Published in Soil Sci. Soc. Am. J. 68:562-566 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
DIVISION S-5—NOTES
A COMPARISON OF SEVERAL APPROACHES TO MONITOR WATER-TABLE FLUCTUATIONS
C. P. Morgan and
M. H. Stolt*
Dep. of Natural Resources Science, Univ. of Rhode Island, Kingston, RI 02881
* Corresponding author (mstolt{at}uri.edu).
 |
ABSTRACT
|
|---|
Relationships established between redoximorphic features and the seasonal high water table should be based on the most accurate representation of water-table fluctuations. In this study, we compared hydrographs developed using water-table readings made at weekly intervals over a 12-wk period to those developed over the same period for an adjacent water-table well using measurements recorded every half hour by a data logger. The hydrograph developed using the weekly readings underestimated the height of the water table for 33% of the study period. A simple inexpensive maximum water-table recording device (MWTRD) was developed to record the highest level the water table reached during the interval between site visits. Two approaches are demonstrated for improving the accuracy of the weekly hydrograph using data collected by the MWTRD along with a limited amount of logger data. These adjusted hydrographs accounted for >80% of the underestimation of the height of the water table compared with the weekly measurements.
Abbreviations: MWTRD, maximum water-table recording device • SHWT, seasonal high water table
|
INTRODUCTION
|
|---|
DEPTH TO THE SEASONAL high water table (SHWT) is an important factor in many land use decisions. Soil interpretation rating guides, developed by the Natural Resources Conservation Service, list over 30 land use decisions that are dependent upon knowing the depth to the SHWT (Soil Survey Staff, 1996). Some of the more important interpretations based on the SHWT include the construction of buildings and roads, surface application of waste, and the use of on-site waste disposal systems. The depth to the SHWT is also a key factor in identification of wetlands (Environmental Laboratory, 1987).
The standard method for developing water-table hydrographs is to install monitoring wells and measure the depth to the water table at regular intervals. These hydrographs are examined relative to the depth, abundance, and type of redoximorphic features so that the redoximorphic features can be used as an indicator of the SHWT. Measurement intervals commonly used in soil morphology–water table studies include weekly (Galusky et al., 1998; Genthner et al., 1998), biweekly (Hyde and Ford, 1989; Khan and Fenton, 1994; Elless et al., 1996; Griffin et al., 1998; Rabenhorst and Lindbo, 1998; Jacobs et al., 2002), and monthly (Zampella, 1994). Other studies have used measurement intervals that varied over the length of the study (Daniels et al., 1971; Veneman et al., 1976; Franzmeier et al., 1983; James and Fenton, 1993).
Water table response to precipitation generally occurs quickly with the water table reaching its maximum height for each event within the span of 1 d, and often times within a few hours following the precipitation event. The water table generally falls to the pre-event level over the course of a few days, depending on the amount of rain (McDaniel et al., 2001). The dynamic nature of the ground water system makes timing and frequency of the water-table measurements an important factor to consider when making interpretations concerning the relationship between redoximorphic features and water-table depth. Therefore, data collected at monthly, biweekly, or weekly intervals may be inadequate to truly represent water-table fluctuations on a site.
An alternative to manual observations is to equip each well with a data logger. Data loggers can be set to record the depth to the water table at a predetermined interval and allow for a large amount of detailed information to be collected with a minimum number of site visits. Data loggers designed to measure water-table levels are quite costly; thus, if a soil morphology–water table study involves many wells, the cost of outfitting each well with a data logger may be prohibitive. The objectives of this study are: (i) to demonstrate a simple, inexpensive device that can be used to record high water-table levels between site visits and; (ii) to show that with this simple device a more accurate depiction of water-table fluctuations can be generated.
|
Materials and Methods
|
|---|
This study examined water-table fluctuations in a soil formed in glacial fluvial parent materials. The soil classified as a coarse-loamy, mixed, mesic, Aquic Dystrudept. The site was instrumented with two water-table wells, a device to record the maximum water level between site visits (MWTRD), and a water-table data logger set to record the water level at 30-min intervals. Wells were installed using a hand auger to the depth of 120 cm and sealed at the ground surface with bentonite to prevent water from infiltrating along the side of the well. The logger was used to check the accuracy of the MWTRD as well as to monitor the rise and fall of the water table (spike) following a precipitation event.
The MWTRD consists of a steel rod on which a float can move up and down with the fluctuating water table within the well (Fig. 1)
. The rod is 0.5 cm in diameter and painted with a rust inhibiting paint (Rustoleum, Rust-Oleum Brands, Vernon Hills, IL). A 3.2-cm diameter washer (slightly smaller diameter than the inside diameter of the well) is secured at the bottom of the rod by two push-nuts, one above and one below the washer. The washer helps keep the device centered in the well and keeps the float from sliding off the bottom of the rod when the MWTRD is removed from the well. The float is a wine bottle cork 2 cm in diameter and 4 cm high that is drilled to slide freely up and down the rod. A 2-cm diameter thin-plastic washer sits on top of the cork and keeps the magnet from getting hung up in the rough edges created by drilling the cork. Washers can be made from a plastic coffee can lid using a hole-saw of the same diameter as the cork. The magnet is a microstirring bar, 2 mm in diameter and 7 mm in length (Fisherbrand #14-511-67, Fischer Scientific, Pittsburgh, PA). The MWTRD was installed in a 3.4-cm inside diameter polyvinyl chloride (PVC) pipe slotted to within 30 cm of the soil surface. The MWTRD was centered in the well by the washer at the bottom of the rod and by a hole drilled in the well cap through which the rod is extended 2 cm. Following a rainfall event the cork rises with the water table and the magnet is pushed up along the steel rod. When the water table and cork drop, the magnet remains stationary marking the location of the highest water-table level.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 1. Device for recording maximum water-table level between site visits. The maximum water-table recording device (MWTRD) is placed inside a 3.2-cm i.d. slotted polyvinyl chloride (PVC) well with a hole in the cap. As the water table rises in the well, the float pushes the magnet up the metal rod. When the water table and the float decline the magnet remains in place recording the highest level the water table reached.
|
|
Wells were monitored on a weekly basis from early April to mid June 2002 (Fig. 2)
. This time frame was chosen because it represented a time period during the growing season that the water table would be at the highest level. Daily precipitation data were obtained from the New Shoreham Wastewater Treatment Facility, located 4.5 km south of the study site. To read the MWTRD, the device is lifted from the well and the distance from the magnet to a mark on the rod indicating the soil surface is measured. The distance is a record of the highest level the water table reached between readings. During each weekly measurement two numbers were recorded, the current depth to the water table and the depth to the magnet from the soil surface. If the magnet was recorded at a depth that was shallower than the depth of the water table recorded during the previous week's visit, we assumed that at some point during the previous week the water table rose to the position of the magnet in response to precipitation. If the magnet was at a level equal to the depth of the measured water table recorded the previous week, we assumed that no rise in the water table occurred since the last reading. The MWTRD is reset following each reading by positioning the float and the magnet at the bottom of the device and reinstalling the device in the well. At this point the float and magnet will rise to the current water-table level within the well or remain at the bottom of the device if the well is dry. Laboratory studies found that the water table needs to rise at least 0.5 cm before the cork responds enough to move the magnet indicating a rise in the water table.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 2. Water-table fluctuations from 6 Apr. to 13 June 2002 in a coarse loamy Aquic Dystrudept. The solid line represents water-table fluctuations based on data measured with a logger at half hour intervals. The dashed line represents water-table fluctuations based on manual monitoring on a weekly basis. The black triangles represent the highest level the water table reached during the weekly interval as recorded by the magnet on the maximum water table recording device. The "T" under the last water-table inflection indicates the spike used to construct Fig. 3 and 4. The "S" between the vertical dotted lines shows the section of logger data used to calculate the deep seepage rate (0.1 cm h–1).
|
|
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 3. Estimated water-table fluctuations 6 Apr. to 13 June 2002 based on weekly measurements adjusted with the use of the water-table recording device. Dashed line represents water-table data recorded weekly. Solid line shows the single-spike method hydrograph; where a single spike, inferred from the position of the magnet on the maximum water-table recording device (MWTRD) as an increase in the water-table level due to a precipitation event, is added to the weekly data. The height of the spike is determined using the magnet data, while the shape is that of the template spike.
|
|
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 4. Hydrographs depicting water-table fluctuations from 6 Apr. to 13 June 2002 in a coarse loamy Aquic Dystrudept. The thin line is the hydrograph produced using the logger data. The heavy line is the hydrograph constructed using the precipitation–deep seepage approach where weekly water-table readings are used in conjunction with the magnet data, the spike and deep seepage templates, and daily precipitation data.
|
|
Two methods were explored to refine the hydrograph created using the weekly water-table measurements. Both methods use the magnet measurement, a template of a spike in the water table created using a portion of the logger data (see June data in Fig. 2), and EXCEL spreadsheet software (Microsoft Corp., Redmond, WA) to compile the data sets. The first method uses the weekly data as a base line to which template spikes of the same magnitude of the height of the maximum water table are added. The template represents the typical pattern of the water table rising and falling in response to a precipitation event (Fig. 2). This method is the more conservative of the two approaches because only a single spike is used; accounting for only the largest fluctuation in the water table that occurs between the weekly site visits. In the single spike method, the magnet data and the shape of the spike from the logger determine the height of the template spike. Water-table logger data from a number of sites have shown that although the magnitude of a spike in the water table varies depending upon the size of the precipitation event, the shapes of the spikes, for an individual well, are very similar. The similarity in the shapes of the spikes makes it possible to use a single spike from a given site as a template from which other spikes in the hydrograph can be developed for that site. It should be noted that spikes representing very small (generally <5–10 cm) fluctuations in the water table might be the exception. Therefore, these smaller spikes may not be appropriate choices for the template.
The single spike hydrograph was created by "pasting" the template spike data into a spreadsheet containing the weekly readings for those weeks where the magnet indicated the water table rose in response to precipitation (Fig. 3)
. Dates were converted to a number format to simplify the combining of the weekly data with the half-hour interval data generated by the logger that corresponds to the template spike. An appropriate value was applied in the spreadsheet to the template data so that the spike would be pasted in the interval representing the appropriate week. Placement of the template peak within the weekly hydrograph was such that the trailing end of the peak ended at the next weekly reading. In the same manner, the depth component of the template data was fit so that the top of the peak matched the height the magnet reached during the weeks where water table rises were recorded by the magnet. Standardizing the depth and date components of the data allows the template to retain the same shape while being inserted at the correct depth and time interval (Fig. 3).
In weeks where more than one precipitation event occurs, only the event resulting in the largest rise in the water table is recorded by the MWTRD. Thus, fluctuations in the water table caused by smaller precipitation events that occur over the same weekly interval are not accounted for when the single spike method is used to create a hydrograph. A second approach, that uses precipitation data and the rate at which the water table declines (deep seepage rate, Boersma et al., 1972), was developed to account for these missing spikes and to further refine the weekly hydrograph. In the precipitation–seepage method, the magnitude of the rise in water table for each rainfall event was estimated by comparing the logger data with precipitation amounts. The 60-cm spike in the water table, used to create the template spike (Fig. 2), resulted from a 5-cm precipitation event. The peak heights of additional spikes were calculated using the same ratio of 1 cm of rain resulting in a 12-cm rise in the water table. The height the water table reaches following precipitation is also dependent upon the water-table depth before the precipitation event. Therefore, we also calculated the deep seepage rate between precipitation events to estimate the correct starting depth to add the water-table spike. The deep seepage rate can be calculated as the rate of decline in the water table that occurs after the water table has risen and then fallen to approximately its preprecipitation event level, or as the rate of decline in the water table if precipitation is absent for the time interval. The deep seepage rate at this site was calculated from late May data that represent the end of the longest continuous dry period of the study (Fig. 2). The water table declined at a rate of 0.1 cm h–1 over this period.
The precipitation–seepage method combines the weekly readings, the magnet data, the smaller weekly peaks based on precipitation amounts, and the deep seepage rate to create the hydrograph (Fig. 4)
. The weekly readings provide a check of the actual water-table level and prevent cumulative error in the constructed hydrographs. These checks, however, may result in a hydrograph that shows artifacts from the construction process. For example, there is a vertical line on the declining side of largest spike (see mid May, Fig. 4) and a time-offset between many equivalent peaks when comparing the precipitation–seepage hydrograph with the logger data hydrograph (Fig. 4). The vertical decline in the water table, seen at the trailing end of the large spike, is a result of the template spike not descending to a deep enough level by the time the next weekly measurement was made. To account for the difference in water-table levels for the same period in time, a vertical line was added to join the bottom of the spike to the actual water-table level recorded during the weekly reading. Precipitation data were recorded daily. To be consistent when using these data in the precipitation–seepage method, we inserted the spikes at midnight on the last day of a measured precipitation event. This approach resulted in an offset between the timing of the spikes in the hydrograph representing the actual water-table fluctuations (logger data) and the hydrograph created using the precipitation–seepage method. The vertical line and the offsets are minor flaws in the method and the precipitation–seepage hydrograph still shows nearly the identical pattern, maximum water-table heights, and durations as the hydrograph constructed from the logger data.
|
Results and Discussion
|
|---|
Comparison of the weekly data to the logger data showed the weekly data underestimated the height of the water table about 33% of the time (Fig. 2). The weekly observations underestimated the height of the water table by a few centimeters to >55 cm. Weekly measurements were ineffective in describing the cumulative duration of the water table above 60 cm even though the water table was above this depth for 17% of the study period (Table 1). Using the conservative single-spike approach (magnet data and the water-table spike template), we accounted for approximately 82% of the underestimation in the water-table height (Fig. 3). The single-spike approach slightly underestimated the cumulative duration of the water table above 60 cm and very effectively described the cumulative duration above 40 cm (Table 1). This approach has the advantages of taking less time to construct the hydrograph compared with the precipitation–seepage method, as well as not requiring precipitation data. However, two drawbacks to the single-spike approach are of concern. First, the magnet only records the largest precipitation event between site visits. Second, the single-spike hydrograph at times overestimates the height of the water table leading to a slight overestimation of the cumulative duration of the water table (see the 80-cm depth, Table 1).
View this table:
[in this window]
[in a new window]
|
Table 1. Cumulative duration of the water table above specified depths during the 12-wk study period for an Aquic Dystrudept based on four methods to construct hydrographs. Water-table levels were recorded at 30-min intervals to construct the logger hydrograph.
|
|
The precipitation–seepage approach (which uses daily precipitation data in conjunction with the magnet data and the water-table spike template) accounted for approximately 88% of the underestimation of the water table (Fig. 4). This approach used daily precipitation data to account for fluctuations in the water table that were not recorded by the magnet. In addition, the deep seepage rate was used to estimate water table decline between precipitation events. Both the deep seepage rate and daily precipitation data are easily obtained and the application of these data result in a graph that looks very similar to the hydrograph constructed from the logger data (Fig. 4). Regardless of the depth, cumulative durations calculated using the precipitation–seepage method were within 3% of the cumulative durations calculated from the logger data (Table 1) suggesting that this approach provides an excellent estimation of water-table fluctuations.
The decision to use one method over the other depends on the application of the information. For soil morphology–water-table relationships, the more that is known about the depth and duration of the water table the better. For such cases, the method applying the deep seepage rate and daily precipitation data appears to be the most appropriate. For direct applications, such as SHWT determinations for siting an on-site waste disposal system, there is more concern about the duration of the water table above a certain level, than for the amount of time it may be deeper than predicted. In such cases, the simpler more conservative single-spike approach may be sufficient.
Another factor in the decision to use one approach over the other is the length of time between site visits. As this interval increases the chance of additional precipitation events occurring between site visits also increases. Therefore, if the interval between site visits is longer, such as a month, the precipitation–seepage approach might be the more appropriate method. If the interval between site visits is relatively short, such as the 1-wk interval used in this study, the single-spike approach may prove to be adequate.
The magnet measurements along with measurements taken during site visits are effective checks that limit distortion caused by cumulative error, allowing a water-table hydrograph to be refined using precipitation data. When the template is used in conjunction with the magnet measurements, not only is the maximum level of the water table known, but it is also possible to estimate the cumulative duration of the water table at various depths. Both methods require the data logger to be on-site only long enough to create the template, therefore, a single logger can be moved from site to site to create the templates for multiple wells. The data for this study were collected during the months of April through June. Water-table fluctuations may vary with the season in response to changes in evapotranspiration rates and resulting antecedent water-table levels. Therefore, different template spikes and deep seepage rates may need to be developed for different times of the year.
The frequency at which water-table measurements are made greatly influences the ability of a hydrograph to depict the dynamic nature of the water table. Even making observations once a week can create a hydrograph that inaccurately depicts water-table fluctuations. The simple device described in this paper can be used to construct water-table hydrographs that are more accurate and therefore more useful in developing soil based interpretations of the SHWT.
|
ACKNOWLEDGMENTS
|
|---|
This work is a contribution (No. 3979) of the Rhode Island Agricultural Experiment Station. This work was funded through the USEPA Block Island and Green Hill Pond Watershed National Decentralized Wastewater Treatment Demonstration Project.
Received for publication August 9, 2002.
|
REFERENCES
|
|---|
- Boersma, L., G.H. Simonson, and D.G. Watts. 1972. Soil morphology and water table relations: I. Annual water table fluctuations. Soil Sci. Soc. Am. Proc. 36:644–648.
- Daniels, R.B., E.E. Gamble, and L.A. Nelson. 1971. Relations between soil morphology and water-table levels on a dissected North Carolina Coastal Plain surface. Soil Sci. Soc. Am. Proc. 35:781–784.
- Elless, M.P, M.C. Rabenhorst, and B.R. James. 1996. Redoximorphic features in soils of the Triassic Culpeper Basin. Soil Sci. 161:58–69.
- Environmental Laboratory. 1987. Corps of Engineers wetland delineation manual. Tech Rep. Y-87–1. U.S. Army Corps of Eng., Waterways Exp. Stn., Vicksburg, MS.
- Franzmeier, D.P., J.E. Yahner, G.C. Steinhardt and H.R. Sinclair, Jr. 1983. Color patterns and water table levels in some Indiana soils. Soil Sci. Soc. Am. J. 47:1196–1202.[Abstract/Free Full Text]
- Galusky, L.P., Jr., M.C. Rabenhorst, and R.L. Hill. 1998. Toward the development of quantitative soil morphological indicators of water table behavior. p. 77–93. In M.C. Rabenhorst et al. (ed.) Quantifying soil hydromorphology. SSSA Spec. Publ. 54. SSSA, Madison, WI.
- Genthner, M.H., W.L. Daniels, R.L. Hodges, and P.J. Thomas. 1998. Redoximorphic features and seasonal water table relations, upper coastal plain Virginia. p. 43–60. In M.C. Rabenhorst et al. (ed.) Quantifying soil hydromorphology. SSSA Spec. Publ. 54. SSSA, Madison, WI.
- Griffin, R.W., L.P. Wilding, and L.R. Drees. 1992. Relating morphological properties to wetness conditions in the Gulf Coast Prairie of Texas. p.126–134. In J.M. Kimble (ed.) Characterization, classification, and utilization of wet soils. Proc. Eighth International Soil Correlation Meeting, Baton Rouge, LA, and Victoria, TX. 6–21 Oct. 1990. USDS-SCS Publ. Natl. Soil Surv. Center, Lincoln, NE.
- Hyde, A.G., and R.D. Ford. 1989. Water table fluctuations in representative Immokalee and Zolfo soils of Florida. Soil Sci. Soc. Am. J. 53:1475–1478.[Abstract/Free Full Text]
- Jacobs, P.M., L.T. West, and J.N. Shaw. 2002. Redoximorphic features as indicators of seasonal saturation, Lowndes County, Georgia. Soil Sci. Soc. Am. J. 66:315–323.[Abstract/Free Full Text]
- James, H.R., and T.E. Fenton. 1993. Water tables in paired artificially drained and undrained soil catenas in Iowa. Soil Sci. Soc. Am. J. 57:774–781.[Abstract/Free Full Text]
- Kahn, F.A., and T.E. Fenton. 1994. Saturated zones and soil morphology in a Mollisol catena of central Iowa. Soil Sci. Soc. Am. J. 58:1457–1464.[Abstract/Free Full Text]
- McDaniel, P.A., R.W. Gabehart, A.L. Falen, J.E. Hammel, and R.J. Reuter. 2001. Perched water tables on Argixeroll and Fragixeralf hillslopes. Soil Sci. Soc. Am. J. 65: 805–810.[Abstract/Free Full Text]
- Rabenhorst, M.C., and D.L. Lindbo. 1998. Micromorphology of sandy epipedons along an upland-wetland transect. p.195–208. In M.C. Rabenhorst et al. (ed.) Quantifying Soil Hydromorphology. SSSA Spec. Publ. 54. SSSA, Madison, WI.
- Soil Survey Staff. 1996. National soil survey handbook, 430-VI, Natural Resources Conservation Service, U.S. Gov. Print. Office. Washington, DC.
- Veneman, P.L.M., M.J. Vepraskas, and J. Bouma. 1976. The physical significance of soil mottling in a Wisconsin toposequence. Geoderma 15:103–118.
- Zampella, Robert A. 1994. Morphologic and color pattern indicators of water table levels in sandy pineland soils. Soil Sci. 157:312–317.
This article has been cited by other articles:
|
|
|
|
 
S. D. Logsdon, G. Hernandez-Ramirez, J. L. Hatfield, T. J. Sauer, J.H. Prueger, and K. E. Schilling
Soil Water and Shallow Groundwater Relations in an Agricultural Hillslope
Soil Sci. Soc. Am. J.,
July 14, 2009;
73(5):
1461 - 1468.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. W. Donohue, M. H. Stolt, A. Gold, and P. Groffman
Human-Transported Material Soils of Urbanizing Estuarine Landscapes
Soil Sci. Soc. Am. J.,
July 14, 2009;
73(5):
1587 - 1596.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. P. Morgan and M. H. Stolt
Using Hydrologic Patterns and Precipitation Data to Construct an Empirical Model for Understanding Cumulative Saturation
Soil Sci. Soc. Am. J.,
March 1, 2009;
73(2):
598 - 604.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. P. Morgan and M. H. Stolt
Soil Morphology-Water Table Cumulative Duration Relationships in Southern New England
Soil Sci. Soc. Am. J.,
March 29, 2006;
70(3):
816 - 824.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
