RESEARCH PAPER

 

The prospects for stormwater harvesting in Cape Town: Part 2 - catchment-scale managed aquifer recharge with stormwater

 

 

John Okedi; Neil Philip Armitage

Department of Civil Engineering, University of Cape Town, Private Bag X3, Rondebosch 7701, Cape Town, South Africa

Correspondence

 

 

---

ABSTRACT

The City of Cape Town in South Africa faced the possibility of taps running dry in 2018 due to a prolonged drought that commenced in 2015. With such droughts expected to reoccur frequently in future, this study investigated the prospects for managed aquifer recharge (MAR) with stormwater. This would require temporary storage to collect and hold the stormwater during and immediately after rainfall events while it seeps into the aquifer. The 89 km² Zeekoe Catchment located in the southern part of Cape Town was selected as a case study as it had both existing surface storage (61 stormwater ponds) and was lying above a large unconfined aquifer. As the stormwater ponds were largely designed for flood control, they would need to be modified for MAR. In this desktop study, the main objective was to model temporary detention of stormwater in the ponds with the aim of predicting infiltration into, and thus augmentation of, the aquifer. The requirement that the flood control function be maintained, combined with the limited capacity in the ponds, was a key consideration. The study determined that the physical characteristics in the Zeekoe Catchment, i.e., largely flat terrain, pervious sandy soils, and a relatively deep (20-50 m) unconfined aquifer, could support managed aquifer recharge and borehole abstraction rates of 3.5-8.1 L/s per borehole from some 140 boreholes. This could provide a mean annual groundwater yield of 29-33 Mm³ (about 15% of Cape Town water demand in 2018).

Keywords: catchment scale; stormwater harvesting; managed aquifer recharge; Cape Town

---

 

 

INTRODUCTION

In the past 30 years since the 1990s, a water cycle management approach aimed at environmental protection has emerged, variously called, inter alia, water sensitive urban design (WSUD) in Australia, low impact development (LID) in the United States and sustainable drainage systems (SuDS) in the United Kingdom (Fletcher et al., 2014). The approaches are based on a more holistic water cycle management philosophy that aims to minimise net outflow of water from an urban catchment (Fletcher et al., 2014). Wong (2007) notes that cities are potential catchment areas in their own right which, if well managed, would meet a substantial proportion of their water needs. Marsden and Pickering (2006) determined that the mean cost per kilolitre of SWH was lower than many other sources, including, inter alia, seawater desalination, rainwater harvesting and long-distance pipelines. The focus of this study was to investigate the prospects for catchment-scale stormwater harvesting using a groundwater aquifer in Cape Town, South Africa, for medium to long-term storage. The City of Cape Town (CCT) already has plans to implement groundwater extraction from the Table Mountain Group (TMG) and Cape Flats Aquifer (CFA). If not managed well, this could potentially result in severe and irreversible environmental impacts, such as ground subsidence and saltwater intrusion in coastal catchments. In Mexico City, excessive abstraction of groundwater over a long period of time (since the 1950s and greatly increased in the 1980s) has resulted in subsidence of 0.4 m/year since 1984, with the total subsidence reaching 8 m in some areas by 2010 (Ortiz-Zamora and Ortega-Guerrero, 2010). The proposed introduction of stormwater into the CFA could mitigate the negative effects of groundwater extraction - essentially replicating the historic water cycle in the area where little rainwater reached the ocean but rather seeped into the underlying aquifers.

This study investigated the potential for catchment-scale SWH in CCT. Since CCT covers 2 445 km² with more than 800 stormwater ponds, the study was restricted to the 89 km² Zeekoe Catchment situated on the Cape Flats in the southern part of the city. The stormwater harvesting technique investigated involved underground storage in the local aquifer using 'managed aquifer recharge' (MAR). The technique is coupled with various demand scenarios: (i) non-potable water for use within the catchment including urban agriculture, public open spaces, domestic gardens, and toilet flushing; (ii) potable water for general use with treatment provided by small 'package' plants situated within the catchment and injected directly into the existing reticulation system; and (iii) potable water for general use with treatment provided at one of the nearby bulk water treatment plants where the stormwater is blended in with the water coming from the surface reservoirs.

 

LITERATURE REVIEW

Managed aquifer recharge (MAR) has been implemented in many countries, such as Australia (Dillon et al.,2009; Page et al., 2009; Miotlinski et al., 2014), the United States of America (USA) (Murray et al., 2007), Namibia (Murray et al., 2007; Tredoux et al., 2009), and South Africa (Bugan et al., 2016). In Australia, MAR has been implemented in Perth, Adelaide, and Melbourne, with aquifer storage capacities of 250 Mm³, 80 Mm³ and 100 Mm³ respectively (Dillon et al., 2009). Some other examples of MAR projects in Australia include Salisbury near Adelaide where stormwater is treated in a wetland and injected into the aquifer; and the Burdekin Delta in North Queensland where 45 Mm³ of water is recharged and abstracted for irrigation of sugarcane and other crops; (Dillon et al., 2009; Page et al., 2009; Miotlinski et al., 2014). Some examples of MAR in the USA include Peace River in Florida and the Kerrville in Texas (Murray et al., 2007). The Peace River and Kerrville schemes comprise injection of treated water into the groundwater aquifer and recovery of about 68 000 m³/day and 9 500 m³/day, respectively. In Namibia, MAR is a water resource for the city of Windhoek, with artificial recharge from the Von Bach Dam and reclaimed treated wastewater injected into the Auas aquifer with a yield of 2-8 Mm³/year (Murray et al., 2007; Tredoux et al., 2009). In South Africa, the Atlantis Water Resource Management Scheme (AWRMS) is an MAR system with about 7 500 m³/day of stormwater and wastewater is infiltrated into the aquifer to boost the groundwater resource by more than 2.7 x 10⁶ m³/year (Bugan et al., 2016; DWA, 2010). The system was established to supply water to the town of Atlantis located 50 km north of the Cape Town CBD on the west coast, which was not originally linked to the city reticulation system (DWA, 2010).

 

METHOD

Site selection

In this study, the Zeekoe Catchment (Fig. 1) was chosen from the various catchments in Cape Town as it has many stormwater ponds (some 61 ponds) and is located over the Cape Flats Aquifer. The area has a flat, sandy terrain while the aquifer is unconfined with thickness ranging from 20-50 m and considerable potential for groundwater abstraction to meet CCT demand.

Selection of the infiltration model

PCSWWM was used to model surface to groundwater transfer in the stormwater ponds, which had been assumed to be transformed into infiltration basins that would be ordinarily dry until temporarily filled with rainwater (Okedi and Armitage, 2026; Okedi, 2019). Infiltration in PCSWMM can be represented by Horton, Green-Ampt or Curve Number methods (James et al., 2010). The selection of the Horton method for this study area was based on the best match, with field-measured infiltration rates with the assistance of a double ring infiltrometer (DRI) in accordance with the ASTM D3385-09 at 3 sites across the study area (Fig. 1 and Table 1) (Mavundla et al, 2025).

Shallow surface core-samples (300 mm) were also retrieved from the sites and tested in the laboratory in accordance with the ASTM D2216-10 to determine the saturated hydraulic conductivity, bulk density, volumetric water content, porosity, saturation, residual water content, particle density, and particle-size distribution analysis (Table 2). Figure 2 shows the particle size distributions.

The results show similarities in the soil particle distribution for all the selected ponds across the study area. Furthermore, other characteristics, such as porosity and coefficient of uniformity and curvature, specific gravity and natural moisture content were also similar. These similarities justify the reliance on a limited number of test sites to provide general infiltration parameters for the study area. The infiltration rates measured with the DRI experiments at the three ponds were then compared with the values estimated with the Green-Ampt and Horton methods to determine the most appropriate approach to be used in the model.

In Horton's method, the decay of infiltration rate with time is expressed by the exponential relationship presented in Eq. 1 (Horton, 1933).

where: f is infiltration rate at any time t (cm/h); f_o is initial infiltration rate at t-0 (cm/h)f_c is final infiltration rate (after equilibrium at steady state) at t = t_c (cm/h); λ is Horton's decay coefficient, which depends on soil characteristics and vegetation cover (h^-1).

The Horton parameters for the three sites are presented in Table 3.

In the Green-Ampt method, the determination of infiltration rate is based on Darcy's law using Eq. 2 (Green and Ampt, 1911).

where: f is infiltration rate (cm/h); F is cumulative infiltration (cm); k is hydraulic conductivity (cm/h); S_c is capillary suction at the wetting front (cm); γ is porosity of the soil (%).

The infiltration rates measured in the DRI experiments were compared with values estimated using the Green-Ampt and Horton methods and plotted in Fig. 3. The statistical descriptors of the parameters in Green-Ampt method needed for model development were determined from the experiment and are presented in Table 4.

As shown in Fig. 3, both the Green-Ampt and Horton models visually fit the measured data reasonably well, except in the case of Pond 3. The Nash-Sutcliffe efficiency (NSE) and correlation (R²) coefficients were used to determine which method provided a better match (Table 5).

Both methods provided reasonably good results, i.e., above 0.5, except NSE for Green-Ampt in Pond 2 and all cases in Pond 3.

However, while the plots in Fig. 3 show that the Green-Ampt method better represented the initial infiltration values, the Horton method provided a better estimation ofthe final infiltration rates. The final selection of Horton method for modelling the infiltration component of the model was based on its superior performance over the longer time periods that might better represent the conditions in stormwater ponds used for aquifer recharge.

The modelling first determined the catchment-scale infiltration component, i.e., the portion of rainfall that is transferred directly to the groundwater aquifer and thus not directly contributing to runoff. The enhanced surface to groundwater transfer was then modelled through the simulation of extended detention in the existing stormwater ponds to enhance the infiltration there. In the model, the stormwater ponds had to be represented as infiltration sources - and not merely temporary storage. The closest available option is 'bio-retention', which allows temporary storage combined with high infiltration rates through a vegetated surface followed by stretches where the surface was allowed to dry out. The model thus considered the stormwater ponds as bio-retention cells on the assumption that they would function in a similar manner. A typical bio-retention cell is composed of 3 horizontal layers, i.e., surface, soil and storage layers, with an underdrain at the bottom as shown in Fig. 4 (Brown et al., 2011):

• Surface layer - the top section of the bio-retention cell that receives both direct rainfall and runoff from the catchment. The modelling of water balance in this section is based on the continuity equation (Eq. 3) (James et al., 2010). The surface layer properties are specific to the bio-retention cells' geometric characteristics, i.e., surface area and depth (consistent with the stormwater pond shape) and vegetation cover (100% of the pond surface area).

where: is fraction of freeboard above the surface not filled with vegetation; is depth of water in the surface layer (mm); i is rate of precipitation falling directly on the surface layer (mm/h); q_o is inflow to the surface layer from runoff' captured from other areas (mm/h); e_l is surface layer evapotranspiration rate (mm/h); is infiltration rate of surface water into the soil layer (mm/h); is surface layer runoff' or overflow rate (mm/h).

• Soil layer - the middle section of a bio-retention cell generally consists of an engineered soil mixture with organic matter and a thickness of 450-900 mm (James et al., 2010). For the model development and calibration, however, the site-specific soil data presented in Table 2 were used, i.e., porosity (0.30-0.44), field capacity (15.6-17.6%), wilting point (4.4-5.2), conductivity (4.8-19.9 cm/h), conductivity slope (9.7-9.9 cm/cm) and suction head (5.9-114.5 cm). Modelling of the water balance in the soil layer used Eq. 4.

where: D₂ is thickness of the soil layer (mm); 0₂ is soil layer moisture content (fraction); f₁ is infiltration rate of surface water into the soil layer (mm/h); e₂ is soil layer evapotranspiration rate (mm/h); f₂ is percolation rate of water through the soil layer into the storage layer (mm/h).

• Storage layer - the bottom of a bio-retention cell generally consists of crushed stone or gravel with thickness from 150-450 mm. In the model development, site-specific data were used, i.e., void ratio (0.47-0.78) as shown in Table 2 and filtration rate (2-10 cm/h) as shown in Fig. 3. The modelling of the storage layer was based on Eq. 5:

where: Φ₃ is voids fraction of storage layer (fraction); d₃ - depth of water in the storage layer (mm); f₂ is percolation rate of water through the soil layer into the storage layer (mm/h); e₃ is storage layer evapotranspiration rate (mm/h); f₃ is exfiltration rate of water from the storage layer to native in-situ soil (mm/h); q₂ is layer runoff or overflow rate (mm/h).

Various values were allocated to the surface, soil, and storage layers based on the field and laboratory tests and supplemented by recommendations from various publications (e.g. James et al., 2010; Brown et al., 2011). The following assumptions were made in the modelling:

• Surface layer - the plan area was assumed to be constant for the entire depth, the inflow was assumed to be uniformly distributed over the entire surface area, and water movement inside the bio-retention cell was assumed to be one-dimensional in the vertical direction.

• Soil layer - the moisture content was assumed constant throughout the soil layer.

The modelled mean annual water balance values in PCSWMM for evaporation, evapotranspiration, surface runoff and infiltration for the existing land uses with an average catchment imperviousness of 45% are presented in Fig. 5.

As shown in Fig. 5, there was significant infiltration even before the modelling of extended detention in the stormwater ponds. This can be attributed to large sections of the study area having rural/farmland characteristics where natural recharge takes place. Furthermore, the study area possesses physical characteristics that support natural infiltration, including, inter alia, sandy soils (pervious) and reasonably flat terrain. However, with population growth and the associated land-use change, natural 'greenfield' areas are being converted to impervious surfaces in the absence of sustainable drainage systems (SuDS), and hence extended detention in ponds would be important in enhancing surface to groundwater transfer.

Modelling groundwater flow field

In this study, generalised groundwater flow field due to abstraction, and also pollution transport, was modelled from first principles using the hybrid optimization approach presented in Mahinthakumar and Sayeed, (2005; 2006) following the steps presented in Fig. 6.

 

 

The approach combines genetic algorithms (GA) with local search methods to solve the groundwater flow equations that were further developed with assistance of the principal author. They solve the two-dimensional steady-state partial differential equation with a time-step component commonly known as Richards' equation (Richards, 1931) (Eq. 6).

where: v_x and v_y are velocity (flux) in x-direction and y-direction, respectively; K(x, y) is hydraulic conductivity in 2 dimensions; dh is hydraulic head; dx and dy are spatial steps in x and y directions, respectively.

In the model, Eq. 6 was represented in a finite difference form (Eq. 7) with discrete nodes defined along a grid covering the study area. The hydraulic head (h) was represented in both the 2-dimensional spatial domain and time interval, i.e., (hi+1,j,t, hi,j,t, hi_−1,j,t); (h_i,j+1,t, 2h_i,j,t, h_i,j−1,t); (h_i,j,t, h_i,j,t−1) and estimated using algebraic equations containing finite differences and values from nearby points.

The simulation was initially undertaken on a trial section (1.44 km² with a single pond) and then finally at the catchment scale (89 km² with 61 ponds). Likely groundwater abstraction rates and suitable locations of the boreholes relative to the infiltration basins were estimated through a particle tracking model of the groundwater flow paths in the aquifer. The aim was to ensure that the stormwater ponds were the origin of the groundwater flow fields.

To minimise simulation run-time, discrete nodes were placed in the centre of 1 x 1 km resolution squares of constant size covering the entire catchment, as shown in Fig. 7. To further reduce simulation run-time, the catchment was subdivided into 4 rectangular domains / computational areas (CA) (labelled CA1, CA2, CA3 and CA4 in Fig. 7) that approximated the extent of the four discrete study areas. The groundwater abstraction modelling aimed to determine the appropriate number of abstraction boreholes (N_w), distance of the boreholes from stormwater ponds (D_p), and abstraction rates (Q_w) for the study areas that would maximise the quantity of water abstracted from the supplementary groundwater resource, assumed to be equal to the stormwater transferred to the aquifer through the infiltration process, i.e., SWH. Furthermore, the managed extraction of groundwater from the vicinity of the ponds should hopefully reduce the risk of subsidence resulting from a general drop in the groundwater table.

Modelling water quality improvement within the aquifer

It was also desirable that the aquifer provides sufficient improvement of the quality of the stormwater, to minimise the subsequent treatment process. One significant advantage of storing stormwater underground is that aquifers act, to a large extent, as sand filters. There is improvement through infiltration, adsorption (the process whereby pollutants bind to the surface of fine sand particles), biodegradation and volatilisation (the conversion of some compounds to gases or vapour). There is a trade-off between maximising the quantity of the harvested stormwater by extracting from the regions closest to the ponds and enhancing the water quality improvement for pathogens by extending the flow path and residence time. The pollution decay associated with groundwater transport from the stormwater pond to abstraction boreholes was thus assessed using the first-order relationship presented in Eq. 8 (Delleur, 2007) for E. coli (an indicator organism for faecal pollution) using a decay rate informed by the published values listed in Table 6.

where: C_t is concentration or quantity at time t; C_o is initial quantity at the start of assessment (t = 0); λ is pollution decay rate (day^-1). The units of C_t and C_o depend on the pollution.

Equation 8 is a simplification of the process, as E. coli removal typically depends on various factors including, inter alia, the availability of nutrients in the water, the exposure to UV radiation and temperature (Delleur, 2007). Nevertheless, the simplification was deemed adequate for the study, as the goal was to demonstrate the potential of water quality improvement from the process of stormwater recharge and recovery.

Determination of borehole distribution and determination of abstraction rates

The ideal borehole site selection and determination of the recommended abstraction rates was based on the following criteria:

• Visual inspection of the particle tracks provided by the model was used as guidance to ensure that the groundwater flow paths originated from stormwater ponds.

• The mean groundwater levels are generally deepest at the end of the dry summer, then rise to the surface in many areas with the natural recharge from winter rainfall. Physical data on the seasonal groundwater level fluctuations provided the basis for the setting of the initial conditions and hydraulic heads.

• The location of wells and groundwater abstraction rates were set such that the interference between the various drawdown curves was minimised.

• The total abstraction quantity was made approximately equal to the anticipated infiltration with the modified ponds as estimated in PCSWMM.

• The retention time of the water in the aquifer was kept at around 1 year to ensure that pathogenic organisms, indicated by E. coli, would be largely eliminated.

 

RESULTS AND DISCUSSION

Edith Stephens

Initially, a stormwater recharge and recovery model was made of a trial section incorporating only one stormwater pond - the Edith Stephens Wetland (Figs 8 and 9). Examples of the borehole model simulation are presented in Fig. 10.

 

 

 

 

An initial abstraction rate of 1.2 L/s per borehole was implemented in the model (Fig. 10 left). With this, all flow fields originated from the Edith Stephens Wetland. The borehole abstraction rates were then increased stepwise until 5.8 L/s (Fig. 10 right), which was determined to be the optimal abstraction rate with the flow fields visually observed to originate mainly from the pond. When the abstraction rate was increased above 5.8 L/s per borehole, the flow fields increasingly originate from areas beyond the stormwater pond.

A decay rate of 0.025 - the smallest value of those listed in Table 6 - was selected as a conservative value associated with slow organism inactivation and prolonged survival times, i.e., worstcase conditions. Simulations were undertaken in the model with Eq. 8 to determine transport and decay of the E. coli as an indicator organism, using the estimated abstraction rate of 5.8 L/s per borehole. The model was run with an initial mean value of 12 000 CFU/100 mL at the pond, based on the monthly grab samples collected by CCT from various locations in Edith Stephens Wetland over the period 2006-2017 (Fig. 11).

 

 

The very high initial E. coli values are consistent with major pollution sources, such as on-site sanitation, upstream of the Edith Stephens Wetland, and direct discharge of grey- and blackwater into the drainage channel from informal settlements. Since the grab samples were not collected at regular intervals (i.e., the sample collection date in the month was inconsistent, and some values were missing), the data could only provide an indication of the contamination in the stormwater drainage and values for modelling purposes. The results of the water quality modelling are presented in Fig. 12. The results from the trial section with a single pond (Edith Stephens Wetland), using E.coli as an indicator organism for faecal pollution, showed that the sandy aquifer in the study area had the potential to remove very high levels of E.coli, i.e., from between 1 x 10³ and 1 x 10⁵ counts per 100 mL to values below 1 count per 100 mL (Fig. 12).

The South African National Standard for drinking water (SANS 241-1:2015, SABS, 2015) requires zero E.coli count; however it was evident from the model that stormwater harvested from groundwater storage would be suitable for potable water uses with minimal additional disinfection treatment.

Zeekoe Catchment

The findings from the trial section model and additional information from literature were then used to develop the catchment-scale model. Vandoolaeghe (1989) had determined that a total of 10 Mm³/yr could be abstracted, with 27 boreholes each pumping at an abstraction rate of 12 L/s from the CFA, while Fraser et al. (2001) suggested a total groundwater yield of 18 Mm³/yr in the same area. Abstraction rates of 6 L/s per borehole were determined to be most suitable, with higher values potentially extending the groundwater cone of depression to the coastline and resulting in possible seawater intrusion (DWA, 2008). In a more recent study, 6 scenarios were assessed, consisting of 3 arrangements of 9, 18 and 27 boreholes with abstraction rates of 3 L/s and 5 L/s (Mauck, 2017). One of the key aims of that study was flood mitigation through drawdown of the water table to values lower than a pre-determined threshold of 1.5 m below the surface through groundwater abstraction (Mauck, 2017). Mauck (2017) determined that an abstraction rate of 3 L/s would not draw down the water table to below the 1.5 m threshold for flood mitigation in all three borehole arrangements. With the borehole pumping rates increased to 5 L/s, the simulated groundwater drawdown exceeded the 1.5 m threshold only 5% of the time for the 18 boreholes and completely for the 27 boreholes (Mauck, 2017).

The various recommended abstraction rate values were used to guide the borehole abstraction rates in each of the four rectangular computational areas (CA), CA1, CA2, CA3 and CA4 presented in Fig. 7. Four boreholes were randomly placed around the stormwater ponds in each of the computational areas and initially simulated using an abstraction rate suggested in Fig. 13.

 

 

The number of boreholes and abstraction rates were then adjusted to ensure that the flow fields originated from the nearest stormwater ponds (i.e., harvesting 'stormwater' from the ponds). The assumption was that infiltration rates were not diminished by high water tables, i.e., abstraction rates lowered groundwater levels to sufficient depth such that the pond floor was above capillary fringe (i.e., the zone of saturated soil directly above the water table where groundwater is pulled upwards from the saturated zone by capillary action). A summary of the key parameters, the final optimised modelled number of boreholes, the abstraction rates per borehole and the mean annual groundwater yields for each computational area, are presented in Table 7.

The modelling suggested that the likely optimal abstraction rates per borehole range from 3.5-8.1 L/s if the flow was to be restricted to that from the ponds. When the abstraction rates were increased beyond these values, the origin of the groundwater flow fields was increasingly from outside the pond area, i.e., not a consequence of MAR at the ponds.

A further assessment was undertaken to determine the likely water quality improvement associated with stormwater recharge and recovery. The CCT collects grab samples from various locations in the study area indicated in Fig. 14.

 

 

A theoretical assessment of stormwater quality improvement, with E.coli as the indicator organism for pathogens, was undertaken for CA1, CA2, CA3 and CA4 based on the CCT data, using Eq. 8 to estimate the likely decay with time. The modelled values for CA1 (Fig. 7) with 10 stormwater ponds and 20 abstraction boreholes suggest that the sandy aquifer in the study area has the potential to remove very high levels of E.coli, i.e., between 1 x 10^-6 counts per 100 mL and less than 8 counts per 100 mL. In CA1 and CA2, E.coli counts were detected at boreholes after about 100 days in the model - with the count increasing rapidly over a 200-day period. The counts stabilised in the range of 1 to 8 counts per 100 mL at about 350 days (about 1 year). In CA3 and CA4, E.coli counts were detected at boreholes after about 200 days in the model, and increased rapidly over a 400-day period before stabilising after about 600 days (a little less than 2 years).

In summary, with the South African National Standard for drinking water (SANS 241-1:2015, SABS, 2015) requiring a zero count of E.coli per 100 mL, the findings from the catchment-scale model show that the stormwater harvested from groundwater storage would theoretically be adequate for potable water uses with minimal additional disinfection treatment. The abstracted water would not require additional treatment for non-potable water demands, i.e., irrigation of urban agriculture, public open parks and residential gardens. Lim et al. (2015) reported similar results, i.e., that microbial pollution in stormwater from groundwater storage was significantly reduced to levels where the water could be directly used for some indoor residential needs with a limited level of contact, e.g., machine washing and toilet flushing. Vanderalm et al. (2010) showed that stormwater recovered from an aquifer after a mean residence time of 240 days was suitable for non-potable water applications. Continuous monitoring would be necessary while post-recovery disinfection and aeration for iron removal might be required.

 

CONCLUSIONS

In this study, it was determined that the Zeekoe Catchment has the potential for relatively high borehole abstraction rates compared with other areas in Cape Town. The mean annual natural infiltration for the 89 km² was estimated to be in the range of 20-21 Mm³. With the 61 stormwater ponds available in the study area adapted to function as bio-retention cells, the mean annual infiltration could likely increase the groundwater resource to 29-33 Mm³. The actual additional groundwater resource due to stormwater infiltration could thus be 9-12 Mm³. The impact of land-use change was also assessed with a hypothetical future general catchment imperviousness of 75%. It was determined that this could decrease the natural mean annual infiltration volume to 10-13 Mm³. Managed aquifer recharge with stormwater to enhance groundwater augmentation could, however, increase the groundwater resource to about 21 Mm³. The results from modelling various potential groundwater abstraction scenarios in the Zeekoe Catchment show that, depending on the aquifer parameters - i.e., conductivity, porosity and aquifer depth - suitable borehole pumping rates for later abstraction typically range from 3.5-8.1 L/s for the anticipated 140 boreholes in the 89 km² catchment. MAR provides water quality improvement benefits. The study area contains several informal settlements (slums, shanty towns), that generate wastewater and litter discharges into the drainage channels, particularly in the upper reaches of the catchment, and the CCT monthly grab samples of stormwater quality showed that the drainage system in the study area is highly impacted by pollution. MAR could substantially improve the water quality. A preliminary assessment suggested that a residence time of about a year should provide die-off of pathogens in the abstracted water to values less than 10 E.coli counts/100 mL.

 

AUTHOR CONTRIBUTIONS

John Okedi was responsible for the collection of the data, the construction and running of the various models, the analysis of the model outputs, and the writing of the draft paper. Neil Armitage was responsible for the conceptualisation of the project, critical intellectual input during the research, and the final editing of the paper.

 

REFERENCES

ARMITAGE N, FISHER-JEFFES L, CARDEN K, WINTER K, NAIDOO V, SPIEGEL A, MAUCK B and COULSON D (2014) Water sensitive urban design (WSUD) for South Africa. Framework and Guidelines. WRC Report No. TT 588/14. Water Research Commission, Pretoria.         [ Links ]

ARMITAGE N, VICE M, FISHER-JEFFES L, WINTER K, SPIEGEL A and DUNSTAN J (2013) Alternative technology for stormwater management: The South African guidelines for sustainable drainage systems. WRC Report No. TT 558/13. Water Research Commission, Pretoria. ISBN: 9781431204137.         [ Links ]

BOSKIDIS I, PISINARAS V, PETALAS C and TSIHRINTZIS V (2012) Monitoring and modeling of two alluvial aquifers in lower Nestos river basin, Northern Greece. J. Environ. Sci. Health Part A Toxic/ Hazardous Substances Environ. Eng. 47 (12) 1849-1868. https://doi.org/10.1080/10934529.2012.689552        [ Links ]

BUGAN RDH, JOVANOVIC N, ISRAEL S, TREDOUX G, GENTHE B, STEYN M, ALLPASS D, BISHOP R and MARINUS V (2016) Four decades of water recycling in Atlantis (Western Cape, South Africa): Past, present and future. Water SA 42 (4) 577-594. https://doi.org/10.4314/wsa.v42i4.08        [ Links ]

BROWN R, HUNT WF and KENNEDY SG (2011) Designing bioretention with an internal water storage ( IWS ) layer. Design guidance for an innovative bioretention feature. AG-588-19W. North Carolina Cooperative Extension, NC State University, Raleigh.         [ Links ]

CCT (City of Cape Town) (2012) City of Cape Town GIS Database: General land use layer shapefile. Strategic Information Department, City of Cape Town Metropolitan Municipality, Cape Town.         [ Links ]

CCT (City of Cape Town) (2015) City of Cape Town GIS Database: Water quality data and monitoring points. Strategic Information Department, City of Cape Town Metropolitan Municipality, Cape Town.         [ Links ]

CCT (City of Cape Town) (2017) Annexure 4. Revised Consumptive Tariffs, Rates and Basic Charges for Electricity Services, Water Services and Waste. City of Cape Town Metropolitan Municipality, Cape Town.         [ Links ]

DELLEUR JW (2007) The Handbook of Groundwater Engineering (2^nd edn). CRC Press, Boca Raton.         [ Links ]

DWAF (Department of Water Affairs and Forestry, South Africa) (2008) The assessment of water availabilty in the Berg Catchment (WMA19) by means of Water Resources Related Models (Vol. 5). Department of Water Affairs and Forestry, Pretoria.         [ Links ]

DWA (Department of Water Affairs, South Africa) (2010) The Atlantis Water Resource Management Scheme: 30 years of artificial groundwater recharge. Department of Water Affairs, Pretoria.         [ Links ]

ENGELBRECHT JFP (1998) Groundwater pollution from cemeteries. Proc. WISA Biennial Conference, 4-7 May 1998, Cape Town: Water Institute of Southern Africa, Midrand.         [ Links ]

FLETCHER T, SHUSTER W, HUNT WF, ASHLEY R, BUTLER D, ARTHUR S, TROWSDALE S, BARRAUD S, SEMADENI-DAVIES A, BERTRAND-KRAJEWSKI J-L, MIKKELSEN PS, RIVARD G, UHL M, DAGENAIS D, and VIKLANDER M (2014) SUDS, LID, BMPs, WSUD and more - The evolution and application of terminology surrounding urban drainage. Urb. Water J. 12 (7) 525-542. https://doi.org/10.1080/1573062X.2014.916314        [ Links ]

FRASER L and WEAVER J (2000) Groundwater impact scoping for the Cape Flats Aquifer. CSIR, Pretoria        [ Links ]

GREEN WH and AMPT GA (1911) Studies on soil physics, 1: the flow of air and water through soils. J. Agric. Sci. 4 1-24. https://doi.org/10.1017/S0021859600001441        [ Links ]

HORTON R (1933) The role of infiltration in the hydrologic cycle. Trans. Am. Geophys. Union 14 (1) 446-460. https://doi.org/10.1029/TR014i001p00446        [ Links ]

JAMES W, ROSSMAN LE and JAMES RC (2010) User Guide to SWMM5 (13^th edn). CHI, Guelph. ISBN: 9780980885330.         [ Links ]

JOVANOVIC N, BUGAN RDH, TREDOUX G, ISRAEL S, BISHOP R and MARINUS V (2017) Hydrogeological modelling of the atlantis aquifer for management support to the atlantis water supply scheme. Water SA 43 (1) 122-138. https://doi.org/10.4314/wsa.v43i1.15        [ Links ]

KALBACHER T, DELFS J-O, SHAO H, WANG W, WALTHER M, SAMANIEGO L, SCHNEIDER C, KUMAR R, MUSOLFF A, CENTLER F, SUN F, HILDEBRANDT A, LIEDL R, BORCHARDT D, KREBS P and KOLDITZ O (2012) The IWAS-ToolBox: Software coupling for an integrated water resources management. Environ. Earth Sci. 65 (5) 1367-1380. https://doi.org/10.1007/s12665-011-1270-y        [ Links ]

LIM KY, HAMILTON AJ and JIANG SC (2015) Assessment of public health risk associated with viral contamination in harvested urban stormwater for domestic applications. Sci. Total Environ. 523 95-108. https://doi.org/10.1016/j.scitotenv.2015.03.077        [ Links ]

MAHINTHAKUMAR GK and SAYEED M (2006) Reconstructing groundwater source release histories using hybrid optimization approaches. Environ. Forens. 7 (1) 45-54. https://doi.org/10.1080/15275920500506774        [ Links ]

MARSDEN J and PICKERING P (2006) Securing Australia's urban water supplies: opportunities and impediments. Marsden Jacob Associates (Pty) Ltd, Victoria, Australia.         [ Links ]

MAUCK B (2017) Managed aquifer recharge potential for the Cape Flats Aquifer. PhD thesis, University of Cape Town.         [ Links ]

MAVUNDLA KP, OKEDI J, KALUMBA D and ARMITAGE NP (2025) Estimation of infiltration parameters for groundwater augmentation in Cape Town, South Africa. Hydrology 12 87. https://doi.org/10.3390/hydrology12040087        [ Links ]

OKEDI J (2019) The prospects for stormwater harvesting in Cape Town, South Africa using the Zeekoe Catchment as a case study. PhD thesis. University of Cape Town.         [ Links ]

OKEDI J and ARMITAGE NP (2026) The prospects for stormwater harvesting in Cape Town: Part 1 - transforming existing stormwater ponds into reservoirs using real time control. Water SA 52 (1) 35-48.         [ Links ]

ORTIZ-ZAMORA D and ORTEGA-GUERRERO A (2010) Evolution of long-term land subsidence near Mexico City: review, field investigations, and predictive simulations. Water Resour. Res. 46 1-15. https://doi.org/10.1029/2008WR007398        [ Links ]

PAGE D, VANDERZALM J, BARRY K, LEVETT K, KREMER S, AYUSO-GABELLA N, DILLON P, TOZE S, SIDHU J, SHACKLETON M and co-authors (2009) Operational residual risk assessment for the Salisbury stormwater ASTR project. Water for a Healthy Country Flagship Report Adelaide, South Australia. ISSN: 1835-095X.         [ Links ]

QUAYLE T (2012) Atlantis, South Africa - recycling wastewater to bolster ground water supply: the Atlantis story. ACCESSanitation Case Study, CSIR and City of Cape Town, South Africa. URL: http://www.awasla.org/Docs/yearCCESS_Case_study_Atlantis_SA.pdf.         [ Links ]

RICHARDS LA (1931) Capillary conduction of liquids through porous mediums. Physics 1 318-333. https://doi.org/10.1063/1.1745010        [ Links ]

SABS (South African Bureau of Standards) (2015) SANS 241-1:2015. South African National Standard SANS 241-1: 2015 Drinking Water Specification. SABS, Pretoria.         [ Links ]

SIDHU JPS, HODGERS L, AHMED W, CHONG MN and TOZE S (2012) Prevalence of human pathogens and indicators in stormwater runoff in Brisbane, Australia. Water Res. 46 6652-6660. https://doi.org/10.1016/j.watres.2012.03.012        [ Links ]

TOZE S (2006) Reuse of effluent water - benefits and risks. Agric. Water Manage. 80 (1-3) 147-159. https://doi.org/10.1016/j.agwat.2005.07.010        [ Links ]

VANDERZALM J, LEVETT K, PAGE D, DILLON P, TOZE S, MIOTLINSKI K, SIDHU J, BARRY K, ALEXANDER K., HYDE K and REGEL R (2010) Assessing the risks of recycling urban stormwater for potable supply via an aquifer. Illawarra Health and Medical Research Institute. 189. URL: https://ro.uow.edu.au/ihmri/189        [ Links ]

WONG T (2007) Water sensitive urban design - the journey thus far. Aust. J. Water Resour. 110 213-222. https://doi.org/10.1080/13241583.2006.11465296        [ Links ]

 

 

Correspondence:
Neil Philip Armitage
Email: neil.armitage@uct.ac.za

Received: 8 August 2024
Accepted: 8 December 2025