Including Pathogen Risk in Life Cycle Assessment ... - ACS Publications

Jul 24, 2014 - 96 Gothenburg, Sweden. •S Supporting Information ... (LCA) of wastewater and sludge management systems, as this is commonly omitted f...
0 downloads 14 Views 1MB Size
Article pubs.acs.org/est

Including Pathogen Risk in Life Cycle Assessment of Wastewater Management. 2. Quantitative Comparison of Pathogen Risk to Other Impacts on Human Health Sara Heimersson,* Robin Harder, Gregory M. Peters, and Magdalena Svanström Chemical Environmental Science, Department of Chemical and Biological Engineering, Chalmers University of Technology, SE-412 96 Gothenburg, Sweden S Supporting Information *

ABSTRACT: Resource recovery from sewage sludge has the potential to save natural resources, but the potential risks connected to human exposure to heavy metals, organic micropollutants, and pathogenic microorganisms attract stakeholder concern. The purpose of the presented study was to include pathogen risks to human health in life cycle assessment (LCA) of wastewater and sludge management systems, as this is commonly omitted from LCAs due to methodological limitations. Part 1 of this article series estimated the overall pathogen risk for such a system with agricultural use of the sludge, in a way that enables the results to be integrated in LCA. This article (part 2) presents a full LCA for two model systems (with agricultural utilization or incineration of sludge) to reveal the relative importance of pathogen risk in relation to other potential impacts on human health. The study showed that, for both model systems, pathogen risk can constitute an important part (in this study up to 20%) of the total life cycle impacts on human health (expressed in disability adjusted life years) which include other important impacts such as human toxicity potential, global warming potential, and photochemical oxidant formation potential.



INTRODUCTION AND BACKGROUND Agricultural use of sewage sludge is often promoted in order to recycle nutrients (primarily phosphorus and nitrogen) and organic matter. In many European countries, however, agricultural use of sewage sludge is much debated. Most countries restrict agricultural use of sludge (see e.g. EU Council directive 86/278/EEC that regulates the quality of sludge to be used in agriculture), and some countries (e.g., Switzerland and The Netherlands) even prohibit it. Among the concerns expressed by different stakeholders are potentially increasing soil concentrations of heavy metals and organic micropollutants, exposure to pathogenic microorganisms during sludge handling or after spreading, and odor release during transport and spreading of sludge. In order to evaluate the environmental benefits and drawbacks of agricultural use of sludge, it is important to assess these different potential environmental impacts and understand to what extent they contribute to total impacts. Life cycle assessment (LCA) is a valuable tool for the assessment of impacts on human health and the environment from a life cycle perspective. Many LCA studies have been performed for different wastewater and sludge management systems (see the work of Corominas and colleagues1 and Yoshida and colleagues2 for comprehensive reviews). The choice of impact categories (to be assessed in the life cycle impact assessment (LCIA) phase of the LCA) in previous © 2014 American Chemical Society

LCAs reflects the availability of consensus methodology and inventory data for the respective impact categories: often only impact categories for which internationally agreed-upon LCA methodology exists (e.g., climate change and acidification) are assessed while some impacts are more seldom (e.g., human toxicity and odor3) or never (e.g., pathogen risk) covered. The fact that human toxicity and pathogen risk heavily depend on site-specific conditions is one reason they are complicated to assess, and the availability of quantitative risk assessment (QRA) methodology for those categories has historically covered the need for assessment in systems analyses to some extent. Pathogen Risk. Pathogen risk, while the main object of quantitative microbial risk assessment (QMRA), has until now almost never been assessed in LCA studies.4 An attempt to develop a methodology for the inclusion of pathogen risk in LCA was made by Larsen and colleagues;5 the methodology was built on QMRA principles, encompassing human exposure to two bacteria through bathing, but has not been used in studies reported in literature. Aramaki and colleagues6 compared the reduction of pathogen risk associated with the Received: Revised: Accepted: Published: 9446

March 27, 2014 July 11, 2014 July 24, 2014 July 24, 2014 dx.doi.org/10.1021/es501481m | Environ. Sci. Technol. 2014, 48, 9446−9453

Environmental Science & Technology

Article

Figure 1. Flowchart showing the studied model systems. White boxes are common for both model systems. Light gray boxes describe processes specific to model system A and the dark gray box the process specific to model system B.

Aim of Study. This article is the second in an article series based on a study that aimed to investigate the potential for and importance of including pathogen risk in LCA of wastewater and sludge management systems. The approach followed in the study was to hybridize LCA and QMRA, using QMRA methodology for estimating pathogen risk and LCA to cover other relevant potential impacts. Such hybridization requires that QMRA be performed with reference to the same functional unit and system boundaries as LCA. The first part of the article series,15 henceforth referred to as part 1, reported on the estimation of the cumulative pathogen risk associated with a model wastewater and sludge management system by combining disease burden estimates obtained from a QMRA model with disease burden estimates calculated based on data on pathogen risk from previous QMRA studies in the literature. The work started from the same principal idea as previous work by Larsen and colleagues5 and Aramaki and colleagues6 but was more developed as it aimed at including as many relevant pathogens and exposure pathways as possible. It included impacts related to occupational, residential, and recreational exposure during wastewater treatment, from exposure to the effluent, and during and after sludge spreading on agricultural fields. For a system with a wastewater treatment plant (WWTP) treating 12 500 m3/day (serving 28 000 persons) and agricultural use of sludge, the pathogen risk corresponded to 0.2−9 disability adjusted life years (DALYs) per year of operation of the system. The part of the study reported in this article (part 2) aimed at further adding to the research presented in part 1 by quantitatively evaluating the importance of pathogen risk compared to other life cycle impacts on human health associated with wastewater and sludge management and assessed by the use of LCA. The importance of pathogen risk was evaluated for two model systems consisting of wastewater and sludge treatment followed by either agricultural sludge use (model system A) or sludge incineration (model system B).

installation of an urban wastewater system, assessed using QMRA to increased health risks resulting from construction and operation of the treatment plant (assessed using end point indicators in LCA). LCA and QRA have substantial similarities, but also fundamental differences.7,8 LCA assesses impacts relative to the life cycle of a product or a service while QRA assesses risks related to specific substances and is time and sitespecific.9,10 Olsen and colleagues10 stressed the demand for more specific and detailed information on exposure conditions in QRA. A very different approach for assessing a variety of impacts (among them pathogen risk) from wastewater effluent was developed by Papa and colleagues,11 utilizing microbiological assays to quantify the pathogen risk, and comparing those to other impacts in monetary terms, connected to societal costs. Other Life Cycle Environmental Impacts on Human Health. Risks to human health can arise for several reasons, related to e.g. accidents and natural disasters. This study focuses on impacts on human health resulting from utilization of environmental resources and from emissions to the environment in the life cycle of a wastewater treatment service. Climate change, human toxicity, stratospheric ozone depletion, particulate matter, photochemical ozone formation, and ionizing radiation can all be considered in LCA to have an impact on human health.12 For many of these impact categories internationally agreed upon methodology exists. Human toxicity, however, is an impact category where consensus is still strived for, for example through the development of the consensus method USEtox.13 Yoshida and colleagues2 reviewed 35 LCAs on sludge management systems and found that 17 of those assessed human toxicity, as compared to e.g. climate change, acidification and eutrophication which were assessed in two-thirds or more of the studies. Renou and colleagues14 concluded that the choice of LCIA method for human toxicity was critical to the result, especially for where in the life-cycle that hot-spots appeared. 9447

dx.doi.org/10.1021/es501481m | Environ. Sci. Technol. 2014, 48, 9446−9453

Environmental Science & Technology



Article

METHOD In this study, pathogen risk was compared to other impacts on human health using LCA. The procedure for LCAs set by ISO 14044:200616 and the recommendations for LCA in a European context in the International Reference Life Cycle Data Systems (ILCD) Handbook17 were used as guidance. Goal and Scope. The performed LCA was intended to enable an appraisal of whether or not pathogen risk constitutes a significant impact compared to other impacts on human health in the context of wastewater management. To this end, two sludge management configurations were investigated. The LCA was performed for two model systems (Figure 1) both consisting of a WWTP with primary as well as secondary treatment (activated sludge with nitrification/denitrification), chemical phosphorus removal and anaerobic digestion of the mixed sludge. Generated biogas was assumed to be combusted, producing electricity and heat that were used internally at the WWTP. The WWTP configuration was based on a general system described by Foley and colleagues.18 In model system A in this study, sludge is used for agricultural purposes, while in model system B, it is incinerated. Model system A was chosen due to the often expressed stakeholder concerns about potential adverse outcomes related to pathogen exposure in such systems. In model system B, sludge is incinerated which leads to a fully sanitized ash and hence presumably lower overall pathogen risk. The focus was not on comparing different system configurations, but rather on assessing the importance of pathogen risk in different system configurations. The functional unit was chosen to be treatment of 10 000 m3 wastewater per day during 1 day, reflecting the treatment of wastewater as the main service of the studied systems. A European geographical scope was selected. The multifunctionality of model system A (producing both treated wastewater and sludge used as fertilizer) was solved using substitution, as suggested by the ILCD Handbook for LCAs of this type, in this case substitution of mineral fertilizers based on nitrogen and phosphorus content. The thermal energy generated during incineration is assumed to lower the net thermal heat consumption in model system B. The electricity produced from biogas combustion in both systems was considered to replace grid-mix electricity production. In accordance with the aim of the study, to put pathogen risk into the context of a full LCA, end point indicators were used to describe life cycle impacts affecting the safeguard subject human health. Expressing all midpoint impact categories as their impact on human health at end point level, in DALYs, makes it possible to compare their relative contributions to the total results (see the further discussion on this later in this section). Results from using existing LCA characterization methods for the life cycle impacts climate change (global warming potential, GWP), stratospheric ozone depletion (ozone depletion potential, ODP), human toxicity (human toxicity potential, HTP), particulate matter (particulate matter formation potential, PMFP), ionizing radiation (ionizing radiation potential, IRP), and photochemical oxidant formation (photochemical oxidant formation potential, POFP) on human health19,12 were complemented with the pathogen risk results estimated through QMRA (as detailed in part 1) and adjusted to be used in this LCA. The ILCD Handbook does not provide any recommendations on characterization methods for end point indicators, but suggests for human toxicity an interim approach for recalculating USEtox results into end point results

for human toxicity and ReCiPe for many of the other impacts, and therefore, these were chosen to be used in this study. Life Cycle Inventory. Life cycle inventory data was primarily obtained from Foley and colleagues18 (their case 4 Bi corresponds to model system A), whose data set comprises data on net consumption of energy and chemicals consumption in the WWTP, emissions to air from WWTP activities, emissions to water through the effluent, emissions to agricultural soil from the agricultural application of sludge, transports of sludge and chemicals, fuel consumption for sludge spreading on agricultural fields, as well as data on the amount of mineral fertilizer replaced by the sludge when used in agriculture (see the Supporting Information of the work of Foley and colleagues18). Municipal wastewater and sludge contain a wide variety of different pollutants. The described data set contains data on emission of heavy metals in effluent and sludge, while data on emissions of other substances, like organic micropollutants, are lacking. In addition, nitrogen leakage from agricultural fields is missing. For model system B, monoincineration of sludge was assumed. The modeling of the incineration is based on data from Larsen and colleagues;20 the data used can be found in the Supporting Information. Electricity consumption and emissions to air and water was modeled starting from sludge with 20% dry content.18 Production of chemicals and electricity in the WWTP was assumed to be average EU-25, with data from the GaBi Professional database 2013 unless otherwise stated. Ferric chloride production was modeled based on Frohagen.21 For the production of aluminum sulfate, calcium chloride, sodium persulfate, and flocculant and data on the production of general inorganic chemicals, the Ecoinvent database 2.2 was used. Data on mineral fertilizer production was from Davis and Haglund.22 The polyelectrolyte production was modeled as acetonitrile production. Data for pathogen modeling was entirely based on part 1, though adjusted to the functional unit used in the presented study. Note that the pathogen risk estimated in part 1 was not generated for the specific model system A but for a model system comparable in size and in wastewater and sludge management approach. Furthermore, only pathogen risk associated with wastewater treatment and sludge handling (foreground system) were considered. For model system B, the same pathogen risk data was used as for model system A, but excluding exposure routes specific for the agricultural sludge use (other than occupational exposure during truck loading which is assumed to occur also in model system B). Incineration is assumed to result in a pathogen-free ash. Calculations included a sensitivity analysis for the choice of LCIA method for human toxicity, as this has earlier been shown to have a large impact on LCA results.14 Expressing End Point Indicators in DALYs. The use of end point indicators instead of midpoint indicators introduces uncertainties, connected to assumptions regarding the severity of the damage caused per unit of the midpoint indicator. But despite this, end points are sometimes used in order to make results more meaningful, e.g as these “constructed attributes” facilitates weighing of different impacts against each other which may facilitate decision-making.23 The choice to use an end point indicator in this study was motivated partly by the difficulties in coming up with a relevant midpoint indicator for pathogen risk, and partly by the desire to investigate the importance of pathogen risk compared to other potential 9448

dx.doi.org/10.1021/es501481m | Environ. Sci. Technol. 2014, 48, 9446−9453

Environmental Science & Technology

Article

Figure 2. Schematic view of the calculation procedure for calculating end points causing impacts on human health for the specific characterization methods selected for this study: ix = interventions (emissions), ipx = interventions in the form of pathogens, m = midpoints indicators, eh = end point indicators causing impacts on human health, npx = number of infections per specific pathogen type, eh,px = end point indicator for a specific pathogen type, Q = conversion factor. Dashed arrows indicate that the selected worldview has potential to impact the calculations according to ReCiPe (for its indicators). One intervention can contribute to several impacts, and several interventions can contribute to the same impact. Either the end point indicator is calculated via aggregated midpoint results (as for e.g. GWP), directly for each intervention, and then summarized (e.g., for pathogens) or via subgroups of interventions (e.g., as for ODP).

generally based on Qem characterization factors calculated from average DALYs for different types of medical conditions, weighted by incidence cases, as is done, e.g., by Huijbregts and colleagues.24 Goedkoop and colleagues12 point out the problems of assessing region-specific human health impacts with world average DALYs, as is commonly done in LCA, as DALYs can differ substantially depending on the geographical boundaries of the DALY calculations, about 2-fold for cancer diseases and 5-fold for noncancer diseases.12,25 The ReCiPe method also offers possibilities for adjusting characterization factors for some of the impact categories, e.g. the factor for calculating the end point levels for GWP and ODP, based on three different worldviews (individualist, hierarchist and egalitarian). These affect the choices made regarding uncertain parameters e.g. the choice of impact time horizon. In the assessment presented in this article, the default hierarchist worldview is used, which is based on the most common policy principles with regard to issues like time frame, and assumes mean relative impact on human life years for all diseases, and very small consequences from diarrhea for people in developed countries.12 In part 1, pathogen emissions and exposure were converted into pathogen risk using pathogen-specific severity factors representing the burden of disease per case of infection (or illness) in DALYs. These severity factors are generated in a similar way as those used in LCIA and represent averages for different medical conditions weighted based on the number of incidences without applying age-weighting or discounting of future health damages. The severity factors used in Part 1 are nation-specific (e.g., data for The Netherlands reported by Havelaar and colleagues26). Since Europe is set as the geographical scope for this study, the choice of severity factors calculated for Europe and North America are considered sufficiently accurate. As the severity factors are pathogenspecific, they are comparable to the Qei characterization factors

impacts on human health, which called for a common denominator. As described in part 1, DALYs are used both in LCA and QMRA methodology as a measure of the burden of disease. But, as the practical use of this concept varies slightly, it is important to ensure that the generated results are consistent enough to be put in the same LCA framework. This section describes how end point indicators are generated in the LCA method ReCiPe (used in this study for many of the impact categories) and how corresponding end point indicators were generated for pathogens in part 1. Figure 2 provides a schematic view of the process for calculating end points for the different impacts assessed in this study. In the ReCiPe method,12 a distinction is made between two ways of generating end points. The midpoint characterization factors (Qmi) couple the intervention i (e.g., emissions or resource extractions) to the midpoint impact indicator m. End point characterization factors (Qem) couple the midpoint impact indicator m to the end point e. Combined characterization factors (Qei) directly couple the intervention i to the end point e. ReCiPe offers (Qmi + Qem) or Qei for each impact category, but not both. The Qem is specific for each midpoint impact category but not for each specific intervention, except for ODP for which specific end point characterization factors for subgroups of substances exist.12 The end point impact for human health is normally described by the number of DALYs, which is a measure of the years of life lost (YLL) and the years of life as disabled (YLD) in relation to the studied unit. The ReCiPe method12 includes both of these parameters in their characterization factors for human health, unlike some other LCIA methods which only include YLL in order to avoid the need to weight health disabilities for different diseases. As for most other LCIA methods, ReCiPe does not apply ageweighted DALYs and does not discount future health damages. Human health impacts expressed as DALYs in LCA are 9449

dx.doi.org/10.1021/es501481m | Environ. Sci. Technol. 2014, 48, 9446−9453

Environmental Science & Technology

Article

for ODP as described above, but are more specific since specific sludge exposure pathways and DALY characterization factors are used, see Figure 2. Finally, the pathogen risk assessment was made without any specific worldview in mind, so the choice of the hierarchist default is a natural choice for the other assessed impacts. The comparability of the results for the different impact categories are further commented on in the Discussion section.



RESULTS For both assessed model systems, pathogen risk was found to be important compared to other life cycle impacts on human health, as illustrated by their relative contributions in Figure 3.

Figure 4. Overall impact on human health for model systems A and B using USEtox or ReCiPe (based on USES-LCA) for calculating human toxicity. Calculated with the higher risk estimate.

Thus, it seems to be important to include pathogen risk in LCAs of wastewater and sludge management systems, regardless of whether agricultural use or incineration of the sludge is assumed. In fact, although the actual pathogen risk was assumed to be lower in model system B, the contribution from pathogen risk to the overall impact on human health for that system is in the same range as for model system A, as the overall impact from model system B is smaller (see the first two bars in Figure 4), mainly depending on a lower contribution from HTP.



DISCUSSION The inclusion of pathogen risk in environmental assessments of sludge management systems is strongly motivated by stakeholder concerns. This inclusion could be achieved either by separately presenting the results of two independent assessments (i.e., a QMRA assessing pathogen risk and an LCA assessing other impacts on human health) or by integrating pathogen risk in an LCA framework as in this study. The question of which approach is preferable can be debated. The consideration of two sets of results in parallel introduces the need for decision-makers to weigh different indicators against each other. This either happens tacitly or in a multi criteria decision analysis (MCDA) framework with its own needs for subjective weighting factors.23 This study explored the application of QMRA with the intent to include pathogen risk in LCA (part 1), and evaluated whether pathogen risk can be important in terms of its contribution to the total life cycle impacts on human health for wastewater and sludge management systems (part 2). It draws no conclusions on whether or not an integration of QMRA and LCA is preferable in different situations, but in general, a key challenge to MCDA use is the availability of an appropriate, representative stakeholder group for weighting processes, so our alternative approach may be easier to apply than MCDA for future-oriented studies, for example. An extensive discussion on the sensitivity of the pathogen risk results to different input parameters are provided in Part 1. The most relevant parameters for the results presented in this article are the assumed pathogen concentration in the sludge, the selection of pathogens assessed and the exposure assumptions made throughout the assessment. Evaluation of the Pathogen Risk Assessment As a Characterization Method. The ILCD Handbook27 contains criteria for evaluating characterization methods to be used in LCIA. The method used for calculating pathogen risk as described in part 1 is an attempt to adjust QMRA methodology

Figure 3. Relative contributions from different types of impacts to the overall impact on human health from the studied model systems. The first set of bars (a and b) show pathogen risk contribution for model system A depending on whether the pathogen risk is modeled with (a) minimum or (b) maximum risk estimates as calculated in part 1. The second set of bars (c and d) show the corresponding results for model system B.

For each model system, two bars are shown, representing the minimum or maximum pathogen risk (reflecting variations in parameter values such as pathogen concentrations in various compartments as well as fate and transport, exposure, dose− response and severity assessment parameters) as calculated in Part 1. When using the higher pathogen risk estimate, for model system A, the largest contributions are from HTP (to which the largest contributor is the effects of heavy metal emissions on agricultural land), GWP and PMFP (both connected mainly to the electricity production) and pathogen risk; for model system B, the largest contributions are from GWP, pathogen risk, PMFP and HTP (for the same reasons as in model system A). When using the lower pathogen risk estimate, the pathogen risk becomes almost negligible. However, this is partly a result of the LCIA method that was selected for human toxicity. Figure 4 shows the total impacts on human health including the higher pathogen risk estimate in absolute values for the two model systems, calculated using two alternative LCIA methods for human toxicity (ReCiPe based on USES-LCA and USEtox). When ReCiPe (i.e., USES-LCA) is applied to human toxicity, its contribution becomes very small for model system A, which causes pathogen risk to have greater relevance. This also has the effect that the two systems become more comparable in terms of their total impact on human health. 9450

dx.doi.org/10.1021/es501481m | Environ. Sci. Technol. 2014, 48, 9446−9453

Environmental Science & Technology

Article

Figure 5. Human toxicity (in DALYs) for model systems A and B, calculated using (a) USEtox as suggested in the ILCD Handbook and (b) USESLCA (ReCiPe) characterization methods. Contributions from different life cycle stages are shown. Note the different scales of the y-axes in the two graphs.

difference between the pathogen risk as calculated in part 1 and other impacts on human health as calculated in part 2, namely the spatial scale of the model. The QMRA model relies on specific exposure pathways and is sensitive to the location and timing of the exposure to pathogens. The LCIA models, in contrast, operate on the continental scale rather than the local scale. The comparability of the QMRA results specific to the model system and conventional LCA results can therefore be questioned. The calculations in part 1, however, represent a certain extent of generality since the parameters used in the calculations represent general values representative of a given wastewater and sludge management option rather than a specific wastewater and sludge treatment site. This study only includes pathogen risk resulting from the foreground system, in contrast to the other assessed impact categories for which the background system is included to a larger extent. The foreground system is, however, likely to be the main source of pathogen risk (unless there are major inputs of agricultural products to the foreground system). Sensitivity to Human Toxicity Impact Assessment Method. The relative importance of the contribution from pathogen risk to the overall human health impact is sensitive to the choice of LCIA method for human toxicity. The difference of several orders of magnitude for the HTP depending on characterization method chosen (Figure 3) is interesting but not unexpected. USEtox is a consensus model, developed based on several earlier methods, one of them is USES-LCA. Rosenbaum and colleagues13 tested the different characterization methods for a model scenario and concluded that USEtox results and results generated with earlier methods on which USEtox is based, differs with a factor of 100−1000 for human health impacts. Owsianiak and collegues28 also demonstrated large differences between USEtox and ReCiPe (USES-LCA) results for their case study, and partly explained this difference by the fact that certain substances that add major contributions in USEtox are less important in ReCiPe (USESLCA). The authors pointed out Zn2+ as one such substance, which is likely to be an explanation also to the large differences between the USEtox and ReCiPe results in Figure 4, as Zn2+ is one of the major contributors to the USEtox results for model system A. It is also interesting that the relative difference between the model systems’ contributions to the impact category human toxicity changes. In particular, if one uses ReCiPe (USES-LCA), model system A actually gets a better performance than model system B, as can be seen in Figure 5. This mainly depends on the characterization factors for Zn2+ to agriculture and Hg2+ to air in the different methods, as these are the main contributors to the results in Figure 5 for model

to generate results that can be introduced in an LCA framework. As the model is only a first attempt toward developing a generic characterization method that is fully compatible with impact assessment in LCA, a full evaluation based on these criteria is not useful at this point. The general criteria and subcriteria have here instead been used as a basis for a qualitative evaluation. Completeness of Scope. The developed method can be used for comparative purposes and it covers the most relevant impact mechanisms for the human health area of protection. However, the method was developed and applied only for pathogen risks related to wastewater and sludge management. It cannot yet be used for assessing pathogen risks from the background system or from other types of systems. Environmental Relevance. The model is based on established QMRA methodology and considers fate, exposure and effect in a quantitative way, with the delimitations described above. The main impact pathways are covered. Scientific Robustness and Certainty. The fact that the method is based on an earlier accepted QMRA method suggests that the method relies on a sound scientific basis. However, in the context of LCA the method is in an early state of development and has only been peer-reviewed to a limited extent. Documentation, Transparency, and Reproducibility. The principal approach and calculation method is published in part 1 of this article series. No other public documentation exists at this stage. Applicability. The method is not a classical impact assessment method in LCA where characterization factors for interventions are used to calculate an impact potential. Instead, the method requires specific modeling for specific systems and does therefore require some additional effort from the LCA analyst. This introduces the opportunity for a higher level of specificity, however, at the cost of a greater data collection and calculation effort. The method only generates end point results. It can be concluded that despite shortcomings inherent to novel methods in general, the pathogen risk assessment method is useful for the purpose of this specific study. A more general application of the model requires further method development and application in different case studies. Comparability of End Point Indicators for Different Life Cycle Impact Categories. The pathogen risk, expressed in DALYs, calculated in part 1, and the end point impact for the other impact categories generated by the ReCiPe characterization methods and USEtox for the model systems, also expressed in DALY, are considered to be comparable for the purpose of this study. However, there is one important 9451

dx.doi.org/10.1021/es501481m | Environ. Sci. Technol. 2014, 48, 9446−9453

Environmental Science & Technology



system A and model system B, respectively. In connection to this, it is important to remember that the LCA described in this article is limited to human toxicity caused by heavy metals, ignoring potentially important organic micropollutants, due to lack of inventory data and characterization factors. The contribution to HTP should, thus, be higher than estimated in this study, but there are, so many other issues involved in HTP assessment today that it is difficult to estimate the importance of this. In addition, characterization factors for heavy metals are identified as having a relatively high degree of uncertainty compared to other characterization factors, for both tested methods,13,29 and results should therefore be interpreted with care. Results similar to the ones in Figure 4 have been shown in literature. Hospido and colleagues30 assessed human toxicity of different sludge treatments combined with agricultural sludge utilization, using CML 2 (2000) (which used USES-LCA for the toxicity modeling) and concluded that the impact was dominated by the emissions of heavy metals to soil. Peters and Rowley31 showed the opposite outcome: in their study of alternative sludge management scenarios with agricultural sludge application, heavy metals to soil gave a much smaller contribution than the one resulting from electricity and transports. Peters and Rowley31 argued that the toxicity model they applied (EUSES 2.0 under Australian conditions) did not take the speciation of heavy metals into account, and they therefore assumed that only the exchangeable fraction (leached out by a CaCl2-solution in a sequential extraction scheme32) of the heavy metal concentrations should be regarded as bioavailable and thus used as input for the EUSES model. If all heavy metals had been assumed to be bioavailable, the result would have been closer to that of Hospido and colleagues (see the Supporting Information in the work of Peters and Rowley31). Comparison of Model Systems. Although it was not the purpose of this study, it is tempting to use the results of this study for comparing the different sludge management scenarios in model systems A and B. As shown in Figure 4, there seems to be a large difference in the overall impact between the model systems, which is expected. Whether the possible difference is important or not is largely value-based. Decision makers need to be able to, in a structured way, compare pathogen risk to other potential impacts on human health, and we argue that the approach described and tested in this article series is one suitable way of doing this, and better than ignoring the trade-off in analyses. When applying such methods in different case studies, insights might lead to further method development and to more data being generated or gathered that will improve future assessments. Outlook. This study, applying a combination of adapted QMRA and LCIA methodology, showed that pathogen risk can potentially represent an important part of the overall impacts on human health in the context of wastewater and sludge management, regardless of the sludge handling approach (i.e., agricultural use or incineration). Therefore, pathogen risk should not be neglected in environmental assessments of wastewater and sludge management systems. However, when sludge is used in agriculture, pathogen risk may in some situations be dwarfed by potential impacts of exposure to heavy metals and chemicals. This highlights the great importance of including also the impacts of these pollutants in such assessments. The results were sensitive to the pathogen data used and to the human toxicity LCIA method chosen.

Article

ASSOCIATED CONTENT

S Supporting Information *

Details on the inventory for modeling sludge incineration in model system B. This material is available free of charge via the Internet at http://pubs.acs.org/.



AUTHOR INFORMATION

Corresponding Author

*Phone: +46 (0)31 772 28 32. E-mail: sara.heimersson@ chalmers.se. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This project has received funding from the European Union’s Seventh Programme for research, technological development, and demonstration under grant agreement no. 265156 ROUTES. This research at the same time served as a prestudy for further research funded by the Swedish Research Council Formas under grant agreement no. 2012-1122.



REFERENCES

(1) Corominas, L.; Foley, J.; Guest, J. S.; Hospido, A.; Larsen, H. F.; Morera, S.; Shaw, A. Life cycle assessment applied to wastewater treatment: State of the art. Water Res. 2013, 47 (15), 5480−5492. (2) Yoshida, H.; Christensen, T. H.; Scheutz, C. Life cycle assessment of sewage sludge management: A review. Waste Manage. Res. 2013, 31 (11), 1083−1101. (3) Peters, G. M.; Murphy, K. R.; Adamsen, A. P.; Bruun, S.; Svanström, M.; ten Hoeve, M. Improving odour assessment in LCA the odour footprint. Int. J. Life Cycle Assess, DOI: 10.1007/s11367-0140782-6. (4) Larsen, H. F.; Hauschild, M. Z. Review of existing LCA studies on waste water treatment technologies. SETAC Europe 18th Annual Meeting, Warsaw, Poland, 2008. (5) Larsen, H. F.; Olsen, S. I.; Hauschild, M.; Laurent, A. NEPTUNE, New sustainable concepts and processes for optimization and upgrading municipal wastewater and sludge treatment, Work Package 4−Assessment of environmental sustainability and best practice, Deliverable 4.2− Methodology for including specific biological effects and pathogen aspects into LCA; Technical University of Denmark: Lyngby, Denmark,2009. (6) Aramaki, T.; Galal, M.; Hanaki, K. Estimation of reduced and increasing health risks by installation of urban wastewater systems. Water Sci. Technol. 2006, 53, 247−252. (7) De Haes, H. A. U.; Sleeswijk, A. W.; Heijungs, R. Similarities, differences and synergisms between HERA and LCA - An analysis at three levels. Hum. Ecol. Risk Assess. 2006, 12 (3), 431−449. (8) Sleeswijk, A. W.; Heijungs, R.; Erler, S. T. Risk assessment and life-cycle assessment: Fundamentally different yet reconcilable. Greener Manage. Int. 2003, 41, 77−87. (9) Russell, A. J. Human and ecological risk assessment and life cycle assessment: Intersections, collisions, and future directions. Hum. Ecol. Risk Assess. 2006, 12 (3), 427−430. (10) Olsen, S. I.; Christensen, F. M.; Hauschild, M.; Pedersen, F.; Larsen, H. F.; Tørsløv, J. Life cycle impact assessment and risk assessment of chemicals - A methodological comparison. Environ. Impact Asses. 2001, 21 (4), 385−404. (11) Papa, M.; Pedrazzani, R.; Bertanza, G. How green are environmental technologies? A new approach for a global evaluation: The case of WWTP effluents ozonation. Water Res. 2013, 47 (11), 3679−3687. (12) Goedkoop, M.; Heijungs, R.; Huijbregts, M. A.; De Schryver, A.; Struijs, J.; van Zelm, R. ReCiPe 2008 - A life cycle impact assessment method which comprises harmonised category indicators at the midoint and the endpoint level, Report 1: Characterisation; Ruimte en Milieu, 9452

dx.doi.org/10.1021/es501481m | Environ. Sci. Technol. 2014, 48, 9446−9453

Environmental Science & Technology

Article

Ministerie von Volkhuisvesting, Ruimtelijke Ordening en Milieubeheer, 2013. (13) Rosenbaum, R. K.; Bachmann, T. M.; Gold, L. S.; Huijbregts, M. A. J.; Jolliet, O.; Juraske, R.; Koehler, A.; Larsen, H. F.; MacLeod, M.; Margni, M.; McKone, T. E.; Payet, J.; Schuhmacher, M.; Van De Meent, D.; Hauschild, M. Z. USEtox - The UNEP-SETAC toxicity model: Recommended characterisation factors for human toxicity and freshwater ecotoxicity in life cycle impact assessment. Int. J. Life Cycle Ass. 2008, 13 (7), 532−546. (14) Renou, S.; Thomas, J. S.; Aoustin, E.; Pons, M. N. Influence of impact assessment methods in wastewater treatment LCA. J. Clean. Prod. 2008, 16 (10), 1098−1105. (15) Harder, R.; Heimersson, S.; Svanström, M.; Peters, G. Including pathogen risk in life cycle assessment of wastewater management. 1. Estimating the burden of disease associated with pathogens. Environ. Sci. Technol. 2014, DOI: 10.1021/es501480q. (16) ISO 14044:2006 Environmental management−life cycle assessment−requirements and guidelines; International Organization for Standardization: Geneva, Switzerland, 2006. (17) European Commission Joint Research Centre, ILCD Handbook−International Reference Life Cycle Data System, first ed.; European Union: 2010. (18) Foley, J.; de Haas, D.; Hartley, K.; Lant, P. Comprehensive life cycle inventories of alternative wastewater treatment systems. Water Res. 2010, 44 (5), 1654−1666. (19) European Commission Joint Research Centre, ILCD Handbook−International Reference Life Cycle Data System−Recommmendation for Life Cycle Impact Assessment in the Euroean context, first ed.; European Union: 2011. (20) Larsen, H. F.; Hansen, P. A.; Boyer-Souchet, F. NEPTUNE− New sustainable concepts and processes for optimization and upgrading municipal wastewater and sludge treatment, Work Package 4−Assessment of environmental sustainability and best practice, Deliverable 4.3−Decision support guideline based on LCA and cost/efficiency assessment, Technical University of Denmark: Lyngby, Denmark, 2010. (21) Frohagen, J. Livscykelanalys på tre järnbaserade fällningskemikalier (Life cycle assessment of three ferric based precipitation chemicals). Chalmers Tekniska Högskola, Avdelningen för teknisk miljöplanering, 1997. (22) Davis, J.; Haglund, C. Life cycle inventory (LCI) on fertiliser production. Chalmers University of Technology, 1999. (23) Rowley, H. V.; Peters, G. M.; Lundie, S.; Moore, S. J. Aggregating sustainability indicators: Beyond the weighted sum. J. Environ. Manage. 2012, 111, 24−33. (24) Huijbregts, M. A.; Rombouts, L. J.; Ragas, A. M.; van de Meent, D. Human-toxicological effect and damage factors of carcinogenic and noncarcinogenic chemicals for life cycle impact assessment. Integr. Environ. Assess. Manage 2005, 1 (3), 181−244. (25) Murray, C.; Lopez, A. Global Burden of Disease: A comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020; Harvard School of Public Health, World Bank, World Health Organisation, 1996; Vol. 1. (26) Havelaar, A. H.; Haagsma, J. A.; Mangen, M. J. J.; Kemmeren, J. M.; Verhoef, L. P. B.; Vijgen, S. M. C.; Wilson, M.; Friesema, I. H. M.; Kortbeek, L. M.; van Duynhoven, Y. T. H. P.; van Pelt, W. Disease burden of foodborne pathogens in the Netherlands, 2009. Int. J. Food Microbiol. 2012, 156 (3), 231−238. (27) European Commission Joint Research Centre, ILCD Handbook−International Reference Life Cycle Data System−Framework and requirements for Life Cycle Impact Assessment models and indicators; first ed.; European Union: 2010. (28) Owsianiak, M.; Laurent, A.; Bjørn, A.; Hauschild, M. Z. IMPACT 2002+, ReCiPe 2008 and ILCD’s recommended practice for characterization modelling in life cycle impact assessment: a case study-based comparison. Int. J. Life Cycle Assess. 2014, 19 (5), 1007− 1021. (29) Van Zelm, R.; Huijbregts, M. A. J.; Van De Meent, D. USESLCA 2.0-a global nested multi-media fate, exposure, and effects model. Int. J. Life Cycle Asses. 2009, 14 (3), 282−284.

(30) Hospido, A.; Carballa, M.; Moreira, M.; Omil, F.; Lema, J. M.; Feijoo, G. Environmental assessment of anaerobically digested sludge reuse in agriculture: Potential impacts of emerging micropollutants. Water Res. 2010, 44 (10), 3225−3233. (31) Peters, G. M.; Rowley, H. V. Environmental comparison of biosolids management systems using life cycle assessment. Environ. Sci. Technol. 2009, 43 (8), 2674−2679. (32) Peters, G. M.; Maher, W. A.; Barford, J. P.; Gomes, V. G. Selenium associations in estuarine sediments: Redox effects. Water, Air, and Soil Poll. 1997, 99 (1−4), 275−282.

9453

dx.doi.org/10.1021/es501481m | Environ. Sci. Technol. 2014, 48, 9446−9453