Water Well Journal Article

Reducing Seawater Intrusion via Low-tide Pumping

Over an eight-year period, the water association on Camano Island was able to minimize chloride in community wells and protect the island's ground water supply.

By Jerry Purdum and Glenn R. Engel

This article describes efforts to mitigate the effects of seawater intrusion on small water systems on Camano Island, Washington. With demonstrated success of these efforts over an 8 year period, it is hoped that others can benefit from the techniques utilized by these water systems. Some history of the efforts will be explained as well as presentation and analysis of data collected over a 17-year period.

Background: The Seawater Problem

Camano Island, Washington, is a small island located in Puget Sound, surrounded by saltwater. Drinking water is mostly obtained from a narrow or thin aquifer located 80 feet below mean sea level, while the surrounding Puget Sound waters are 300 plus feet deep.

This water source suffers from 'lateral seawater intrusion'. This aquifer was deposited by glacial flows, is generally 5 to 30 feet thick, and consists of fine to coarse sand and small gravel. The sole source of replenishment or recharge is from an average annual rainfall of 25 inches.

In 1994, Sierra Vista Water Association (SVWA) experienced problems with seawater intrusion for the first time. At that time, SVWA was a small two-well water system of less than 100 connections located in the southern panhandle of Camano Island, a section 1 mile wide by 8 miles long. In this area of Camano Island (Figure 1), four adjacent community water systems supply drinking water to an area 3-1/2 miles long by mile wide including 500+ connections. Private wells in the same area also draw water from the same aquifer.

In 1994 the SVWA water system was pumping from two wells with screens 85 feet below MSL into a 30,000 gallon water storage tank at elevation 200 to 210 feet above MSL. Average summer chloride concentrations at that time approached 150 mg/L (parts per million). One inorganic water test revealed a chloride test result of 254 ppm, slightly above the 250 ppm EPA maximum contaminant level (MCL) for this secondary contaminant. The accepted threshold for identifying seawater intrusion is 100 ppm.

The SVWA wells were controlled by a mechanical timer which spread the pumping over a 24-hour period. With two days of water storage available, the timer allowed some mitigation of the very large spikes in demand on the water source, usually in the morning and again in the evening when people return from work.

An extensive collection of data[1], history, research, and consultations with hydrogeologists, representatives from the U.S. Geological Survey, Department of Energy, Department of Health, and water managers/engineers inspired us to study our problem for the purposes of mitigation. A chloride field test kit and static water measurement equipment was acquired to observe chlorides as well as effects of tides on static water elevations. Varied pumping rates, total pumping over time and seasonal variations were observed.

The following observations were made during this investigation:

Seawater intrusion increased with total volume pumped as well as with increased rates of pumping.
Static well water elevations through tide cycles varied 40 percent of the tide changes. A tide change of 10 feet reflected a change in static well casing water of 4 feet.
Chloride concentrations in the well casings were higher at high tides versus low tides. Fresh water floats on salt water and thus the interface between fresh water and seawater goes up and down with the tide. At low-tide this interface point is at its lowest elevation.
Drawdown and recovery tests revealed very little drawdown with rapid recovery.
Neighboring community water systems were also experiencing seawater intrusion.
One neighboring community had chloride tests of 1,400 ppm in 1986 and 548 in 1996, almost 8 percent seawater. (seawater is 18,000 ppm)
One community water system was surrounded by private wells that pumped on demand during daylight hours to pressure tanks. This community system tried pumping only at night to avoid pumping simultaneously with its nearby private well owners. This system experienced a significant drop in chloride concentrations.

Mediation Efforts

The observations were presented to the association members at an annual meeting in August 1994. The membership was requested to immediately practice conservation, authorize and finance the addition of a third well, and to increase water storage. These proposals were accepted by the membership.

After the meeting Glenn suggested using a computer to control pumping to take advantage of the observation that the chlorides were apparently lower at low tides. After some discussion it was decided to give the idea a try. Since no commercial product was available at that time to do this type of control, a custom program running on a palmtop computer was developed in conjunction with a custom control circuit.

In January 1995 the computer and control circuitry was installed and test measurements of the system continued. Initial measurements were promising and showed a decline in chlorides over the following year in excess of 50 ppm.

Figure 1. A cross section of the Camano Island aquifer. From this aquifer, four adjacent community water systems supply drinking water to an area 3-1/2 miles long by 1/2 mile wide, including 500-plus connections. Private wells in the same area also draw water from the aquifer. (enlarge)

Figure 2. The palmtop installation. The computer allows entry of system and pump configuration settings. (enlarge)

Figure 3. A computer screen shot showing tides. (enlarge)

The computer program allowed entry of the total system water demand (daily), the number of wells, and the pump rate of each pump as well as other configuration data such as tidal lag (time between high tide in the ocean and high tide in the well casing), and pump sequencing and timing algorithms. To compute tide levels, the program used tide prediction coefficients provided by the National Ocean Survey for our locality. The calculations are based on astronomical effects of the moon and sun and are valid as long as a significant geologic change to the area does not occur. The computer program calculated the number of pumping hours necessary each day to satisfy system demand and pumping to be accomplished during low-tide intervals to avoid pumping at high tide. When low-tide emphasis is not desired, the program can be used to automatically spread out the pumping over an entire day.

Using the computer to control pumping also provided another important advantage by leveling out draw from the aquifer over an extended period of time. In the case of SVWA, water usage on weekends could easily double over similar periods during the week. By using the computer in conjunction with multi-day water storage, these types of spikes can be averaged over an entire week. This averaging works for both tide based emphasis as well as traditional linear sequential pumping.

In April 1996 a third well was added followed by a 60,000 gallon storage tank in April 1997. All three wells are within a 500 meter radius. During this time the chlorides continued to decline even though consumption was increasing. The added storage supplied one week of storage during peak summer demands and two weeks of storage during the winter.

Reducing Chlorides

Figure 4 shows the history of the chloride measurements, the system water consumption, number of active connections, and important system changes such as initiation of low-tide pumping, adding wells, and system storage capacity changes.

Figure 4. SVWA historical data: The history of the chloride measurements, the system water consumption, number of active connections, and important system changes such as initiation of low-tide pumping, adding wells, and system storage capacity changes.(enlarge)

During the period from 1985 to 1989 the chlorides generally followed the consumption growth until the second well was added in 1989. Chlorides declined for a short time after the well was added and then began to track the consumption again until 1995, when pumping at low tides was initiated. After commencing pumping only at low tides, the chlorides declined for the next year and continued to decline after the third well was added in 1996. This was in spite of continuing growth in consumption during this time. Construction of the additional storage tank provided the needed equalizing and standby storage to allow full use of low-tide pumping even considering high weekend demands. The objective was to achieve a weekly amount of total pumping even though some days had consumption that exceeded the amount of new water added to the storage tanks during low-tide pumping.

Thus, the chloride curve shows a loss of 50 ppm from January 1, 1995 through April 1996. Continued low-tide pumping, installation of the well # 3 in 1996 and the added water storage tank in 1997 has kept the average annual chlorides between 50 and 100 ppm for the last seven years with continued water demand from consumer growth.

These results are consistent with several other water systems on Camano Island that have implemented tide based pumping subsequent to reports from SVWA on its preliminary results. One system had their two wells go from chloride readings of 160 and 180 ppm to 60 and 40 ppm, respectively, over a two month period after utilizing low-tide pumping.

It also should be noted that educating the island's residents on the need for conservation in 1994 has had a long-lasting positive effect on per-connection consumption.

Summary

Many factors affect seawater intrusion such as usage patterns, tides, aquifer recharge, and adjacent well utilization. By utilizing a computer to control pumping it is possible to totally remove demand based spikes in pumping by taking a longer term view and pumping for a daily or weekly average rather than attempting to follow usage spikes. In addition, use of a computer (see figure 4) allows for advanced algorithms to be utilized such as low-tide pumping and 24 hour evenly distributed pumping.

Authors

Jerry Purdum, PE (email) has been water manager of SVWA for 12 years. He also is operations manager for the Camano Water Systems Association, which is an island-wide grass roots organization on Camano Island dedicated to "helping small water systems help themselves." Jerry has been active as a registered professional civil engineer in Washington, Alaska, Oregon, and other areas.

Glenn Engel (email) is a research scientist for Agilent Technologies Inc. He has degrees in electrical engineering and computer science from Kansas State University and has been involved with advanced instrumentation of various sorts for more than 24 years. His interest in computer-based pumping originated from being a member of SVWA at the time of the initial chloride problems developed. He is the author of the computer-based pumping program used by SVWA. More information on the program and other related information can be found at http://www.engel.org/welltimerpro/.

[1] The SVWA data collection included 171 certified lab chloride/specific conductivity tests from 1982 to 2003.