New coastal science for
South-east Queensland December 2004
www.coastal.crc.org.au
The Brisbane River and urban surroundings (Photo: Chris Roelfsema)
View of Lota Creek from key a sampling site, “The Jetty” on the southern end of the Esplanade, Lota, a coastal suburb of Brisbane (Photo: Eloise Larsen)
The impact of sewage overflows on an urban creek: a case study of Lota Creek in Brisbane
S
ewerage systems in Australia are designed with emergency overflow structures to discharge untreated sewage into local waterways in case of system overload or breakdown. Our aim was to determine the impacts of sewage overflows and potential risks to public and ecosystem health in the tidal waterways of the coastal suburb of Lota in Brisbane. Expectations for improved sewerage system performance are driven by the belief that sewage overflows cause significant environmental harm. However, our research shows in a large wet weather event, stormwater and not the overflow was the dominant stressor of ecological health. Even in the dry weather overflow event, impacts were likely to be low and restricted to the mixing zone at the point of overflow. However, sewage overflows, in dry and wet weather, did pose an unacceptably high public health hazard to potential swimmers in Lota Creek until the overflow stopped and there was a complete tidal exchange with the estuary. This Coastal CRC project has provided a sound scientific basis for quantifying overflow impacts and prioritising management of overflows and stormwater. It has also identified the need to develop programs to educate and raise community awareness, assess waterway use and potential pathways that can pose a risk to human health from wet and dry weather overflows both locally and into Moreton Bay.
Peter Pollard
*,1,2,3
1,4
1,5
1,3
, Rhys Leeming , Sam Bagraith , Margaret Greenway
* Corresponding author:
[email protected] 1 Coastal CRC 2 Centre for Riverine Landscapes 3 School of Environmental Engineering, Griffith University, Nathan, QLD 4111 4 CSIRO Marine Research, Hobart, Tasmania 5 Brisbane Water, TC Beirne Building, Fortitude Valley, Brisbane 6 CRC for Water Quality and Treatment, School of Civil and Environmental Engineering, UNSW-Sydney, NSW 2052
6
and Nicholas Ashbolt
The impact of sewage overflows to an urban creek: a case study of Lota Creek in Brisbane Copyright © 2004 by: Cooperative Research Centre for Coastal Zone, Estuary and Waterway Management
Written by: Peter Pollard* Rhys Leeming Sam Bagraith Margaret Greenway Nicholas Ashbolt Published by the Cooperative Research Centre for Coastal Zone, Estuary and Waterway Management (Coastal CRC). Images cannot be reproduced without permission of the copyright holder.
*Centre for Riverine Landscapes School of Environmental Engineering Griffith University Nathan campus, QLD 4111 Indooroopilly Sciences Centre 80 Meiers Road Indooroopilly Qld 4068 Australia www.coastalcrc.org.au
The text of this document may be copied and distributed for research and educational purposes with proper acknowledgement. Disclaimer: The information in this document was current at the time of publication. While the document was prepared with care by the authors, the Coastal CRC and its partner organisations accept no liability for any matters arising from its contents
National Library of Australia Cataloguing-in-Publication data The impact of sewage overflows to an urban creek: a case study of Lota Creek in Brisbane ISBN 1 921017 01 5 (print) ISBN 1 921017 02 3 (electronic data) QNRM05164
2
The impact of sewage overflows to an urban creek: a case study of Lota Creek in Brisbane
Peter Pollard Rhys Leeming Sam Bagraith Margaret Greenway Nicholas Ashbolt
Cooperative Research Centre for Coastal Zone, Estuary and Waterway Management BRISBANE
O
AUSTRALIA
3
At a glance, what we found
Legend No impact
The concerns and observations of this research are summarised in the following table.
Concern
Inconclusive
Low and local impact Unacceptably high Extremely unacceptably high
Observation Ambient Dry W eather Weather No Overflow
Stormwater Run-off No Overflow
Wet W eather with Weather Overflow
Dry W eather with Weather Overflow
Public Health (Swimabillity) Risk to public health from human faecal bacteria and viruses Loss of amenity for recreational activities Ecosystem Health Impacts of increased turbidity, nutrients: nitrogen, phosphorus and dissolved organic carbon and to the water column Reduced oxygen concentrations in the water column Toxicants Adding hormone disruptors to the water column Adding metals to the water column
What are sewage overflows? Sewerage systems in Australia are designed with wet weather overflow structures that discharge into local waterways when the capacity of the system is exceeded. Their purpose is to prevent untreated sewage backingup and flowing into people’s homes and onto private property. A dry weather overflow occurs when there is a system breakdown due to blockages or pump failure. Wet weather overflows are mainly caused by the infiltration of water into the sewerage system during heavy rainfall to a point where the capacity of the system is exceeded. Nearly half of the sewer pipe network in Brisbane is privately owned and water can enter the sewer through the illegal connection of roof and stormwater drains (inflows). The other
half is managed by Brisbane City Council where water also enters through poorly sealed access chambers, cracked pipes and defective joints (infiltration). The Coastal CRC has been working with Brisbane City Council to determine the impacts of sewage overflows on public and ecosystem health in a study of the waterways of the coastal suburb of Lota.
overflows that impact the aesthetics of the receiving waters are obvious. However, nutrients, pathogens, organic toxicants and metals also enter the waterway during an overflow event. The aim of this research was to assess the ecological and public health impacts and potential risks of sewage overflows in the Lota Creek catchment area (including Bowering Street tributary).
The public, generally, is unaware of overflows; their purpose or function. However, they are acutely aware of the gross pollutants associated with them. The appearance of litter in our rivers during heavy rain can stir the emotions of even the most placid of Brisbane City residents. These physical signs of wet weather
This was a case study in an eastern coastal suburb of Brisbane. While this was an intensive study of a single wet-weather event, there is sufficient hydrological detail coupled to public and ecosystem health indicators for the information and management options to be translated into other, similar, coastal environments.
4
What did we do? The study was conducted in the lower Lota catchment waterways. The seven overflow structures in the study area (Planning Unit LT/010) are shown in the numbered yellow circles in Figure 1. Overflow monitors were installed in all overflow structures with alerts at SP 20 and O/F 717 (Figure 2) as described by Millar and Gardner (2002) and Millar et al., (2002). Samples were collected manually from the “T” sites and by autosamplers at “A” as shown in Figure 1 in four situations: O
O
O
O
Ambient dry weather event: No rain or overflow in the study area. Samples taken at high and low tide. Dry weather with overflow event: Equipment failure as the only cause of overflow. Wet weather with overflow event: Water infiltration of sewerage network was the only cause of overflow. Wet weather without overflow event: Stormwater not influenced by sewerage overflow. Rain event
Figure 1. The seven overflow structures in the study area (Planning Unit LT/010) are shown in the numbered yellow circles. The circled numbers represent the codes used to identify the overflow structures (O/F = Overflow; SP = Sewage pump station). Due to heavy rain (60 mm.d-1 on 15 May 2003) sewage filled the wet well at SP 20. It backed up behind SP20 and overflowed at O/F 717. “T” shows manual sampling sites, “A” shows autosampler sites.
large enough to cause a wetweather overflow event. We first characterised the pollutants in the sewage of Lota to determine its physical, chemical, toxicological and microbiological character (Pollard and Chapman, 2002). The analytes included: organic and inorganic nutrients, sterol biomarkers, microbial faecal indicators, pathogens indicators (bacteria, viruses,
Overflow structure
Figure 2. Cross-sectional diagram of overflow structure, sewer and alert/monitoring system on flap-valve.
5
protozoa), toxicants, 69 metals, exotic chemicals, radioisotopes and endocrine disrupters. Although the project was not a study of the stormwater impacts, we looked for the pollutants common to both to distinguish the impacts of the overflow during the wet-weather event. The next stage of the project determined the ambient (normal) water-quality of Lota’s waterways in relation to the pollutants found in the untreated sewage. This was the most cost-effective approach. CSIRO Mathematics and Information Service designed a sampling strategy based on the “weight of evidence” for ambient and wet-weather sampling (Harch and Toscas, 2002). As wetweather events were infrequent (one in two years), repetitive event sampling was not possible. The seven overflow structures were monitored for two years. Of these, six had no impact on the local waterway because they did not overflow.
Pathogen indicator analysis is expensive. We found the most costeffective approach was to test samples for faecal coliform1 contamination during the event at all locations and times. Faecal coliform tests were completed within 10 hours so that areas (samples) of greatest contamination were identified. The more expensive and time-consuming analyses, such as for faecal sterols, enterococci, clostridium and coliphage were then used as confirmation tests for the presence of human faecal contamination.
potential risk to ambient public health caused by the overflow events. Human faecal contamination in the waterway was used to assess the public health hazard and potential risk assessment based on pathogenic indicators and faecal sterol biomarkers. Quantitative microbial risk assessment (QMRA) was then based on the WHO (2003) microbial water-quality guidelines. The enterococci abundance was used in the risk assessment model to determine potential human public health risks.
Knowing the ambient water-quality of the receiving waterway, we were able to compare changes in the
Risk assessment requires knowledge of how the waterways are used, e.g. swimming. While we observed
children swimming and canoeing in the study area, a quantitative assessment of the use (exposure to risk) of the waterway was beyond the scope of this study. Two different systems were used for assessing public and ecosystem health impacts. The dose-response research for QMRA is well advanced compared to that of ecosystem health assessment. Hence the public health hazard is well quantified and based on key pathogen surrogates. However, to carry out dose-response studies for assessing ecosystem 1 Faecal coliform is the colloquial name for thermotolerant coliforms
Table 1. Research* summary.
Concern
Observation Ambient Dry W eather Weather No Overflow
Stormwater Run-off No Overflow
Wet W eather with Weather Overflow
Dry W eather with Weather Overflow
Faecal coliforms high but low risk from human fecal contamination
Unsafe for recreation during overflow, even Extremely high as no though 80% of the faecal stormwater or in-pipe coliform count was due dilution to stormwater runoff and was not of human origin
Faecal coliforms high but low risk from human fecal contamination
High during overflow
Extremely high
Unacceptably High
Low and restricted to point of release, stormwater runoff was the main contributer
Low and confined to BST close to point of release
Low in some locations
Low in some locations
Not reduced: physical reaeration rates of the water column were greater than microbial respiration rates
Little reduction: physical re-aeration rates of the water column were greater than microbial respiration rates
Not measured
Inconclusive (below detection limit)
Based on hydrological dilution - may have impact
Adding metals to the water column
None
Not measured
Below detection limit however based on hydrological dilution may have impact in BST Below trigger values for aquatic ecosystem health
Likely cumulative effects
None
Low based on dry weather re-suspension of sediments
Public Health Risk to public health from human enteric bacteria and viruses
None
Loss of amenity for recreational activities
None
Ecosystem Health Impacts of increased turbidity, nutrients: nitrogen, phosphorus and dissolved organic carbon and to the water column
Reduced oxygen concentrations in the water column Toxicants Adding hormone disruptors to the water column
None
Low and localised, dilution puts all below trigger values outside BST Low based on dry Low based on dry weather re-suspension of weather re-suspension of sediments sediments
*Stormwater impacts were only related to contaminants found in the untreated sewage overflow BST = Bowering Street tributary Enteric = of the intestine Faecal coliform = Thermotolerant coliform
6
health would require waterways to be dosed with a range of pollutants to assess their impacts. Therefore we compared changes to ambient physical and chemical conditions caused by the overflow and pollutants exceeding trigger values recommended by ANZECC (2000) and BCC water-quality guidelines and objectives for aquatic ecosystem protection. This study took into account dry- and wet-weather overflows focusing on the impacts and hazards. Table 1 summarises the concerns and observations from the research.
What did we find? Overflow events The cause of the dry-weather overflow was the loss of power to the main pump station (SP 20) at Hindes Street Lota (power pole collapsed) on 4 June 2002 (Figure 1). Sewage backed-up in the sewer pipe network behind the wet well at SP 20 and overflowed at Coolana Street (O/F 717) at 08:10 when untreated sewage entered the Bowering Street tributary (BST) at a rate of approximately 65 m3.h-1 for 5 hours (0.33 ML). The overflow coincided with the falling tide in the receiving waters of Bowering Street. At low tide the volume of water in the Bowering street tributary was 2.6 ML (neap tide) and 5.6 ML (spring tides) on the high tide. Details of the hydrological tidal influence are described by Hargraves (2002), as part of the urban rainfall run-off model task of the project. The overflow stopped at 13:30 on 4 June 2002. The untreated sewage was diluted 1:6 with the tributary water at the
Overflow pipe Toilet paper
Figure 3. Picture of untreated sewage leaving the overflow structure (O/F 717) shows the impact on visual recreation quality.
overflow site. The first obvious sign of untreated sewage was toilet paper in the waterway, an obvious loss of the visual amenity (Figure 3). The wet-weather overflow occurred in May 2003. Untreated sewage overflowed into the Bowering Street tributary and Lota Creek for 36 hours because of a rainfall event of some 60 mm in two sharp bursts over 12 hours. Sewage effluent (1.1 ML) was released into the Bowering Street tributary where it mixed with 53 ML of stormwater run-off. A “mixing zone” immediately in front of the overflow structure was difficult to define because of tidal movement.
This study showed the plume of untreated sewage did not readily disperse in the receiving environment. Compared to untreated sewage in dry weather, the sewage was diluted 1:3 (v/v) in the sewer at the start of the overflow and this increased to 1:6 (v/v) for the in-pipe dilution. The dilution rate of the untreated sewage varied as it entered the Bowering Street tributary from 1% to 5.5% (v/v) on 15 May 2003. Stormwater and the tidal prism contributed to the dilution of the sewage as it was released; however, the tide contributed only 10% (v/v) of the total volume.
Public health risks Microbial water quality assessment Details of the microbial water quality results can be found in Pollard and Leeming (2002; 2003) and Pollard et al., (2004). Generally, during the dry and wet weather overflow event the faecal indicators (faecal coliform, E. coli, Clostridium, enterococci and coliphage) increased several orders
Figure 4. Faecal coliform (cfu.100mL-1) at the time of peak sewage overflow during the dry weather event 4 June 2002). The public health risk remained unacceptably high during and 12 h after the overflow had stopped. The movement of the sewage plume into Moreton Bay needs investigating.
7
of magnitude above those observed under ambient conditions and increased above public health guidelines for primary contact (swimming). Based on the faecal sterol biomarker studies, all of the faecal contamination (faecal coliforms) was of human origin during the dry-weather overflow event (Figure 4). During the wetweather overflow, stormwater contributed 80% of these indicators, the remainder being of human origin (Figure 5 a, b and c). However, the unacceptable public health hazard remained during the wet-weather overflow. The dry- and wet-weather overflows produced a sewage plume that did not readily mix with seawater. Risk to public health was of greatest concern following the initial overflow. For areas affected, the unacceptably high risk to potential swimmers and loss of recreational amenity persisted until the overflow stopped and there was a complete tidal exchange. The plume contaminants were exported beyond the limits of Lota Creek Study area into Moreton Bay (Figures 4 and 5).
Figure 5. Most (80%) of the faecal coliform count was associated with the stormwater run-off and was not of human faecal origin. The distribution of faecal coliforms (cfu.100mL-1) during the overflow on 15 May 2003; (A) morning (start), (B) afternoon (peak) and (C) on 16 May 2003 in the evening (after 36 hours) is shown. The sewage plume moved into Moreton Bay and these impacts need investigating.
8
Water quality guidelines are often based on faecal indicators that cannot separate human faecal contamination from that of other warm-blooded animals. Hence stormwater run-off often causes waterways and coastlines to exceed these standards (Nobel et al., 2003) and in tropical waters may result from growth and erosion of faecal coliforms (thermotolerant coliforms) and enterococci in soils (Fujioka et al., 1999). We have found that much of these breaches may not be due to human faecal matter based on faecal sterol analysis, during an overflow. A lesser public health hazard remains
from potential bacterial and parasitic protozoan pathogens.
Quantitative microbial risk assessment (QMRA) QMRA is a means of identifying a potential public health risk. For Lota Creek, the human risks are from exposure to faecal pathogens by recreational swimmers and people eating local shellfish. However, there is no known shellfish harvesting in the area. QMRA identifies the groups most at risk and defines the pathways. The selected reference pathogens suggested for the Lota Creek study showed viral risks were highest, as previously seen for Sydney’s coastal beaches (Ashbolt et al., 1997). To provide a ‘second opinion’ on the risk estimates, an estimate of potential gastrointestinal disease was obtained by assuming that only 20% of the enterococci in the bathing waters came directly from sewage (the primary material expected to contain infectious human enteric viruses). This was based on the faecal sterol work that showed most of the faecal indicators were not of human origin. This gave expected infection risks in the range of 10% to 100% indicating the waterway was unsuitable for recreation for at least the duration of the overflow (36 hours) with a further 24 hours before the waterway returned to background. For the dry weather event this unacceptably high risk would have persisted during the overflow, and for 12 hours after the overflow had stopped. Both approaches suggest that the risks to bathers during overflow events during both high rainfall and in dryweather should be considered unacceptably high. One management option could be signs erected to warn
swimmers of risks during these (rain) periods. The areas most affected are the Bowering Street tributary, lower reaches of Lota Creek and into Moreton Bay (depending on the direction of the sewage plume). How far the plume extended and persisted as a potential hazard in Moreton Bay and the bathing area at Wynnum and Manly was beyond the scope of this study, but warrants investigating (Figures 4 and 5).
Ecological health Details of the abiotic water quality can be found in Pollard and Leeming (2002, 2003) and Pollard et al., (2004). During the wet-weather overflow event, inorganic (e.g. nitrogen and phosphorus) and organic nutrients and suspended solids in the overflow effluent were rapidly diluted in the sewerage network and in the waterway. Concentrations were either below those associated with a healthy aquatic ecosystem and/or below those of the stormwater concentrations. Possible adverse impacts of chemical and physical stressors on ecosystem health were due primarily to stormwater run-off and not the overflow. Dissolved oxygen (DO) concentrations in the water column are often the basis of judging ecosystem health. As DO is ultimately governed by the growth rates of heterotrophic bacteria, we were interested in their abundance and growth. Lota Creek and the Bowering Street tributary had high numbers of bacteria and viruses with the bacteria growing extremely fast compared to ambient normal conditions.
9
The organic carbon released from the overflow increased heterotrophic bacterial growth rates in the waterways downstream of the overflow above ambient concentrations. This growth parameter was the first to detect an influence of the overflow effluent on biological processes in the waterway. Despite this, there was no deterioration in the water column’s dissolved oxygen due to the overflow event. The rate of water column physical re-aeration was likely to be always greater than the rate at which the bacteria used up the dissolved oxygen. During the dry-weather overflow, nutrient water quality parameters (nitrogen and phosphorus) in the Bowering Street tributary below the overflow site were at concentrations 10 fold higher than the water quality objectives of Brisbane City Council. However, once these nutrients reached Lota Creek, they were diluted by more than a factor of 100 and approached the ambient water quality concentrations and objectives of Brisbane City Council for Lota Creek and ANZECC (2000) guidelines. Overall, the water quality changes were confined to the period of the overflow within 500 m of the overflow structure with little changes to the dissolved oxygen concentrations due to the overflow itself. The Lota Creek and Bowering Street tributary results suggest that the ecological health of other tidally exchanged creeks can be determined in part by measuring the rate of the physical re-aeration of the water column and the rate of sewage dilution. These need to be in
conjunction with knowledge of the nutrient-loading rate in the waterway that could potentially drive the heterotrophic bacterial growth rates in the sewage to extremes. Some metals (aluminium, copper and zinc) were of concern in the sewage before in-pipe dilution. However, the combined in-pipe and stormwater dilution left them below ANZECC trigger values in wet-weather. Estrogenic hormones were found in the undiluted sewage and were monitored in the waterways of this study. However, they were below the detection limit of the assay i.e. 5 ng. L-1. The dilution of the estrone found in the untreated sewage suggested the concentrations could have ranged from 0.1 to 2 ng. L-1 in the Bowering Street tributary in the wet and
