OR/17/007 Urban groundwater monitoring: identifying good practice - (Roelof Stuurmann and Helen Bonsor)

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 groundwater monitoring; monitoring design; monitoring
drivers; monitoring installation; groundwater monitoring infrastructure

Introduction

This section provides a review of examples of good practice concerning city-scale groundwater monitoring, as wells as an overview of some of the key drivers for groundwater monitoring currently performed in urban areas in Europe. The section discusses key guiding principles for effective city-scale monitoring, comparing different approaches according to different drivers for monitoring, and pre-existing monitoring data and infrastructure in cities.

The saying ‘you can’t manage what you don’t measure’ applies well to groundwater management. At the same time, measurement is useless without a vision of why the data is worth collecting and how it will enable improvements in groundwater management. Such a vision requires a basic understanding of the urban groundwater system. Why, how, where, when, and to whom is groundwater important?

There are a large range of drivers for groundwater monitoring, at city-scales. These include, but are not restricted, to:

• Need to understand the characteristics of the urban groundwater resource — this is typically found to be key driver for groundwater monitoring in cities across Europe which traditionally have not used, or managed the groundwater resource (e.g. cities which do not have issues with flooding and shallow groundwater-levels, and cities which have not historically used groundwater for drinking water supply)
• Need to protect the groundwater resource from over-abstraction and contamination — especially if used for public water supply
• Need to manage flooding (including flooding of building basements)
• Need to manage and redevelop contaminated soil and land
• Need to manage and regulate increasing use of shallow geothermal heat source schemes — both for heating and cooling — in cities

Understanding the depth to the water-table in a city, how this varies spatially and temporally, is essential for informed city planning, so that correct building foundation design is developed, there is appropriate use of infiltration schemes (i.e. they are not installed in areas where the depth to groundwater is less than 1 or 2 metres), and there can be informed utilisation and management of shallow geothermal energy schemes and private and public water supplies in an urban conurbation. Downstream impacts of the utilisation of these subsurface opportunities in parts of the city and redevelopment and regeneration schemes need to understood and incorporated into the wider above ground spatial planning work, to ensure utilise of the groundwater resource upstream, does need lead to negative downstream effects in other Strategic Development Frameworks in urban areas.

Cities which have historically relied on groundwater for industry or public water supply generally have a large amount of existing groundwater data and monitoring infrastructure at a city-scale. In these cities, the key work currently is to revise and systemise the monitoring network to be of an appropriate design and spatial distribution for current monitoring and data drivers, rather than historical drivers. Re-design and systemisation of city-scale groundwater monitoring network has been done very efficiently in Hamburg using the city’s 3D geological and groundwater models to meet the current drivers for understanding the groundwater resource (Bricker 2013).

In cities which have had traditionally very little historical use of the urban groundwater, the main aim of monitoring is to provide regulators and city authorities a general understanding of the characteristics of the groundwater resource — e.g. the depth to groundwater across the city — and how groundwater may impact flooding, building infrastructure, drainage and energy infrastructure. Installing a new monitoring network in these cities, which is appropriate to capture all the important variations in the resource, in the absence of significant existing, is fraught with difficulty. A pilot approach has been trialled in Glasgow and key lessons of good practice learnt from this which other cities can learn from.

Good practices

There is no one good practice for the development of city-scale groundwater monitoring — different data and monitoring network designs are determined by the city planning or regulator needs of the data collated, and also the complexity and variability (both spatially and temporally) of the urban groundwater resource. What is ‘good practice’ in monitoring network design and data sampling is very much dependent on the objectives of the modelling. Table 2.1 illustrates the range of different spatial densities of monitoring required for different drivers within the Netherlands.

Key guiding principles of good practice exist though. For example, it is essential that a monitoring network 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. If there are multiple drivers for groundwater monitoring in a city, it may not be possible for sufficient data (either spatially or temporally) to be collated from one monitoring network. An example of this would be if city planning authorities wanted to collate data to understanding how groundwater is contributing to complex flooding issues across a city, and a water supplier wants to understand regional groundwater flow and seasonal water-table variations across the city. Either a nested monitoring network would have to be installed, so that the city planning authorities could collate more detailed data in some areas, or two separate city-scale networks would have to be developed.

Table 2.1    Monitoring objectives and density in a number of Dutch cities and municipalities

Municipality	Main monitoring objective	Number of wells	outreach
Amsterdam	Protection wooden pile foundations related to leaking/draining sewers. Control high water levels.	>3000 (6 times/year by hand), ca.250 using sensors	Public website
The Hague	Manage high groundwater levels. Takes action (drainage) when groundwater level exceeds 70 cm — surface level. Monitoring by hand every 6 weeks. City contacts complain owners within 3 days!	Hundreds.	Public website
Rotterdam	Wooden piles protection. No other specific objectives determined. Monitoring by hand	ca.2000	Public website.
Gouda	Subsidence control and groundwater flooding	tens	Public website
Vlaardingen	Insight in risks of groundwater flooding related to land subsidence	hundreds	Public website
Breda	Groundwater flow patterns in relation to spreading of groundwater contamination	tens	none
Roosendaal, Bergen op Zoom	Insight in groundwater regimes, reference/aid in responding to complaints of citizens	60–80	none
De Bilt	Possibilities for infiltration of rain water in built up areas (disconnection from the sewer)	ca.40	website
Hoogeveen	Manage groundwater flooding	73 (all sensors)	Public website
Bloemendaal	Manage groundwater flooding due to stopped groundwater extraction and climate change	262 wells, 27 surface water level sites	report

In designing urban groundwater monitoring networks one must start at the back of the process: the finished product. What kind of information is needed, and how often should it be updated? Do you want graphs, tables, or other displaying formats? This amounts to an effective implementation of five aspects: (1) clear monitoring objectives, (2) data storage, (3) data analysis, (4) action plan, and (5) data presentation. Starting at the back, Data presentation: An urban (ground)water helpdesk requires some kind of periodical update of groundwater information in an accessible manner. Here we address three methods. Firstly, a groundwater annual report is a means to meet this goal. It is also an effective tool to grow groundwater awareness, both with the public and within the municipality. Groundwater awareness is crucial for the interpretation of the urban groundwater management, and therefore for a vital urban groundwater monitoring network. Secondly, displaying the results via the internet increases the visibility. A number of Dutch municipalities have already public groundwater monitoring websites: Amsterdam, 2500 locations measured 6 times/year and several hundred continuously (www.maps.waternet.nl/kaarten/peilbuizen.html), see Figure 2.1. The main groundwater monitoring objective in Amsterdam is protecting wooden piles in relation to damaged, and therefore leaking and sewer pipes draining groundwater: The Hague, hundreds of observation wells measured by hand every 6 weeks (www.wareco-denhaag-public.munisense.net/). The wells are distributed around the city, without specific objective; Utrecht presents in addition to the monitoring data also interpretations like isohypse-maps (lines with equal hydraulic heads). Thirdly, operating public groundwater observation wells can help to make groundwater visible. In that case the groundwater level can be read above ground using a recording output from a device which floats on groundwater in an observation well (Figure 2.1).

Figure 2.1    Internet map showing the observation wells in Amsterdam with photo’s showing the observation wells in the streets.

Action plan. An action plan is essential for an effective urban groundwater monitoring network. The plan should comprise the monitoring variables determining whether action must be taken, signal and intervention levels, the type of action required, who is to take action and who is to pay for it.

Data analysis. The method of analysis depends on the measuring objective and the criterion on which action is to be based, e.g., a jump or trend in the groundwater level or exceedance of a certain value.

Data storage. Precious data require careful storage. The availability of long and consistent measurement series delivers a lot of information, e.g. about the effects of climate change on groundwater level fluctuation. The central Dutch DINO-database at TNO (www.Dinoloket.nl) includes all observation well measurements of provinces, water companies, water boards and a large number of municipalities. The measurements are stored, processed and presented in a uniform format, so that the dataset is immune for administration boundary reclassifications. The results are public accessible at Dinoloket, and now from the comprehensive national BRO database for all subsurface data in the Netherlands.

A dedicated monitoring network is designed for each objective individually. The dedicated networks can be combined into an integrated groundwater network.

Integration of the proposed objective-based monitoring networks leads to an integrated network consisting of shallow and deep observation wells. Different measuring targets determine each individually specific boundary condition to the monitoring network. For each observation well must be known what the measurement objective is. It is rare that all objectives can be served with one universal monitoring network.

Historically, many urban groundwater monitoring networks are the result of a steadily growing number of observation wells installed for project-related objectives (e.g. construction of infrastructure). Project monitoring networks are almost always clusters of monitoring wells. Only some of these observation wells should be included into a city-scale monitoring network infrastructure where they match with the general city-scale monitoring objectives. If a city-scale monitoring network incorporates all projects monitoring network infrastructure, the city-scale network will be too costly to maintain and operate, and it will also collate too much data, which is largely non-targeted to monitoring objectives at city-scale. Maintenance and operation costs of groundwater monitoring networks are easily and often underestimated, with tasks including, but not limited to: data collation from loggers and manual dipping; measuring point inspections; repair of incremental damage and tear; labelling; cleaning and purging of observation wells every few years; and water levelling checks every five years to update to absolute height of the tube and measurement datum.

Hamburg is good practice example of how monitoring networks can be reduced and systemised very effectively to match current monitoring data needs, and reduce operation costs (see Case studies). Similarly in The Netherlands, where historically many municipalities have installed monitoring networks over time, a common concern is how to prevent a groundwater monitoring networks from becoming a costly investment that delivers little more than a collection of measurements where one hardly understands how to use these. This is becoming particularly pertinent following the substantial increase in the number of monitoring networks installed in municipalities following the 2006 groundwater duty-of-care legislation (Figure 2.2), and consequents shifts in monitoring requirements.

The need for monitoring arises from the need to successfully manage a resource or system. Monitoring is a cyclic process, where the kind of information needed determines monitoring strategy and design, enabling data to be collected, analysed, and translated into useful information. Besides being used for policy-making or operational management, monitoring results can also be used to refine the monitoring cycle itself.

Figure 2.2    The number of urban groundwater monitoring networks, including used observation wells, in The Netherlands, prior to groundwater duty-of-care legislation (2006). Figures based on responses to a questionnaire (at about 1 observation well per 1000 inhabitants).

Workflows of good practice

Good practice in the design of groundwater monitoring network should include each of the following worksteps (the same worksteps are true for the optimisation of an existing monitoring work):

Statistical approaches and geo-spatial statistics can then be used to determine the minimum or optimum number of monitoring points required — both spatially, and within each aquifer horizon — to capture sufficient data to be able to manage the groundwater resource effectively according to the key drivers for the monitoring.

For the installation of monitoring points there are also some guiding principles of good practice as captured in the Ten Commandments box below.

Case studies

Hamburg

Hamburg forms a key example of good practice in reviewing and systemising a city monitoring infrastructure to ensure the monitoring network is: cost effective to maintain and operate; and, the network design and data generated are appropriate for current monitoring needs.

The key drivers for monitoring in Hamburg are to manage public and private water abstraction, so to mitigate flooding issues from shallow groundwater-levels which floods basements, and ensure protection of the water quality. Prior to the review of the city monitoring, the network consisted of over 2000s points.

Re-design and systemisation of city-scale groundwater monitoring network has been done very efficiently in Hamburg using the city’s 3D geological and groundwater models (Bricker 2013). The use of the models meant that the city municipality and key stakeholders (such as the public water supply utility company Hamburg Wasser) could work from an agreed conceptual model of the urban groundwater system, and identify where higher/lower monitoring density was required according to the location of public supply well fields, interaction of competing uses of the resource, and where there was greater geological and/or aquifer complexities (e.g. adjacent to the tidally influenced estuary river). This approach meant a complex task could be done very efficiently, without different stakeholder’s groundwater data (often held in different formats) having to be systemised and collated before the city analysis and review could be undertaken.

The 3D geological and groundwater models were used by the state geological survey (BUE) and public water authorities to determine the variability of the resource and the minimum resolution of the data required from the optimised monitoring network. Only 40 of the 650 monitoring points are required to supply sufficient data to meet the needs of the Water Framework Directive (WFD) — which only requires a general understanding of the resource.

Monitoring points were reviewed on the basis of their: construction quality (i.e. the boreholes were known to be properly cased and screened; the well-head had good sanitary seal); age; operation performance; and location in the aquifer (both spatially and vertically.

This led to the city’s monitoring network being reduced from over 2000 monitoring points of variable construction quality to just 650.

Glasgow

Glasgow provides an example of approaches which can be used to set up city monitoring networks in cities which have little historical groundwater data or city-scale monitoring.

The need for city-scale monitoring of the urban groundwater resource in Glasgow has arisen due to increasing issues of flooding; the need to be able to use infiltration drainage in the city where appropriate to alleviate the at-capacity sewer system; and to need to protect and mitigate groundwater quality in areas of contaminated soils in the city. The city municipality and the environmental regulator therefore have a need to better understand the general characteristics of the groundwater resource across the city (e.g. depth to groundwater; baseline quality).

There is very little historical groundwater data for the city of Glasgow, and the only recent data available are from very specific sites within the city where there have been major infrastructure projects, major redevelopment work, or remediation of large contaminated land sites. There is no city-scale monitoring infrastructure. To be able understand how shallow groundwater in the city contributes to flooding issues, and might restrict the use of infiltration drainage schemes in the city, city-scale monitoring data are required.

In the absence of significant groundwater data (point data or time series) developing an appropriate monitoring network design which would generate sufficient data was very difficult. Glasgow City Council and the geological survey therefore took another approach — developing a pilot monitoring network within a small area of the city centre where there is a large amount of existing groundwater monitoring data from site investigations, as well as 3D geological data and model. The pilot monitoring network was strategically designed so that groundwater-levels were monitored close to, and also away from the influence of rivers and infiltration schemes. The data and understanding of the groundwater resource gained from the pilot monitoring network was then used to assess what minimum spatial and temporal density of monitoring is required across the city. This knowledge can then be used to develop an appropriate city-scale monitoring network to develop a better understanding of the general characteristics of the resource.

In the absence of significant resources to drill new boreholes either by the geological survey, city authorities or national regulators, the approach being taken to develop the city-scale monitoring network is to adopt existing monitoring boreholes where available (and if of an appropriate location and construction quality) from specific sites and then install new monitoring infrastructure only where there are spatial gaps.

The application of the 3D geological information underpinned the design of the initial pilot monitoring network, and ensured a strategic and appropriate ‘pilot’ was used to inform a city scale design. Analysis of the groundwater monitoring identified that 1 borehole per 1.1 km² is required to capture significant temporal and spatial variations in the shallow groundwater to be able to characterise the general characteristics of resource at a city-scale.

Knowledge gaps

ID	Current State	Desired State	Gap Description	Gap Reason	Remedies
1	Urban groundwater monitoring systems have been developed over time and are ad-hoc, and do not capture appropriate data for current monitoring needs.	Systemised and optimised city-scale urban groundwater monitoring networks for individual city monitoring requirements and legislation.	Lack of systemised urban monitoring.	Lack of monitoring and networks and regulation. The overall responsible for arranging is not clear. Communication geologist-city planner must improve.	Monitoring, research, case studies, legislation.
2	No formal legislation or regulation on specification of monitoring infrastructure.	Greater guidance and/or regulation of required minima monitoring infrastructure required.	Limited urban groundwater monitoring legislative/regulative in some countries.	Lack of recognition of importance of monitoring, and monitoring requirements. Difficulties in implementation in many cities due to lack of monitoring infrastructure.	Monitoring, scientific work, case studies.
3	Lack of monitoring infrastructure and historical records in many cities.	City-scale groundwater monitoring network and data collection to help inform city development opportunities and risks.	Lack of systemised urban monitoring.	Lack of finance, leglislation drivers in some countries, as well as lack of available monitoring data and research.	Investment, research, policy.