OR/17/007 Modelling urban groundwater and geothermal resources - (Lothar Moosmann and Nikolaus Classen)

Bonsor, H C, Dahlqvist, P, Moosmann, L, Classen, N, Epting, J, Huggenberger, P, Garica-Gil, A, Janźa, M, Laursen, G, Stuurman, R and Gogu, C R. 2017. Groundwater, geothermal modelling and monitoring at city-scale: reviewing European practice and knowledge exchange. British Geological Survey Internal Report, OR/17/007.

Key words: city-scale modelling; optimisation; drivers; cost-benefit;
data availability; modelling for regulation and management

Introduction

This section provides a review of examples of good practice concerning development and application of groundwater and geothermal modelling to help support sustainable utilisation and planning of the resources at city-scales. The review also includes discussion to costs and benefits of groundwater and geothermal modelling over investment in monitoring data and infrastructure in cities. Key future requirements in technical development and functionality of urban groundwater and geothermal models are also reviewed.

City needs for groundwater and geothermal modelling

Although groundwater is often referred to as a relatively stable and better-protected water resource compared to surface water (Volker and Henry 1988; Zektser and Everett 2004), the protection and management of groundwater is a complex task, even more so in urban areas where there are intense competing demands of the resource and surrounding subsurface. An understanding of natural hydrological systems and the processes of pollution transport form the basis for efficient water management is essential for appropriate groundwater management and planning in cities, as well as some understanding of the influence of urban infrastructure to the resource.

‘Modelling is a helpful and essential tool to describe and understand especially
the groundwater and geothermal processes on the City-scale today and also
to predict future scenarios with different use of the resources.’

The ability of city municipalities and stakeholders to simulate the natural system and scenarios of the impacts of proposed changes to the groundwater enhances the adaptive capacity of water management. A key use of groundwater modelling within urban areas is to assist understand develop these different areas of understanding, and to help delineate what groundwater protection areas should be within the more complex and heterogeneous urban environment. Other general drivers for the modelling include:

• Groundwater management (groundwater resources, groundwater use, catchment areas, groundwater-infrastructure interaction, remediation)
• Groundwater protection (contaminant transport, groundwater salinization)
• Thermal management (interaction of geothermal system, groundwater quality issues)
• Underground city planning (infrastructure development)

Different stakeholder prioritise some of the above drivers over others:

• Water suppliers (managing groundwater use, catchment areas, extraction concepts)
• Authorities (protection and management of groundwater, groundwater resources, shallow geothermal resources, sustainable development)
• Private (groundwater use, geothermal systems)
• City operators (water, transportation)

City municipality priorities are:

• How much groundwater can be sustainability abstracted without significant negative impacts to other city resources or infrastructure? (e.g. heritage building foundations, basement infrastructure, and soil and water chemistry?
• Does shallow groundwater influence flooding within the city, and if so in what areas?
• Where can infiltration drainage schemes be used sustainably in the city? And what at volume of water can be infiltrated?
• What are the impact of infiltration schemes to groundwater levels and flooding?
• What is the likely impact of proposed SGE schemes?
• Do new private abstraction schemes impact industry or public water supply schemes?
• Models need to accessible and understandable to city planning, so that the knowledge and data within them can be used to inform the planning process.

Good practices

There is no one good practice for the development of city-scale groundwater modelling — different modelling requirements are determined by: the end use of the model (which can be very different in different cities, according the underlying driver for development of the model); and, the availability of input data.

For example, a groundwater model being develop to help improve the conceptual understanding of the urban groundwater resource characteristics will have very different resolution and input data requirements than, a model being developed to inform a management of public water supply abstraction and defining well field protection areas. Groundwater models aimed at stimulating processes and problems of groundwater flooding will in turn have yet another modelling framework, and different set of priorities for the data and processes modelled and output requirements.

Good practice in model development is very much dependent on the objectives of the modelling and also the availability of input data.

Key guiding principles of good practice exist though. For example, it is essential that a model is developed for a clear objective — otherwise the data or understanding developed from the model will not be of an appropriate scale or resolution to inform the driver or objective behind the model.

Good practice in the construction and development of a model should include each of the following worksteps (Figure 4.1):

• Identify the clear objectives for modelling, as basis to design the conceptual or numerical model (communication between the contracting authority and the contractor and the interests of third parties or authorities).
• State the minimum key input data required and minimum amount/quality of these data (analysis of the available data, the framework for additional data collection and statements on how best to model quality that can be achieved with it).
• Identification of knowledge gaps and requirements for filling them (extension of groundwater and surface water monitoring systems; requirements on the documentation of groundwater use data and subsurface structures; definition of field investigations to study specific processes).
• Show cost-benefit of numerical modelling (When is modelling more cost-efficient or indispensable compared to analytical analysis of monitoring data or solely conceptual modelling approaches).
• Identify valid modelling assumptions (general acceptance in the model assumptions). Defining the requirements for the suitable setup of model geometries (geology and subsurface structures) as well as necessary data for parameterizing the various natural and anthropogenic boundary conditions.
• An agreed conceptual city-scale model by all users. This avoids different users making decisions based on different conceptual and numerical groundwater models with different input data.

Guiding points of good practice for model accuracy and resolution of input data

It is good practice to use input data and parameters of a scale and resolution only appropriate to the modelling objective. Some includes of input data which should be represented within groundwater modelling are listed. Not all groundwater models would have to include all of these data to be ‘good practice’, depending on the modelling objectives.

• Geology (drilling-cross section — surface map — 3D structure model; Geophysical data)
• Hydraulic parameters/pumping tests (hydraulic conductivity, storativity, permeability)
• Groundwater level data (hydraulic head, river and groundwater hydrographs, piezometric surface)
• Groundwater quality (pollution, salinity, temperature)
• Groundwater recharge (including water supply system losses), Interaction with surface waters and atmosphere (steady-state/transient)
• Infrastructure (water supply network, sewer system, tunnel, underground transportation network, subsurface building structures, permanent and temporary dewatering systems, district heating)
• Consumptive use of groundwater (pumping rates, demands)
• Geothermal use (pumping rates, production-injection temperatures, heat power)
• Transport parameters: Tracer test (solute) & thermal response test (heat) (porosity, dispersivities, molecular diffusion)
• Groundwater supply systems (pumping rates, exploitation schemes and schedules)

Input data required should be organized and made available in a database and/or GIS-system. At low data availability the limited data-sets should be supplemented by extensive data research. The available data widely open in a database improves the acceptance for data delivery. Data coming from different sources should be filtered and homogenized. Under cost efficiency — so much as necessary, so little as possible — is a useful approach in ensuring an appropriate balance in input data is met.

Development of an agreed city-scale groundwater model: good practice

Development of an agreed city-scale model which is coordinated with all users (e.g. Cooperation between authorities and water companies) is also a key point of good practice where multiple users have the same needs or objectives for using a groundwater model (e.g. managing flooding and infiltration drainage in a city) — otherwise individual users develop individual models which might parametrize the groundwater system slightly differently and reach different model outputs. Having a single agreed model between the users in a city avoid different user decisions based on different conceptual and numerical groundwater models with different inputs.

Workflows of good practice

Good practice in the design of groundwater modelling should include each of the following work-steps — in dependence of the available data and the required accuracy of the model outputs.

Figure 4.1    A generalised good practice workflows for high and low data availability.

Case studies

Different good practice strategies are determined by different availability of input data in cities. The examples of groundwater models developed for Glasgow and Hamburg are used to highlight the different good practice steps in each case, and these are separated into ‘good practice’ steps and ‘best effort’ workflow examples.

Hamburg (Germany) — Good practice developing a standardised city groundwater model between multiple stakeholders.

The case study is an example of how groundwater modelling can be used effectively to support city planning and management of subsurface resources. It is also a benchmark example of the interaction and cooperation required between multiple users of a model (e.g. city authorities and the water supplier) to develop an agreed city-scale model for all users.

The groundwater model was developed by the city’s public Water Supply Company (Hamburg Wasser) in collaboration with the State Geological Survey (BSU) in the city municipality. The aquifer framework used within the groundwater model is taken from the cities 3D geological model, developed and held by BSU for third party use, and the aquifer parameterization in the groundwater model was developed using data held by both BSU and Hamburg Wasser. The geological model, which was used to inform the geometry and stratigraphy of the aquifer is based on approx. 200 000 boreholes alone, and the 3D numerical groundwater modelling with SPRING (delta-h) software parameterized this geological framework with the cities extensive groundwater monitoring data. However, the outstanding elements of good practice to be taken from the case study are the integration of both public and private datasets within the city to develop a coherent and agreed understanding of the aquifer properties in the city, and how the regional groundwater system should be demarcated to develop appropriate and management groundwater catchment protection areas. Had the groundwater model been developed by either the State Geological Survey, or the Utility company in isolation, the model was not be treated as an accepted or agreed groundwater model by the other, and groundwater management decisions, or city development decisions would be based on separate groundwater models and understanding.

The groundwater model is used to help approve new groundwater abstraction schemes in the city, and also manage the protection of the public supply well fields. The improved understanding of the cities groundwater resource which has been gained from the model, has also underpinned an optimization program of groundwater and surface water monitoring network in the city, which has now been streamlined to 650 monitoring sites for quality and quantity. The Model will be used in the long-term for groundwater protection issues, and as a decision making tool for the groundwater use — public and domestic.

Glasgow (United Kingdom) — An example of good practice in developing an urban groundwater with low data availability to improve conceptual understanding.

The case study is an example of how a robust conceptual groundwater model can be developed to help improve general understanding of urban groundwater resource in the absence of significant aquifer properties data (e.g. urban-scale groundwater flow patterns, general depth to groundwater and characteristics).

Glasgow is underlain by complex quaternary deposits 30 m thick, up to >50 m thick across the city. These form a complex shallow aquifer, wherein higher permeability sand and gravel dominated units are laterally discontinuous over 100s meters and are of significant variation in thickness where present. The urban groundwater resource held within these deposits is vulnerable to contamination as a result of the shallow depth to groundwater (<5 m) within the city centre, and the presence of multiple sources of potential pollution (e.g. from shallow mine shafts underneath the city centre, heavy metal soil contamination and buried waste). To be able to manage and protect the resource, as well as meet future legislative requirements of the EU Water Framework Directive, the city municipality and national environmental regulator (SEPA) require a better conceptual understanding of the general characteristics of the urban groundwater resource within the superficial aquifer (e.g. general regional groundwater flow pattern across the city, depth to water table, and seasonal variability of the resource).

There are, however, few observed groundwater levels or aquifer properties data to parameterize an urban groundwater model. The city forms a good case study in this situation where a ‘best effort’ approach can be taken to develop a lower resolution groundwater model, using the input data available to develop a ‘conceptual model’ of the groundwater regime in the city, specifically focused to representing groundwater-levels and recharge — key parameters needed to be understood better to manage and protect the resource in Glasgow. The complex geometry of the superficial aquifer were modeled using the detailed 3D geological-framework modeled from the city’s comprehensive geological model, based on the information of 50 000 boreholes. Aquifer properties were approximated using the few aquifer properties data available, and from more extensive datasets from similar superficial aquifers elsewhere in Scotland. The model was calibrated to groundwater-level observations available (most one-off observations, rather than time series) and river levels (time series).

Whilst the model is not of sufficient resolution to delineate groundwater catchment areas, or protection zones around any potential future public water abstraction in the city — the model is robust enough to enable stakeholder to develop an improved conceptual understanding of the urban groundwater resource to ensure appropriate use and management of the resource at a city-scale (e.g. to help inform where infiltration drainage is inappropriate), and to help inform where future monitoring of the groundwater resource should be focused. With more local data (groundwater level network) and validation of the conceptual groundwater model the information quality provided by the model could be significantly improved in the future. By using the 3D geological model to inform the aquifer geometry and properties modeled ensured the groundwater model was developed from an existing coherent understanding of the cities geology and aquifer, which has been developed by the British Geological Survey, City Council, and several other public and private stakeholders in the city.

Integration of groundwater models to resource management and city planning: good practice

As well as Hamburg, the cities of Ljubljana and Bucharest provide good practice examples of developing groundwater models which are appropriate to city planning needs. These are discussed in turn below, with the different strategies necessitated by the availability of groundwater data in the city highlighted.

Ljubljana (Slovenia) — good practice example of translation and use of the cities groundwater monitoring network data within a decision support tool to support management and protection of the urban resource.

Following a potentially significant contamination event to the city’s urban groundwater resource, which supplies drinking water to the city, the city municipality and geological survey have developed a new decision support system (DSS) for the management of the aquifer. This forms a good practice case study in two respects: 1) that the DDS is based on time series groundwater observation data from the cities monitoring network, and 2) providing a clear example of how groundwater data and knowledge can be translated to support decision making for management of the subsurface and cities resilience.

The groundwater model is based on the MIKE SHE/MIKE 11 modeling software, and simulates the groundwater dynamics and transport of pollutants in the aquifer based on an integrated groundwater/surface water model. A user-friendly graphical interface enables water managers to utilize the database, numerical modeling techniques and expert knowledge, and thus gives them fast and easy access to supporting information for mitigating groundwater pollution.

Bucharest (Romania) — one of the few cities to have developed a city-scale groundwater recharge and flow model which has integrated subsurface urban infrastructure, to better understand the impacts and interaction of the infrastructure to the urban groundwater resource.

The city of Bucharest forms a best effort example of developing an integrated groundwater recharge and flow model with subsurface urban infrastructure, to understand how the urban recharge regimes are impacted and altered by subsurface infrastructure. This model incorporates the available monitoring network data in the city, and as such Bucharest also form an example of good practice in using and translating groundwater monitoring to support decision making and management of the city’s groundwater resource.

In Bucharest urban sewer infrastructure is known to impact the groundwater resource through leakage and new infrastructure has altered the groundwater flow regime in the city. Understanding the impacts of this infrastructure on the urban groundwater resource is important, to mitigate flooding and protect the quantity and quality of the groundwater resource, which supplies the city’s public water supply. Urban infrastructure in Bucharest has: 1) led to reduced natural aquifer recharge; and 2) several subsurface infrastructure components provide focused groundwater recharge, such as water supply network losses, and leaky sewer systems.

To help understand these impacts better, and to mitigate impacts of future subsurface infrastructure Bucharest University and the city municipality has developed a groundwater numerical flow model focused on simulating the interaction between urban infrastructure and the groundwater system of Bucharest city. The groundwater recharge regime and influence of different urban infrastructure to the recharge were modelled by parameterizing preferential flow paths and leakage into the groundwater model to simulate the pipelines and conduits of the urban drainage infrastructure (e.g. 3D sewer system and the 3D subway network).

The model has enabled: (1) the detection of the sewer segments susceptible to groundwater infiltration; (2) the detection of the sewer segments susceptible of exfiltration into groundwater; (3) identification of the sewer segments immersed in groundwater totally (about 17km representing 3.5%) or partially (about 80 km representing 16.5%); and (4) the quantification of the water exchange between groundwater and the city sewer network.

For more detail to Bucharest city and groundwater issues and resource management please refer to Appendix A3.

Application of groundwater models to understand impact of shallow geothermal schemes in cities: good practice

The cities of Basel and Zaragoza provide good practice examples for the development and parameterisation of groundwater heat-transport models to help understand subsurface geothermal energy potential, and what impact new shallow geothermal schemes may have on the overall resource (Epting et al 2013; Garcia-Gill et al. 2014).

Hamburg is also a key bench example of good practice in this (Bonsor et al. 2013) — here the agreed-city scale groundwater model has been attributed with groundwater temperature data from over 100 boreholes in the city, and the model is used by the city municipality water department as key resource in determining new shallow geothermal scheme applications. The thermal influence and elevated groundwater temperature from deep salt domes in the city’s subsurface can be clearly visualized and quantified. The examples of Basel and Zaragosa are discussed in turn below, with the different strategies necessitated by the different drivers and data availability in the cities.

Basel (Switzerland)

3D numerical groundwater flow and heat-transport modeling (FEFLOW ©) have enabled quantified understanding of the thermal influences on the shallow unconsolidated urban groundwater body in the city (Epting et al. 2013; Epting and Huggenberger 2013). Using the model it can be demonstrated that the urban thermal groundwater regime is influenced by: (1) urbanization and annual heating periods; (2) thermal groundwater use; (3) seasonal trends; (4) river-groundwater interaction; and (5) climate change and consequences thereof. The model output facilitate the ‘present state’ of the urban thermal groundwater regime to be described, and also to derive the ‘potential natural state’ of the groundwater body. Furthermore, scenario development facilitates the evaluation of new thermal groundwater and subsurface use as well as potential mitigation measures for the future thermal management of specific regions within the groundwater body. Currently, the developed tools are extended for the thermal management of the various groundwater bodies of the whole city of Basel. The model is also used to help support city management of the groundwater resource (both quantitative and quality) as well as help begin to understand the impact and development of subsurface structures (tunnels, buildings).

Zaragoza (Spain)

A regional 2D model (developed using TRANSIN-IV model code) has allowed the groundwater flow and heat transport processes in the southern part of the city of Zaragoza (80 km²) to be simulated (Garcia-Gill et al. 2014). The modelling has provided understanding and some quantification of the thermal interferences between 28 existent shallow geothermal exploitation systems and the dimensions and cinematics of the heat plumes generated (thermal contamination). The model is calibrated and validated against a high resolution data set obtained from two monitoring networks. The first one is a standard network designed for the management of groundwater resources (head measures and chemical sampling) and the second one consists of a test monitoring network to control the main shallow geothermal installations (temperature). The model has proved useful for the evaluation of the thermal river-groundwater interaction, test concession process protocols for new exploitations and remediation strategies (Garcia-Gill et al. 2014).

Cost-benefit analysis of groundwater modelling for city planning needs: high versus low input data modelling

In view of the small budgets of the most municipalities, it is essential to have cost-efficient tools to assess the service capacity of the underground and manage the subsurface sustainably.

The best-practice approach is an idealised approach with the assumption, that there are extensive data for the model input as well as a lot of time and money. In reality there are good practice and best effort approaches with high and low input hydrogeological data availability, respectively. With high input data availability, detailed high resolution groundwater models can be developed. The output benefit include wider relevance and use of the models for a lot of applications, as well as a greater level of validation of the model and thereby more accurate model outputs and information for different stakeholders, ranging from water managers, city planners and geothermal regulators. The best-effort approach enables process-understanding from only few input-data under the assumption of common principals. However, the limited validation which is possible to these models in the absence of much observational groundwater data means the model outputs cannot be used to support critical decisions on resource management.

Which kind of model will be used of a stakeholder depends first of all on the objectives. If the model is to be used to plan the wastewater pipes for a city for example, a best-effort approach is perhaps sufficient. If there is a need for a model to simulate the complex interactions between surface water and groundwater management, or the impacts of different groundwater abstractions, a higher resolution model with greater input data is necessary.

Figure 4.2    The needs and results of the different approaches of modelling; Figure (a) is the idealised scheme for a best-effort and a best-practice approach; Figure (b) shows the scheme for the mentioned models from Hamburg (best-practice) and Glasgow (best-effort).

[cost-efficiency = relationship between costs for modelling/data acquisition and model output; process understanding = can the model contribute for the common process understanding: accuracy = the accuracy of the model output; less demand of data = amount of needed input data; low time need = time consumption for the modelling; output information = amount of model information as output; questionnaires (amount) = how many different kind of questionnaires (water, heat city planning) can be answered by the model]

Knowledge gaps

Critical knowledge gaps in modelling capabilities limit understanding of urban groundwater resources and subsurface planning at present and include:

• Inclusion of Made (artificial) Ground and subsurface infrastructure into groundwater models
• Modelling the linkage of sea-level change to groundwater-levels in coastal cities
• Integrating real time monitoring data into groundwater models, to enable forecasting and prediction for city planning
• Adequate monitoring systems to provide required data to develop calibrated and validated flow and heat transport models

ID	Current State	Desired State	Gap Description	Gap Reason	Remedies
1	The made ground (artificial) as well as the subsurface infrastructure is not implemented often in the groundwater models.	Groundwater models with all required input data to model the groundwater behaviour more accurate.	Inaccurate groundwater models concerning the missing input data.	The location of artificial made grounds are not mapped and the data of the infrastructure are distributed to different authorities or public utilities.	The mapping of the made ground and the data exchange with the public utilities and the implementation of these data into the model.
2	Most models do not consider the linkage of sea-level change to groundwater-levels in coastal cities or the interaction between the surface water and the groundwater.	Consideration and modelling of the interaction between the groundwater and surface water bodies.	Fluxes (e.g. heat or mass) cannot be modelled accurate at the boundary layers.	The data collection is difficult because the fluxes are transient.	It is necessary to examine and describe the influences and interactions of the neighbouring water bodies to the groundwater bodies by a further conceptual model.
3	The most city-scale models are steady-state models and forecasts and predictions are not possible.	Integrating real time monitoring data into groundwater models, to enable forecasting and prediction for city planning.	Potential changes of the input parameters cannot modelled as a prediction or forecast.	Change the steady-state model to a transient/dynamic model.	For the transient model dynamic data are required and the knowledge of the processes and interactions must be present.
4	The monitoring systems in the most cities are insufficient to get accurate groundwater models.	Adequate monitoring systems to provide required data to develop calibrated and validated flow and heat transport models.	The monitoring network is not dense enough to get spatially highly resolved input data for accurate models.	Insufficient monitoring network for the groundwater.	Implement a monitoring network in the cities with a high temporally and spatially resolution.