The estimation of microbial parameter removal was undertaken through the consecutive application of the percentage removal in each stage of the TTs to the initial stormwater composition with regards to key parameters. Due to the lack of available data pertaining stormwater in the case study area, the Werribee catchment in the west of Melbourne, the initial stormwater composition was assumed to be that given in the collation of literature data in the Stormwater Harvesting Guidelines [
5]. The 95th percentile statistical summary chemical water quality data from these Stormwater Harvesting Guidelines was compared to the ADWG values [
17], and the concentration of the parameters that exceeded or were close to the ADWG values were estimated: polyaromatic hydrocarbons (PAH), As, Cd, and Pb. Biological oxygen demand (BOD) removal and chemical oxygen demand (COD) removal were also estimated as general indicators of water quality.
2.3.1. PAH Removal
The vast majority of PAH content in stormwater is removed in the wetland and RO stages (TTA and TTB) or in the wetland and O3/BAF stages (TTC and TTD). The advanced oxidation (TTA and TTC), UF (TTC and TTD), and chlorination (all TTs) stages act as polishing stages.
Wetland treatment: The modelling in the current study used a PAH removal rate of 68% for the free-water-surface constructed wetland [
18]. The PAH removal attributed to constructed wetlands in the current modelling was based on a study conducted in China, Greece (35.3° north of the equator) [
18]. This study was selected due to the similarity in distance from the equator to Melbourne (37.5° south of the equator). The efficiency of the removal of polycyclic aromatic compounds (PAHs) was evaluated in a pilot-scale constructed wetland system combining a free-water-surface wetland, a subsurface wetland, and a gravel filter in parallel. The average PAH removal rates were 79.2% for the subsurface-flow constructed wetland, 68.2% for the free-water-surface constructed wetland, and 73.3% for the gravel filter, respectively.
RO treatment: The modelling used a conservative value of 80% for RO removal of PAH (benzo(a)pyrene) [
6]. Reverse osmosis has been described as one of the, if not the, most effective single-unit process for the removal of chemicals of concern in water treatment. It typically removes >90% and often >99% of wastewater organics, depending on the compound. The indicative RO removal of benz(a)pyrene (3,4-benzopyrene), a polyaromatic hydrocarbon (PAH), is given, in the AGWR—Augmentation of Potable Water Supplies [
6]—as >80%. The RO rejection of smaller polyaromatic hydrocarbons, naphthalene, anthracene, and phenathrene, was found to be 98–99% (USEPA 1987) [
34]. Similarly, the rates of RO removal of other polyaromatic hydrocarbons, such as acenanthrene, fluoranthene, naphthalene, and phenthrene, of 99%, 86%, 99%, and 99%, respectively, were reported [
35].
Table 3.
Literature treatment conditions.
Table 3.
Literature treatment conditions.
Treatment Train Stage | Chemical | Water Type | Key Conditions | Comment and Removal | Reference and Reference Type |
---|
Wetland | PAH | Stormwater | Feed concentration of 786 ± 514 ng/L; 42 m2 wetland; southern Greece (35°19__N and 25°10__E); mixed cultures of two species of reed, Phragmites australis and Arudo donax; mean water temperature: 12.1 and 34.1 °C | Free-water-surface constructed wetland; 68% removal | [18], experimental |
Wetland | Heavy metals | Stormwater | Various constructions and conditions | Typical removal: 30–60% removal, 30% used in modelling | [19], literature review |
Wetland | BOD, COD | Stormwater | No media, with floating, submerged, and emergent plants; continuous water supply | Free-water-surface constructed wetland; around 89% BOD removal and 72% COD removal | [11], literature review |
RO | PAH | Wastewater | Validation of specific application and operational conditions required | Indicative, >80% removal; intended to be informative and not to be used as the design basis for schemes | [6] (Table 4.10 of water recycling guidelines) |
RO | As(III) | Synthetic brackish water | Feed concentration between 0.36 and 1.2 mg/L As(III), spiked drinking water, Dow 5K membranes, tests at manufacturer operating specifications | 85% removal | [20] |
RO | Cd | Drinking water | Feed concentrations between 0.47 mg/L and 1.9 mg/L, spiked drinking water, Dow 5K membranes, tests at manufacturer operating specifications | 98% removal | [20] |
RO | Pb | Municipal and industrial wastewater | Laboratory and on-site pilot-scale tests, feed concentration of 1.5 mg/L, various conditions | 89 to 100% removal, 95% used in modelling | [21], experimental |
RO | COD, BOD | Hospital wastewater | Feed concentrations: 200–235 mg/L COD, 95–115 mg/L BOD; feed rate: 10–14 L/h; variable pressure to 13.6 Bar maximum; specific flux: 90–190 L/m2/h/bar | More than 99% removal for both COD and BOD | [22], experimental |
UV | PAH | Natural water | Feed concentration: 3.9 to 5.6 µg/L; 3 different PAHs; 3 different water matrices; fluence between 40 and 1500 mJ/cm2 | Negligible PAH removal at 40 mJ/cm2 | [23], experimental |
UV | COD, BOD | Raw and biotreated textile dye bath effluent | Feed concentrations: 760 mg/L COD, 261 mg/L BOD; UV dose: 5 mW/cm2; exposure time between 5 and 25 min | Negligible COD and BOD removal at 12 mJ/cm2, 35% BOD and 25% COD removal at 7500 mJ/cm2 | [24], experimental |
UV/H2O2 | PAH | Natural water | Feed concentration: 3 µg/L; 30 min contact time; UV radiation: 170 µW/cm2; 10 mg/L H2O2 | 99% removal at 306 mJ/cm2 | [25], experimental |
UV/H2O2 | COD, BOD | Raw and biotreated textile dye bath effluent | Feed concentrations: 760 mg/L COD, 261 mg/L BOD; UV intensity: 5 mW/cm2 and 254 nm; 150–200 mg/L H2O2 for raw wastewater, 100–150 mg/L for biotreated wastewater | Negligible BOD and COD removal at 800 mJ/cm2 Raw wastewater: 35% COD removal, 44% BOD removal at 7500 mJ/cm2 Biotreated: ~85% COD removal and ~90% BOD removal at 7500 mJ/cm2 | [24], experimental |
Cl2 | PAH | Wastewater | Various conditions | Indicative, >80% removal | [6] |
Cl2 | COD | Industrial wastewater | Feed concentration: 39 mg/L; chlorination after coagulation and flocculation; 1.2 mg/L free chlorine; 30 min contact time | 10% removal | [33], experimental |
Cl2 | BOD | Secondary-effluent wastewater | Various feed concentrations: 12–30 mg/L, 5 mg/L residual Cl2; 15 min contact time | 67% to 20% depending on starting concentration; conservative setting of 30% removal chosen for modelling | [32], experimental |
O3 + biological treatment | PAH | Contaminated water | Feed PAH concentration: ~5000 µg/L; 0.5 mg/L ozone; 30 min ozone treatment; 24 h biological treatment in flask | 91% PAH removal overall | [36], experimental |
BAC | PAH | Diesel and petrol Synthetic wastewater | Petroleum content: 5 mg/L; ~1100 µg/L PAH; 8 L BAC to 300 L contaminated water; aerobic conditions; 12–24 h contact time | 97% PAH removal | [27], experimental |
O3 + BAC | As | Groundwater | Feed As concentration: 14–27 µg/L; 43 min contact time | 99% removal | [28], experimental |
BAC | Pb, Cd | Wastewater | Feed Pb and Cd concentrations: ~ 200 µg/L; 50–150 mg/L activated carbon; 2 h contact time | 99% Pb and 86% Cd removal | [29], experimental |
O3 | COD, BOD | Secondary-effluent wastewater | Full scale, variable feed concentrations (~10–80 mgO2/L COD, ~2–10 mgO2/L BOD), 11–13 mg/L ozone | 8%–88% COD removal with most results in 10–20% removal range, ~0% BOD removal | [37], experimental |
O3 + BAF | COD | Surface water | Biological sand filter, full-scale plant, 30–60 min contact time, 17 mg/L O3 concentration | Two different plants: one achieved ~50% COD removal, the other, ~20% COD removal | [38], experimental |
O3 + BAF | BOD, COD | Textile effluent | Biological aerated filtration; feed COD ≤ 110 mg/L, BOD ≤ 30 mg/L; 20–25 mg/L ozone dose; 3.3 h hydraulic retention time; 6 air-to-water flow ratio | Approximately 64% COD removal, 67% BOD removal | [26], experimental |
UF | PAH | Biologically treated wastewater | Feed PAH concentration: 22–38 µg/L; 0.04 µm pore size | 67% removal | [30], experimental |
UF | COD, BOD | Hospital wastewater | Feed concentrations: 200–235 mg/L COD, 95–115 mg/L BOD; 0.01 µm pore size (1 kDa molecular-weight cutoff) | 97% removal for both COD and BOD | [22], experimental |
| COD, BOD | Stormwater | Feed concentrations: 11–32 mg/L BOD, 28–60 mg/L COD; 50 kDa molecular-weight cutoff UF | | [39], experimental |
RO, however, cannot be relied upon as the only process stage, as it is inefficient in the removal of low-molecular-weight organics, such as formic acid, methanol, formaldehyde, and urea. It also performs poorly in the rejection of boron [
40]. Reverse osmosis rejection of neutral organic pollutants has been found to increase with the increase in compound length and width and to decrease with the increase in compound hydrophobicity [
41]. Some non-polar, low-molecular-weight organics, such as N-Nitroso-dimethylamine (NDMA) and 1,4-dioxane, can pass through RO membranes [
42]. Another disadvantage of RO treatment is the challenge of the disposal of the concentrate. For potentially very harmful organics such as PAHs, ocean disposal would require further treatment, and surface-water disposal in inland regions is likely to be prohibited. O
3/BAF TTs may be more suited to potable reuse in inland regions.
UV treatment: The modelling used a PAH (benzo(a)pyrene) removal rate of 0% for UV alone. The UV dose used in the modelling of the effect of UV treatment was 12 mJ/cm
2, as this is a dose that is consistent with traditional wastewater disinfection and has been used in the literature to estimate the risk of potable reuse [
12]. Little or no PAH degradation is expected at this low fluence. Sanches et al. [
23] investigated the effect of UV dose on the degradation of three different PAHs (anthracene, fluoranthene, and benz(a)pyrene) in three different matrices (laboratory groundwater, groundwater, and surface water) and found very low removal at these low UV doses. PAH removal was found to require high fluence and was found to depend on the water matrix. Approximately 10% benzo(a)pyrene reduction in the groundwater types and approximately 0% reduction in surface water were reported at the fluence of 40 mJ/cm
2. Less degradation was reported for the other PAHs. Using much higher UV fluence, 1500 mJ/cm
2, anthracene and benzo(a)pyrene were efficiently degraded, with much higher percent removal being obtained when present in groundwater (83–93%) compared with surface water (36–48%). The removal percentages obtained for fluoranthene were lower and ranged from 13 to 54% in the different water matrices tested.
UV/H
2O
2 treatment: The modelling at the fluence of 800 mJ/cm
2 used a PAH (benzo(a)pyrene) removal rate of 99% for UV/H
2O
2. A study on the UV/H
2O
2 treatment of natural water by Rubio Clemente [
25] found that anthracene and benzo(a)pyrene removal rates of approximately 88% and 78%, respectively, were achieved using H
2O
2 at 10 mg/L and irradiance of 0.17 mW/cm
2 for 1 min, equating to the fluence of 10.2 mJ/cm
2. Increasing the exposure time to 30 min, equal to the fluence of 306 mJ/cm
2, increased the removal of the two PAHs to approximately 99%.
O
3/BAF treatment: The modelling in the current study used a 97% PAH removal rate for the O
3/BAF combination [
27]. Ozonation followed by biological treatment has been found to be very effective for PAH removal. An overall PAH (benzo(a)pyrene) removal rate of 91% after 30 min ozonation at 0.5 mg/L O
3 and 24 h biotreatment in stirred flasks was achieved [
36]. Degradation using ozone alone under the same dose/time regime only achieved 63% degradation. The use of BAF rather than stirred flasks is expected to yield timelier PAH removal. Overall PAH removal efficiency rates of 96.9% to 99.7% were achieved within 24 h. The major contributor to removal was sorption rather than biodegradation [
27].
UF treatment: The modelling in the current study used 67% PAH removal for UF treatment [
30]. O
3/BAF-based TTs (TTC and TTD) heavily rely on ultrafiltration (UF) for PAH removal. Smol and Wlodarczk-Mekula [
30] investigated the use of ultrafiltration to remove PAHs from highly polluted water from a coke process. The total concentration of 16 PAHs in the process of ultrafiltration was in the range of 8.9–19.3 mg/L. The efficiency of removal of PAHs from coke wastewater in the process of ultrafiltration equalled 66.6%. Taking into account the initial filtration, the total degree of removal of PAHs reached 85% [
30].
Chlorination treatment: Oxidation using chlorine can achieve PAH removal rates similar to those achieved with UV/H
2O
2. The indicative removal of benz(a)pyrene, a polyaromatic hydrocarbon (PAH), using chlorination is given, in the AGWR—Augmentation of Potable Water Supplies—as >80% [
6]. The modelling in the current study used a PAH removal rate using chlorination of 80% [
6].
2.3.2. Arsenic, Cadmium, and Lead Removal
Removal of these metals is limited to the wetland, RO, and BAF stages of TTs.
Wetlands: A 2012 literature review by Haarstad et al. (2012) [
19] shows the occurrence of more than 500 organic and metallic pollutants in wetlands. The removal of heavy metals is typically reported in the order of 30 to 60%, but it can reach 80 to 90%. A removal value of 30% was used in the modelling for constructed wetlands.
RO treatment: The RO removal of Cd from drinking water with Cd concentrations ranging from 0.02 mg/L to 0.54 mg/L with an average concentration of 0.23 mg/L using a Toray SC 3100 Membrane was reported to be 95% to 99%, with an average removal of 99%. Similarly, the use of a Dow 5K membrane achieved an average Cd removal rate of 98% [
20]. The modelling used 98% Cd removal with RO [
20].
The RO removal of arsenic (As(III)) has been found to be highly variable. This variability has been attributed to membrane type, matrix effects, and test conditions [
20]. The average As(III) removal rates over three separate one-week periods for the Dow 5K membrane were found to be 98%, 75%, and 83%. The modelling used 85% As(III) removal with RO [
20].
Ozbey-Unal et al. [
21] studied MF-RO treatment of industrial wastewater with a Pb concentration of 1.5 mg/L and found removal efficiency that ranged from 89.3 to 100% for the Pb ion. The modelling used 95% Pb removal for the RO stage in TTA and TTB.
BAF treatment: Arsenic removal using BAF is expected to be higher than removal using RO. The removal of As(III) with BAC has been found to be effective. Pokhrel et al. evaluated As(III) removal from groundwater and found 99% removal using BAC [
28].
Biologically active filtration was reported to be able to remove up to 86% of Cd(II) from simulated wastewater with a Cd(II) concentration of 0.2 mg/L (Dong 2018). Adsorption experiments also showed that BAC is able to reduce Pb(II) concentrations by 95% for starting concentrations of less than 0.2 mg/L (Dong 2018).
2.3.3. BOD and COD Removal
BOD removal and COD removal were modelled, as these are general indicators of the potential for the generation of harmful disinfection by-products (DBPs) in the advanced oxidation and chlorination stages of TTs. Most of this removal takes place in the early TT stages, thus minimising DBP formation later, in advanced oxidation and/or chlorination.
Wetland treatment: The removal of dissolved organics in free-water-surface wetlands is attributed to phytodegradation, phytovolatilization, phytostimulation, phytoextraction, and microbial degradation. Dissolved heavy metal removal is attributed to precipitation, adsorption, and plant uptake. The removal of undissolved pollutants is attributed to sedimentation. The performance of constructed wetlands depends upon various factors, such as hydraulic and organic loading rates, pH, dissolved oxygen, temperature, plant species, and growth phase [
11]. The selection of literature data on BOD, COD, and heavy metal removal using constructed wetlands was based on general conclusions of literature reviews. A recent literature review on the performance of constructed wetlands in tropical and cold climates concluded that low temperature has the most antagonistic effect on the performance of constructed wetlands [
21]. Free-water-surface constructed wetlands were estimated to exhibit around 89% BOD removal and 72% COD removal. These COD and BOD removal values were used in the modelling. Considerably higher BOD and COD removal rates were reported for constructed wetlands (98% and 97%, respectively) [
43], but these figures were not used in the modelling, as the Ethiopian study area of that research is subject to a very different climate from that of the study area of the current research. The Ethiopian study compared the performance of constructed wetlands and natural wetlands and found that natural wetlands yielded lower BOD and COD removal (92% for both BOD and COD).
RO treatment: The modelling used 99% for COD and BOD removal using RO. RO has been found to be very effective in the removal of COD and BOD. Jadhao et al. investigated the RO and UF treatment of hospital wastewater that contained 200–235 mg/L COD, 95–115 mg/L BOD and found that the percentage removal efficiency rates of COD and BOD were more than 99% with RO [
22].
O
3/BAF treatment: The removal of COD and BOD using ozone alone at full scale has been found to be highly variable. Martinez et al. used ozone to treat secondary effluent and achieved 8% to 88% removal of COD, with most of the results being between 10% and 20%, and ~0% to 68% removal of BOD, with most of the results indicating ~0% BOD removal [
37]. Similarly, in the case of COD, Zanacic et al. found that two different full-scale O
3/BAF plants achieved different COD removal rates [
38]. One achieved ~50% COD removal, and the other, ~20% removal. Less variable results were achieved by He et al. [
26]. The study investigated the combined use of ozone and biological aerated filtration to treat textile effluent containing ≤110 mg/L COD and ≤30 mg/L BOD. A 20–25 mg/L ozone dose with 3.3 hr hydraulic retention time and an air-to-water flow ratio of 6 was found to result in approximately 64% COD removal and 67% BOD removal [
26]. The modelling used 64% COD removal and 67% BOD removal for the O
3/BAF combination.
UF treatment: Ultrafiltration has also been found to be able to retain large percentages of COD and BOD. The percentages of removal efficiency of COD and BOD from hospital wastewater using a tight UF membrane with 0.01 µm pore size, approximately equating to a 1 kDa molecular-weight cutoff [
44], were found to be more than 99% for RO and more than 97% for UF [
22]. For stormwater treatment using a membrane with a 50 kDa molecular-weight cutoff, however, average rates of UF removal of COD and BOD of 42% and 66% were observed [
39]. Ultrafiltration does not provide a physical barrier for the separation of dissolved salts and heavy metals, as these are smaller than the membrane pores [
31]. The modelling in the current study used 97% COD and BOD removal with UF [
22], as this was achieved with a UF membrane with a pore size closer to that of the literature study used to estimate PAH removal [
30] (see
Table 3).
UV and UV/H
2O
2 treatment: The modelling in the current study used conservative BOD and COD removal values of 0% with UV alone and UV/H
2O
2. The UV and UV/H
2O
2 treatment modelling in the current study was limited to the fluence doses adopted of 12 mJ/cm
2 for treatment with UV alone and 800 mJ/cm
2 for treatment with UV/H
2O
2 [
12]. BOD and COD removal is expected to be very low at both these fluence levels. Muhammad et al. [
24] found approximately 5% BOD and COD reductions from raw textile bath wastewater for UV alone and approximately 5% COD removal and 15% BOD removal with UV/H
2O
2 treatment using irradiance of 5 mW/cm
2 (254 nm) for 5 min exposure time, equating to the UV fluence of 1500 mJ/cm
2. The maximum reported BOD and COD removal rates were approximately 35% BOD and 25% COD for UV alone and approximately 45% BOD and 30% COD for the UV/H
2O
2 treatment after 25 min of exposure at 5 mW/cm
2, equating to the fluence of 7500 mJ/cm
2. Similarly, Rubio-Clemente [
25] found little or no TOC removal from natural water using irradiance of 0.46 mW/cm
2 for 5 min, equating to 138 mJ/cm
2. TOC removal rates of approximately 20% were achieved using 10 mg/L H
2O
2 and irradiance of 0.46 mW/cm
2 for 30 min, equating to approximately 800 mJ/cm
2.
Chlorination treatment: The modelling in the current study used conservative chlorination treatment BOD and COD removal values of 30% [
32] and 10% [
33], respectively. The strong oxidizing ability of chlorine decreases the amount of residual organic substances and thus can decrease BOD in the effluent. The net effect on BOD and COD is dependent on the chlorine dose and contact time. Chlorine doses up to 5 mg/L added to water from treated sewage and allowed to react for 30 min were found to decrease BOD, but when 30 mg/L or more of chlorine was added, the organic matter was progressively decomposed to low molecular weights, causing BOD and COD to double in the effluent [
45]. Another study by Ishihara et al. [
32] on chlorination of secondary effluent found that higher percentage removal rates were achieved with chlorination in samples with higher starting BOD. When three different samples with starting BOD concentrations of ~30, 21, and 13 mg/L were treated to a residual chlorine concentration of 5 mg/L, they exhibited BOD reductions of 67%, 38%, and 20%, respectively. Ustun et al. treated industrial wastewater with 25 mg/L COD and 1–3 mg/L NaOCl (0.3 to 1.2 mg/L free chlorine) for 30 min, giving rise to 10% removal of COD [
33].