Introduction

“Harvesting experience (i.e., systematically and rigorously validating and documenting examples of success in Sustainable Urban Water Management) is needed to build a robust evidence base for the implementation of “what has already worked” in new locations so that resources can be directed toward appropriate adaptation to local contexts rather than wholesale reinvention.”

J. Hering and K. Vairavamoorthy (2018)

Sustainable management of urban water supply, stormwater, and wastewater, is critical to the health and well-being of people in urban areas, and to ecosystems affected by urban land and water use. The demands of population growth, aging infrastructure and changing climate will increase pressure on all of these systems in the future. Here, we describe an approach that builds on the concept of the city blueprint (a framework to assess urban water sustainability of cities globally) to identify vulnerabilities in the urban water system at the spatial extent of entire cities, and then uses multi-scale design and evaluation of alternative future scenarios to explore and inform decision makers of the potential outcomes of different options. Our approach uses basin-scale modeling to characterize the regional context for the scenarios. Finer-scale modeling and evaluation of the scenarios at the spatial extent of watersheds and urbanizing neighborhoods is used to evaluate options and provide information to managers at scales relevant to their decision-making needs.

While urban areas occupy only about 3% of the world’s land surface (Liu et al. 2014), they have a disproportionate impact on the hydrologic regimes of associated river systems and their water quality. The proportion of the world’s population living in cities is expected to reach 66% by 2050. More than 70% of global greenhouse gas emissions are contributed by energy use in urban areas, and just 50 of the world’s largest cities draw their water supply from watersheds covering 41% of the world’s land surface (NSF ACERE 2018). In coastal regions, these impacts extend to estuarine and coastal environments, exacerbating the formidable pressures from human activities on coastal ecosystems.

With increasing pressures on urban water systems (hereafter called UWS), and a growing emphasis on sustainable solutions that integrate economic, environmental and social needs, urban water management is undergoing a transition. Water managers are advocating for integrated approaches that link different elements of the urban water cycle, and emphasize the need to break down management silos (Owen 2015; US Water Alliance 2016; Roth and Mazza 2017). For example, wastewater and stormwater can become a resource through water reuse, and flood mitigation can be integrated with urban park and habitat management. These approaches have been called variously “water sensitive cities” (Wong and Brown 2009), “sustainable urban water management” (Marlow et al. 2013), and “One Water” approaches (Mukheibir et al. 2014; US Water Alliance 2016; Bolson et al. 2018).

The City Water Blueprint was developed as a framework to assess the sustainability of urban water management for cities at a global scale (van Leeuwen et al. 2012; Koop and van Leeuwen 2015). City Water Blueprints can help assess vulnerabilities of UWS, distinguish barriers to effective water management solutions and inform policy with the goal of improving water management solutions through integrated planning, i.e., “One Water Solutions”. However, translating “One Water Solutions” into regionally-relevant actions for water managers, and involving the general public to build understanding and support for sustainable water management and planning in their cities requires more than a blueprint. Region and location-specific approaches are needed to 1) engage diverse water managers to help identify key concerns and develop locally-relevant, cross-scale solutions, 2) identify important system attributes and responses and 3) quantify those responses under alternative future management and climate scenarios.

Future planning for managing UWS requires not only the ability to assess existing systems but also methods to imagine, design, and evaluate the response of innovative UWS that anticipate coming change. To be most effective, planning and modeling approaches must include not only conventional practices and infrastructure, but also innovative water management options and new types of infrastructure that conceive natural and constructed parts of water systems as an integrated whole (Mitchell 2006; Owen 2015). However, innovative methods for water management pose a challenge to planners and modelers precisely because they are novel (Santelmann et al. 2001). In many regions, we lack existing analogs or long-term data sets needed to model system behavior and effectiveness of a given solution across different contexts and scales. Few studies have assessed the cumulative effects of grey and green infrastructure solutions at the spatial extent of watersheds on the order of 100 to 1000 ha in size (Jefferson et al. 2017) and fewer still have integrated across spatial extents from parcel to watershed to basin. Yet, scale-dependency of impacts and trade-offs among those impacts across different sectors and scales has been identified as a key element of urbanization. Seto et al. (2017) note that “… urban sustainability is inherently multidimensional and multiscalar … Solutions at one spatial scale or along one dimension (e.g., energy, land use) may not have the same effect at a different spatial scale or dimension. One example of this is the spatial pattern of urban land use, which results in varying trade-offs at different scales”.

Urban ecosystems are well-suited to the application of a geodesign conceptual framework (Steinitz 2012), in part because the process requires explicit considerations of scale and context. Geodesign incorporates interdisciplinary perspectives, design across specific geographic scales, and uses GIS-based analytic and design tools to explore alternative future scenarios and their impacts. The geodesign process incorporates six different types of models, including: 1) Decision models, 2) Representation models, 3) Process models, 4) Evaluation models, 5) Change models, and 6) Impact models. Sustainable UWS will require not only innovative design across sectors and scales, but also the use of effective and appropriate models to assess the system response across multiple sectors, dimensions and scales (Steinitz 2012). Because urban ecosystems are complex, and system response must be assessed for multiple systems, the development and implementation of innovative UWS poses major challenges to planners and decision makers.

These challenges, and an approach to develop the capacity to model whole system responses in meeting them, are identified and discussed here. We group the steps of this approach under three broad categories, as follows:

  1. 1.

    Building Collaborations for the Future: Engagement of stakeholders– envisioning water solutions that are likely to be applied by diverse water managers across whole water systems requires that research teams work with knowledgeable stakeholders to clarify priorities, which often include exploring a variety of integrated policy and technological solutions.

  2. 2.

    Field Testing the Future: Transdisciplinary design of systems– an iterative approach to the design of future scenarios can help water managers, planners and the public explore high priority UWS alternatives and depict them across scales. This process allows combined Researcher/Stakeholder teams to incrementally agree on water system priorities at each scale, and to better understand the linkages among them. Use of an incremental approach allows the time necessary for development of alternative futures that are plausible, acceptable to end-users, and feasible for the research team to evaluate.

  3. 3.

    Simulating the Future: Modeling and evaluation of whole UWS performance requires both data and an array of modeling and evaluative tools that can be used to investigate system response at appropriate spatial and temporal scales. Models must be able to evaluate both current systems and alternative futures at the multiple scales relevant to different end-users.

Multiscale, spatially- and temporally-explicit scenario approaches are powerful methods for working with stakeholders. These approaches can be used to focus attention on key systems of interest, and then illustrate existing conditions and future trajectories in familiar ways, ways compatible with how people actually experience places. Here, we define a spatially-explicit scenario as one that is specific to a place, and linked to georeferenced data that describe the attributes of that place (e.g., topography, land cover, climate, and census data). Scenarios, and the models used to generate representations of these scenarios, are both spatially- and temporally-explicit when they simulate and specify the changes in the spatial attributes of place over specific time frames.

Building collaborations for the future: identifying priorities in transdisciplinary projects

“Alone we can do so little, together we can do so much.”

Helen Keller (Lash 1980)

To illustrate our multiscale, multi-sector approach and its application to the challenges inherent in development of innovative UWS, we report here on one part of the national Urban Water Innovation Network or UWIN, a collaborative effort of water stakeholders and researchers from across the U.S.A. (Bolson et al. 2018). The summary that follows uses work from one UWIN study region in the U.S. Pacific Northwest, western Oregon’s Willamette River Basin (Fig. 1). We highlight efforts of a joint researcher/stakeholder team that defined priorities for urban water management in this region. Using input from an iterative series of stakeholder/researcher collaborations that took place over the course of two years, the team developed three ca. 2060 mapped alternative futures, shown at three nested spatial extents; the 3 million ha. Willamette River Basin, the 4000 ha Chicken Creek watershed, and the 100 ha. West Sherwood urban neighborhood. The scenarios depicted responses to future pressures and trends that reflected stakeholder priorities and were considered plausible based on their expert knowledge.

Fig. 1
figure 1

Context Map of Willamette River Basin within U.S.A. and Oregon; showing spatial extent of the nested study regions; Willamette River Basin, Chicken Creek Watershed, and Sherwood Focal Neighborhood

Full size image

The stakeholder group included representatives of city, county, and state governments, citizens, coalitions, nonprofits and businesses. Members volunteered their time. They were selected based on a combination of professional affiliation and demonstrated expertise in four important water sectors (Table 1), as well as interest in exploring new ideas and sustainability. We purposely sought to engage members who had demonstrated leadership in the various water sectors, and were willing to assist the research team in vetting indicators of water sustainability.

Table 1 Key system requirements and goals for the four urban water sectors, and the opportunities that could be provided by integration across sectors
Full size table

The earliest engagements with the Oregon stakeholder group occurred as part of a series of national meetings intended to clarify regional similarities and differences in water management challenges across the U.S.A. During regional meetings, data were collected through pre- and post-meeting surveys and observations of open-ended discussions to identify stakeholder perspectives on key urban water challenges, drivers and pressures propelling these challenges, and promising solutions (Bolson et al. 2018).

Bolson et al. (2018) identified common concerns among water managers across the country, but also highlighted regional differences (Table 2). For example, climate change and ecosystem degradation were frequently cited challenges across regions, but the specifics of those concerns differed. Concerns about climate change in the Florida stakeholders centered around risks and impacts associated with sea level rise, while concerns in the Pacific Northwest focused on reduced water availability during hotter, drier summers and competing demands from a growing population. Ideas on how to overcome those perceived challenges varied by region. Among the Pacific Northwest stakeholders, there was strong interest in collaborative planning and looking for win-win opportunities where a single project would derive multiple benefits for diverse sectors or groups. There was also a desire to think across spatial scales. For example, stormwater managers advocated for exurban riparian protection as a cost-effective way to reduce thermal loads before streamflow entered developed areas, and water providers placed water availability in the context of water needs for agriculture and endangered fish.

Table 2 Summary of national and region-specific challenges and solutions for urban water management identified by national UWIN stakeholder meetings (Bolson et al. 2018). National themes gain specificity when considered in a regional context
Full size table

Following the initial meetings that contributed to the national survey, we held a series of quarterly Oregon researcher/stakeholder meetings, each lasting approximately three hours. These meetings were dedicated to more detailed discussions of scenario design, leading to spatially- and temporally-explicit depictions of plausible alternative futures. In the sections that follow, we highlight key qualities of the alternative futures developed in this collaborative process, and summarize them in Table 3.

Table 3 Multi-scale alternative future scenario assumption matrix. Working with a regionally representative group of stakeholders, researchers clarify assumptions about key concerns that distinguish plausible, water-relevant scenarios at river basin (~30,000 km2), watershed (~4000 ha) and neighborhood (~100 ha) spatial extents, all with a ca. 2060 planning horizon
Full size table

Field testing the future: Designing innovative, multi-scale alternative futures

“Premier among the consequences of being human is the capacity to imagine possible futures, and to plan and choose among them.”

E. O. Wilson (2014)

Scenario-driven alternative futures analysis is an approach to anticipatory assessments designed to inform community decisions about the likely effects of different options for future land and water use (Santelmann et al. 2001; Hulse et al. 2004; Liu et al. 2007; Mahmoud et al. 2009; Steinitz 2012; Wu et al. 2015, Rastandeh 2015). The primary advantage of this approach is as an integrative procedure for connecting the different ways of understanding the processes that shape landscapes. These range from the biophysical processes in the disciplinary realm of the sciences and engineering to the social policy and management decisions that influence land and water use. While the specific characteristics of scenario-driven alternative future studies are as diverse as the situations in which they are applied, common qualities arise. These studies begin by defining discrete, coherent assumptions about how conditions of interest unfold in some bounded place over some specified period of time. A logically coherent group of these assumptions comprise a scenario (Wodak and Neale 2015; IPBES 2016). For many scenario studies it suffices to represent scenarios with narrative descriptions alone, but, in our usage, a spatially-explicit representation of a scenario’s land and water use at temporally-explicit time steps comprise an alternative future.

The next step in scenario-driven alternative future studies subjects these alternative future representations to a series of computational evaluation models to learn what effects each alternative future may have on some defined set of system responses. Continued collaboration with stakeholders helps focus this evaluation on the elements of the system that people care about. This evaluative phase is often of prominent concern to stakeholders, because it allows them to compare and contrast the modeled implications of choosing one alternative over another. Where the focus is on anticipating trajectories of change caused by human use of land and water, scenario-driven approaches provide a framework for effectively incorporating science into community-based decision-making and, in the process, for fostering a more sophisticated dialogue of the facts, values and perceptions that underpin informed land and water management. And when this is done well, E.O. Wilson argues, it allows us to ‘field test the future’.

Defining the roles of stakeholders and experts

Where innovation is a motivating purpose of scenario-driven alternative futures projects, especially innovation concerning complex systems such as those needed to protect, provide and treat urban water, there is a central need for engagement of both knowledgeable stakeholders and scientific experts. Such transdisciplinary engagement has inevitable transaction costs (Mahmoud et al. 2009; Musacchio 2009). One source of high transaction costs are the different ways of thinking and working prevalent among professionally-diverse project participants, and the varying incentive systems in their professional lives that motivate them to know a little about a lot, or a lot about a little (Ferguson et al. 2016).

In the case of urban water innovations, there are often trade-offs between a participant’s a) technical expertise, b) degree of familiarity with the various sectors and systems (e.g. drinking water, stormwater, wastewater, natural waterbodies) and associated rate payer and user fee-driven business models that may fund these sectors, as well as c) imaginative capacity to conceive truly innovative futures (Steinitz 2012). It can be difficult for those who know existing systems well to imagine a radically different future, that is, to think beyond what they have personally known and experienced (Baker et al. 2004; Hulse et al. 2016). As a result, a central challenge when designing and modeling alternative urban water futures is to ensure that both alternative futures and evaluative models bracket innovative practices, and that they do not unduly constrain options and choices for future UWS.

Highlights of the Willamette River Basin alternative futures effort

The Willamette River Basin is the Pacific Northwest case study area of the national UWIN project. As was the case across all the national UWIN study areas, Oregon stakeholders (called the Stakeholder Advisory Committee (SAC)) concluded that climate change and population growth were key pressures and trends for the future (Bolson et al. 2018). Because so many biophysical processes that effect water are driven by climate forcings, the research team built on prior work in the Willamette Basin (Hulse et al. 2016; Jaeger et al. 2017) and recommended using future climate assumptions from spatially-downscaled regional outputs of atmosphere-ocean general circulation models (GCM) in combination with Representative Concentration Pathways (RCP) that integrate emissions concentrations, land-use change and socio-economic responses. For this effort we selected two GCM x RCP combinations representing medium and high impacts on vegetation, evapotranspiration and wildfire at the spatial extent of the Willamette Basin. The chosen climate models from which these assumptions were derived have been shown to perform well in replicating the historical climate of the U.S. Pacific Northwest (Rupp et al. 2013).

Population and future housing need assumptions were derived from state-sanctioned projections at the county level for the river basin extent (Jaeger et al. 2017), augmented with transportation planning population projections for the watershed extent (City of Sherwood 2014) and local preliminary growth concept plans at the neighborhood extent (City of Sherwood 2016a). Use of local population and housing projections for local spatial extents, in tandem with coarser-extent population projections from state-level sources, proved important in convincing stakeholders that projections were sufficiently realistic and responsive to local constraints to merit consideration.

With future climate and population assumptions agreed, the Research Team worked iteratively with the stakeholders over two years to depict three 2060 alternative futures at each of three spatial extents - basin, watershed and neighborhood. The watershed (~4000 ha) and neighborhood (~100 ha) extents and locations were identified jointly with the stakeholders over a series of meetings as a consensus emerged regarding key water concerns and the spatial grains and extents at which alternative futures addressing these concerns would need to be explored and depicted. The three futures were named: Current Course ca. 2060, Stressed Resources ca. 2060 and Integrated Water Future ca. 2060, each corresponding to a central organizing scenario theme. In Current Course ca. 2060, the theme is that the current land and water regulatory regime is assumed to remain in force through 2060. In Stressed Resources ca. 2060, water treatment and conveyance systems are assumed to be stressed as a result of the combined pressures of population growth, changing climate, and policies that respond to, but do not anticipate the need for change. In Integrated Water Future ca. 2060, water quality and quantity management are assumed to be coordinated at the watershed extent in ways that incorporate innovation and anticipate future pressures. Reduced to its essentials, the process of working with the stakeholders began with 1) agreeing on overarching scenario themes and then progressively clarifying scenario assumptions for the same set of key concerns for each future at basin, watershed and neighborhood extents; then 2) iteratively presenting, discussing and revising, based on stakeholder guidance, the mapped spatial patterns of land use and land cover at each spatial extent. The futures were deemed complete once a consensus was reached among the stakeholders that the mapped patterns adequately represented the stated assumptions regarding each concern.

The key concerns are noted in the left column of Table 3. Definition of assumptions for each key concern progressed from general concepts, such as the assumed degree of future coordination across discrete water management sectors, to agreeing on parameter values (e.g. number of new households), to detailed spatial allocations (e.g. where should new homes be sited) based on local conditions. The focus in this portion of the effort was on producing depictions of land and water use that, in the judgement of the combined stakeholder/researcher team, were consistent with the stated assumptions for each key concern of each alternative future scenario at each extent. These scenario assumptions are listed in the body of Table 3 and, for illustration, the corresponding mapped patterns for watershed and neighborhood extents are shown in Figs. 2 and 3, respectively.

Fig. 2
figure 2

Land use and land cover depictions at the Chicken Creek Watershed spatial extent for the existing (circa 2010) and three alternative future landscapes (circa 2060)

Full size image
Fig. 3
figure 3

Historic air photo comparison of Sherwood area in 1977 (top left), and 2011 (lower left); and three ca. 2060 alternative future land use and land cover depictions at the Sherwood Focal Neighborhood spatial extent

Full size image

We next offer a description of the study area, briefly characterizing each spatial extent. The Willamette River Basin contains about 12% of the land area of Oregon, is home to over 69% of the state’s population and is more than 60% forested. Forests occur predominantly in the higher elevation montane area. These upland forests evolved in a fire-dependent natural disturbance regime. The spatial extent of these forests and their rates of evapotranspiration are important factors in the basin’s water budget (Turner et al. 2017). The Mediterranean climate is typified by cool, wet winters and warm, dry summers. Summer rainfall makes up <5% of the annual average total of 1370 mm, which ranges from 950 mm to more than 2000 mm across the basin. The combination of warm summers, ample spring moisture and fertile volcanic soils supports significant agricultural production within the valley, estimated by the Oregon State University Extension Service (Sorte and Rahe 2015) as influencing more than 13% of sales of products and jobs in the state.

The Chicken Creek Watershed spans the southwest edge of Portland Metro’s Urban Growth Boundary (UGB) (Fig. 1). Although the majority of the watershed is currently outside of Portland’s UGB, 60% of the City of Sherwood, which is inside the UGB, lies within the Chicken Creek Watershed. Portland has identified an Urban Reserve area, called the West Sherwood Urban Reserve, adjacent to the City of Sherwood within the Chicken Creek Watershed. Urban Reserves are an adopted regional growth management device which delimits a bounded territory outside the current UGB showing land designated to accommodate all urban growth for the next 50 years. At present, the primary land uses within the Chicken Creek watershed that are outside of the urbanized City of Sherwood but inside the Urban Reserve are in agricultural and rural residential uses. As of 2010, approximately 40% of the watershed was forested land cover.

The Sherwood Focal Neighborhood, as yet unurbanized, is adjacent to the western edge of the Sherwood UGB in the West Sherwood Urban Reserve. In 2010, the primary land uses in the Sherwood Focal Neighborhood were agriculture and rural residential. The City of Sherwood has already planned and sited (although not yet constructed) a new high school in the northeast corner of the neighborhood. Stakeholders requested that this high school be incorporated into the alternative futures for all three circa. 2060 Sherwood focal neighborhood futures.

The alternative futures

We next briefly describe the three alternative futures, focusing here on the watershed and neighborhood extents because of the heightened attention stakeholders devoted to them. The watershed and neighborhood spatial extents are where, in the inherited land tenure system, people and organizations own and manage land and actively alter land use and land cover in areas they control. Prior Willamette Basin efforts and stakeholder experience had previously produced a well-developed basis for scenarios at the basin extent (Hulse et al. 2002, 2009, 2016; Baker et al. 2004; Wu et al. 2015; Jaeger et al. 2017). Of particular note here are the logical and spatial intersections of i) the mapped representations of changes that impact water when lands are converted from nonurban to urban uses, ii) the mapped locations where these impacts affect more than one major water management sector (Table 1), and iii) the capacity of evaluative models to defensibly detect these impacts across scales as portions of the landscape change from nonurban to urban use. By way of reminder, an alternative future consists of a spatial arrangement at a particular time of land use and land cover types in mapped representations in which locations are associated with site attributes at a specific time. This arrangement represents scenario assumptions in spatially explicit (i.e. a neighborhood, a watershed) and temporally explicit (i.e. ca. 2060) form. Regardless of the ways stakeholders may wish to conceive futures, or the ways researchers may wish to evaluate them, the conversation, depiction and evaluation occurs largely through a vocabulary of mapped classes of land use and land cover at discrete spatial extents and time steps, a vocabulary that necessarily constrains evaluative models, since it is primarily the depiction of landscape conditions in space and time that the models evaluate.

Current course ca. 2060

The Current Course ca. 2060 scenario assumes a continuation of current trends in population, climate change and land use patterns. At the Willamette River Basin extent, the population reaches 4.51 million by 2060 and the average annual temperature is 2.5 °C. higher than the 1950–2005 average. These underlying assumptions and guidance from stakeholders resulted in expanded residential and road patterns consistent with the ca. 2010 landscape of the already developed adjacent neighborhood of Sherwood – i.e. single family detached homes and a road network to accommodate private vehicle access to each tax lot. The Current Course ca. 2060 Sherwood Neighborhood accommodates 1218 new dwellings (36% of these dwellings at a residential density of 0–10 dwelling units per hectare (du/ha), 28% at a density of 23–40 du/ha and 36% at a density of >40 du/ha), resulting in 63% of the neighborhood area in impervious cover. This number and mix of residential dwelling densities allowed creation of 32 ha of green infrastructure facilities, occupying 28% of the neighborhood extent, with 12 of these 32 ha in native riparian vegetation.

Stressed resources ca. 2060

Stressed Resources ca. 2060 assumes a 37% increase in population over Current Course ca. 2060 (at the Willamette River Basin extent, a total of 6.16 million) and greater increases in average annual temperature (an increase of 4.0 °C). The primary assumption in Stressed Resources ca. 2060 is that water treatment and conveyance infrastructure is stressed. More conventional approaches to water management are employed rather than innovations that anticipate the need for flexible response to future change. At the neighborhood extent this means minimum protection of high priority conservation lands (riparian function in particular) and no constraints on the amount of impervious surface. Residential development is still focused on single family detached homes with road networks to accommodate private vehicle ownership. Stressed Resources ca. 2060 accommodates 1610 new dwelling units, 56% at a density of 0–10 du/ha and 44% at a density > 40 du/ha. Accommodating the greater number of dwelling units with more than half again as many single family detached homes (0–10 du/ha) required a significant increase of dwelling units at the highest density (>40 du/ha) and resulted in a residential development pattern that occupies a greater percentage of the neighborhood relative to Current Course ca. 2060, thus increasing impervious cover to 72% of the neighborhood. This number and mix of residential dwelling densities allowed creation of 16 ha of green infrastructure facilities, occupying 14% of the neighborhood extent, with 8 of these 16 ha in native riparian vegetation.

Integrated water future ca. 2060

Integrated Water Future ca. 2060 assumes the same increases in population and average annual temperature as Stressed Resources ca. 2060 (6.16 million people at the Willamette River Basin extent, an average annual temperature increase of 4.0 °C). Integrated Water Future ca. 2060 is guided by the central organizing principle of coordinated management of water quality and quantity at the watershed extent. Implicit in this scenario is the expectation that innovation will be incorporated into UWS in anticipation of the need for flexibility in the face of future change. At the neighborhood extent this means that protection of watershed riparian function has priority and is considered before the siting of new built features. Integrated Water Future ca. 2060 accommodates 1494 new dwellings, 57% at a density of 23–40 du/ha and 43% at a density of >40 du/ha, with 55% of the neighborhood in impervious cover. This number and mix of residential dwelling densities allowed creation of over 55 ha of green infrastructure facilities, occupying 47% of the neighborhood extent, with 16 of these 55 ha in native riparian vegetation. The primary road network still accommodates private vehicle ownership but Integrated Water Future also incorporates green boulevards that by 2060 have transitioned to focus on alternative modes of transportation (Fig. 4).

Fig. 4
figure 4

Detail of Integrated Water Future Neighborhood Green Boulevard evolving from ca. 2020 to ca. 2060

Full size image

Table 4 contrasts selected landscape metrics for the three alternative futures at the watershed and neighborhood extents. Higher density development in the Integrated Water Future Scenario accommodated about 60% more dwelling units than the Current Course scenario, while maintaining about the same developed land area as Current Course. The emphasis on riparian protection in this scenario also created more habitat for native species, when estimated using a habitat suitability model described by White et al. (1997). In contrast, the emphasis on single family detached homes and limited protection of natural areas, created higher imperviousness and reduced habitat for terrestrial native species in this scenario.

Table 4 Selected landscape metrics for the three alternative futures
Full size table

Simulating the future: modeling across scales

“Creativity of nature and human creativity cannot be separated. But creativity implies surprises: David of Michelangelo or relativity theory. Uncertainty and surprise are part of human destiny.”

Ilya Prigogine (2005)

While the national UWIN stakeholder meetings helped identify general stakeholder concerns about the future of urban water management, the subsequent alternative futures process helped the Pacific Northwest stakeholders articulate those concerns with specificity, and group them into thematic, regionally plausible narratives. The next step to make the scenarios meaningful in a decision-making context, is to use evaluative models to compare and contrast the water management outcomes of the three alternative futures. Integrated evaluative models provide a tool for both pragmatic exploration of alternatives and theory building (Bach et al. 2014; Seppelt et al. 2009). When planners and decisions makers are involved, integrated models can become a tool for communities to ask “what if” questions about the future, and to articulate and visualize information helpful for long-range planning (Hulse et al. 2004; Nassauer and Opdam 2008; Steinitz 2012). In this section of the paper we highlight two challenges for modeling integrated UWS that are illustrated by this case study.

Integrated modeling across scales

Work with the SAC highlighted the need for water system models that represent both built and natural systems and can measure the cumulative effects of infrastructure choices. For example, members of the SAC advocated for a watershed perspective, arguing that it takes both site-scale stormwater management and basin-scale approaches such as preservation of riparian corridors, to minimize stormwater impacts on natural stream systems. They also argued that it is more cost-effective to preserve and restore exurban riparian areas, than to retrofit green infrastructure into developed areas with high real estate values. They noted that the riparian areas provided ecosystem and recreational co-benefits and specifically requested that the research team model and evaluate those benefits. In 2004, Clean Water Services, the regional water resources management utility for the Tualatin River Basin, received the nation’s first watershed-based NPDES permit that combines wastewater and stormwater (City of Sherwood 2016b). The 2016 renewal of the stormwater permit also included a requirement that CWS assess and address watershed level hydromodification – changes in hydrologic characteristics which degrade water resources including stream erosion and changes in channel morphology (City of Sherwood 2016b).

Modeling the effectiveness of both site-scale and watershed-scale stormwater management is challenging because it spans typical modeling approaches and scales. Stormwater systems are typically designed and evaluated at the neighborhood or city scale with a focus on infrastructure performance up to the point of connection with the natural stream system. Conventional modeling tools developed for the purpose of stormwater engineering design, such as the EPA’s Storm Water Management Model (SWMM) (Rossman 2015) facilitate site-scale stormwater design choices such as location and sizing of pipes, low-impact development features, and other water quality control measures. However, the cumulative effect of the stormwater control measures on natural channels, watershed hydrology and aquatic ecosystems is less frequently evaluated and less well understood (Jefferson et al. 2017). While storm drainage design models like SWMM are referred to as urban hydrologic models, Salvadore et al. (2015) comment that they are “not designed to analyze catchment-sized problems”. Such models often simplify or do not represent processes such as sub-surface flow and evapotranspiration, processes that are important for determining streamflow over larger space and time scales. In contrast, models that prioritize a watershed-scale perspective better represent water flow paths in the natural environment, but typically do not represent the details of flow through the stormwater network or the spatial heterogeneity of the urban environment (Salvadore et al. 2015). A lack of detailed data and the complexity of process interactions in the built environment, create uncertainty about predictions from these watershed-scale models (Salvadore et al. 2015; Bach et al. 2014).

If UWS of the future are to integrate across these systems, coupling the stormwater system and the natural stream system, for example, or linking re-used wastewater into the urban water supply system, then research and modeling tools will need to span these sectors, too. They will also need to employ modeling tools that managers consider sufficiently precise and reliable to use in real-world settings.

Integrated modeling across disciplines

Work with the SAC also highlighted the need for models that take into account both biophysical and socioeconomic controls on the water system, and that consider regional factors that operate beyond the city scale. For example, one of the initial concerns expressed in stakeholder meetings (Table 2) was the influence of exurban summer water demands on future water availability. Several stakeholders were engaged in the Willamette Basin Review, a process to reallocate water stored in federal reservoirs in the upper Willamette River Basin (USACE 2017). The reallocation process will determine how much stored water will be reserved for future municipal uses as well as for other water needs such as agricultural irrigation, and summer flow augmentation for water quality and habitat. Future water availability then is not only a function of urban water needs, but also of the regional water supply context, and the laws and institutions that control access to water.

These metrics of landscape response at each spatial extent (shown in Table 4) are the outcomes of process models and resulting representational models, intermediate stages in the workflow of a geodesign process. The next steps in that process include application of models to evaluate change and impact in the study units at each scale, with models appropriate to the purpose and scale of each evaluation.

Interest in the connection between biophysical and socioeconomic processes were part of the rationale for development of a coupled human-natural systems model for the Willamette River Basin called Willamette Envision (Turner et al. 2017; Jaeger et al. 2017). One of the goals of this model is to link biophysical controls on water availability such as snowpack accumulation and upland forest dynamics, with institutional controls on water availability such as water rights and reservoir operations. Integration of both human and natural systems in the model helped researchers identify unexpected interactions and feedbacks that affect water supply and demand. For example, simulations with warmer climate conditions created more wildfire in the 70% of the basin covered by forests (Hulse et al. 2016). The reduction in forest leaf area contributed to reduced evapotranspiration, and higher streamflows, despite the warmer climate conditions (Turner et al. 2017).

The Willamette Envision model illustrates the potential gains from modeling across disciplines and sub-systems. Bach et al. (2014) describe a classification scheme for integrated urban water models that corresponds with the types of infrastructure and water systems that the models represent. Those at the highest level of complexity are the coupled human-natural systems models, which Bach et al. (2014) label integrated urban water system models (IUWSMs). These models integrate across both disciplinary boundaries (e.g. economics, sociology, engineering, hydrology, ecology) and urban water infrastructure sectors (e.g. wastewater, stormwater, drinking water, natural stream systems). They note that few existing models achieve this level of integration and call for more interdisciplinary work to advance integrated modeling research. This case study illustrates that the interests and concerns of water managers support the need for this work.

Conclusions

Transdisciplinary research projects involving academics and agency professionals are well suited to exploring and testing the efficacy of new, innovative design options to evaluate multiple impacts on the urban water system in the broadest sense, including natural stream systems and the public perceptions of the livability of their neighborhoods. Such projects are often of interest to professional stakeholders precisely because they allow exploration of options many years in advance of when expensive commitments must be made to publicly-funded infrastructure assets.

Numerous challenges must be overcome to develop the sustainable UWS of the future; among them the need to allow exploration of both locally intensive and spatially-extensive options for real innovation. The multi-scale alternative futures process described here highlights the benefits of a participatory process that engages partners across water management sectors to identify and explore integrative practices. Participation by local experts in the scenario design process helps identify innovations that are regionally plausible and feasible and creates a forum for stakeholders to explore and share innovative practices. The scenarios developed with this process also highlight the need to develop integrative models that consider links and feedbacks between water systems, and integrate across the biophysical and socioeconomic systems that influence urban water management.