Charting a Path for Innovative Toilet Technology Using Multicriteria Decision Analysis

Policy Analysis

Charting a Path for Innovative ToiletTechnology Using MulticriteriaDecision AnalysisM A R K E . B O R S U K , * , † , ‡ M A X M A U R E R , †

J U D I T L I E N E R T , † A N D T O V E A . L A R S E N †

Eawag, Swiss Federal Institute of Aquatic Science andTechnology, 8600 Dübendorf, Switzerland

Received August 30, 2007. Revised manuscript receivedJanuary 03, 2008. Accepted January 04, 2008.

Practical and theoretically sound methods for analyzinginnovative environmental technologies are needed to informpublic and private decisions regarding research and development,risk management, and stakeholder communication. Byintegrating scientific assessments with a characterization ofvalues,multicriteriadecisionanalysis (MCDA)supports therankingof alternative technology pathways on the basis of technical,financial, and social concerns. We applied MCDA to evaluate theuseofNoMixurineseparatingtoilets formanagingenvironmentalrisk and postponing expensive upgrades to a large wastewatertreatment plant near Zürich, Switzerland. Results indicate that,given current priorities, no single, fixed course of action(including the status quo) will be desirable to all stakeholdersover the considered time horizon. However, a path forwardis suggested that is not significantly disadvantageous to anystakeholder now and leaves open future options, allowing societyto achieve overall greater benefits if priorities change, newenvironmental risks are revealed, or technology improves. Whileour analysis focuses on a particular catchment in Switzerland,many communities worldwide are faced with an aging andinefficient wastewater treatment infrastructure while alsoexperiencing growth and development. Our framework canhelp these communities balance the conflicting objectives ofdiverse stakeholders and gain insight into the role that urineseparation can play in transitioning to a more comprehensiveand sustainable urban water management system.

IntroductionEmerging environmental concerns often generate techno-logical innovation (1). However, if conventional technologiesare deeply entrenched, innovative substitutes may not beadopted, despite their superiority. This technological “lock-in” may occur for any number of reasons, includinginstitutional, political, and economic commitments (2) orhabits of behavior, learning, and culture (3). It may also bethat those stakeholders in the best position to affect adoptionof the new technology do not have a view of the long-termadvantages nor a clear path for implementation.

The sanitary system of the modern Western world providesa case in point. The conventional methods of wastewater

collection, transport, and treatment are extremely wastefulof nutrients, which could otherwise be recycled for agricul-tural use. They are also very water inefficient, using largeamounts of water for transport. In combined sewer systems,this leads to overflow problems during storm events andassociated water pollution. Finally, micropollutants, thosecompounds, such as pharmaceuticals, which may havesignificant health effects at low concentration, go largelyuntreated in the current system (4). All of these problems areespecially grievous when this Western technology is exportedto developing countries with the promise that it will solveexisting water pollution problems.

Recently, the NoMix (or urine separating) toilet has beendeveloped as an innovative approach to sustainable waste-water management (5–7). The NoMix approach is an exampleof both a source separation and waste design strategy (8, 9).The technology consists of a specially designed toilet in whichurine is collected separately (Figure 1), a urine storage tankin or near the household, a treatment unit located eitherlocally or centrally that processes the urine, possibly recov-ering nutrients in the process, and a means for transportingthe urine from the storage tanks to the treatment unit.

Urine is the targeted waste type of the NoMix technologybecause it contains most of the nutrients (5) and 60–70% ofmicropollutants, including natural and synthetic hormones,antibiotics, and other pharmaceuticals (10, 11). Thus, sepa-rating urine at the source, rather than handling it as part ofa mixed waste stream at the treatment plant, can be an elegantand efficient way of dealing with micropollutants andfacilitating nutrient recycling (12). Furthermore, wastewaterthat does not contain urine has a biologically balanced C/P/Nratio, making it possible to achieve excellent treatment resultswith compact treatment plants focused on COD removal(5, 13, 14).

While conceptually simple, the NoMix technology concepthas proven significantly more difficult to implement (15). Aswith other innovations, this may be because successfuladoption depends not only on the priorities of users, but onthose of developers and suppliers as well. Surveys have shownhousehold users to be receptive to the NoMix concept, solong as the current standards of bathroom convenience andhygiene are maintained and no additional costs are incurred(16, 17). However, while toilet manufacturers believe theycan develop NoMix units that meet these requirements, theydo not see an immediate market for their product (18). Thewastewater treatment industry sees the potential to use theNoMix technology to help them meet environmental stan-dards at lower cost but is unsure of the appropriateimplementation strategy (19). Finally, scientists recognizean interesting research opportunity but would like to knowwhat projects would most likely make the emerging technol-ogy practical before investing significant time and effort.Clearly, all these stakeholders have a role in the successfulrealization of the NoMix concept.

In addition to differing stakeholder concerns, there arediffering types and levels of NoMix technology implementa-tion. For example, urine storage can be either short- or long-term at either a local or a centralized location. This choiceaffects the type and cost of treatment options that areavailable. Furthermore, it is unlikely to be cost-effective toreplace all existing toilets in an area at one time. Thus, varioustransition scenarios are possible, some of which may havebenefits that are quite different from those that will ultimatelyaccrue. For example, even before nutrient recycling ormicropollutant removal begin to take place on a large scale,

* Corresponding author phone: 1-603-646-9944; fax: 1-603-646-2277; e-mail: [email protected].

† Eawag.‡ Current address: Thayer School of Engineering, Dartmouth

College, Hanover, New Hampshire.

10.1021/es702184p CCC: $40.75 2008 American Chemical Society VOL. 42, NO. 6, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1855

Published on Web 02/08/2008

separately collected and stored urine can be used to smoothout the daily fluctuations in nutrients that occur in the presenttreatment systems. The peaks of these fluctuations, whicharise from society’s daily cycle of sleeping, eating, drinking,and using the toilet, largely determine the size, cost, andefficiency of treatment plants. “Shaving” these peaks canlead to significant near-term savings by decreasing therequired design capacity of new plants or prolonging thelifetime of existing plants (5, 20).

Another consideration is that some of the issues involvedin deciding whether to implement NoMix technology arenot yet resolvable. For example, there are many uncertaintiessurrounding the level and consequences of micropollutantsin wastewater emissions (4). Without knowing the presentrisks, it is impossible to accurately assess the future potentialbenefits of pollutant reduction. Thus, later stages of imple-mentation may need to be decided adaptively, as we learnfrom initial stages. However, decisions made now may limitavailable options in the future. This aspect of path depen-dence is a widely recognized challenge for technology futuresanalysis (21).

We believe that an application of multicriteria decisionanalysis (MCDA) with an emphasis on value-focused thinking(22) can help to structure the NoMix evaluation process andpossibly clarify the important issues of implementation.MCDA has been proposed previously for analyzing suchemerging technologies and their implications (23, 24).However, it has been pointed out that multicriteria modelsare usually static in their temporal structure, making itdifficult to address issues that pertain to different or varyingtime horizons (23). Another difficulty is that decisionalternatives of an MCDA are often assumed to be mutuallyexclusive, an assumption that may not be reasonable fortechnologies that evolve with time in unknown ways.Implementing a less than optimal technology now, forexample, may allow society to achieve greater benefits in thefuture by leaving open more options. This was explicitlyrecognized by Keeney and McDaniels (25) in the context ofclimate change policy. We hope to build upon their meth-odological development.

One of the goals of our analysis is to exemplify theassessment of a particular technology using the informationthat is currently available. We acknowledge that there aregaps in our knowledge and many uncertainties that are bothaleatory and epistemic. However, we are fortunate enoughto be conducting this analysis near the end of a six-yearinterdisciplinary research project on the social and technicalaspects of the NoMix system (www.novaquatis.eawag.ch).Therefore, we can expect that our project team knows aboutas much as there is to know about this technology at this

time and that there will always be some uncertainty that cannever be fully resolved. Thus, rather than expend undue effortto gather more information, we make use of sensitivityanalysis to learn what information could make the mostdifference for the decision. Collecting this information willthen provide the basis for future work.

Problem DescriptionStudy Location. Although the NoMix toilet technology ispotentially promising for a wide variety of conditions andsettings, to make our analysis concrete, we will focus on asituation concerning a proposed mixed use developmentnear Zürich, Switzerland, called Glattpark (http://www.glat-tpark.ch). Glattpark is being developed on a 67 ha site andincludes multifamily housing for 7000 inhabitants as well asoffice space for another 7000 workers. The wastewater ofGlattpark is to be handled by the treatment plant in nearbyKloten/Opfikon, which is already operating near full capacity.Thus, there is hope that urine source separation in Glattparkcan be used to postpone or avoid an expensive expansionof the Kloten/Opfikon plant. Additionally, the developersand city government have committed to environmentallysustainable management of the site, thus making it attractiveto pursue efforts that will help close the nutrient cycle.

Stakeholders. Certainly, to be successful, NoMix toiletsmust be acceptable to household users. However, the fullpathway to implementation still requires significant researchand development effort. Therefore, in addition to households,we consider wastewater treatment plant managers, sanitaryequipment manufacturers, engineering consultants, andresearch scientists to all be key stakeholders in the NoMixtechnology. While the government at various levels mightalso be considered a stakeholder, because of the public natureof the wastewater treatment industry in Switzerland, weconsider the views of the treatment plant managers to reflectthe requirements imposed by the government.

A previous study of NoMix (26) also considered farmersas stakeholders because the nutrients potentially recoveredfrom urine can be used to produce agricultural fertilizer. Theresults of the study revealed that farmers would use the urine-based fertilizer so long as it did not cost more than theircurrent fertilizer product nor present any risks to humanhealth. Therefore, rather than include the farmers as stake-holders, we decided to make their requirements a precondi-tion to the rest of the analysis. If a urine-based fertilizer canbe produced that is as cheap, safe, and convenient as thecurrent product, then the details of the urine collection andrecycling process will be immaterial to the farmers.

Decision Alternatives. Unfortunately, the decision re-garding NoMix technology for the Glattpark development isnot independent of how the rest of the catchment of theKloten/Opfikon treatment plant is handled. This is becauseseparate urine treatment with subsequent nutrient recyclingis only economically practical in the long term if most of thecatchment implements urine source separation. This wouldallow substantial downsizing and simplification of thetreatment plant with attendant long-term cost savings.Therefore, we conceive of the NoMix decision for our casestudy as a hierarchy of decisions, stated as follows: (1) Isurine collection and conversion to fertilizer an appropriatelong-term goal for the entire Kloten/Opfikon catchment? (2)If the answer to question 1 is “No,” then can urine separationin Glattpark alone be used as a peak-shaving strategy topostpone expansion of the Kloten/Opfikon plant? (3) If theanswer to question 1 is “Yes,” then how should urine fromGlattpark be handled during the transition period? Inparticular, should peak shaving be implemented as anintermediate step? (4) If peak shaving is an appropriate partof either the long-term or transition strategy, then how exactlyshould it be implemented?

FIGURE 1. Various types of NoMix toilets are available on themarket today. This is a model by Roediger (photo: YvonneLehnhard).

1856 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 6, 2008

To address the questions above, we originally formulated12 strategic alternatives, which we narrowed down to theeight most attractive or interesting for the present analysis:a status quo (or “business as usual”) alternative, five peak-shaving alternatives for Glattpark that contain variouscombinations of some key options, and two nutrient recyclingalternatives for the full Kloten/Opfikon catchment (Figure2). The details of these alternatives follow:

0. Status Quo. This alternative assumes that Glattparkwill be built and the wastewater will be handled in thetraditional manner at the Kloten/Opfikon wastewater treat-ment plant (WWTP). This implies that the plant will need tobe upgraded in 2011.

1. Peak Shaving. In this alternative, NoMix toilets (suchas those available from the German firm, Roediger, www.roevac.com) are installed in all Glattpark residences. Thesewill each be connected to a 5 L storage tank located in theWC, which is sufficient for two to three days of urine. Deliveryto the WWTP will occur via the existing sewer line usingtimers that release the stored urine at appropriate intervalsto level out the delivery rate.

2.A. Peak Shaving with Limited CSO Control. Alternative1 will not address the significant problem of combined seweroverflow (CSO) during storm events. Therefore, alternative2.A employs a larger 10 L storage tank in each WC that willallow sewer managers to delay urine release for up to fivedays during storms. This will involve the installation of amore sophisticated release system that can be centrallycontrolled via communication lines (e.g., telephone or cable).

2.B. Peak Shaving with Conservative CSO Control. Al-ternative 2.A may still be a problem during extended rainyperiods. Therefore, this alternative expands the storage tankto 30 L/WC and locates it in the cellar of multifamilyapartment buildings. Releases would be centrally controlled.

2.C.i. Peak Shaving with No CSO Risk 1. To completelyeliminate the risk of CSO events, all urine could be directedthrough dedicated lines to an intermediate storage tanklocated in Glattpark. Trucks would be used to pick up storedurine and deliver it to a storage tank at the WWTP to be usedfor peak shaving. This would further increase the operator’scontrol over the process. To allow for truck transport, tankswould need to be sized large enough for a five day storagetime at both Glattpark and the WWTP.

2.C.ii. Peak Shaving with No CSO Risk 2. This alternativeis similar to the previous one but has urine stored onsite intanks in the basements of multifamily buildings rather thanat a central Glattpark location. Smaller and more evenlyspaced truck deliveries of urine would allow a smaller storagetank at the WWTP.

3.A. Nutrient Recycling 1. Of the alternatives describedabove, only the last one (2.C.ii) can be reliably implementedthroughout the entire Kloten/Opfikon catchment for nutrientrecycling. This is because, although it has been suggested (5)that the centralized release mechanism of alternatives 1-2.Bmight be used to transport pure urine through existing sewersat night, this is still highly speculative. Additionally, althoughGlattpark could be constructed to have separate local linesfor urine (alt. 2.C.i), this would not be the case throughoutthe entire catchment. Therefore, alternative 3.A extends thebuilding storage and truck transport of alternative 2.C.ii tothe rest of the catchment. Urine would then be stored in alarge tank at the WWTP and recycled into a fertilizer product.For example, a combination of electrodialysis and ozonationis currently being tested for this purpose in a pilot plantproject (27). The remaining wastewater would continue tobe transported in the existing sewers and treated at theWWTP. Because this wastewater would have a biologicallybalanced C/P/N ratio, the process of treatment would bemuch simpler and only require removal of organics (COD)(5, 13, 14).

3.B. Nutrient Recycling 2. Another option is to recoverthe nutrients in urine onsite. This would involve smalltreatment units designed to process the urine of one or twomultifamily buildings. We assume that the responsibility andcost of operating the urine treatment units would be borneby the existing treatment plant managers or their contractors.They would also continue to handle the wastewater arrivingat the WWTP in existing sewers, as described for alternative3.A, above.

Multicriteria Decision Analysis

To evaluate the alternatives described in the previous section,we performed an MCDA with an emphasis on value-focusedthinking (22). This means involving stakeholders in theprocess of structuring the decision by carefully discussingthe values that matter to them regarding the decision at hand.The process started with the construction of hierarchies toestablish and structure the decision objectives of eachstakeholder (22). Quantitative attributes were then definedfor measuring the achievement of each detailed objective.An assessment was then made of the expected outcome ofeach alternative in terms of the defined attributes. Next, amultiattribute value function was derived for each stakeholderto describe the relative importance of the various attributesand the levels of those attributes (28). Finally, the multiat-tribute value of each alternative was calculated for eachstakeholder to determine how well each of the alternativessatisfies the stakeholders’ objectives, and sensitivity anduncertainty analyses were performed on these results.

Utility functions are often used instead of value functionsin situations involving uncertainty because they incorporatestakeholders’ attitudes toward risk (29). Decision alternativesare then ranked according to calculations of expected utility.However, we decided that the nature of uncertainties inthe NoMix decision was too vague to properly formulaterisk attitudes or outcome probabilities. Therefore, wegenerated ranges of multiattribute values that attempt torepresent the range of possible outcomes of the variousalternatives. These ranges can then be considered infor-mally in the decision-making process and revised as newknowledge becomes available.

FIGURE 2. Schematic of alternatives considered.

VOL. 42, NO. 6, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1857

Objective Hierarchies. An objectives hierarchy wasdeveloped for each stakeholder (Figure 3; Detailed definitionsand methods can be found in the Supporting Information).

Outcome Assessment. Each of the identified alternativeswas evaluated with respect to its predicted ability to meeteach of the stakeholder objectives identified in the previoussection. This was done by assessing the levels of quantitativeattributes characterizing the detailed subobjectives. For theattributes associated with WWTP operation, sanitary equip-ment manufacture, and engineering consultancy, the quan-titative outcome assessment was performed by the profes-sional stakeholder representatives. The outcomes on all otherattributes were assessed by the authors using the combinedresults of the Novaquatis research projects as well as thebackground literature that these projects assembled. Muchuncertainty remained, thus requiring the adoption of variousassumptions and estimations. However, as mentioned in theIntroduction, we believe that our estimations represent thecurrent state of the science (see Supporting Information forassumptions and a table of outcomes.)

Some of the critical issues in our outcome assessmentwere the following: (1) The currently available NoMix toilethas some drawbacks, as revealed in surveys of users. Theseinclude the requirement that men sit when urinating, theawkward sitting position for some women and all children,difficulty with cleaning the two separate bowls, the need forseveral flushes to remove wastes, occasional problems withtoilet paper blockage, the need for regular flushing with citricacid to prevent pipe scaling, and some dislike of the toiletappearance. (2) Some users of NoMix toilets have the sensethat it smells more than conventional toilets. This smell couldcome from the toilet itself or the storage tank. Therefore,

onsite smell was scored according to the presence of a NoMixtoilet and the size and location of the storage tanks. (3) It ishypothesized that NoMix toilets might be further improvedto overcome the drawbacks of the current design. Therefore,we also considered a scenario in which a “high-tech” toiletis developed that eliminates all problems of aesthetics,hygiene, cleaning, convenience, and toilet smell. We assumedthat this toilet would have a cost that is 50% higher and areduced lifetime. Of course, because such a toilet does notyet exist, these assumptions are particularly uncertain. (4)Economically, the largest advantage of the peak-shavingalternatives is in postponing the need to upgrade the Kloten/Opfikon WWTP. It is estimated that an investment of 10million Swiss Francs (CHF) can be delayed for up to sevenyears (from 2011 to 2018) (30). Once built, the new WWTPcan be expected to last for 35 years. Therefore, we chose a48 year time horizon (2005 to 2053) for all economiccalculations and annualized all costs and savings with a 3%interest rate. The upgraded WWTP is expected to have higherrunning costs, which are also delayed by seven years. For thenutrient recycling alternatives, it is assumed that the WWTPwill be downsized in 2023 when the life of the current plantexpires, resulting in significant long-term cost savings. Ofcourse, some of the savings at the WWTP will be offset bythe cost of the equipment and operations of the NoMixsystem, which we assume will be borne by the WWTP budget.The production of fertilizer is assumed to be cost neutral.

Value Function Assessment. The next step was to developthe value function that allows us to rank the alternativesaccording to how well they achieve the multiple objectivesof each stakeholder group. We used an additive functionalform: v(x1,... ,xn) ) w1v(x1) + wnv(xn) , where ν(x1,. . ., xn) is

FIGURE 3. Objective hierarchies for the five stakeholder groups followed by detailed attributes for the Environmental Qualityobjective, as elicited from research scientists. Attribute weights, which were assessed at a later stage of the analysis, are shown inthe right column. When any of the higher level objectives consist of two or more subobjectives, the total weight on thesubobjectives is given in parentheses. Because of the sensitivity of profit estimates, weights were not assessed for the SanitaryEquipment Manufacturers.


the multidimensional value function for a particular stake-holder group, the ν(x1) are linear, one-dimensional valuefunctions that convert each of the n attributes x1 to a commonscale, and the w1 are the tradeoff weights for each attribute(see Supporting Information for a justification of assumptionsand stakeholder elicitation details). Weights for each attributefor each stakeholder are given in Figure 3.

Model Results and Sensitivity AnalysisHouseholds are overwhelmingly expected to prefer the StatusQuo to the implementation of any peak-shaving or nutrientrecycling alternatives (Figure 4a). This is because the instal-lation of a NoMix toilet introduces personal bathroom qualityconcerns, health risk potential, and significant costs that arenot outweighed by the environmental benefits.

Even if a “high-tech” toilet could be developed thatovercomes all the problems of aesthetics, hygiene, cleaning,convenience, and smell, the rankings of households wouldnot change (not shown). However, if, in addition, the highercosts to the households were subsidized by the government

or offset by reduced taxes, the Nutrient Recycling 1 alternativewould be very nearly preferable (Figure 4b). In fact, sensitivityanalysis shows that if the weight on the Environmental Qualityattribute were to increase beyond 0.25, this alternative wouldactually be the most preferred (Figure 5).

Because the Environmental Quality attribute has a muchgreater weight to the WWTP manager than to householdsand because of the potential for significant cost savings atthe plant, WWTP managers are likely to be motivated topursue the NoMix alternatives over the Status Quo (Figure6). The Nutrient Recycling 1 alternative ranks the highest,primarily because of its environmental benefit. There arenot large differences in the ability of the various alternativesto meet the Plant Operation objectives, suggesting that thepotential problems involved with the new technology areexpected to be outweighed by the interesting new challengesand benefits to employee motivation.

Unfortunately, if the WWTP must subsidize householdsto gain their acceptance of NoMix, then much of the costsavings at the plant will be negated. As mentioned in theprevious section, these costs, which include the price of ahypothetical “high-tech” toilet, are particularly uncertain.We used a price ranging from that of the existing NoMixtoilet to a price 1.5 times greater and a lifetime that is halfas long. Even with this significant uncertainty, it seems that,because of the large number of households that would needto be subsidized, the Nutrient Recycling alternatives are nolonger desirable (Figure 7). Only some of the more modestpeak shaving alternatives might be preferred to the StatusQuo. However, if the weight on the Environmental Qualityattribute were to be greater than 0.45, then either NutrientRecycling alternative would be preferred to the Status Quo(Figure 8).

The sanitary equipment manufacturers would be mostmotivated by either of the Nutrient Recycling alternativesimplemented using the “high-tech” toilets (Figure 9), eventhough they expect that this would require a five millionfranc research investment to develop. This preference arisesbecause many more toilets can be sold throughout the entirecatchment than for the peak-shaving alternatives, andapparently, they expect the profit margin to be higher for

FIGURE 4. (a) Overall value of each of the alternatives toprivate households and bar graphs depicting the contributionsof each of the four upper-level objectives to total value. (b)Results for households after assuming that a high-tech toiletcan be developed to eliminate problems of aesthetics, hygiene,cleaning, convenience, and smell and that additional costs ofthe NoMix technology are offset by subsidies or reduced taxes.

FIGURE 5. Sensitivity of the households’ ranking of alternativesto the weight on Environmental Quality. For clarity, not allalternatives are shown. Those that are not shown have a trendsimilar to the Peak Shave alternative. Vertical line indicatescurrent weight.

FIGURE 6. Overall value of each of the alternatives to WWTPmanagers. Also shown are bar graphs depicting the contributionsof each of the three upper-level objectives to total value.

FIGURE 7. Overall value of each of the alternatives to WWTPmanagers after incurring the cost of subsidies to households.Because of the high costs involved, the values wererescaled relative to those in Figure 6. Therefore, only therelative rankings and not the actual numbers can becompared between figures. Error bars indicate uncertainty inthe cost of “high-tech” toilets.


high-tech toilets than for the current design. However, itseems that they do not expect this higher profit to outweighthe investment cost if toilets were sold to Glattpark alone.Still, any of the alternatives would be more profitable thanthe Status Quo for this stakeholder group.

Similarly, both the engineering consultants (Figure 10)and the research scientists (not shown) should like to pursueany alternative over the Status Quo, and the Nutrient Recyclingalternatives would be most preferred. This is because theselast two dominate the others on all objectives for bothstakeholders.

Synthesis and Path ConstructionThe present stakeholder situation presents quite a dilemma.WWTP managers can expect to benefit greatly from separatelycollecting and processing urine into fertilizer. These benefitswill accrue to financial, environmental, and operationalobjectives (Figure 6). However, they can only expect publicacceptance of urine separation if they are willing to fullysubsidize the additional costs to households, including thecurrently uncertain costs of a hypothetical “high-tech” toilet(Figure 4b). Unfortunately, these subsidies are likely tooverwhelm the value of the benefits for all alternatives butthe most moderate of peak-shaving options (Figure 7). Evenin these cases, peak-shaving would only be justifiable if thehigh-tech toilet were to not cost any more than the currentlyavailable NoMix model. According to the assessments of theequipment manufacturers, this seems unlikely, as they expect

to obtain a higher profit from a high-tech toilet than fromthe current model to cover the significant expenses requiredto develop it. Further, if urine source separation is imple-mented only in Glattpark, the manufacturers would preferto sell the existing model rather than spend the money todevelop a more advanced design (Figure 9). This occurs, inpart, because the manufacturers did not declare having anynonmonetary objectives for their decision, unlike the WWTPmanagers or engineering consultants, for example.

Overall, this troublesome situation suggests that, giventhe present priorities, separate urine collection for use inpeak-shaving or nutrient recycling does not appear to be atenable long-term goal for the Kloten/Opfikon catchment.However, the emerging risks of micropollutants and otherwater quality impairments hint that the weight stakeholdersplace on the Environmental Quality objective could signifi-cantly change in the future. Sensitivity analyses show thatif heightened concerns lead to a doubling of the relativeimportance of this objective for households and WWTPmanagers, the Nutrient Recycling 1 alternative would becomemost preferred (Figures 5 and 8).

Because of the significant uncertainty surrounding therisk of micropollutants and the need to make a decision forGlattpark now, it might be argued that a precautionaryapproach should be adopted. A review of our results showsthat such a path can be identified. Alternative 2.B., PeakShaving with Conservative CSO Control, is currently the mostacceptable of the peak-shaving alternatives for households(Figure 4b). This is because the location of a large storagetank in the cellar reduces the smell, health risk, and spacerequirements of the NoMix technology while providingeffective protection against CSO events and associatedmicropollutant releases. The lack of additional intermediateor central storage tanks has practical benefits for WWTPoperations, which make it desirable for plant managers, evenwhen the costs of subsidies are included (Figure 7). Thisalternative is also relatively easily converted to either of theNutrient Recycling options without wasted expendituresbecause it channels urine to an onsite storage tank withoutthe need for WC storage, intermediate storage, or separatesewer lines beyond the building. Therefore, alternative 2.B.makes for a desirable transition scenario that leaves openfuture options should increased health risks or resourceshortages force society to place more weight on environ-mental protection relative to cost or convenience objectives.

To actually forge this path will require some carefulcommunication and negotiation among the stakeholdergroups. The equipment manufacturers will first need to beurged to pursue development of a high-tech toilet that meetsconsumers’ needs and then sell it at a reasonable price to thedevelopers of Glattpark. This will not necessarily be the mostdesirable alternative for the manufacturers, but should stillbe preferable to the Status Quo (Figure 9). It will also bemore desirable if it comes with an agreement that, ifdetermined to be appropriate in the future, a larger contractfor the entire Kloten/Opfikon catchment will be forthcoming.

Households in Glattpark will only accept the high-techurine separating toilet if they are reimbursed for theirexpenses. This will likely be an amount between 50 and 120CHF per WC per year, depending on the selling price andanticipated lifetime of the toilet. We suggest that this amountactually be paid to the residents, and not to the developer,to help compensate them for any minor inconveniences. Itwould also be useful to remind residents that, while theenvironmental benefits may not be great now, the technologyprovides insurance against currently unforeseen risks orresource shortages.

The WWTP managers will likely be just barely breakingeven with this alternative, both in terms of expenses andeffort. However, we anticipate that there will be additional

FIGURE 8. Sensitivity of the WWTP manager’s ranking of alter-natives to the weight on Environmental Quality. For clarity, not allalternatives are shown. Those that are not shown have a trendsimilar to the Peak Shave alternative. Vertical line indicatescurrent weight.

FIGURE 9. Overall value of each of the alternatives to sanitaryequipment manufacturers.

FIGURE 10. Overall value of each of the alternatives to engi-neering consultants. Also shown are bar graphs depicting thecontributions of each of the three upper-level objectives tototal value.


benefits not accounted for in our analysis related to thegaining of experience and momentum. Having a high-techtoilet available as a result of the Glattpark experience shouldfacilitate future technology transfer to other catchments.

While the selection of alternative 2.B. may not be idealfrom the perspective of engineering consultants and researchscientists, it will have significant benefits relative to the StatusQuo and comes with little risk for these groups. Therefore,it would be worth their while to invest some time and energyinto activities that will facilitate the NoMix decision andimplementation process. Appropriate tasks might be to (1)conduct tests and surveys related to the newly developedhigh-tech toilet to ensure that it meets user requirements,(2) aid in the development and testing of signaling devicesto coordinate the release of urine from the onsite storagetanks, (3) perform detailed cost studies to determine theappropriate amount and timing of household subsidies, (4)investigate the human and ecological risks associated withCSOs to determine if further control measures are necessary,and (5) work on developing a urine-based fertilizer productthat is cheap, safe, and convenient to use. These activitieswill both encourage implementation and provide interestingresearch opportunities.

DiscussionWhile our analysis focuses on a particular development ina particular catchment in Switzerland, we believe that theresults can be applied more broadly. Many communities inWestern Europe and the United States are faced with anaging wastewater treatment infrastructure, while also ex-periencing growth and development. This situation makeslong-term planning and the definition of priorities muchmore demanding than in the past (16). Despite someremaining uncertainties in our analysis (see SupportingInformation), we believe the application of MCDA helps toclarify the role that peak-shaving can play in allowing thesecommunities to postpone or avoid expensive WWTP expan-sions. It can also give them a head start in testing urineseparation technology and transitioning to a more compre-hensive, decentralized, and sustainable wastewater treatmentsystem (31).

MCDA also gives insight into the possible barriers totechnology adoption. In this case, it seems that manufacturersdo not want to make the investment necessary to solveremaining consumer problems with NoMix toilets becausethey are waiting for action by the wastewater treatmentcommunity to generate a market. Meanwhile, although ouranalysis suggests that based on current priorities and costaccounting WWTP managers should promote the NoMixtechnology, the problem comes when they must considerpaying for the household installations. Then the NoMixalternatives are only preferable over the Status Quo at thelower end of the possible range in installation costs. Thismeans that if the WWTP managers are at all risk averse, thenthey should logically maintain the existing technology ratherthan pursue NoMix. This seems to be the sticking point ofthe whole situation.

This source of technological “lock-in”, excess inertiacaused by an unbalanced distribution of benefits or a short-term period of lower payoffs for some stakeholders, has beenrecognized as a common barrier to innovative technologies(32). However, as we learned in a previous analysis (33) andapplied to the present case, using MCDA to analyze thepreferences of each stakeholder group separately rather thancombining results across all groups can be especiallyinstructive. In particular, it can suggest a path for imple-mentation that suffices for all groups, even if it is notnecessarily optimal for all. We believe that the precautionarypath we laid out above will do this given adequate coordina-tion among stakeholders.

Even when the specific results of our analysis are notdirectly applicable to a particular community, whetherfor financial, cultural, or practical reasons, the decisionanalytic approach to quantitatively balancing conflictingobjectives among a diversity of stakeholders should stillbe appropriate. In fact, the objectives hierarchy shown inFigure 4 may be comprehensive enough to apply to mostsituations. Only the weights and outcome assessments willthen need to be revised.

AcknowledgmentsWe thank the stakeholder representatives and the scientistsof the Novaquatis project for contributing their time andexpertise to our analysis. We also thank Peter Reichert andtwo anonymous reviewers for their constructive commentson an earlier version of this manuscript. Funding wasprovided by Eawag, the Swiss Federal Institute of AquaticScience and Technology.

Supporting Information AvailableDetails on the creation of objectives hierarchies; consider-ations and assumptions used in assessing outcomes; tableof assessed outcomes (Table S-1); method and assumptionsused in value function construction; and discussion ofremaining uncertainties. This material is available free ofcharge via the Internet at http://pubs.acs.org.

Literature Cited(1) Grübler, A.; Nakicenovic, N.; Nordhaus, W. D. Technological

Change and the Environment. RFF Press: Washington, DC, 2002.(2) Walker, W. Entrapment in large technology systems: institutional

commitment and power relations. Res. Policy 2000, 29, 833–846.

(3) Barnes, W.; Gartland, M.; Stack, M. Old habits die hard: pathdependency and behavioral lock-in. J. Econ. Issues 2004, 38 (2),371–377.

(4) Daughton, C. G.; Jones-Lepp, T. L. Pharmaceuticals and PersonalCare Products in the Environment: Scientific and RegulatoryIssues; American Chemical Society: Washington, DC, 2001.

(5) Larsen, T. A.; Gujer, W. Separate management of anthropogenicnutrient solutions (human urine). Water Sci. Technol. 1996, 34(3–4), 87–94.

(6) Larsen, T. A.; Gujer, W. The concept of sustainable urban watermanagement. Water Sci. Technol. 1997, 35 (9), 3–10.

(7) Otterpohl, R. Options for alternative types of sewerage andtreatment systems directed to improvement of the overallperformance. Water Sci. Technol. 2002, 45 (3), 149–158.

(8) Henze, M. Waste design for households with respect to water,organics, and nutrients. Water Sci. Technol. 1997, 35 (9), 113–120.

(9) Larsen, T. A.; Gujer, W. Waste design and source control leadto flexibility in wastewater management. Water Sci. Technol.2001, 43 (5), 309–317.

(10) Lienert, J.; Gudel, K.; Escher, B. I. Screening method forecotoxicological hazard assessment of 42 pharmaceuticalsconsidering human metabolism and excretory routes. Environ.Sci. Technol. 2007, 41 (12), 4471–4478.

(11) Lienert, J.; Bürki, T.; Escher, B. I. Reducing micropollutants withsource control: Substance flow analysis of 212 pharmaceuticalsin feces and urine. Water Sci. Technol. 2007, 56, 5.

(12) Lundin, M.; Bengtsson, M.; Molander, S. Life cycle assessmentof wastewater systems: influence of system boundaries and scaleon calculated environmental loads. Environ. Sci. Technol. 2000,34, 180–196.

(13) Wilsenach, J. A.; Loosdrecht, M. C. M. V. Effects of separateurine collection on advanced nutrient removal processes.Environ. Sci. Technol. 2004, 38, 1208–1215.

(14) Wilsenach, J. A.; Loosdrecht, M. C. M. V. Integration of processesto treat wastewater and source-separated urine. J. Environ. Eng.2006, 132 (3), 331–341.

(15) Berndtsson, J. C. Experiences from the implementation of aurine separation system: Goals, planning, reality. Build. Environ.2006, 41, 427–437.

(16) Lienert, J.; Larsen, T. A. Considering user attitude in earlydevelopment of environmentally friendly technology: A casestudy of NoMix toilets. Environ. Sci. Technol. 2006, 40 (16),4838–4844.


(17) Pahl-Wostl, C.; Schonborn, A.; Willi, N.; Muncke, J.; Larsen, T. A.Investigating consumer attitudes towards the new technologyof urine separation. Water Sci. Technol. 2003, 48 (1), 57–65.

(18) Larsen, T. A.; Lienert, J., Societal implications of re-engineeringthe toilet. In Water Intelligence Online, UNIQUE ID: 200303006,www.iwaponline.com/wio/2003/03/default001.htm, 2003.

(19) Kompetenzzentrum Wasser Berlin Annual Report; Kompetenz-zentrum Wasser Berlin: Berlin, Germany, 2006; report availableat http://www.kompetenz-wasser.de/.

(20) Rauch, W.; Brockmann, D.; Peters, I.; Larsen, T. A.; Gujer, W.Combining urine separation with waste design: An analysis usinga stochastic model for urine production. Water Res. 2003, 37(3), 681–689.

(21) Araujo, L.; Harrison, D. Path dependence, agency, and tech-nological evolution. Technol. Anal. Strategic Manage. 2002, 14,5–19.

(22) Keeney, R. L. Value-Focused Thinking; Harvard University Press:Cambridge, MA, 1992.

(23) Salo, A.; Gustafsson, T.; Ramanathan, R. Multicriteria methodsfor technology foresight. J. Forecast. 2003, 22, 235–255.

(24) Porter, A. L.; Ashton, W. B.; Clar, G.; Coats, J. F.; Cuhls, K.;Cunningham, S. W.; Ducatel, K.; Duin, P. v. d.; et al. Technologyfutures analysis: Toward integration of the field and newmethods. Technol. Forecast. Social Change 2004, 71, 287–303.

(25) Keeney, R. L.; McDaniels, T. L. A framework to guide thinkingand analysis regarding climate change policies. Risk Anal.2001, 21 (6), 989–1000.

(26) Lienert, J.; Haller, M.; Berner, A.; Stauffacher, M.; Larsen, T. A.How farmers in Switzerland perceive fertilizers from recycledanthropogenic nutrients (urine). Water Sci. Technol. 2003, 48(1), 47–56.

(27) Pronk, W.; Zuleeg, S.; Lienert, J.; Escher, B.; Koller, M.; Berner,A.; Koch, G.; Boller, M. Pilot experiments with electrodialysisand ozonation for the production of a fertilizer from urine. WaterSci. Technol. 2007, 56, 5.

(28) Keeney, R. L.; Raiffa, H. Decisions with Multiple Objectives; Wiley:New York, NY, 1976.

(29) von Neumann, J.; Morgenstern, O. Theory of Games andEconomic Behavior. Princeton University Press: Princeton, NJ,1944.

(30) Ingenieurbüro Gujer Urinseparierung im Glattpark. Auswirkun-gen auf die ARA Kloten/Opfikon (in German); Available at:www.gujeragch: Rümlang, Switzerland,2005.

(31) Larsen, T. A.; Peters, I.; Alder, A.; Eggen, R.; Maurer, M.; Muncke,J. Re-engineering the toilet for sustainable wastewater manage-ment. Environ. Sci. Technol. 2001, 35 (9), 192A–197A.

(32) Cowan, R.; Gunby, P. Sprayed to death: Path dependence, lock-in, and pest control strategies. Econ. J. 1996, 106, 521–542.

(33) Hostmann, M.; Truffer, B.; Reichert, P.; Borsuk, M. E. Stakeholdervalues in decision support for river rehabilitation. Arch. Hy-drobiol. 2005, 155, 491–506 (Large Rivers Supplement 15).

ES702184P


Charting a Path for Innovative Toilet Technology Using Multicriteria Decision Analysis

Documents