Elsevier

Journal of Hydrology

Volume 544, January 2017, Pages 524-537
Journal of Hydrology

Research papers
Hydrologic impact of urbanization with extensive stormwater infiltration

https://doi.org/10.1016/j.jhydrol.2016.11.030Get rights and content

Highlights

  • Monitoring data and model used to analyze the hydrologic impact of urbanization.

  • Groundwater observations used to calibrate a model of urban stormwater infiltration.

  • Local stormwater infiltration increases groundwater levels throughout a catchment.

  • Urbanization reduces evapotranspiration from shallow aquifers.

  • Urbanization alters the water balance and can lead to increased groundwater levels.

Abstract

This paper presents a novel modeling analysis of a 40-year-long dataset to examine the impact of urbanization, with widespread stormwater infiltration, on groundwater levels and the water balance of a watershed. A dataset on the hydrologic impact of urbanization with extensive stormwater infiltration is not widely available, and is important because many municipalities are considering infiltration as an alternative to traditional stormwater systems. This study analyzes groundwater level observations from an urban catchment located in Perth, Western Australia. The groundwater observation data cover approximately a 40-year-long period where land use changes (particularly due to urbanization) occurred; moreover, the monitored area contains both undeveloped and urbanized areas where stormwater infiltration is common practice via soakwells (shallow vertical infiltration wells). The data is analyzed using a distributed and dynamic hydrological model to simulate the groundwater response. The model explicitly couples a soakwell model with a groundwater model so that the performance of the soakwells is reduced by the increase of groundwater levels. The groundwater observation data is used to setup, calibrate and validate a coupled MIKE SHE-MIKE URBAN groundwater model and the model is used to quantify the extent of groundwater rise as a result of the urbanization process. The modeled urbanization processes included the irrigation of new established private and public gardens, the reduction of evapotranspiration due to a decrease in green areas, and the development of artificial stormwater infiltration. The study demonstrates that urbanization with stormwater infiltration affects the whole catchment water balance, increasing recharge and decreasing evapotranspiration. These changes lead to a rise in the groundwater table and an increase in the probability of groundwater seepage above terrain.

Introduction

Stormwater infiltration through infiltration trenches, swale-trench systems, infiltration areas, permeable pavements, rain gardens and bioretention basins is part of Water Sensitive Urban Design (WSUD) (Fletcher et al., 2015) which aims at improving urban stormwater management and building more livable and resilient urban environments that are adapted to future uncertainties caused by climate change (Wong and Brown, 2009).

Stormwater infiltration (rather than stormwater drainage pipe systems) increases groundwater recharge and groundwater levels (Jeppesen, 2010, Göbel et al., 2004) which can cause an increased risk of groundwater seepage above terrain, decreased infiltration capacity of stormwater infiltration systems, and an increase of water flowing into foundation drains and infiltrating sewer systems (Antia, 2008, Endreny and Collins, 2009). This is particularly important in areas with shallow water tables, where even modest increases in groundwater levels can cause the water table to rise to the ground surface, causing groundwater flooding.

The implementation of stormwater infiltration in an urban area is constrained by existing built environment and underground infrastructures, soil pollution, groundwater levels, local drinking water assets, and the quality of the stormwater runoff (Göbel et al., 2004, Mikkelsen et al., 1994, Revitt et al., 2003). It is important to predict the consequences of stormwater infiltration on subsurface flows and the interaction with underground structures (Dietz, 2007). This requires an understanding of the mutual interaction between infiltration devices and groundwater (Bouwer, 2002). To predict the impact of WSUD systems (WSUDs) hydrological models are often used (Elliott and Trowsdale, 2007).

Several studies have analyzed the impact of stormwater infiltration on infiltration rates, groundwater mounding and groundwater table focusing on the impact of single systems. Duchene et al. (1994) presented a numerical model based on Richards’equation to simulate infiltration from infiltration trenches. Bouwer (2002) described the main processeses affecting the performance of different infiltration systems, and Guo (1998) presented a steady-state surface-subsurface model to design trench infiltration basins. Dussaillant et al. (2004) presented a three layer model with subsurface flow described by Richards equations for designing rain gardens, while Browne et al., 2008, Browne et al., 2013 presented a one-dimensional and a two-dimensional model to calculate infiltration rates from stormwater infiltration systems and Thompson et al. (2010) used the software HYDRUS-2D to predict the water-table mounding beneath infiltration basins. Carleton (2010) simulated the effect of local stormwater infiltration on local groundwater mounding. Roldin et al. (2013) presented a model to predict infiltration rates and groundwater mounding from single soakaways. Bergman et al. (2011) and Warnaars et al. (1999) reported the performance of observed infiltration trenches in an urban area, and Machusick et al. (2011) presented the groundwater observations from an experimental stormwater infiltration basin on a shallow unconfined aquifer

The larger scale impact of widespread stormwater infiltration has been examined by several authors. Markussen et al. (2004) modeled the impact of rain gardens on a 50 ha catchment in Denmark, and concluded that annual combined sewer overflow volume could be reduced by 75–85%. Roldin et al. (2012a) modeled the impact of widespread stormwater infiltration through soakaways on a 3 km2 catchment in Denmark and concluded that annual combined sewer overflow volume could be reduced by 24–68%; Petersen et al. (1994) modeled theoretical catchments with stormwater infiltration; Xiao et al. (2007) modeled a real neighborhood scale and Peters et al. (2007) modeled the city scale. These studies showed that stormwater infiltration can reduce stormwater runoff and combined sewer overflows, and that the efficiency of infiltration systems is greatly influenced by the soil characteristics.

Several studies have modeled the behaviour of infiltration units in urban areas. Manglik et al. (2004) presented a model to simulate the aquifer response to time varying recharge and pumping from multiple basins and wells. Elliott et al. (2009) analyzed the effect of aggregation of single storm-water control devices in an urban catchment model. Antia (2008) presented a model to predict the infiltration rates, mounding and seepage zones resulting from the interaction between multiple infiltration devices; he also reported cases in the UK where infiltration devices caused unexpected downstream flooding. Endreny and Collins (2009) used the software MODFLOW to predict the impact of different spatial arrangements of infiltration units on stormwater recharge and groundwater mounding for a 8 ha catchment in New York, they predicted a water table rise of 1.1 m after 30 years of recharge through bioretention basins. Maimone et al. (2011) used the software DYNFLOW to predict the effect of widespread stormwater infiltration in Philadelphia. Ku et al. (1992) modeled the impact of widespread stormwater infiltration basins on Nassau County, New York and concluded that recharge would increase by up to 12% and the groundwater table would rise up to 1.5 m compared to a traditional scenario without stormwater infiltration. Jeppesen (2010) modeled the impact of widespread storm water infiltration in Copenhagen concluding that it would lead to a surface near groundwater table and require increased groundwater drainage. Göbel et al. (2004) modeled the effect of stormwater infiltration on the groundwater table of an 11.5 km2 semi-urban catchment. Kidmose et al. (2015) modeled the impact of stormwater infiltration scenarios on groundwater levels and stream flows in Silkeborg in Denmark. Barron et al., 2013a, Barron et al., 2013b quantified the impact of urbanization and stormwater infiltration on the hydrological water balance in the Southern River catchment in Western Australia.

This paper addresses two gaps in the literature. Firstly, while some studies have used models to predict the hydrologic impact of urbanization with widespread stormwater infiltration, none of the models have been calibrated against observation data from an urbanized catchment with stormwater infiltration. In fact, data sets including groundwater level observations in a large catchment undergoing urbanization with extensive stormwater infiltration are not currently available in the literature. Secondly, there is a lack of studies where shallow groundwater levels affect the infiltration from WSUDs, and the two-way coupling between groundwater and the infiltration devices is explicitly modeled (Roldin et al., 2012a). Most of the published literature simulates catchment scale stormwater infiltration without explicitly modeling the interaction of groundwater levels with the stormwater infiltration systems (Maimone et al., 2011, Göbel et al., 2004, Barron et al., 2013a, Barron et al., 2013b).

This paper presents the development of an urban hydrologic model that was calibrated and validated using groundwater observation data from an urban area where widespread stormwater infiltration has been practiced for decades and which has been subject to major land use changes. The data set is from a subcatchment in Perth, Western Australia, where groundwater level data has been collected since 1960. Since that time, the area has been progressively urbanized, and so the dataset is ideal for analyzing the hydrologic impact of urbanization with stormwater infiltration. The area is particularly interesting because of the widespread implementation of direct stormwater infiltration. Groundwater flooding is a problem in the area, and there is a suspicion that its incidence has increased as a consequence of urbanization and stormwater infiltration. The aim of this paper is to analyze the catchment scale hydrologic impact of urbanization with stormwater infiltration. This was done by fitting an integrated hydrological model to a groundwater observations dataset from Perth and by simulating the groundwater response during different land use scenarios. The model was used to analyze (1) the changes in groundwater levels, (2) the changes in the hydrologic balance and (3) the performance of soakwells in areas with a shallow groundwater table. The novelty of this paper is the use of groundwater head observations from a large catchment with gradual urbanization including widespread stormwater infiltration. The study illustrates some of the long term issues arising from urbanization that may lead to increased urban stormwater infiltration and changes in evapotranspiration and irrigation supplied to new developed areas.

Section snippets

The study area

The study area is located in the south of Perth (Western Australia) and covers an area of about 112 km2 (Fig. 1). The area is mainly flat with topographical heights ranging between 0 and 35 m above Australian Height Datum (AHD) and is characterized by mild terrain slopes (<0.5%).

The area has a Mediterranean climate with very dry summers and wet winters. The mean annual rainfall was 852 mm in the period 1980–2014. However, the south-west of the Western Australia region has experienced up to 20%

Hydrological model

This study uses the software MIKE SHE-MIKE URBAN (DHI, 2012a, DHI, 2013) to set up a physically-based distributed hydrological model. The MIKE SHE tool simulates the processes of interception, ponding, evapotranspiration, overland flow, infiltration and unsaturated and saturated flow. Streamflow is computed by MIKE 11 (DHI, 2012b). The MIKE URBAN model was used to simulate the water level and the infiltration processes of soakwells. The rainfall falling on rooftops is diverted into soakwells

Model calibration and validation

Table 2 shows the fixed, the dependent (‘Tied’ in the table) and calibrated model parameters and their values. The values of the calibrated hydraulic conductivity and storage coefficients for the sandy layer of the superficial aquifer (Layer 2) are similar to those of Barr and Barron (2009).

Table 3 shows the RMSE, MAE, SRMS and ME for the calibration and validation period and Fig. 6 shows the ME for selected wells in the calibration period. Fig. 7 shows the simulated and observed time series of

Single event response

The model was designed for long term simulation of changed land use and captures annual trends, but does not reproduce the dynamics of single events. Fig. 12 shows an example of simulated and observed heads for a number of rainfall events. The daily groundwater observations collected at the CoG and SR wells (Fig. 3c) clearly show short term groundwater responses driven by the rainfall, with sharp peaks and fast groundwater recessions that were not fully captured by the model. Daily groundwater

Conclusions

This study presents a calibrated and validated hydrological model to analyze the catchment scale impact of urbanization with stormwater infiltration on groundwater levels and water balance and the performance of soakwells in the presence of a shallow groundwater. The novelty of this paper is the use of groundwater head observations from a large catchment with gradual urbanization including widespread stormwater infiltration.

Different land use scenarios were implemented to analyze the hydrologic

Acknowledgments

The authors thank the Danish Council for Research that financed the present research through the project BIV (Byer i Vandbalance/Cities in water balance) and the Department of Water in Perth, particularly Krish Seewray, Joel Hall and Belinda Quinton for providing the data and comments on the paper.

References (61)

  • M. Roldin et al.

    A simplified model of soakaway infiltration interaction with a shallow groundwater table

    J. Hydrol.

    (2013)
  • E. Warnaars et al.

    Hydrologic behaviour of stormwater infiltration trenches in a central urban area during 2 3/4 years of operation

    Water Sci. Technol.

    (1999)
  • A.D. Allen

    Outline of the hydrogeology of the superficial formations of the Swan Coastal Plain: Western Australia Geological Survey

    Ann. Rep.

    (1976)
  • H.E. Anderson et al.

    Applied Groundwater Modeling

    (1992)
  • D.D.J. Antia

    Prediction of overland flow and seepage zones associated with the interaction of multiple infiltration infiltration devices (cascading devices)

    Hydrol. Process.

    (2008)
  • AutoCal, 2012. Auto Calibration Tool; User Guide. Dhi Water & Environment. In: Agern Allé 5, Hørsholm,...
  • A. Barr et al.

    Application of a Coupled Surface Water-Groundwater Model to Evaluate Environmental Conditions in the Southern River Catchment

    (2009)
  • O.V. Barron et al.

    Urbanisation and shallow groundwater: predicting changes in catchment hydrological responses

    Water Resour. Manage

    (2013)
  • M. Bergman et al.

    Evaluation of two stormwater infiltration trenches in central Copenhagen after 15 years of operation

    Water Sci. Technol.

    (2011)
  • M. Botte et al.

    A decade of sustainable stormwater management approaches – do we need to go back to basics?

  • H. Bouwer

    Artificial recharge of groundwater: hydrogeology and engineering

    Hydrogeol. J.

    (2002)
  • D. Browne et al.

    A new saturated/unsaturated model for stormwater infiltration systems

    Hydrol. Process.

    (2008)
  • D. Browne et al.

    A two-dimensional model of hydraulic performance of stormwater infiltration systems

    Hydrol. Process.

    (2013)
  • G.B. Carleton

    Simulation of Groundwater Mounding Beneath Hypothetical Stormwater Infiltration Basins

    (2010)
  • R.S. Crosbie et al.

    A time series approach to inferring groundwater recharge using the water table fluctuation method

    Water Resour. Res.

    (2005)
  • W.A. Davidson

    Hydrogeology and Groundwater Resources of the Perth Region Western Australia

    (1995)
  • Department of Water

    Wungong Brook Catchment Area Drinking Water Source Protection Plan

    (2007)
  • Department of Water, 2004–2007. Stormwater Management Manual for Western...
  • Department of Water, 2009. Strategic policy 2.03 – Managing unlicensed groundwater...
  • Department of Water, 2013. Water resource considerations when controlling groundwater levels in urban...
  • Cited by (94)

    • Analyzing the impact of hydrological storage and connected impervious area on the performance of distributed kerbside infiltration systems in an urban catchment

      2022, Journal of Hydrology
      Citation Excerpt :

      A large timestep can compromise the ability to simulate the runoff peak flow rate (Ball et al., 2019). Another problem with prior modelling-based impact assessment studies is that the majority has not documented calibration procedures for their catchment models containing WSUD elements against observed outflows (Locatelli et al., 2017). Moreover, few impact assessment studies based on hydrological modelling approaches have evaluated the ability of models to represent WSUD systems with an objective statistical framework (Roldin et al., 2012).

    View all citing articles on Scopus
    View full text