ECONOMIC BENEFITS ARISING FROM USE OF WATER SENSITIVE URBAN DEVELOPMENT SOURCE CONTROL MEASURES.

Peter J. Coombes, George Kuczera and Jetse D. Kalma; Department of Civil, Surveying and Environmental Engineering. University of Newcastle, New South Wales.
ABSTRACT The economic benefits of implementing Water Sensitive Urban Development source control measures are evaluated for two case studies. The Lower Hunter and Central Coast water supply networks in New South Wales, Australia have been investigated using a systems approach for different levels of implementation of source control measures including rainwater tanks. Use of rainwater tanks to supplement mains water supply for toilet, hot water and outdoor uses can reduce annual regional water demand by up to 24,700 ML, delay construction of new water supply headworks infrastructure by up to 38 years and, in some cases, eliminate the need for construction of new water supply infrastructure resulting in economic benefits to the community of up to $6B during the next 100 years. 1. INTRODUCTION

Urban development and its constructed hydraulic systems cause profound changes to the natural water cycle. The area of impervious surfaces is increased whilst natural watercourses are replaced with hydraulically efficient pipes and channels. Water demand resulting from urban development is met by importing large volumes of treated water, across large distances and at considerable cost, from neighboring catchments. At the same time similar volumes of stormwater from roofs are discharged unused from urban developments via expensive stormwater systems. Water Sensitive Urban Development (WSUD) source controls include reuse of rainwater, stormwater and wastewater. Research into WSUD from the urban water cycle management [Andoh and Declerck, 1999 and McAlister, 1998] perspective shows that significant economic, social and environmental benefits to the community may be derived from more efficient use of water resources and infrastructure. However, a major impediment to the use of the WSUD approach is a perception that it is expensive to implement and that it has limited economic benefits. This study considers the benefits of the use of rainwater stored in tanks to supplement domestic toilet flushing, hot water and outdoor water use in the Lower Hunter and Central Coast regions, New South Wales, Australia. It shows how various scenarios that include rainwater tanks influence drought security of the water supply headworks system. The impact on delaying augmentation of the headworks system is examined. A comparative economic analysis is also presented. 2. METHOD

The drought security of the Lower Hunter water supply headworks system was assessed using WATHNET [Kuczera, 1992] for different household demand scenarios. WATHNET is a suite of programs for generalised water supply headworks simulation using network linear programming. Data from the Hunter Water Corporation (HWC), Gosford/Wyong Water Authority, Australian Bureau of Statistics (2000) and from the Figtree Place development was used to develop inhouse and exhouse water use profiles for the regions [Coombes et al., 2000 and 2000a]. Five housing (H1-H5) and two unit cluster scenarios (C1 and C2) (Table 1) were considered for each region. Roof areas were assumed to increase with the number of occupants and tank sizes were estimated as a function of available land area in typical housing allotments. Table 1. Housing scenarios with roof areas
Item Occupants H1 1 H2 2 H3 3 H4 4 H5 5+ C1 11 C2 24 Dwellings Roof area (m 2 ) 1 100 1 135 1 175 1 215 1 250 4 415 9 600 Tank size (kL) 10 10 10 10 10 20 30

The simple model of commercial and industrial water use, and unaccounted-for consumption and losses developed by Kuczera and Ng (1992) was modified for this study to preserve the ratio between domestic and total water use in both regions. This model accounts for about 57% and 38% respectively of total mains water use in the Lower Hunter and Central Coast regions. It is used identically in all the scenarios reported below. A number of demand and streamflow scenarios were simulated using 1000 100-year replicates based on recorded climate sequences and monthly demand generated for each dwelling [Coombes et al., 2000a]. The base scenario considers the status quo. This is provision of traditional stormwater systems to areas undergoing urbanisation and provision of additional mains water supply by further regulation of river systems. Alternative scenarios consider several WSUD approaches that include the use of rainwater tanks. Seven alternative scenarios are considered: In the growth scenario (denoted as G), WSUD source controls (including rainwater tanks, infiltration and detention strategies) are installed for all new housing. In the other six scenarios (denoted as G +0.25% to G +3%), source controls are installed for all new housing and existing housing is retrofitted with water tanks at rates varying from 0.25% to 3% per year until 90% of dwellings have a water tank. The economic efficiency of the WSUD and traditional base scenarios is compared using the methods of annual and present equivalence [Smith, 1979]. All costs or benefits considered are in year 2000 dollars. Although little information is available about the cost of WSUD, this study has used information from demonstration projects, published studies and desktop analysis to estimate costs [Coombes et al., 2000 and 2000a]. In the annual equivalence analysis each alternative scenario starts in year 2000 with enough funds to ensure economic viability of the base scenarios. That is a balance of $17M and $36M respectively for the Lower Hunter and Central Coast regions. Each year expenses are deducted, income is added and interest is earned on the balance. The following assumptions are used: Real interest rate = Commonwealth bonds interest consumer price index = 4.78%; Maintenance/replacement/operation of the rainwater tank and retention system costs $73.30 per year; WSUD in new development areas results in a saving, conservatively set at $959 per dwelling; Cost to install a rainwater tank and retention system is $1500 per house; and A kilolitre of mains water costs $0.92 (Lower Hunter Region) or $0.65 (Central Coast Region). The analysis considers comparative costs and benefits using the base scenario as the reference. For the base scenario, the balance of the initial investment carried forward from year t is Balt+1, is: Balt+1 = (i+1) (Balt - augCostt) (1) where i is the real interest rate and augCostt is the augmentation cost (if any) in year t. For the other scenarios, the balance in year t is Balt+1 = (i+1) (Balt - augCostt - conCostt - mainCostt + waterSavt + savDevt) (2)

where for year t conCostt and mainCostt are the construction, maintenance and operation costs for the rainwater tank and retention system, waterSavt is the savings in foregone mains water consumption, and savDevt is the saving in new stormwater infrastructure from using WSUD. 3. LOWER HUNTER WATER SUPPLY

The Lower Hunter Region has a population of 455,000 people with an overall growth rate of 0.9%. Domestic water demand accounts for approximately 43% of total water demand. The region spans five local government areas, namely Newcastle, Lake Macquarie, Maitland, Cessnock and Port Stephens. The region has been divided into nine zones to facilitate water supply modelling. A schematic of the regions water supply headworks system is presented in Figure 1. The system presently consists of two major surface reservoirs that harvest water from the Williams River catchment and a sub-surface reservoir, the Tomago Sand Beds unconfined aquifer. The Chichester Reservoir has a capacity of 21500 megalitres (ML). Current environmental flow constraints for the Chichester River require all flows below 14 ML/day to be released from the reservoir. A gravity pipeline delivers up to 95 ML/day to the region via the Dungog treatment plant. The Seaham diversion weir is used to divert water from the Williams River into Grahamstown reservoir via Ballickera pumping station and canal. The pumping station has a maximum capacity of 1330 ML/day. At low river flows water quality and environmental constraints limit pumping from Seaham Weir. The peaky nature of high flows in the Williams River also limits the volume of water that can be pumped from Seaham Weir. The final stage of Grahamstown Reservoir with a capacity of 198,200 ML has been used in the study. It is supplied with water via diversions from Seaham Weir and local runoff from a 99 km2 catchment. Water from the reservoir is distributed to the region via the Tomago treatment plant.

Tomago Sand Beds is an extensive unconfined aquifer with a storage capacity of approximately 232,800 ML. Water from the Sand Beds is used to supplement water from the Williams River system. When the combined capacity of Chichester and Grahamstown reservoirs falls below 80% up to 100 ML per day is drawn from the Tomago Sand Beds.
Tillegra flow Proposed Tillegra reservoir Upper Williams river Chichester pipeline Evap Chichester flow Chichester reservoir Chichester river

Residential demand Industrial and commercial demand Port Stephens demand

Proposed duplicate of Chichester pipeline

Evap.

Grahamstown flow

Middle Williams river Seaham diversion weir Lower Williams river

Grahamstown reservoir

Balikerra canal Spill Rainfall recharge

Tomago Sand Beds

Baseflow

Ocean

Figure 1. Schematic of Lower Hunter Region water supply headworks system Reliability of the headworks system is defined as the probability that water restrictions will not be imposed in a particular year. Restrictions on demand are triggered when the combined storage of Chichester and Grahamstown reservoirs falls below 60%. When the reliability of the system falls below 90% current policy requires that the system be augmented. In this study following the recommendation of Hunter District Water Board (1982), augmentation consists of construction of a reservoir on the Williams River at Tillegra. The Tillegra reservoir (Figure 1) is to be constructed in two stages. The augmentation strategy is in two stages: Stage 1/. Construct Tillegra Reservoir (stage 1) with a capacity of 240,000 ML combined with installation of pumps to increase the hydraulic capacity of Chichester Pipeline at a cost of $103.7M. Stage 2/. Construction of a water supply pipeline parallel to the Chichester Pipeline at a cost of $101.7M Table 2 summarises the timing of the water supply augmentation required to maintain reliability at or above 90% for each of the demand scenarios. The symbol NR indicates augmentation was not required. The introduction of rainwater tanks delays augmentation from 8 to 34 years depending on the rate of rainwater tank adoption. Figure 2 displays the growth in annual mains water demand for selected demand scenarios. The results demonstrate the very significant reduction in mains water demand that can be achieved by use of rainwater tanks. By the year 2032 the G+3% scenario has reduced annual mains consumption from 95,800 ML to 77,700 ML, a 19% reduction in demand on the headworks system. For scenarios G+0.75% to G+3% reliability of the water supply headworks system remains at 99% after augmentation stage 1 therefore the need for further augmentation of the system in the twenty-second century is eliminated. Table 2. Years in which augmentation of water supply is required
Augmentation required for scenario (year) Augment stage Base Growth G+0.25% G+0.5% G+0.75% G+0.9% G+2% G+3% 1 2041 2049 2050 2055 2064 2067 2075 2075 2 NR NR NR NR NR NR NR NR

Figure 3 presents a time series of annual equivalent economic position (Balt+1, equations 1 and 2) for the different scenarios. The analysis reveals that the growth scenario provides the greatest benefit to the community. This benefit will be greater than $6B during this century. The present equivalence analysis (Table 3) shows that for the base, G+2% and G+3% scenarios capital investment is required to ensure viability of the strategy. However for the scenarios growth to G+0.9% capital investment is not required and additional funds are generated.

140000 130000

7000

Balance of $17M investment ($M)

Annual (ML/yr)

120000 110000 100000 90000 80000 70000 2000

Base Growth G + 0.9% G + 3%

6000 5000 4000 3000 2000 1000 0 2000 -1000 Base Growth G+0.9% G+3%

2020

2040

2060

2080

2100

2020

2040

2060

2080

2100

Year

Figure 2: Mains water demand with different levels of rainwater tank use
S c e n a rio B a se G ro w th G + 0 .2 5 % G + 0 .5 % G + 0 .7 5 % G + 0 .9 % G + 2% G + 3%

Year

Figure 3: Annual equivalence for the different demand scenarios

In v e s tm e n t ($ M ) 17 -5 0 -3 3 -2 3 -1 3 -6 62 102

Table 3. Present equivalence results for the water supply scenarios

The results show that scenarios using installation rates of greater than 0.9% per year of rainwater tanks to existing dwellings and to all new dwellings were not financially viable when compared to the base scenario. However, scenarios using installation rates of up to 0.9% per year provides greater benefits to the community than the traditional water supply option. 4. CENTRAL COAST WATER SUPPLY

The Central Coast region has a population of 289,100 people with a growth rate of 1.37% that is expected to slow to 0.24% after the year 2030. Domestic water demand accounts for approximately 62% of total water demand. The region includes two local government areas, namely Gosford and Wyong. The region has been divided into two zones to facilitate water supply modelling. A schematic of the Central Coast region water supply headworks system is presented in Figure 4. The system presently includes a major surface reservoir on Mangrove Creek and harvests water from the Mangrove Creek, Mooney Mooney Creek, Ourimbah Creek and Wyong River catchments. Mangrove and Mooney Mooney Creeks discharge to the Hawkesbury River. Ourimbah Creek and Wyong River discharge to the Tuggarah Lakes. Mangrove Dam has a storage capacity of 195,000 ML. Gosford/Wyong demand is supplied by releasing water from Mangrove dam into Mangrove creek that is pumped from Mangrove Weir to the Somersby Transfer system. Water is also pumped from the Mooney Dam on Mooney Mooney Creek via the Somersby Transfer System. Water from the Wyong River and Ourimbah Creek is pumped to the Mardi Transfer system via Mardi dam to also supply Gosford/Wyong demand. Reliability of the headworks system is defined as the probability that storage in Mangrove Dam will not be fall below 20% in a particular year. When the reliability of the system falls below 99% current policy requires that the system be augmented. The augmentation strategy is in two stages: Stage 1: construction of a Wyong River to Mangrove dam transfer system (Figure 4) at a cost of $76.4M, and Stage 2: construction of a McDonald River to Mangrove dam transfer system (Figure 4) and increase the capacity of Mangrove dam by 455,000 ML at a cost of $194.5M. Table 4 summarises the timing of the water supply augmentation required to maintain reliability at or above 99% for each of the demand scenarios. The symbol NR indicates augmentation was not required. Introduction of rainwater tanks delays augmentation stage 1 from 28 to 36 years for the growth to G+0.5% scenarios and the requirement for augmentation stage 1 is eliminated for scenarios greater than G+0.5%. The use of rainwater tanks also eliminates the need for augmentation stage 2.

McDonald River Proposed Weir

Mangrove Creek

Wyong River Proposed Transfer Tunnel

Proposed Upper Wyong Weir

Gosford/Wyong Demand Mardi Transfer Wyong Weir and Pump Stn.

Mangrove Dam Mooney Mooney Creek Somersby Transfer Mardi Dam Mooney Dam and Pump Stn.

Proposed Transfer System Mangrove Weir and Pump stn.

Ourimbah Creek

Tuggarah Lakes Ourimbah Weir and Pump Stn.

Hawkesbury River

Figure 4: Schematic of Central Coast water supply headworks system Figure 6 displays the growth in annual mains water demand for selected demand scenarios. Use of rainwater tanks results in significant demand reduction. By the year 2042 the G+2% scenario has reduced annual mains consumption from 49,700 ML to 35,500 ML, a 28% reduction in demand on the headworks system. Table 4: Years in which augmentation of water supply is required Augmentation required for scenario (year) Augment Base Growth G+0.25% G+0.5% G+0.75% G+0.9% G+2% stage 2026 2054 2058 2062 NR NR NR 1 NR NR NR NR NR 2062 NR 2
55000
6000

Balance of $36M investment ($M)

Demand (ML/yr)

50000

Base G+0.9% G+2% Growth

5000 4000 3000 2000 1000 0 2000 -1000

Base Growth G+0.9% G+2%

45000

40000

35000 2000

2020

2040

Figure 6: Mains water demand with different levels of rainwater tank use

Year

2060

2080

2100

2020

2040

2060

2080

2100

Figure 7: Annual equivalence for the different demand scenarios

Year

Figure 7 presents a time series of annual equivalent economic position for the different scenarios. The growth scenario provides the greatest benefit to the community. This benefit will be greater than $5B during the next 100 years. The present equivalence analysis (Table 5) shows that for the base and G+0.5% to G+2% scenarios capital investment is required to ensure viability of the strategy. The growth and G+0.25% scenarios do not require capital investment and generate additional funds. Table 5: Present equivalence results for the water supply scenarios
Scenario Investment ($M) Base 36 Growth -14 G+ 0.25% -4 G + 0.5% 8 G + 0.75% 16 G + 0.9% 24 G + 2% 85

The results show that scenarios using installation rates of greater than or equal to 2% per year of rainwater tanks to existing dwellings and to all new dwellings were not financially viable when compared to the base scenario. However, scenarios using installation rates of up to 0.9% per year provides greater benefits to the community than the traditional water supply option. 5. CONCLUSIONS

The benefits of WSUD source control approaches arise from reduced mains water use and reduced stormwater infrastructure. Rainwater tanks contribute significantly to these benefits. Water levels in rainwater tanks used to supply domestic toilet flushing, exhouse and hot water uses are constantly drawn down. This ensures that the tank regularly has storage capacity available to accept roof runoff resulting in reduced mains water use and stormwater discharge. The rainwater tank is also more efficient at harvesting water than the traditional water supply system because roof catchments are not subject to the large losses experienced by water supply catchments. This study demonstrates that the use of rainwater tanks to supply exhouse, hot water and toilet flushing demand can delay construction of new water supply headworks infrastructure by up to 34 years and reduce annual regional water demand by up to 24,700 ML in the Lower Hunter region. In the Central Coast Region construction of new water supply headworks infrastructure can be delayed by up to 38 years and, in some cases, eliminated. Annual regional water demand is reduced by up to 14,500 ML. In addition, the case study demonstrates that use of WSUD source controls including rainwater tanks in new urban developments offers the economically most efficient infrastructure solution providing benefits to the community of up to $6B in the Lower Hunter Region and up to $5B in the Central Coast Region. However, these conclusions need to be tempered by the limitations of the study. This study has not valued the environmental benefit associated with delaying the construction of dams to augment water supply. Moreover, the construction and lifecycle costs of WSUD approaches have only been assessed approximately, albeit conservatively. Therefore, the benefits of WSUD source control approaches have most likely been understated. Current work is directed at addressing these limitations. 6. ACKNOWLEDGEMENTS

The authors gratefully acknowledge the assistance provided by Brendan Berghout of the Hunter Water Corporation and Ian Grimster from the Gosford/Wyong Water Authority. Funding for the study was provided by Newcastle City Council and the New South Wales Government Stormwater Trust. 7. REFERENCES

Andoh R. Y. G., and Declerck C., (1999). Source control and distributed storage a cost effective approach to urban drainage in the new millennium. Proceedings of the 8th International Conference on Urban Storm Drainage. pp.1319-1326. Sydney. Australia. Australian Bureau of Statistics (2000). Australia in profile: A regional analysis. Coombes P. J., Argue J. R., and Kuczera G., (2000). Figtree Place: A case study in Water Sensitive Urban Development. Urban Water Journal, Vol. 1, No. 4, London, UK. Coombes P. J., Kuczera G., Argue J. R., and Kalma J. D., (2000a). Costing of water cycle infrastructure savings arising from Water Sensitive Urban Design source control measures. Second International Conference on Decision making in Urban and Civil Engineering. Lyon, France. Coombes P. J., Kuczera G., Kalma J. D., (2000b). A probabilistic behavioural model for prediction of exhouse water demand. 3rd International Hydrology and Water Resource Symposium. Perth, Australia. Hunter District Water Board (1982). Water development plan 2010. Unpublished report. Kuczera G., (1992). Water supply headworks simulation using network linear programing. Advances in Engineering Software. 14, pp.55-60. Kuczera G., and Ng W. S., (1994). Stochastic economic approach to headworks augmentation timing. Research report No.72. Urban Water Research Association of Australia. Mc Alister A., (1998). Brisbane City Council water sensitive urban design case study. Hydrastorm98, Adelaide. Australia. Smith W. G., (1979). Engineering Economy: Analysis of capital expenditures. 3rd ed. The Iowa State University Press. Iowa, USA