison between South-East and Central Melbourne Monzur Alam Imteaz a,∗ , Ataur Rahman b , Amimul Ahsan c a b c
Faculty of Engineering and Industrial Sciences, Swinburne University of Technology, Melbourne, Australia School of Computing, Engineering and Mathematics, University of Western Sydney, NSW, Australia Green Engineering & Sustainable Technology Lab, Institute of Advanced Technology, University Putra Malaysia (UPM), Malaysia
a r t i c l e
i n f o
Article history: Received 10 November 2011 Received in revised form 8 March 2012 Accepted 25 May 2012 Keywords: Rainwater tank Daily water balance Climatic conditions Reliability Threshold tank size
a b s t r a c t A spreadsheet based daily water balance model is used for the performance analysis and design optimisation of rainwater tanks at two different regions of Melbourne, South-East and Central Melbourne. These two different regions of Melbourne are characterised by notable different topography and rainfall characteristics. From historical rainfall data, three representative years (dry, average and wet) are selected. Reliability is defined as percentage of days in a year when rainwater tank is able to supply the intended partial demand for a particular condition. For the three climatic conditions, a number of reliability charts are produced for domestic rainwater tanks in relations to tank volume, roof area, number of people in a house (i.e. water demand) and percentage of total water demand to be satisfied by the harvested rainwater. It is found that for a relatively small roof size (100 m2 ), 100% reliability cannot be achieved even with a very large tank (10,000 L). Reliability becomes independent of tank size for tank sizes larger than 5000–7000 L depending on the location. This is defined as threshold tank size, relationships with threshold tank sizes and annual rainfall amounts are then established for both the locations. It is found that for a particular annual rainfall amount, threshold tank size for South-East Melbourne is higher than that of Central Melbourne. However, for a relatively large roof size (200 m2 ), approximately 90% reliability can be achieved with a tank size of 10,000 L and with further increase in tank size, a 100% reliability is achievable, except in a dry year. Furthermore, it is found that reliabilities for South-East Melbourne maintain consistent higher values as compared to Central Melbourne. © 2012 Elsevier B.V. All rights reserved.
1. Introduction With increasing population and changing climate regime, water supply systems in many cities of the world are under stress. To tackle this problem, water authorities are adopting several measures including demand management and identifying alternative water sources such as stormwater harvesting, greywater and wastewater reuse and desalination. Among all the alternative water sources, stormwater harvesting perhaps has received the highest level of attention. In Australia, government authorities have been promoting stormwater harvesting through campaigns, as well as offering incentives and grants to promote water saving ideas and innovations. Among all the stormwater harvesting options, rainwater tanks have been most widely studied. Fewkes (1999) conducted studies on residential rainwater tanks in the United Kingdom, producing a series of dimensionless design curves which allows estimation of
∗ Corresponding author. E-mail address: [email protected] (M.A. Imteaz). 0921-3449/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.resconrec.2012.05.009
the rainwater tank size required to obtain a desired performance measure given the roof area and water demand patterns. Vaes and Berlamont (2001) developed a model to determine the effectiveness of rainwater tanks and stormwater runoff using long term historical rainfall data. Coombes and Kuczera (2003) found that for an individual building with a 150 m2 roof area and 1–5 kL tank in Sydney can yield 10–58% mains water savings (depending on the number of people using the building). According to Coombes and Kuczera (2003), depending on roof area and number of occupants, rainwater tank use can result in mains water annual savings of 18–55 kL for 1 kL sized tanks and 25–144 kL for 10 kL sized tanks. In Sweden, Villarreal and Dixon (2005) investigated water savings potential of stormwater harvesting systems from roof areas. Villarreal and Dixon (2005) noted that a mains water saving of 30% can be achieved using a 40 m3 sized tank (toilet and washing machine use only). Ghisi et al. (2007, 2009) investigated the water savings potential from rainwater harvesting systems in Brazil and found that average potential for potable water savings to be 12–79% per year for the cities analysed. Muthukumaran et al. (2011) found that use of rainwater inside a home in regional Victoria in Australia can save up to 40% of potable water use. Farreny et al. (2011)
2
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
examined the quantity and quality of rainwater harvesting in Spain and found that sloping smooth roofs may harvest up to about 50% more rainwater than flat rough roofs. Mun and Han (2012) developed a design and evaluation method for a rainwater harvesting system on the basis of water balance equation and found that a design based on sensitivity analysis and proper management of a rainwater harvesting system should be emphasized to improve the operation efficiency. Imteaz et al. (2012) presented rainwater savings potentials and reliability of rainwater tanks in dry years for Southwest Nigeria. Coombes (2007) conducted studies on the modelling of the rainwater tanks and the opportunities for effective retention storage using the PURRS (Probabilistic Urban Rainwater and Wastewater Reuse Simulator) water balance model. Following over a decade of research into the quality of rainwater collected from roofs, Coombes (2007) identified the potential for rainwater to be utilised far more extensively than many government regulators do recommend. Despite positive outcome from many studies, there remains a general community reluctance to adopt stormwater harvesting including rainwater tanks on a wider scale. Part of the reason for this reluctance can be attributed to lack of information about the effectiveness of a stormwater harvesting system and the optimum storage size required to satisfy the performance requirements under the specific site conditions (Imteaz et al., 2011b). A proper in-depth understanding of the effectiveness of any proposed onsite stormwater harvesting system is often lacking. The predicted change in rainfall patterns in Australia as a result of global warming adds further complexity to planning adequate rainwater harvesting schemes. Furthermore, many studies have used mean annual rainfall data or generated rainfall data in modelling rainwater harvesting system. In an area of highly inter-annual rainfall variability, analysis considering long term mean annual rainfall may not be useful. Eroksuz and Rahman (2010) investigated the water savings potential of rainwater tanks in multi-storied residential buildings for three cities in eastern Australia. They concluded that rainwater tank of appropriate size in a multi-storied building can provide significant water savings even in dry years. They also developed equations for predicting annual rainwater savings potentials for those cities. Khastagir and Jayasuriya (2010) analysed reliability of rainwater tanks and calculated the reliability using a daily water balance model. They produced contours of optimum tank sizes for surrounding areas of Melbourne, considering the historical daily rainfall, the demand for rainwater and the roof area for a supply reliability of 90%. All of these analyses were based on historical daily rainfall data, making an average of cumulative historical savings and other model variables. Through such analysis of averaged variables/parameters rainwater tank users may not get an adequate insight of the expected realistic situation(s) in regards to variability and outcomes. With the impacts of climate change, such ranges of realistic outcomes are expected to be widen further. Furthermore, design charts (contours) presented by Khastagir and Jayasuriya (2010) are based on an expected reliability of 90%. In reality, different user may opt for different reliability. Imteaz et al. (2011a) developed a daily water balance model for the optimisation of rainwater tank size. This model considered daily rainfall, losses due to leakage, spillage and evaporation, roof area, tank volume, rainwater demand, overflow losses and tank top-up requirements in case of shortage. Using the developed model, they presented a range of reliability charts for a central Melbourne location in relations to tank volume, roof area, total water demand and rainwater demand for three different climatic conditions (driest, average and wettest years). However, as the driest and wettest years are very rare events, this paper presents a more realistic representative dry and wet
years. Through statistical analysis of historical annual rainfall data, dry year is defined as having an annual rainfall value closer to 10th percentile value of historical annual rainfall value, and wet year is defined as having an annual rainfall value closer to 90th percentile value. Reliability charts as a function of tank size are produced for dry, average and wet years under different demand scenarios. Furthermore, unlike traditional analysis where a single average rainfall for a large city is widely adopted, this study assesses the impact of spatial variability of rainfall on the reliability of rainwater tanks for Melbourne, the second largest city of Australia. Melbourne has considerable rainfall gradient where rainfall shows remarkable spatial variability. 2. Methodology A spreadsheet based daily water balance model was developed considering daily rainfall, contributing catchment (roof) area, losses due to leakage, spillage and evaporation, storage (tank) volume and water uses. In this model, the primary input value was the daily rainfall amount for three different years (dry, average and wet years). Historical daily rainfall data was collected from Bureau of Meteorology, Australia. For central Melbourne and South-East Melbourne, the selected rainfall stations were located at Scotch College and Scoresby Research Institute, respectively. These two stations are located 19 km apart. The daily runoff volume was calculated from daily rainfall amount by multiplying the rainfall amount with the contributing roof area and deducting the losses. For this study, from the produced runoff, a 10% deduction was applied to account for several losses (leakage, spilling and evaporation). Generated runoff was diverted to the connected available storage tank. Available storage capacity was compared with the accumulated daily runoff. If the accumulated runoff was bigger than available storage volume, excess water (overflow) was deducted from the accumulated runoff. Amount of water use(s) is deducted from the daily accumulated/stored runoff amount, if sufficient amount of water is available in the storage. In a situation, when sufficient amount of water is not available in the storage, model assumes that the remaining water demand is met from the town water supply. The model calculates daily stormwater use, daily water storage in the tank, daily overflow and daily town water use. In addition, model calculates accumulated annual rainwater use, accumulated annual overflow and accumulated annual town water use. The overall procedure can be mathematically described as below: Cumulative water storage equation: St = Vt + St−1 − D
(1)
St = 0, for St < 0
(2)
St = C, for St > C
(3)
where St is the cumulative water stored in the rainwater tank (L) after the end of tth day; VT is the harvested rainwater (L) on the tth day; St−1 is the storage in the tank (L) at the beginning of tth day; D is the daily rainwater demand (L); and C is the capacity of rainwater tank (L). Townwater use equation: TW = D − St , for St < D
(4)
where TW is the townwater use on tth day (L). Overflow equation: OF = St−C , for St > C
(5)
where OF is the overflow on tth day (L). Reliability is calculated with the equation: Re =
N−U × 100 N
(6)
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7 Table 1 Physical features of the study locations. South-East Melbourne
37◦ 50 2.2 S 145◦ 1 45.2 E 18 m Flat
37◦ 53 55.3 S 145◦ 13 49.5 E 77 m Hilly
Dry year 60% demand
Table 2 Historical rainfall characteristics of the study locations. Central Melbourne Mean (mm) Median (mm) Standard deviation Maximum (mm) Minimum (mm)
South-East Melbourne
657 648 144 914 387
Dry year 80% demand
80
Reliability %
Latitude Longitude Average elevation Topography
2 People, Roof: 100m2
100
Central Melbourne
3
60
Average year 80% demand
40
Average year 60% demand
20
Wet year 80% demand
0
Wet year 60% demand
1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Tank Size (L)
862 847 160 1237 513
Fig. 1. Reliability curves for a two-people household with a roof size of 100 m2 for South-East Melbourne.
where Re is the reliability of the tank to be able to supply intended demand (%); U is the number of days in a year the tank was unable to meet the demand; and N is the total number of days in a particular year. 3. Data Physical features of the two selected regions are presented in Table 1, which shows South-East Melbourne is located at higher elevation than the Central Melbourne. For central Melbourne and South-East Melbourne daily rainfall data of 41 years (969–2009) and 62 years (1948–2009) were available, respectively. Table 2 summarises the important rainfall statistics for the two study locations, which shows that South-East Melbourne has much higher rainfall than Central Melbourne. Through statistical analysis of total annual rainfall data, three separate years were selected as dry year, average year and wet year. Selected years and corresponding annual rainfall values are shown in Table 3. Developed daily water balance model was simulated with the daily rainfall data for the above-mentioned years for several different tank sizes (1000–10,000 L) and percentage of total water demand to be fulfilled by rainwater tank. To assess the effect of roof area, two roof areas were considered: 100 m2 and 200 m2 . In regards to number of occupants (i.e. total water demand), two scenarios were examined: 2 people and 4 people. An average nonpotable water demand of 185 L/person/day was considered based on an earlier study by Imteaz et al. (2011b). In regards to per cent of total water demand to be fulfilled by rainwater tank, two demand scenarios (60% and 80%) were examined. Several relationship curves are presented to show the effects of different parameters on reliability of rainwater tanks under different climatic conditions.
of 100 m2 for south-eastern Melbourne. Fig. 2 shows the reliability curves as a function of different tank sizes under same climatic, roof size and demand conditions for central Melbourne. From both the figures it is clear that 100% reliability is not achievable with a roof area of 100 m2 , which is quite small to collect an adequate amount of rainfall. In this case, the collected rainfall will be utilised too quickly and the tank will remain empty until the arrival of a new rainfall event. For larger tanks, if the roof area does not increase, reliability would not increase as in that case the tank will remain partially or full empty most of the time as the roof area is too small to catch enough water and/or demand is too high to utilise the water too quickly. With the current conditions, maximum achievable reliability is almost same for both the Central and South-East Melbourne. However, with higher water demand, maximum reliability (in a wet year) for South-East Melbourne (∼75%) is higher than the maximum reliability of Central Melbourne (∼60%). Also, in a dry year, maximum reliabilities in South-East Melbourne (36% and 56%) is much higher than the maximum reliabilities in Central Melbourne (22% and 38%). For South-East Melbourne, it is found that reliabilities cannot be increased further with the increase in tank size for a tank size larger than 5000 L. Whereas, for the same conditions in Central Melbourne, reliabilities can be increased with the increase in tank size for a tank size of up to 7000 L. However, in a dry year in Central Melbourne, reliability remains the same for a tank size larger than 3000 L. In such situations, reliability may be increased with the increase in roof size or decrease in demand. To visualize distinction between reliabilities in Central and South-East Melbourne, Fig. 3a–c shows reliability chart for a twopeople household connected with a roof area of 100 m2 under different demands (80% and 60%) for a dry year, average year and wet year respectively. From the figures, it is found that for both 2 People, Roof: 100m2
4. Results
100
Table 3 Selected rainfalls and corresponding years for the dry, average and wet years. Central Melbourne
Dry year (10 percentile) Average year (50 percentile) Wet year (90 percentile)
Dry year 80% demand
80
Reliability %
Fig. 1 shows the reliability curves as a function of different tank sizes under different demand scenarios (60% and 80%) for a two-people household (i.e. total non-potable water demand of 370 L/day) for a dry year and wet year connected with a roof size
Dry year 60% demand
60
Average year 80% demand Average year 60% demand
40
Wet year 80% demand
20
Wet year 60% demand
South-East Melbourne
Rainfall (mm)
Year
Rainfall (mm)
Year
465 654 858
2007 1983 1974
588 835 1081
1967 1983 1992
0 1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank Size (L) Fig. 2. Reliability curves for a two-people household with a roof size of 100 m2 for Central Melbourne.
4
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
Dry Year, 2 people, Roof: 100m2 60
Reliability (%)
50 40 30 20
Southeast - 80% demand Southeast - 60% demand
10
Central - 80% demand Central - 60% demand
0 1000
2000
3000
4000
5000
6000
7000
8000
900010000
Tank size (L) Average Year, 2 people, Roof: 100m2
90 80
Reliability (%)
70 60 50 40 Southeast - 80% demand
30
Southeast - 60% demand
20
Central - 80% demand 10 0
Central - 60% demand 1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Wet Year, 2 people, Roof: 100m2 100 90
Reliability (%)
80 70 60 50 Southeast - 80% demand
40
Southeast - 60% demand
30
Central - 80% demand
20
Central - 60% demand
10 0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Fig. 3. Comparison of reliabilities for a two-people household connected with a roof area of 100 m2 within south-east and central Melbourne.
the dry year and average year, reliabilities are very similar for 60% demand in Central Melbourne and 80% demand in South-East Melbourne. This reflects the fact that South-East Melbourne gets much higher rainfall than the Central Melbourne. Also, it is found that in a dry year, reliabilities remain same for a tank size larger than 5000 L. However, for an average year and dry year, this threshold tank size is 6000 L and 8000 L, respectively. With the same conditions having four-people household scenario, reliabilities decrease significantly as shown in Fig. 4a–c. From the figures, it is found that maximum reliabilities which can be achieved (for a 60% demand in South-East Melbourne) are 17%, 32% and 50% in a dry year, average year and wet year, respectively. Whereas, for a two-people household scenario, maximum reliabilities of 57%, 78% and 88% can be achieved (for a 60% demand in South-East Melbourne) in
a dry year, average year and wet year, respectively. These figures reveal a drastic change in maximum achievable reliabilities due to change in number of people in a household. Furthermore, it is found that threshold tank sizes (maximum tank size above which reliabilities remain same) are 4000 L, 5000 L and 7000 L in a dry year, average year and wet year respectively. In all the above mentioned scenarios, it can be concluded that reliabilities are limited by either climate condition or roof size. Fig. 5 shows relationships between threshold tank sizes and annual total rainfall amounts. From this figure it is evident that a linear relationship exists with the annual rainfall amount and threshold tank size, i.e. with the increase in annual rainfall, threshold tank size also increases linearly. Also, within these relationship graphs, there are consistent gaps between two-people demand and four-people demand
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
5
Dry Year, 4 people, Roof: 100m2
20 18
Reliability (%)
16 14 12 10 8 Southeast - 80% demand
6
Southeast - 60% demand
4
Central - 80% demand
2
Central - 60% demand
0 1000
2000
3000
4000
5000
6000
7000
8000
900010000
Tank size (L) Average Year, 4 people, Roof: 100m2 35
Reliability (%)
30 25 20 15 Southeast - 80% demand 10
Southeast - 60% demand Central - 80% demand
5
Central - 60% demand 0 1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Wet Year, 4 people, Roof: 100m2 60
Reliability (%)
50 40 30 20
Southeast - 80% demand Southeast - 60% demand
10
Central - 80% demand Central - 60% demand
0 1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Fig. 4. Comparison of reliabilities for a four-people household connected with a roof area of 100 m2 within south-east and central Melbourne.
scenarios. From these plots, threshold tank size for an intermediate demand and for any other annual rainfall amount can be estimated. Also, it is found that lines for threshold tank sizes for Central Melbourne is shifted (and slightly tilted) upward compared to the lines for South-East Melbourne. This reveals that for a same annual rainfall, threshold tank size for South-East Melbourne would be higher than that of entral Melbourne. As the roof area has a significant influence on the expected reliabilities, same model simulations were performed with a larger roof area (200 m2 ) for a four-people household scenario. Fig. 6a–c shows reliability charts under different demands (80% and 60%) for a dry year, average year and wet year, respectively. From these figures, it is found that reliabilities keep on increasing even with a tank size of up to 10,000 L, except for a dry year in Central Melbourne. For Central Melbourne, in a dry year threshold tank size is 5000 L. Also, it is found that for a four-people household scenario, maximum reliabilities of 57%, 78% and 88% can be achieved (for a 60% demand in South-East Melbourne) in a dry year, average
Fig. 5. Relationships between threshold tank sizes and annual total rainfall amounts.
6
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
Dry Year, 4 people, Roof: 200m2 60
Reliability (%)
50 40 30 20
Southeast - 80% demand Southeast - 60% demand
10
Central - 80% demand Central - 60% demand
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Average Year, 4 people, Roof: 200m2 80
Reliability (%)
60
40 Southeast - 80% demand Southeast - 60% demand
20
Central - 80% demand Central - 60% demand 0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Wet Year, 4 people, Roof: 200m2 100
Reliability (%)
80
60
Southeast - 80% demand
40
Southeast - 60% demand Central - 80% demand
20
Central - 60% demand 0
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Tank size (L) Fig. 6. Comparison of reliabilities for a four-people household connected with a roof area of 200 m2 within south-east and central Melbourne.
year and wet year respectively. These maximum achievable reliabilities are exactly same for a scenario with a half roof size (100 m2 ) and half the demand (2 people). However, the patterns are not same; for a 2 people-100 m2 roof size scenario, reliabilities become independent of the tank size for a tank size larger than threshold tank size. Whereas, for a 4 people, 200 m2 roof size scenario in most cases, reliabilities keep on increasing even with a tank size of 10,000 L, which indicates that in some cases a reliability of 100% may be achieved with a tank size larger than 10,000 L. It should be noted here that reliability is not the only guiding factor in selecting a tank size as all the houses in a big city like Melbourne have townwater supply and hence factors such as tank
space and cost would also influence the selection of a particular tank size. Reliability becomes more important when townwater supply is not available and when a city experiences long drought and struggles to cope with limited water availability and high water demand. 5. Summary and conclusions This paper investigates reliability of rainwater tanks at two different regions (Central and South-East) of Melbourne, under different scenarios i.e. climatic conditions, roof areas, tank volumes, household water demands and portion of total demand to be fulfilled by rainwater. It is found that for a relatively small roof
M.A. Imteaz et al. / Resources, Conservation and Recycling 66 (2012) 1–7
size (100 m2 ), 100% reliability cannot be achieved even with a very large tank (10,000 L). Reliability becomes independent of tank size for tank sizes larger than 5000–7000 L depending on geographical location within Melbourne. This is defined as threshold tank size. With the increased rainwater demand and same smaller tank size, reliability drops down to significant low levels (about 32%). Also, it is found that reliabilities for South-East Melbourne are much higher than reliabilities for Central Melbourne and South-East Melbourne maintains consistent higher reliability values for most of the cases (except in a wet year having higher demands). However, for a relatively large roof size (200 m2 ), approximately 90% reliability can be achieved with a tank size of 10,000 L and it is clear that with further increase in tank size, a 100% reliability is achievable, except in a dry year. Similar to the case with smaller roof size, it is found that reliabilities for South-East Melbourne maintains consistent higher values compared to the reliabilities for Central Melbourne. In general, to achieve a 100% reliability seems to be difficult, because of the fact that a high amount of non-potable water demand (185 L/person/day) has been assumed for this study. In reality, in many parts of the world, the non-potable water demand would be much lower. To identify a threshold tank size for a particular region, which might be the ultimate interest of the end users, relationships among threshold tank sizes and annual rainfall amounts are produced. A linear relationship exists with the annual rainfall amount and threshold tank size, i.e. with the increase in annual rainfall, threshold tank size also increases linearly. Also, within these relationship graphs, there are consistent gaps between two-people demand and four-people demand scenarios. From these fitted lines, threshold tank size for an intermediate demand and for any other annual rainfall amount can be estimated. Also, it is found that the fitted lines for threshold tank sizes for Central Melbourne is shifted upward compared to that of the South-East Melbourne. This reveals that for the same annual rainfall, threshold tank size for South-East Melbourne would be higher than that of Central Melbourne. The considered rainfall stations (which are located 19 km apart) produced significantly different results due to the difference in topographic conditions and rainfall variability. The results reveal that rainwater tank reliability can vary significantly within a large city like Melbourne and emphasises the need to change the design practice of considering a single annual rainfall value for the purpose of rainwater tank sizing.
7
The results outlined in this study would vary with geographical locations i.e. with different climatic conditions or in general with different rainfall intensities and pattern. In reality, there are numerous optimal solutions with different combinations of storage volumes, roof sizes, rainwater demand and number of people in the household. A comprehensive decision support tool would be able to produce reliability, cumulative water savings, cumulative overflow losses and cumulative townwater supply used for any geographical location. References Coombes P. Rainwater research 2007. www.rainwaterharvesting.com.au [viewed 13.03.10]. Coombes P, Kuczera G. Analysis of the performance of rainwater tanks in Australian capital cities. In: 28th International Hydrology and Water Resources Symposium; 2003. Eroksuz E, Rahman A. Rainwater tanks in multi-unit buildings: a case study for three Australian cities. Resources, Conservation and Recycling 2010;54: 1449–52. Farreny R, Morales-Pinzon T, Guisasola A, Taya C, Rieradevall J, Gabarrell X. Roof selection for rainwater harvesting: quantity and quality assessments in Spain. Water Research 2011;45:3245–54. Fewkes A. The use of rainwater for WC flushing: the field testing of a collection system. Building and Environment 1999;34(6):765–72. Ghisi E, Bressan DL, Martini M. Rainwater tank capacity and potential for potable water savings by using rainwater in the residential sector of southeastern Brazil. Building and Environment 2007;42:1654–66. Ghisi E, Tavares DF, Rocha VL. Rainwater harvesting in petrol stations in Brasilia: potential for potable water savings and investment feasibility analysis. Resources, Conservation and Recycling 2009;54:79–85. Imteaz MA, Adeboye O, Rayburg S, Shanableh A. Rainwater harvesting potential for southwest Nigeria using daily water balance model. Resources, Conservation and Recycling 2012;62:51–5. Imteaz MA, Ahsan A, Naser J, Rahman A. Reliability analysis of rainwater tanks in Melbourne using daily water balance model. Resources, Conservation and Recycling 2011a;56:80–6. Imteaz MA, Shanableh A, Rahman A, Ahsan A. Optimisation of rainwater tank design from large roofs: a case study in Melbourne, Australia. Resources, Conservation and Recycling 2011b;55:1022–9. Khastagir A, Jayasuriya N. Optimal sizing of rain water tanks for domestic water conservation. Journal of Hydrology 2010;381:181–8. Mun JS, Han MY. Design and operational parameters of a rooftop rainwater harvesting system: definition, sensitivity and verification. Journal of Environmental Management 2012;93:147–53. Muthukumaran S, Baskaran K, Sexton N. Quantification of potable water savings by residential water conservation and reuse – a case study. Resources, Conservation and Recycling 2011;55:945–52. Vaes G, Berlamont J. The effect of rainwater storage tank on design storms. Urban Water 2001;3:303–7. Villarreal EL, Dixon A. Analysis of a rainwater collection system for domestic water supply in Ringdansen, Norrkoping, Sweden. Building and Environment 2005;40:1174–84.