Abstract
In this article, 35 published studies on life cycle assessment (LCA) of sewage sludge were reviewed for their methodological and technological assumptions. Overall, LCA has been providing a flexible framework to quantify environmental impacts of wastewater and sewage sludge treatment and disposal processes for multiple scales, ranging from process selection to policy evaluation. The results of LCA are, in principle, unique to the goal and scope of each study, reflecting its local conditions and comparison between different LCAs is not intended. Furthermore, the assessments are limited by the methodological development of the life cycle impact assessment (LCIA) and the advancement of research in quantifying environmental emissions associated with wastewater and sewage sludge treatment processes. Thus, large discrepancies were found in the selection of the environmental emissions to be included and how they were estimated in the analysis. In order to reduce these choice uncertainties, consolidation of the modelling approach in the following area are recommended: quantification of fugitive gas emissions and modelling of disposal practices. Besides harmonization of the key technical assumptions, clear documentation of the modelling approach and the uncertainties associating with each assumption is encouraged so as to improve the integrity and robustness of assessment.
Keywords
Introduction
Sewage sludge management poses not only financial but also planning challenges to wastewater treatment plants (LeBlanc et al., 2008). Wastewater treatment processes tend to accumulate pathogens, heavy metals and trace organic pollutants in the sludge, but recognizing its nutrient value, stabilized sewage sludge is commonly applied on the land as a fertilizer and soil conditioner (Singh and Agrawal, 2008). Prior to land application, regular monitoring and stabilization of sewage sludge is required to ensure the protection of public health (Iranpour et al., 2004). Properly stabilized sludge, also known as biosolids, is generally regarded as acceptable to apply on land. However, the application of stabilized sludge on land remains controversial (Busetti et al., 2005; Carballa et al., 2004). Stabilization processes aim to destroy pathogens, eliminate offensive odours and improve aesthetics and transportability. Stabilization technologies utilize physical (heat, desiccation, pressure, irradiation and ultrasound), biological (anaerobic digestion and composting) and chemical processes (pH adjustment, oxidants, biochemical by-products). Some stabilization measures also provide an opportunity for energy recovery, such as anaerobic digestion and pyrolysis, making the sewage sludge management increasingly multi-focused and, hence, regulated by a number of policies whose goals sometimes contradict each other.
In order to untangle this policy and planning complexity, life cycle assessment (LCA) of sewage sludge management has been gaining ground in recent years. LCA is a framework to quantify the potential environmental impacts of targeted systems and provide a basis for informed decision making. The concept of the LCA was first developed in the 1960s and 1970s to assess industrial products along with the rise of sustainable consumerism its application was further expanded from 1980 to 1990 (Guinée et al., 2011). There are several guidelines for conducting LCAs [ISO 14040 (ISO, 2006); EC-JRC, 2012]. LCA has been applied extensively to the solid waste management sector, including sewage sludge management. More than 300 peer-reviewed articles on LCA on various aspects of solid waste management issues have been published and almost 50 LCA models specifically targeting solid waste treatment technologies have been developed in Europe alone (Gentil et al., 2010; Pires et al., 2011).
Few studies have been conducted with the aim of comparing methodological aspects of LCA modelling of sewage sludge management. Lundin et al. (2000) compared the effects of the choice of system boundaries and suggested that system expansion was necessary for fully addressing the environmental benefits and burdens of wastewater treatment systems. Renou et al. (2008) compared the influences of the choice of LCIA methodology and called for future research concerning local impact categories such as human and eco-toxicity. Friedrich et al. (2007) conducted a short review of the use of LCA in the water industry, including wastewater and sewage sludge management. However, the aspects of technological modelling are often overlooked, although previous studies indicate that the results of an LCA are affected greatly by the underlying assumptions, such as the choice of environmental emissions and their quantification method (Gentil et al,. 2010; Pires et al., 2011; Winkler and Bilitewski, 2007). There is need to revisit the technical assumptions in sewage sludge LCA and to consolidate the modelling approaches for the key technologies for sewage sludge management.
Thirty-five sewage sludge LCA studies were reviewed in this study. This review study only covers articles published in peer-reviewed journals since these studies were subject to third party critical review, as recommended in ISO 14040/44 (ISO, 2006) and they reflect the state of the knowledge of environmental impacts associated with sludge management practices. Table 1 lists the studies included. They were published in 13 different peer-reviewed journals between 1998 and 2012. The reviewed studies were identified through a keyword search for ‘sewage sludge and life cycle assessment’, ‘sewage sludge and life cycle analysis’ and ‘sewage sludge and environmental analysis’ on the ISI Web of Knowledge. Some of the early works were identified through the reference lists of later articles. For the purpose of comparison, the review focuses on sewage sludge of municipal origin. Some LCA studies were excluded due to the lack of sufficient documentation of the technical assumptions behind the sewage sludge treatment processes modelled (i.e. Dennison et al. (1998); Emmerson et al. (1995); Lundie et al. (2004)) or the absence of the impact assessment phase (i.e. Hara and Mino (2008)). Other studies were drawing on results directly from previous LCA studies (Palme et al., 2005) and were thus excluded from the study. The review covers three aspects of LCA: drivers and goals, methodological assumptions and technical modelling. The drivers and goal were examined in a chronological manner to illustrate the trend in the use of LCA for sewage sludge management. Methodological and technical modelling approaches were reviewed to identify similarities and differences between the LCA studies. It is the aim of this review article to illustrate the current status and provide a basis for harmonization of environmental assessment of sewage sludge LCA.
Study area, scale and drivers for selected studies.
N.R., not reported; N.A., not applicable; d.w., dry weight; w.w., wet weight.
Drivers and goals of the reviewed studies
As listed in Table 1, it is evident that sewage sludge LCAs have been conducted on various scales from treatment process selection to formulation of national policies and legislation. These projects were concentrated in European countries (21 out of the 35 studies reviewed), as several key regulations affecting sewage sludge have been enacted and implemented in Europe.
The implementation of the Urban Water Directive has increased sewage sludge production and has initiated the search for environmentally viable sewage sludge treatment options (Gallego et al., 2008; Hospido et al., 2008; Suh and Rousseaux, 2002). The search for cost- and energy-efficient alternatives has also fostered an investigation of alternative wastewater treatment practices, such as constructed wetlands and decentralized treatment systems (Burian et al., 2000; Uggetti et al., 2011) and source-separated collection of wastewater (i.e. urine diversion, grey water separation). These alternative methods were analysed in the context of small (Tillman et al., 1998) to mid-size communities (Remy and Jekel, 2008).
Meanwhile, wastewater treatment plants already equipped with secondary or even tertiary treatment processes are now facing increasingly stringent nutrient discharge standards to prevent surface water pollution. Foley et al. (2010) conducted an LCA on ten different scenarios for nutrient discharge limits in an Australian context. Carbon contained in wastewater and sewage sludge is largely regarded as 100% biogenic (Fellner and Rechberger, 2009). However, the direct measurement of radiocarbon (14C) in municipal wastewater treatment plants effluent indicated that 25% of dissolved organic carbon and 14% of particulate carbon could be of fossil origin (Griffith et al, 2009). Carballa et al. (2011) conducted an assessment taking the findings from Griffith et al. (2009) into consideration and found that the direct emission from sludge treatment processes could be tripled when considering the fossil carbon content of the sludge.
For instance, anaerobic digestion is one of the most commonly utilized methods for sewage stabilization and renewable energy production, and was included in 19 studies. The generated biogas can be utilized for on-site energy and heat consumption, upgraded to augment natural gas supply, or compressed for use as a transportation fuel. Novel technologies for energy recovery such as pyrolysis, wet oxidation, super critical wet oxidation, ash melting and various pretreatment processes were also assessed (Bridle and Skrypski-Mantle, 2000; Hong et al., 2009; Hospido et al., 2005; Houillon and Jolliet, 2005; Nakakubo et al., 2012; Soda et al., 2010; Svanstrom et al., 2005).
Use of sewage sludge on land could provide an opportunity for nutrient recovery. In Europe, where landfilling of sewage sludge is discouraged by statutory requirements, land application is becoming increasingly popular. For instance, together with the ban on ocean dumping, the use of sewage sludge on land is expected to rise from 65 to 85% in Spain by 2015 (Hospido et al., 2010). However, critics argue that the use of sewage sludge on land could result in the accumulation of heavy metals and other pollutants in soils and potentially open new transmission path-ways for infectious disease (Gerba and Smith, 2005; He et al., 2005). In order to balance trade-offs between nutrient recovery and associated health risks, application of sewage sludge on land has been the subjected of vigorous public debates in many countries (LeBlanc et al., 2008). Considering these competing interests, Sablayrolles et al. (2010) and Hospido et al. (2010) focused their LCA studies on toxicity aspects of sewage sludge use on land.
Furthermore, among all the benefits of the utilization of sludge for agricultural purposes, phosphorus recovery has been the focus on sewage sludge treatment (Johansson et al., 2008; Lederer and Rechberger, 2010; Linderholm et al. 2012; Lundin et al., 2000, 2004; Nakakubo et al., 2012; Svanstrom et al., 2005). Phosphorus is un-substitutable and a non-renewable resource, whereas modern agriculture relies heavily on input of phosphate fertilizer for maintaining or increasing its productivity (Cordell et al., 2009; Fresco, 2009). The idea of phosphorus depletion has been challenged, as the phosphate rock reserve was revised in 2011. Experts now claim that the reserve is likely to last for 300–400 years at the current rate of exploitation (Kauwenbergh, 2010; USGS, 2012). Nonetheless, high grade and economically viable phosphate rock reserve is centered in a few countries, with Morocco and Western Sahara holding over 70% of the global known reserve (USGS, 2012). Recovery of phosphorus is still a sensible approach for ensuring food security (Cordell et al., 2011). Besides being used on land as a fertilizer, sludge could be used as a raw material for other industrial activities. Murray et al. (2008) compared an option for use of sludge for brick manufacturing, while Hong and Li (2011) evaluated its use in the cement industry.
In order to facilitate life cycle thinking in the decision-making process at the plant level, LCA tools for wastewater and sewage sludge treatment processes were also developed. (Brown et al., 2010; Pasqualino et al., 2009). Yet, the use of LCA is not limited to the products or process selection at the plant level; LCA has also been used in the context of regional infrastructure planning and policy analysis. In Lundin et al. (2004), it was used for developing sustainability indicators for the Stockholm metropolitan area (Palme et al. 2005). Cartmell et al. (2006) used LCA as a tool to assess the environmental, economic and social viability of co-combustion of sewage sludge and municipal solid waste. Along with LCA, cost benefit analysis, risk characterization and sustainability appraisal methods were also used in this study. Lastly, sewage sludge management has also been used as a case study to improve LCA methodologies themselves. Poulsen and Hansen (2003) focused on the methodological development to quantify the use of non-renewable resources. Renou et al. (2008) compared the impact of LCIA methodologies on the outcomes of LCA. Hospido et al. (2010) assessed the contribution of emerging pollutants such as pharmaceutical and personal care products to the overall toxicity potential.
Methodological assumptions
This section reviews the methodological assumptions, such as functional units, system boundaries and the choice of LCIA methods. Table 2 summarizes the functional units, LCI database, LCA modelling tool, LCIA methods and approach to sensitivity analysis adopted by the reviewed studies. Each component of the methodological assumptions is discussed in detail below.
Summary of functional units, LCI database, LCA modelling tool, LCAI methods and the approach for sensitivity analysis of the reviewed articles.
Those models are used only for inventory phase and not LCA tools. d.w., dry weight; w.w., wet weight.
Functional unit
The functional unit provides the basis of comparison for the LCA studies; the amount and type of sludge managed to reach a certain level of stabilization. A mass-based approach was the most common practice (17 of 35 studies), whereas volume-based functional units were used to a lesser degree (seven of 35 studies). A person equivalent (the amount of wastewater or sludge generated in a specific time period by one individual) was also used to represent both the quality and quantity of the functional unit. The provisioning of wastewater/sewage sludge treatment services (Nakakubo et al., 2012; Remy and Jekel, 2008; Renou et al., 2008) or the capacity to remove chemical oxygen demand (Poulsen and Hansen, 2003) were used. More recently, the service or good provided by treatment of sewage sludge was considered as a functional unit. Hong and Li (2011) focused on the production of final products (Portland cement), while one tera-joule (TJ) of steam production from sludge incineration was studied by Liu et al. (2011) and the use of 11 kg of pure P to agricultural land was introduced by Linderholm et al. (2012).
The majority of the studies included a quality component such as total solids, organic carbon, or nutrient content in the definition of the functional units, although it was often poorly documented.
System boundary
The system boundary defines which technical systems will be included in the assessment. Guinée et al. (2002) discussed those system boundaries in three levels: the boundary between the technical system and the environment, the boundary between different technical systems (especially for the choice of upstream and downstream systems) and the differentiation of significant and insignificant contributions (cut-off criteria). Each aspect of the system boundaries was discussed in detail in the following sub-sections.
A geographic boundary was mentioned in 32 out of 35 studies (Table 1). On the contrary, the time horizon of the technical system was not clearly stated in most of the sewage sludge LCA studies. Only three studies explicitly mentioned the time horizon. Suh and Rousseaux (2002) and Pasqualino et al. (2009) used 100 years for the evaluation of global warming potential (GWP), and Tarantini et al. (2007) used the same period for emissions from landfilled sewage sludge. Tillman et al. (1998) specified the number of decades as a time horizon to address the long lifetime of built infrastructures.
Technical systems
Fifteen studies used raw sludge as a starting point. Seven studies omitted some of the sludge treatment processes as they were shared by all the alternatives considered. Fourteen studies included wastewater treatment processes, giving equal or greater emphasis to the sewage sludge treatment processes. Among these, Remy and Jekel (2008) and Tillman et al. (1998) expanded the system boundary to include the collection system, since they considered the options for source separation of wastewater at the point of production.
The selection of sludge end use is also reflected in the goal and scope of the studies, but others also stated the lack of knowledge of those practices. Cartmell et al. (2006), Bravo and Ferrer (2011) and Liu et al. (2011) omitted the disposal and reuse of incineration ash. Uggetti et al. (2011) disregarded the treatment after dewatering processes and did not consider the subsequent composting or other sludge treatment processes following dewatering of sludge.
The construction and demolition phase of infrastructure was not included in the majority of the studies. Five studies considered the construction phase, although Foley et al. (2010) only included the volume of concrete used in each alternative as an indicator of resource use intensity. Three studies included the construction and demolition phase to highlight a reduction in materials used for decentralized wastewater treatment processes (Lundin et al., 2000; Remy and Jekel, 2008; Tillman et al., 1998). Some studies targeting medium to large service populations also included the construction phase, but they mostly focused on the GWP (Hong et al., 2009; Stokes and Horvath, 2010). The end-of-life treatment of equipment was partially included by Uggetti et al. (2011).
Others cited previous studies for wastewater and water infrastructure LCA studies to illustrate that the contribution of the construction and demolition phase was insignificant or negligible. However, the studies including the construction phase reached different conclusions as to the importance of the construction phase. Whereas Renou et al. (2008), Tillman et al. (1998) and Remy and Jekel (2008) concluded that the construction phase only contributed 0.1–16% of the total environmental impact of the wastewater treatment process, Lundin et al. (2000), Hong et al. (2009) and Stokes and Horvath (2010) emphasized that depending on the boundary condition, the construction and machinery production could be a significant, if not major, contributor to some environmental impacts categories.
Both Tillman et al. (1998) and Lundin et al. (2000) compared the central and decentralized systems of wastewater treatment for two different service populations. Both of them agreed that the environmental loads for construction of facilities are lower for large-scale than for small-scale systems, and economies of scale also applied not only to the investment cost but to the environmental burden.
Upstream and downstream boundaries
Out of the 35 studies, 34 took a zero burden approach and assumed no environmental emissions were associated with the generation of input wastewater or sludge. Tillman et al. (1998) included provisioning of potable water in the upstream emission; yet the embedded emissions for other constituents of wastewater such as detergent and human excreta were not included. Provisioning of electricity for the processing operation was included in all of the studies, although some studies lacked documentation. Polymers for the dewatering process, lime for air pollution and pH control, and ferric salts for nutrient precipitation were the major chemical inputs.
Downstream boundaries include avoided energy and fuel and chemical production. Two Swedish studies used the regional district heating model (MARTES) to identify the least efficient means of heat production to account for the energy substitution (Johansson et al., 2008; Svanstrom et al., 2005). Other studies took an attribution approach and used the average energy mix for a given geographical area for energy substitution.
As recovery of nutrients was listed as being one of the focal topics of sewage sludge LCA, 24 studies included fertilizer substitution. This will be discussed further in the section ‘Disposal and utilization of stabilized sludge’ below.
Twelve studies drew process-based LCI data for upstream and downstream systems from existing databases such as BUWAL, Ecoinvent, IDEMAT, AQUASAVE, IFU&IFEU, Probas, GEMIS and JLCA-LCA. Others used data for larger geographical units obtained through literature search. Murray et al. (2008), Hong et al. (2009) and Stokes and Horvath (2010) took a hybrid approach by using economy input–output LCA for the background and process-based LCA for core systems. Some studies mentioned the use of built-in databases within LCA software packages such as SimaPro, but did not specify which database was used.
Boundary condition for cut-off criteria
Processes or emissions which are found to be insignificant in the life cycle stages, activities and specific processes and products can be omitted or ‘cut-off’ from the assessment (IRLCD, 2010). A cut-off criterion is often based on an expert’s judgment, but some guidelines advise applying quantitative cut-off criteria. For example, carbon trading schemes such as the Clean Development Mechanism or PAS 2050 state that greenhouse gas (GHG) emissions which are believed to be less than 5% of the calculated total emission, could be considered to be de minims, and hence are excluded from the assessment (Sinden, 2009). None of the studies had a clear discussion about the cut-off criteria used. Rather, they followed the cut-off rules embedded in the LCI database or literature cited.
Life cycle impact assessment
LCIA is a process to translate emissions into defined impacts by assigning each emission with a specific characterization factor and sum the characterized impact for each impact category. Several LCIA methods are available and commonly used by researchers and LCA practitioners. In this study, CML, IMPACT and the guideline by the Intergovernmental Panel for Climate Change (IPCC) were identified as the most favoured LCIA methods among sewage sludge LCAs. The choice of LCIA methodologies is largely based on the location of the study: Environmental Priority Strategies in product design (EPS: Steen (1999)) were adapted by Swedish studies, the Centrum voor Milieukunde Leiden Method (CML: Guinée et al. (2002)) by Spanish studies, and IMPACT 2002+ (Jolliet et al., 2003) by studies in other areas. When GHG emissions were the sole focus of the study, the national emission reporting guideline set by the IPCC was also used. The Uniform System for the Evaluation of Substance (USES-LCA) was used to evaluate the fates and exposure rates of trace pollutants by Sablayrolles et al. (2010), Peters and Rowley (2009), Hong et al. (2009) and Tarantini et al. (2007). Yet, nine studies did not specify the LCIA methodologies adapted.
These methods vary in the selection of environmental impact categories and characterization modelling approaches (e.g. midpoints approach vs. endpoints approach). Renou et al. (2008) conducted a comparative study of five impact characterization methods. Their study indicated that consistent results can be obtained for global impact categories (i.e. global warming, resource depletion and ozone depletion). However, discrepancies were found in more regional and local impact categories (i.e. acidification, eutrophication, human and eco-toxicity).
Figure 1 presents the impact categories used in the 35 reviewed papers and the number of studies that included each impact category. By far the most common impact category considered was GWP in kg CO2 eq., which was included in 30 studies. Ten studies were based on energy balances. Acidification potential (AP), eutrophication potential (EP), human toxicity (HT) and eco-toxicity (ET) were also common impact categories assessed, reflecting the function of and public concerns with sewage sludge treatment practices. Murray et al. (2008) and Stokes and Horvath (2010) used the US EIO-LCA table and included NOx, SOx, particulate matter (PM) and volatile organic carbon (VOC) emissions as surrogates of human toxicity impacts.
Impacts categories included in the reviewed sludge LCA: GWP, global warming potential; AP, acidification potential; EP, eutrophication potential; POP, photochemical oxidation potential; ODP, ozone depletion potential; HT, human toxicity; ET, ecotoxicity; DAR, depletion of abiotic resource; EB, energy balance.
Eleven studies included the normalization process and normalization references were taken from the chosen LCIA methodologies. Contributions by one average Western European person during one year including all activities in life was often used but an average contribution from one Swedish person in studies was also used in the studies adopting the EPS method. Poulsen and Hansen (2003) and Hong et al. (2009) also performed normalization, but normalization factors were not documented. Only three studies weighted results (Liu et al., 2011; Lundin et al., 2004; Suh and Rousseaux, 2002) but the reasoning behind the weighting was not clearly documented (Lundin et al., 2004; Suh and Rousseaux, 2002). Liu et al. (2011) cited the work of Lin et al. (2005) which provides a weighting system based on the environmental policies issued in the tenth five-year (2001–2005) plan.
Technical assumptions
There are several approaches available for modelling sewage sludge management systems. Lederer and Rechberger (2010) and Cartmell et al. (2006) took a largely mechanistic approach following the flow of elements in sludge through treatment processes. On the other hand, Hospido et al. (2004, 2005, 2008) and Gallego et al. (2008) considered sludge treatment as a single process and evaluated inputs and outputs from sewage sludge treatment processes. In this case, emission data were collected from existing wastewater treatment plants or bench scale studies, and were correlated with input mass or with volume entering the targeted system. The majority of the reviewed studies were hybrids of the two approaches.
Three studies were dedicated to the development of LCA tools for the wastewater treatment sector; SiSOSTAQUA by Pasqualino et al. (2009), WWEST by Stokes and Horvath (2010) and BEAM by Brown et al. (2010). BEAM is available to the public for free (CCME, 2011) and the other two are used mainly for research purposes. SiSOSTAQUA has been used for the environmental assessment of wastewater treatment and reuse (Meneses et al., 2010; Pasqualino et al., 2009, 2011). SiSOSTAQUA includes a full range of impact categories, while BEAM and WWEST are Excel-based models and mainly focus on global warming. Besides global warming, WWEST calculates nitrogen oxides (NOx), particulate matter (PM), sulfur oxides (SOx) and volatile organic compounds (VOC).
Thirteen studies used a generic model for LCA and/or a material flow analysis (MFA), such as SimaPro, Gabi, LCA iT, TEAM by Ecobilan, UMBERTO GEMIS and STAN. Tillman et al. (1998) used ORWARE (Organic Waste Research model), an LCA modelling tool for organic waste. While the majority of the studies treat the wastewater treatment process as a static system, Foley et al. (2010) used a commercial wastewater treatment process simulation model BioWin® to encompass the dynamic features of the wastewater treatment process.
Table 3 presents the number of technologies covered in the reviewed articles. In total, 31 treatment technologies were assessed ranging from traditional stabilization methods to novel technologies for phosphorus extraction recovery (i.e. SUSAN by Lederer and Rechberger (2010), Bio-con and Cambi-KREPRO by Lundin et al. (2004) and Svanstorm et al. (2005), struvite precipitation, hydroapatite precipitation, alkali extraction by Nakakubo et al. (2012) and struvite precipitation and Ash Dec by Linderholm et al. (2012) and thermal conversion (i.e. ENERSLUDGE by Bridle and Skrypski-Mantle (2000)).
Technologies considered in each of the studies.
In the following subsections, an in-depth analysis of some key technologies is presented. It should be noted that the lack of sufficient documentation poses difficulties in fully reconstructing the technical assumptions behind each LCA study. Parameters for emissions, chemical consumptions, energy demand, and process outputs were included in the tables only when clear documentation was provided in the articles.
Thickening and dewatering
Thickening and dewatering processes are used to increase the solids content of sewage sludge, which improves the treatability and transportability of the material. Common thickening processes, such as gravity thickening, dissolved air flotation and gravity belt thickeners, achieve a solids content of 3–6%. Centrifuges, belt filter presses and sludge drying beds are used as a dewatering process, producing a sludge cake with a solids content of 10–30% (Tchobanoglous et al. 2003).
Eighteen studies took thickening and dewatering processes into consideration. The dewatering and thickening processes were largely modelled by increasing the total solids content. Energy and polymer inputs were also accounted for in the environmental burden of thickening and dewatering. The energy use depends strongly on the choice of thickening and dewatering technology: ranging from 0.72 kWh per tonne of sludge treated (d.w.) for reed bed mineralization process to 101 kWh using a centrifugal machine. Yet large discrepancies were also observed within the same technology. For instance, Uggetti et al. (2011) reported an energy consumption rate of 5 kWh per tonne of sludge treated (d.w.) for the centrifugal process, which is one-twentieth of the amount reported by Brown et al. (2010).
Thickening and dewatering processes are known to change the composition of the sludge. Constituents in wastewater are present in either dissolved or particulate form. For instance, a large amount of nitrogen is present in digested sludge as dissolved nitrogen in terms of NH3+ or NO3−, while a range of heavy metals and phosphorus is more likely to be bound to particulates (Hjorth et al., 2009). Thickening and dewatering selectively capture particulates from sludge and increase the concentration of particulate- bound constituents. Houillon and Jolliet (2005) and Lederer and Rechberger (2010) took the removal of these constituents into consideration by applying a specific transfer coefficient for each constituent contained in the sludge. Besides the change in sludge composition, Soda et al. (2010) included the fugitive emissions of CH4 and N2O from thickening and dewatering processes.
Stabilization
Stabilization of sewage sludge can be achieved chemically, biologically, or thermally. The most common stabilization practices are lime stabilization, composting and anaerobic digestion, which were included in 9, 11 and 19 studies, respectively.
Lime stabilization
Nine studies included a lime stabilization process. Lime stabilization was modelled by accounting for chemical and energy inputs. The studies assumed 200–300 kg lime addition per tonne of sludge treated (d.w.). In some studies the dry matter content was increased to 20–25%. Lime addition increased the pH of the sludge to above 11 and could be used to mitigate acidification of soil when applied on land. This will be discussed further in the section ‘Disposal and utilization of stabilized sludge’ below.
Although lime addition is less capital- and labor-intensive than other options, it could have a profound impact on the overall environmental performance, due to the high embedded energy and material requirements for lime production. Murray et al. (2008) reported that lime addition made up 93% of overall fuel consumption and 50% of air pollutant (NOx) emissions.
Composting
Eleven studies included composting, which was modelled by input of fuel and electricity for material turning during the composting process, screening of final products, ventilation and odour control. The energy use ranged from 51 kWh per tonne of sludge treated (d.w.) by Murray et al. (2008) to 305 kWh per tonne of sludge treated (d.w.) by Sablayrolles et al. (2010). Murray et al. (2008) and Sablayrolles et al. (2010) also included the transportation of bulking agents such as yard waste to the composting facilities.
Loss of nitrogen through denitrification and ammonia volatilization was included by Lundin et al. (2000) and Poulsen and Hansen (2002), who assumed losses of 50 and 33%, respectively. They also included a 6% loss of phosphorus during composting, although no documentation was provided for this loss. Brown et al. (2010) recognized CH4 and N2O emissions and assumed 0–2.5% of input C would be emitted as CH4 and 0–4.6% of input N would be emitted as N2O based on the review of 12 field measurements (Brown et al., 2008).
Anaerobic digestion
Nineteen studies included anaerobic digestion. Anaerobic digestion was often modelled as simple energy input and output, but sometimes more elaborate modelling was used. Table 4 summarizes the key assumptions behind the modelling of the anaerobic digestion process. The produced biogas was mainly used for process heating and electricity production, although four studies also considered the portion flared and Bridle and Skrypski-Mantle (2000) and Nakakubo et al. (2012) did not consider substitution by biogas utilization. The biogas generation rate was mainly correlated to the input sludge quality by either a volatile solids destruction rate or a COD removal rate. Others adopted the operational data collected at WWTPs. Fugitive emissions were included in only five studies; CO2, CO, CH4, N2O and non-methane volatile organic compounds. The gas emission rates were either based on plant operational data or taken from literature. In addition, lime for pH control was included by Murray et al. (2008). Biogas often has high levels of hydrogen sulfide and siloxanes, which cause damage to gas engines (Schweiglofler and Niessner, 2001; Syed et al., 2006). Gas scrubbing processes were used to improve the quality of biogas, but none of the studies explicitly mentioned biogas pretreatment processes and chemical input.
Parameters included in anaerobic digestion process.
Y, included; N, not included or documented.
Thermal treatment
Nineteen studies included thermal treatment processes, such as mono-incineration (10), co-incineration (10), wet oxidation (3), pyrolysis (3), melting (2) and two emerging technologies, SUSAN and ENERSLUDGE. Co-incineration of sludge in cement kilns, municipal solid waste incinerators, or coal- fired power plants was considered. Table 5 tabulates the technical assumptions behind mono- and co-incineration processes. Energy input and outputs were largely estimated based on the water content and the heating value of sludge, or were obtained from operational data of existing plants. All the studies except for Nakakubo et al. (2012) considered the substitution of recovered heat.
Parameters included to model sludge incineration process (Y- included; N-not included or documented).
More emissions were indicated but not listed in the articles. Y, included; N, not included or documented.
All studies evaluated the fuel value of sewage sludge as a source of electricity, district heating and process heating. Six studies included the chemical inputs (lime, NaOH, ammonia and activated carbon) for the flue gas treatment processes. Outputs from sludge incineration include bottom ash, fly ash, stack gas emission, filter cake and wastewater from flue gas treatment. The distinction between fly ash and bottom ash was made by only two studies and both of them were taking a mechanistic approach (Cartmell et al. 2006; Lederer and Rechberger, 2010); heavy metal emissions were correlated with the initial sludge quality.
A wide range of stack emissions were evaluated in some studies, including CO2, CO, NO2, NH3, N2, HCl, HF, Hg, particulate matter, SOx, NOx, dioxin and furan. Murray et al. (2008) included the air emissions of supplementary fuel (coal/natural gas) but did not include the emissions from the combustion of the sludge itself.
Incinerated ash was either sent to landfill or utilized as concrete aggregates or raw material for brick production. Lederer and Rechberger (2010), Nakakubo et al. (2012) and Linderholm et al. (2012) considered the option of recovering phosphorus from incinerated ash.
Disposal and utilization of stabilized sludge
Use of sewage sludge on land (UOL) was incorporated in 28 studies. Other disposal practices were also considered: six studies considered sludge landfilling, and five studies included other options for end-use of incinerator ash such as concrete aggregate and brick manufacturing. Sludge end-products were also used for landfill covers, mine land rehabilitation and soil amendment (Hong et al., 2009; Johansson et al., 2008; Murray et al., 2008).
Table 6 shows the technical assumptions for modelling UOL. Most studies assumed that the nutrients (P and N) in the sludge could be substituted for use of conventional fertilizer. Remy and Jekel (2008) used a different nitrogen substitution for each sewage sludge product (50% for sewage sludge, 30% for digestate, 10% for compost). Tarantini et al. (2007) and Renou et al. (2008) did not recognize any benefit of compost to soil and state that compost was applied on land only to improve the soil’s physical properties.
Parameters included to model sludge use on land.
Y, included; N, not included or documented.
Nakakubo et al. (2012), Linderholm et al. (2012), Houillon and Jolliet (2005) and Bridle and Skrypski-Mantle (2000) also included potassium substitution, although both stated that its contribution to the overall results was minor, if not negligible. Chemically stabilized sludge is appreciated for its residual lime and could be used to replace the addition of lime to mitigate acidification of soil. This was counted as a benefit of land application of sewage sludge by reducing the consumption of lime for soil pH adjustment (Houillon and Jolliet, 2005; Peters and Rowley, 2009). The other aspect of UOL is the introduction of heavy metals to arable soil. Ten studies included the toxicity impact assessment of heavy metals, although two studies did not explicitly mention the quantification method. There is no consensus on which elements should be included as ‘heavy metals’ and the number of metals included ranged between four and 11 in the different studies. Trace organic pollutants [polycyclic aromatic hydrocarbons (PAHs), polychlorinated biphenyl (PCBs), bis(2-ethylhexyl)phthalates (DEHPs), nonylphenol ethoxylates (NPEs)] were also included in a recent study by Sablayrolles et al. (2010) who concluded that PAHs contributed the most to human toxicity and eco-toxicity after lead. Hospido et al. (2010) included 12 organic compounds originating from pharmaceutical and personal care products and pointed out that the contribution of trace organics to the overall toxic impact of the target systems ranged from 6.0 to18.6% for human toxicity and from 0.8 to 54.5% for terrestrial eco-toxicity. Foley et al. (2010) considered not only the embedded energy and material use for fertilizer production, but also introduction of heavy metals from application of conventional fertilizer. Although the concentration of some heavy metals (As, Cd, Cr, Ni) in synthetic fertilizer can be higher, the lower nutrient content of stabilized sludge necessitates application of large amounts of biosolids, which results in larger heavy metal loadings to arable land.
Only a few studies considered the nitrogen and carbon balance in agricultural systems. Nitrogen loss by ammonia evaporation and denitrification in UOL were included in seven studies. Two of them also considered the loss from surface water runoff and leaching to ground water. Air emissions of methane from soil were included in four studies. Emissions factors were taken from either Ecoinvent (Doka, 2003, 2007), the IPCC National Greenhouse Gas Inventory Reporting Guideline, or other literature. Sablayrolles et al. (2010) included an extensive list of air emissions from sewage sludge application: benzene, hexane, CO2, CO, CH4, VOCs, NH3, HCl, NOx, SO2, N2O. The included emissions were partly from the operation of heavy equipment in the field. Johansson et al. (2008) included the fugitive emissions of N2O from application of conventional fertilizer, although the amount was almost negligible. While it is common to include carbon storage in soil in solid waste LCAs, only two studies on sludge UOL evaluated carbon sequestration (Brown et al., 2011; Foley et al., 2010).
Discussion
The LCA framework has been applied to a wide range of planning challenges related to sewage sludge management. Sewage sludge management faces a number of regulatory and societal challenges. Stringent water discharge standards increase sludge production and increase the economic burden of wastewater treatment. An attempt to mitigate global climate change has led to limitations on the landfilling of sewage sludge and an expansion of energy recovery from sewage sludge. Conservation of non-renewable resources, such as phosphate ore, has been addressed.
The review also revealed the lack of a common modelling approach for sewage sludge LCA, which could amplify the choice uncertainties. To illustrate this issue, the results for GWP for 10 studies were extracted in Table 7 below. Those studies share similar functional units: the amount of sludge in dry weight treated by similar process configurations. GWP was chosen not only because it is the most commonly assessed impact category by the reviewed articles, but also because the results are less likely to be affected by the choice of LCIA methods, as proven by Renou et al. (2008). First, the geographical area and other local conditions affected the GWP for each study. Brown et al. (2011) calculated the GHG emissions for scenarios with almost identical process configuration but in different geographical areas. The GHG emissions associated with anaerobic digestion followed by land application ranged from −26 to 43 kg CO2-eq per dry tonne of sludge. This difference was largely explained by variations in embedded emissions of electricity production: 0.733 kg CO2 eq. per kWh electricity generation in the Canadian province of Nova Scotia, but 0.01 kg CO2 eq. per kWh electricity in Quebec.
Global warming potential in selected studies in kg CO2-eq per dry tonne of sludge treated.
UOL, use on land; INC, incineration; AD, anaerobic digestion; LF, landfill.
Secondly, variation exists in process and emissions data. As discussed in the section ‘Technical assumptions’ above, the assumptions made for energy and chemical consumption vary greatly between the LCA studies. While those differences in parameter values could be incorporated by statistical uncertainty analysis, some of the noted differences are found to be epistemic.
One example is methane leakage from the anaerobic digestion process. The studies considering methane leakage (Brown et al., 2011; Hong et al., 2009; Hospido et al., 2010; Poulsen and Hansen, 2003) generally found a higher impact on GWP. The field measurements of fugitive emissions from sewage sludge treatment processes are limited and when included, they are largely based on the assumption provided by other GHG accounting guidelines (i.e. IPCC guideline).
Large discrepancies were also found after measurements of sewage sludge leaving the wastewater treatment plant. GWP for landfilling following anaerobic digestion was −745 kg CO2-eq per dry tonne of sludge in one case (Poulsen and Hansen, 2003) and 1173 kg CO2-eq per dry tonne of sludge in another (Peters and Rowley, 2009). Both studies found anaerobic digestion of sewage sludge to produce a negative emission as the biogas was used for producing electricity that replaced electricity from a coal fired power plant, hence avoiding the use of fossil fuels. The difference in the final result stems from the assumption made for modelling sewage sludge landfilling. Poulsen and Hansen (2003) excluded methane emission from landfilled sewage sludge, stating that it was insignificant compared to methane losses from the digestion process. Fugitive emission of GHG after land application of sewage sludge is also another parameter contributing large uncertainty to the outcome of the LCA. Johansson et al. (2008) discussed the uncertainties of N2O and CH4 emissions from the application of sewage sludge on land based on field measurements. Large discrepancies were found in the results: 3200 kg CO2-eq. per tonne of sludge treated at the high end of the estimate and 65 kg CO2-eq. for the low end. The dataset for long-term environmental emission from disposal and utilization of sewage sludge is still scarcely available. The evaluation of the long term consequences of biogenic carbon is also a focal area for methodological development (i.e. Jørgensen and Hauschild, 2013). As for other impact categories, discrepancies can be explained by the limited scientific knowledge on environmental emissions or the absence of LCIA methodologies at the time each assessment was commenced. For instance, Hospido et al. (2010) was the first study which included the toxicological effect of pharmaceutical and personal care products (PPCPs). This study was conducted concurrently with a novel experiment to determine the removal efficiency of PPCPs by anaerobic digestion. Still, little is known about the occurrence and fate of trace organic pollutants such as PPCPs during stabilization processes, and site-specific emissions data is limited. The characterization factor for some organic pollutants present in wastewater are already available (Munoz et al., 2008; Larsen et al., 2007) and further expansion of coverage are anticipated.
Although the methodology of LCIA itself is still evolving to embrace a yet wider range of environmental impacts, many authors also mentioned the insufficiency of current impact assessment methods. Pathogen reduction is the major goal for sanitation practices; other methodologies to quantify the risk of pathogen spread for sewage sludge are still under development, and existing studies have greatly neglected the benefit of pathogen reduction. Sludge is also known to improve the chemical and physical properties of soil such as cationic exchange capacity, soil bulk density, field capacity of water, micro-organism diversity and soil organic carbon contents (Singh and Agrawal, 2008). Few attempts have been made to include these effects; Peters and Rowley (2009) included the water content of sludge as a benefit for rice cultivation. However, efforts are needed in both field data collection and development of the framework for characterizing these benefits.
LCA is based on a number of assumptions and the results of LCA are, in principle, unique to each study’s goal and scope, and comparison between different studies is not typically intended. The integrity of LCA studies depends largely on the transparency, and thorough documentation of underlying assumptions and limitations; doing so is vital to ensure reproducibility of the studies. During the review, the authors encountered a number of problems regarding documentation. Perhaps due to word limits for journals, some articles did not document the procedures to the extent that reconstruction of the studies could be completed. Some studies were supported by more comprehensive project reports, yet those reports were often difficult to access.
Poulsen and Hansen (2003) and Johansson et al. (2008) presented a table listing almost all the parameters chosen for the study. Foley et al. (2010), Stokes and Horvath (2010), Murray et al. (2008) and Brown et al. (2011) included supporting documents describing details of the modelling approaches. Hospido et al. (2004) provided an LCI table with its data source, measurement period and acquisition method of data. Inclusion of an uncertainty analysis is another way to improve the robustness of the assessment. Only seven out of 30 studies documented the sensitivity analysis and only Hong and Li (2011) included the formal evaluation of uncertainty (Table 2). Thorough documentation of assumptions and assessment of uncertainties should be encouraged to assure the quality of LCA modelling of sewage sludge management.
There are already some initiatives to establish a guideline for LCA of solid waste and wastewater management practices. European Commission Joint Research Center has published a technical guideline on the solid waste LCA for waste expert LCA practitioners (EC-JRC, 2011) and International Water Association Working Group for Life Cycle Assessment of Water and Wastewater Treatment (IWA, 2013). The steering committee of the IWA LCA working group has recently published a review of wastewater LCA studies on methodological grounds and is now calling for the harmonization of assessment approaches (Corominas et al., 2013). These efforts will provide a standard for conducting and documenting LCA studies, which would help in reducing the choice uncertainties and ensuring the integrity of studies.
Conclusion
In this study, 35 published studies on sewage sludge LCA were reviewed. It is evident that LCA has provided a significant and flexible framework for assessing the planning and policy challenges related to sewage sludge management. The number of sewage sludge LCAs has increased dramatically during the past decade and this trend is likely to continue, as life cycle thinking is adopted as a core principle for waste management and resource conservation around the World (i.e. Canada, Japan, New Zealand and EU member states (EC-JRC, 2012)).
LCA gives flexibility for researchers to explore a wide variety of goals and scopes but also leads to large discrepancies in both methodological and technical assumptions. While some differences are representative of local conditions or the modelling approach used, it is evident that disagreements in technical assumptions had considerable effects on the outcomes of the studies. There is a need to harmonize the key technological assumptions and create a common platform for documenting the assumptions to improve the reproducibility and consistency of the studies.
LCA studies reflect the current scientific understanding and data availability of the target system. Reliable emission data, such as fugitive emissions from sludge and wastewater treatment and disposal practices, as well as the fate and behavior of emerging pollutants in the sludge treatment processes, need to be determined in order to strengthen the inventory of the LCA. Improvements are also needed in LCIA methodology in order to evaluate benefits of the sludge treatment processes, such as pathogen reduction and recycling of organic matter back to agricultural soil. With the suggested improvements, LCA modelling of sewage sludge management is set to gain further acceptance in assessing the environmental aspects of alternative sludge management.
Footnotes
Acknowledgements
The authors would like to express their gratitude to Senior Researcher Anders Damgaard, (Department of Environmental Engineering, Technical University of Denmark) for his insightful input to discussions. Hiroko Yoshida also acknowledges the generous offer of Technical University of Denmark and 3R Institute for providing PhD Scholarship.
Declaration of conflicting interests
The author declares that there is no conflict of interest.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.