There have been repeated requests for “periodic, high-quality, nonpartisan reporting on the condition and use of U.S. ecosystems, the goal being a stable set of broadly accepted and well-tested indicators” (H. John Heinz III Center for Science, Economics and the Environment, 2008). In particular, reporting on the Nation’s major aquifers in which water levels are declining, increasing, or stable is one of the 10 highest priority data gaps identified for reporting on the Nation’s ecosystems (H. John Heinz III Center for Science, Economics and the Environment, 2006). At the present time, there are no regularly-updated indicators for the Nation’s groundwater resource. One possible approach for a national indicator is to use composite water-level hydrographs of the Nation’s Principal Aquifers.
A composite water level is an average water level calculated from selected wells. The composite groundwater level uses the median water level for the period of interest (either annual or monthly) for each index well and then averages all the index wells for the particular year or month to calculate a composite water level. The composite water levels representing the average or mean water level of all the index wells is then presented on a hydrograph.
As an example, an average annual groundwater level composite (figure 1) for 14 key observation wells in the uppermost unconfined aquifer was used to show the drought of the mid-1960s on Long Island by Cohen, Franke, and McClymonds (1969). This composite groundwater hydrograph clearly shows the effect of the drought in the mid-1960s that occurred on Long Island, NY.
Figure 1 – Average annual groundwater levels in 14 key observation wells on Long Island, NY (from Cohen, Franke, and McClymonds, 1969, figure 6)
Composite average monthly groundwater levels have also been calculated (Franke and McClymonds, 1972) as shown in figure 2. The monthly composite hydrograph shows the annual cycles that occur in addition to the long term trend.
Figure 2. Composite average monthly groundwater levels in selected wells and composite monthly discharge of selected streams in Nassau County for the period of 1940-1950 (from Franke and McClymonds, 1972)
The composite hydrographs tend to smooth out the effects of local or random fluctuations in the record and present an “average” response of the area represented by the “index” wells selected to be included in the composite. Busciolano (2005, p. 29) stated the purpose of his use of composite hydrographs on Long Island, NY as:
“These composites were created to simplify comparison among data sets, to minimize the effects of local fluctuations in the record, and to present the relative condition of Long Island’s water-level conditions in an easy-to-view format.”
Composite hydrographs have been used throughout the Nation. The most readily available references are for Long Island, NY, but average water levels (or composite water levels) have been used to evaluate many other areas, for example, areas in the High Plains aquifer of Kansas (Woods and others, 2000), areas in Monterey County, CA (Monterey County Water Resources Agency, 1997), and areas in San Benito County Water district, CA (Todd Groundwater, 2014).
An important step in creating a composite hydrograph is the identification of the “index” wells to be used in calculating the average composite water level. What is the reason for grouping the specific wells into a composite? The “index” wells should all come from an area that is similar in some hydrologic condition so that it can be used as a representation of that hydrologic condition. It takes a knowledgeable individual to select the appropriate “index” wells.
The audience for the composite hydrograph is twofold. The first audience is the knowledgeable hydrologist who understands the strengths and weaknesses of the approach and can supply additional hydrologic insight into their interpretation. The second audience is those individuals who do not have a technical hydrologic background but understand the general implications of the trends shown in the composite hydrographs.
As stated above, the “index” wells used in the calculation of the composite water level hydrograph should all come from an area that is similar in some hydrologic condition. In order to develop an indicator of groundwater levels in the United States, composite hydrographs have been generated for the Principal Aquifers of the United States (http://groundwaterwatch.usgs.gov/Composite.asp) using the data available in USGS Groundwater Watch (http://groundwaterwatch.usgs.gov/).
The purpose of the composite hydrographs of the Principal Aquifers of the United States is to show the “average” response of water levels in the Principal Aquifer over time. This should indicate if there is ongoing depletion or show longer term climatic changes in each Principal Aquifer. To this end, Composite hydrographs were calculated based on annual median measurements from each index well.
In order to produce composite water levels, decisions had to be made regarding the period of record, the amount of missing record allowable, whether to use the mean or the median, whether to use below land surface (BLS) or above sea level (ASL) as the variable of interest, and other issues.
The period of record used for this analysis is a moving 30-year hydrograph from the most recent complete year of record. The value used for each year for an index well is the median water level.
Many wells do not have a measurement in every year for the 30 year period of record. These missing measurements cause noise in the data when the water levels are averaged to generate a composite hydrograph. Exploratory data analysis was conducted to evaluate the allowable amount of missing record that would minimize the amount of noise in the composite hydrograph and maximize the number of index wells available. After consideration of the results, it was decided to use only index wells with no missing record. That is, every index well has at least one measurement for every year in the 30 year period. This minimizes the noise in the composite hydrographs.
An additional requirement for the calculation of a composite hydrograph for a Principal Aquifer is that there must be at least 9 index wells that meet the period of record criteria, and those 9 wells must have a reasonable areal distribution throughout the Principal Aquifer. The reasonable areal distribution was defined qualitatively. At the present time, this approach limits the composite hydrographs that are available. But over time, more and more wells will meet the criteria and the population of available composite hydrographs will increase. Table 1 provides a list of the number of index wells available, as of June 2016, for each Principal Aquifer for the 30 year period. Cells highlighted in green indicate that there are enough index wells available to calculate a composite hydrograph. In addition, two other Principal Aquifers, the Alluvial Aquifers and the High Plains Aquifer, were excluded because 1) the Alluvial Aquifers are discontinuous aquifers covering the entire United States and one composite hydrograph would not be meaningful and 2) the High Plains Aquifer well distribution was clustered in Nebraska and additional resources from this intensively studied aquifer are available to determine trends.
Table 1. List of Principal Aquifers and the number of wells with a one-hundred percent complete period of record available for a 30 year composite hydrograph. Composite hydrographs are computed for the aquifers highlighted in green.
For each index well, the variable calculated for each year is the median water level for the year. For annual composite hydrographs using continuous data, the median is calculated from all daily values and periodic measurements in the year. For annual composite hydrographs with only periodic data, the annual median is the median of all periodic measurements available for the year. For example, if there is only one measurement in a year, that measurement is the annual median water level. If there are 3 periodic measurements in a year, then the annual median water level is the median of the 3 periodic measurements. Thus, the annual median water level is always the median of available daily values and periodic measurements.
For the composite hydrograph, the median water levels from each index well are used to calculate both a mean composite water level and a median composite water level. For a specific year, the median water levels for each index well are averaged to compute the mean composite water level and the median of the medians for all index wells is used as the median composite water level.
Most water level measurements in Groundwater Watch are reported as Depth Below Land Surface Datum. Some locations, however, report water levels as Elevation Above Sea Level. Either variable can be plotted to make a composite hydrograph. In most instances the depth below land surface has a smaller range of values for an entire Principal Aquifer. If water levels above sea level are used for areas with large elevation changes, the larger values may bias the mean water level calculations. For the composite hydrographs presented in Groundwater Watch, all water levels are presented as Below Land Surface. Thus, water levels reported as Above Sea Level have been converted to Below Land Surface values.
Data from selected wells in the Mississippi Embayment aquifer system illustrates the calculation and plotting of the annual composite hydrographs. Figure 3 shows the data table for the 30 year annual composite hydrograph. The data table has 34 index wells. The index wells were selected from all wells in Groundwater Watch that were identified in the Mississippi Embayment Principal Aquifer and had at least one measurement in each of the 30 years (a 100 percent complete period of record). Each horizontal row represents an index well. The annual median water levels are presented under each year. Approved and provisional data are included in the statistic. The bottom two rows are the calculated annual composite mean and median water levels for each year. The data table is available from the individual Principal Aquifer Composite Hydrograph web page by selecting the “View Source Data” button at the bottom of the page.
Figure 3. The data table for the 34 index wells identified for the 30 year annual composite hydrograph.
Figure 4 is the hydrograph of mean and median composite water levels that were calculated in the bottom two rows of the data table. A map of the location of the 34 index wells is shown in figure 5. In addition, a “Google Earth” map of the index well locations is available by selecting the “View Google Earth Map” at the bottom of an individual Principal Aquifer Composite Hydrograph web page.
Figure 4. A 30 year annual composite hydrograph of water levels below land surface for the Mississippi Embayment Principal Aquifer. The green line shows the mean values and the red line shows the median values.
Figure 5. Map of the location of the 34 wells used to calculate the 30-Year Annual Composite Hydrograph in Figure 4.
Figure 6 shows a zoomed in portion of the data table shown in figure 3. Selection of any index well site ID by clicking on it will direct the user to the Groundwater Watch page for that well. Hovering over the icon of a well hydrograph under the “Well Info” column shows the hydrograph for that particular index well in a pop-up window. If the BLS (which stands for Below Land Surface) is highlighted in yellow under the “Well Info” column, it indicates that the water levels were converted from Above Sea Level (ASL) measurements and the conversion equation is shown by hovering over the yellow highlighted BLS. The colored symbol under the “Well Info” column refers to the type of well and the percentile group for the latest measurement compared to all historical measurements in the current month. An explanation of the symbol shapes and colors is provided on the bottom of figure 5.
Data that were used to calculate the median value for any well and year can be obtained by selecting (clicking on) the median value of interest in the table. Median values that were calculated using provisional data will be indicated by a red “P” next to the value.
Figure 6. Zoomed in portion of the data table for the Mississippi Embayment Aquifer System.
The only criteria for the hydrographs as shown are that the well is identified in a Principal aquifer and the well record has at least one measurement in each of the previous 30 years (a 100 percent complete period of record). The advantage of this approach is that there is no need to individually vet each potential well in the aquifer system, the selection is automatic. The composite hydrographs speak for themselves. In one sense the hydrographs are unbiased.
The disadvantage of this approach is that the distribution of wells may not be evenly spaced, and multiple wells in one area may have undue influence on the composite hydrograph. In addition, some of the wells might be monitored for specific stresses, which might not be appropriate to represent the aquifer in general and thus may not provide an index of the entire aquifer.
Nevertheless, the composite hydrographs generated provide a broad overview of water levels in the Principal Aquifer. As such they are useful indicators of the trends in the Principal Aquifers.
A composite water level is an average water level calculated from selected wells. The composite groundwater level uses the median water level for each year for each index well and then averages all the index wells for the particular year to determine a composite water level. The composite water levels representing the average or mean water level of all the index wells is then presented on a hydrograph.
Composite groundwater level hydrographs have been calculated for the Principal Aquifers of the United States. The length of the composite hydrograph is 30 years. The composite hydrographs generated provide a broad overview of water levels in the Principal Aquifer. As such they are useful indicators of the trends in the Principal Aquifers.
Busciolano, R., 2005, Statistical Analysis of Long-Term Hydrologic Records for Selection of Drought-Monitoring Sites on Long Island, New York: U.S. Geological Survey Scientific Investigations Report 2005-5152, 14 p. (available online at http://pubs.er.usgs.gov/publication/sir20045152)
Cohen, Philip, Franke, O. L., and McClymonds, N. E., 1969, Hydrologic effects of the 1962-66 drought on Long Island, New York: U.S. Geol. Survey Water-Supply Paper 1879-F, 18 p. (available online at http://pubs.er.usgs.gov/publication/wsp1879F)
Franke, O.L., and McClymonds, N.E., 1972, Survey of the hydrologic situation on Long Island, N.Y, as a guide to water Management alternatives: U.S. Geological Survey Professional Paper 627-F, 59 p. (available online at http://pubs.er.usgs.gov/publication/pp627F)
H. John Heinz III Center for Science, Economics and the Environment, 2006, Filling the gaps—Priority data needs and key management challenges for national reporting on ecosystem condition: Washington, D.C., H. John Heinz III Center for Science, Economics and the Environment.
H. John Heinz III Center for Science, Economics and the Environment, 2008, Highlights The State of the Nation’s Ecosystems 2008: The H. john Heinz III Center for Science, Economics and the Environment, Washington, D.C., 34 p. (available online at http://www.ksg.harvard.edu/index.php/content/download/76058/1709084/version/1/file/Heinz_Center_2008_Highlights_State_of_the_Nations_Ecosystems.pdf)
Monterey County Water Resources Agency, 1997, Water Resources Data Report – Water Year 1994-1995: Monterey County Water Resources Agency, Salinas, CA, 70 p.
Todd Groundwater, 2014, San Benito County Water District – Annual Groundwater Report: San Benito County Water District, Hollister, CA, 134 p.
Woods, J.J., Schloss, J.A., and Macfarlane, P.A., 2000, January 2000 Kansas Water Levels and Data Related to Water-Level Changes: Kansas Geological Survey Technical Series 15, 21 p.