Range of application of pumped groundwater-control methods in soils

Groundwater control is typically associated with temporary works to facilitate underground construction, for example of basements, substructures and tunnels, that extend below groundwater level. The objective of groundwater control is to allow excavations to be made in stable and workably dry conditions. If groundwater is not adequately controlled, problems can occur including flooding of the excavation by groundwater and groundwater-induced instability of the base or side slopes of the excavation. The role of groundwater control in civil engineering temporary works is discussed by Roberts et al. (2023); good practice is described in the Construction Industry Research and Information Association (CIRIA) report C750 (Preene et al., 2016).

Groundwater control can be achieved by a variety of techniques, many of which have only a limited range of application in relation to factors such as drawdown of groundwater level and strata hydraulic conductivity (permeability). Selection of the most appropriate technology for the site requirements is a key part of groundwater control system design.

The current paper focusses on pumping techniques and gives the background to the development of guidance for the range of application of groundwater control methods, specifically for pumping methods in soils, including some historical context. The paper presents an updated guidance range of application diagram for pumped methods in soils, which has been based on recent case studies from the authors’ experience. Implications for practice are then discussed.

There are three principal groups of groundwater control methods (which can be used in combination), based on

• pumping (Figure 1(a))

• exclusion (Figure 1(b))

• the application of fluid counter pressures (Figure 1(c)).

Further details are given by Cashman and Preene (2021).

The first group of methods, groundwater control by pumping (Figure 1(a)), is based on pumping from an array of wells, located in or around the excavation, in order to lower groundwater levels.

Exclusion methods (Figure 1(b)) use low-permeability groundwater barriers (either physical cut-off walls or a zone of treated ground) to partly exclude groundwater from the excavation. This minimises the groundwater pumping requirement because the only groundwater to be dealt with is that trapped within the soil/rock in the enclosed area, together with any seepage or leakage through or around the groundwater barriers.

Where the excavation geometry is relatively confined, as in a tunnel or shaft, fluid counter pressure methods (Figure 1(c)) can potentially be used. In this approach the space close to the exposed soil or rock is pressurised with a fluid such as compressed air, water or bentonite slurry. This is a very specialised approach but is the basis of operation for many modern tunnel boring machines (Shirlaw, 2012; Warren et al., 2018) and the wet caisson method of shaft sinking (Allenby and Kilburn, 2015; Smith, 2018).

Groundwater control is not a highly codified activity. For example, there is no British Standard on the execution of special geotechnical works that covers pumped dewatering methods (in contrast there are execution standards for grouting (BSI, 2010) and jet grouting (BSI, 2001) among others). In terms of design standards, the section on dewatering in Eurocode 7 (BSI, 2004) is very brief, comprising only one side of A4 (it is expected that more detail on dewatering will be included in the forthcoming update of Eurocode 7).

In the absence of formal standards, much of the publicly available guidance has been developed by practitioners and published in guidance documents and textbooks (Cashman and Preene, 2021; DOD, 2004; Powers et al., 2007). In the UK, the main guidance on groundwater control is CIRIA report C750 Groundwater Control – Design and Practice (Preene et al., 2016).

It is useful to consider why this is – why is it so difficult to provide generic guidance for groundwater control methods? A key factor is the diverse range of possible methods (including the three categories of methods shown in Figure 1), each of which has different controlling factors, that can be applied in a wide range of geometries in terms of depths and lateral extent of excavations. A further key factor is the very wide range of soil hydraulic conductivity (in engineering applications often termed ‘permeability’) within which groundwater control methods may be required. Some of the guidance diagrams cover an applicable range of hydraulic conductivity of seven orders of magnitude – this is much greater than the potential range of other controlling parameters in geotechnical engineering.

Consequently, it is very difficult to provide wide-ranging comparative guidance between different groundwater control methods. Therefore, without hands-on experience or access to a ‘back catalogue’ of project case histories, it can be difficult for designers and analysts to develop an appreciation of the limitations of the different techniques; hence there is a risk that an inappropriate technique may be specified and as a result the groundwater control system may not perform as required.

The remainder of this paper discusses available guidance that focuses on one category of groundwater control techniques – pumped methods. The guidance that has evolved is defined in terms of soil hydraulic conductivity and drawdown (the vertical distance by which groundwater level is lowered from its original level), a combination of parameters that can be very useful to designers.

These techniques use the cumulative effect of pumping from an array of wells or sumps to lower the groundwater level below the base of the excavation (Figure 1(a)). This means that (in appropriate ground conditions) the excavation can be carried out in the dry, above the temporarily lowered groundwater level. The pumping approach is sometimes known as ‘dewatering’ or ‘construction dewatering’.

Groundwater control by pumping is a long-established method, with civil engineering applications dating back to the early 19th century. It was used by Marc Isambard Brunel when sinking the Rotherhithe shaft on the Thames Tunnel in 1824 (Skempton and Chrimes, 1994). In 1830, Robert Abraham specified that groundwater levels should be lowered using an 8 hp steam engine to provide dry working conditions for foundation construction at the Westminster New Bridewell prison in London (Skempton, 2002). However, the first large-scale application of the approach on a civil engineering project is generally recognised as being the work between 1835 and 1838 on the Kilsby Tunnel on the London-to-Birmingham railway. In this case, pumping from multiple shafts was used to lower the groundwater level and overcome a section of unstable ‘quicksand’ conditions (also known as ‘running sand’) encountered by the tunnel (Preene and Chrimes, 2022).

There is an important distinction within the various pumping methods.

• Open pumping. The principle of these methods is that groundwater is allowed to enter an excavation, from where it is removed by pumping (Figure 2(a)). A key disadvantage is that instability can result either from the proximity of excess groundwater pressures or from the wash-out of fine particles from soil or rock in the slopes and base of the excavation. Entrainment of fines could also compromise the quality of the pumped groundwater.

• Pre-drainage. These methods use wells to lower groundwater levels in advance of excavation works (Figure 2(b)). The method draws water towards the wells, not into the excavation, avoiding troublesome seepages and reducing the risk of groundwater-induced instability. Also, the wells are screened such that the pumped groundwater should be free of fines when correctly specified and installed.

A range of different techniques can be deployed to control groundwater by pumping. The most common methods, summarised in Table 1, include artificial recharge which can be used either as a means of disposing of pumped water, or as means of limiting potential environmental impacts (including ground settlement). Table 1 is based on information from Groundwater Lowering in Construction by Cashman and Preene (2021), which also provides more details of these and other methods.

Other dewatering techniques have characteristics and associated limitations which relate to the most common techniques given in Table 1. Examples include the following.

• Horizontal wellpoints. These comprise horizontal drains installed with a trenching machine and pumped with surface wellpoint pumps. They have the same characteristics as a wellpoint system with the benefit of no connecting pipework at surface.

• Passive relief wells. These comprise wells installed within the footprint of a shaft or excavation to provide artificial pathways for groundwater to flow upward into the excavation without causing soil instability. The objective is to relieve excess pressures in a confined aquifer below the excavation. To avoid the excavation flooding, the inflow from the relief wells is removed by open pumping. Passive relief wells have similar characteristics to deepwells but their application is restricted to a limited range of situations because: they are only applicable for pressure relief of a confined aquifer which is not exposed in the excavation and where the target drawdown level is above dig level; they are generally not viable for medium to high hydraulic conductivities because of the challenge of managing significant inflows during excavation.

• Deepwells pumped with surface pumps. These have similar characteristics to wellpoints and are sometimes used where a shallow drawdown (<6 m) is required in soils with a high hydraulic conductivity.

Structural retaining walls such as concrete diaphragm walls or secant piles are commonly used as temporary works to support excavations and may subsequently be incorporated into the permanent works. Where the retaining wall is not deep enough to reach a low-permeability horizon they cannot form a complete groundwater barrier, and a potential groundwater flow path will remain below the retaining wall into the excavation. In these circumstances, extending the depth of the retaining wall or installing a grout plug carries significant cost, programme and embedded carbon dioxide penalties – see for example Casey et al. (2015). Combining the partial cut-off afforded by the retaining wall and an internal pumped groundwater-control scheme can be an efficient and effective strategy. Note that the retaining wall will need to resist external hydrostatic loads and the internal pumping system will need to remain operational until the permanent works have sufficient structural integrity and ground anchors or weight to resist uplift pressures.

An alternative strategy – see Figure 2(b) – is to use an external pumped groundwater system to control both hydrostatic and uplift pressures during construction. The drawdown outside the retaining walls will in principle lead to some soil settlement owing to the accompanying increase in effective stress (Preene, 2000). This could be an issue if the external soils are compressible, and/or the surrounding buildings are fragile with shallow foundations. Artificial recharge is sometimes used as a mitigation measure to reduce the risk of damaging settlements.

The design of a pumped groundwater-control system starts, but does not end, with a groundwater seepage problem applied to a hydrogeological conceptual model for the site. The objective is to pump from wells or sumps at specified locations, with the intention of manipulating groundwater levels or pore water pressures so the proposed excavation can be made in dry and stable conditions (Figure 3).

Seepage problems are usually analysed using methods in guidance such as CIRIA report C750 (Preene et al., 2016), which mainly comprise closed-form analytical or graphical solutions (such as those based on the Dupuit–Forchheimer equations or flownets) or numerical groundwater models using proprietary software packages. These methods are all based on the use of Darcy’s law, which relates rates of groundwater flow to applied hydraulic gradients. Darcy’s law is relevant to pumping methods because, in essence, such systems work by generating groundwater flow in response to hydraulic gradients (Preene, 2025).

Darcy’s law, developed from experiments by the public health engineer Henri Darcy (1856), relates flow rate Q to hydraulic gradient i

where Q is the volumetric flow of water per unit time (the ‘flow rate’); k is the hydraulic conductivity of a porous medium; A is the cross-sectional area through which the water flows; and i is the hydraulic gradient.

The negative sign in Equation 1 shows that groundwater flows down the hydraulic gradient (i.e. from high head to low head). In most geotechnical problems, groundwater flow is slow enough not to be turbulent and the relationship between flow rate and hydraulic gradient is linear, defined by a constant hydraulic conductivity k.

Darcys’ law shows that for a given hydraulic conductivity, the flow rate is controlled by the area of flow and the hydraulic gradient, which in turn depends on the drawdown of groundwater levels achieved by pumping. In some conditions, discussed below, the hydraulic gradient can be increased by applying a vacuum. This is why different pumping methods have different capabilities – the particular geometries and drawdown limitations of the different techniques will control their potential performance.

Equation 1 shows that as hydraulic conductivity reduces, larger and larger hydraulic gradients or areas of flow are required to achieve significant flow rates. However, the range of hydraulic conductivity of soils that might require groundwater control is around five orders of magnitude – perhaps 10⁻² to 10⁻⁷ m/s – covering the range from very coarse gravels to sandy silt, respectively, and the hydraulic gradients, even with the application of a vacuum, and area of flow cannot be increased by a sufficient factor to compensate. Therefore, at relatively low hydraulic conductivities (below around 10⁻⁷ m/s), pumped well systems may be ineffective because they cannot generate a large enough hydraulic gradient to cause sufficient water flow to adequately lower groundwater levels.

At very low hydraulic conductivities, groundwater may flow more readily in response to electrical potential gradients (i.e. applied voltages) than to hydraulic gradients. This is the principle of the specialised technique of electro-osmosis, which is sometimes used in very-low-permeability soils such as silts or clays with no permeable fissures or fabric (Alder et al., 2015; Casagrande, 1952). In this method direct current (DC) is passed through the soil between an array of anodes and cathodes installed in the ground; water is driven to the cathodes, which are typically installed as wells and used to remove excess water that builds up there.

The fact that different pumping methods influence hydraulic gradients differently means that the type of pumping system is very relevant to the design. Not every pumping technique works for every combination of ground conditions and construction requirements. Therefore, a designer should have an understanding of both the ground conditions and the principal attributes, limitations and constraints of the different pumping methods. These constraints are discussed below.

It is instructive to consider how existing guidance on the application of pumped groundwater-control systems was developed by practitioners. One of the earliest published examples is by Glossop and Skempton (1945), who recognised the value of classifying soils based on particle size and the added value of relating the divisions between grades (being the descriptive terms of fine/medium/coarse, clay/silt/sand/gravel) to recognised engineering properties. They illustrated the usefulness of this approach by relating their proposed grade divisions to various geotechnical processes including for drainage of soils and compressed air tunnelling, as reproduced in Figure 4.

This diagram must have been considered useful by practitioners because it was used in published guidance for several decades. It was included by Harding (1947) in a practical guide to the choice of geotechnical processes and techniques and later in the British Standard Code of Practice for Foundations (BSI, 1972, 1986). The diagram was subsequently annotated to include the range for gravity drainage and vacuum drainage by Mansur and Kaufman (1962), attributed to Moretrench Corp. (a version of this diagram is included here as Figure 5, reversed to match the style of Figure 4). This version was then used in several practical guides to dewatering techniques including Powers (1981), US Army (1983) and a simplified unattributed version in CIRIA report R113, a now superseded UK guidance document on dewatering practice (Somerville, 1986).

A new guidance diagram became available in the 1990s, based on the work by Roberts (1988) which recognised that, in addition to the soil hydraulic conductivity, which can be related to particle size, the choice of groundwater control strategy also depends on the drawdown. Hydrogeological analysis which considered hydraulic gradient, time for drawdown, air entry/capillary rise and gravity/vacuum drainage led to the development of a new diagram using hydraulic conductivity for the horizontal axis and drawdown for the vertical axis. The resulting diagram was found to be consistent with empirical monitoring data derived from back-analysis of a database of completed dewatering projects. The first recognisable version of the current range of application of dewatering systems diagram was published by Roberts and Preene (1994) and is reproduced in Figure 6 together with the supporting empirical data (each of which is a project case history from the authors’ experience).

The work summarised in Figure 6 together with observations made by Cashman (1994) provided the basis for Figure 7, the diagram used in the dewatering best practice guide CIRIA C515 (Preene et al., 2000) and the second edition of CIRIA C750 (Preene et al., 2016). Figure 7 (or adaptations thereof) has been used in dewatering guidance in Qatar (Ashghal, 2014), New Zealand (Christchurch City Council, 2016), South Africa (van der Merwe and Areff, 2021) and Canada (CGS, 2023).

This range of application diagram was subsequently updated (and presented, without detailed explanation, in the paper by Roberts et al. (2023)) based on more recent project experience to better reflect modern design and installation practice. The background to, and justification for, the updated diagram is given later in this paper.

Engineering practitioners can use diagrams such as Figure 7 to give an initial indication of the options for groundwater control for a project. Although the excavation depth and drawdown may be known at concept design stage, the range of uncertainty in potential hydraulic conductivity may be significant and if this range crosses a boundary between different techniques, Figure 7 provides an indication of the nature of potential risks for the groundwater control strategy. In this respect it is worth exploring the underlying drivers for the divisions and why the boundaries are relatively imprecise. The upper hydraulic conductivity limit for pumped groundwater control (dewatering), 10⁻³ to 10⁻² m/s, is characterised by high abstraction flows which increase for larger drawdowns. There are certainly practical issues in abstracting, conveying and discharging high abstraction flows, with the effective limit generally controlled by a combination of factors including the following.

• Cost concerns: Darcy’s law shows that for the same excavation geometry, drawdown and hydraulic gradient, the flow rate is proportional to hydraulic conductivity. Pumping costs and energy demand also increase in line with flow and, at some point, it will become more cost effective to exclude the groundwater with a partial or full cut-off. The economic trade-offs are also sensitive to the plan size of the excavation and geological profile, among other factors. Note that in the absence of a fixed recharge boundary, the hydraulic gradient will tend to reduce at higher hydraulic conductivities (reflecting increased distance of influence), giving a flow rate that is approximately proportional to the square root of the hydraulic conductivity.

• Environmental risk: high abstraction flows, albeit perhaps short term, raise obvious environmental concerns with the potential for adverse impact on groundwater-dependent habitats, water resources, potential mobilisation of legacy contamination in the vicinity or saline intrusion when near the coast.

• Discharge: both environmental and flood risk concerns may arise from the discharge of such high pumped flows.

Note that an appropriately designed and instrumented pumping test can provide valuable information to narrow the range of hydraulic conductivity and to clarify and quantify any risks and concerns.

The lower hydraulic conductivity limit for dewatering schemes, taken to be approximately 10⁻⁷ m/s, is characterised by steep hydraulic gradients, which limit well yields and the extent to which individual wells can interact to generate a drawdown. Furthermore, very loose or soft soils are susceptible to compression as a result of the increase in effective stress that accompanies a reduction in pore water pressure at constant total stress. Compression manifests as a tightening of the packing of the soil particles in the immediate vicinity of the abstraction well, hence a reduction in void ratio. This reduces the hydraulic conductivity and yield of the wells, limiting the drawdown that can be achieved. Hence effective drainage of soft, normally consolidated low-permeability soils is difficult. On the other hand, the increases in effective stress associated with pumping have minimal impact on the void ratio of overconsolidated medium-dense to dense granular soils.

A further factor to consider is that many soils are anisotropic, owing to natural layering when deposited, such that vertical hydraulic conductivity, k_v, is often much less than the horizontal hydraulic conductivity, k_h. The flow to a dewatering system is primarily horizontal and the hydraulic conductivity scale in Figure 7 should normally be taken as the horizontal conductivity, k_h. Note also that layering at soil fabric scale may complicate the determination of conductivity from borehole samples, particularly when these are disturbed. As already noted for the upper hydraulic conductivity limit, an appropriately designed and instrumented pumping test can provide valuable information to narrow the range of hydraulic conductivity and to clarify and quantify the risks. The impact of anisotropy on the design and performance of a dewatering system will depend on both the geometry of the flow path and the degree of anisotropy. A partial cut-off barrier will introduce some vertical flow, which may be impeded by a reduced k_v. However, any reduction in flow will be accompanied by an increase in hydrostatic loads on the partial cut-off and potentially below the base of the excavation. Care is required to ensure that the resulting loads and excess pore pressures do not risk inducing instability. Anisotropy can be beneficial, particularly for fine soils where discrete layers of higher-conductivity sand can enhance flow and drainage of a predominantly silt/clay stratum. Anisotropy generally results in k_h ≫ k_v with a ratio, k_h/k_v, of 10 being a common default assumption where no other information is available. Note that due to geological processes such as folding, there are, albeit relatively uncommon, scenarios where the reverse may be true.

The internal boundaries between the various techniques in Figure 7 are controlled by a combination of factors including the following.

• Sump pumping (a type of open pumping) from within an excavation tends to be more viable in gravelly soils which can form a natural filter, limiting the risk of continuous abstraction of fines.

• The boundaries between single-stage wellpoints, two-stage wellpoints and deepwells relate to limitations on suction lift but also on cost, with wellpoints typically of lower cost than deepwells.

• Wellpoints and deepwells rely primarily on gravity to drive the flow through the soils to the wells, whereas ejector wells, also known as eductors or jet pumps, can pump both water and air, providing vacuum-assisted drainage. It is generally necessary to seal the annulus between the borehole wall and well liner and to seal the well head to allow the vacuum to develop in the well casing. The vacuum increases the hydraulic gradient to each well allowing drainage of lower-permeability soils than can be drained by gravity alone.

The vertical scale in Figures 6 and 7 reaches to 20 m of drawdown, which was the approximate limit of experience and empirical data at the time these figures were developed. This is not a true limitation, and much greater drawdowns have been achieved. However, such projects invariably require high-quality geotechnical and hydrogeological investigation together with expert advice to confirm that conditions are suitable. Note that the techniques that have depth limitations, such as suction lift for wellpoints, require pumps to be installed at or below the standing groundwater level to achieve the maximum drawdowns shown.

In recent decades, contractors have increasingly recognised that vacuum-assisted drainage can be achieved using wellpoints installed with an annular seal (Figure 8(a)), although this may limit the effective suction lift. Furthermore, deepwells can be adapted to provide vacuum-assisted drainage by providing an annular seal around the well casing and an airtight well head cap with a connection to a surface vacuum pump (Figure 8(c)). This arrangement is unlikely to meet the submersible pump manufacturers’ net positive suction head (NPSH) requirements, which may result in a reduced pump service life. Also, the slimline electric submersible pumps used in deepwells rely on the groundwater flow to provide cooling and lubrication, which means that there is a lower flow limit for the reliable application of vacuum deepwells. The minimum yield for a small electric submersible borehole pump is approximately 0.15 m³/h; assessment of well yields using empirical methods from Preene and Powrie (1993) indicates a requirement for hydraulic conductivity >2 × 10⁻⁶ m/s for a 5 m thick aquifer to achieve this.

Over the last 20 years it has become apparent that contractors have increasingly favoured the flexibility, ease of control (and typically lower cost) of vacuum deepwells and wellpoints over ejectors (Figure 8(b)). Much experience has been gained from dewatering projects for real estate developments and urban light railway schemes in Toronto, Canada, where the superficial geology is dominated by dense water-bearing medium- to very-low-hydraulic-conductivity sands and silts.

Figure 9 is an update of the range of application of groundwater control pumping methods in granular soils which takes account of the observations from recent project experience as noted above. The main changes are as follows.

• Addition of vacuum deepwells as an established technique (in addition to the existing methods of open pumping, wellpoints, deepwells and ejector wells). This additional technique is applicable for some of the lower hydraulic conductivity range formerly seen as the exclusive preserve of ejector wells.

• Extending the application of single-stage wellpoints further into the zone of relatively low hydraulic conductivity. This is based on recent project experience which shows that this technique, if designed and installed appropriately, is typically effective and more efficient than ejector wells for shallow drawdowns in soils with a hydraulic conductivity <1 × 10⁻⁵ m/s.

• Recognition from project experience that open pumping (which is typically applied by pumping from sumps and drains) is effective for drawdowns of 1 − 2 m over a wider range of hydraulic conductivity than was indicated in the previous guidance of Figure 7. Recent experience has shown that, for shallow drawdowns, open pumping, if designed and installed appropriately, can be effective provided that the hydraulic conductivity is not so low that vacuum-assisted drainage is required.

An important limitation is that the updated guidance in Figure 9 (and the earlier guidance in Figure 7), applies to uncemented soils because there are relationships between particle size, hydraulic characteristics including conductivity, and seepage-induced instability. The guidance in the figures is not applicable to groundwater control in cemented soils or fractured rock, where open pumping is often significantly more widely applicable, although other concerns, such as poor discharge water quality due to entrainment of fine particles, can be a factor in selection of methods. Some of these issues are explored by Eid et al. (2024).

The updated guidance in Figure 9 provides a useful bridge between theory and practice for the general practitioner and offers a framework for understanding the range of application of the various pumped groundwater-control techniques. The figure provides initial guidance on the selection of suitable methods but gives no information on well spacing, well depth, flow rates or the many other parameters required to fully specify a dewatering scheme.

Figure 9 is useful as part of the decision making process to select an appropriate method of pumped-well groundwater control. Figure 10 presents a flow chart, based on the ‘problem–solution–technology–impact’ framework of Preene (2021). The flow chart is intended to be a useful aide memoire for the development of pumped groundwater-control schemes; the relevance of each stage is discussed below.

• Problem. Clarity is needed on the groundwater issue(s) being addressed. This requires a focus on the interaction between the ground conditions and the proposed excavation (based on the conceptual model for the site) to understand what groundwater control problems must be controlled in which aquifers. In many jurisdictions advance permission (often termed an ‘abstraction licence’) must be obtained from statutory regulatory bodies before groundwater can be pumped at any significant rate. The need for permissions to pump (and discharge) groundwater should be identified as soon as possible, to make sure that sufficient time is allowed to obtain them.

• Solution. The role of the solution is to set out what must be achieved in terms of groundwater control and must be defined at an early stage, to aid the later selection of a groundwater control technology. The solution should identify the strategy for dealing with groundwater –whether to use pumping methods, exclusion or fluid counter pressure methods; which aquifers are to be targeted; and the constraints associated with the necessary regulatory permissions.

• Technology. The selected technology should be capable of delivering the required performance in the expected hydrogeological conditions (based on seepage analysis or groundwater modelling of the proposed system). For strategies based on pumping, this is where the guidance of Figure 9 should be applied to check that the proposed technology will be effective for the required depths and predicted hydraulic conductivity. If this step indicates that the chosen technology may not be effective, then this should prompt a review of alternative technologies or even different solutions (e.g. changing from a pumping to an exclusion solution).

• Impacts. Potential external impacts of groundwater control must be considered. Ideally, the risk of potential impacts should have been identified at the ‘Problems’ stage of the process, so it can be considered during the development of the solution and selection of the technology. In any event the potential for impacts should be reviewed once the proposed solution and technology are clear. If there is potential for problematic impacts, such as ground settlement or migration of contamination, it may be necessary to change the solution and technology, or to add mitigation measures such as artificial recharge.

The design of pumped groundwater-control systems in soils might start with, but is much more than, a hydrogeological modelling and seepage analysis problem. Design of a successful groundwater control system involves the selection of the appropriate pumping method. Therefore, designers must have some understanding of the attributes and constraints of each method. The range of application diagram presented in Figure 9 summarises the latest understanding of the capabilities of commonly used pumped groundwater-control methods, based on records from more than 40 years of projects. The diagram does not remove the need for a detailed groundwater control design (to determine factors such as required dewatering flow rate, well depth and spacing, etc.), but is intended to aid selection of potentially appropriate pumping methods in granular soils as soon as the required drawdown and assessed soil hydraulic conductivity have been estimated.