geology to the streetscape: Evaluating urban underground resource potential

Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Tunnelling and Underground Space Technology journal homepage: www.elsevier.com/locate/tust

From hydro/geology to the streetscape: Evaluating urban underground resource potential Michael Robert Doyle Architecture and the Sciences of the City, Deep City Project, Laboratory for Environmental and Urban Economics (LEURE), Swiss Federal Institute of Technology, Lausanne (EPFL), Station 16, Lausanne CH-1015, Switzerland

a r t i c l e

i n f o

Article history: Received 6 July 2015 Received in revised form 11 January 2016 Accepted 12 January 2016 Available online xxxx Keywords: Underground resource potential Urban morphology Multi-criteria evaluation Spatial network analysis

a b s t r a c t Despite a persistent call for a greater recognition of the underground in urban planning practices, cities still tend to address underground resources only when the need arises. Historically, this has proven costly for cities that have neglected the potential synergies and conflicts between, for instance, urban aquifers and underground infrastructure systems or building foundations. For urban planning to remain in a paradigm of needs to resources risks rendering conflicts between urban underground activities irreversible and possible synergies unattainable. Researchers and practitioners from multiple disciplines argue for the many benefits of underground development—alternative renewable energy and drinking water sources, additional urban space and reusable geomaterials. Visualizing resource potential is a first step in raising awareness among planners of the capacities of the underground. Existing mapping methods tend to focus only on underground space development in contexts where the needs for the underground are already urgent and do not explicitly engage with the distribution of existing land uses. As an alternative to existing methods, this paper will present a procedure for mapping underground resource potential that incorporates four resources—space, groundwater, geothermal energy and geomaterials—developed by the Deep City project at the Swiss Federal Institute of Technology in Lausanne. San Antonio, Texas, a city with a complex relationship to an underground aquifer system but current little need and support for underground space, serves to illustrate the mapping method. Two future surface light rail and bus rapid transit lines, presented in recent planning reports, are examined in light of a latent but as yet untapped multi-resource underground potential. The paper concludes with a discussion of the applicability of the method to other cities and possible opportunities for improvement. Ó 2016 Elsevier Ltd. All rights reserved.

1. The underground as an integral part of the urban Since the first official geological investigations in the 19th century, urban planners, architects, geologists and engineers have called for a greater inclusion of the underground in urban planning. Eugène Hénard’s early 20th century vision for the Paris street of the future rethought the way urban infrastructure was situated vertically (Hénard, 1982). Édouard Utudjian’s proposal in the mid-20th century for an ‘underground urbanism’ offered the underground as a solution to the problems of congestion and pollution of rapidly urbanizing Western cities (Utudjian, 1952). As many of the contributions to this journal’s recent retrospective on its predecessor Underground Space attest, the relationship between the urban activities of the surface and those of the subsurface remains the source of an intense interdisciplinary focus. Urban areas experiencing population growth and investment in the

E-mail address: [email protected]

construction industry, particularly in Asia, are looking toward cities whose histories are marked by over a century of management of underground resources. At the same time, population growth in urban areas, particularly in the Europe and North America, is slow and situated on historically dispersed territories. Both cases beg the question: what is the underground potential of these areas and what possible alternative urban forms can it foster? Much of the current interest in the underground remains focused on the excavation of large volumes of space. Although space as an underground resource contributes in multiple ways to improving urban quality of life (International Tunnelling and Underground Space Association, 2012), the underground is also an important source of drinking water, geomaterials for construction or infill and geothermal energy. Experiences in cities like Paris, Mexico City and Tokyo demonstrate the problematic relationship excavation and construction practices in these cities have had with their urban aquifer, particularly where they provide a significant source of drinking water (Blunier, 2009). Unlike geomaterials

http://dx.doi.org/10.1016/j.tust.2016.01.021 0886-7798/Ó 2016 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

2

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

(the extraction of which has only a limited geographical impact), groundwater extraction and pollution have consequences that can reach a much larger scale (Morris et al., 2003). At least 122 cities worldwide with a population of greater than one million obtain at least 25% of their total water consumption from groundwater (Struckmeier and Richts, 2008). This provides both a challenge and an opportunity to develop diagnostic tools that can integrate multiple resources into the evaluation of underground potential, for both urban areas whose needs for additional space are pressing and for those where the management of other resources (geothermal or groundwater) may be of greater interest in the short term. For these latter cities, short term efforts can identify and reserve volumes of underground space for future excavation. Progress has been made in recent years in developing methods for evaluating and visualizing underground potential. The Helsinki Underground Master Plan presents existing and reserved areas for caverns and tunnels, indicating potential entry locations and depths of tunnels and spaces, overlaid by city blocks and the street network (Vähäaho, 2009). The map’s objective is to manage publicly owned underground real estate, in a manner synonymous with a land preservation plan. In Hong Kong, the cavern suitability maps currently under development present locations for potential cavern development classified according to geotechnical and land use characteristics of the surrounding context (Wallace et al., 2014). Researchers in China have tested a similar method on the city of Changzhou, and reported results at the parcel level (Peng et al., 2014). Whereas the cavern suitability map is specifically focused on the development potential of specific zones, the method tested on Changzhou provides a suitability score for different layers of the underground and for each parcel in the study area. This latter strategy is interesting because it provides information on resource potential at every location in the study area. Although information concerning groundwater or geomaterials factors into the classification of suitability, the maps presented above only focus on potential for the construction of underground space. Potential, understood as ‘‘latent qualities or abilities” (Oxford University Press, 2015), is presented in Helsinki as either built, planned or reserved, in Hong Kong on a 5-point scale from ‘not suitable’ to ‘high suitability’, and on a similar 4-point scale for the Chinese researchers. While these scales help communicate results quickly to decision-makers, the mapping method itself does not provide for competing but equally possible potentials—for instance, the parallel development of geothermal (for heating and cooling systems) or the risk of groundwater pollution. The evaluation of underground potential that takes into consideration multiple resources will need to account for the potential conflicts and synergies between uses. Furthermore, for all the importance placed on creating new urban spaces underground, the degree to which surface urban activities factor into underground space potential suggests that additional progress can still be made. Urban theory is increasingly arguing for a more relational approach to urbanism, meaning the potential of the existing urban fabric (its forms and functions) depends upon the relationships between land uses and the overall structure of the transport network, rather than as a set of static land uses dispersed over a passive transport infrastructure (Hillier, 2007). For example, the underground potential of a park is dependent not only on the geotechnical properties of the ground beneath it, but also on the strategic location of each surface entry point for potential clientele. Centrality is an important metric for underground space, because certain activities require being central to capture potential clientele (like commercial spaces) or being central for the easy distribution of goods or materials. A storage facility or water treatment center slated for development in a location outside of existing networks would require restructuring the

network of which they are a part (roads, distribution pipes, etc.). The evaluation of underground space potential can benefit from an integration of such network properties of the urban fabric. This paper presents a mapping method for evaluating the underground potential of an urban area for four resources (space, geothermal energy, groundwater and geomaterials) with a particular focus on the role of the surface urban morphology. This method is an extension of the one developed by the Deep City project at the Swiss Federal Institute of Technology in Lausanne (EPFL) since 2005 (Li et al., 2013b; Parriaux et al., 2004) and constitutes a portion the author’s doctoral work in architecture and urban planning, conducted under the supervision of a professor of geology and a professor of economics. The following section will present the scope, aims and main applications of the Deep City project, providing a brief overview of the mapping method and the contributions of the author. Then, a case study of San Antonio, Texas, will present each step of the calculation of the maps from data collection to interpretation in the context of the city’s current transportation plans. San Antonio is an interesting case in that it has a relatively complex relationship to its urban aquifer, which is the main source of drinking water for the region, but has no plans and very little political or economic support for developing underground spaces, either isolated (e.g. underground parking) or as part of a network system (e.g. public transport). The conclusion will return to the initial questions posed in the introduction and discuss future directions for the research.

2. Deep City: from resources to needs The Laboratory of Engineering and Environmental Geology (GEOLEP) at the EPFL received funding from the Swiss National Research Foundation in 2005 to launch the Deep City project, which sought to respond to the tendency for the urban underground to be managed on a project-by-project basis with often disastrous or risk-laden effects on other uses both on the surface and subsurface (Parriaux et al., 2010; Piguet et al., 2011). The underground is addressed as a strategy to increase urban compactness, increase walkability and the accessibility of urban activities. It is an alternative to building upwards, with the objective of concentrating urban activities above and below the street level at lower surface densities. For this reason, the project has been particularly interested in urban activities like food and retail, which are often located in the underground in pedestrian passages or metro stations or in underground conditions with only electric or zenithal lighting (e.g. malls, cinemas, theaters). The project is now hosted by the Laboratory for Environmental and Urban Economics (LEURE), following the retirement of the principal researcher who nevertheless remains an active contributor to Deep City. The mapping method (Fig. 1) has evolved significantly through case studies of Geneva, Switzerland, (Parriaux et al., 2010; Piguet et al., 2011) and Suzhou, China (Li et al., 2013a,b). First, geological data either as formations or boreholes is compiled in a GIS software package. The author’s contribution incorporates information on the distribution of the built environment (as buildings, parcels, or streets) into the GIS model. When possible, information concerning the buildings or parcels (resident population, jobs or activity) is included, but depends on the format in which such data is compiled. The second step classifies the geological formations into families of characteristics and evaluates the relative suitability of each family for each of the four resources, either using expert interviews or scientific evidence. The built environment is analyzed according to a series of centrality metrics, which in turn are given each a relative weight using pairwise judgments based on evidence in the scientific literature. The application of the method to San

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

Fig. 1. A summary of the Deep City methodology with specific contributions of the author’s dissertation research on the right hand side.

Antonio, Texas, in Section 3 will explain in more detail the evaluation process. The third step of the Deep City method entails two stages of aggregation. The first produces resource-specific maps—overall potential for groundwater, geothermal energy, space and geomaterials. The second produces a single potential map that captures the combined potentials. This double aggregation process builds on the mapping method in two ways. Both for Geneva and Suzhou, the resulting maps were oriented toward presenting the underground space development potential. In the case of Geneva, potential conflicts were managed through restriction maps delimiting problematic areas to avoid, but synergies were not explicitly expressed. By aggregating the four resources into a single map, a combined relative potential can serve as a reference point for comparison between locations in the analysis area, which can be interpreted using the single resource maps. Furthermore, the project proposes explicitly incorporating uncertainty in the aggregation process using Ordered Weighted Averaging. This was not pursued further after the Geneva case study, but would enable a calculation of risk that emphasizes either the strongest criteria or the weakest criteria during calculation. The San Antonio case study will illustrate the results and utility of this modification to the aggregation process.

3. San Antonio: a city of untapped underground potential

3

clay soils in the south require additional measures to stabilize building foundations (Ewing, 2008). The Balcones Fault Zone contains the Edwards Aquifer, which crosses Bexar County from southwest to northeast, and is the primary water supply source for many residents of San Antonio and neighboring communities. The limit of the artesian zone to the south is determined by the presence of subsurface saline water, that threatens to enter and contaminate local wells if not monitored properly (Thomas et al., 2012). The San Antonio 2020 visioning process conducted through a series of public forums and over 5000 surveys from 2010 to 2011 laid out the urban planning objectives for the city for the next ten years (Byrd et al., 2011), which include increasing walkability and the number of pedestrian-oriented neighborhoods as well as concentrating development in already urbanized areas. The transportation plan for 2035 developed by VIA Metro Transit (Jacobs Engineering, 2011) identifies seven main corridors for future public transportation projects, through analysis of traffic and origin–destination data in addition to population and demographic forecasts. These corridors seek to link the main activity hubs of the city, while decreasing congestion on existing arteries. The North/South Central Corridor line runs from the San Antonio International Airport (Fig. 4) near the northern segment of the inner 410 highway loop down through the downtown (including the central business district) and south to an important mixed use development project. The transportation plan recommends light rail for this corridor given its importance in connecting ‘‘various southern and northern neighborhoods with major employers throughout the corridor, especially the core” (Jacobs Engineering, 2011, p. 206). The East/West Central corridor, recommended for bus rapid transit or light rail crosses the downtown, connecting Fort Sam Houston to the east and Lackland Air Force Base to the west (the US Department of Defense being an important employer in the region), passes through the downtown and connects the North/South Central Corridor to the Westside Multimodal Transit Center (WMTC), a future regional transit hub. San Antonio’s relationship to its underground resources is currently dominated by the management of the aquifer system and transportation objectives do not take into consideration any underground potential for the proposed lines or stations. Investigating the underground resource potential of the city will therefore serve to highlight other complementary underground potentials, examine as an example the two lines and stations proposed by the transportation plan and question the role underground resources could play in current and future planning practices, even if significant underground development is not a viable option in the short term.

3.1. The urban and hydrogeological context of San Antonio San Antonio, Texas, is located in Bexar County in south central Texas and is the seventh largest city in the United States with an urban area population in 2014 of 1.98 million on an urbanized area estimated at approximately 1546 square kilometers. In comparison to cities of a similar population size, it is half the density of Hamburg (Germany), a third the density of Vienna (Austria) and less than a fourth the density of Minsk (Belarus) (Demographia, 2015). About 35% of the population lives within the highway 410 (Connally) loop and another 40% lives between 410 and the 1604 (Charles W. Anderson) loop (U.S. Census Bureau, 2010). Resident population densities are concentrated within the 410 loop and from the northwest to northeast areas between 410 and 1604 (Fig. 2). San Antonio is located on the Balcones Fault Zone, which explains the variety of geological strata exposed in the Bexar county area, with older limestone and granite in the Northwest and younger clay and sand in the southeast (Fig. 3). The ease of building on the older strata may explain the greater concentration of resident population densities to the north of the downtown: the

3.2. Data sources and evaluation Digital versions of the San Antonio Sheet of the Geologic Atlas of Texas (Bureau of Economic Geology, 1983) and a geological map of the substratum produced in the 1950s (Arnow, 1963) provided information on the location of geological formations at two depths, 0–15 m and 15–30 m. In order to make it easier for non-specialists to return to the geological information if necessary during the planning process, the formations were classified into a series of geological families called geotypes.1 For example, four geological formations characterized by a predominance of chalk or marl and possible karstic voids, which increases the likelihood of a higher water content, but also affects their stability and thermal 1 The classification method was developed by the GEOLEP laboratory at the EPFL as a strategy to express the principal properties of the formations for the purposes of urban planning without relaying genealogical or geographical information typical of existing classification methods (Parriaux and Turberg, 2007). They are not meant to replace geological maps or site-specific investigations, but rather to simplify the description of the geology and render it easier to work with for other disciplines.

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

4

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

Fig. 2. Distribution of population densities over the San Antonio Metropolitan Area. Data source: 2010 Census.

conductivity, correspond to the Alternating Marls and Limestones geotype (MC). This geotype is found in the northern part of the city and constitutes the areas where most of the urbanization has occurred. The main geological formations of San Antonio (fifteen occurred most frequently according to the atlas) correspond to seven geotypes (Table 1, Fig. 5): a series of younger superficial formations varying from six to fifteen meters deep (APA and APT) following the alluvial plains of the Medina and San Antonio rivers and the Salado and Leon creeks, two limestone types (C and CS) situated around the northern sector of 1604 near the municipalities of Hollywood Park and Shavano Park, a hard chalk and limestone (MC) type already mentioned and found on either side of the C and CS formations and two clay (MGR) and sand (GR) geo-types found mostly in the southern half of the city from Lackland and Fort Sam Houston military bases to the municipalities of Somerset and Elmendorf (Ewing, 2008). The next step evaluated the relative suitability of geotypes for construction of space, extraction of groundwater, heating and cooling from geothermal energy and reuse as geomaterials. Suitability is relative because there is no reliable way to measure absolute suitability for groups of geological formations on the scale at which the mapping process is operating. The resulting scale is a series of ratios, where the level of suitability can be anywhere from greater than zero to less than 100 (since compared elements are considered of some importance and no single element is of sole importance). The Deep City project generates this scale using the Analytic Hierarchy Process (AHP), a multi-criteria decisionmaking method that relies on simple pairwise comparisons using

linguistic qualifiers (‘less suitable’, ‘equally suitable’, ‘more suitable’) to generate a comparison matrix.2 Pairwise comparisons are carried out using scientific or expert (tacit) knowledge on a scale from one to nine (Saaty, 1990). For example, in calculating space suitability of the geotypes, each type is compared through pairwise comparisons to other types. The CS geotype was judged by the team (the author, a professor of geology and a professor of economics) to be more suitable for construction of space than the GR type due to the greater ease of construction in a limestone type than in sandstone. Due to possible holes in the karstic Edwards Limestone, the CS was considered to be slightly less suitable than the C type. These initial impressions were crossed with mineral resource locations (United States Geological Survey, 2014) and exploitation (Texas State Historical Association, Internet) as well as geothermal data (Blackwell and Richards, 2004). The relative importance of each geotype was further developed through discussions with geologists in San Antonio in the winter of 2014.3 Pairwise comparisons were performed for each of the geotypes according to the four resources on a scale from one to nine. The ratio scale is calculated by compiling the pairwise comparisons as an n by n matrix (where n is the number of elements), which

2 AHP is a method that is still popular in the literature and is being constantly debated and improved upon (Ishizaka and Labib, 2011). 3 Ideally, the pairwise comparisons are carried out by the local specialists themselves, something that was not organized for San Antonio, but has been tested by the author in later case studies.

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

5

Fig. 3. Geological map of San Antonio (Bexar County). Data source: USGS.

is normalized and solved to produce the principal eigenvector, whose consistency can then be evaluated mathematically (Saaty, 1990). For instance, the comparison matrix for suitability for space development (Table 2), after being normalized using the sum of the matrix, places the geotypes on a ratio scale in decreasing level of suitability: C (31%), MC (28%), CS (18%), APT (9%), MGR (6%), GR (5%) and APA (3%). The consistency of the matrix can be determined by taking the average of the eigenvalues of each geotype (basically a type’s overall relationship to each of the other geotypes), subtracting from it the number of elements (here, five) and dividing it by one less the number of elements (four). This number is then compared to the average consistency of around 500 randomly-generated matrices of the same dimension to verify that the perturbations do not deviate more than 10 percent (Saaty, 1990), which for the matrix here is 4 percent. Consistency evaluates how well the rule of transitivity is respected in the pairwise comparisons—for instance, whether the relationship between geotypes C and GR is proportional to the relationships between C and APT and APT and GR. The Deep City method seeks to identify the areas where it is most economically feasible to place activities underground, focusing thus far on commercial activities. Research conducted in Suzhou, China, suggests that around 30–60 percent of land acquisition and construction costs can be saved by placing a commercial or mixed-use project underground in favorable geological conditions where land value is high (Li et al., 2013b). In Switzerland, cost comparisons performed on different hypothetical surface and subsurface commercial building projects in different geological

conditions found that the most economically competitive alternatives (in terms of construction and operational costs) are those that are built both on the surface and underground in locations where land value is relatively high (Maire, 2011). Although further research is necessary, geological potential for the construction of underground spaces is therefore complimented by an urban one, understood as accessibility of the surface urban form. Commercial activities like food and retail tend to cluster near each other or within short distances of potential clientele, either jobs or residents, but they also benefit from being on shortest routes between adjacent activities (Porta et al., 2009; Sevtsuk, 2010). GIS software makes it relatively easy to estimate where current or future centers may be, using a street or transit network and a unit of analysis such as street segments, parcels or buildings. Urban potential for San Antonio was calculated using accessibility to resident population and commercial floor area of Bexar county parcels acquired as shapefiles from the Bexar County Appraisal Office’s 2013 dataset.4 Resident population per census tract was proportionally distributed over the parcels identified as residential within the tract boundary. Because employment data was not available at the parcel or tract level, commercial value per square meter served as a proxy for the relative pull of each commercial location. Commercial area for each parcel was available in the Appraisal Office’s dataset as gross built floor area. The degree of

4 Acquired through a written request made directly to the BCAD office: http:// www.bcad.org.

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

6

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

Fig. 4. Forecasted surface rail or bus rapid transit corridors in the inner 410 loop (adapted from Jacobs Engineering, 2011).

Table 1 Geotypes and corresponding geological formations for San Antonio (N.B. the acronyms correspond to the French nomenclature). Code

Geotype name

Formations in San Antonio (map label)

APA APT

Recent plain alluvium Terrace plain alluvium

CS MC

Siliceous limestone Alternating marls and limestones

MGR

Marls with some sandstones

GR C

Sandstone Limestone

Alluvium (Qal) Fluviale terrace deposits (Qt) Leona formation (Qle) Uvalde gravel (T-Qu) Edwards limestone (Ked) Pecan gap chalk (Kpg) Anacacho limestone (Kac) Austin chalk (Kau) Glen rose formation (Kgru-C-Kgrl) Wilcox group (Ewi) Midway group (Emi) Navarro group and marlbrook marl (Kknm) Del rio clay (Kdr) Carrizo sand (Ec) Buda limestone (Kbu)

centrality and accessibility on foot to resident population and commercial area, measured on the street network, was calculated for each parcel (n = 420,339). In ESRI’s ArcMap, each parcel was represented

by a single point placed approximately along the street network according to its address. The commercial centrality of each parcel is calculated as the sum of the distance to all commercial activities within 800 m along the street network. As the travel mode of interest here is walking, the street network dataset built in ArcGIS excluded highways and highway ramps inaccessible to pedestrians. Furthermore, closer activities were given a higher weight using an exponential distance decay for pedestrian travel (Handy and Niemeier, 1997), which considers that nearly ninety percent of trips completed on foot are within ten minutes (or about 800 m) from a point of departure. The resulting value is referred to in this paper as commercial destination potential. Centrality to residential population, or residential destination potential was calculated in the same way. In addition to being centrally located, parcels may lie along the shortest paths between origins and destinations. This property, known as betweenness, evaluates the potential of a particular location to be passed in front of by trips occurring between locations at a given distance.5 The resulting value, commercial path potential, is the sum of the number of trips that would pass by a parcel from all commercial origin and destination pairs at a given distance. Weighting the parcels according to resident population

5

For one of many mathematical descriptions, see Porta et al. (2009).

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

7

Fig. 5. San Antonio Geotypes, Texas (Bexar County).

Table 2 Comparison matrix for the suitability of each geotype for space construction (San Antonio, TX). Geotypes

APA

APT

CS

MC

MGR

GR

C

Priority vector (%)

APA APT CS MC MGR GR C

1.00 4.00 3.00 6.00 2.00 2.00 6.00

0.25 1.00 3.00 5.00 0.50 0.50 6.00

0.33 0.33 1.00 2.00 0.25 0.20 2.00

0.17 0.20 0.50 1.00 0.25 0.20 1.00

0.50 2.00 4.00 4.00 1.00 0.50 6.00

0.50 2.00 5.00 5.00 2.00 1.00 5.00

0.17 0.17 0.50 1.00 0.17 0.20 1.00

3 9 18 28 6 5 31

Consistency index = 4%.

approximates residential path potential, or the potential numbers of passers-by going from residential location to commercial location. Residential path and destination potential and commercial path and destination potential are the four metrics used to account for the potential of the current built environment to accommodate underground commercial spaces. Like geotypes, relative weights

are established using pairwise comparisons where experience is difficult to quantify. The rare studies that have investigated retail location behavior allow us to hypothesize that being situated along shortest paths (betweenness) will be between 1.3 and 1.4 times as important as being nearby (gravity).6 In the case of the relationship between the metric of proximity (gravity) and betweenness, preliminary analysis of Montreal’s Indoor City, conducted by the author (Doyle, Forthcoming), suggests that the relative importance of access to residents for the rental value of underground spaces may be dependent upon the existing distribution of housing units as well as zoning laws. A similar study conducted in Cambridge, Massachusetts, of surface food and retail stores argues that where housing is more evenly distributed, it may make less of a difference

6 Sevtsuk’s work in Cambridge, Massachusetts specifically looked at the influence of the betweenness, closeness (similar to gravity but without decay) of street segments on the probability that a location would be chosen by a food or retail owner (Sevtsuk, 2010). Porta and colleagues’ research on Bologna, Italy, investigated the same metrics looking this time at retail densities (Porta et al., 2009). The fact that their results are similar (1.3 and 1.4 respectively) is interesting. Of course, more research is needed to show that this is not simply coincidental.

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

8

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

for stores to be near residential population (i.e. demand maximizing favors existing commercial areas over residential ones). The research team therefore decided to qualify commercial as significantly more important than residential population in a comparison matrix (Table 3). In addition to the geology and built environment, other available data was included. Given the significance of groundwater for the San Antonio area, local well data acquired from the Texas Water Development Board permitted the interpolation of approximate percentage of groundwater saturation at 15 and 30 m depths. The ground saturation level may be advantageous for increasing thermal conductivity (for geothermal systems) or for local irrigation, but it may render excavation and construction activities more difficult or costly. The degree of conflict or synergy was considered to increase or decrease linearly and is represented as a standard scale normalized between 0 (0% groundwater saturation) to 1 (100% saturation). The 100 year flood plain surrounding lakes and rivers, which accompanied the Appraisal Office’s data was included as a possible source of conflict for excavation and construction. This is represented as a binary value, either in the flood zone, which we wish to avoid (value of 0) or outside of it, which is preferable (value of 1). The Edwards aquifer, as one of the main if not the sole drinking water source for much of Central Texas, including San Antonio, has been under different forms of protection since the 1950s (‘‘Regulatory History of the Edwards Aquifer,” n.d.). The Edwards Aquifer Protection Program regulates the types of activities and construction permits necessary in order to operate in three different zones. These zones include the recharge zone (where ‘‘stratigraphic units constituting the Edwards crop out”, 30 TAC § 213.3 (27)), the transition zone (where ‘‘faults, fractures and other geologic features present a possible avenue for recharge of surface water to the Edwards Aquifer” (30 TAC § 213.3(36)) and the contributing zone (where ‘‘runoff from precipitations flows downgradient to the recharge zone” (30 TAC § 213.22(2)). The chapter of the Texas Administrative Code that specifically addresses the Edwards Aquifer (30 TAC 213) suggests that the recharge and transition zones are the highest risk areas (with the recharge zone riskier than the transition zone) with moderate risk to be taken into account in the contributing zone. This appreciation of risk was referred to in producing a comparison matrix and generating relative weights for the three zones plus the unregulated area (Table 4). The priority given to being outside the protection is much higher than being in any of the three zones with a decreasing priority from the contributing to transition to recharge zones. The aggregation process is illustrated as a hierarchy in which each resource potential is to be calculated individually before being combined (Fig. 6). The geotypes, the built environment, the 100 year flood plain, the percent of groundwater at zero to 15 and 16–30 m deep, as well as the Edwards aquifer protection zones constitute the six main criteria for generating the potential maps for each resource. For example, space potential combines the excavation and construction potential of the geotypes (understood according to their general compactness, granularity and consistency), urban centralities and the flood and aquifer protection zones. Each criterion’s relative importance (presented in the diagram as a percentage) is calculated through an additional series of pairwise comparisons, asking, for example, whether it is preferable to be outside the flood plain or outside the aquifer protection are when choosing a site for underground space development. In the case of San Antonio, we are specifically considering the potential for commercial space development, which is why the centrality metrics have such a high importance. For other land use types, the relative weights might vary for the built environment criterion. For San Antonio, the values are calculated at the two different depths in ArcGIS on cells in a 50 by 50-m vector grid. They are combined

Table 3 Comparison matrix for urban potential measures.

Commercial potential Residential potential

Commercial

Residential

Metric

Priority vector (%)

1.00

7.00

0.14

1.00

Path (1.3) Destination Path (1.3) Destination

50 38 7 5

N.B. Consistency is not calculated for a 2  2 matrix.

Table 4 Comparison matrix for the Edwards Protection Zones.

Outside protection area Contributing Transition zone Recharge zone

Outside

Contributing

Transition

Recharge

Priority vector (%)

1.00

3.00

5.00

9.00

55

0.33 0.20

1.00 0.33

3.00 1.00

7.00 6.00

26 14

0.11

0.14

0.17

1.00

5

Consistency index = 0% (fully consistent).

after being normalized so that each criteria sits between zero and one.7 The individual resource potential maps indicate the areas that score higher for each resource. At a depth of 15 m (Fig. 7), the part of San Antonio lying within the 410 ring highway has several zones in the northern half where space potential is relatively high and where geomaterial and groundwater potentials overlap. Moving south however into the area that includes the CBD, the potential for geothermal energy systems and groundwater extraction is higher. Space remains relatively high (in the upper 50th percentile), but only along particular streets where the commercial and residential path potentials are high. The influence of the percent of groundwater in the first 15 m is evident in the geomaterial and groundwater potential maps where two zones of low potential (geomaterial) and medium potential (groundwater) are visible in the south western and eastern quadrants of the inner 410 loop. There is also an obvious and abrupt change in potential at the fault lines separating the marl/sandstone (MGR) and marl/limestone (MC) geotypes found in the northern section of the study area. Presenting the four resources separately risks their being considered in isolation when referred to by the planning process—one of the main problematics the Deep City method seeks to overcome. The classic solution, borrowing from AHP, would be to combine the four maps according to a series of criteria weights. This is best illustrated by returning to the transit lines mentioned in Section 3.1 (Fig. 8). Although the envisioned transportation scenarios are entirely aboveground, it is interesting to examine the underground potentials of the two proposed stations. When combining the potential maps with an equal priority for each (consulting local decision-makers for this step was beyond the scope of the study), the Woodlawn station clearly has a relatively high combined potential (score of 81 out of 100), while the WMTC is situated in a zone of slightly higher than medium potential (combined score of 62). The separate scores for each resource potential (Fig. 9) suggest that Woodlawn provides an opportunity for underground space construction, contains potentially interesting geomaterials (mostly limestone) and may support local groundwater extraction (for drinking or irrigating). Geothermal 7 A more detailed explanation of the aggregation process using the Analytic Hierarchy Process in GIS can be found in Boroushaki and Malczewski (2008).

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

9

Fig. 6. Aggregation procedure (from top to bottom) for mapping the four resource potentials and then combining them using order-weighted averaging.

Fig. 7. Individual resource potential maps for the urban area within the 410 loop (depth 0–15 m).

energy systems are not impossible, but may not be practical here, as the low score suggests. The WMTC does have some opportunity for underground space development, although it is situated over more clay-heavy ground, which reduces the potential for reuse of local geomaterials. Groundwater extraction or protection (the site is situated over local aquifers according to well data) and geothermal energy systems may be of higher priority. The centrality metrics reveal the potential of the sites to support underground commercial activities. The Woodlawn station is located at the intersection of two four-lane roads, San Pedro Avenue, which is north–south and mostly commercial, and Woodlawn avenue, a mostly residential east–west street comprised of twostory single-family homes. The parcels around the station score in the top 90th to 100th percentile for accessibility to commercial property values and residential population at a distance of about a 10-min walk (800 m) (Fig. 10). Although with such low built

density the construction of underground space is not yet a viable economic alternative to building on the surface, it remains a site of interest if the recommended transit line could be placed underground in the future. In contrast to the Woodlawn station that is a local transit stop, the WMTC is a regional hub, located in an area just outside the western edge of the Central Business District next to a rail line. While it is not far from high-value commercial areas, it is less likely to be a chance stop for pedestrians on their way from one commercial location to another. From a standpoint of potential clientele, the WMTC would have to count on passengers being the main clientele to support underground commercial spaces and even then it may be more difficult to argue given its less favorable geological conditions. Of course, the mapping method proposed here is meant to give a general overview and not to replace local investigations of soil conditions, the built environment or economic capacity. The maps

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

10

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

Fig. 8. Aggregate resource potential at 0–15 m deep overlaid by two future transit lines.

Fig. 9. Comparison of resource potentials for two future transit stations (San Antonio, TX).

Fig. 10. Comparison of urban centrality metrics for two future transit stations (San Antonio, TX).

are intended to help inform a long term planning process that must deal with uncertainty, not only in the information upon which decisions are being made (data and its analyses), but also and most importantly in the number of criteria that must be satisfied in the

evaluation of potential. The classic AHP calculation considers each criterion equally, meaning, for the mapping method proposed here, that aggregate potential is the sum of that of each resource (Fig. 8). In order to capture the varying attitudes of decision-makers and

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

11

Fig. 11. Aggregate maps with emphasis on the strongest (a) and weakest (b) resource potentials.

stakeholders towards the relative importance of the aggregated criteria, researchers have proposed incorporating order weights into the aggregation process (Boroushaki and Malczewski, 2008; Yager and Kelman, 1999). Concretely, this means that for each location the mapping process reorders from highest to lowest the resource potential scores and multiplies them by an order vector that either places all the weight on the first (highest) or last (lowest) resource. This process, called Ordered Weighted Averaging (OWA), seeks to account for varying degrees of participation of criteria in the decision-making process. OWA allows the calculation of aggregate potential to present on a one map the locations where there are a few, but not all the same, high resource potentials. A second map can highlight locations where all four resources are similarly strong. For instance, Woodlawn station and the WMTC’s potential scores of 81 and 62, respectively, are the result of simply adding the four resource potentials together and normalizing their result (synonymous with a dummy order vector of [0.250, 0.250, 0.250, 0.250]8). If potential is calculated with only an appreciation for its highest potentials (using an order vector of [0.764, 0.182, 0.043, 0.010]), then Woodlawn’s score moves to 98 and the WMTC’s increases by 10 points to 72. Mapping these scores for the area around the 410 loop (Fig. 11a) reveals those areas where there are one or several very high resource potentials. Areas of high geomaterial and geothermal potential emerge in the southern and western parts of the study area and the potential for space and groundwater is highest in the northern section. Using the reverse order vector [0.010, 0.043, 0.182, 0.764], which emphasizes the weakest resource potentials, Woodlawn and the WMTC have nearly the same score (51 and 48 respectively), because their weakest potentials (geothermal and geomaterials) are nearly equal (as illustrated in Fig. 10). The regional map of these scores highlights those areas where the resource potentials are collectively high (Fig. 11b). The highest

8 For a mathematical description and examples of using OWA, please see Yager and Kelman (1999).

potentials (red zones) with this calculation are outside the 410 loop, over the aquifer recharge zone in the northern part of Bexar County. For the planning process, the analysis produces a series of maps that tell a different, but complementary story, which can be revealed by stepping backwards through the aggregation process (a reversal of Fig. 6). The three aggregation maps represent the mid-point and two extremes of a particular attitude toward the priority levels of the four underground resources without limiting the map to only one resource. Emphasizing the weakest resource potentials reveals the zones where potential conflicts and synergies are likely to be the most critical and where their interactions have to be planned for more carefully. Emphasizing the strongest potentials reveals areas where one or two resource potentials are higher than the rest and where targeted efforts for the management of a particular resource could look first in priority. Furthermore, specific weights could be given to each resource so that the OWA calculation multiplies the resource scores by criteria weights in addition to the order weights. The maps can be crossverified with those that represent each individual resource for comparison and, if necessary, the planning team could return to the individual criteria. The metrics for evaluating the built environment should always also be examined separately, because the results will be more familiar to planners or architects. 4. Generalizability of the method and further research As a method that combines both analysis and synthesis, the success of the mapping procedure relies on being adaptable to the specific conditions of each context. The analytical phase that compiles spatial data on the geology and natural and built environments can remain relatively stable from case to case, keeping in mind that the criteria that matter for one city may be different or absent in another. Not every urban area has collected data on nor modeled their geology or the built environment in the same way. The method of reducing the geological formations to families of geotypes can be easily applied elsewhere by geologists and new geotypes can always be created for contexts with unique geological

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

12

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx

conditions. Local planners, of course, may be comfortable using current local geological categorizations. Similarly, the spatial analysis of the urban form also can be easily applied to different locations, because centralities form locally. Only the specific relationships between certain activities (commercial, residential, logistic, etc.) will vary slightly—either evidence from elsewhere can serve as rules of thumb or the planning team can conduct local investigations of urban spatial structure.9 Regardless of the diversity of data, the minimal requirements for map production remain information on the geological formations and some indicator of the built environment (either streets, buildings or parcels). The synthetic phase, in which the mapping layers begin to be combined, follows a trajectory whose outcomes are placespecific. Even if the relationships between the criteria may in some cases be similar, their quantification cannot be considered an absolute, particularly where they rely on expert knowledge gathered through local experience. Further work is necessary to improve the means by which the expert knowledge is collected and quantified. The Analytic Hierarchy Process is a useful tool for quantifying intangibles, but the case study of San Antonio did not fully integrate local experts into the pairwise comparisons. Work by the author in 2015 on the city of Hong Kong, China, integrated experts more fully into the AHP process and an additional case study in 2016 in Dakar, Senegal, will apply a strategy to render the consultation process more robust and engage local experts even further. For instance, the Delphi method proposes a series of phases for the definition of criteria and pairwise comparisons (Novakowski and Wellar, 2009). Further work will also be conducted on the way the aggregate maps are presented. The addition of Ordered Weighted Averaging imports tools from fuzzy set theory and provides the planning team with maps that highlight zones where the relationships between the resources may be the most sensitive and the most likely to result in conflicts or synergies. More specific maps may be useful, particularly once the planning team develops specific questions for current or future projects. The aggregation maps still require quite a bit of interpretation by the planning team and further work can illustrate how aggregation maps can be transformed into maps of specific conflicts or synergies. Currently, the potential of each location is calculated relative to the entire study area, but calculating the potentials at a ‘neighborhood’ level would establish a local metric for potential. This kind of AHP-OWA calculation has not yet been conducted on a GIS data set including the geological scale and will raise the question of the scale of such a territorial division. This will be the object of future work by the author. This paper presented a procedure for mapping underground resource potential, taking the city of San Antonio, Texas, as a case study. Starting from a method developed by the Deep City project at the Swiss Federal Institute of Technology in Lausanne, the author’s doctoral work seeks to advance the method to a regional scale, to incorporate spatial analytics of the surface built environment, to test and render more rigorous the quantification of expert, tacit, knowledge and to generate a specific series of maps that allow a planning team to better account for the region’s underground resource potential. Contrary to current practices and other mapping methods, the one presented here pays particular attention to four underground resources rather than only to space and is intended as much for urban areas with little to no envisioned underground space development as for those where the need is more urgent. The inclusion of the built environment also provides an important bridge between surface and subsurface activities and between the domains of geologists, urban planners and architects. 9 Further research on the relationship between underground spaces and accessibility is necessary and is being carried out by the author in a parallel study of Montreal’s Indoor City (Doyle, Forthcoming).

The method reflects the Deep City project’s commitment to a paradigm of ‘resources to needs’—conceiving of future urban projects in response to rather than exploitative of underground resource potential. Whether the underground is an interesting option today, tomorrow or never, an urban area should be able to make an informed, calculated and responsible decision. Acknowledgments An understanding of San Antonio and its underground would not have been possible without the generosity of Dr. Thomas E. Ewing, geological consultant, John Waugh, hydrogeologist at the San Antonio Water Services and Razi Hosseini, assistant director for the Transportation and Capital Improvements department of the City of San Antonio. The author would also like to thank the three anonymous reviewers whose generous comments were greatly appreciated. References Arnow, T., 1963. Ground-Water Geology of Bexar County, Texas (Geological Survey Water-Supply Paper No. 1588). US Government Printing Office, Washington, D. C.. Blackwell, D.D., Richards, M., 2004. Geothermal Map of North America. American Assoc. Petroleum Geologist (AAPG). . Blunier, P., 2009. Méthodologie de gestion durable des ressources du sous-sol urbain (Doctoral Dissertation). EPFL, Lausanne. . Boroushaki, S., Malczewski, J., 2008. Implementing an extension of the analytical hierarchy process using ordered weighted averaging operators with fuzzy quantifiers in ArcGIS. Comput. Geosci. 34 (4), 399–410. http://dx.doi.org/ 10.1016/j.cageo.2007.04.003. Bureau of Economic Geology, 1983. Geologic Atlas of Texas, San Antonio Sheet. Bureau of Economic Geology, Austin. Byrd, D., Rodriguez, S.M., Weston, G., 2011. SA2020. San Antonio, TX. Demographia, 2015. Demographia World Urban Areas, 11th Annual ed. Demographia, Belleville, Illinois, . Doyle, M., Forthcoming. Mapping Underground Resource Potential: From Hydrogeology to Urban Form (Preliminary Title) (PhD in Architecture and the Sciences of the City). École polytechnique fédérale de Lausanne, Lausanne, Switzerland. Ewing, T.E., 2008. Landscapes, Water and Man: Geology and History in the San Antonio area of Texas. South Texas Geological Society, San Antonio, Tex.. Handy, S.L., Niemeier, D.A., 1997. Measuring accessibility: an exploration of issues and alternatives. Environ. Plann. A 29 (7), 1175–1194. http://dx.doi.org/ 10.1068/a291175. Hénard, E., 1982. Etudes sur les transformations de Paris, et autres écrits sur l’urbanisme. L’Equerre, Paris. Hillier, B., 2007. Space is the Machine: A Configurational Theory of Architecture (Electronic Edition, orig. 1996). Cambridge University Press, Cambridge. International Tunnelling and Underground Space Association, 2012. Report on Underground Solutions for Urban Problems (No. 011). Ishizaka, A., Labib, A., 2011. Review of the main developments in the analytic hierarchy process. Expert Syst. Appl. http://dx.doi.org/10.1016/j. eswa.2011.04.143. Jacobs Engineering, 2011. 2035 Long Range Comprehensive Transportation Plan. VIA Metropolitan Transit, San Antonio, TX. Li, H., Parriaux, A., Thalmann, P., Li, X., 2013a. An integrated planning concept for the emerging underground urbanism: Deep City Method Part 1 concept, process and application. Tunnell. Underground Space Technol. 38, 559–568. http://dx. doi.org/10.1016/j.tust.2013.04.010. Li, H., Li, X., Parriaux, A., Thalmann, P., 2013b. An integrated planning concept for the emerging underground urbanism: Deep City Method Part 2 case study for resource supply and project valuation. Tunnell. Underground Space Technol. 38, 569–580. http://dx.doi.org/10.1016/j.tust.2013.04.009. Maire, P., 2011. Étude multidisciplinaire d’un développement durable du sous-sol urbain. Aspects socio-économiques, juridiques et de politique urbaine (Doctoral Dissertation). Swiss Federal Institute of Technology Lausanne (EPFL), Lausanne. Morris, B.L., Lawrence, A.R.L., Chilton, P.J.C., Adams, B., Calow, R.C., Klinck, B.A., 2003. Groundwater and its Susceptiblity to Degration a Global Assessment of the Problem and Options for Management. Division of Early Warning and Assessment, United Nations Environment Program, Nairobi, Kenya, . Novakowski, N., Wellar, B., 2009. Using the Delphi technique in normative planning research: methodological design considerations. Environ. Plann. A. http://dx. doi.org/10.1068/a39267. Oxford University Press, 2015. Oxford English Dictionary. .

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021

M.R. Doyle / Tunnelling and Underground Space Technology xxx (2016) xxx–xxx Parriaux, A., Blunier, P., Maire, P., Dekkil, G., Tacher, L., 2010. Projet Deep City: Ressources Du Sous-Sol Et Développement Durable Des Espaces Urbains. Lausanne: vdf Hochschulverlag AG an der ETH Zürich. Parriaux, A., Tacher, L., Joliquin, P., 2004. The hidden side of cities—towards threedimensional land planning. Energy Build. 36 (4), 335–341. http://dx.doi.org/ 10.1016/j.enbuild.2004.01.026. Parriaux, A., Turberg, P., 2007. Les géotypes pour une représentation géologique du territoire. Tracés 133 (15–16), 11–17. http://dx.doi.org/10.5169/seals-99593. Peng, J., Wang, Y., Peng, F., 2014. Evaluation of underground space resource for urban master plan in Chanzhou City. In: Underground Space: Planning, Administration and Design Challenges. Seoul, South Korea. pp. 30–35. Piguet, P., Blunier, P., Lepage, L., Mayer, A., Ouzilou, O., 2011. A new energy and natural resources investigation method: Geneva case studies. Cities 28 (6), 567– 575. http://dx.doi.org/10.1016/j.cities.2011.06.002. Porta, S., Latora, V., Wang, F., Strano, E., Cardillo, A., Scellato, S., Iacovelli, V., Messora, R., 2009. Street centrality and densities of retail and services in Bologna, Italy. Environ. Plann. B: Plann. Des. 36 (3), 450–465. http://dx.doi.org/10.1068/ b34098. Regulatory History of the Edwards Aquifer. (n.d.). July 1, 2015, . Saaty, T.L., 1990. How to make a decision: the analytic hierarchy process. Eur. J. Operational Res. 48 (1), 9–26. http://dx.doi.org/10.1016/0377-2217(90)90057-I. Sevtsuk, A., 2010. Path and place: a study of urban geometry and retail activity in Cambridge and Somerville, MA (Doctoral Dissertation). Massachusetts Institute of Technology, Dept of Urban Studies and Planning, Cambridge, MA.

13

Struckmeier, W., Richts, A., 2008. Groundwater Resources of the World. Electronic Map. BGR/UNESCO, Hannover. Texas State Historical Association, Internet. Texas Almanac: Nonpetroleum Materials. Texas State Historical Association. . Thomas, J.V., Stanton, G.P., Lambert, R.B., 2012. Borehole Geophysical, Fluid, and Hydraulic Properties within and Surrounding the Freshwater/Saline–Water Transition Zone, San Antonio Segment of the Edwards Aquiferr, South-Central Texas, 2010–2011. Texas Water Science Center, Austin, Texas. United States Geological Survey, 2014. Mineral Resources on Line Spatial Data. Internet Map, USGS. . U.S. Census Bureau, 2010. 2010 Demographic Census Profile 1 – Shapefile Format. . Utudjian, E., 1952. L’urbanisme Souterrain. Presses universitaires de France, Paris. Vähäaho, I., 2009. Underground Masterplan of Helsinki: A City Growing Inside Bedrock (Extract from the Underground Masterplan of Helsinki). Helsinki, City of Helsinki. Wallace, M., Roberts, K.J., Lau, V., 2014. A geographic information systems approach for regional cavern suitability mapping for Hong Kong. In: Underground Space: Planning, Administration and Design Challenges. Seoul. Yager, R.R., Kelman, A., 1999. An extension of the analytical hierarchy process using OWA operators. J. Intelligent Fuzzy Syst. 7, 401–417.

Please cite this article in press as: Doyle, M.R. From hydro/geology to the streetscape: Evaluating urban underground resource potential. Tunnel. Underg. Space Technol. (2016), http://dx.doi.org/10.1016/j.tust.2016.01.021