Volume 43, Issue 9
Regular Article
Free Access

Long-term water chemistry database, Little River Experimental Watershed, southeast Coastal Plain, United States

G. W. Feyereisen

G. W. Feyereisen

Southeast Watershed Research Laboratory, Agricultural Research Service, U.S. Department of Agriculture, Tifton, Georgia, USA

Search for more papers by this author
R. Lowrance

R. Lowrance

Southeast Watershed Research Laboratory, Agricultural Research Service, U.S. Department of Agriculture, Tifton, Georgia, USA

Search for more papers by this author
T. C. Strickland

T. C. Strickland

Southeast Watershed Research Laboratory, Agricultural Research Service, U.S. Department of Agriculture, Tifton, Georgia, USA

Search for more papers by this author
J. M. Sheridan

J. M. Sheridan

Southeast Watershed Research Laboratory, Agricultural Research Service, U.S. Department of Agriculture, Tifton, Georgia, USA

Search for more papers by this author
R. K. Hubbard

R. K. Hubbard

Southeast Watershed Research Laboratory, Agricultural Research Service, U.S. Department of Agriculture, Tifton, Georgia, USA

Search for more papers by this author
D. D. Bosch

D. D. Bosch

Southeast Watershed Research Laboratory, Agricultural Research Service, U.S. Department of Agriculture, Tifton, Georgia, USA

Search for more papers by this author
First published: 07 September 2007
Citations: 15


[1] A water quality sampling program was initiated in 1974 by the U.S. Department of Agriculture Agricultural Research Service on the 334 km2 Little River Experimental Watershed (LREW) near Tifton in south Georgia to monitor the effects of changing land use and agricultural practices over time and to support development of simulation models capable of predicting future impacts of agricultural land use and management changes. Stream samples were taken on a weekly or more frequent basis and were analyzed for chloride, ammonium nitrogen, nitrate plus nitrite nitrogen, total kjeldahl nitrogen, total phosphorus, and dissolved molybdate reactive phosphorus. Monitoring began in 1974 on the entire watershed and four nested subwatersheds, ranging in size from 16.7 to 114.9 km2, and continues until present. Partial records of 7, 10, and 19 years exist for three additional subwatersheds. Suspended solids data are available for all eight subwatersheds for 1974–1978 and 1979–1981, three subwatersheds for 1982–1986, and all eight subwatersheds again beginning in the year 2000. The concentration and associated load data are being published on the LREW database anonymous ftp site (ftp://www.tiftonars.org/).

1. Introduction

[2] Hydrologic and water quality monitoring have been in progress on the 334 km2 Little River Experimental Watershed (LREW) in the western headwaters of the Suwannee River since the late 1960s [Asmussen et al., 1975; Yates, 1976; Mills et al., 1984]. Portions of the LREW water chemistry record have been published previously [Lowrance et al., 1985; Sheridan and Hubbard, 1987; Lowrance and Leonard, 1988; Hubbard et al., 1990]; however, it has been over a decade since nutrient concentrations and loads have been published in the scientific literature. The purpose of this paper is to document the water chemistry data set for the first 30 years of record for the LREW and the ftp site address where these unique databases may be accessed.

2. Little River Experimental Watershed

[3] The Little River Experimental Watershed (LREW) is a 334 km2 area that stretches from the headwaters of the Little River, located approximately 9 km west of Ashburn, Georgia, southeast approximately 36 km to 6 km west of Tifton, Georgia [Bosch et al., 2007, Figure 1]. The LREW is located in the Tifton-Vidalia Upland of the Gulf-Atlantic Coastal Plain. Water chemistry data are available for the eight nested stream gauges located within the LREW [Bosch et al., 2007, Figure 2]. Table 1 contains characteristics of each of the subwatersheds. The first full year of a water quality monitoring program was 1974.

Table 1. Characteristics of Each Subwatershed in the LREW and Duration of Record for Nutrients and Solids
Subwatershed Area, km2 Elevation Range, m Sampling Point UTM Zone 17 Years of Record: Nutrients Years of Record: Solids
B 334.3 82–146 254518E, 3485875N 1974–present 1974–1978, 1979–1981, 1982–1986, 2000–present
F 114.9 92–146 250365E, 3499631N 1974–present 1974–1978, 1979–1981, 1982–1986, 2000–present
I 49.9 102–146 244933E, 3507512N 1974–present 1974–1978, 1979–1981, 2000–present
J 22.1 105–146 243864E, 3509522N 1974–present 1974–1978, 1979–1981, 2000–present
K 16.7 106–146 244320E, 3509946N 1974–present 1974–1978, 1979–1981 1982–1986, 2000–present
M 2.6 123–146 241910E, 3514711N 1982–1986 2002–present 1974–1978, 1979–1981, 2002–present
N 15.7 87–123 254403E, 3489904N 1974–1981 2002–present 1974–1978, 1979–1981, 2002–present
O 15.9 86–122 256119E, 3487152N 1974–1981 1993–present 1974–1978, 1979–1981, 2000–present

[4] Agricultural land use in the LREW affects water quality [Lowrance et al., 1984]. Land cover over the entire LREW has been classified recently as 50% forest, 41% mixed agricultural, 7% urban, and 2% water [Bosch et al., 2006]. On the basis of satellite imagery, the range of estimates for forest cover was 40 to 56% and for mixed agricultural land use was 31 to 52% over the past three decades [Bosch et al., 2006]. Forested lands consist of pine on the upland areas and hardwoods in the dense riparian vegetation in the flat, broad swamp areas. Agricultural cropping rotations have changed over the decades from peanut, corn, soybean, winter wheat, and tobacco, to cotton, peanut, and vegetable crops. Beef cattle are sparsely distributed across the watershed on pastures. There are two chicken (broiler) growers in the watershed, two dairies, and a few small swine operations [Suttles et al., 2003]. There are no permitted point source discharges in the watershed.

3. Period of Collection

[5] Continuous water quality sampling in the LREW began in 1974 with monitoring of in-stream chloride, nitrate plus nitrite nitrogen, and dissolved molybdate reactive phosphorus concentrations on seven of the eight subwatersheds. In 1979, ammonium, total kjeldahl nitrogen, and total phosphorus were added to the list of analytes monitored. Sampling of these six analytes has continued using various methods and frequencies, depending upon specific research objectives, until the current time. Table 1 shows the period of record for each subwatershed for the nutrients analyzed. The intermittent record for the three smallest watersheds is attributable to shifts in research priorities and budgetary constraints.

[6] Collection of suspended solids data in the LREW has been intermittent since 1974. Studies were published for records spanning the time frames of: August 1974 to August 1978 on total solids and January 1979 to April 1981 on suspended solids [Sheridan and Hubbard, 1987]; and January 1984 to March 1986 on suspended plus dissolved solids [Hubbard et al., 1990]. Monitoring of suspended solids concentrations was reinitiated in January 2000 on subwatersheds K, J, I, F, O, and B and in January 2002 on subwatersheds M and N, and continues to present.

4. Data Collection Methods

[7] The methods of sample collection were modified over the 30-year period, reflecting changes in research needs, specific study objectives, laboratory funding levels, and technological advances. Samples were taken weekly or more often by one of the following methods: manual grab (Grab); automated, timed, discrete (ATD); automated, flow weighted composite, nonrefrigerated (AFCN); and automated, flow weighted composite, refrigerated (AFCR). Table 2 identifies the timeframes the various sampling methods were used on each subwatershed.

Table 2. Dates of Various Sampling Methods for Each Subwatershed in the LREW
Subwatershed Grab Automated, Timed, Discrete Automated, Flow Composite, Nonrefrigerated Automated, Flow Composite, Refrigerated
B 28 Jan 1974 to 12 Jan 1995 20 Jan 1995 to 30 Dec 2002 6 Jan 2003 to present
F 1 Feb 1974 to 26 Mar 1982, 5 Dec 1986 to 12 Jan 1995 2 Apr1982 to 2 Apr 1986 20 Jan 1995 to 29 Dec 2003 5 Jan 2004 to present
I 24 Jan 1974 to 2 Apr 1982, 12 Dec 1986 to 12 Jan 1995, 14 Jan 2002 to 31 Mar 2003 6 Apr 1982 to 30 Mar 1986 20 Jan 1995 to 29 Dec 2003 5 Jan 2004 to present
J 24 Jan 1974 to 18 Aug 1978, 12 Jan 1990 to 30 Dec 1994 28 Dec 1978 to 2 Jan 1990 20 Jan 1995 to 30 Dec 2002 6 Jan 2003 to present
K 24 Jan to 15 Aug 1974, 2 Feb 1993 to 12 Jan 1995 Aug 21 1974 to 26 Mar 1993 20 to Jan 1995 to 30 Dec 2002 6 Jan 2003 to present
M 15 Aug 1974 to 28 Dec 1977, 25 Jan 1979 to 24 Apr 1981, 8 Jan 1982 to 9 Apr 1982, Dec 1986, 7 Jan 2002 to 29 Apr 2002 13 Apr 1982 to 2 Apr 1986 6 May 2002 to present
N 28 Jan 1974 to 14 Dec 1978, 7 Jan 2002 to 29 Apr 2002 22 Dec 1978 to 25 Dec 1981 6 May 2002 to present
O 28 Jan 1974 to 22 Dec1978, 11 Feb 1993 to 12 Jan 1995 23 Dec 1978 to 30 Dec 1981 20 Jan 1995 30 Dec 2002 6 Jan 2003 to present

[8] Manual grab samples during the 1974–1978 period were collected from the flow over the weir. Manual grab samples collected after that were obtained from the stilling area immediately upstream of the control weir at each gauging station. All pumped samples were also taken from the stilling area immediately upstream of the control weir at the gauging stations. Analyte concentrations measured for the manual grab and timed discrete pumped samples represent instantaneous concentrations. The ATD samples were selected for analysis depending on the goals of particular studies. For 1979–1986, ATD samples were collected at 12 hour intervals. All samples during a storm event (rising or falling limb of hydrograph) were analyzed. Selected base flow samples were analyzed [Lowrance et al., 1985; Lowrance and Leonard, 1988]. For one study period (19 December 1985 to 27 February 1986), ATD samples were taken every 3 hours on subwatersheds J and I and all samples were analyzed [Lowrance and Leonard, 1988]. Additional details of ATD sampling methods have been published by Lowrance and Leonard [1988]. Automated, flow weighted composite samples have been obtained by programming automated water quality samplers to draw sample quantities at intervals of equal streamflow volume. The composite samples are collected weekly. Each composite sample consists of 14 to 180 subsamples depending on the flow rate. Within a week, as flow increases, pumping frequency increases. Thus the automated, flow weighted composite samples represent the mean concentration over the sampling period. The automated, flow weighted composite samples are currently maintained in small refrigerators housed in secured and ventilated structures that have been installed at the gauging stations over the years 2002 to 2004.

[9] Concentrations of dissolved NO3 + NO2–N, NH4–N, Cl, and dissolved molybdate reactive P (DMRP) were determined using EPA approved colorimetric techniques [American Public Health Association, 1976; Clesceri et al., 1998]. Total kjeldahl N and total P were determined on digestates of unfiltered samples [Technicon Industrial Instruments, 1977; Lachat Instruments, 1997]. From the beginning of the record period through 1986, these analyses were conducted on a Technicon Autoanalyzer II instrument. Beginning in 1987 through the present, these analyses were conducted using Lachat flow injection analyzers.

[10] Details of the sediment sample collection and analysis methods were published for the 1974–1978 and 1979–1981 time periods by Sheridan and Hubbard [1987], and for the 1982–1986 time period of Hubbard et al. [1990].

[11] Under the current measurement regimen suspended solids were determined by passing a sample through a Whatman 934AH (0.45 μm effective pore size) filter and measuring the oven-dry weight of the material trapped on the filter. The concentration of suspended solids, which includes organic and inorganic constituents, was calculated by dividing the mass of material by the sample volume from which it was filtered, and was expressed in units of mg L−1.

5. Concentrations and Loads

[12] Nutrient concentrations recorded for each sample test and suspended solids concentrations since January 2000 are available on the LREW database anonymous ftp site (ftp://www.tiftonars.org/). Streamflow nutrient loads were calculated by summing the product of the nutrient concentration and volume of streamflow corresponding to the sampling period. Nutrient loads for each subwatershed and suspended sediment loads since January 2000 are available in eight separate files. Nutrient loads have been estimated for each day to allow direct comparison to watershed simulation modeling. Assumptions were made about the stream analyte concentrations on nonsample dates based upon the sample type. Grab samples were assumed to represent stream concentrations for the latter half of the preceding sampling interval and the first half of the succeeding sampling interval. The automated, flow weighted composite, nonrefrigerated or refrigerated (AFCN/R) samples represent the flow weighted mean concentration over a sampling period, beginning with the day following the previous sample date and ending with the sample date. The sampling intervals for the automated, timed discrete (ATD) sampling regime ranged from a few days to three hours. The daily loads reported for periods when the sampling interval was greater than 1 day, e.g., grab sampling at 1-week intervals and weekly composite samples, are quasi-daily loads. Because of the dependence of loads on streamflow volumes, comparison of simulated and observed hydrology will largely govern model goodness of fit for loads.

[13] The file “streamchemistry_readme.doc” describes the method used for filling missing concentration values. Table 3 contains the percent missing concentration readings by subwatershed for each analyte. Table 4 contains the percent of calculated load by subwatershed for each analyte that resulted from a missing laboratory concentration value that was estimated.

Table 3. Percent Missing Concentration Values
Analyte Percent of Sample Concentrations Missing, %
All B F I J K M N O
Chloride 1.29 0.54 0.48 0.86 1.33 0.86 0.00 3.69 3.48
Ammonium nitrogen 3.20 2.09 2.41 2.31 2.84 1.80 2.38 10.05 7.09
Nitrate plus nitrite nitrogen 1.41 1.15 1.01 0.64 1.43 0.86 0.00 3.57 3.84
Total Kjeldahl nitrogen 2.83 1.32 2.10 2.66 3.57 2.04 2.54 4.87 4.38
Total phosphorus 4.89 2.09 5.99 5.27 5.48 4.14 7.76 4.87 4.29
Ortho phosphate 3.82 1.81 5.11 4.35 4.14 2.78 7.89 3.69 3.55
All 2.85 1.47 2.80 2.64 3.10 2.01 3.43 4.93 4.38
Table 4. Percent of Loads Calculated Using Concentrations Filled for Missing Values
Analyte Percent of 30-Year Load Calculated Using Estimated Concentrations for Missing Values, %
All B F I J K M N O
Chloride 0.27 0.15 0.44 0.50 0.91 0.02 0.00 0.00 0.00
Ammonium nitrogen 0.57 0.38 0.46 2.04 1.16 0.07 4.33 0.52 0.30
Nitrate plus nitrite nitrogen 0.10 0.05 0.24 0.00 0.28 0.00 0.00 0.00 0.01
Total Kjeldahl nitrogen 0.61 0.54 0.79 0.47 1.20 0.14 5.77 0.03 0.56
Total phosphorus 2.37 1.24 4.11 3.33 8.35 0.18 15.11 0.07 0.78
Ortho phosphate 2.33 1.45 3.02 6.33 3.71 0.06 18.83 0.00 0.36
All 0.34 0.22 0.54 0.55 1.04 0.04 0.80 0.00 0.06

6. Suspended Solids Concentrations and Loads: Older Studies

[14] Suspended solids concentrations are available for each sample test during the three older studies, 1974–1978, 1979–1981, and 1982–1986 [Sheridan and Hubbard, 1987; Hubbard et al., 1990]. The solids loads for all subwatersheds are also available. Solids loads were calculated for the various sampling regimes in the same manner that the nutrient loads were calculated. Grab sample concentrations were used to calculate loads for the days one half sampling interval prior to and after the sample date. Concentrations obtained from ATD samples taken at a daily interval were multiplied by daily streamflow. Subdaily ATD sample concentrations were averaged to obtain a daily concentration value, then multiplied by daily streamflow. If concentrations were missing for all fractions of a sample taken on a day with multiple samples, the sample was deleted from the record. Automated, timed discrete sample concentrations taken at an interval greater than 1 day were averaged to fill the nonsample days. The types of sampling used to obtain solids concentrations are identified in Table 2. Load data are presented in daily format as a convenience, but in most cases represent quasi-daily loads because the sampling intervals were routinely longer than 1 day.

7. Data Availability

[15] The data presented in this paper are available on the anonymous ftp site: ftp://www.tiftonars.org/. The site is maintained by the USDA-ARS Southeast Watershed Research Laboratory in Tifton, Georgia. Concentration and load data for a given year are typically available 2 years after having been collected. More recent data may be made available to researchers collaborating with a USDA-ARS Southeast Watershed Laboratory scientist.


[16] Many persons at the Southeast Watershed Research Laboratory have contributed their expertise and time to the collection and analysis of water chemistry data. We express our gratitude and thanks for their dedication and contribution to this long-term database. The current employees who provide the support for collection of these nutrient and suspended solids data are Leila Hargett, Chemist; Chris Clegg, Physical Science Technician; and Rex Blanchett, Physical Science Technician.