Aquifer recharge with reclaimed water: life-cycle assessment of hybrid concepts for non-potable reuse

Against a background of population growth, increasing industrialization and urbanization as well as climate change, wastewater reuse is increasingly considered as a possible alternative water source for diverse non-potable uses (Asano 1998). A crucial factor is to define the water quality requirements for specific uses. However, it also needs to be considered that reclaimed water is usually available at relatively constant flows throughout the year, while the demand for reclaimed water is seasonal. Therefore, the question of reclaimed water storage is also essential.

Most of the aquifer recharge applications of wastewater reuse so far rely on high-pressure membrane systems or even double-membrane and advanced oxidation processes as pre-treatment. However, when non-potable reuse is targeted or the replenishment of a threatened aquifer is planned, recharge with high-quality non-potable water could be envisaged (Aertgeerts & Angelakis 2003), as acknowledged by the legislation of several countries. The goal of this study was to identify treatment schemes including aquifer recharge (Figure 1) that meet the requirements for non-potable re-use and to analyse the environmental impact in comparison to traditional water re-use schemes using double membranes.

A defined secondary effluent (SE) was considered as input flow for this conceptual study on the basis of a worldwide survey of typical SE water qualities (total suspended solids, TSS: 19 mg/L; turbidity: 9 NTU; chemical oxygen demand, COD: 61 mg/L; N-NH₄: 7 mg/L; N-NO₃: 9 mg/L; P-PO₄: 6 mg/L; total coliforms (TCF): 5 × 10⁵/100 mL; Escherichia coli: 10⁵/100 mL). The major legislations from WHO, USEPA and Australian guidelines were considered to define the water quality to be reached by these hybrid treatment schemes (NRMMC-EPHC-AHMC 2006; WHO 2006; NRMMC-EPHC-NHMRC 2009; USEPA 2012). Particulate pollution (TSS), biological pollution (biochemical oxygen demand, BOD₅) and microbiological contamination (TCF) were the only quality requirements for unrestricted non-potable reuse which were consistently defined in the above-mentioned guidelines (TSS < 10 mg/L, BOD₅ < 20 mg/L and less than 1 faecal coliform (FCF)/100 mL). Besides pathogen removal, all other processes occurring during saturated subsurface passage in the aquifer were not taken into account due to their site specificity. Especially as the targeted values in suspended solids (10 mg/L) and microbiological contaminants (1/100 mL) require significant disinfection and filtration processes as pre-treatment. The targeted value corresponds to calculated 5-log removal of microbiological pollution. This includes 2 log for ozonation (Bahr et al. 2007), >3 log for UV applying a fluence of 400 J/m² (Lazarova et al. 1999), up to >4 log for ultrafiltration (UF) (Wang et al. 2005), >1.5 log from infiltration ponds/slow sand filtration (Bali et al. 2011) and another 1.5 log removal for a subsurface passage of minimum 5 days (Schijven et al. 1998, 1999, 2000; DeBorde et al. 1999; Partinoudi & Collins 2007). Five schemes that fulfil the requirements for unrestricted non-potable reuse were selected on the basis of a large review of typical pollutant removal efficiencies found in the literature (Asano 1998; Salgot et al. 2002; Amy et al. 2006; Dillon et al. 2008; DWA 2008) and compared with a double-membrane system (UF + NF) operating with similar input qualities, but with direct reuse without aquifer storage (Table 1). Thus, a direct comparison is not possible, other than to say UF − NF achieves better water quality (e.g. also reduces salinity).

Owing to the lack of existing data, and to enable an easier comparison of different schemes, a hypothetical case study was considered based on real data from Veolia-operated sites. The case study was chosen to be located in Morocco as the Mediterranean region is expected to be a potential area for application. This choice impacts the emission factors (e.g. for the electricity mix) as well as the transport distances for raw and engineered material supply.

The plant was designed with a size of 50,000 population equivalent (PE), with a considered design wastewater flow of 6,250 m³/d (2.3 Mm³/y) based on a 125 L/day/PE wastewater generation. The wastewater quality is based on the 75th percentile of the SE quality review performed, as listed above. Additional boundary conditions are no need for salinity reduction, e.g. due to the specific crop to be irrigated and no major heavy metal input into the municipal wastewater treatment plant. The hypothetical case study consists of a tertiary water reclamation and aquifer recharge facility next to an existing wastewater treatment plant (for main design parameters, see Table 2). The possibility to recharge the aquifer and sufficient land availability are taken for granted.

To compare the different hybrid solutions, a life-cycle assessment (LCA) was conducted. LCA is a standardized method to quantify various environmental impacts of a process or product. It enables monitoring of all direct and indirect impacts of a given process and reveals a shift of environmental burdens to other areas of the environment or to other geographical areas. The life cycle of a system is the ‘consecutive and interlinked stages of a product system, from raw material acquisition or generation from natural resources to final disposal’ (ISO 14040 2006).

As applied to water and wastewater systems, the LCA evaluates the different stages of the life of the infrastructure (mainly construction, use and decommissioning) and includes the linked indirect activities for operation, such as electricity production, transport and chemicals used (Renou 2006). For assessing the life cycle impacts of a given process, aggregated inputs and outputs of the system are evaluated with specific environmental indicators. In the present paper, only the impact on climate change (carbon footprint) will be presented, while the study also assessed human toxicity and terrestrial acidification. Substance flow models for all assessed treatment schemes are implemented and evaluated using the LCA software ‘Umberto^®5.6’ (IFU & IFEU 2009). The considered emission factors for background processes (electricity, chemicals and infrastructure) are compiled from the ecoinvent database (Ecoinvent 2010) complemented by a Veolia-internal database.

In addition to traditional LCA indicators, the land footprint (ha of land/Mm³ of reclaimed water per year) and the electricity demand or energy intensity (kWh/m³) of the different treatment trains were assessed. Figure 2 gives an overview of the system boundaries, the processes considered and the indicators evaluated.

The electricity demand of treatment trains #1–4 is approximately 0.20–0.25 kWh/m³ and thus up to five times less than combined ultrafiltration and nanofiltration (#6) (Figure 3). The combination of UV with infiltration ponds (#5) has even lower energy demand levels with 0.08 kWh/m³. It is noteworthy that water pumping from the recharged aquifer can significantly increase the electricity demand of the treatment trains. Here, a 25 m pump total dynamic head was considered (Table 2), and the wells' energy demand may amount to up to 30% (#1–4) or even 60% of the total energy demand (#5). Thus, storing the water into aquifers with a deep piezometric surface will not be economically favourable compared to using more shallow or surface water resources. At an abstraction depth of 50 m, the electricity demand for water abstraction will exceed that of the treatment itself (schemes 1–5).

Most hybrid treatment trains have comparable CO₂ emissions of around 0.20 kg CO₂eq/m³ (Figure 4), with UV disinfection and infiltration (#5) emitting less CO₂ (around 0.1 kg CO₂eq/m³). Nanofiltration (#6) increases the CO₂ emissions three to five fold (around 0.7 kg CO₂eq/m³). Electricity is clearly the dominant factor for CO₂ emissions related to these treatment trains. It represents 70–75% of the total life cycle carbon footprint of the treatment trains, except for UV disinfection and infiltration (#5) where it represents only 51% of the total carbon footprint, and nanofiltration (#6), with 94% of life cycle emissions. Another important contribution to CO₂ emissions for the treatment trains with infiltration ponds (#1, #4 and #5) is construction – amounting up to 46% for UV disinfection and infiltration (#5). Chemicals also represent a significant source of CO₂ emissions for the treatment trains involving membranes (#2, #4 and #6). Similar conclusions were obtained for LCA of advanced phosphorous removal during wastewater treatment by Remy et al. (2014).

 Figure 5 shows the land footprint of each proposed scheme. If space availability is an issue, solutions involving membranes (#2 and #6) or slow sand filters (#3) should be preferred as they need only very little space. On the whole, highly urbanized areas may not be the primary target of hybrid reuse schemes, which will probably choose high-technology, high-energy demanding equipment providing reclaimed water of potable quality.

All five proposed hybrid treatment trains are capable of supplying water of high-quality fit for all non-potable reuses, and the combination of disinfection, filtration and aquifer passage proved to be an efficient combination for removing suspended solids, residual BOD and microbiological contaminants to the required degree. The environmental performance of the treatment trains was compared in terms of carbon footprint (Figure 4), but also electricity demand (Figure 3) and land footprint (Figure 5). Both the electricity demand and carbon footprint of hybrid schemes were found to be considerably lower than for a double-membrane system, besides offering an additional storage solution in the aquifer without evaporation losses and direct anthropogenic or climatic impact.

Thus, there is a significant margin for lowering the environmental impact, energy demand and operational costs (not shown) if non-potable water quality is sufficient for the reuse goal. While the legal context and social acceptability may represent barriers for this intended recharge of non-potable water to the aquifer, one may question the necessity to use water of potable quality for non-potable reuse, saline intrusion control or land subsidence mitigation if alternative high-quality non-potable water solutions are available.

The project OXIMAR-1 was funded by Veolia Water. The authors gratefully acknowledge not only financial support but also technical guidance and fruitful discussions with colleagues of Veolia Eau Technical Direction (VE-DT), VWS&T Corporate Marketing Department and Veolia Environnement Recherche et Innovation (VERI) in the course of the project.