Instrumentation for Measuring Runoff, Sediment ... - Semantic Scholar
Published online January 3, 2006
Instrumentation for Measuring Runoff, Sediment, and Chemical Losses from Agricultural Fields
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
Carlos A. Bonilla,* David G. Kroll, John M. Norman, Daniel C. Yoder, Christine C. Molling, Paul S. Miller, John C. Panuska, Jeffrey B. Topel, Peter L. Wakeman, and K. G. Karthikeyan lection tanks are suitable, then the areas may not be representative of typical field conditions (Toy et al., 2002). Another common instrumentation technique continuously measures and records the runoff rate using a weir or flume in conjunction with a water-level sensor and datalogger. An automated pumping sampler is used to draw samples from just downstream of the weir or flume, communicating with the datalogger to tie the sample concentration to the flow volume it represents, permitting a mass movement calculation. Such systems are valuable for studies needing time-varying concentration values, but may be overkill for studies requiring only total storm mass values. A system like this can be expensive because it consists of a flume with substantial installation effort, water level sensor, datalogger, and automatic sequential sampler. In addition, such systems assume that samples extracted at specific times are representative of larger time frames, and may cause excessive sedimentation due to the flume constriction (Hamlett et al., 1984). Slot samplers offer an alternative to flumes and total collection tanks and can provide a compromise solution of modest cost for a wide range of plot sizes, from small watersheds up to an area of several km2 (PAP/RAC, 1997). Slot samplers obtain a representative portion of the runoff–sediment mixture. Two common slot samplers are the stationary multislot divisor and the Coshocton wheel sampler, which is equipped with a revolving slot. The stationary multislot divisor, first suggested by Geib (1933), routes the runoff from a collector through the multislot divisor, where a sample is obtained from a single slot and routed to a sample storage tank. A second or third sample storage tank may be connected to the first, if additional sample storage is needed (Brakensiek et al., 1979; PAP/RAC, 1997). The design of slot samplers has been refined over the years to overcome some of their limitations and make them more suitable for sediment transport and chemical loss studies. Sheridan et al. (1996) designed a low-impact flow event (LIFE) sampler to minimize disturbance and cost. This slot-type sampler was used to quantify nutrient concentrations in runoff flowing through riparian buffers in the Coastal Plain, where slopes are gentle and sheet flow is likely. Later, Franklin et al. (2001) modified the LIFE sampler to accommodate steep slopes (5–12%), and larger flow rates (1–5.5 L min21). The design consists of two sets of sample splitters (10 splitters for each set) connected in series. One-tenth of the inlet flow exits from the first set of 10 splitters and one-hundredth of the inlet flow from the second set, to be finally collected in a holding tank. Recently, Pinson et al. (2004) designed an improved stationary multislot divisor for sampling runoff plots. The system is built using a commercially available 19-L
ABSTRACT This work describes a simple, passive sampling system for measuring runoff, sediment, and chemical losses from typical agricultural fields. The sampler consists of a 5 to 7 m wide runoff collector connected to a series of multislot divisors. These divisors split the flow into aliquots, providing a continuous sampling during the runoff event. Divisors were located in a wooden box below ground level. With an adequate pump, this system can operate in fields with a slope gradient as low as 2%, and can stay in the field during winter to record first snowmelt-generated runoff. A radio transmitter reports by telemetry the occurrence and magnitude of any runoff event, and indicates when the system should be sampled and emptied. This article includes a description of the equipment, advantages, and disadvantages based on 2 yr of operation, and examples of data collected.
T
HE MEASUREMENT of sediment transport and chemical losses from farm fields has been a major research interest because of the need to characterize practices that conserve soil and water resources. Although many instruments have been developed in an attempt to measure runoff and sediment transport, at the present time none has proven to be fully satisfactory (PAP/RAC, 1997). For plots and small watersheds, total collection tanks, flumes and samplers, and slot divisors have commonly been used (Brakensiek et al., 1979; PAP/RAC, 1997; Toy et al., 2002). Total collection tanks may be constructed to measure erosion in small test plots, and must be large enough to contain the total runoff (water and sediment) expected in a 24- or 48-h period. The volume of the water–sediment mixture is then determined and the solid sediment material sampled for subsequent laboratory analysis and computation of its weight and volume. Unfortunately, total collection devices are often unsuitable for erosion studies because the runoff even for small plots can be excessive (Brakensiek et al., 1979; PAP/RAC, 1997). If plots are small enough so total col-
C.A. Bonilla, D.G. Kroll, J.M. Norman, J.B. Topel, and P.L. Wakeman, Department of Soil Science, University of Wisconsin-Madison, 1525 Observatory Drive, Madison, WI 53706; D.C. Yoder, Biosystems Engineering and Soil Science; University of Tennessee, 2506 E.J. Chapman Drive, Knoxville, TN 37996-4531; C.C. Molling, Cooperative Institute for Meteorological Satellite Studies, Space Science and Engineering Center, University of Wisconsin-Madison, 1225 W. Dayton Street, Madison, WI 53706; P.S. Miller, J.C. Panuska, and K.G. Karthikeyan, Biological Systems Engineering, University of Wisconsin-Madison, 460 Henry Mall, Madison, WI 53706. The company names are for the convenience of the readers and do not imply an endorsement by the University of Wisconsin-Madison. Received 19 Apr. 2005. *Corresponding author ([email protected]). Published in J. Environ. Qual. 35:216–223 (2006). Short Communications doi:10.2134/jeq2005.0130 ª ASA, CSSA, SSSA 677 S. Segoe Rd., Madison, WI 53711 USA
216
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BONILLA ET AL.: RUNOFF, SEDIMENT & CHEMICAL LOSS FIELD MEASUREMENTS
(5-gallon) bucket with a screw-top lid. A series of 22.5j V-slot weirs are precision laser cut into a strip of metal, which is then rolled into a ‘‘crown’’ and fastened to the top of the lid, from which the center has been removed. When used, plot runoff is routed into the bucket. The bucket fills completely and overflowing water–sediment is evenly divided among many slots. Flow from one of the overflowing slots is then collected in another bucket, and more buckets with crowns can be placed in series to divide runoff for larger storm events or larger size plots. The objective of this article is to describe the design, installation, and operation of an effective system suitable for measuring runoff, sediment, and chemical losses from within agricultural fields or near their edges. The system is a significant advance in that it makes use of the Pinson et al. (2004) multislot divisor; furthermore, we believe this new design to be superior to that of Franklin et al. (2001). This new design accommodates a wide range of runoff rates and amounts, is convenient because of the use of 19-L buckets, can handle collector widths of 5 m or more for better landscape sampling, allows studies near discharge outlets where slopes are small, and can be used for contributing areas as large as 0.4 ha (1 acre).
DESIGN AND INSTALLATION The runoff–sediment sampler has three main components: (i) the collector, which intercepts and routes the runoff into the multislot divisors; (ii) the multislot
217
divisors, which provide a flow proportional sample of sediment and runoff; and (iii) the electronic and data storage devices that record the occurrence and magnitude of the runoff events.
The Collector The collector, typically 5 to 7 m wide (Fig. 1 and 2), intercepts and routes the runoff into the multislot divisors. The water flow is directed to the collector outflow pipe by two boards that form the sides of a funnel. A plastic sheet over the collector floor stops infiltration losses, and a heavy duty tarp installed as a roof prevents direct rainfall from entering the system. The plastic sheet that forms the collector floor is installed in a trench dug down-slope of the 5 to 7 m long inlet side and buried so that the inlet to the collector forms a smooth transition from the soil to the plastic with minimal disturbance. After the inlet side is installed, the side boards are ‘‘rolled’’ up in the plastic sheet until they are vertical and extend from the corner of the inlet side to the outlet pipe of the collector (see Fig. 3 for pictures of a field installation). Rolling the plastic on the side-boards ensures an excellent seal all along the sides of the collector walls. The side boards are fastened to metal stakes and stretched out to keep the floor straight and flat, avoiding wrinkles and reducing sedimentation in the collector. The stakes are pounded in the ground about 2 or 3 cm back from the side board so that 5-cm long bolts can be used to pull the side boards toward the stakes, stretching the plastic floor. The roof tarp is
Fig. 1. Field installation. The collector (on the left) shows the wooden-frame and the plastic floor. When finished, the collector is covered with a heavy-duty tarp. The bottom truss on the front support for the collector roof comes down to 8 cm above the soil and is 15 cm upslope of the front edge of the plastic floor. A PVC pipe (in the center) carries the flow to the multislot divisors (on the right). Only the first divisor is above grade level so that the excess water is routed directly down the slope. The excess water from the other divisor heads flows to the sump, where a battery operated pump (114 L min21 [30 gallons min21] capacity) lifts the sump water to grade level for discharge.
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
218
J. ENVIRON. QUAL., VOL. 35, JANUARY–FEBRUARY 2006
Fig. 2. An illustration of the runoff collector used with the edge-of-field runoff collection system. The entire frame is covered with a heavy-duty tarp including the front gable area. The front of the collector cover extends to about 8 cm above the soil and 15 cm beyond the plastic floor to prevent rain-drop splash from entering the collector.
attached on the lower board of the ‘‘A-frame’’ (Fig. 1), stretched over a wooden frame, and a series of bungee cords keep the tarp tight, as well as facilitating removal
Fig. 3. Pictures of a field installation. (A) Transition of the collector floor and walls to the 15-cm diameter PVC pipe where the plastic floor is inserted into the pipe and retained with a split PVC ring within the pipe. A screen prevents field debris from getting to the divisor heads. (B) Picture of four-bucket divisor system under construction with plastic liner, pump circuit and battery, and neoprene boot construction around Bucket 1 (see text). (C) Completed divisor assembly with cover, pump outlet pipe, supply pipe with elbow, rain gauge, and container that protects the battery and pump circuit for sump operation. (D) Field installation for winter and spring snowmelt operation. The rain gauge and solar panel for charging the pump battery are visible next to the divisor collection system and the covered collector is at left.
of the roof tarp for checking and servicing (cleaning or patching) the collector interior. A 15-cm diameter (6-inch) PVC pipe is used to route the runoff from the collector to the multislot divisors, and the plastic floor is fed into the PVC pipe and held in place with a split ring inside the 15-cm pipe (Fig. 3A). Six vertical slots opened at the end of the elbow on the pipe avoid the accumulation of water in the PVC pipe if water freezes in the first bucket. Each slot is 6 cm high and 0.3 cm wide (see Fig. 4, detailed in Fig. 4A). The slots must be located in such a way that 3 cm are below the lower apex of the V-notch of the bucket, and the other 3 cm are above. These slots ensure normal operation of the divisor system when intermittent spring runoff occurs interspersed with freezing conditions that can lead to a thin layer of ice on the top of the water within the first bucket and inside the PVC pipe (which extends about 7 cm below the water surface level in the first bucket). Two conditions are of importance when installing the collector. First, the collector entrance must be located in an area where the topography concentrates the runoff, or where the occurrence of rills is evident. The selection of the site can be facilitated by using a detailed digital elevation map or visiting the field after a major rainfall event to check for rill occurrence. Alternatively, the collector may be used at the bottom of a field plot isolated by some type of borders, with those borders connecting directly to the ends of the collector. The second recommendation is that the collector must be located in a place that provides at least 2% slope along the side-boards and also along the PVC pipe. A
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BONILLA ET AL.: RUNOFF, SEDIMENT & CHEMICAL LOSS FIELD MEASUREMENTS
219
Fig. 4. Box design and dimensions.
slope equal to or greater than 2% provides enough energy to the water to keep the sediment in suspension (Pinson et al., 2004), reducing the chance of deposition inside the collector or along the pipe. After 2 yr of operation, no sediment has ever been observed in the PVC pipe, even for lengths of 15 m.
runoff rate that can be handled with the 12-notch divisor is about 30 L s21, and 7 L s21 when using the 24-notch divisor (Pinson et al., 2004). The maximum measurable runoff volume with this configuration is close to 540 m3.
The Multislot Divisors From the collector, the runoff is routed into multislot divisors specially designed for this purpose (Fig. 4 and 5). This type of divisor system provides a storm-integrated, discharge-weighted sample for determining runoff and sediment yield by using a stationary slotted crown developed by Pinson et al. (2004). The crown is installed on the top of a 19-L bucket. When the first bucket is full, most of the effluent is discarded (11⁄12) and a small fraction (1⁄12) is routed to a second bucket (Fig. 3B). When the second bucket fills, a small fraction of the overflow is routed to a third bucket (1⁄24) and the remainder is discarded. Finally, when the third bucket fills, a fraction of the overflow (1⁄24) is routed to a 79-L (21-gallon) basin and the remainder is discarded. The maximum peak
Fig. 5. Illustration of (left) 24-slot and (right) 12-slot multislot divisors (all units are in cm).
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
220
J. ENVIRON. QUAL., VOL. 35, JANUARY–FEBRUARY 2006
The four buckets are located in a pressure-treated wooden box, buried with its top slightly above ground level (Fig. 3C). The box design and dimensions are given in Fig. 4. The frame is built with 3.5 by 3.5 cm (2 by 2 inch) boards but 3.5 by 8.5 cm (2 by 4 inch) boards form a base at the bottom to maintain buckets within 0.05j of level as they fill with water. This 3.5 by 8.5 cm base, which contains a sturdy shelf for Bucket 3, is installed first and leveled in the bottom of the hole so that the box rests on top of it. Three solidly braced shelves are installed in the box to support the buckets. The first bucket protrudes through the PVC tray so that excess water from the 11 divisor slots is drained down slope. A neoprene boot, which is attached to the tray and tightened around the neck of the bucket with a bungee cord, prevents excess water that is leaving the divisor slots and exiting the system from draining to the sump pump, thereby exceeding the 113 L min21 capacity of the pump (Fig. 4A). This neoprene seal (0.8 mm thick, 1⁄32 inches) is attached to the tray by sandwiching it between the tray and a PVC ring 36 cm i.d. When sampling Bucket 1, it can be lifted from above without disturbing the PVC tray by loosening the neoprene boot, lifting the 15-cm PVC supply pipe, and removing the elbow. Easy removal of Bucket 1 is important because sediment cannot be sampled directly from Bucket 1 in the field with satisfactory results (discussed later). Since Buckets 2 and 3 are positioned below grade level, the excess water from these buckets drains to a 79-L (21-gallon) bucket beneath the box, where it is pumped out with a sump pump. The power supply for the pump is a 12-V, deep cycle battery connected to a solar panel for continuous operation (Fig. 3B). A large deep-cycle battery with 82 amp-hour capacity can operate the pump continuously for 6 h. If the field topography provides enough slope gradient, water can be discharged from a hole in the bottom of the box and the pump is not needed. This system has been operating at a number of sites year-round in Wisconsin for 2 yr, quantifying runoff, sediment, and P losses from winter and spring snowmelt as well as spring, summer, and fall storms. Winter operation is accomplished by constructing a treated plywood box around the divisor head enclosure, insulated on the inside with 5-cm thick sheets of urethane foam (Fig. 3D). The wall of the wooden box is dug into the soil to a depth of about 5 cm. The dimensions of this box are 2.4 m long by 1.5 m wide by 0.6 m high, and it has a urethane foam cover consisting of two 5-cm thick sheets reinforced with thin metal strips. A tarp 3.7 m long and 2.4 m wide completes the installation cover, using staked bungee cords to stretch it tightly over the insulated frame. This design makes use of heat stored in the soil and prevents freezing inside if there is ample snow cover. Although a cold winter with little snow will result in subfreezing temperatures inside the double enclosure during the coldest time of the winter, in 2 yr this has not occurred during the times when snowmelt occurred. Thus, this passive, unheated system has worked for snowmelt monitoring with no apparent loss of data if samples are removed promptly (within 1 d) following the snowmelt event.
The precision of the runoff and sediment measurements depends on how well the V-notch weirs in each bucket are leveled. When building the divisor, the notches must be cut as precisely as possible, ideally with a laser device. Once in the field, the buckets must be leveled in place. In this setup, each shelf is provided with a triangular metal frame with a bolt in each corner (Fig. 4D). By turning the bolts, the divisor heads are leveled with a special machine level tool that is placed in the notches of the divisor head to obtain level within 60.05j. This device consists of an upper flat surface on which sits a precision level (30-5 cm [12-inch] machinist’s improved level no. 98-12, Starrett Co., Athol, MA), connected to a lower stainless steel rod (10 mm diam.) that sits in the divisor head slots. The device must be carefully assembled and machined to achieve the 60.05j leveling accuracy. Laboratory evaluation indicates that this system divides runoff with accuracies within 65% over most of the flow range and within 616% at very low and very high flows (Pinson et al., 2004). Laboratory measurements of sediment division accuracy indicated smaller errors than flow division (Pinson et al., 2004).
The Electronic and Data Storage Devices The divisor provides a storm-integrated, dischargeweighted sample for runoff. However, by installing a water content sensor (ECH2O-20 probe, Decagon Devices, Pullman, WA), in each bucket, the time-varying runoff rate can be estimated. If it is installed vertically, the water sensor reading can be converted to water volume for each 19-L (5-gallon) bucket with a sensor (Masarik et al., 2004), and once connected to a logger the flow rate over time can be calculated. If the logger includes a radio system, the runoff occurrence can be checked by telemetry. Alternatively, the logger could be connected to a cell phone and monitored from any location. On rare occasions the maximum capacity of a divisor system may be exceeded, especially with snowmelt on ground with concrete frost. When this occurs, information about the size of the event is lost, and one only knows that the event exceeded the maximum capacity of the system. By installing an ECH2O-20 probe vertically in the 15-cm PVC supply pipe about 50 cm upslope from the elbow that empties water into the first bucket, the water depth in the PVC pipe can be monitored. Knowing the water depth, the water flow rate can be calculated from Manning’s equation (Haan et al., 1994) if the slope of the pipe is known. Fiveminute averages of water depth can be recorded on a datalogger and calibration of the ECH2O-20 probe maintained using data that does not exceed the divisor system capacity. This system is being tested but insufficient experience has been obtained to verify its robustness.
Cost The total cost for supplies, labor, and instrumentation is about $5000 per installation (Table 1). About onethird of this cost is for the datalogger system to record water levels in buckets, rainfall, soil temperature, and soil moisture. About one-seventh of the total cost is divisor heads and buckets, which were purchased from
221
BONILLA ET AL.: RUNOFF, SEDIMENT & CHEMICAL LOSS FIELD MEASUREMENTS
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
Table 1. Materials, instruments, and labor required to install a runoff collector unit. Item
Cost
Instruments Multislot divisors (Biosystems Engineering and Environmental Science, Univ. of Tennessee, Knoxville, TN 37996-4531) Pump system Solar panel, 20 W (Campbell Scientific, Logan, UT 84321-1784) Heavy duty sump pump, 12 VDC, 11 amps (Basement Watchdog Special model BWSP, 1730 GPH @ 10 ft. Glentronics, Wheeling, IL 60090) Deep cycle marine battery, 82 amp-hour (local hardware store) Data collection system Data logger (EM5R logger, Decagon Devices, Pullman, WA 99163) Water content sensor (ECH2O, 20 probe, Decagon Devices, Pullman, WA 99163) Radio transceiver module (Rm1 radio module, Decagon Devices, Pullman, WA 99163) Tipping bucket rain gauge (Model TR-525I, Texas Electronics, Dallas, TX 75235) Pocket PC (iPAQ Pocket PC h2215, Hewlett-Packard Co., Palo Alto, CA 94304) Materials Pipes, elbows, metal stakes, boards, tarps (local hardware store) Labor Wooden box, installation Total
$700
the University of Tennessee. A considerable amount of labor is required to build the boxes and install the collector and divisor head box in the field.
OPERATION After every runoff event the depth of water in each bucket is measured using an aluminum meter stick graduated in millimeters (to allow for a volume calculation), and water samples are taken for analysis. Once in the laboratory, sediment samples are dried and weighed to determine the sediment mass. The total runoff volume for the entire event at the time of sampling is calculated by R 5 V1 1 (12 3 V2 ) 1 (12 3 24 3 V3 ) 1 (12 3 24 3 24 3 V4 )
[1] 3
where R is the total runoff volume (m ), and V1 to V4 (m3) are the volumes collected in Buckets 1 through 4, respectively. Total sediment mass for the entire runoff event at the time of sampling is calculated by S 5 (V1 3 C1 ) 1 (12 3 V2 3 C2 ) 1 (12 3 24 3 V3 3 C3 ) [2] 1 (12 3 24 3 24 3 V4 3 C4 ) where S is the total sediment (kg), and C1 through C4 (kg m23) are the concentration of total solids measured in Buckets 1 to 4, respectively. Chemical loads can also be calculated by using Eq. [2], but with chemical concentrations replacing sediment concentrations. When sampling sediment, experience has shown that it is not possible to obtain a representative sample from the soil suspension in the first 19-L (5-gallon) bucket in the field. The largest particles are in this bucket, and it is not uncommon to have a 5- to 10-cm layer of sediment in the bottom of the bucket. Thus, the first bucket is removed from the system entirely and returned to the laboratory for measurements; a new first bucket is then installed in the system each time samples are removed.
Alternative Configurations Various configurations of buckets and divisor heads provide flexibility for accommodating different sized contributing areas. For small contributing areas, the ac-
$800
$1700
$400 $1200 $4800
curacy of the measurements can be improved by reconfiguring the buckets and divisor heads based on the maximum runoff rate and the maximum runoff volume that are expected to occur. The maximum flow rate that can be handled with a 12slot divisor head is 29.7 L s21 (Pinson et al., 2004). In the field the maximum runoff rate can be estimated from the contributing area, the topographic features, and rainfall intensity. When deciding about a configuration, the peak flow should be obtained from a hydrograph; however, using a worse-case scenario approach, the maximum runoff rate can be predicted as the product of the rainfall intensity, the runoff coefficient for the field conditions, and the size of the contributing area. The maximum runoff volume that can be sampled depends on the number of buckets and the configuration of divisor heads. Table 2 and Fig. 6 show three of the multiple combinations that can be implemented in the field. Configuration 1 corresponds to the design that has been presented previously in the article, which accommodates large runoff rates and volumes (29.7 L s21, runoff volume . 30 m3). This configuration consists of three 19-L (5-gallon) buckets (Fig. 6A), with a 12-slot divisor in Bucket 1, and 24-slot divisors in Buckets 2 and 3, plus a 79-L (21-gallon) bucket to store 1⁄24 of the water exiting from Bucket 3. In particular for this configuration, as a result of surface tension effects at the bottom of the slots, flows lower than 0.02 L s21 in Bucket 3 produce errors larger than 616% when using the 24-slot divisor bucket (Pinson et al., 2004). Table 2 shows that the lowest flow rate into Bucket 1 that will assure a minimum divisor-head flow rate of 0.02 L s21 (error less than 616%) in Buckets 3 is 5.76 L s21. Thus, on a 0.5-ha field, the runoff rate should be larger than about 4 mm h21 to ensure a flow rate into Bucket 3, which is . 0.02 L s21. For small fields the system can be modified to decrease the minimum runoff rate. The slots in Bucket 3 can be Table 2. Maximum runoff capacity and minimum flow rate into Bucket 1 for the three configurations shown in Fig. 6. Configuration 1, Midsize fields 2, Small fields 3, Plots
Maximum runoff capacity
Minimum flow rate
L 554 000 28 600 5 690
L s21 5.76 0.24 0.24
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
222
J. ENVIRON. QUAL., VOL. 35, JANUARY–FEBRUARY 2006
Fig. 6. Three divisor heads configurations suitable for different sized contributing areas: (A) Mid-size fields (. 0.4 ha), (B) small fields (0.04–0.4 ha), and (C) plots (, 0.04 ha).
covered (e.g., using tarp tape), connecting directly Buckets 3 and 4 (Fig. 6B). This bridge essentially gives the third bucket a 99-L (26-gallon) capacity, reducing the minimum runoff rate requirement from 5.76 L s21 (configuration 1) to 0.24 L s21 (Table 2, configuration 2). The maximum runoff capacity for this configuration is close to 30 m3. With configuration 2 the runoff rate on a 0.5-ha field should be larger than about 0.17 mm h21 to ensure a flow rate into Bucket 3 that is . 0.02 L s21. For plots with runoff volumes smaller than 6 m3, a combination of three 19-L (5-gallon) buckets can be used (Fig. 6C). A 12-slot divisor is used in Bucket 1, and a 24-slot divisor in Bucket 2. The third bucket is used to store 1⁄24 of the water exiting from Bucket 2. This
configuration operates with a minimum flow rate into Bucket 1 as low as 0.24 L s21 (Table 2, configuration 3).
EXAMPLE APPLICATION Eight of these runoff–sediment collectors have been installed in different agricultural fields in Wisconsin, and they have been operational since July 2003. More than 300 site events have been collected to date, and Table 3 contains a sampling of events from various sites. To illustrate the operation of the system, one additional event is shown below in detail. The runoff event took place on 11 June 2004 in a field located in west-central Wisconsin. The field consists of a 0.5-ha
Table 3. Sampling of runoff events collected from various sites. Contributing area
Texture
ha
Slope
Crop
silt loam
13
0.10
silt loam
0.03
alfalfa
62.0
7
corn
96.6
clay loam
5
corn
28.9
0.05
loam
8
alfalfa
128.3
0.21
loam
8
alfalfa
109.6
0.17
sandy loam
10
Event
mg kg21
%
0.50
Bray P-1
corn
63.6
Runoff volume
Total P
no. buckets
g L21
kg ha21
0.01 2.66 7.23
0.5 2.4 3.0
0.83 1.52 3.10
0.02 8.07 44.60
3.70 0.18 0.26
6.39 2.43 1.94
0.01 4.26 7.23
0.4 2.6 3.0
1.23 3.55 1.27
0.10 150.00 91.60
0.37 0.62 0.37
1.81 1.44 1.28
0.01 10.48 0.59
0.6 3.0 2.1
0.14 2.56 0.32
0.06 829.40 5.82
0.18 0.08 0.03
0.58 1.13 0.34
0.03 0.23 1.73
1.1 1.8 2.2
0.78 0.74 0.35
0.53 3.47 12.16
1.35 1.05 0.38
2.03 1.72 1.06
0.02 1.72 272.70
1.0 2.2 3.5
3.10 38.00 0.58
0.33 1.23 757.80
0.52 0.80 0.50
7.47 1.48 1.05
0.01 3.18 101.80
0.3 2.5 3.2
0.61 13.90 0.00
0.03 262.30 2.11
1.29 0.69 0.68
1.59 5.64 0.90
m3 Site 1 storm 24 May 2004 storm 11 June 2004 snowmelt 22 Mar. 2004 Site 2 storm 15 June 2004 storm 30 Mar. 2004 snowmelt 22 Mar. 2004 Site 3 storm 23 Nov. 2004 storm 11 June 2004 snowmelt 9 Mar. 2004 Site 4 storm 23 May 2004 storm 31 Dec. 2003 snowmelt 2 Mar. 2004 Site 5 storm 26 Mar. 2004 storm 26 Mar. 2004 snowmelt 3 Jan. 2005 Site 6 storm 22 Nov. 2004 storm 22 May 2004 snowmelt 9 Feb. 2005
Dissolved reactive P
Sediment
mg L21
BONILLA ET AL.: RUNOFF, SEDIMENT & CHEMICAL LOSS FIELD MEASUREMENTS
Table 4. Example of data collected during an actual rainfallrunoff event. Collected volume
Sampled volume
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
L Bucket 1 Bucket 2 Bucket 3 Bucket 4 Total
23.3† 23.3† 8.18 0.0
23.3 280 2360 0.0 2663
Sediment g L21 7.50 2.13 1.39
Dissolved reactive P
Total P
mg L21 0.13 7.51 0.17 2.88 0.18 2.33
† The volume is larger than 19 L (5 gallons) because of the additional depth associated with the divisor head.
223
This instrumentation has been used successfully for almost 2 yr in farm fields ranging in size from 0.04 to 0.5 ha. Since a small weighted fraction of the runoff is collected during the entire runoff event, this system can sample from larger areas than the systems designed for the smaller standard unit plots (4 by 22 m, ,0.01 ha). Thus, this proposed system works on fields that are more representative of typical farm fields; however, determining the contributing area of the collector using detailed topographic maps can be challenging. ACKNOWLEDGMENTS
contributing area planted with alfalfa under no tillage for many years. The average slope is 12%, with a minimum of 0.6% and a maximum of 23%. The typical soil texture is silt loam, with 4% organic matter. The rainfall event totaled 35 mm, with a maximum intensity of 6.4 mm in 30 min. The erosivity index for the storm (EI30) was calculated to be 75.3 MJ mm ha21 h21 in metric units or 4.4 hundreds ft tonf in acre21 h21 in U.S. customary units. At the end of the event the first two buckets were full of water, and the third bucket was half full. The water volume, sediment, dissolved reactive P, and total P measured in each bucket are presented in Table 4. The total runoff from the contributing area was 2.7 m3 (calculated using Eq. [1]). A total of 4.05 kg of sediment was transported through the collector during the event, with an average concentration of 1.52 g L21. For P, 0.47 g of dissolved reactive P were transported, with an average concentration of 0.18 mg L21. Finally, 6.46 g of total P were moved during this event, with an average concentration of 2.43 mg L21. When expressed as a function of contributing area, 8.10 kg ha21 of sediment was eroded (approximately 7.2 lb acre21), and 0.93 g ha21 and 12.87 g ha21 of dissolved reactive P and total P were transported, respectively.
CONCLUSIONS The instrumentation presented in this article provides a cost-effective passive sampling system for measuring runoff, sediment, and chemical losses under snowmelt or storm conditions in farm fields. Measurements can be made where there is no external source of power and in remote locations on operational farms, even near a discharge outlet where slopes are usually small.
This research was partially supported with funds from the USDA Cooperative State Research, Education, and Extension Service’s Water Quality Program under grant no. 2002-5113001951, and the Wisconsin Buffer Initiative funded by the Natural Resources Conservation Service under grant no. 685F48-3-085 and 68-5F48-4-082 and the Wisconsin Department of Natural Resources with funding from the Bureau of Watershed Management, Runoff Management Section for NonPoint Source Pollution Abatement.
REFERENCES Brakensiek, D.L., H.B. Osborn, and W.J. Rawls (ed.). 1979. Field manual for research in agricultural hydrology. USDA Agric. Handb. 224. USDA, Washington, DC. Franklin, D.H., M.L. Cabrera, J.L. Steiner, D.M. Endale, and W.P. Miller. 2001. Evaluation of percent flow captured by a small in-field runoff collector. Trans. ASAE 44:551–554. Geib, H.V. 1933. A new type of installation for measuring soil and water losses from control plots. J. Am. Soc. Agron. 25:429–440. Haan, C.T., B.J. Barfield, and J.C. Hayes. 1994. Design hydrology and sedimentology for small catchments. Academic Press, San Diego, CA. Hamlett, J.M., J.L. Baker, S.C. Kimes, and H.P. Johnson. 1984. Runoff and sediment transport within and from small agricultural watersheds. Trans. ASAE 27:1355–1363, 1369. Masarik, K.C., J.M. Norman, K.R. Brye, and J.M. Baker. 2004. Improvements to measuring water flux in the vadose zone. J. Environ. Qual. 33:1152–1158. PAP/RAC. 1997. Guidelines for mapping and measurement of rainfallinduced erosion processes in the Mediterranean Coastal Areas. PAP-8/PP/GL.1. Split, Priority Actions Programme Regional Activity Centre (MAP/UNEP), with the cooperation of FAO, Rome. Pinson, W.T., D.C. Yoder, J.R. Buchanan, W.C. Wright, and J.B. Wilkerson. 2004. Design and evaluation of an improved flow divider for sampling runoff plots. Trans. ASAE 20:433–438. Sheridan, J.M., R.R. Lowrance, and H.H. Henry. 1996. Surface flow sampler for riparian studies. Appl. Eng. Agric. 12:183–188. Toy, T.J., G.R. Foster, and K.G. Renard. 2002. Soil erosion: Processes, prediction, measurement, and control. John Wiley & Sons, New York.
Instrumentation for Measuring Runoff, Sediment, and Chemical Losses from Agricultural Fields
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
Carlos A. Bonilla,* David G. Kroll, John M. Norman, Daniel C. Yoder, Christine C. Molling, Paul S. Miller, John C. Panuska, Jeffrey B. Topel, Peter L. Wakeman, and K. G. Karthikeyan lection tanks are suitable, then the areas may not be representative of typical field conditions (Toy et al., 2002). Another common instrumentation technique continuously measures and records the runoff rate using a weir or flume in conjunction with a water-level sensor and datalogger. An automated pumping sampler is used to draw samples from just downstream of the weir or flume, communicating with the datalogger to tie the sample concentration to the flow volume it represents, permitting a mass movement calculation. Such systems are valuable for studies needing time-varying concentration values, but may be overkill for studies requiring only total storm mass values. A system like this can be expensive because it consists of a flume with substantial installation effort, water level sensor, datalogger, and automatic sequential sampler. In addition, such systems assume that samples extracted at specific times are representative of larger time frames, and may cause excessive sedimentation due to the flume constriction (Hamlett et al., 1984). Slot samplers offer an alternative to flumes and total collection tanks and can provide a compromise solution of modest cost for a wide range of plot sizes, from small watersheds up to an area of several km2 (PAP/RAC, 1997). Slot samplers obtain a representative portion of the runoff–sediment mixture. Two common slot samplers are the stationary multislot divisor and the Coshocton wheel sampler, which is equipped with a revolving slot. The stationary multislot divisor, first suggested by Geib (1933), routes the runoff from a collector through the multislot divisor, where a sample is obtained from a single slot and routed to a sample storage tank. A second or third sample storage tank may be connected to the first, if additional sample storage is needed (Brakensiek et al., 1979; PAP/RAC, 1997). The design of slot samplers has been refined over the years to overcome some of their limitations and make them more suitable for sediment transport and chemical loss studies. Sheridan et al. (1996) designed a low-impact flow event (LIFE) sampler to minimize disturbance and cost. This slot-type sampler was used to quantify nutrient concentrations in runoff flowing through riparian buffers in the Coastal Plain, where slopes are gentle and sheet flow is likely. Later, Franklin et al. (2001) modified the LIFE sampler to accommodate steep slopes (5–12%), and larger flow rates (1–5.5 L min21). The design consists of two sets of sample splitters (10 splitters for each set) connected in series. One-tenth of the inlet flow exits from the first set of 10 splitters and one-hundredth of the inlet flow from the second set, to be finally collected in a holding tank. Recently, Pinson et al. (2004) designed an improved stationary multislot divisor for sampling runoff plots. The system is built using a commercially available 19-L
ABSTRACT This work describes a simple, passive sampling system for measuring runoff, sediment, and chemical losses from typical agricultural fields. The sampler consists of a 5 to 7 m wide runoff collector connected to a series of multislot divisors. These divisors split the flow into aliquots, providing a continuous sampling during the runoff event. Divisors were located in a wooden box below ground level. With an adequate pump, this system can operate in fields with a slope gradient as low as 2%, and can stay in the field during winter to record first snowmelt-generated runoff. A radio transmitter reports by telemetry the occurrence and magnitude of any runoff event, and indicates when the system should be sampled and emptied. This article includes a description of the equipment, advantages, and disadvantages based on 2 yr of operation, and examples of data collected.
T
HE MEASUREMENT of sediment transport and chemical losses from farm fields has been a major research interest because of the need to characterize practices that conserve soil and water resources. Although many instruments have been developed in an attempt to measure runoff and sediment transport, at the present time none has proven to be fully satisfactory (PAP/RAC, 1997). For plots and small watersheds, total collection tanks, flumes and samplers, and slot divisors have commonly been used (Brakensiek et al., 1979; PAP/RAC, 1997; Toy et al., 2002). Total collection tanks may be constructed to measure erosion in small test plots, and must be large enough to contain the total runoff (water and sediment) expected in a 24- or 48-h period. The volume of the water–sediment mixture is then determined and the solid sediment material sampled for subsequent laboratory analysis and computation of its weight and volume. Unfortunately, total collection devices are often unsuitable for erosion studies because the runoff even for small plots can be excessive (Brakensiek et al., 1979; PAP/RAC, 1997). If plots are small enough so total col-
C.A. Bonilla, D.G. Kroll, J.M. Norman, J.B. Topel, and P.L. Wakeman, Department of Soil Science, University of Wisconsin-Madison, 1525 Observatory Drive, Madison, WI 53706; D.C. Yoder, Biosystems Engineering and Soil Science; University of Tennessee, 2506 E.J. Chapman Drive, Knoxville, TN 37996-4531; C.C. Molling, Cooperative Institute for Meteorological Satellite Studies, Space Science and Engineering Center, University of Wisconsin-Madison, 1225 W. Dayton Street, Madison, WI 53706; P.S. Miller, J.C. Panuska, and K.G. Karthikeyan, Biological Systems Engineering, University of Wisconsin-Madison, 460 Henry Mall, Madison, WI 53706. The company names are for the convenience of the readers and do not imply an endorsement by the University of Wisconsin-Madison. Received 19 Apr. 2005. *Corresponding author ([email protected]). Published in J. Environ. Qual. 35:216–223 (2006). Short Communications doi:10.2134/jeq2005.0130 ª ASA, CSSA, SSSA 677 S. Segoe Rd., Madison, WI 53711 USA
216
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BONILLA ET AL.: RUNOFF, SEDIMENT & CHEMICAL LOSS FIELD MEASUREMENTS
(5-gallon) bucket with a screw-top lid. A series of 22.5j V-slot weirs are precision laser cut into a strip of metal, which is then rolled into a ‘‘crown’’ and fastened to the top of the lid, from which the center has been removed. When used, plot runoff is routed into the bucket. The bucket fills completely and overflowing water–sediment is evenly divided among many slots. Flow from one of the overflowing slots is then collected in another bucket, and more buckets with crowns can be placed in series to divide runoff for larger storm events or larger size plots. The objective of this article is to describe the design, installation, and operation of an effective system suitable for measuring runoff, sediment, and chemical losses from within agricultural fields or near their edges. The system is a significant advance in that it makes use of the Pinson et al. (2004) multislot divisor; furthermore, we believe this new design to be superior to that of Franklin et al. (2001). This new design accommodates a wide range of runoff rates and amounts, is convenient because of the use of 19-L buckets, can handle collector widths of 5 m or more for better landscape sampling, allows studies near discharge outlets where slopes are small, and can be used for contributing areas as large as 0.4 ha (1 acre).
DESIGN AND INSTALLATION The runoff–sediment sampler has three main components: (i) the collector, which intercepts and routes the runoff into the multislot divisors; (ii) the multislot
217
divisors, which provide a flow proportional sample of sediment and runoff; and (iii) the electronic and data storage devices that record the occurrence and magnitude of the runoff events.
The Collector The collector, typically 5 to 7 m wide (Fig. 1 and 2), intercepts and routes the runoff into the multislot divisors. The water flow is directed to the collector outflow pipe by two boards that form the sides of a funnel. A plastic sheet over the collector floor stops infiltration losses, and a heavy duty tarp installed as a roof prevents direct rainfall from entering the system. The plastic sheet that forms the collector floor is installed in a trench dug down-slope of the 5 to 7 m long inlet side and buried so that the inlet to the collector forms a smooth transition from the soil to the plastic with minimal disturbance. After the inlet side is installed, the side boards are ‘‘rolled’’ up in the plastic sheet until they are vertical and extend from the corner of the inlet side to the outlet pipe of the collector (see Fig. 3 for pictures of a field installation). Rolling the plastic on the side-boards ensures an excellent seal all along the sides of the collector walls. The side boards are fastened to metal stakes and stretched out to keep the floor straight and flat, avoiding wrinkles and reducing sedimentation in the collector. The stakes are pounded in the ground about 2 or 3 cm back from the side board so that 5-cm long bolts can be used to pull the side boards toward the stakes, stretching the plastic floor. The roof tarp is
Fig. 1. Field installation. The collector (on the left) shows the wooden-frame and the plastic floor. When finished, the collector is covered with a heavy-duty tarp. The bottom truss on the front support for the collector roof comes down to 8 cm above the soil and is 15 cm upslope of the front edge of the plastic floor. A PVC pipe (in the center) carries the flow to the multislot divisors (on the right). Only the first divisor is above grade level so that the excess water is routed directly down the slope. The excess water from the other divisor heads flows to the sump, where a battery operated pump (114 L min21 [30 gallons min21] capacity) lifts the sump water to grade level for discharge.
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
218
J. ENVIRON. QUAL., VOL. 35, JANUARY–FEBRUARY 2006
Fig. 2. An illustration of the runoff collector used with the edge-of-field runoff collection system. The entire frame is covered with a heavy-duty tarp including the front gable area. The front of the collector cover extends to about 8 cm above the soil and 15 cm beyond the plastic floor to prevent rain-drop splash from entering the collector.
attached on the lower board of the ‘‘A-frame’’ (Fig. 1), stretched over a wooden frame, and a series of bungee cords keep the tarp tight, as well as facilitating removal
Fig. 3. Pictures of a field installation. (A) Transition of the collector floor and walls to the 15-cm diameter PVC pipe where the plastic floor is inserted into the pipe and retained with a split PVC ring within the pipe. A screen prevents field debris from getting to the divisor heads. (B) Picture of four-bucket divisor system under construction with plastic liner, pump circuit and battery, and neoprene boot construction around Bucket 1 (see text). (C) Completed divisor assembly with cover, pump outlet pipe, supply pipe with elbow, rain gauge, and container that protects the battery and pump circuit for sump operation. (D) Field installation for winter and spring snowmelt operation. The rain gauge and solar panel for charging the pump battery are visible next to the divisor collection system and the covered collector is at left.
of the roof tarp for checking and servicing (cleaning or patching) the collector interior. A 15-cm diameter (6-inch) PVC pipe is used to route the runoff from the collector to the multislot divisors, and the plastic floor is fed into the PVC pipe and held in place with a split ring inside the 15-cm pipe (Fig. 3A). Six vertical slots opened at the end of the elbow on the pipe avoid the accumulation of water in the PVC pipe if water freezes in the first bucket. Each slot is 6 cm high and 0.3 cm wide (see Fig. 4, detailed in Fig. 4A). The slots must be located in such a way that 3 cm are below the lower apex of the V-notch of the bucket, and the other 3 cm are above. These slots ensure normal operation of the divisor system when intermittent spring runoff occurs interspersed with freezing conditions that can lead to a thin layer of ice on the top of the water within the first bucket and inside the PVC pipe (which extends about 7 cm below the water surface level in the first bucket). Two conditions are of importance when installing the collector. First, the collector entrance must be located in an area where the topography concentrates the runoff, or where the occurrence of rills is evident. The selection of the site can be facilitated by using a detailed digital elevation map or visiting the field after a major rainfall event to check for rill occurrence. Alternatively, the collector may be used at the bottom of a field plot isolated by some type of borders, with those borders connecting directly to the ends of the collector. The second recommendation is that the collector must be located in a place that provides at least 2% slope along the side-boards and also along the PVC pipe. A
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
BONILLA ET AL.: RUNOFF, SEDIMENT & CHEMICAL LOSS FIELD MEASUREMENTS
219
Fig. 4. Box design and dimensions.
slope equal to or greater than 2% provides enough energy to the water to keep the sediment in suspension (Pinson et al., 2004), reducing the chance of deposition inside the collector or along the pipe. After 2 yr of operation, no sediment has ever been observed in the PVC pipe, even for lengths of 15 m.
runoff rate that can be handled with the 12-notch divisor is about 30 L s21, and 7 L s21 when using the 24-notch divisor (Pinson et al., 2004). The maximum measurable runoff volume with this configuration is close to 540 m3.
The Multislot Divisors From the collector, the runoff is routed into multislot divisors specially designed for this purpose (Fig. 4 and 5). This type of divisor system provides a storm-integrated, discharge-weighted sample for determining runoff and sediment yield by using a stationary slotted crown developed by Pinson et al. (2004). The crown is installed on the top of a 19-L bucket. When the first bucket is full, most of the effluent is discarded (11⁄12) and a small fraction (1⁄12) is routed to a second bucket (Fig. 3B). When the second bucket fills, a small fraction of the overflow is routed to a third bucket (1⁄24) and the remainder is discarded. Finally, when the third bucket fills, a fraction of the overflow (1⁄24) is routed to a 79-L (21-gallon) basin and the remainder is discarded. The maximum peak
Fig. 5. Illustration of (left) 24-slot and (right) 12-slot multislot divisors (all units are in cm).
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
220
J. ENVIRON. QUAL., VOL. 35, JANUARY–FEBRUARY 2006
The four buckets are located in a pressure-treated wooden box, buried with its top slightly above ground level (Fig. 3C). The box design and dimensions are given in Fig. 4. The frame is built with 3.5 by 3.5 cm (2 by 2 inch) boards but 3.5 by 8.5 cm (2 by 4 inch) boards form a base at the bottom to maintain buckets within 0.05j of level as they fill with water. This 3.5 by 8.5 cm base, which contains a sturdy shelf for Bucket 3, is installed first and leveled in the bottom of the hole so that the box rests on top of it. Three solidly braced shelves are installed in the box to support the buckets. The first bucket protrudes through the PVC tray so that excess water from the 11 divisor slots is drained down slope. A neoprene boot, which is attached to the tray and tightened around the neck of the bucket with a bungee cord, prevents excess water that is leaving the divisor slots and exiting the system from draining to the sump pump, thereby exceeding the 113 L min21 capacity of the pump (Fig. 4A). This neoprene seal (0.8 mm thick, 1⁄32 inches) is attached to the tray by sandwiching it between the tray and a PVC ring 36 cm i.d. When sampling Bucket 1, it can be lifted from above without disturbing the PVC tray by loosening the neoprene boot, lifting the 15-cm PVC supply pipe, and removing the elbow. Easy removal of Bucket 1 is important because sediment cannot be sampled directly from Bucket 1 in the field with satisfactory results (discussed later). Since Buckets 2 and 3 are positioned below grade level, the excess water from these buckets drains to a 79-L (21-gallon) bucket beneath the box, where it is pumped out with a sump pump. The power supply for the pump is a 12-V, deep cycle battery connected to a solar panel for continuous operation (Fig. 3B). A large deep-cycle battery with 82 amp-hour capacity can operate the pump continuously for 6 h. If the field topography provides enough slope gradient, water can be discharged from a hole in the bottom of the box and the pump is not needed. This system has been operating at a number of sites year-round in Wisconsin for 2 yr, quantifying runoff, sediment, and P losses from winter and spring snowmelt as well as spring, summer, and fall storms. Winter operation is accomplished by constructing a treated plywood box around the divisor head enclosure, insulated on the inside with 5-cm thick sheets of urethane foam (Fig. 3D). The wall of the wooden box is dug into the soil to a depth of about 5 cm. The dimensions of this box are 2.4 m long by 1.5 m wide by 0.6 m high, and it has a urethane foam cover consisting of two 5-cm thick sheets reinforced with thin metal strips. A tarp 3.7 m long and 2.4 m wide completes the installation cover, using staked bungee cords to stretch it tightly over the insulated frame. This design makes use of heat stored in the soil and prevents freezing inside if there is ample snow cover. Although a cold winter with little snow will result in subfreezing temperatures inside the double enclosure during the coldest time of the winter, in 2 yr this has not occurred during the times when snowmelt occurred. Thus, this passive, unheated system has worked for snowmelt monitoring with no apparent loss of data if samples are removed promptly (within 1 d) following the snowmelt event.
The precision of the runoff and sediment measurements depends on how well the V-notch weirs in each bucket are leveled. When building the divisor, the notches must be cut as precisely as possible, ideally with a laser device. Once in the field, the buckets must be leveled in place. In this setup, each shelf is provided with a triangular metal frame with a bolt in each corner (Fig. 4D). By turning the bolts, the divisor heads are leveled with a special machine level tool that is placed in the notches of the divisor head to obtain level within 60.05j. This device consists of an upper flat surface on which sits a precision level (30-5 cm [12-inch] machinist’s improved level no. 98-12, Starrett Co., Athol, MA), connected to a lower stainless steel rod (10 mm diam.) that sits in the divisor head slots. The device must be carefully assembled and machined to achieve the 60.05j leveling accuracy. Laboratory evaluation indicates that this system divides runoff with accuracies within 65% over most of the flow range and within 616% at very low and very high flows (Pinson et al., 2004). Laboratory measurements of sediment division accuracy indicated smaller errors than flow division (Pinson et al., 2004).
The Electronic and Data Storage Devices The divisor provides a storm-integrated, dischargeweighted sample for runoff. However, by installing a water content sensor (ECH2O-20 probe, Decagon Devices, Pullman, WA), in each bucket, the time-varying runoff rate can be estimated. If it is installed vertically, the water sensor reading can be converted to water volume for each 19-L (5-gallon) bucket with a sensor (Masarik et al., 2004), and once connected to a logger the flow rate over time can be calculated. If the logger includes a radio system, the runoff occurrence can be checked by telemetry. Alternatively, the logger could be connected to a cell phone and monitored from any location. On rare occasions the maximum capacity of a divisor system may be exceeded, especially with snowmelt on ground with concrete frost. When this occurs, information about the size of the event is lost, and one only knows that the event exceeded the maximum capacity of the system. By installing an ECH2O-20 probe vertically in the 15-cm PVC supply pipe about 50 cm upslope from the elbow that empties water into the first bucket, the water depth in the PVC pipe can be monitored. Knowing the water depth, the water flow rate can be calculated from Manning’s equation (Haan et al., 1994) if the slope of the pipe is known. Fiveminute averages of water depth can be recorded on a datalogger and calibration of the ECH2O-20 probe maintained using data that does not exceed the divisor system capacity. This system is being tested but insufficient experience has been obtained to verify its robustness.
Cost The total cost for supplies, labor, and instrumentation is about $5000 per installation (Table 1). About onethird of this cost is for the datalogger system to record water levels in buckets, rainfall, soil temperature, and soil moisture. About one-seventh of the total cost is divisor heads and buckets, which were purchased from
221
BONILLA ET AL.: RUNOFF, SEDIMENT & CHEMICAL LOSS FIELD MEASUREMENTS
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
Table 1. Materials, instruments, and labor required to install a runoff collector unit. Item
Cost
Instruments Multislot divisors (Biosystems Engineering and Environmental Science, Univ. of Tennessee, Knoxville, TN 37996-4531) Pump system Solar panel, 20 W (Campbell Scientific, Logan, UT 84321-1784) Heavy duty sump pump, 12 VDC, 11 amps (Basement Watchdog Special model BWSP, 1730 GPH @ 10 ft. Glentronics, Wheeling, IL 60090) Deep cycle marine battery, 82 amp-hour (local hardware store) Data collection system Data logger (EM5R logger, Decagon Devices, Pullman, WA 99163) Water content sensor (ECH2O, 20 probe, Decagon Devices, Pullman, WA 99163) Radio transceiver module (Rm1 radio module, Decagon Devices, Pullman, WA 99163) Tipping bucket rain gauge (Model TR-525I, Texas Electronics, Dallas, TX 75235) Pocket PC (iPAQ Pocket PC h2215, Hewlett-Packard Co., Palo Alto, CA 94304) Materials Pipes, elbows, metal stakes, boards, tarps (local hardware store) Labor Wooden box, installation Total
$700
the University of Tennessee. A considerable amount of labor is required to build the boxes and install the collector and divisor head box in the field.
OPERATION After every runoff event the depth of water in each bucket is measured using an aluminum meter stick graduated in millimeters (to allow for a volume calculation), and water samples are taken for analysis. Once in the laboratory, sediment samples are dried and weighed to determine the sediment mass. The total runoff volume for the entire event at the time of sampling is calculated by R 5 V1 1 (12 3 V2 ) 1 (12 3 24 3 V3 ) 1 (12 3 24 3 24 3 V4 )
[1] 3
where R is the total runoff volume (m ), and V1 to V4 (m3) are the volumes collected in Buckets 1 through 4, respectively. Total sediment mass for the entire runoff event at the time of sampling is calculated by S 5 (V1 3 C1 ) 1 (12 3 V2 3 C2 ) 1 (12 3 24 3 V3 3 C3 ) [2] 1 (12 3 24 3 24 3 V4 3 C4 ) where S is the total sediment (kg), and C1 through C4 (kg m23) are the concentration of total solids measured in Buckets 1 to 4, respectively. Chemical loads can also be calculated by using Eq. [2], but with chemical concentrations replacing sediment concentrations. When sampling sediment, experience has shown that it is not possible to obtain a representative sample from the soil suspension in the first 19-L (5-gallon) bucket in the field. The largest particles are in this bucket, and it is not uncommon to have a 5- to 10-cm layer of sediment in the bottom of the bucket. Thus, the first bucket is removed from the system entirely and returned to the laboratory for measurements; a new first bucket is then installed in the system each time samples are removed.
Alternative Configurations Various configurations of buckets and divisor heads provide flexibility for accommodating different sized contributing areas. For small contributing areas, the ac-
$800
$1700
$400 $1200 $4800
curacy of the measurements can be improved by reconfiguring the buckets and divisor heads based on the maximum runoff rate and the maximum runoff volume that are expected to occur. The maximum flow rate that can be handled with a 12slot divisor head is 29.7 L s21 (Pinson et al., 2004). In the field the maximum runoff rate can be estimated from the contributing area, the topographic features, and rainfall intensity. When deciding about a configuration, the peak flow should be obtained from a hydrograph; however, using a worse-case scenario approach, the maximum runoff rate can be predicted as the product of the rainfall intensity, the runoff coefficient for the field conditions, and the size of the contributing area. The maximum runoff volume that can be sampled depends on the number of buckets and the configuration of divisor heads. Table 2 and Fig. 6 show three of the multiple combinations that can be implemented in the field. Configuration 1 corresponds to the design that has been presented previously in the article, which accommodates large runoff rates and volumes (29.7 L s21, runoff volume . 30 m3). This configuration consists of three 19-L (5-gallon) buckets (Fig. 6A), with a 12-slot divisor in Bucket 1, and 24-slot divisors in Buckets 2 and 3, plus a 79-L (21-gallon) bucket to store 1⁄24 of the water exiting from Bucket 3. In particular for this configuration, as a result of surface tension effects at the bottom of the slots, flows lower than 0.02 L s21 in Bucket 3 produce errors larger than 616% when using the 24-slot divisor bucket (Pinson et al., 2004). Table 2 shows that the lowest flow rate into Bucket 1 that will assure a minimum divisor-head flow rate of 0.02 L s21 (error less than 616%) in Buckets 3 is 5.76 L s21. Thus, on a 0.5-ha field, the runoff rate should be larger than about 4 mm h21 to ensure a flow rate into Bucket 3, which is . 0.02 L s21. For small fields the system can be modified to decrease the minimum runoff rate. The slots in Bucket 3 can be Table 2. Maximum runoff capacity and minimum flow rate into Bucket 1 for the three configurations shown in Fig. 6. Configuration 1, Midsize fields 2, Small fields 3, Plots
Maximum runoff capacity
Minimum flow rate
L 554 000 28 600 5 690
L s21 5.76 0.24 0.24
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
222
J. ENVIRON. QUAL., VOL. 35, JANUARY–FEBRUARY 2006
Fig. 6. Three divisor heads configurations suitable for different sized contributing areas: (A) Mid-size fields (. 0.4 ha), (B) small fields (0.04–0.4 ha), and (C) plots (, 0.04 ha).
covered (e.g., using tarp tape), connecting directly Buckets 3 and 4 (Fig. 6B). This bridge essentially gives the third bucket a 99-L (26-gallon) capacity, reducing the minimum runoff rate requirement from 5.76 L s21 (configuration 1) to 0.24 L s21 (Table 2, configuration 2). The maximum runoff capacity for this configuration is close to 30 m3. With configuration 2 the runoff rate on a 0.5-ha field should be larger than about 0.17 mm h21 to ensure a flow rate into Bucket 3 that is . 0.02 L s21. For plots with runoff volumes smaller than 6 m3, a combination of three 19-L (5-gallon) buckets can be used (Fig. 6C). A 12-slot divisor is used in Bucket 1, and a 24-slot divisor in Bucket 2. The third bucket is used to store 1⁄24 of the water exiting from Bucket 2. This
configuration operates with a minimum flow rate into Bucket 1 as low as 0.24 L s21 (Table 2, configuration 3).
EXAMPLE APPLICATION Eight of these runoff–sediment collectors have been installed in different agricultural fields in Wisconsin, and they have been operational since July 2003. More than 300 site events have been collected to date, and Table 3 contains a sampling of events from various sites. To illustrate the operation of the system, one additional event is shown below in detail. The runoff event took place on 11 June 2004 in a field located in west-central Wisconsin. The field consists of a 0.5-ha
Table 3. Sampling of runoff events collected from various sites. Contributing area
Texture
ha
Slope
Crop
silt loam
13
0.10
silt loam
0.03
alfalfa
62.0
7
corn
96.6
clay loam
5
corn
28.9
0.05
loam
8
alfalfa
128.3
0.21
loam
8
alfalfa
109.6
0.17
sandy loam
10
Event
mg kg21
%
0.50
Bray P-1
corn
63.6
Runoff volume
Total P
no. buckets
g L21
kg ha21
0.01 2.66 7.23
0.5 2.4 3.0
0.83 1.52 3.10
0.02 8.07 44.60
3.70 0.18 0.26
6.39 2.43 1.94
0.01 4.26 7.23
0.4 2.6 3.0
1.23 3.55 1.27
0.10 150.00 91.60
0.37 0.62 0.37
1.81 1.44 1.28
0.01 10.48 0.59
0.6 3.0 2.1
0.14 2.56 0.32
0.06 829.40 5.82
0.18 0.08 0.03
0.58 1.13 0.34
0.03 0.23 1.73
1.1 1.8 2.2
0.78 0.74 0.35
0.53 3.47 12.16
1.35 1.05 0.38
2.03 1.72 1.06
0.02 1.72 272.70
1.0 2.2 3.5
3.10 38.00 0.58
0.33 1.23 757.80
0.52 0.80 0.50
7.47 1.48 1.05
0.01 3.18 101.80
0.3 2.5 3.2
0.61 13.90 0.00
0.03 262.30 2.11
1.29 0.69 0.68
1.59 5.64 0.90
m3 Site 1 storm 24 May 2004 storm 11 June 2004 snowmelt 22 Mar. 2004 Site 2 storm 15 June 2004 storm 30 Mar. 2004 snowmelt 22 Mar. 2004 Site 3 storm 23 Nov. 2004 storm 11 June 2004 snowmelt 9 Mar. 2004 Site 4 storm 23 May 2004 storm 31 Dec. 2003 snowmelt 2 Mar. 2004 Site 5 storm 26 Mar. 2004 storm 26 Mar. 2004 snowmelt 3 Jan. 2005 Site 6 storm 22 Nov. 2004 storm 22 May 2004 snowmelt 9 Feb. 2005
Dissolved reactive P
Sediment
mg L21
BONILLA ET AL.: RUNOFF, SEDIMENT & CHEMICAL LOSS FIELD MEASUREMENTS
Table 4. Example of data collected during an actual rainfallrunoff event. Collected volume
Sampled volume
Reproduced from Journal of Environmental Quality. Published by ASA, CSSA, and SSSA. All copyrights reserved.
L Bucket 1 Bucket 2 Bucket 3 Bucket 4 Total
23.3† 23.3† 8.18 0.0
23.3 280 2360 0.0 2663
Sediment g L21 7.50 2.13 1.39
Dissolved reactive P
Total P
mg L21 0.13 7.51 0.17 2.88 0.18 2.33
† The volume is larger than 19 L (5 gallons) because of the additional depth associated with the divisor head.
223
This instrumentation has been used successfully for almost 2 yr in farm fields ranging in size from 0.04 to 0.5 ha. Since a small weighted fraction of the runoff is collected during the entire runoff event, this system can sample from larger areas than the systems designed for the smaller standard unit plots (4 by 22 m, ,0.01 ha). Thus, this proposed system works on fields that are more representative of typical farm fields; however, determining the contributing area of the collector using detailed topographic maps can be challenging. ACKNOWLEDGMENTS
contributing area planted with alfalfa under no tillage for many years. The average slope is 12%, with a minimum of 0.6% and a maximum of 23%. The typical soil texture is silt loam, with 4% organic matter. The rainfall event totaled 35 mm, with a maximum intensity of 6.4 mm in 30 min. The erosivity index for the storm (EI30) was calculated to be 75.3 MJ mm ha21 h21 in metric units or 4.4 hundreds ft tonf in acre21 h21 in U.S. customary units. At the end of the event the first two buckets were full of water, and the third bucket was half full. The water volume, sediment, dissolved reactive P, and total P measured in each bucket are presented in Table 4. The total runoff from the contributing area was 2.7 m3 (calculated using Eq. [1]). A total of 4.05 kg of sediment was transported through the collector during the event, with an average concentration of 1.52 g L21. For P, 0.47 g of dissolved reactive P were transported, with an average concentration of 0.18 mg L21. Finally, 6.46 g of total P were moved during this event, with an average concentration of 2.43 mg L21. When expressed as a function of contributing area, 8.10 kg ha21 of sediment was eroded (approximately 7.2 lb acre21), and 0.93 g ha21 and 12.87 g ha21 of dissolved reactive P and total P were transported, respectively.
CONCLUSIONS The instrumentation presented in this article provides a cost-effective passive sampling system for measuring runoff, sediment, and chemical losses under snowmelt or storm conditions in farm fields. Measurements can be made where there is no external source of power and in remote locations on operational farms, even near a discharge outlet where slopes are usually small.
This research was partially supported with funds from the USDA Cooperative State Research, Education, and Extension Service’s Water Quality Program under grant no. 2002-5113001951, and the Wisconsin Buffer Initiative funded by the Natural Resources Conservation Service under grant no. 685F48-3-085 and 68-5F48-4-082 and the Wisconsin Department of Natural Resources with funding from the Bureau of Watershed Management, Runoff Management Section for NonPoint Source Pollution Abatement.
REFERENCES Brakensiek, D.L., H.B. Osborn, and W.J. Rawls (ed.). 1979. Field manual for research in agricultural hydrology. USDA Agric. Handb. 224. USDA, Washington, DC. Franklin, D.H., M.L. Cabrera, J.L. Steiner, D.M. Endale, and W.P. Miller. 2001. Evaluation of percent flow captured by a small in-field runoff collector. Trans. ASAE 44:551–554. Geib, H.V. 1933. A new type of installation for measuring soil and water losses from control plots. J. Am. Soc. Agron. 25:429–440. Haan, C.T., B.J. Barfield, and J.C. Hayes. 1994. Design hydrology and sedimentology for small catchments. Academic Press, San Diego, CA. Hamlett, J.M., J.L. Baker, S.C. Kimes, and H.P. Johnson. 1984. Runoff and sediment transport within and from small agricultural watersheds. Trans. ASAE 27:1355–1363, 1369. Masarik, K.C., J.M. Norman, K.R. Brye, and J.M. Baker. 2004. Improvements to measuring water flux in the vadose zone. J. Environ. Qual. 33:1152–1158. PAP/RAC. 1997. Guidelines for mapping and measurement of rainfallinduced erosion processes in the Mediterranean Coastal Areas. PAP-8/PP/GL.1. Split, Priority Actions Programme Regional Activity Centre (MAP/UNEP), with the cooperation of FAO, Rome. Pinson, W.T., D.C. Yoder, J.R. Buchanan, W.C. Wright, and J.B. Wilkerson. 2004. Design and evaluation of an improved flow divider for sampling runoff plots. Trans. ASAE 20:433–438. Sheridan, J.M., R.R. Lowrance, and H.H. Henry. 1996. Surface flow sampler for riparian studies. Appl. Eng. Agric. 12:183–188. Toy, T.J., G.R. Foster, and K.G. Renard. 2002. Soil erosion: Processes, prediction, measurement, and control. John Wiley & Sons, New York.
