Pumping Tests

by Glenn M. Duffield, President, HydroSOLVE, Inc.

What Is A Pumping Test?

Typical well configuration for pumping test in a nonleaky confined aquifer

Figure 1. Typical well configuration for pumping test in nonleaky confined aquifer.

In hydrogeology, a pumping test is a controlled field experiment in which a well is pumped at a controlled rate and water-level response (drawdown) is measured in one or more surrounding observation wells and optionally in the pumped well (control well) itself. Aquifer test and aquifer performance test (APT) are alternate designations for a pumping test.

Pumping Test or Pump Test? Although pumping test and pump test are often used interchangeably, pumping test is the preferred term (Woessner and Anderson 2002). If you're testing the performance of a pump, use pump test; if you're testing the performance of an aquifer through the action of pumping a well, use pumping test.

The goal of a pumping test, as in any aquifer test, is to estimate hydraulic properties of an aquifer system. For the pumped aquifer, one seeks to determine transmissivity, hydraulic conductivity (horizontal and vertical) and storativity (storage coefficient). In layered systems, one also uses pumping tests to estimate the properties of aquitards (vertical hydraulic conductivity and specific storage). Pumping tests can identify and locate recharge and no-flow boundaries that may limit the lateral extent of aquifers as well.

Pumping test, Theis 1935 method, Gridley, IL

Figure 2. Estimation of aquifer properties by matching Theis (1935) type-curve solution to time-drawdown data collected in an observation well during a constant-rate pumping test in a nonleaky confined aquifer (data from Walton 1962).

Typically, aquifer properties are estimated from a constant-rate pumping test by fitting mathematical models (type curves) to drawdown data through a procedure known as curve matching (Figure 2). Diagnostic tools such as derivative analysis are useful for identifying flow regimes and aquifer boundaries from a pumping test prior to performing curve matching.

Besides pumping tests, other types of aquifer tests used to determine the hydraulic properties of aquifers and aquitards include constant-head tests and slug tests.

Methods

Common types of pumping tests include the following:

Constant-rate tests maintain pumping at the control well at a constant rate.
Step-drawdown tests proceed through a sequence of constant-rate steps at the control well to determine its well loss and well efficiency characteristics.
Recovery tests use water-level (residual drawdown) measurements after the termination of pumping. Should normally be considered an essential element of any pumping test.

Designing a Pumping Test

The following concepts should factor into the design of a pumping test:

time of year for pumping test
natural agents of groundwater fluctuation such as barometric pressure changes, earth tides and tidal variations which may affect water levels in observation wells during the pumping test
off-site groundwater use which may influence water levels during the pumping test
effects of pumping test on surrounding water users
depth setting and type of pump in control well
pumping duration and rate
pumping rate measurement and control
water-level measurement and frequency
disposal of pumped water
collection of water quality samples during pumping
location and orientation of streams, faults, lithologic contacts and other potential aquifer boundaries
potential for salt water intrusion in coastal areas

Pumping Rate Measurement and Control

Carefully controlling and recording the pumping rate are essential for obtaining high-quality results from a pumping test.

Measuring the flow rate during a pumping test can be accomplished in a number of ways including the following:

container and stopwatch
in-line flowmeter
orifice weir
weir or flume

Figure 3. Circular orifice weir for discharge measurement (USEPA 1993).

Flow rates should be recorded with sufficient frequency to demonstrate a constant rate or to monitor planned rate changes. In the event of temporary test interruption (e.g., power failure), pumping stop and restart times should be noted to allow for proper interpretation of the test. Bear in mind that the discharge rate often decreases with time as the water level in the control well drops. Kruseman and de Ridder (1994) recommend checking and, if necessary, adjusting the flow rate at least once every hour; however, one should consider more frequent measurements until it becomes evident how often rate adjustments are required (Stallman 1971).

As a general rule, Stallman (1971) recommends controlling the pumping rate within �10% for a constant-rate test or for each constant-rate step. USBR (1995) suggests a tolerance of �5% for the pumping rate. State and local regulations may offer additional guidance regarding an acceptable tolerance for rate variation during a pumping test.

Pumping test analysis software such as AQTESOLV can account for variable pumping rates, but results will depend on how closely recorded flow rates correspond to actual rate changes.

Water-Level Measurement and Frequency

During a pumping test, water levels in observation wells may be measured by means of manual techniques or through the use of sensors with data loggers.

Manual measurement techniques include chalked steel tape, electric water-level sounders and air line methods.
Pressure transducers combined with data loggers provide rapid and accurate measurements.

Even when working with reliable transducers and data loggers, it's good practice to obtain periodic manual measurements at each observation well to (1) confirm transducer readings and (2) provide backup readings in the event of accidental data loss.

When taking manual measurements in the pumped well, be sure to secure the measuring device when sounding the well to prevent accidental entanglement with the pump.

A logarithmic schedule is appropriate for recording water levels a during a constant-rate pumping test. For variable-rate tests (e.g., during recovery or step-drawdown tests), logarithmic sampling should restart for each change in rate. You can find recommended schedules for manual water-level measurements in Driscoll (1986) and Kruseman and de Ridder (1994).

Pressure transducers are available in vented and nonvented models. Vented transducers measure pressure relative to the ambient barometric pressure. Nonvented transducers measure absolute pressure including the pressure of the air column above the sensor. A barometric sensor is required to correct readings from nonvented transducers for changes in barometric pressure.

Most modern data loggers have an option for logarithmic sampling; however, you may convert measurements from a linear to a logarithmic schedule with software such as AQTESOLV.

Pre- and post-test water-level measurements are essential for the identification of trends (e.g., barometric fluctuations) that may require correction prior to pumping test data interpretation. A linear schedule is recommended for monitoring water levels before and after the test. Pre-test water-level measurements should begin at least several days before pumping begins (Figure 4).

Removal of water-level trend from pumping test data

Figure 4. Identification of regional water-level trend from pre- and post-test monitoring (Heath 1983).

Disposal of Pumped Water

In the case of a pumping test that extracts water from an aquifer, care should be exercised in the disposal of the pumped water. First and foremost, be certain that you dispose of the water in accordance with any applicable laws and regulations.

Bear in mind the following considerations as you weigh options for the disposal of discharge water:

Avoid direct discharge of water on the ground surface if the water is likely to recharge the pumped aquifer (e.g., a shallow unconfined aquifer or karst aquifer with sinkholes). Recycling of the pumped water through recharge can result in the false identification of a constant-head boundary or leakage.
Discharge of water to a surface water feature such as a stream is a viable option if it is anticipated that the surface water body is hydraulically disconnected from the pumped aquifer or if it behaves as a constant-head boundary.
Avoid adverse thermal, biological, water-quality, erosion or sediment mobilization impacts when discharging to groundwater or surface water and obtain appropriate discharge permits.

In the design phase of a pumping test, you may use a software tool like AQTESOLV to evaluate the effect of pumped water disposal on observation well response by performing forward simulations for different well/aquifer scenarios.

Test Duration

The decision to terminate a pumping test is best made on the basis of hydrogeologic conditions at the test site and the objectives of the test. Longer tests may be necessary to estimate specific yield in an unconfined aquifer or to observe boundary effects.

Plotting and inspecting pumping test data as they are collected in the field can help you to decide when it's appropriate to end a pumping test.

Have the objectives of the test been achieved?
Have sufficient late-time data been collected to estimate specific yield in an unconfined aquifer?
Has the test continued long enough to detect and locate aquifer boundaries?

Other factors may influence the planned duration of a pumping test including budgetary constraints and regulatory requirements. Often, applicable regulations may establish a minimum duration required for a test, but longer tests may be necessary to achieve other test objectives.

Links to Guidance Documents

Find additional guidelines and procedures in the list of guidance documents.