Elsevier

Environmental Modelling & Software

Volume 72, October 2015, Pages 117-125
Environmental Modelling & Software

Reliability of water supply from stormwater harvesting and managed aquifer recharge with a brackish aquifer in an urbanising catchment and changing climate

https://doi.org/10.1016/j.envsoft.2015.07.009Get rights and content

Highlights

  • Annual demand equating 12.8% catchment rainfall met with 99.5% reliability.

  • 7% less rainfall under climate change reduced supply by 3%.

  • 20% impervious area increase resulted in 8% more supply.

  • Increased supply volumes through urbanisation would offset effects of climate change.

  • Aquifer losses through mixing reduced supply by 10% due to short storage times.

Abstract

Recent research shows stormwater harvesting with Managed Aquifer Recharge (MAR) and complementary treatment can deliver safe potable water supplies. To address supply reliability the “WaterCress” hydrological model was used iteratively to simulate runoff, recharge and recovery for different rainfall, catchment and aquifer conditions, and operational scenarios based on the Parafield scheme in Salisbury, South Australia. Using historical rainfall and current catchment and operating conditions, annual demand equating to 12.8% of catchment rainfall could be met with 99.5% volumetric reliability. Using projected rainfalls from a high emission climate scenario resulted in a smaller harvestable volume decline than the increase expected from urban consolidation. Freshwater storage depletion in the brackish aquifer was expected to reduce the supply by 10% with 99.5% reliability compared with zero depletion. This simple generic modelling approach was useful for estimating reliability of stormwater MAR systems to assist planning and design and provide a basis for investor confidence.

Introduction

A stormwater harvesting and aquifer storage and recovery (ASR) system has been operated at Parafield Airport in Adelaide, South Australia, since 2003 by the City of Salisbury. Urban runoff from a residential and industrial catchment of 1590 ha has been diverted from a stormwater drain into an adjacent capture basin set below the drain invert level. High rate pumps lift the captured water into a holding basin from where it is released via a constant flow rate reedbed for treatment and injection into an underlying confined limestone aquifer. Injection and subsequent recovery has been via a combination of two ASR wells (where water is recovered from the injection wells), or a field of aquifer storage transfer and recovery (ASTR) wells where four injection wells surround two separate recovery wells. Recovered water is used for irrigation and industrial supplies and to dilute a separate recycled water supply for non-potable household use.

Urban runoff and stormwater system modelling has historically focussed on modelling runoff generation and performing flood risk analyses, modelling hydraulics and stormwater infrastructure, and stormwater quality modelling. The nature of these modelling approaches were examined by Zoppou (2001) and Elliot and Trowsdale (2007) reviewed and compared the features of a range of commonly used stormwater modelling codes. The stormwater recycling system modelling approach applied in this study differs from published studies in that stormwater hydrology was coupled with a subsurface storage component and customer demand and volumetric reliability of supply using different scenarios were determined through an iterative process. The modelling code chosen is similar in many ways to other water balance based hydrological models (Elliot and Trowsdale, 2007) but contains additional components to account for storage losses from the subsurface and collection of statistics from multiple simulations of each scenario.

MAR operations at Parafield have been studied extensively including for public health and environmental risk assessment, hydraulic modelling, operational management, impacts to infrastructure and water aesthetics, economics and public acceptance (Dillon et al., 2014). Much of this research addressed water quality assurance to support potable and non-potable supplies and this paper focuses only on the reliability of supply. Water utilities are generally required to meet a prescribed level of supply reliability. Gao et al. (2014) used a Monte Carlo technique accounting for variations in reservoir storage to model municipal water supply to determine the most cost effective supply configurations of groundwater replenishment with recycled water or desalination that met a 99.5% supply reliability. In this current study the relationship between water demand and reliability of supply was determined for the various scenarios. For potable supplies a 99.5% volumetric reliability standard was adopted, and for non-potable supplies, 95%. During a recent drought affecting most Australian capital cities, seawater desalination plants were built to secure water supplies. Levelised costs were considerably higher than the pre-existing average costs of water supplies, mostly from rural catchments within those cities, showing the high value placed on water security. Although urban stormwater is a climate-dependent source, impervious catchments are much less vulnerable to reduced runoff than pervious catchments during drought, and if coupled with aquifer storage, their reliability could be high.

The observed and projected drying climate for southern Australia reduces confidence in historical rainfall reliability as a guide to future water availability (Charles et al., 2003). In this paper daily rainfall was simulated based on historical records and also downscaled from a global circulation model (GCM) representing one of the more severe climate change scenarios. In growing cities the impervious area also grows through infill developments that may generate more runoff. The contrasting effects of more runoff from increased impervious area and reduced runoff from lower rainfall were compared. Furthermore the effects on the yield-reliability relationship were evaluated for several stormwater harvesting design and operating parameters, and for the rate of depletion of recoverable fresh water due to mixing in an originally brackish aquifer.

Section snippets

Modelling scenarios

A hydrologic model (WaterCress as described in Section 2.3) with a daily time step was used to:

  • Convert daily rainfall to stormwater runoff at the harvesting location at Parafield

  • Route the stormwater that can be accommodated through the harvesting system based on current and proposed designs and operating criteria for recharge to the aquifer

  • Account for the water balance in the aquifer, based on the sequences of seasonal recharge and recovery and depreciating the residual fresh water storage to

Hydrological insights

The functioning of the WaterCress model is revealed for a period when supply failure occurred. A four month period was selected over summer and autumn following a dry winter and a time series plot is shown for a single run of the model for the base case (scenario 1) with an annual demand of 960 × 103 m3 (61 mm or 13.1% catchment mean rainfall) (Fig. 5). The freshwater storage was already low at the start of summer when the demand for water was highest at 5.8 × 103 m3/day drawing down storage by

Conclusions and information gaps

This work demonstrated the volumetric reliability of supply of urban stormwater captured and stored in an initially brackish aquifer. The method for addressing freshwater storage depletion through mixing in the brackish aquifer was very simple but its utility here was based on ten years of data to enable calibration. Results revealed that as a percentage of mean annual rainfall over the catchment for the base case scenario, 18% became stormwater runoff at the harvesting point, 15% was

Acknowledgements

The authors acknowledge support of the partners to the Managed Aquifer Recharge and Stormwater Use Options research project. These are the National Water Commission through the Raising National Water Standards Program, the South Australian Government through the Goyder Institute for Water Research, CSIRO Land and Water Flagship, City of Salisbury, the Adelaide and Mt Lofty Ranges Natural Resources Management Board, University of Adelaide, University of South Australia, South Australian Water

References (30)

  • R.S. Crosbie et al.

    Differences in future recharge estimates due to GCMs, downscaling methods and hydrological models

    Geophys. Res. Lett.

    (2011)
  • P.J. Dillon

    Water security for Adelaide, south Australia

  • P. Dillon et al.

    Managed Aquifer Recharge Stormwater Use Options: Summary of Research Findings

    (2014)
  • Engineers Australia

    Australian Rainfall and Runoff. Stage 2 Revision Project 6. Loss Models for Catchment Simulation in Urban Catchments

    (2014)
  • L. Gao et al.

    The economics of groundwater replenishment for reliable urban water supply

    Water

    (2014)
  • Cited by (0)

    View full text