Domestic Scale Rainwater ASR Demonstration Project Project Status Report July 2003 – June 2005 Peter Dillon and Karen Barry

Progress Report No. 2

21 December 2005

For Patawolonga Catchment Water Management Board

Copyright and Disclaimer © 2005 CSIRO To the extent permitted by law, all rights are reserved and no part of this publication covered by copyright may be reproduced or copied in any form or by any means except with the written permission of CSIRO Land and Water. Important Disclaimer: CSIRO advises that the information contained in this publication comprises general statements based on scientific research. The reader is advised and needs to be aware that such information may be incomplete or unable to be used in any specific situation. No reliance or actions must therefore be made on that information without seeking prior expert professional, scientific and technical advice. To the extent permitted by law, CSIRO (including its employees and consultants) excludes all liability to any person for any consequences, including but not limited to all losses, damages, costs, expenses and any other compensation, arising directly or indirectly from using this publication (in part or in whole) and any information or material contained in it.

Domestic Scale Rainwater ASR Demonstration Project Project Status Report July 2003- June 2005

Peter Dillon and Karen Barry CSIRO Land and Water

CSIRO Land and Water Client Report December 2005

Acknowledgements The monitoring reported in this paper was made possible through the support of the Patawalonga Catchment Water Management Board. Flow meters were provided by SA Water Corporation. The authors also thank Paul Pavelic of CSIRO Land & Water for his assistance with pumping and flow metering tests.

i

Executive Summary

This report describes results of two years of operating a domestic scale ASR demonstration project at Kingswood, in the Brown Hill Creek urban catchment that flows to the Patawolonga Basin. In 2004/05 571 mm rainfall at the site resulted in 133 KL rainwater injection into a shallow well. Salinity of recovered water continued to be a problem at the site. Swapping ASR to the observation well used in the previous year improved recovery efficiency, but not enough to make the system useful for irrigation. Further improvements were identified and funding is sought for a third year, to allow performance evaluation of the refined system.

ii

Table of Contents

Executive Summary i List of Tables iii List of Figures iii

1. Introduction

1

2. Site Description and Operation

1

3. Results and Discussion

3

3.1. Water balance

3

3.2. Piezometric water levels

6

3.3. Monitoring the ASR well and quality of recovered water

7

3.4. Well profiles

10

3.5. Quality of ambient groundwater and rainwater

13

4. Summary and Recommendations

16

5. References

17

APPENDIX I – Rainfall and Tank level plots, EC and Turbidity data

18

APPENDIX II – ISMAR5 paper 20

iii

List of Tables

Table 1(a). Monthly water volumes Year 2003 – 2004

4

Table 1(b). Monthly water volumes Year 2004 – 2005

4

Table 2. The efficiency of water capture for 2003/04 and 2004/05

5

Table 3. Quality of ambient groundwater and rainwater 15

List of Figures

Figure 1. Location of Site

2

Figure 2. Site layout looking south

2

Figure 3. Cumulative recharge and recovery volumes

5

Figure 4. Monitored standing water levels from July 2002 to July 2005

6

Figure 5(a). EC monitored in ASR well

7

Figure 5(b). Temperature monitored in ASR well

7

Figure 6. Changes in temperature and EC following injection events in the 100 mm well during September 2003

8

Figure 7(a). Electrical conductivity of recovered water from the 100 mm well after recharge

9

Figure 7(b). Electrical conductivity of recovered water from the 125 mm well after recharge

9

Figure 8(a). Electrical conductivity and temperature profiles during the 1st year of ASR

10

Figure 8(b). Electrical conductivity and temperature profiles during the 2nd year of ASR

11

Figure 9(a). Flowmeter and geophysics log for the 100 mm well

12

Figure 9(b). Flowmeter and geophysics log for the 125 mm well

13

Figure 10. Weight of material retrieved from rooftop gutters 14

1

1. Introduction Successful Aquifer Storage and Recovery (ASR), enables water to be stored below ground for recovery at a later date. Following a research project on stormwater ASR at municipal scale in a Tertiary limestone aquifer (Dillon et al, 1997), that led to guidelines for this practice (Dillon and Pavelic, 1996), ASR has been taken up widely in Adelaide where there are now 22 sites recycling 2 million m3/year of stormwater (Hodgkin, 2004). This is a small but growing practice, which is contributing to the security of Adelaide’s water supply (Dillon et al 2004). An evaluation of a shallow Quaternary system of alluvial aquifers of the Adelaide Plains (Pavelic et al, 1992) revealed that much of the metropolitan area has potential for storage of water on a smaller scale in brackish aquifers. For example, by recharging domestic roof runoff for use in garden watering in summer to further reduce demand on a stressed water supply system. However, it is recognised that distributed management by householders of many small ASR schemes also has potential for creating problems, through inadequate knowledge on how to operate and maintain collection systems and ASR wells and the potential for groundwater levels to rise to problematic levels and for pollution of groundwater (eg. see Dillon et al 2004). This is one of the activities being explored to support improved stormwater management as part of Patawolonga Catchment Water Management Plan (2002) which is coherent with the South Australian Government (2004) strategy ‘Water Proofing Adelaide’. Since June 2003 an ASR Demonstration site has been in operation at Kingswood, South Australia draining roof runoff from a domestic dwelling by gravity feed into a shallow alluvial aquifer. The objectives of the study are:

• to measure operational performance of domestic scale ASR, determine departures from predicted performance, and identify solutions to any problems that arise, so that in future appropriate advice can be given when permitting such installations

• to provide a rainwater ASR comparison for an intended operational ASR site with treated stormwater runoff in shallow alluvium at the Urrbrae Agricultural High School wetland, (approx 300m north east)

• to provide a site to determine whether recovered water quality from a nearby well meets all drinking water criteria, after the plume of fresh water reaches that well.

2. Site Description and Operation In April 2002 two bores (Unit No. 6628-20967, Permit No. 58120 and Unit No. 6628-20967, Permit No. 57956) were drilled in the rear garden of a residential dwelling in Kingswood, South Australia. The suburb of Kingswood is located in the base of the foothills 6 km’s south east of the city of Adelaide (figure 1). The two bores with respective diameters of 125 mm (#6628-20967, southern well) and 100 mm (#6628-20968, northern well), located 5 m apart, were installed to a depth of 24 m in the upper quaternary aquifer. Below surficial clay the profile is a mixture of clay, sand and gravel to 21 m depth with a stiff clay base. Pumping tests’ conducted on 12 December 2002 showed well yields to be 1.1 L/s (#6628-20968) and 0.4L/s (#6628-20967) with an ambient groundwater salinity of 2500 mg/L TDS, and depth to ambient groundwater level of 12 m. For many years roof runoff had drained into a sump adjacent to the southern well, but there was no evidence of freshening of the aquifer, suggesting that surficial and deeper layers were not well connected. From these results it was decided to install the injection line in the 100 mm (#6628-20968) well with the higher hydraulic conductivity and hence greater capacity to receive injected water. The roof area of 285 m2 was connected to a 4 m3 tank (3 m3 active storage) by a syphonic drainage system. That is, drainage pipes run underground and at the tank, rise and discharge into the top of the tank. This has implications for water quality which will be described later. There is an off

2

take from the tank to the ASR well via a 100 um filter, allowing gravity feed when the volume of the water in the tank rises above the outlet level. Two low-head flow meters to monitor inflows and outflows to and from the ASR well were installed in the injection and recovery pipes (figure 2).

Groundwater levels have been monitored using a combination of pressure transducers and capacitance probes, with occasional manual measurements, since 11 July 2002. Recharge to the ASR well (injection of rainwater from roof run-off) began on the 29 June 2003 and on the 30 July 2003 a YSI 600XLM water quality sonde was installed in the ASR well, which continuously monitors the electrical conductivity (EC) and temperature. Daily rainfall is recorded using a rain gauge. Periodic samples of the roof run-off have also been collected before, during and after injection events, since the start of recharge. There has been on average 0.6 KL recovered per month to purge the ASR well of any accumulated biofilm or sediments and to assess the quality of the stored water. During the first year of operation (29 June 2003 to 21 July 2004) the injection line was installed in the 100 mm diameter well. Although 142 KL

Figure 2. Site layout looking south

Figure 1. Location of Site

3

of rainwater was injected during this first year, the first 0.6 KL of water recovered during purging events showed the salinity of water was still close to background level and hence not suitable for garden irrigation. Following assessment of the recovered water quality data, a second set of pumping tests was carried out on 22 July 2004 to assess the permeability changes and it was decided to immediately move the injection line to the less permeable well 125 mm diameter well (#6628-20967), to determine if the reduced aquifer permeability would keep the fresh injectant closer to the well and thereby reduce the salinity of the recovered water (ie. improve recovery efficiency). On 20 September 2004, to further automate the monitoring of the ASR site, a logger unit was connected to two pressure transducers and a pluviometer, to monitor standing water levels in the ASR well and the rainwater tank and measure rainfall. The unit also has the capability to monitor the volumes and temperature of inflow and outflow and the electrical conductivity (EC) of recovered water. The sensors for these parameters were installed on the 4 August 2005. Electromagnetic vertical flow velocity metering, downhole resistivity, gamma logging and EC profiling were carried out on 31 May 2005 to enable more detailed analysis of the well permeability with respect to depth and hence ways in which recovery efficiency could be further improved. 3. Results and Discussion 3.1 Water balance This report presents the data from the first two years of the study, 29 June 2003 – 30 June 2005. Table 1 summarises the monthly volumes of injection and recovery as well as rainfall and household mains water consumption. Occasionally the 100 µm filter became blocked or air within the injection line inhibited injection following the cleaning of the filter. If this happened prior or during a large storm event, the tank overflowed to the sump. This occurred on 6 separate occasions, during December 2003, June, July (twice), November 2004 and June 2005 and volumes are recorded as estimated spill in Table 1, to help account for variations in the relationship between monthly rainfall and injection. Tank levels also varied following each recharge event, depending on clogging of the filter and intermittency of rain during recharge. Plots of daily rainfall and tank levels are included in Appendix 1. Figure 3 shows the relative potential contribution of harvested water to the household supply. Over the 24 month period a total volume of 275 KL was injected and 22 KL was recovered, leaving a net increase in storage of 253 KL. Recovery consisted mostly of intermittently purging the well to remove any accumulated particulate matter. This water was discharged onto the garden or into a nearby sump and soaked into the ground as additional recharge to shallower alluvium, along with a small amount of runoff from paved areas adjacent the house.

4

Table 1(a). Monthly water volumes Year 2003 – 2004

Injection

Volume (KL) Recovered

Volume (KL) Rainfall (mm) Mains (KL) Spill (KL)

June a 2.8 July 11.6 1.1 42.5 20.4 August 22.4 0.5 85.3 22.3 September 15.0 0.6 59.4 20.3 October 12.9 0.3 62.4 19.5 November 1.7 0.6 7.4 29.6 December 7.3 0.0 35.2 38.5 1.6 January 3.2 0.0 15.0 40.8 February 1.3 0.0 5.1 54.3 March 7.8 0.7 29.3 33.8 April 2.9 0.6 18.7 23.3 May 21.9 0.6 74.6 24.4 June 34.6 0.6 128.2 26.7 0.5

Year 03/04 145.3 5.4 563.1 353.8 2.1

a injection started 29 June 2003

Table 1(b). Monthly water volumes Year 2004 – 2005

Injection

Volume (KL) Recovered

Volume (KL) Rainfall (mm) Mains (KL) Spill (KL) July a 23.5 9.2 101.6 33.3 0.5 August 25.2 1.5 90.0 21.1 2.8 September 18.9 0.9 65.6 18.9 October 1.1 0.6 16.2 22.8 November 19.2 0.6 74.7 22.8 4.4 December 4.0 0.6 17.4 36.6 January 6.4 0.6 37.7 41.7 February 5.1 0.6 17.4 37.6 March 2.7 0.6 12.6 36.1 April 0.1 0.6 6.1 37.1 May 0.9 0.6 4.6 22.9 June 26.1 0.6 126.7 21.6 6.1 Year 04/05 133.2 17.0 570.6 352.4 13.8 100 mm well 12.0 6.6 42.9 22.2 125 mm well 121.3 10.5 527.7 330.2

a injection line installed in 125 mm well on 22nd July 2004

5

0

100

200

300

400

500

600

700

800

Jul-03 Sep-03 Nov-03 Jan-04 Mar-04 May-04 Jul-04 Sep-04 Nov-04 Jan-05 Mar-05 May-05 Jul-05

Volu

me

(KL)

Mains water consumption

Net recharge (100 mm well)

Net recharge (125 mm well)

ASR well (100 mm) ASR well (125 mm)

Figure 3. Cumulative recharge and recovery volumes During the first two years of the study, rainfall data from the nearest Bureau of Meteorology station at the Waite Institute (Lat: -34.97o Long: 138.63o) recorded 546 mm and 565 mm during the years 2003/04 and 2004/05 respectively, compared with the average rainfall of 622 mm. The study site has recorded slightly higher yearly rainfalls at 563 mm and 571 mm for each of the years so far. Using the data from Table 1, the efficiency of water capture can be calculated for each of the years to date, 2003/04 and 2004/05 (Table 2).

Table 2. The efficiency of water capture for 2003/04 and 2004/05

2003/04

2004/05

Total runoff / total rainfall 0.901 0.903 Total recharge / total runoff 0.985 0.906 Net recharge / total recharge 0.949 0.946 Net recharge / total rainfall 0.843 0.775 Net recharge (KL)* 135.3 126.0 Total rainfall (KL) 160.5 162.6

* Allowing 0.6 KL/month recovery for flushing well Note that the proportion of water that can be recovered at a suitable salinity for use cannot be determined yet. When the 100 mm well north (#6628-20968) was used it was found to be < 0.1 % of net recharge, but this was only after one season. The initial recovery efficiency at an equivalent stage for the 125 m south ASR wel (#6628-20967) will be determined in the coming summer.

6

3.2 Piezometric water levels Standing water levels have been monitored for the last three years using a combination of loggers and manual measurements. Figure 4 plots all available recorded data, with gaps indicating there were some occasions of logger stoppage. The annual groundwater level fluctuation is approx 1m, with the peak in December and the lowest levels in June, approximately a 2-3 month lag on rainfall fluctuations. Surprisingly there is a downward trend in groundwater levels that eclipses the effect of the increased recharge from the ASR system. Both wells record the same level when there is no recharge or recovery. #6628-20968 (100 mm well north) In the 1st year of ASR operation in the 100 mm well, it had been observed that the head only rose by about 1 metre during injection periods, with injection rates typically at 0.3L/s (figure 4). Recovery events in the 100 mm well, were run at up to 1 L/s for approximately 8.5 minutes, with the well never running dry. During this period, data was logged at 1 hour intervals and consequently recovery events were not recorded. #6628-20967 (125 mm well south) During the 2nd year of operation in the lower yielding well, the heads were seen to rise up to 6 m (though not all events were captured due to the length of the log interval). With the installation of new flow meters for the injection and recovery lines on 4 August 2005, event triggered logger intervals at 20 L during injection and 10 L during recovery will enable more detailed recording of the head changes during particular events and allow analysis of any changes in conductivity of the well. Drawdown during initial recovery events reached the level of the pump (18 m) rapidly at 1 L/s, causing the pump to shunt. Consequently, recovery events in the 2nd year of operation were run at a lower rate by pumping at 1 L/s until shunting occurs (normally within 2 minutes) then reducing the rate of pumping to between 0.15 and 0.25 L/s until a total of 510 L was recovered. Then intermittent pumping at 1 L/s was carried out until shunting occurred and was allowed to recover for 1 mintute before continuing again to purge a much residual material as possible from the well. At approximately 600 KL the recovery was stopped (generally after 2 to 4 cycles of 1 L/s).

11

11.5

12

12.5

13

13.5

14Jul-02 Oct-02 Jan-03 Apr-03 Jul-03 Oct-03 Jan-04 Apr-04 Jul-04 Oct-04 Jan-05 Apr-05 Jul-05

SW

L - T

OC

(m)

100 mm well

125 mm well

ASR well (100 mm) ASR well (125 mm)

1 m injection head

injection heads rise at least 6 m

Figure 4. Monitored standing water levels from July 2002 to July 2005

7

3.3 Monitoring in the ASR well and quality of recovered water Electrical conductivity (EC) and temperature have been continuously logged in the ASR well for 2 years using a YSI 600XLM© water quality monitoring sonde, Figure 5 (a&b). The ambient salinity (>4000 uS/cm) is diluted with fresh rainwater (<100 uS/cm) during injection events. With regular injection events in the winter months, the salinity in the well is maintained below 1000 uS/cm. However once injection events become less frequent (during summer) the EC of water in the ASR well increases rapidly towards ambient conditions, Figure 5(a). The temperature of ambient groundwater is 18˚ C, and during winter rainwater can be less than 10˚ C and in summer up to 25˚ C, Figure 5(b).

0

1000

2000

3000

4000

5000

Jul-03 Sep-03 Nov-03 Jan-04 Mar-04 May-04 Jul-04 Sep-04 Nov-04 Jan-05 Mar-05 May-05 Jul-05

EC (u

S/cm

)

EC 100 mm well

EC 125 mm well

ASR well (100 mm) ASR well (125 mm)

Irrigation threshold (equiv. to 1500 mg/L TDS)

Figure 5(a) EC monitored in ASR well

5

10

15

20

25

30

Jul-03 Sep-03 Nov-03 Jan-04 Mar-04 May-04 Jul-04 Sep-04 Nov-04 Jan-05 Mar-05 May-05 Jul-05

Tem

pera

ture

(deg

rees

C)

Temp 100 mm well

Temp 125 mm well

ASR well (100 mm) ASR well (125 mm)

Figure 5(b) Temperature monitored in ASR well

8

In Figure 6, electrical conductivity, temperature and standing water level are plotted over a one month period (September 03), during a period of high recharge. This shows that with sufficient recharge the average temperature of the ASR well water can be depressed by up to 5˚ C in winter and that it returns to equilibrium more quickly than salinity due to the thermal mass of aquifer material. This also shows that water remains fresh when recharge events are frequent but as soon as the frequency diminishes the EC increases in the well. This is further confirmed with the data obtained during recovery events.

3

6

9

12

15

18

1/09/2003 8/09/2003 15/09/2003 22/09/2003 29/09/2003

swl d

epth

from

TO

C (m

)/Tem

p (d

egre

es C

0

500

1000

1500

2000

2500

3000

EC (u

S/cm

)

swl depth from TOC (m)Temp (degrees C)EC (uS/cm)

Figure 6. Changes in temperature and EC following injection events in the 100 mm well

during September 2003 During recovery events, samples have been collected for EC and turbidity measurement. The rate of salinity increase depends on the number of days since the last recharge event and the volume of water injected. These have been plotted for each well in Figure 7 (excluding the pump test event in July 2004). The volume of rainwater injected between recovery events was more evenly distributed during the first year of operation in the 100 mm north well (#20968) than during the second year in the 125 mm south well (#20967). While there are differences in frequency and volume of recharge events, it appears that in the 2nd year of operation, when the 125 mm well was used as the ASR well, the water was fresher in general at the start of recovery events. However EC still rapidly increases during recovery. It is apparent that there is still an insufficient volume of freshwater accumulated around the well to create a buffer against the saline ambient groundwater. Appendix 1 tables all turbidity and EC data from each of the monthly ‘pump-out’ events.

9

0

10

20

30

4 4 7 2 1 14 7 1 1 4

Days since last injection prior to recovery event

Volu

me

Inje

cted

(KL

)

0

1000

2000

3000

4000

5000

0 1 2 3 4 5 6 7 8

Cummulative Volume Recovered (KL)

EC (u

S/cm

)

Figure 7(a). Electrical conductivity of recovered water from the 100 mm well after recharge

0

10

20

30

1 6 5 1 5 23 5 1 28 15 7

Days since last injection prior to recovery event

Volu

me

inje

cted

(KL)

0

1000

2000

3000

4000

5000

0 1 2 3 4 5 6 7 8

Cummulative Volume Recovered (KL)

EC (u

S/cm

)

Figure 7(b). Electrical conductivity of recovered water from the 125 mm well after recharge

10

3.4 Well Profiles Conductivity and temperature profiles have been carried out several times during the trial using a YSI 600XL© profiling sonde. In the first year of operation, the electrical conductivity profile was measured 3 times (22/09/03, 18/03/04 & 4/06/04) in the initial observation well (125 mm, PN 58120) to detect any possible breakthrough and once (4/06/04) in the ASR well (100 mm, PN 57956) when the pump was temporarily removed. The profiles showed distinct stratification within the 24 m depth of each well with the fresher injected water overlying saltier water, Figure 8(a).

10

14

18

22

26

0 1000 2000 3000 4000 5000

EC (uS/cm)

Dep

th T

OC

(m)

22.09.03 18.03.04 4.06.04 4.06.04 (100 mm)

10

14

18

22

26

15 16 17 18 19 20

Temperature (degrees C)

Dep

th T

OC

(m)

22.09.03 18.03.04 4.06.04 4.06.04 (100 mm)

Figure 8(a). Electrical conductivity and temperature profiles during the 1st year of ASR On 22 July 2004, the injection line was moved to the 125 mm well and during this 2nd year of ASR operation, three sets of profiles have been carried out in each well. The first was near the end of winter, 21 days (11/08/04) since installation of the injection line in the 125 mm well, during which 30 KL had been injected (avg 1.4 KL/day). The second was in the middle of summer after 192 days (30/01/05) and injection of 86 KL (avg 0.45 KL/day) and the third was near the start of winter after 312 days (30/05/05) and injection of 95 KL (avg 0.3 KL/day). Figure 8(b) shows the profiles in the 125 mm and 100 mm wells for each of these 3 dates, it is observed that the EC profile is initially fresh (<500 uS/cm) throughout the profile during periods of regular injection during winter, however as the injection events decrease during summer and to the start of the following winter, the EC at the base of the profile

11

between 20 – 24 meters moves back to ambient conditions. The temperature of the water is also seen to move rapidly back to ambient conditions, throughout the profile. It is a possibility that by limiting the ASR well’s depth to 20 m, an improvement in recovery efficiencies could be observed. This was assessed by performing downhole velocity profiling in May 2005.

10

14

18

22

26

0 1000 2000 3000 4000 5000

EC (uS/cm)D

epth

TO

C (m

)

11.08.04 (100 mm) 11.08.04 (125 mm)31.01.05 (100 mm) 31.01.05 (125 mm)30.05.05 (100 mm) 30.05.05 (125 mm)

10

14

18

22

26

15 16 17 18 19 20Temperature (degrees C)

Dep

th T

OC

(m)

11.08.04 (100 mm) 11.08.04 (125 mm)31.01.05 (100 mm) 31.01.05 (125 mm)30.05.05 (100 mm) 30.05.05 (125 mm)

Figure 8(b). Electrical conductivity and temperature profiles during the 2nd year of ASR Downhole flow metering under ambient conditions and during pumping at a low flow rate from just below the water table, as well as Gamma and Neutron Logs were carried out in each of the wells on 31 May 2005. From previous experience and by piezometry it was shown that pumping to produce the first profile had no impact on results for the second profile as hydraulic equilibrium had been re-established prior to the second test. Although both wells have slotted casings from their base to 2 m below ground level, with the annulus gravel-packed, the length of the packer used for downhole flow metering exceeded the length of the individual slots. Short-circuiting of flow via the annulus did not occur, as can be seen from the flowmeter results in the bottom 4 m of each well, where no flow was induced. The results for each well have been plotted against recent EC profiles (30/05/05) in Figure 9. The zone of lowest flow (<0.005) corresponds to the bottom 3 meters of the well where the stratigraphic log indicates clay. Interestingly the area of highest flow in the 100 mm well is were the EC is at its highest level, whereas in the 125 mm well, the higher EC correlates with an area of low flow. Two main mechanisms for salinity stratification are possible. One is that the fresh injectant having lower density than the brackish native groundwater is more buoyant and rises to the top of the saturated zone and spreads laterally. This would result in a salinity gradient near the base of the well which could be expected to rise with time as the plume spreads. Given the relatively small salinity contrast, the cooler temperature of most injectant than native groundwater, and

12

the expected anisotropy in the hydraulic properties of the layered porous media within the saturated zone, it was thought that this mechanism was unlikely to be significant. An alternative explanation is that although permeability of the lower part of the formation is low, the brackish native groundwater from this lower layer contributes to the salinity of water recovered from the well. In this event, backfilling the base of the 125 mm well to 20 m may prevent the brackish water entering the well between injection events and would be likely to reduce the salinity of recovered water without measurably reducing the well yield. The 125 mm well was backfilled with sand and bentonite to a depth of 19.8 m on 2 September 2005, further observations will be required to assess the effectiveness of this treatment. For the 100 mm well, no such treatment was possible without seriously impacting on the yield of the well.

Depth Vs Flow and EC for 100 mm well

10

12

14

16

18

20

22

24

0 1 2 3 4 5

Dep

th T

OC

(m)

EC (mS/cm)Induced Q

(L/min)

SWL @ 13.2 m

Clay

Stiff clay

Depth Vs Gamma/Neutron for 100 mm well

10

12

14

16

18

20

22

24

0 50 100 150 200 250

Gamma (API)Neutron (CPS)soil profile

Figure 9(a) Flowmeter and geophysics log for the 100 mm well (31/05/05)

13

Depth Vs Flow and EC for 125 mm Well

10

12

14

16

18

20

22

24

0 1 2 3 4 5

Dep

th T

OC

(m)

EC (mS/cm)

Induced Q (L/min)

SWL @ 13.1 m

Clay

Stiff clay

Depth Vs Gamma/Neutron for 125 mm well

10

12

14

16

18

20

22

24

0 50 100 150 200 250

Gamma (API)

Neutron (CPS)

soil profile

Figure 9(b) Flowmeter and geophysics log for the 125 mm well (31/05/05)

3.5 Quality of ambient groundwater and rainwater Aside from electrical conductivity, turbidity and temperature measurements, no detailed water quality analyses in groundwater have extended beyond the initial sampling in December 2002 from each bore. However, there have been five grab samples collected from the ‘house tank’ (injectant) and one from the ‘shed tank’ since recharge commenced on 29 June 03 to end of June 2005. Table 3 assembles these data, (most of which have been reported previously).The ambient groundwater is saline with low organic carbon and phosphorus but significant organic nitrogen. The rainwater is fresh with minimal nutrient levels but detectable zinc levels. The shed tank sample was collected as a comparison with the current injectant water quality, to assess its use as an extra source of injectant water to increase the volume of water which can be recharged. The leaf trap sample was taken on 19 May 03 from a roof gutter where the down-pipe leaf guard had become choked with decaying leaves and water had been sitting in the gutter for some time. This sample contained coloured water with a musty odour and is taken to be a worst case for the potential for nutrient loading in the rainwater available for injection. This is a consideration for evaluating the potential bio-clogging of the well (which has not occurred yet). There is no first-flush diversion in the rainwater collection system, and with syphonic

14

delivery of water from the downpipes to the head of the tank, these data demonstrate that anaerobic conditions could establish in the underground pipe leading to the tank. This explains the sulphidic odours smelled in the head space of the tank during autumn 2004 and 2005 after small showers of rain. This led to an expansion of the study in 2004/2005 to include monthly gutter clearing, with drying and weighing of leaves and gutter detritus, with a view to developing an approximate mass balance for organic carbon. Figure 10 shows a monthly summary of net dry weights of material removed from the gutters between April 2004 and June 2005. This material included varying mixtures of leaf matter and silt with occasional bees, feathers and bird droppings. As expected, the peak period corresponds with the Autumn months, when deciduous trees drop their leaves. Taking into account the roof area of each of the gutter zones, the area ‘front west’ is shown to have contributed at least half the litter load, followed by ‘front east’ and ‘carport’, with the ‘back’ only about 5 percent of the total load. There are two reasons for this. There are deciduous trees in the front yard and street trees that contribute loading, whereas at the rear of the house, there are no overhanging trees. Secondly, only front west and east have flywire screens inside the 12 mm screen used on all the downpipes to trap leaves. The carport and front east are considered to have the same proximity to trees and the contrast between these is considered to be primarily due to the presence or absence of flywire screen (1.4 mm aperture).

0

100

200

300

400

500

600

Apr-04

May-04

Jun-04

Jul-04

Aug-04

Sep-04

Oct-04

Nov-04

Dec-04

Jan-05

Feb-05

Mar-05

Apr-05

May-05

Jun-05

Net

dry

wt.

(mg/

m2/

day)

carport

front west

front east

back

April 04 'front west' had a load of 1500 mg/m2/day

Figure 10. Weight of material retrieved from rooftop gutters

15

Table 3. Quality of ambient groundwater and rainwater

100mm Well 125mm Well Leaf trap Shed TankParameters (mg/L) 12/12/02 12/12/02 19/05/03 21/05/03 17/08/03 1/10/03 12/11/04 1/01/05 17/08/03

EC (uS/cm) 4005 4150 42.3 21.5 28.2 23 21 70.5pH-field 6.64 6.63 7.82 5.8Temp-field (oC) 19.6 19.8 16.7DO-field 2.17 2.31 4.32Eh-field (mV SHE) 365 372 609Turbidity (NTU) 2.12 0.75 1.05 1.2 0.99 1.9Total Organic Carbon 2.0 2.6 694 <1.0 2.2 0.8 3.0 0.9 0.8Calcium 151 149 57 1.1 0.7 0.4 0.4 1.8Magnesium 173 165 12.4 0.5 0.4 0.4 0.3 0.7Potassium 12.1 12.4 13.1 <1 <1 <1 <0.1 0.6Sodium 503 530 14 2.9 1.8 2.3 2.8 10.5Bicarbonate 675 717 288 7 7 6 19Chloride 990 1000 20 7 9 8 4.2 3.4 6Sulphate 140 140 33.1 <1.5 <1.5 <1.5 0.1a 9.2Ammonia as N 0.027 0.022 30.1 0.116 0.347 0.053 0.12 0.016Phosphorus - Total as P <0.025 <0.025 39.7 0.018 0.026 0.047 <0.1 0.029 0.156Kjeldahl Nitrogen as N <0.25 <0.25 275 0.17 0.39 0.2 0.64 0.21Nitrate + Nitrite as N 5.82 5.87 0.006 0.202 0.389 0.075 0.2 0.416Total Nitrogen 0.8Total Arsenic <0.001 <0.001 0.066 <0.001 <0.001Total Aluminium <0.1Boron <0.1Total Copper 0.162 <0.001 0.002 <0.02Total Iron <0.03 <0.03 148 <0.03 <0.03 <.03 <0.1 <0.03Total Manganese <0.05Total Lead 3.361 <0.0005 0.0006Total Strontium <0.003 <0.003 <0.1Total Zinc 16.99 0.138 0.248 0.124 0.08 6.17E.coli ( /100 mL) 0 0 2400 7Coliforms ( /mL) 680 520 >2400 >2400Faecal Coliforms (/100mL) 7Colony Count (20oC) Aerobic ( /mL) >10000 >10000True Colour @ 456 nm (HU) 2 0.8a as S only

House Tank (Injectant)

16

4. Summary and recommendations The total recharge in two years of 275 KL has been achieved from a total of 1134 mm rain, from a roof area of 285 m2. Operation has continued to be very intensive with the 100 micron filter requiring flushing every 2 to 10 KL depending on the turbidity and nutrients in the runoff. In its current form, the energy required to maintain recharge is greater than is likely to occur more broadly among well-owners. It is proposed to improve filtration so that maintenance is not as onerous by installing either a bag filter beneath the tank inlet screen or a second (larger) 100 um filter downstream of the tank. A leaf guard mesh has been purchased by the owner to apply to the section of guttering receiving the highest leaf loadings, on conclusion of the monthly leaf weighing study. The disappointing aspect of the trial so far is still the high salinity of recovered water due to the very low recovery efficiency achieved. After the injection line was moved to the 125 mm well, salinity of recovered water during redevelopment has continued to increase beyond values acceptable for irrigation supplies. However, flow metering carried out in May 05 has shown promising results. Analysis of the data along with neutron and gamma logs has shown there is potential for improvement of the recovery efficiency of the 125 mm well. The highest salinity zone is bottom three meters of the well, by back filling this bottom layer of the well it is hoped this saline zone will be isolated from the remainder of the bore volume, allowing higher recovery efficiency. Once this has been done, it is planned to increase the available volume of rainwater that can be injected by connecting up rainwater from a neighbours property in order to maximise the chance of retaining a freshwater plume in the aquifer to enable recovery for irrigation. It is expected that recovery efficiency will increase in future years, and data collected will be used to predict recovery efficiency under various scenarios, including relating to increased volumes via increasing the catchment. With the remaining equipment for the logger unit being commissioned in August 2005 to provide automated logging of flows, temperature of injectant and recovered waters and EC of recovered waters, there should be less reliance on daily manual measurements, once these data are correlated with the manual data readings. This increase in data capacity in the next 12 months will provide a detailed database to enable a completed water balance to be calculated as well as determining if the site can maintain a viable ASR operation. A detailed proposal for Stage 3 is attached. In June 2005 a poster paper was presented at ISMAR5 in Berlin, as a case study entitled “Domestic-scale ASR with rainwater at Kingswood, South Australia”. The costs of travel and attendance were covered by a CSIRO Chief Executive Study Award to enable continued development of further collaborative activities with researchers in Berlin. Support of the project by the Catchment Water Management Board was acknowledged. A copy of that paper, covering the study period July 2003 to December 2004, is attached.

17

5. References Dillon P.J. and Pavelic P. (1996). Guidelines on the quality of stormwater and treated

wastewater for injection into aquifers for storage and reuse. Urban Water Research Assoc. of Aust. Research Report No 109.

Dillon P., Pavelic P., Sibenaler X., Gerges N. and Clark R. (1997). Aquifer storage and recovery of stormwater runoff. Aust. Water & Wastewater Assoc. J. Water 24(4), 7-11.

Dillon P., Toze S. and Pavelic P. (2004). Conjunctive use of urban surface water and groundwater for improved urban water security. Proc. IAH Congress Zacatecas, Mexico 11-15 Oct 2004.

Hodgkin T. (2004). Aquifer storage capacities of the Adelaide region. South Australia Department of Water, Land and Biodiversity Conservation Report 2004/47.

Patawalonga Catchment Water Management Board (2002). Catchment Water Management Plan 2002-2007. http://www.cwmb.sa.gov.au/patawalonga/plans/Pat01_Cover.pdf

Pavelic P., Gerges N.Z., Dillon P.J. and Armstrong D. (1992). The potential for storage and re-use of Adelaide stormwater runoff using the upper Quaternary groundwater system. Centre for Groundwater Studies Report No. 40.

South Australian Government (2004). Water Proofing Adelaide Draft Strategy: A thirst for change. http://www.waterproofingadelaide.sa.gov.au/pdf/Strategy_full.pdf

18

APPENDIX I

0

5

10

15

20

25

30

35

40

45Ju

l-03

Aug

-03

Sep

-03

Oct

-03

Nov

-03

Dec

-03

Jan-

04

Feb-

04

Mar

-04

Apr

-04

May

-04

Jun-

04

Jul-0

4

Aug

-04

Sep

-04

Oct

-04

Nov

-04

Dec

-04

Jan-

05

Feb-

05

Mar

-05

Apr

-05

May

-05

Jun-

05

Dai

ly ra

infa

ll (m

m)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Jun-03 Aug-03 Oct-03 Dec-03 Feb-04 Apr-04 Jun-04 Aug-04 Oct-04 Dec-04 Feb-05 Apr-05 Jun-05

Wat

er le

vel i

n ra

inw

ater

tank

(m)

Logger DataManual Data

ctf of tank overflow (1.68 m)

base of tank

19

Quality of recovered water from the ASR well during monthly ‘pump-out’ events

Volume KL EC (uS/cm) Turbidity (NTU) Volume KL EC (uS/cm) Turbidity (NTU)

4/07/2003 1168 9.98 7/08/2004 0.335 73.5 6350.648 230 246

17/08/2003 0.912 875 29.4 0.891 361 49.11.197 1550 15.4 1.640 605 247

7/09/2003 0.014 195 64.1 5/09/2004 0.334 90.9 18.10.058 257 79.8 0.404 136.9 2520.115 640 69.3 0.620 271 14.80.230 863 34.1 0.876 358 12.90.285 1575 7.4

3/10/2004 0.171 113.8 14021/09/2003 0.000 282 3.52 0.303 233 683

0.058 238 55.5 0.530 627 220.115 512 159 0.629 839 4.880.173 755 39.30.230 895 28.6 7/11/2004 0.090 108.7 35.70.296 1011 15.6 0.151 114.2 629

0.300 344 932/11/2003 0.060 180 90.5 0.510 505 49

0.300 585 530.360 766 14.4 5/12/2004 0.090 108.2 1540.570 934 10.5 0.150 367 >1000

0.300 852 1066/03/2004 0.090 3380 170 0.510 1255 13.5

0.507 4460 201/01/2005 0.090 216 95.2

4/04/2004 0.087 1835 346 0.189 1184 2640.491 2850 20 0.296 1872 24.5

0.500 2280 4.742/05/2004 0.087 662

0.493 1552 6/02/2005 0.056 100 3.050.091 323 937

14/06/2004 0.086 192 875 0.303 1189 1700.489 585 49 0.526 1657 91

0.599 2000 76.94/07/2004 0.090 330 554

0.510 1326 41.6 5/03/2005 0.120 263 2650.238 1846 46.80.295 2040 40.10.501 2470 59.50.600 2470 25.2

2/04/2005 0.108 1316 >10000.168 2290 640.300 2510 40.507 3110 50.616 3200 8

1/05/2005 0.129 668 4260.204 2550 22.70.293 3380 100.509 3650 6.77

4/06/2005 0.100 1658 1760.157 2420 3180.300 3600 34.60.506 3650 1.350.610 3720 321

100 mm well 125 mm well

20

APPENDIX II Barry, K. and Dillon, P. (2005). Domestic-scale ASR with rainwater at Kingswood, South Australia. Proc. ISMAR5, Berlin, June 2005.

Domestic-scale ASR with rainwater at Kingswood, South Australia Karen Barry and Peter Dillon CSIRO Land and Water, PMB 2 Glen Osmond, SA 5064 Australia (Karen.Barry@csiro.au; Peter.Dillon@csiro.au).

Abstract Rainwater derived from roof surfaces can offer sufficient quantities of good quality water if it can be successfully harvested, particularly in semi-arid areas where alternative sources are scarce. A demonstration project in its second year is currently underway to investigate the operational performance of domestic-scale aquifer storage and recovery (ASR) with rainwater. A shallow alluvial aquifer in the southeast Adelaide metropolitan area was targeted after an earlier regional assessment had suggested this area has potential for ASR. The mean annual rainfall here is approximately 550 mm with most of this falling in winter between May and September. Annual evaporation (A class pan) is about 2000 mm, most of which occurs in the warm, dry summer months creating a significant irrigation demand. Two wells, located 5 m apart, were installed to a depth of 24 m. Their ambient salinity of 2500 mg/L TDS is considered too brackish for irrigation of the clayey soils. The run-off from a 250 m2 roof area of a single dwelling has been plumbed to the ASR well under gravity feed via a 4 m3 storage tank and 100 μm filter. Flow rate, volume, piezometric head, electrical conductivity and temperature are monitored continuously in the injection well and the samples of the influent and recovered water analysed periodically. After the first full 12 months of operation, with 563 mm rainfall, 142 KL had been injected with no observable clogging. Although the salinity of the groundwater has been reduced, it is not yet sufficiently fresh to be useful for irrigation supplies due to mixing losses and the effect of regional groundwater flow. Productive use of the groundwater may be possible once greater volumes have been injected, however the supply is limited by available roof area and rainfall variability. The main challenge is therefore to meet operational performance criteria under the current constraints.

Keywords ASR; rainwater harvesting; salinity; water quality

INTRODUCTION In semi-arid areas, rainwater derived from roof surfaces may offer useful quantities of good quality water if it can be successfully harvested and stored. This may have value in developing countries where alternative sources of stormwater are often contaminated. Successful ASR allows water to be stored below ground for recovery at a later date. Following a research project on stormwater ASR at municipal scale in a Tertiary limestone aquifer (Dillon et al, 1997), that led to guidelines for this practice (Dillon and Pavelic, 1996), ASR has been taken up widely in Adelaide where there are now 22 sites recycling 2 million m3/year of stormwater (Hodgkin, 2004). This is a small but growing practice, which is contributing to the security of Adelaide’s water supply (Dillon et al 2004). An evaluation of a shallow Quaternary system of alluvial aquifers of the Adelaide Plains (Pavelic et al, 1992), revealed that much of the metropolitan area has potential for storage of water on a smaller scale in brackish aquifers. For example, by recharging domestic roof runoff for use in garden watering in summer to further reduce demand on a stressed water supply system. This is one of the activities being explored in support of improved stormwater management as part of plans by the Patawalonga Catchment Water Management Board (2002) and conforms with the South Australian Government (2004) strategy ‘Water Proofing Adelaide’. However, it was recognised that distributed management by householders of many small ASR schemes also has potential for creating problems, through inadequate knowledge on how to operate and

maintain collection systems and ASR wells and the potential for groundwater levels to rise to problematic levels and for pollution of groundwater (eg. see Dillon et al 2005). It was also not clear whether recovered water quality would meet the water quality requirements for its intended uses and whether, if water was recovered from nearby wells, the pathogen attenuation that aquifers provide could be relied on for producing supplies to substitute for some in-house uses. Hence the objectives of this study are, (i) to measure operational performance of domestic scale ASR, determine departures from predicted performance, and identify solutions to any problems that arise, so that in future appropriate advice can be given when licensing such installations, (ii) to provide a rainwater ASR comparison for an intended operational treated stormwater runoff ASR site in shallow alluvium nearby, and (iii) to provide a site to determine whether recovered water quality from a nearby well meets all drinking water criteria, after the plume of freshwater reaches that well. SITE DESCRIPTION AND OPERATION In June 2003 a Domestic Scale Rainwater ASR Demonstration site was established in the rear garden of a residential dwelling in Kingswood, South Australia. The suburb of Kingswood is located in the base of the foothills 6 km south east of the central business district of Adelaide. In 2002 two wells (100 and 125 mm diameter) located 5 m apart, were installed to a depth of 24 m in a Quaternary alluvial aquifer. The profile is largely made up of clay containing several layers of sand and gravel to 21 m depth and underlain by more clay. Results from ‘pumping-tests’ showed their well yields to be 1.1 L/s (100 mm, northern) and 0.3L/s (125 mm, southern) with an ambient groundwater salinity of 2500 mg/L TDS, and depth to ambient groundwater level of 12 m. From these results it was decided to install the injection line in the 100 mm (northern) well with the higher hydraulic conductivity and hence greater capacity to receive and recover injected water. The run-off from a single residential dwelling of 250 m2 was then plumbed to this ASR well under gravity feed via a 4 m3 tank (with 3 m3 active storage) and 100 μm filter. Two low-head flow meters were installed to monitor inflows and outflows to and from the ASR well. Groundwater levels have been monitored using a combination of pressure transducers and capacitance probes, with a backup of regular manual measurements. In July 2003 a YSI© water quality sonde was installed in the ASR well to continuously monitor the electrical conductivity and temperature. Sampling for basic water quality parameters for each well was done in December 2002, prior to the start of any injection to the well. Periodic samples of the roof run-off have also been collected before, during and after injection events. Since the start of injection there have been on average 0.6 KL recovered per month to purge the well of suspended solids and to assess the quality of the recovered water. During the first year of operation (July 03 to June 04) the injection line was installed in the 100 mm diameter well. On the 22 July 2004 following a second set of pumping tests the injection line was moved to the lower yielding well (125 mm) with the aim of increasing the volume of fresh water that could be recovered for irrigation. RESULTS AND DISCUSSION This report presents data from the first 18 months of the study, July 03 – December 05. Table 1 shows the monthly volumes of injection and recovery in each well. Figure 1 compares the household water consumption of mains water with net volume recharged as an indicator of the relative potential contribution of harvested water to household water demand. Over the 18 month period a total volume of 235 KL was injected and 18 KL was recovered, leaving a net increase in storage of 217 KL.

Recovery consisted mostly of intermittently purging the well to remove any accumulated particulate matter. This water was discharged into a nearby sump and soaked into the ground as additional recharge to shallower alluvium, along with a small amount of runoff from paved areas adjacent the house. The salinity of recovered water in this first year of operation was too high to use the banked water for garden irrigation in summer. In the first year the overall ratio of injected water to mains water usage was 40%. It was observed that during the wet winter months injection rates approximated mains usage rates, however in the drier summer months, when water demand peaks, there is little runoff, showing the value of having substantial storage to meet summer demand (Figure 1). Daily standing water levels have been monitored since July 2002 using a combination of loggers and manual measurements (Figure 2). During injection periods it has been observed that the head only rose by about 1 metre (with injection rates typically at 0.3L/s). Drawdown during recovery events has been logged with the pressure transducers during pumping tests. Purging events in the 100 mm well, were run at up to 1 L/s for 10 minutes, with the well never running dry. Pumping at the same rate in the lower yielding well (125 mm) drawdown rapidly reached the level of the pump at 18 m. Recovery events in the 2nd year of operation are now run at a lower average rate of 0.3 L/s for 35 minutes, including intermittent short periods of pumping at 1 L/s. The annual groundwater level fluctuation is approximately 1 m, with the peak in December and the lowest levels in June, approximately a 3-4 month lag on seasonal rainfall fluctuations. In spite of rainfall being near average, surprisingly there is a downward trend in groundwater levels that eclipses the effect of the increased recharge from the ASR system. Both wells record the same level when there is no recharge or recovery.

Table 1. Monthly water volumes

In jection Volum e

(KL)

Recovered Volum e

(KL)Rainfa ll

(m m )Mains (KL)

July 11.56 1.12 42.50 20.39August 22.38 0.46 85.30 22.29Septem ber 15.03 0.58 59.40 20.29O ctober 12.87 0.29 62.40 19.51Novem ber 1.66 0.57 7.40 29.62Decem ber 7.31 0 35.20 38.50January 3.24 0 15.00 40.77February 1.25 0 5.10 54.32M arch 7.80 0.66 29.30 33.77April 2.86 0.58 18.70 23.26M ay 21.94 0.58 74.60 24.39June 34.60 0.58 128.20 26.67July a 23.51 9.20 101.60 33.29August 25.18 1.48 90.00 21.13Septem ber 18.89 0.86 65.60 18.93O ctober 1.06 0.63 16.20 22.76Novem ber 19.18 0.60 74.70 22.76Decem ber 4.04 0.60 17.40 36.57

Total Volum es July 03 - Dec 04 234.3 18.8 928.6 509.2Volum es 100 m m well 154.4 12.0 606.0 376.0Volum es 125 m m well 79.9 6.8 322.6 133.2

a conducted 'pum p-tests' on both wells 21/22 July04 & m oved pum p from 100 m m to 125 m m well

Figure 1. Net volume of recharge and mains water consumption

0

50

100

150

200

250

300

350

400

Jul-03 Sep-03 Nov-03 Jan-04 Mar-04 May-04 Jul-04 Sep-04 Nov-04 Jan-05

Vol

ume

(KL)

Mains water consumption

Net recharge (100 mm well)

Net recharge (125 mm well)

Electrical conductivity (EC) and temperature have been continuously logged in the ASR well for 18 months, using a YSI water quality monitoring sonde (Figure 3). During injection events in the first year of operation, the ambient EC (>4000 uS/cm) was diluted with fresh rainwater (<100 uS/cm). With regular events in winter the EC in the well was maintained below 1000 uS/cm. However once injection events become less frequent, as in summer and during pumping, the EC of water in the ASR well increased rapidly towards ambient conditions. Water recovered from the well also showed a similar pattern, with salinity increasing during recovery. Insufficient freshwater has accumulated around the well in this first year to create a buffer against the saline ambient groundwater. The temperature of ambient groundwater is 18˚ C, and during winter rainwater can be less than 10˚ C and in summer up to 25˚ C. There is sufficient recharge in winter to depress the average temperature of ASR well water by 5˚ C. Temperature returns to equilibrium more quickly than salinity due to the thermal mass of aquifer material. For the second year of operation the injection line was installed in the 125 mm well, though recovery rates were slower, the plume of fresh water around the well has still risen to the irrigation threshold of 2500 uS/cm (Figure 3) during purging events following periods of low injection.

Figure 2. Monitored daily standing water levels from July 2002 to January 2005

Figure 3. Electrical conductivity and temperature measurements in the ASR well

0

5

10

15

20

25

Jul-03 Sep-03 Nov-03 Jan-04 Mar-04 May-04 Jul-04 Sep-04 Nov-04 Jan-05

Tem

pera

ture

(deg

rees

C)

0

1000

2000

3000

4000

5000

EC

(uS

/cm

)

Temp 100 mm wellTemp 125 mm wellEC 100 mm wellEC 125 mm well

Irrigation threshold 2500 uS/cm

0

5

10

15

20

25

Jul-03 Sep-03 Nov-03 Jan-04 Mar-04 May-04 Jul-04 Sep-04 Nov-04 Jan-05

Tem

pera

ture

(deg

rees

C)

0

1000

2000

3000

4000

5000

EC

(uS

/cm

)

Temp 100 mm wellTemp 125 mm wellEC 100 mm wellEC 125 mm well

Irrigation threshold 2500 uS/cm

11

11 .5

12

12 .5

13

13 .5

14Ju l-02 O ct-0 2 Ja n -03 A pr-03 Ju l-0 3 O ct-03 Jan -04 A pr-04 Ju l-04 O ct-0 4 Jan -05

SW

L fro

m g

roun

d le

vel (

m)

1 00 m m w e ll

1 25 m m w e ll

p um p m oved to 125 m m w e llIn jec tion in

100 m m w ell

In jec tion in

125 m m w ell

11

11 .5

12

12 .5

13

13 .5

14Ju l-02 O ct-0 2 Ja n -03 A pr-03 Ju l-0 3 O ct-03 Jan -04 A pr-04 Ju l-04 O ct-0 4 Jan -05

SW

L fro

m g

roun

d le

vel (

m)

1 00 m m w e ll

1 25 m m w e ll

p um p m oved to 125 m m w e llIn jec tion in

100 m m w ell

In jec tion in

125 m m w ell

Conductivity and temperature profiles have been carried out periodically through the trial using a YSI profiling sonde. In the first year of operation, the electrical conductivity and temperature profiles were measured 3 times in the initial observation well (125 mm) to detect any possible breakthrough and once in the ASR well (100 mm) when the pump was temporarily removed. The profiles show distinct stratification within the 24 m depth of each well with the fresher water overlying saltier water. This may explain why the salinity of recovered water increases so quickly. It is possible that the low salinity at the watertable in the observation well in September 2003 was due to injectant spreading laterally above the saline ambient water, however a pre-existing salinity gradient due to the presence of the nearby stormwater drainage sump cannot be ruled out. Since injection has occurred in the 125 mm well two sets of profiles have been carried out in each well, one near the end of winter and one mid summer. Here it is observed, that the water remains fresh (<500 uS/cm) above a depth of 20 m and below this salinity increase rapidly, reaching ambient values (>4000 uS/cm) at the base of the well (24 m). It is possible that by limiting the ASR well’s depth to 20 m, an improvement in recovery efficiencies could be observed. This will be assessed by performing down hole flow meter profiling, to establish the hydraulic conductivity profile of each well and allow development of strategies to enhance and maintain recovery efficiencies. Detailed water quality analyses in groundwater have not extended beyond the initial sampling in December 2002 from each bore before the start of injection and grab samples collected from the rainwater tank since injection commenced on 29 June 03. Table 2 assembles these data. The ambient groundwater is saline with low organic carbon and phosphorus but has significant nitrate. The rainwater is fresh with minimal nutrient levels but contained detectable zinc levels, from a galvanized steel roof. The leaf trap sample was taken in May 03 from a roof gutter where the down-pipe leaf guard had become choked with decaying leaves and water had been sitting in the gutter for some time. This sample contained coloured water with a musty odour and is taken to be a worst case for the potential for nutrient loading in the rainwater available for injection. A siphonic pipe system results in accumulation of detritus in the pipes so that rainwater discharging to the tank is sometimes depleted in oxygen.

Table 2. Water quality data 1 0 0 m m W e ll 1 2 5 m m W e ll L e a f t ra p

P a r a m e te r s (m g /L ) 1 2 /1 2 /0 2 1 2 /1 2 /0 2 1 9 /0 5 /0 3 2 1 /0 5 /0 3 1 7 /0 8 /0 3 1 /1 0 /0 3 1 2 /1 1 /0 4

E C (u S /c m ) 4 0 0 5 4 1 5 0 4 2 .3 2 1 .5 2 8 .2 2 3p H - f ie ld 6 .6 4 6 .6 3 7 .8 2 5 .8T e m p - f ie ld ( o C ) 1 9 .6 1 9 .8 1 6 .7D O - f ie ld 2 .1 7 2 .3 1 4 .3 2E h - f ie ld (m V S H E ) 3 6 5 3 7 2 6 0 9T u rb id ity (N T U ) 2 .1 2 0 .7 5 1 .0 5 1 .2T o ta l O rg a n ic C a rb o n 2 .0 2 .6 6 9 4 < 1 .0 2 .2 0 .8 3 .0C a lc iu m 1 5 1 1 4 9 5 7 1 .1 0 .7 0 .4 0 .4M a g n e s iu m 1 7 3 1 6 5 1 2 .4 0 .5 0 .4 0 .4 0 .3P o ta s s iu m 1 2 .1 1 2 .4 1 3 .1 < 1 < 1 < 1 < 0 .1S o d iu m 5 0 3 5 3 0 1 4 2 .9 1 .8 2 .3 2 .8B ic a rb o n a te 6 7 5 7 1 7 2 8 8 7 7 6C h lo r id e 9 9 0 1 0 0 0 2 0 7 9 8 4 .2S u lp h a te a s S 1 4 0 1 4 0 3 3 .1 < 1 .5 < 1 .5 < 1 .5 0 .1 a

A m m o n ia a s N 0 .0 2 7 0 .0 2 2 3 0 .1 0 .1 1 6 0 .3 4 7 0 .0 5 3P h o s p h o ru s - T o ta l a s P < 0 .0 2 5 < 0 .0 2 5 3 9 .7 0 .0 1 8 0 .0 2 6 0 .0 4 7 < 0 .1K je ld a h l N it ro g e n a s N < 0 .2 5 < 0 .2 5 2 7 5 0 .1 7 0 .3 9 0 .2N itra te + N itr i te a s N 5 .8 2 5 .8 7 0 .0 0 6 0 .2 0 2 0 .3 8 9 0 .0 7 5T o ta l N it ro g e n 0 .8T o ta l A rs e n ic < 0 .0 0 1 < 0 .0 0 1 0 .0 6 6 < 0 .0 0 1 < 0 .0 0 1T o ta l A lu m in iu m < 0 .1B o ro n < 0 .1T o ta l C o p p e r 0 .1 6 2 < 0 .0 0 1 0 .0 0 2 < 0 .0 2T o ta l I ro n < 0 .0 3 < 0 .0 3 1 4 8 < 0 .0 3 < 0 .0 3 < .0 3 < 0 .1T o ta l M a n g a n e s e < 0 .0 5T o ta l L e a d 3 .3 6 1 < 0 .0 0 0 5 0 .0 0 0 6T o ta l S tro n t iu m < 0 .0 0 3 < 0 .0 0 3 < 0 .1T o ta l Z in c 1 6 .9 9 0 .1 3 8 0 .2 4 8 0 .1 2 4 0 .0 8E .c o li ( /1 0 0 m L ) 0 0 2 4 0 0 7C o lifo rm s ( /m L ) 6 8 0 5 2 0 > 2 4 0 0 > 2 4 0 0F a e c a l C o li fo rm s ( /1 0 0 m L ) 7C o lo n y C o u n t (2 0 o C ) A e ro b ic ( /m L ) > 1 0 0 0 0 > 1 0 0 0 0

a a s S o n ly

R a in w a te rT a n k

CONCLUSIONS Operation and maintenance inputs required so far have been at a level beyond that which could reasonably be expected of normal householders. This is not unexpected for a first trial, and demonstrates the value of running trials to iron out problems before such methods are permitted, advocated or promoted. No clogging has been observed in the well, with the current regime of water quality management and with pumping out approximately 0.6 KL each month during months with moderate recharge. Initially quite turbid water is recovered on each cycle and discharged onto the garden or into the sump. This suggests that purging has real value in helping to prevent clogging. Salinity of recovered water has increased quickly beyond values acceptable for irrigation supplies in the first year of operation. The decision in July 2004 to assess the use of the 125 mm well, with its lower hydraulic conductivity as the ASR well, has to date resulted in fresher water on recovery and has not been restricted by the volume injected from rainfall events. The EC profile data suggests that density stratification could contribute to the low recovery efficiency. The trial is intended to run for three years to enable the objectives to be met. Results to date suggest some changes are required in order to achieve a viable ASR system. In the next 12 months it is proposed to obtain additional information and operational experience to define requirements for a viable system. ACKNOWLEDGEMENTS The monitoring reported in this paper was made possible through the support of the Patawalonga Catchment Water Management Board. Flow meters were provided by SA Water Corporation. The authors thank Paul Pavelic of CSIRO Land & Water for his assistance with pumping tests. REFERENCES Dillon P., Pavelic P., Barry K., Fildebrandt S. and Prawoto N. (2005). Nomograms to predict water quality improvement for managed recharge of aquifers (these proceedings). Dillon P.J. and Pavelic P. (1996). Guidelines on the quality of stormwater and treated wastewater for injection into aquifers for storage and reuse. Urban Water Research Assoc. of Aust. Research Report No 109. Dillon P., Pavelic P., Sibenaler X., Gerges N. and Clark R. (1997). Aquifer storage and recovery of stormwater runoff. Aust. Water & Wastewater Assoc. J. Water 24(4), 7-11. Dillon P., Toze S. and Pavelic P. (2004). Conjunctive use of urban surface water and groundwater for improved urban water security. Proc. IAH Congress Zacatecas, Mexico 11-15 Oct 2004. Hodgkin T. (2004). Aquifer storage capacities of the Adelaide region. South Australia Department of Water, Land and Biodiversity Conservation Report 2004/47. Patawalonga Catchment Water Management Board (2002). Catchment Water Management Plan 2002-2007. http://www.cwmb.sa.gov.au/patawalonga/plans/Pat01_Cover.pdf Pavelic P., Gerges N.Z., Dillon P.J. and Armstrong D. (1992). The potential for storage and re-use of Adelaide stormwater runoff using the upper Quaternary groundwater system. Centre for Groundwater Studies Report No. 40. South Australian Government (2004). Water Proofing Adelaide Draft Strategy: A thirst for change. http://www.waterproofingadelaide.sa.gov.au/pdf/Strategy_full.pdf