Journal of Environmental Management 151 (2015) 404e415

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Adaptation in hindsight: Dynamics and drivers shaping urban wastewater systems € rg Rieckermann c, Thomas Hug d, Willi Gujer e Marc B. Neumann a, b, *, Jo Basque Centre for Climate Change, Alameda Urquijo, 4 - 4
, 48008 Bilbao, Spain IKERBASQUE, Basque Foundation for Science, Bilbao, Spain c Eawag, Swiss Federal Institute of Aquatic Science and Technology, 8600 Dübendorf, Switzerland d Hunziker Betatech AG, 8411 Winterthur, Switzerland e Emeritus, ETH Zurich, Zurich, Switzerland a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 August 2014 Received in revised form 22 December 2014 Accepted 27 December 2014 Available online 13 January 2015

Well-planned urban infrastructure should meet critical loads during its design lifetime. In order to proceed with design, engineers are forced to make numerous assumptions with very little supporting information about the development of various drivers. For the wastewater sector, these drivers include the future amount and composition of the generated wastewater, effluent requirements, technologies, prices of inputs such as energy or chemicals, and the value of outputs produced such as nutrients for fertilizer use. When planning wastewater systems, there is a lack of methods to address discrepancies between the timescales at which fundamental changes in these drivers can occur, and the long physical life expectancy of infrastructure (on the order of 25e80 years). To explore these discrepancies, we take a hindsight perspective of the long-term development of wastewater infrastructure and assess the stability of assumptions made during previous designs. Repeatedly we find that the drivers influencing wastewater loads, environmental requirements or technological innovation can change at smaller timescales than the infrastructure design lifetime, often in less than a decade. Our analysis shows that i) built infrastructure is continuously confronted with challenges it was not conceived for, ii) significant adaptation occurs during a structure's lifetime, and iii) “muddling-through” is the pre-dominant strategy for adaptive management. As a consequence, we argue, there is a need to explore robust design strategies which require the systematic use of scenario planning methods and instruments to increase operational, structural, managerial, institutional and financial flexibility. Hindsight studies, such as this one, may inform the development of robust design strategies and assist in the transition to more explicit forms of adaptive management for urban infrastructures. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Urban infrastructure Wastewater Planning Adaptive management Uncertainty Complexity Hindsight

1. Introduction Urban infrastructure is normally developed on a project-byproject basis. It is not typical for agents to review and compare the long-term historical development of their systems and to confront assumptions and decisions from past designs with actual outcomes. Few scientific studies systematically review the long term development of urban infrastructure systems. For example, Flyvbjerg et al. (2003) uncover systematic bias between projections and actual outcomes in the transport sector. Methods exist to

* Corresponding author. Basque Centre for Climate Change, Alameda Urquijo, 4 e 4, 48008 Bilbao, Spain. E-mail address: [email protected] (M.B. Neumann). http://dx.doi.org/10.1016/j.jenvman.2014.12.047 0301-4797/© 2015 Elsevier Ltd. All rights reserved.

support the long-term planning of infrastructure, including scenario planning (Schoemaker, 1995) or more quantitative approaches such as real options (De Neufville, 2003). To better inform these approaches, we believe there is value in improving the understanding on how infrastructure systems evolve. Identifying the drivers’ temporal characteristics such as smoothness, stochasticity, trend reversals, transitions, and abrupt regime shifts can critically test the assumptions underlying current planning methods. In our study we try to uncover some of these temporal characteristics for forces driving the development of wastewater infrastructure. Socio-economic drivers of urban wastewater systems include among others population growth, urbanisation, and industrialisation. Many current design projects use forecasts based on trend analysis to predict change in these drivers (Dominguez and Gujer,

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

2006). However, political and socio-economic changes are not necessarily smooth transitions and can exhibit both abrupt shifts and trend reversals. Historically, trend extrapolation has mainly taken place within growth scenarios. A pitfall of extrapolation is that it may miss key trends. Extrapolating just prior to turning points can have highly undesired consequences. This was the case for many developed countries, where rising water use and thus wastewater generated in the 1960's and 1970's was followed by sharp declines. In a detailed example Moss (2008) showed how infrastructure expansion taking place in Eastern Germany after unification was succeeded by rapidly declining wastewater loads due to de-industrialisation, population shrinkage and a fall in per capita water consumption. He demonstrated how the resulting overcapacities did not only lead to unnecessarily high costs but also to poor performance due to sediment build-up in oversized sewers. For natural system drivers (rainfall patterns, river dynamics) the design of wastewater systems has in most cases been based on stationarity, the assumption, that the statistical characteristics of historical data will be preserved in the future. Recent studies emphasise that due to climate change, the assumption of stationary rainfall patterns may no longer be valid as a basis for designing long-lived water infrastructure (Milly et al., 2008). Furthermore, changes within the catchment of the receiving water (e.g. land use change or river engineering measures) can change statistical characteristics of background loads and flows and impact the urban infrastructure's ability to achieve the environmental requirements. For the water sector Geldof (1995) described the long-term development of water management using the metaphor of multiple interlinked controllers characterising the physical and social domains and forming a complex adaptive system (Holland, 1992) with the following characteristics:    

It can be characterised as a network of agents acting in parallel. Multiple levels of organisation are present. Agents have the ability to anticipate future developments. The agents are perpetually confronted with novel issues.

In our study we seek to examine these features, specifically the role of the involved principal agents and their corresponding paradigms, their anticipation of future developments and their resulting actions. In the wastewater sector Geels (2006) examined the role of the principal agents in institutional regime dynamics during the hygiene revolution in the Netherlands (1840e1930) and showed how system transformation is not only due to technological substitution but is largely explained by the competition between these agents. For example, he shows how during the mid-19th century, the Dutch city councils and engineers (he uses the term “regime insiders”) initially resisted the pressures of hygienist doctors (“regime outsiders”) and how these overcame the resistance to become the new insiders. In our case study we intend to uncover how the principal agents are acting within these regimes and how the broader developments (urbanisation, industrialisation, and democratisation) alter the stage and provide resources for agents. In this, we are especially interested in the roles of the municipal engineer and the physician, the two central agents in the development of wastewater infrastructure. Whereas the former was the principal agent in enabling and shaping the growth of cities in the nineteenth century (Schultz and McShane (1978)), the latter shaped the hygiene revolution. Dominguez and Gujer (2006) analysed the development of load and capacity of the wastewater treatment of Zurich (Switzerland) during a 14 year time period. We extend their relatively short observation window to 170 years for the same system. Where they focused on how unexpected changes in wastewater loads interacted with upgrading activities and operational adjustments at the

405

wastewater treatment facility, we are able to capture many more phenomena. Specifically, we assess the stability of critical assumptions made by engineers in planning and design of urban wastewater systems by exploring how various drivers influence and interact upon the newly commissioned infrastructure. These drivers relate to sources of uncertainty that are dealt with in the planning stages of engineering projects and have been recognised as being more influential than sources of uncertainty that appear in more detailed design stages (Neumann (2007), Dominguez (2008), Belia et al. (2009), Neumann and Vanrolleghem (2011)). The main goal of our study is to reveal the discrepancies that arise between fast and unforeseeable dynamics in various drivers and long infrastructure design lifetime. We specifically analyse how long it takes for a new system to become inadequate and the time required to identify and to implement new solutions. To our knowledge this is the first scientific analysis of how principal drivers shape an urban wastewater system over a period of more than a century and therefore includes a very wide range of phenomena. We analyse how the system boundaries evolve and explore the dynamics which take place both within each infrastructure generation and during the transition from one generation to the next. The identified driver typology is expected to be transferable to other cities, although the challenges, objectives and principal agents may vary. The methodology applied in this “adaptation in hindsight” study is transferable to other urban infrastructure sectors where the infrastructure lifetime exceeds the timescales at which the principal drivers evolve. Following a general description of components and drivers of current wastewater systems a detailed analysis is conducted on the evolution of components and drivers for Zurich's wastewater system from 1850 to 2020. The discussion then focuses on the hindsight assessment of system performance and on what we can be learnt for improving current design practice. 2. Materials and methods 2.1. Components and drivers of current wastewater systems We start out by giving a generic characterisation of a current centralised wastewater system explaining the main components and drivers. In this way we introduce the reader unfamiliar to wastewater treatment to the main characteristics and explain our perspective of analysis. The task of today's engineer is to provide infrastructure that is able to transform wastewater loads in a way that environmental requirements are fulfilled (rectangular shapes in Fig. 1). These components are influenced by multiple drivers (oval shapes in Fig. 1). 2.1.1. Drivers of load Wastewater loads are determined by urban catchment characteristics and rainfall patterns. Households, commerce and industry produce wastewater which is collected in sewer networks. Lifestyle- and behavioural changes, city growth or shrinkage, and changes in industrial intensity alter the wastewater patterns. Runoff entering the wastewater systems is also subject to change, e.g. due to climate change. Loads can be influenced through demand side management (e.g. increase of water price) or source control efforts that prevent substances from entering the wastewater system in the first place (e.g. disconnection of runoff from sewers). 2.1.2. Drivers of infrastructure The engineer designs using best practice tools (e.g. use of design

406

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

Fig. 1. Today's wastewater system: Infrastructure is provided to transform wastewater loads in order to meet environmental requirements. Various dynamic drivers (oval shapes) influence these components.

guidelines) and projections of future loads. Typical design horizons are 25e40 years for central wastewater treatment plants and around 80 years for sewer pipes (Dominguez, 2008). The infrastructure of a centralised wastewater facility is determined by the implemented technology (e.g. conventional activated sludge system with fine-bubble aeration) and the dimensioning (e.g. tank sizes). Available technologies change in time due to technological innovation and changing engineering paradigms and financial resources. 2.1.3. Drivers of requirements Requirements are anchored in legislation mirroring the values of society (e.g. environmental awareness). Typical requirements for today's wastewater treatment systems are effluent concentration limits. Activities that impact river dynamics (e.g. building of dams, land use changes) can force the water authorities to alter effluent requirements. 2.2. Sources of information for the case study In the period from 1850 to 1980, the work of historian Martin Illi was used as the main source of information. Illi (1987) conducted a detailed survey of primary documents including minutes from parliamentary debates or reports to the municipality and related newspaper coverage. In our study, we summarise and discuss his survey from an engineering perspective. During the period from 1926 to 1932 Brentano (1934), provided information about the relationship between Zurich's wastewater treatment and river water pollution. For the years 1980e1985, detailed technical information and accounts concerning the upgrading of Zurich's wastewater treatment facility were available (Zürich €sserung, 1986). In the period from Hauptabteilung IV Stadtentwa 1989 to 2003 we used information from Dominguez and Gujer (2006) who described the loads and capacity development during this period. Further information about the developments taking place from the mid-1970s were obtained from personal experience by co-author W. Gujer, who has been professionally involved with the development of Zurich's wastewater treatment facility since 1974. To describe the current developments, we also engaged in direct communication with engineers at Zurich's wastewater treatment facility.

2.3. Structure of the case study We segment the examined time window from 1850 to 2020 into six phases (see Table 1). The titles of the six phases describe the central challenge and the implemented engineering solution. For each phase we give a detailed account of the changes taking place with respect to loads, infrastructure and environmental requirements. For each phase, we construct a causal diagram depicting the main drivers, the times at when they act upon the systems and their interactions. At the end of each section we extract the main findings. 3. Evolution of zurich's wastewater management 3.1. Cholera outbreaks due to population growth and poverty e covering of cesspits and municipal sludge collection (phase 1: until 1860s) During the first half of the 19th century, the City of Zurich experiences the onset of industrialisation coincident with population growth (Fig. 2). Wealthy city dwellers leave the town centre and move to new houses in the outskirts. Their flats are taken over by Table 1 Different phases considered in case study. Phase Years

Challenges

Implemented design

1

Cholera outbreaks due to population growth and poverty City development and typhus outbreaks

Covering of cesspits and municipal sludge collection Municipal sanitation with bucket system including liquid waste sewer Water-borne sewer and central mechanical wastewater treatment plant Biological wastewater treatment Enhanced biological treatment and merging of wastewater catchments Heat recovery, sludge incineration, ozonation

2

Until 1860s 1860s e1880s

3

1880s Hygiene movement and river e1930s water pollution (solids)

4

1930s Water pollution (organics) e1980s 1980s North Sea protection and e2000s wastewater load decrease

5

6

2000s Energy efficiency, resource e2020 recycling, climate change, micropollutant removal

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

407

Fig. 2. Phase 1: Until 1860s: Cholera outbreaks due to population growth and poverty e Covering of cesspits and municipal sludge collection.

craftsmen who divide and sublet their apartments. This leads to an increase in population density. The associated surge in poverty and rapidly deteriorating sanitary conditions finally provoke the first outbreaks of epidemics such as the severe Cholera epidemic of 1854. The two main theories of the time are that Cholera is caused either by contact with poisonous air (Miasma theory) or poisonous soil. Therefore the doctors are mainly concerned with reducing the smell coming from open cesspits and ditches (the sanitation system has not significantly changed since medieval times). In 1855, a private society of doctors, builders and engineers approach the city council with a pilot-project to cover the cesspits and to organise a communal collection of human- and solid waste. The pilot-project is deemed unsuccessful by contemporaries as the pits are too small and the emptying procedure itself causes substantial smell. In 1860, the city assumes responsibility for emptying the open cesspits and in 1863 a private company is licensed to undertake this task. This company uses pumps to empty the cesspits and burn off the biogas. The gathered sludge is stored in a reservoir and collected by farmers who use it as fertilizer. Phase 1 e main findings  The coinciding increase in population density and poverty lead to a deterioration of sanitary conditions.  The perceived causes of the cholera epidemics (Miasma theory) are first tackled in private initiatives and then quickly taken over by the municipality.  The municipality licenses a private company to manage the urban cesspits.  As human waste is regarded as highly valuable the sludge from the cesspits is re-used in agriculture.

3.2. City development and typhus outbreaks e municipal sanitation with bucket system including liquid waste sewer (phase 2: 1860se1880s) There are political and economic motives for Zurich to become a modern European city with boulevards (Fig. 3). To facilitate this

transition the citizens of the Canton of Zurich vote for the introduction of cantonal (i.e. county-wide) building regulations in 1863. The new laws require a complete rehabilitation of the existing sanitation system and the construction of sewers for new urban developments. The city council gives the municipal engineer A. Bürkli the mandate to identify appropriate solutions. In 1864 A. Bürkli conducts an explorative tour through France, England, Belgium and Germany to study state-of-the-art water supply- and wastewater systems. However, his report to the city council initially does not receive high priority. However, just a year later (1865/66), Zurich experiences a severe Typhus epidemic which accelerates municipal provisions for the rehabilitation of the sanitation system. A. Bürkli recommends installing the “bucket system” consisting of privy buckets with perforated walls, so-called “fosses mobiles  a  l’e gout”. In 1864 Bürkli has witnessed the operdiviseur filtrant a ation of such a system in Paris. It allows retaining the solids for fertilizer use. Bürkli's recommendations provoke a debate on the choice of appropriate technology: should the city install the proposed bucket system or a water-borne sewer which flushes both the solid and liquid waste to the river? Bürkli's opponents emphasise the smaller capital investments of a water borne-sewer (no need for buckets) and the lower operating costs. Bürkli argues against the waterborne sewer as it would result in the loss of valuable fertilizer (due to the flushing of all the waste to the river) and increase river pollution. He asserts that, if the water-borne sewers were introduced, water closets would be mandatory in households in order to properly flush the solids through the conduits. The city council decides to follow A. Bürkli's suggestion to install the bucket system. A likely reason why citizens prefer the bucket system over the water-borne sewer is that they can control the valuable waste. The bucket system provides each household with the option to empty the buckets for its own use and does not require the instalment of water closets. In 1867, following A. Bürkli's recommendation, the citizens of Zurich pass a municipal bill, the so-called “Kloakenreform”, authorising the rehabilitation of the sanitation system. A cholera epidemic in the same year accelerates the implementation of the proposed measures. The installation of the bucket system

408

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

Fig. 3. Phase 2: 1860se1880s City development and typhus outbreaks e Municipal sanitation with bucket system including liquid waste sewer.

progresses very quickly. By 1871 two thirds of the inner city buildings are connected to the bucket system (1108 buckets). A liquid waste sewer is installed and between 1867 and 1873, the length of the public sewer system increases from 10 to 80 km. Four main trunk sewers discharge into the river Limmat downstream from the city. The city council organises a centralised waste collection for the decentralised bucket system. In order to maintain the liquid waste sewers and to collect the waste from the buckets 20 workers are employed. In 1870 the revenues from selling the sludge to farmers (13.26 CHF per m3 of sludge) just about cover the operational costs (14 CHF per m3 of sludge). However, the revenues steadily decrease to 12 CHF/m3 in 1875 and to 1 CHF/m3 in 1899 due to:  An increasing number of water closets, which leads to the washing out of nutrients from the buckets.  The availability of commercial chemical fertilizer and reduced cost of transport due to railroads.  The increase in animal products due to dietary changes and the drop in wheat prices (1870) leads to a shift from crop farming to cattle farming. This increases the availability of manure as an additional nutrient source. To reduce the deficit of the bucket system, A. Bürkli and the city council suggest that the increasingly nutrient-rich wastewater in the liquid waste sewers should be used to irrigate fields to grow cattle feed. However, the renowned physician K. Reiser publically criticises this irrigation project which also faces land expropriation issues. The project is finally abandoned by the citizens in a vote in 1876. Additionally, in 1882 the city parliament ceases to subsidise the growing deficit of the bucket system and introduces fees and taxes for the waste collection.

Phase 2 e main findings:  The political and economic will of becoming a modern city with boulevards is a main driver for the renewal of the wastewater system.  Typhus outbreaks lead to a quick decision to follow other European cities to build a municipal sanitation system.  The decentralised bucket system with connections to a new sewer for liquid waste disposal is the initial choice of technology. It provides recycling of the solid waste, which at the time is a highly valued fertilizer. The system is introduced rapidly and is operated by the municipality.  Human waste moves from being a valuable resource which covers the operational costs of the bucket system to a worthless waste within a short time. This destroys the main reason for the initial choice of technology.

3.3. Hygiene movement and river water pollution (solids) e waterborne sewer and central mechanical wastewater treatment (phase 3: 1880se1930s) At the end of the 19th century the hygiene revolution comes to be the central driver for the advancement of wastewater treatment (Fig. 4). Physicians and public health experts become the new influential agents and opinion leaders. In 1883 the “Institute for Hygiene” at the University of Zurich is established and in 1888 the “Hygiene Association” of Zurich is formed. Dr. F. Erismann, a professor of hygiene, is elected as city councillor in 1901 and institutes the hygiene movement in Zurich and promotes the introduction of public toilets and bathing houses. The percentage of households with bathrooms increases rapidly

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

409

Fig. 4. Phase 3: 1880se1930s: Hygiene movement and river water pollution (solids) e Water-borne sewer and central mechanical wastewater treatment facility.

from 10% in 1896 to 50% in 1920. Simultaneously, enterprises and citizens in villages downstream of Zurich City begin to complain about river water pollution. As a consequence, in 1881, the Canton of Zurich introduces the first water pollution bill which requires regulation of the release of large amounts of solid waste, untreated factory waste water, and substances which endanger humans and animals. This county-wide legislation is an early example of reducing water pollution through source control. The hygiene movement marks the advent of the bacteriological sciences, which are able to provide objective measures of river water pollution through quantifying pathogen concentration. Bacteriological studies show a deterioration of the river water quality between 1890 and 1900. The intensified industrialisation and the rapid connection of new areas to the sewer system leads to a surge of the dry weather flow from 200 l/s in 1889 to about 1000 l/ s in 1916. Also, the introduction of water closets tends to flush out the contents from the buckets subsequently increasing the pollution load to the river. As it has been predicted earlier, the water closets turn out to not be compatible with the bucket system. In 1909 the city council selects a commission to prepare the replacement of the bucket system, which by this time is operating at a significant financial loss. Municipal engineer V. Wenner visits England and Germany to explore state of the art sewer systems and treatment facilities. Wenner's report from 1916 advises the city council to replace the bucket system with a water-borne sewer and a central wastewater treatment facility. The existing liquid waste sewers that were built for the bucket system are found to be large enough to be used for this purpose. It is supposed that this is because A. Bürkli had either:  Already envisaged the possible future transition to a waterborne sewer 40 years earlier and provided for this option.  Or had chosen a robust design which provided extra capacity.

Initially, however, the beginning of the 1st World War puts all wastewater infrastructure rehabilitation plans on hold. In 1923 Zurich's citizens overwhelmingly approve a 4 Million CHF bond to implement the water-borne sewer and central wastewater treatment facility. Technical standards for house connections are introduced and new constructions require mandatory approval by the authorities. A ten year transition period is introduced to reduce potential opposition by citizens still preferring to re-use their own waste from the currently installed bucket system. Although biological wastewater treatment is already well established in Europe, only a mechanical treatment step is deemed necessary by the engineers due to the dilution ratio and the high assimilative capacity of the river. In 1926 the central wastewater treatment plant starts up. It consists of a grit chamber, a screen and a sedimentation tank and is able to remove 80% of solids. The sludge is digested in a reactor and the supernatant, the liquid residual from digestion, is passed on to a biological trickling filter. The digested sludge is collected by farmers or is dried on “drying fields” for subsequent use as backfill in construction. Five employees operate the treatment plant. However, already in 1927, just one year after start-up, the dry weather flow to the treatment facility reaches two times the predicted design flow. Reasons for this are:  Connections of private water supplies had not been adequately accounted for.  More groundwater infiltrates into the sewers than expected.  More neighbouring villages than originally expected connect to the sewer.  Citizens abandon the bucket system at a higher rate than expected. In 1930, to rehabilitate the overloaded facility, the citizens approve a 1.5 Million CHF bond for the necessary improvements to

410

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

the facility which includes a parallel replication of the sedimentation tank. Pre-sedimentation is also added to remove heavier particles facilitating a smoother operation of the pumps. In addition, mechanical components are introduced to automatize operation. Only two years later, in 1932, the duplicated and improved facility begins operation. Phase 3 e main findings:  The principal agents shift from municipal (civil) engineers such as A. Bürkli to physicians and hygienists such as K. Reiser, F. Erismann.  The scientific development of bacteriology, the hygiene movement and local river water pollution are the dominant driving forces for further development of wastewater infrastructure.  Early source control measures are adopted that efficiently reduce water pollution caused by solids.  The municipality decides to replace the financially strapped bucket system with a water-borne sewer and a centralised wastewater treatment facility. An expansion of the new facility is required immediately after commissioning due to the unforeseen rapid increase of wastewater loads.

3.4. Water pollution (organics) e biological wastewater treatment (phase 4: 1930se1980s) In 1933 a river hydropower station is constructed in Wettingen, a town in the neighbouring Canton of Zurich, just 20 km downstream from the newly commissioned (1932) wastewater treatment facility (Fig. 5). This significantly lowers flow velocity which causes an immediate and significant reduction of assimilative capacity of the river. The sedimentation of solids and the fouling of organic substances intensify water pollution and lead to aestheticand odour problems. The newly installed treatment technology (mechanical only) which relies on the assimilative capacity of the river becomes obsolete only one year after commissioning. To reduce the severe river pollution the municipality recognises the need for biological wastewater treatment. To gain experience with

this technology, a pilot scale trickling filter is installed. In 1936 the Swiss Federal Institute of Technology (ETH Zurich) forms a thinktank to collect information and conduct experiments in view of upgrading the facility to biological treatment. The think-tank is made up of a chemist, a biologist and an engineer. However, the beginning of the 2nd World War halts the initiative of introducing biological treatment. The effects of the war years are clearly visible in the functioning of the treatment facility: By 1940 the wastewater input from households is substantially reduced due to the lower consumption and the recycling of waste (e.g. kitchen waste is fed to pigs). Fat is collected from grease separators and used by industry. Biogas from the digester is used as gas- and fuel substitute. By the 1950's, due to post-war economic growth, the mechanical treatment facility is heavily overloaded and only removes about 30% of the total solids. Colloids and dissolved pollutants are not removed. The ministry of the canton of Zurich approves a project in 1952 to introduce biological treatment. Again, experiments are conducted by the former ETH think-tank which has now matured to the independent Swiss Federal Institute of Aquatic Science and Technology (Eawag). In 1958 voters approve the issuance of a 31 Million CHF bond to finally upgrade to biological treatment. Debates are held whether to install a trickling filter or a suspended activated sludge unit. In the end a partial activated sludge system consisting of an aerated tank and a secondary clarifier is installed to gain much needed experience before completion. In 1969 the partial activated sludge system commences operation. The system is overloaded from the start. Simultaneously other factors are rapidly increasing the amount of wastewater:    

Increase of potable water demand. Introduction of washing machines. Further growth of industry. New phosphorous-containing washing powders. The city council decides in 1972 to completely upgrade the

Fig. 5. Phase 4: 1930se1980s: Water pollution due to organics e Biological wastewater treatment.

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

poorly functioning and overloaded treatment facility. An international request for proposals is launched in 1973 to identify the optimal treatment technology to meet the very stringent effluent requirements (5 mg/l TSS (total suspended solids); 5 mg/l DOC (dissolved organic carbon); 10 mg/l BOD5 (biological oxygen demand); 1 mg/l TP (total phosphorous). Six international engineering consortia compete. Only the consortia proposing chlorination with activated carbon filtration are able to reach the very stringent DOC standard with the downside of being twice as expensive as the other bids. None of the proposals are selected. Follow-up projects are mandated by the city council. In 1974/75 two research engineers from Eawag (the Swiss Federal Institute of Aquatic Sciences and Technology), W. Gujer and M. Boller, conduct studies on ammonium dynamics and nitrification, especially investigating how the conversion of ammonium to nitrate depends on temperature and is inhibited by digester supernatant. In addition they investigate sludge production and filtration and propose a layout which re-uses the existing facility as a pretreatment step in the new facility. Chemical phosphorous removal is introduced in 1975 within the mechanical step of the treatment plant. Citizens overwhelmingly approve funding for a new treatment facility in 1978 and construction begins in 1981. The facility goes online in late 1985 and performs mechanical treatment, biological treatment with nitrification, chemical treatment to eliminate phosphorous, filtration to reduce suspended solids and sludge digestion. Phase 4 e main findings:  The installation of a hydropower facility in 1933 dramatically reduces the river's assimilative capacity which leads to significant water pollution problems rendering the newly installed technology (mechanical treatment only) obsolete after just one year of operation.  The 2nd World War halts initiatives to introduce biological treatment and dramatically influences wastewater amount, and composition as and re-use activities.

411

 Trials to introduce biological treatment in the 1950s and 60s are unsatisfactory due to rapid increase in wastewater loads and uncertainties of the appropriate technology. It takes over 50 years from the desire to install biological treatment until a satisfactory solution is found in 1986.

3.5. North Sea protection and wastewater load decrease e enhanced biological treatment and merging of wastewater catchments (phase 5: 1980se2000s) Zurich is subject to several trend reversals in the mid-1970s (Fig. 6). These include de-industrialisation and population decrease as people move to the suburbs due to increased mobility. These two trends reduce water consumption and therefore also the amount of wastewater entering the treatment facility. Another trend which reduces the amount of wastewater entering the facility, is the rehabilitation of sewers between 1985 and 2003 reducing groundwater intrusion by one third. The amount of urban runoff reaching the combined sewer is reduced through the construction of separate sewer systems and decentralised infiltration sites. All of these trends reduce the flow that reaches the wastewater facility. Changes in environmental legislation occur at the same time impacting the water load entering the facility. The national ban of phosphates for detergents (source control) in 1986 abruptly reduces the phosphorous load to the treatment facility and frees up capacity for other purposes. Another environmental law is introduced to limit nitrates in the Rhine river basin, its coastal zones and the North Sea. This law requires the addition of a de-nitrification step to reduce the nitrate load leaving the facility. Due to the freeing up of capacity reserves mentioned above, the required anoxic zones for this denitrification step can be successively integrated into the existing structure. This treatment step requires 28% of the existing reactor volume and in turn reduces the capacity of the plant to remove COD. However, due to the sewer rehabilitation measures the

Fig. 6. Phase 5: 1980se2000s: North Sea protection and wastewater load decrease e Enhanced biological treatment and merging of wastewater catchments.

412

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

operational maximum dry-weather flow allowed to enter the plant can be lowered from 9 to 6 m3/s. As a consequence the treatment facility can be operated at the higher sludge concentration of 4.5 instead of 3 kg TSS/m3 which in turn increases COD removal capacity. Continued industry shrinkage in the mid 1990's substantially reduces the COD load to the treatment plant. The capacity reserves become large enough to connect an entire neighbouring district to Zurich's treatment facility, corresponding to a 20% wastewater load increase. Additionally, from 2002 onwards, the facility also treats the highly concentrated de-icing wastewater from Zurich airport. Phase 5 e main findings:  De-industrialisation, population shrinkage, water saving measures and sewer rehabilitation measures significantly reduce the amount of wastewater entering the facility.  This reduction of wastewater allows to integrate de-nitrification, to connect an entire district (20% load increase), and to treat the de-icing wastewater from the airport without expanding the facility.

3.6. Energy efficiency, resource recycling, climate change, micropollutant removal e heat recovery, sludge incineration, ozonation (phase 6: 2000s to 2020) A system is installed to recover heat from the effluent for use in the facility and nearby residences in 2007 (Fig. 7). Around the same time pilot tests to recuperate heat directly from the sewers are conducted. However, these tests are not further pursued as this technology can lower the temperature of the wastewater and hence reduce nitrification capacity of the facility. The treatment configuration is upgraded to alternating-

intermittent (A/I) mode in 2013. 12 tanks with separate feed significantly increase process flexibility. Additionally, the A/I mode requires less energy for aeration and this in turn improves nitrogen removal. To recover phosphorus, it is decided to build a central sludge incineration plant for the entire Canton of Zurich at the current wastewater treatment facility site. It is expected that all wastewater treatment plants of the Canton will be required to deliver their sludge to this incineration plant from 2015 onwards. Eventually this will lead to various changes for the operation of the wastewater treatment facility: i.e. the incinerator will provide the required heat for anaerobic fermentation. Therefore less biogas e which is currently processed to provide electricity and heat e will be required. It is expected that the surplus biogas will be sold to a gas company. The Swiss Parliament passes a law in 2014 that requires large facilities to treat micropollutants generated from the use of pharmaceuticals, hormones, personal care products, household products, etc. In Zurich, the planning for micropollutant removal begins in 2014. An ozonation plant will be built and the current sand filters will be used as a bioactive stage following ozonation. As these advanced treatment steps are not economically feasible for smaller facilities, there is increased pressure for neighbouring communities to connect to Zurich's wastewater treatment facility. The requires treatment steps and the residential developments taking place next to the facility lead to a space limitations and to new requirements such as odour control. Climate change is another challenge with increasing impact on wastewater treatment. Next to energy efficiency measures Zurich is exploring mitigation potential of reducing N2O emissions. Also, planning for climate change adaptation is underway: the treatment plant will need to provide protection for more frequent flooding. Phase 6 e main findings:

Fig. 7. Phase 6: 2000s to 2020: Energy efficiency, resource recycling, climate change, micropollutant removal e Heat recovery, sludge incineration, ozonation.

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

 Addressing energy efficiency, resource recovery, micropollutants and climate change requires substantial facility upgrades and introduction of new treatment steps.  Measures within the urban catchment, e.g. heat re-use from sewers may significantly impact the facility's treatment capacity.  More stringent requirements are addressed with highly specialised infrastructure accelerating the trend to further centralisation.

413

treatment steps. The progressive evolution of objectives and the technology transitions imply that during the last 170 years the outer physical boundary has expanded from the private household and neighbourhood to the region and eventually to the entire globe. However, former principal objectives do not disappear but are implicitly accounted for in the evolved infrastructure leading to an increase of the system's complexity.

4.2. Hindsight assessment of system performance

4. Summary and discussion The following section summarises and discusses the results of the hindsight analysis of Zurich's wastewater system. Table 2 summarises the six phases of the evolution of Zurich's wastewater system.

This case study has allowed us to examine how rapid changes influence the performance of the system. We found that the dynamics especially impacted the performance during periods of growth (phases 2, 3, 4). For these periods we summarise in the following section the times that elapsed between following events (see Table 3):

4.1. Evolving challenges, objectives, agents and technologies Evolving challenges imply progressing objectives. The main objectives in the 19th century were to provide systems that would improve public health and allow for re-use of human waste as fertilizer. The implemented systems were intended to support economic growth and city development. In the 20th century the objective transitioned to water pollution control, at first focussing on providing adequate water quality for industries, then reducing nuisances caused by aesthetic impairment and odour, and finally as a means for environmental protection. In the beginning of the 21st century, the scope has widened to enhance sustainability, resource recovery, micropollutant removal and climate change. Concerning the principal agents we observed how regime “outsiders” became “insiders” in regime transitions (Geels, 2006). In Zurich, this was especially evident in the transition of principal agent from municipal engineer to physician. For example at the end of the 19th century physician K. Reiser helped to abandon the field irrigation project of the longstanding municipal engineer A. Bürkli. Following this event, the hygienist F. Erisman took over the role as principal agent at the beginning of the 20th century. Technology transitioned from cesspits and ditches to a bucket system with liquid sewer and then on to centralised wastewater treatment with a water-borne sewer. The central wastewater treatment facility itself experienced many successive additions of

a) Investment decision b) Commissioning c) Detection of inadequacy The city engineer's recommendations and the Typhus epidemic led to the passing of the municipal bill in 1867 to introduce the bucket system. Four years later (4 y: a/b) in 1871, the system had been largely installed. Although it did improve sanitary conditions, there were unintended consequences. Already by the mid-1870s (5 y: b/c) it was becoming financially inadequate due to the devaluation of sludge and the flushing out of nutrients which led to Table 3 Transition times (in years) for the period 1867e1986 following major investment decisions (a). a/b: Time between investment decision and commissioning; b/c: Time between commissioning and detection of inadequacy; c/a: Time between detection of inadequacy to investment decision for upgrade. Transitions Adequate performance Years with major investment 1867 decisions 1923 1930 1958 1978

a/b No

b/c Yes

c/a No

4 3 2 11 8

5 1 1 0 e

47 3 25 9 e

Table 2 The development of Zurich's wastewater system from 1850 to 2020. Phase Time 1

2

3

4

5 6

Challenges

90%!). Our analysis revealed substantial and unexpected developments taking place at much smaller time scales than the physical lifetime of the implemented infrastructure. Within some of the growth periods newly constructed facilities were already obsolete at start-up (e.g. in 1927, 1933 and 1969). Although the shortcomings were typically recognised very quickly and technical solutions available at the time, it often took several decades to implement a solution. We found, that the long periods of inaction may largely be due changing political agendas which results in a competition for financial resources for actions in the public domain: e.g. education, transport, health, security. This was observed in the principal agents engaged in major wastewater infrastructure development: these were typically highly visible public figures (A. Bürkli, F. Erismann) with strong ties to policy makers and influence on setting the political agenda. The substantial trend reversals observed in Zurich in the mid1980s (de-industrialisation, reduced per capita water demand) significantly reduced pressure on the infrastructure. Freed-up capacity was used to connect additional wastewater sources and to integrate advanced treatment processes to meet new environmental legislation without the need for facility expansion. However, it cannot be generalised that underutilised systems will automatically turn out to be of value as was pointed out in the introduction for the recent case of wastewater infrastructure development in Eastern Germany (Moss, 2008). 4.3. Policy implications In summary, we find that changes to drivers acting upon the analysed wastewater system happen quickly and in unexpected and sometimes truly mindboggling ways. Realising this, are there alternatives to the current “predict and provide mode” of wastewater infrastructure design? Due to the evolutionary nature and complex characteristics of wastewater systems and the considerable dynamics of the involved drivers, we observe the predominant strategy to be a reactive form of adaptive management, i.e. “muddling-through”. In other words adaptive management “happens”. Acknowledging this fact we must put into question the current design paradigm that aims to “provide” an optimal infrastructure for a “predicted” likely future for time horizons of 25e80 years. We must therefore ask: are there alternative design strategies that would be better able to cope with the discrepancy between the long infrastructure lifetimes and the fact that various drivers quickly move the system out of the domain

it was conceived for? An alternative to the “predict and provide” mode of design is “robust design” where systems are not conceived with the notion of “working optimally for a single predicted future” but will “work satisfactorily for many possible futures”. Such an approach has been suggested to deal with the deep uncertainties related to climate change impacts on water infrastructure (Dessai et al., 2009). It requires the combined use of scenario planning tools that systematically develop meaningful storylines of possible futures (Schoemaker, 1995) and approaches that increase the system's flexibility (Spiller et al., 2015). In our analysis, we observed several instances where the introduced measures increased the flexibility of the wastewater system. The Alternating/Intermittent (S/I) mode with step feed introduced in Zurich in 2007 is an example of increasing operational flexibility. Introducing such operational flexibility allows the facility to react to changing load conditions in more energy- and resource efficient ways. Managerial flexibility which includes demand side management or source control can significantly increase the solution space of utilities avoiding the emphasis on providing enough capacity for worst case load scenarios (Larsen and Gujer, 2001). Interestingly, Zurich's 1st water pollution bill of 1881 was primarily focused on source control. Structural flexibility can be achieved by increasing modularity (smaller interchangeable components with shorter life spans that can be easily added or removed as demands or requirements change) or by providing a multipleuse infrastructure (Maurer, 2009). Concerning multiple-use infrastructure for the system of Zurich, it is speculated that A. Bürkli indeed designed the liquid waste sewers of the bucket system with diameters large enough for alternative use as water-borne sewers. Many instances were encountered where the existing structures were re-used for different processes in facility upgrades. In addition to engineering oriented types of flexibility introduced above the adaptive capacity of utilities can be further enhanced by increasing their institutional flexibility which may include the freedom to expand the business model or an increase in financial autonomy (Dominguez et al., 2009). Finally, financial flexibility in the sense of financial resources was found to be of high importance in developing wastewater infrastructure. For example, we observed how the two world wars stopped all ongoing infrastructure developments in Zurich. By contrast, Switzerland currently being a wealthy nation is the first country to introduce a law on micropollutant removal from wastewater. The authors believe there is significant scope for exploring the potential for introducing flexibility in more systematic ways for it to become an integral part of a robust design strategy. Applying this hindsight perspective is transferable to other infrastructure systems with long life spans. However, limited data availability may significantly hinder the feasibility of conducting such types of studies. Continuous long-term performance evaluation is a requirement to improve knowledge on infrastructure evolution. Such long-term assessments would allow for improved post-project auditing and performance benchmarking and would increase the accountability in public infrastructure provision (Guimar~ aes et al. (2010), Flyvbjerg et al. (2003)). 5. Conclusions By taking a long-term view of the technological evolution of wastewater infrastructure this study examined the discrepancy between the “predict and provide” mode of design for long-lived structures and the rapid changes occurring in the driving forces. This hindsight analysis demonstrated that the drivers influencing wastewater loads, environmental requirements and technology can change at considerably smaller timescales than the

M.B. Neumann et al. / Journal of Environmental Management 151 (2015) 404e415

infrastructure design lifetime, often in much less than a decade. Repeatedly we found that the dynamics and interactions of drivers rapidly move wastewater infrastructure systems outside of their “planned for” domain. We found that the built infrastructure is continuously confronted with challenges it was not conceived for and that significant adaptation occurs during an infrastructure's lifetime. Many of the drivers, such as changing medical theories, world wars and industry shrinkage are beyond the scope of wastewater professionals. We found that “muddling through” is the dominant strategy for adaptive management. Our analysis of the long-term development of a waste water system, suggests that a hindsight perspective can inform the development of “robust design” strategies. The findings obtained from hindsight studies can support scenario building, in order not to underestimate future changes, and indicate the type and amount of flexibility required. Acknowledgements We especially want to acknowledge Martin Illi without whose substantial effort in recounting the detailed history of urban water management in Zurich our analysis would not have been possible (Illi, 1987). We thank Damian Dominguez for valuable discussions on how to strengthen the wastewater engineer's perspective on socioeconomic processes. We would like to express our gratitude to Entsorgung und Recycling Zürich (ERZ), the wastewater utility of Zurich, and in particular to Christian Abegglen for input on recent and current developments. References Belia, E., Amerlinck, Y., Benedetti, L., Johnson, B., Sin, G., Vanrolleghem, P.A., Gernaey, K.V., Gillot, S., Neumann, M.B., Rieger, L., Shaw, A., Villez, K., 2009. Wastewater treatment modelling: dealing with uncertainties. Water Sci. Technol. 60, 1929e1941. Brentano, M., 1934. Untersuchungen über die Abwasserverh€ altnisse von Zürich (Investigations into issues of wastewater in Zurich). in German. Ph.D. thesis no. 798, ETH Zurich, Switzerland. http://dx.doi.org/10.3929/ethz-a-000091723. De Neufville, R., 2003. Real options: dealing with uncertainty in systems planning and design. Integr. Assess. 4, 26e34. Dessai, S., Hulme, M., Lempert, R., Pielke, R., 2009. Adapting to Climate Change: Thresholds, Values, Governance. In: Chapter 5: climate prediction: a limit to adaptation?. Cambridge University Press, Cambridge, UK, pp. 64e78.

415

Dominguez, D., 2008. Handling Future Uncertainty. PhD Thesis no. 17867, ETH Zurich, Switzerland. http://dx.doi.org/10.3929/ethz-a-005779391. Dominguez, D., Gujer, W., 2006. Evolution of a wastewater treatment plant challenges traditional design concepts. Water Res. 40, 1389e1396. Dominguez, D., Worch, H., Markard, J., Truffer, B., Gujer, W., 2009. Closing the capability gap: strategic planning for the infrastructure sector. Calif. Manage. Rev. 51, 30e53. Flyvbjerg, B., Bruzelius, N., Rothengatter, W., 2003. Megaprojects and Risk: an Anatomy of Ambition. Cambridge University Press, Cambridge, UK. Geels, F.W., 2006. The hygienic transition from cesspools to sewer systems (1840e1930): the dynamics of regime transformation. Res. Policy 35, 1069e1082. Geldof, D.G., 1995. Adaptive water management: Integrated water management on the edge of chaos. Water Sci. Technol. 32, 7e13. ~es, B., Simo ~es, P., Marques, R.C., 2010. Does performance evaluation help Guimara public managers? A balanced scorecard approach in urban waste services. J. Environ. Manage 91, 2632e2638. Holland, J.H., 1992. Adaptation in Natural and Artificial Systems: an Introductory Analysis with Applications in Biology, Control and Artificial Intelligence, first ed. MIT Press, Cambridge, MA, USA. €sserung (From cesspits Illi, M., 1987. Von der Schîssgruob zur modernen Stadtentwa to modern urban drainage). in German. Neue Zürcher Zeitung (NZZ) and €sserung Zürich, Abteilung des Bauamtes I, Zürich, Switzerland, ISBN Stadtentwa 3-85823-387-0. Larsen, T.A., Gujer, W., 2001. Waste design and source control lead to flexibility in wastewater management. Water Sci. Technol. 43, 309e317. Maurer, M., 2009. Specific net present value: an improved method for assessing modularisation costs in water services with growing demand. Water Res. 43, 2121e2130. Milly, P.C.D., Betancourt, J., Falkenmark, M., Hirsch, R.M., Kundzewicz, Z.W., Lettenmaier, D.P., Stouffer, R.J., 2008. Stationarity is dead: whither water management? Science. 319, 573e574. Moss, T., 2008. Cold spots of urban infrastructure: shrinking processes in Eastern Germany and the modern infrastructure ideal. Int. J. Urban Reg. 32, 436e451. Neumann, M.B., 2007. Uncertainty Analysis for Performance Evaluation and Design of Urban Water Infrastructure. PhD Thesis no. 16975, ETH Zurich, Switzerland. http://dx.doi.org/10.3929/ethz-a-005386170. Neumann, M.B., Vanrolleghem, P.A., 2011. Use of variance decomposition in the early stages of WWTP design. In: Proceedings 11th IWA Specialised Conference on Design, Operation and Economics of Large Wastewater Treatment Plants. Budapest, Hungary, September 4-8 2011, pp. 437e440. Schoemaker, P.J.H., 1995. Scenario planning e a tool for Strategic thinking. Sloan Manage. Rev. 36, 25e40. Schultz, K.S., McShane, C., 1978. To engineer the metropolis: sewers, sanitation, and city planning in late-nineteenth-century America. J. Am. Hist. 65, 389e411. Spiller, M., Vreeburg, J.H.G., Leusbrock, I., Zeeman, G., 2015. Flexible design in water and wastewater engineering e Definitions, literature and decision guide. J. Environ. Manage 149, 271e281. €sserung, 1986. Erweiterung Kla €ranlage Zürich Zürich Hauptabteilung IV Stadtentwa €lzli 1980e85 (Expansion of the Zurich Werdho € lzli wastewater treatWerdho ment plant). in German. Tiefbauamt der Stadt Zürich, Zürich, Switzerland.