Top Navigation Bar

ASAE Journal Article

Design and Evaluation of a Discreet Sampler for Waste Treatment Lagoons

J. J. Classen, J. M. Rice, J. P. McNeill, O. D. Simmons III


Published in Applied Engineering in Agriculture Vol. 27(6): 1007-1014 ( Copyright 2011 American Society of Agricultural and Biological Engineers ).

Submitted for review in January 2011 as manuscript number SE 9027; approved for publication by the Structures & Environment Division of ASABE in June 2011.

The authors are John J. Classen, ASABE Member, Associate Professor, J. Mark Rice, ASABE Member, Extension Specialist, John P. McNeill, former Research Assistant, and Otto D. Simmons, III, ASABE Member, Research Assistant Professor, Biological and Agricultural Engineering Department, North Carolina State University, Raleigh, North Carolina. Corresponding author: John J. Classen, Biological and Agricultural Engineering Department, Campus Box 7625, North Carolina State University, Raleigh, NC 27695; phone: 919-515-6800; e-mail: John_classen@ncsu.edu.


Abstract. A lagoon sampler was designed to collect 300-mL samples from depths up to 2.82 m at 15.2-cm intervals. Design parameters required the device to collect discrete samples through the entire lagoon column without cross contamination during raising or lowering of the device through the lagoon and to leave the liquid column undisturbed so subsequent samples could be taken from lower depths at the same location. An evaluation protocol was developed to test the device in a 55.9-cm diameter, 3.35-m tall PVC test column. This test column was of sufficient depth to be representative of lagoons and of sufficient diameter to avoid interference, or "edge effects," associated with the sampling device. The evaluation protocol used dissolved solids (sodium chloride) as the test parameter in 30.5-cm deep layers of sodium chloride of decreasing concentrations to fill the test column from the bottom and simulated different solids concentrations that may be found in a lagoon. Samples were collected at six depths from 122 to 198 cm. Based on electrical conductivity measurements, samples collected with the device were slightly more dilute than the expected value at all sample depths. Further tests showed that operation of the device did not cause mixing of the layered salt solutions. The sampler was used to collect samples from various depths in a lagoon; results suggested there was little interference among samples taken at specific locations through subsequent depths.

Keywords. Sampler, Lagoon, Solids, Protocol.

Anaerobic lagoons are often used for the storage and treatment of manure and wash water from animal production facilities, especially swine operations. If properly designed, constructed, operated, and maintained, anaerobic lagoons are inexpensive and effective systems that stabilize organic matter and reduce odors (Lorimor et al., 2001). One of the most difficult tasks in lagoon management is removing sludge. Current practice for sludge removal is agitation of the lagoon to mix the sludge with the overlying liquid fraction followed by pumping the majority of the contents to tanker trucks and removal off site for land application. However, some operators have complained that after such a sludge removal operation, their lagoon seems to emit more odors which elicit more complaints from neighbors.

An effort to understand the complex microbial communities in lagoons found evidence of different microbial populations at the surface, at mid-depth, and in the sludge layer of lagoons (Devine et al., 2006). Although there was variation over time within each lagoon, these researchers also found less diversity in lagoons that were more odorous, an indication of poor performance (Devine et al., 2006). These results suggested that additional research was needed in an effort to understand the linkages among lagoon performance, sludge management, and microbial diversity but the sampler was not able to obtain the additional undisturbed samples needed from additional depths in the lagoon.

The sampler used in the Devine et al. (2006) study was built in the Biological and Agricultural Engineering (BAE) Department at North Carolina State University (NCSU) and was based on the Kroes-Barth-Dodd (KBD) sampler (Kroes et al., 1987). The sampler was intended to collect discreet samples from different depths throughout the liquid column but demonstrated some problems during operation. On occasion, the sample entry port would not close completely, either allowing lagoon material to enter before the proper depth was reached or allowing sample to exit the port before returning to the surface. The sample chamber was not sealed at the top such that liquid would continue to flow into the tubular handle of the device until the port was closed or hydraulic pressure inside the device equaled that outside the device.

We undertook a redesign of the sampler to address these issues and to meet additional specifications suggested as shortfalls by a team of microbiologists. The sample needed to be as small as possible but at least 275 mL; the device needed to be easily maneuvered to a specific depth during sample collection; and column disturbance needed to be minimized such that samples could be taken from multiple depths at each location within the lagoon. We also developed a testing protocol to verify the performance of the sampler. The redesign and testing are described below.

Literature Review

The work by Kroes et al. (1987) reviewed prior literature and concluded that no device available at the time was well suited to collect samples of thick material from very near the bottom of lagoons. Additionally, no evidence was found of samplers suitable for collecting samples of liquid and sludge throughout the entire depth of anaerobic treatment lagoons.

Hamilton and Rosser (1994) described design improvements to the KBD sampler, including a modular design that was easy to disassemble for cleaning and an internal seal to limit sample size. Although this design was used successfully in their research, the large sample size and reliance on springs to seal the sample chamber were complicating design aspects that we wanted to avoid in our sampler.

Parker et al. (1999) collected stratified samples from a cattle feedlot runoff storage pond by putting dry ice and isopropyl alcohol in a wide, thin container, and pushing the device into the bottom of the storage pond. The sample froze to the device and could be removed undisturbed. However such a device is not practical for deep storage or treatment units like lagoons because the freezing would start as soon as the device entered the water rather than at the desired sampling point. An additional downfall of this concept specifically related to microbial sampling is that freezing could impact different organisms in different ways, thus providing inaccurate information about the community structure.

Anderson et al. (2000) developed a sampling device to collect 10 discrete sludge samples at the same time. The device used a 3-m long and 5-cm diameter PVC column with sample ports placed at 0.3-m intervals. Although the device was used successfully in their study, it is limited to predetermined sample intervals and the authors did not determine the effect of the sampler on the surrounding sludge material.

Fulhage et al. (2005) used a 7.6-cm steel tube sampler with hinged doors to collect a single intact sample of lagoon sludge up to 1.2 m from the lagoon bottom. The sample was sealed in the device by plugging the end with a section of the clay liner. After removal from the lagoon, the sample was accessed by opening the tube, allowing separate collection and analysis of sludge from different depths. While this method would work for a sample from the bottom of a clay-lined lagoon, it would not collect an intact sample from shallower depths in the liquid column. The device was described as heavy and "unwieldy" but provided "consistent" results, although no operational testing procedure was used to verify sample consistency (Fulhage et al., 2005). In fact, there appears to be no documentation of a testing protocol for any of these sludge sampling devices.

Some devices used for sampling soils were investigated such as one used for saturated organic soils described by Caldwell et al. (2005). Although the physical properties of these soils are somewhat similar to sludge from the bottom layers of lagoons, the solids content in lagoons is not sufficient to keep the sample structure intact using such devices.

Based on the literature review and our own experience with the early KBD sampler, we elected to build a new sampler from an updated design rather than modify an existing device. Our objective was to create a sampler that could collect discrete samples at approximately 15-cm intervals from animal waste lagoons up to 2.74 m in depth with minimal mixing of the lagoon material during sampling. Minimizing mixing with such a sampling device is a difficult proposition, especially for samples taken from the sludge layer. Since our goal was to collect individual samples throughout the liquid column, we needed only a single sample at each depth rather than multiple samples; therefore each subsequent sample would be collected at a greater depth than the previous one, minimizing the impact of disturbance.

In addition, we established a testing protocol against which the new device was evaluated. Collection of accurate discrete samples would be based on the difference in total dissolved solids (TDS) in samples collected with the new device and the known value of TDS in the column. The question of mixing by operation of the sampling device would be tested by comparing the value of TDS before and after using the sampling device. While TDS is not the primary parameter of interest in lagoon sampling plans, it is a parameter for which solutions of varying concentrations can be readily prepared and analyzed for use in a controlled evaluation protocol.

Materials and Methods

Sampler Design

The NCSU sampler is a discrete sampling device that appears similar to the KBD sampler (Kroes et al., 1987); an aluminum pole attached to a cylindrical PVC sample chamber with two opposing sample entry ports and a cone-shaped end cap (fig. 1). The sample chamber is of sufficient size to collect a 300-mL sample for microbial and nutrient analyses, while minimizing the volume and the associated disturbance of the liquid column. The physical size of the device was a tradeoff between a small diameter that would be easy to handle when lowered into a lagoon and a short chamber height that would collect a sample representative of a precise depth.

Details of construction are shown in figure 1. The sample chamber was constructed from 8.89-cm diameter PVC with an inside diameter of 8.26 cm. To meet the 300-mL design volume, the chamber height was set at 5.72 cm for an actual volume of 306 mL. A cone-shaped end cap, constructed of solid PVC material, was machined to seal the bottom of the sample chamber and extend 7.62 cm. The top of the sample chamber was also sealed with a machined piece of PVC. The two opposing sample entry ports were cut the full height of the sample chamber (5.72 cm) and 1.91 cm in width in order to minimize turbulence during sample collection. The chamber was attached to a 2.54 cm diameter aluminum handle 2.64 m long.

The sample chamber was sealed by a concentric PVC pipe 9.52 cm tall sliding over the sampler body and overlapping the end of sample chamber by 1.91 cm. A tight clearance between the cover and sampler (0.056 cm) in conjunction with a foam gasket ensured a tight seal and sample integrity (fig. 1).

The sliding cover was controlled by two metal bars attached to a concentric tube with the handle supporting the sample chamber. The user raises and lowers this sliding tube using the metal handles to open and close the sample chamber. The sampler has a mass of 5.90 kg.

SE9027_files/image1.gif

Figure 1. NCSU sludge sampling device and sliding cover; measurements are in mm.

Sampler Seal Evaluation

Initial laboratory testing of the sampler's seal was conducted, first with an empty sample chamber and then with a full sample chamber. The empty sampler was lowered to the bottom of a 1-m column of water, held for 15 s, and removed. In the second test, the sampler was lowered into the water and opened, allowing the chamber to fill, and then removed. After a 15-s delay, the chamber contents were collected and the volume measured.

Protocol Development

The two most important sampler design parameters are representative samples at the depth of interest and minimum disturbance of the lagoon column such that subsequent samples taken at the same location but at different depths are representative of the specific depths at which they are collected. This is a common need among lagoon samplers (Kroes et al., 1987; Anderson et al., 2000) but one that is difficult to verify in the laboratory. A literature review revealed no publications with a testing protocol so we developed a procedure to test both of these design parameters with respect to TDS in a controlled setting using a column of water with varying salt concentrations, "heaviest" to "lightest" from bottom to top of the column. Admittedly, TDS is not as important in lagoon sampling as total solids (TS) or suspended solids; however, with the lack of a well-defined source of suspended solids, we elected to test the device using the dissolved fraction. Since dissolved solids will more easily mix and form a single solution, this is a more rigorous test of column disturbance.

Lab Scale Proof of Concept

A laboratory-scale test was performed to verify that representative samples could be taken at specific depths without mixing the water column and without interference between the sampler and the column wall. A 1-L graduated cylinder was used to represent the liquid column and a pipette was used to represent the sampler. While no laboratory system can duplicate the conditions of the field test, in this proof of concept, we relied on a comparison of the ratio of diameters of the pipette to the graduated cylinder (0.164) and the sampler and column (0.159) to represent the potential for interference and wall effects. Since these ratios were similar, if samples from the graduated cylinder could be taken with the pipette without turbulence, the data from the lab test would be considered to represent the potential for the field test to provide a reasonable test of the device. The procedure was developed in the laboratory using solutions of sodium chloride of zero (deionized water), 100 parts per thousand (ppt, mass basis) and 200 ppt in the one liter graduated cylinder. Red, yellow and blue food coloring was added to each of these solutions, respectively; primary colors were purposefully used so any mixing would be seen as a distinct change in color. Prior to forming the test column, electrical conductivity (EC) (model 09-326-2, Control Company, Friendswood, Tex.) of each solution was measured and used as a surrogate for TDS, after which a peristaltic pump was used to add 300 mL of each solution to the 1-L graduated cylinder, with the 200-ppt solution on bottom, followed by the 100-ppt solution and the deionized water on top. The flow rate of the pump was sufficiently slow to avoid obvious turbulence but visual inspection showed a small amount of mixing of colors at the interface between the bottom two layers (fig. 2).

SE9027_files/image2.jpg

Figure 2. Laboratory column with layered salt solutions. The bottom layer, blue, had a concentration of 200 parts per thousand (ppt) sodium chloride; the middle layer, yellow, had a concentration of 100 ppt; the top layer, red, was tap water with no added sodium chloride.

A 10-mL plastic disposable pipette and electric pipette-aide were used to extract 10-mL samples from the midpoint of each layer in the column and from the interface between each layer. The pipette tip was placed at the appropriate location by hand and was carefully lowered with the changing liquid level. EC of the samples was measured and compared to the EC of the individual solutions taken prior to the formation of the column. After the sampling was complete, no further changes in color were seen, indicating that the sampling process could be completed without disturbing the liquid column and therefore verified the concept of sampling layered salt solutions as a legitimate test of the performance of the sampling device. The higher suspended solids content of lagoon liquid and sludge is expected to be less susceptible to disturbance than salt water solutions, thus making this a more challenging and rigorous test protocol for sampling devices used in lagoons, at least to the point where the sticky nature of sludge suspended solids causes them to adhere to the outside of the sampler.

Field Testing Protocol

As described earlier, sampler performance is based on collecting individual samples from specified depths while leaving the liquid column undisturbed such that additional samples can be taken from lower depths. The sampler volume is sufficient so that only a single sample is needed from each depth. Representative conditions of anaerobic lagoons cannot be reliably recreated in a laboratory setting but testing should be conducted in a column sufficiently large to avoid disturbance of the liquid column due to proximity of the sampling device to the column wall. Although lagoon depths vary, a column should be of sufficient depth to impose a pressure head similar to what the sampler will be exposed to in actual use.

The laboratory proof of concept was scaled into a testing protocol using a 55.9-cm diameter, 3.35-m tall PVC column. This provides a column of sufficient depth to be representative of lagoons in North Carolina and of sufficient diameter to avoid interference with the sampling device such as volume displacement, turbulence, or short circuiting of flow into the sample chamber. The ratio of the diameter of the sampler to that of the column was calculated as 0.159 (less than that of the lab scale proof of concept, 0.164). The bottom of the column was sealed with an end cap that was modified with a 1.9-cm drain and valve and set on a 30.5-cm cinder-block base to provide support and allow access to the drain valve. Tie down straps secured the column vertically against the side of a concrete platform which provided access to the top of the column for all testing procedures (fig. 3).

SE9027_files/image3.jpg

Figure 3. Column test of layered salt solutions.

Seventy-five liters of three different salt solutions (250, 160, and 75 ppt) were prepared on site in separate plastic containers and EC measurements were taken prior to their layering in the column. Although these concentrations exceed the range of TS concentrations expected in anaerobic swine lagoons (Bicudo et al., 1999), they were used to obtain significant concentration differences between layers so any cross contamination would be easily detectable. A sample of each solution was used to determine TDS for comparison to EC measurements. Salt concentrations were higher than in the laboratory test to accommodate an additional layer while maintaining significant concentration differences between layers and avoiding the upper limit of solubility of 264 ppt (Lide, 2005). Appropriate mass of salt was measured and added to each container; tap water was added and mixed by hand for approximately 30 min until all salt dissolved. Food coloring was again added to the solutions as an aid in visualizing the column contents as samples were retrieved. A peristaltic pump was used to slowly pump each solution into the test column, with the most concentrated solution on bottom. Each layer was approximately 30.48 cm thick with the actual depth being measured from the top of the column using a RocTest Model CPR Water Level Indicator (ST-Lambert, Quebec, Canada). The bottom layer was slightly thicker to account for the additional volume of the end cap. A layer of tap water (1.22 m) was then pumped into the column to create an overall depth of 2.13 m. The column was covered to prevent precipitation and contamination overnight; testing was conducted the following day.

Evaluation of the accuracy of the samples consisted of comparing EC of samples taken with the sampler to the actual EC of the water column at the same depths. Although the EC of each solution was measured before filling the column, a more rigorous test would be to obtain independent samples from predetermined sampling depths as a way to determine if any changes to salt concentration had occurred due to mixing or diffusion while forming the test column. Sampling depths were selected to correspond to the mid points of each layer and to the interfaces between the layers.

To measure the actual EC at these depths with the least disturbance and the smallest possible sample volume, six lengths of 0.32-cm Teflon tubing were installed around the perimeter of the column. The length of each tube was selected so the ends were at the desired sample locations. The tubing was suspended from a PVC collar at the top of the column and located around the perimeter to prevent interference with the operation of the sampler (fig. 4). The lower ends of the tubing were held in place by a second collar suspended near the end of the shortest length, approximately 244 cm from the top of the column. Samples (approximately 25 mL) from each location were obtained with a peristaltic pump (Masterflex L/S model 77300-80 manufactured by Barnant Co., Barrington, Ill.) operating at 50 mL/min, a rate sufficient to avoid mixing between column salt layers. Samples were collected from three of the six sample lines at one time (one connected to each of three pump heads) after purging approximately 25 mL to ensure representative samples. Ball valves on each line were closed to prevent flow back into the column and the pump lines were switched to obtain samples from the other three locations. EC of samples was measured in the lab following collection of all samples.

SE9027_files/image4.gif

Figure 4. Graphic depiction of sampling tubes around edge of column (not to scale).

After the initial samples were taken with the peristaltic pump, the sampler was used to collect samples from the same depths. The sampler was manually lowered into the column and stopped at marks on the handle corresponding to the specified depth for each sample. The rate at which the sampler was lowered was neither specified nor measured but was approximately 30 cm in 5 s, slow enough to avoid apparent turbulence. Upon reaching the desired depth, the sampler was opened for approximately 5 s to take the sample and then closed and slowly removed from the column. The contents of the sampler were emptied into a 4-L plastic beaker, mixed for 30 s by hand, and then approximately 30 mL was transferred to a 50-mL conical sample tube and sealed. The sampler and beaker were rinsed three times with tap water followed by three rinses with deionized water before the next sample was taken; the beaker was wiped dry with clean paper towels. The order of sample collection was not random but was consistently from the top to the bottom of the sample range.

To evaluate the mixing effect of the sampler, a second set of samples was collected with the peristaltic pump following sample collection with the lagoon sampler. The change in volume due to previous samples could affect the relative location of the second set of pumped samples and therefore alter the interpretation of results. The volume of the sample tubes was determined to range between 23.5 mL for the shortest tube and 29.6 mL for the longest tube. Considering the purge volume, the tube volume and the sample volume, a total of 460 mL was removed from the column in the process of obtaining the six pumped samples taken before the main samples, lowering the water level in the column by 0.19 cm. The six 300-mL samples taken with the sampler lowered the water level an additional 0.73 cm for a total of 0.92 cm. Therefore the samples pumped at the end of the process were actually taken from less than 1 cm higher than expected. Because of the depth of each layer this offset only presented a possible error for the samples taken from the interface between adjacent layers.

The column was drained and thoroughly rinsed at the end of the testing procedure. The entire test was repeated three times on separate days for a total of three replications. For each replication, the salt solutions were mixed and added to the column on the day prior to sample collection. The column was secured and covered with a tarp overnight. Sampling was completed each day by noon.

Statistical Analysis

All statistical analyses were conducted with SAS software (SAS 9.1.3, 2009, Cary, N.C.). Analysis of variance (ANOVA) was performed on the EC measurements of samples to estimate the effect of depth on the difference between samples. The least square means of the differences were calculated to study the structure of any significant differences found. A confidence level of 95% was used throughout the analysis.

Results and Discussion

Sampler Seal Evaluation

The initial evaluation of the seal showed no measurable water was collected in the chamber when the empty chamber was held in a 1-m column of water for 15 s. Likewise, the second test recovered the entire 300-mL sample when the full sample chamber was held for 15 s before opening.

Lab Scale Proof of Concept

The laboratory proof of concept study was performed simply to determine if salt solutions of different concentrations could be layered without mixing and diffusion and if, under ideal laboratory conditions, representative samples could be taken from several depths within the column. The coloring added to each layer was helpful in determining that the different salt solutions generally remained separate if the graduated cylinder was not disturbed. A small band (<5 mm) of green between the bottom blue layer and the middle yellow layer indicated minimal mixing did occur before samples were taken from the graduated cylinder (fig. 2). This color persisted for several days until the liquid was discarded.

As described in the testing protocol, a single 10-mL sample (each about 1% of the original column volume) was taken from each sample point starting at the top. The EC values of the bulk solution are represented by the solid line in figure 5 with the interface shown as vertical lines. The EC of samples taken from the interface between each layer are slightly lower than the expected values on the midpoint of the vertical lines but are still between the values of the adjacent layers. Since the pipette was positioned by hand during sample collection, human error could account for this difference. Some mixing of the bottom two layers was indicated by a thin layer (<5 mm) of green color but this green band neither increased nor decreased in size during the actual sampling or over several days in the undisturbed column. The general lack of mixing and the close agreement of EC values provided confidence the procedure would work in the large column at field scale.

SE9027_files/image5.gif

Figure 5. Electrical conductivity (EC) measurements from laboratory graduated cylinder experiment. Initial EC measurements were made prior to layering solution into the cylinder; sample EC measurements were made on solution withdrawn from the cylinder with a pipette.

Column Testing

Dissolved solids were determined in all samples over the three days of field column testing; results were related to EC measurements and are represented in figure 6. Because of the high coefficient of determination (R 2 ), EC measurements were used throughout the rest of the experiment.

SE9027_files/image6.gif

Figure 6. Electrical conductivity (EC) measurements represent total dissolved solids concentrations in the column test; based on a total of nine samples taken over three days of testing. Error bars = one standard deviation.

EC measurements were used to answer two questions: Does the redesigned device collect a representative sample at depth as evidenced by similar EC measurements from the sampler and from samples pumped directly from the column? And does the act of collecting samples disturb the water column to an extent such that the samples pumped after using the sampling device have EC measurements that are significantly different from those samples taken prior to using the device?

The question of the NCSU sampler providing representative samples was addressed by comparing the EC values of the initial pumped samples to those samples obtained with the sampler. EC values of the salt solutions measured before filling the column were 93 mS/cm in the 75-ppt solution, 156 mS/cm in the 160-ppt solution, and 188 mS/cm in the 250-ppt solution. EC of all samples collected by the sampler ranged from 22.0 to 189 mS/cm (table 1). Simple inspection of the EC values in table 1 reveals that the EC values of samples using the NCSU sampler were always less than those from samples pumped from the fixed points in the column. The ANOVA showed a significant effect of depth on the EC differences (P < 0.0001) and the least square mean difference was 11.3 mS/cm; the difference was greatest at the first two depths, where the standard deviation was also highest, and near constant at 8.0 mS/cm for the last four depths.

Differences in EC values between pumped and device samples may be due to a dilution effect caused by water on the outside of the device inadvertently added to the beaker used to transfer the sample from the device to the conical sample tubes. Although sufficient time was allowed for the bulk of the water to drip from the exterior of the sampler, the small amount remaining could be sufficient to dilute the samples, influencing the observed results. Upon closer inspection, however, much greater variability and larger differences between pumped and device samples were seen in the first samples taken on each day. These differences are too large to be solely caused by excess water dripping from the sampling device (table 1). Another source of error might have been excess rinse water being retained in the sample chamber.

Table 1. Electrical conductivity from column test, mS/cm (standard deviation; n = 3).

Depth (cm)

Initial (mS/cm)

Device (mS/cm)

122

44.2 (7.43)

22.0 (15.7)

137

102 (7.355)

89.1 (7.07)

152

112 (3.305)

102 (1.258)

168

155 (0.577)

148 (0.577)

183

163 (0.577)

155 (2.517)

198

189 (6.245)

181 (4.509)

While the difference in results at shallower depths seems to be a repeatable phenomenon in the column tests and the mean difference is a statistical reality, the value of the difference may not be important. Excluding the most dilute sample collected, the mean difference in EC was less than 10% of the actual value as estimated by the initial pumped samples.

The question of the sampler disturbing the water column was investigated by comparing the EC of samples taken from the fixed sampling points before and after the sampler was used (table 2). The ANOVA showed no significant effect of depth (P = 0.088) on the difference between EC values collected before and after the sampler was used, indicating that the column was not significantly disturbed or mixed; the least square estimate of the mean difference was 0.36 mS/cm but Student's t-test confirmed this value was not statistically different from zero (P = 0.767).

Table 2. Electrical conductivity as measure of disturbance mS/cm (standard deviation; n = 3).

Depth (cm)

Initial

Final

122

44.2 (7.43)

45.6 (6.50)

137

102 (7.355)

101 (8.445)

152

112 (3.305)

113 (4.007)

168

155 (0.577)

151 (0.577)

183

163 (0.577)

164 (3.215)

198

189 (6.245)

189 (8.327)

Sampler Operation

The redesigned NCSU sampler is relatively simple to operate and our results show it is capable of taking representative samples without disturbing the material being sampled. While the actual lowering of the sampler, opening of the sample chamber, and subsequent raising of the device is a one-person operation, a second person is recommended to accompany the operator. During the column testing procedure on solid ground, a second person was helpful in removing the sample from the chamber and cleaning the sampler in preparation for subsequent samples; safety was also a concern and reason for having a second person present. Since the top of the column was approximately 1.4 m above the platform to which the column was secured (fig. 3), a step ladder with handrails and safety harness were used to prevent the operator from falling; the second person helped assure the operator's safety. A second person should always be on hand when the device is operated from a boat on a lagoon to ensure personal safety and sampling repeatability.

Any analysis is only as good as the sample so when designing a sampling apparatus, cross contamination between samples must be addressed. While extensive washing of the device in the field may not be practical, some means of cleaning the sample chamber must be employed.

The NCSU sampler was used to collect samples from various depths and locations in several different lagoons for a separate project involving microbiological activity and sludge depth. The procedure used a sonar device to determine the depth of the upper layer of sludge; single samples were then taken from about 15 cm above this sludge layer and every 15 cm through the liquid column to a depth well into the thickest part of the sludge. The project involved measuring a number of physical and biological parameters of the samples which generally seemed representative of the different depths; the NCSU sampler met the expectations of the research team. The TS concentrations from two locations in one of those lagoons are shown in figure 7 and make it clear that this lagoon, like many waste lagoons, has a variable depth to sludge and the TS content of that sludge can vary with location within the lagoon. The sludge at location B has a distinct upper boundary at the 2-m depth and consistent TS concentration below that. The upper boundary of sludge at location A is only 0.5 m below the surface and the TS concentration increases gradually with depth. In general, the NCSU sampler collected material that seemed representative of the different depths with little evidence of mixing or interference; one exception was the sample taken just above the 2.0-m depth at location A. The TS of that sample unexpectedly dropped from 12.3% at 1.7 m to 10.3% at 1.8 m. This anomaly in TS concentration could be explained by the sampler disturbing the sludge below the 1.7-m sampling depth. Additional field trials would be necessary to determine if this is a function of the number of prior samples, the TS concentration, or simply operator error such as lost footing or relaxed grip.

SE9027_files/image7.gif

Figure 7. Total solids (TS) measurements from swine lagoon samples collected with the NCSU sampler. Locations are in different parts of the same lagoon collected on the same day; the depth to sludge as indicated by sonar varies across the lagoon.

Protocol Evaluation

The purpose of the sampler described above is to take representative samples from specific depths in an anaerobic lagoon. Although the parameters of interest in such samples will vary, researchers often want to know the concentration of TS, suspended solids, nitrogen, phosphorus, and potassium. The protocol used to evaluate the sampler only compares the value of dissolved solids in samples as measured by EC. This limitation is a justified simplification that is straightforward to set up and easy to implement. The potential problem of repeated sampling at the same location is unrepresentative samples due to mixing of the water column. If disturbed, dissolved materials are less likely to quickly return to any previous state than are suspended materials and are therefore more susceptible to mixing; using dissolved solids as a surrogate measurement is a robust choice for the evaluation.

The specific salt concentrations and the 30.5-cm depth of each layer were chosen to keep the procedure and preparation simple but required that samples be collected from the layer interface as well as from the midpoint of each layer to adequately test the requirement that discrete samples be collected at 15.2-cm intervals. An alternative procedure would be to use six layers of only 15.2-cm depth with six different salt concentrations and collect samples from the midpoint of all six layers, ignoring the interface locations. Although set up would be twice as long, there would be even less influence of changing height of the water column on the relative location of the tubes installed to collect the pumped samples.

Conclusions

We have redesigned a sludge sampler based on current needs for discrete samples from various depths at the same location in a lagoon. The new design is easy for one person to operate up to depths of 3.4 m which is inclusive of most lagoons in North Carolina and other southeast locations; a second person should be on hand for safety considerations and to help with sample transfer.

We have developed a testing protocol for samplers using layers of salt solutions in a 55.9-cm diameter, 3.35-m tall PVC column. A testing column of this size is needed to provide sufficient clearance for operation of the sampler to avoid artificial mixing effects in the column and to impose a realistic hydrostatic pressure to mimic real lagoons. The protocol can be adapted to use a deeper column or more layers to accommodate specific needs.

Testing showed the sample compartment integrity is sufficient to ensure samples are only collected from where the device is opened in a lagoon. Although some variability was seen in sample characteristics at the upper levels of the test column, samples collected with the device from the deeper, more concentrated levels showed good agreement in terms of EC with the expected measurements of independent samples pumped from the same depths. No such verification of sampler performance has been shown in the past.

The sampler was used to collect samples from various depths and various locations in several lagoons and performed well in terms of ease of operation and apparent sample integrity. When results from one of these lagoons were analyzed, there was evidence of distinct samples obtained from each depth. Only in 1 of 19 samples did there appear to be an anomaly in TS data that might indicate that the sampler or its operation had an effect on the sample collected.

Acknowledgements

The authors thank the North Carolina Pork Council for initial funding for this project.

REFERENCES

Anderson, J. B., D. T. Hill, and T. W. Tyson. 2000. Physical and chemical properties of undisturbed stratified sludge in swine waste anaerobic lagoons. In Proc. of 8th Intl. Symposium on Animal, Agricultural and Food Processing Wastes, 171-179. St. Joseph, Mich.: ASAE.

Bicudo, J. R., P. W. Westerman, and L. M. J. Safley. 1999. Nutrient content and sludge volumes in single-cell recycle anaerobic swine lagoons in North Carolina. Trans. ASAE 42(4): 1087-1093.

Caldwell, P. V., A. A. Adams, C. P. Niewoehner, M. J. Vepraskas, and J. D. Gregory. 2005. Sampling device to extract intact cores in saturated organic soils. Soil Sci. Society of America J. 69: 2071-2075.

Devine, A. A., J. C. Ellis, L. S. Bull, and A. M. Grunden. 2006. Utilizing TRFLP analysis for the determination of the microbial community found in hog waste lagoon systems. Abstracts of the General Meeting of the American Society for Microbiology 106402. Washington, D.C.: ASM.

Fulhage, C., A. Schmidt, and J. Lory. 2005. Long-term sludge and nutrient accumulations in swine lagoons: A case study. In 2005 Animal Waste Management Symposium, 416-423. Raleigh, N.C.: North Carolina State University.

Hamilton, D. W., and R. H. Rosser. 1994. The Sidewinder: An improved lagoon sampler. ASAE Paper No. 944064. St. Joseph, Mich.: ASAE.

Kroes, J., C. L. Barth, and R. B. Dodd. 1987. Sludge sampler for waste lagoons. Applied Eng. in Agric. 3(2): 258-260.

Lide, D. R., ed. 2005. CRC Handbook of Chemistry and Physics. 86th ed. Cleveland, Ohio: Chemical Rubber Publishing Co.

Lorimor, J., C. Fulhage, R. Zhang, T. Funk, R. Sheffield, D. C. Sheppard, and G. L. Newton. 2001. Manure Management Strategies / Technologies White Paper. Ames, Iowa: National Center for Manure and Animal Waste Management, Midwest Plan Service.

Parker, D. B., D. D. Schulte, and J. A. Nienaber. 1999. Flat-plate freeze sampler for sludge characterization in a feedlot runoff storage pond. Trans. ASAE 42(1): 123-125.