Flow-Control Plates to Manage Denil Fishways in Irrigation Diversions for Upstream Passage of Arctic Grayling

Recent conservation efforts in the Big Hole River valley in southwestern Montana offer an interesting example of cooperative efforts between irrigators, recreationalists, and conservation agencies to support Arctic Grayling Thymallus arcticus recovery while preserving traditional farming and ranching activities. The Big Hole River and its tributaries form a network of blue-ribbon trout streams that is home to five species of game fish: Brook Trout Salvelinus fontinalis, Brown Trout Salmo trutta, Cutthroat Trout Oncorhynchus clarkii, Rainbow Trout Oncorhynchus mykiss, and the last population of native fluvial Arctic Grayling in the contiguous United States, making it a popular angling destination for locals and tourists. The valley also boasts over 50,000 ha of crop-producing farms (Montana Department of Agriculture 2019), largely devoted to forage production, either alfalfa hay or pasture, and small grains.

Much of the water that supplies valley streamflow comes from winter snowfall in the surrounding mountains. In some years snowfall is abundant and there is gradual melt and release to surface waters through the spring and summer such that conflicting demands on summer stream flow are minimal. In other years the opposite occurs, low snowfall and early runoff, resulting in conflicting water demands for irrigation and in-stream flow. To address this potential water use conflict, conservation agencies worked with irrigators to implement actions that increased instream flows, a key strategy outlined in a voluntary Arctic Grayling Candidate Conservation Agreement with Assurances, or CCAA (Montana Fish, Wildlife and Parks and U.S. Fish and Wildlife Service 2006). Development and implementation of the CCAA in the Big Hole River valley was primarily to address concerns about the status of Arctic Grayling in the watershed, a species of concern in Montana (Montana Natural Heritage Program and Montana Fish, Wildlife and Parks 2022). Ancillary benefits to stream ecosystem function and the trout fishery were a desirable by-product of the CCAA activities.

Irrigators in the valley have adopted or refined many activities to seek a sustainable balance between agricultural production and the health of the fishery because of the CCAA. These activities included enhancing riparian and stream habitat, stock-water facility improvements, advances in irrigation scheduling, enhancements to on-farm irrigation delivery systems, and in-stream diversion management to improve migration routes throughout the watershed (Lamonthe 2009). The CCAA participants were particularly interested in ways to design and operate in-stream irrigation diversions that provided sufficient water for irrigation demands while providing for fish mobility. Fish mobility and stream connectivity is essential to meet life history needs such as spawning, seeking thermal refuge, or pursuing food (Silva et al. 2017).

Irrigation diversions are in-stream structures used to control the water surface elevation near a ditch or canal to maintain flows necessary for irrigation. These diversion structures vary in size and government agencies, irrigation user groups, or individual irrigators may regulate them. In the Big Hole Valley, the diversions are typically manmade structures less than 2 m tall that are regulated by individual irrigators and span the width of the stream (Figure 1). Irrigators can add or remove cross-boards in the weir, or planks, to increase or lower the water surface level in the upstream pool. This adjustability allows the irrigator to respond to stream flow variation while maintaining adequate canal flow. The diversion shown in Figure 1 also has a Denil fishway installed to facilitate upstream movement of fish. CCAA participants have installed 63 Denil fishways at irrigation diversions throughout the Big Hole River watershed as part of the CCAA.

Denil fishways are a type of technical fishway that have closely spaced baffles to reduce water velocity and redirect flow by exerting a retarding shear stress on the fluid moving through the fishway (Katopodis 1992). The resulting flow is highly turbulent but has a lower average velocity than an unbaffled chute. Denil fishways exhibit a low water velocity near the floor of the fishway that increases vertically. Engineers and fisheries personnel believe that the low-velocity zone, which occupies a larger area at higher flow rates, provides an ideal passage route for fish travel in the upstream direction. Researchers have noted successful fish passage through Denil-type fishways for a range of species (Schwalme et al. 1985; Haro et al. 1999; Mallen-Cooper and Stuart 2007) including Arctic Grayling (Tack and Fisher 1977; Blank et al. 2021; Plymesser et al. 2022; Triano et al. 2022).

During times of abundant streamflow in the Big Hole River system there is adequate water to supply the irrigation system while maintaining sufficient flow in the fishway (Figure 1). During times of very low streamflow no water cascades over the diversion weir, the fishway is only partially flowing, and it is possible that the irrigator is not able to maintain the water level required in the pool upstream of the weir to deliver adequate irrigation flows (Figure 2). The challenges presented during times of limited water availability were the motivation for this study. To avoid substantial modifications to existing fishways, irrigators and agency staff proposed installing a plate at the upstream end that restricted the amount of flow to the fishway. Through a series of laboratory-based experiments with treatments consisting of three different flow-control plates and a control (the Denil fishway with no plate served as the control), we investigated 1) the extent to which each treatment reduced flow compared to the control, and 2) the extent to which each treatment impacted Arctic Grayling passage efficiency defined as the ratio of the number of Arctic Grayling that fully ascended the fishway to the number that were attracted to the fishway relative to the control.

Our research team carried out a series of experiments in an open-channel flume at the U.S. Fish and Wildlife Service (USFWS) Bozeman Fish Technology Center (BFTC) in Bozeman, Montana, from July through September of 2017. We tested three different flow-control-plate treatments and a control (Figure 3). Our conversations with representatives of Montana Fish, Wildlife and Parks, USFWS, and the Montana Department of Natural Resources and Conservation who worked with irrigators in the project area informed design of the flow-control plates. We selected the standard (S) plate because irrigators and agency personnel had installed these plates at a few of the Denil fishways in the Big Hole River valley. The Denil slot (DS) plate has the same open-area geometry as a standard Denil fishway baffle and we selected it as a plate that may have a small reduction in flow but retain the fundamental hydraulics of the fishway. We chose the narrowed Denil slot (NDS) plate because there was potential for a marked reduction in flow while retaining some elements of the Denil baffle slot geometry.

The BFTC open-channel flume that we used was approximately 18.3 m in length and 0.9 by 0.9 m in cross-section (Dockery et al. 2019) with a 3.7-m-long Denil fishway installed in the flume at a 12(H):1(V) slope; this drop resulted in a slope of 8.2%, which has effectively passed Arctic Grayling in previous laboratory tests (Blank et al. 2021; Plymesser et al. 2022) and is consistent with design standards used by fish passage practitioners in the Big Hole River watershed (Montana Fish, Wildlife and Parks and Natural Resources Conservation Service, personal communication). Design guidance for Denil fishways recommends maximum slopes of 16.7% (USFWS 2019). Pumps moved fresh water from a nearby spring into the upstream end of the flume and recirculated it at the prescribed flow rate for each trial. The flume design allowed flexible placement of a passive integrated transponder (PIT) antenna system and temperature and water depth data loggers; the flume had a tarp cover to prevent disturbance of test fish and maintain uniform lighting. A detailed description of the research flume that we used for the experiments is in Dockery et al. (2019).

The Arctic Grayling embryos were the offspring of brood stock collected from Red Rock Creek in southwestern Montana, in May 2014. Fisheries personnel transported the embryos to the BFTC where they hatched in a vertical, flow-through incubator and then moved to a 150-L tank supplied with 10–12°C spring water (Davis 2016). A belt-fed system delivered a starter diet to the fry. Once weaned off the starter diet, the fish received 3.5 mm floating feed Classic Trout diet (Skretting USA, Tooele, UT) daily. At age 0+ fisheries personnel transferred the fish to a 20-m-circumference artificial river (fabricated by Hydro Composites, LLC, Stockdale, TX). The oval-shaped artificial river was composed of two 6.42-m-long, 1.56-m-wide straight sections that connected two simulated channel bends with an outside diameter of 5.81 m and a center wall separating the channels. Water was circulated in the tank and velocities varied spatially from 0 to 0.9 m/s. The water source was a natural spring with water temperatures between 8.0 and 13.1°C. Between hatching in 2014 and the restrictor-plate experiments in 2017, researchers used the Arctic Grayling in two other experiments; one in 2015 to quantify swimming performance (Dockery et al. 2019) and another in 2016 to determine passage success through a standard Denil fishway (Blank et al. 2021). The researchers held any Arctic Grayling not actively participating in a research study in the artificial river. BFTC technicians conducted all care and feeding of the Arctic Grayling. In 2017 the Arctic Grayling were age 3+ hatchery fish and represented an adult-size wild counterpart; wild Arctic Grayling are rare and not available for laboratory research purposes. Arctic Grayling reared and exercised in captivity have been swim tested and demonstrated abilities like their wild counterparts (Cahoon et al. 2018; Dockery et al. 2020).

In May 2017 we transferred approximately 400 Arctic Grayling to a 3-m-diameter outdoor tank near the open-channel flume. The holding tank was equipped with external pumps to maintain water velocities that varied spatially from 0 to 0.9 m/s; we maintained water temperature at approximately 12°C and kept feed rations unchanged. To identify and track individual fish, we inserted 12-mm PIT tags (HDX12 Pre-load; Biomark, Boise, ID) into the peritoneal cavity through a small needle incision located just anterior to the anal fin in 2015. During the tagging process, we used MS-222 to lightly anesthetize subject fish. We identified individual subjects using a Destron Fearing 601 Handheld PIT tag reader (Destron Fearing, DFW Airport, TX) prior to each trial.

Prior to each trial we randomly selected 10 fish from the holding tank, recorded the individual PIT tag identification numbers, and transferred them to the staging area. We omitted and replaced subjects from the trial that appeared to have damage to fins or otherwise had indications of poor health. We held the 10 fish we selected for a trial in a staging tank for approximately 24 h without feeding. We controlled water temperatures in the staging tanks and flume by adjusting the amount of cold spring water added to the recirculating system to maintain a target temperature of 12°C (Table 1; range = 11.5–12.9°C) We recorded temperatures before and after each trial using a handheld thermometer.

At the beginning of an experimental trial, we moved fish from the staging tank to the tail-water box at the downstream end of the fishway. We then covered the flume–fishway system with a tarp and left the test subjects to move volitionally for 2 h. During experimental trials we monitored fish movement using a Multi-Antenna HDX Reader (Oregon RFID, Portland OR) attached to two antennas at the upstream and downstream ends of the Denil fishway. Upon completion of the 2-h trial, we inserted screens to restrict further movement of the Arctic Grayling, removed the tarp, and recorded the physical location of the subjects as upstream, downstream, or inside of the Denil fishway. We then removed the fish from the flume and measured, weighed, and placed them into a separate holding tank (Table 1; Data S1, Supplemental Material).

In summary, we selected 200 fish from a holding tank of 400 and swam them in 20 trials, with 10 fish placed in the raceway simultaneously for each trial. We used 10 fish in the trials based on our experience with Arctic Grayling research (Blank et al. 2021) that showed adequate participation while allowing us a reserve number of fish for repeating an interrupted trial or removing fish that appeared unhealthy. We performed five trials for each of the three flow-restrictor plates and the control at replicated depths measured at the notch of the flow plate or first baffle in the case of the control.

In each trial we set a prescribed upstream fishway depth by adjusting the flow rate into the system. The depth of water flowing over the first baffle notch or flow-plate notch at the Denil fishway entrance was the concern; the height of the boards in relation to the Denil fishway placement regulated notch depth. Thus, placement of boards regulates notch depth in the case of the flume research or in an irrigation diversion application; water will back up on the dam boards and fill the Denil fishway and the boards will shunt it to the diversion canal. Flow will pass over the boards once it exceeds the Denil height entrance or volume. Noteworthy is the coupling of depth over the first baffle notch and flow rate through the Denil fishway, in that they increase or decrease in relationship. Researchers have observed Denil ladders placed in irrigation diversions under conditions varying from no water passing over the notch to completely submerged (Triano et al. 2022; Figures 1 and 2). The study notch depth test conditions varied from ∼0.1 to ∼0.35 m. The range of test conditions represent field conditions varying from optimal to suboptimal hydraulics for fishway operation (Katopodis 1992); the minimum tested notch depth provided enough water depth to allow a fish to ascend the fishway without leaping or air exposure, while the maximum depth represented a full-flowing Denil ladder functioning near optimal hydraulic conditions. We measured the notch depth vertically from the lowest point in the open space of the most upstream baffle or flow-control plate to the water surface. We set the prescribed downstream fishway depth using backwater weir boards inserted at the downstream end of the flume. Water depth data loggers (WT-HR, TruTrack; GEO Scientific Ltd., Vancouver, Canada) in the headwater and tail-water boxes provided continuous depth and temperature measurements over the duration of each trial. Note that the upstream depths were similar within the trials for each flow-control-plate treatment but do not match the control where the zero-flow inlet water surface elevation is slightly lower because the first baffle opening within the Denil is recessed into the sloping fishway (Table 1). We used the U.S. Geological Survey midsection method to determine the flow rate in the flume (Rantz 1982; Rantz and Peck 1982) after each trial.

We defined passage success as the ratio of the number of fish that fully ascended the fishway to the number of fish attracted to the fishway and participating in the study (Table 1). We considered fish to be participants if they registered on the PIT antenna at the downstream end of the fishway and considered them to be passage successes if they registered on the upstream PIT antenna. There was only one instance in which we observed a fish upstream of the fishway that did not register on the upstream PIT antenna. Because of its location at the end of the trial we counted it as a successful pass.

We did not intend water temperature and size of the fish as indicated by the fish mass and the fish length (Table 1) to be variables in this experimental design. However, due to random sampling of fish for each trial and the imperfect method of controlling water temperature over a month-long experiment, we checked for lack of homogeneity of these observations using ANOVA as part of the model-building process. We completed this analysis using SigmaPlot (Systat Software, San Jose, CA). The mean water temperature failed the equal variance test, so we used ANOVA on ranks. There was no evidence for differences in water temperature between the control and the DS treatment (P = 0.265), between the control and the S treatment (P = 0.786), and between the control and the NDS treatment (P = 0.841). There was no evidence for difference in fork length (P = 0.265) or fish mass (P = 0.091) from a standard ANOVA. Within trials, we noted no trends in the size of successful fish compared to unsuccessful fish. Based on this analysis we discounted water temperature, fish length, and fish mass as candidate model variables.

All treatments exhibited a nonlinear relationship between flow rate and inlet water depth (Figure 4; Table 2). Examining the P values for the coefficients in equation (1), there is no evidence that the DS (P = 0.305) and the S (P = 0.961) treatments effectively reduced flow when compared to the control. Confirming the visual presentation (Figure 4) there was strong evidence that the NDS treatment had a significantly lower flow capacity than the control (P = 0.005).

In all trials except one (Trial 10) at least 1 of the 10 fish introduced to the downstream pool participated in a passage attempt (Table 2). Over all trials the average participation rate was 77%. Trials 7, 11, 16, and 18 had participants but no successful passage. Over the entire experiment, including treatments and control, 42.2% of fish that participated successfully passed the Denil fishway. The passage rate, defined as the number of fish passing divided by the number fish participating, was 51.4% for the control, 57.4% for the DS treatment, 26.4% for the NDS treatment, and 28.9% for the S treatment. Passage success generally increased with flow for all plate treatments and was 0 for the lowest flow treatment in all plate treatments (flow ≤ 0.014 m³/s; Table 1; Figure 5).

We assumed the mixed effect binomial logistic regression random intercept for trial (α_i) to be normally distributed with a mean of 0 with an estimated variance of 1.035. There was a lack of evidence for an interaction between plate and flow and the interaction term was not included in the inferential model (χ² = 2.32; df = 3; P = 0.51). There was strong evidence (Figure 5; Table 3) that passage success was positively associated with flow (P < 0.0001). There was moderate evidence that the standard (S) plate had reduced passage in comparison to the control (P = 0.03). There was a lack of evidence that passage differed between the Denil plate treatment (DS) and the control (P = 0.92) and the narrowed Denil (NDS) and the control (P = 0.49). Given the limited sample size, lack of replication, large residual deviance, readers should view results cautiously.

We conducted this study to determine if plates installed on the inlet of Denil fishways reduced the flow of water through the fishway and if the plates affected passage success of Arctic Grayling moving in the upstream direction. We first note that it requires a dramatic restriction in cross-section for the flow-control plate to substantially reduce flow (Figure 4). Both visually and statistically it is evident that the control, the DS plate, and the S plate all had similar rating curves when evaluated from an irrigator's standpoint, the flow rate as a function of inlet water depth. This alone may negate any interest in using the S or DS plates to regulate water flow through the fishway. The NDS plate noticeably and statistically reduced flow as would be expected with a weir having a slot width that is only 15.2 cm wide.

A desirable Denil restriction would be a plate that reduces flow through the Denil fishway while maintaining effective fish passage. If we consider the results of the logistic analysis, the S treatment passed fish less effectively than did the control but there was not evidence the DS and NDS treatments reduced passage compared to the control. When seeking ways to reduce water flow through the Denil without compromising fish passage, we recommend that the S plate not be used because it reduces fish passage and the DS plate not be used because it does not substantially reduce water flow. The NDS plate did reduce water flow and there was a lack of evidence that the NDS plate reduced passage in relation to the control. This type of plate design may warrant further research. Alternatively, a scaled Denil fishway is a possible replacement to standard-sized Denil fishways to enhance upstream mobility of Arctic Grayling in small, water-limited streams (Plymesser et al. 2022) to achieve balance between irrigation water needs and fish passage goals.

We contend Figure 5 portrays the irrigation management implication of the study most strongly, where over all treatments there was a positive relationship between water flow rate and Arctic Grayling passage success. This relationship suggests that an adequate supply of water ensures that a Denil fishway is an effective fishway. In the overall analysis the plate treatment was not a significant indicator of passage success, but flow rate was; we contend that it is not prudent to reduce the flow in a Denil fishway using any of the plate geometries we evaluated. We recognize that our recommendation to maintain adequate flow in the fishway may leave some irrigators in low-flow situations with a compromised ability to divert their water. In response, we recommend first that irrigation managers examine the diversion structure itself for effectiveness. Diversion structures can leak or have substantial piping through the coarse streambed on which the diversion rests. Some diversions have flow paths that circumvent the weir by flowing around it as surface flow or have collected debris in a way that reduces the utility of the diversion. Addressing these issues may lessen the need for flow reduction through the Denil and irrigation managers should consider this possibility. There are likely also cases where the diversion itself is well maintained and functional, but the standard-size Denil fishway requires too much flow to ensure adequate fish passage and irrigation flow. In this case modifications to the Denil inlet that allow any fish passage at all are preferable to blocking the inlet entirely. Use of a plate like the NDS treatment may be a good stopgap because it reduces flow and there is evidence that fish pass the plate as well as a lack of evidence that passage differed between the NDS plate and the passageway without a plate attached.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. An archive of data observed in the study trials conducted at the Bozeman Fish Technology Center, Bozeman, Montana, in 2017.

Available: https://doi.org/10.3996/JFWM-22-041.S1 (19 KB XLSX)

Reference S1. Davis MN. 2016. Winter survival and habitat as limiting factors for Arctic Grayling at Red Rock Lakes National Wildlife Refuge. Master's thesis. Bozeman: Montana State University.

Available: https://doi.org/10.3996/JFWM-22-041.S2 (15.575 MB PDF)

Reference S2. Katopodis C. 1992. Introduction to fishway design. Winnipeg, Canada: Freshwater Institute Central and Arctic Region, Department of Fisheries and Oceans.

Available: https://doi.org/10.3996/JFWM-22-041.S3 (864 KB PDF)

Reference S3. Lamonthe P. 2009. Candidate conservation agreement with assurances for fluvial Arctic Grayling in the Upper Big Hole River. Montana Fish, Wildlife and Parks. 2009 Annual Report.

Available: https://doi.org/10.3996/JFWM-22-041.S4 (714 KB PDF)

Reference S4.Montana Fish, Wildlife and Parks and U.S. Fish and Wildlife Service. 2006. Candidate conservation agreement with assurances for fluvial Arctic Grayling in the Upper Big Hole River.

Available: https://doi.org/10.3996/JFWM-22-041.S5 (3.991 MB PDF)

Reference S5. Rantz SE. 1982. Measurement and computation of streamflow. Volume 1, measurement of stage and discharge. Washington, D.C.: U.S. Geological Survey Water Supply Paper 2175.

Available: https://doi.org/10.3996/JFWM-22-041.S6 (21.079 MB PDF)

Reference S6. Rantz SE, Peck DL. 1982. Measurement and computation of streamflow. Volume 2, computation of discharge. Washington, D.C.: U.S. Geological Survey Water Supply Paper 2175.

Available: https://doi.org/10.3996/JFWM-22-041.S7 (21.960 MB PDF)

Reference S7. Tack SL, Fisher JG. 1977. Performance of Arctic Grayling in a twenty foot section of model “A” Alaska steeppass fish ladder. Final Report to the Army Corps of Engineers, Alaska Division.

Available: https://doi.org/10.3996/JFWM-22-041.S8 (812 KB PDF)

Reference S8. [USFWS] U.S. Fish and Wildlife Service. 2019. Fish passage engineering design criteria. Hadley, Massachusetts: USFWS, Northeast Region R5.

Available: https://doi.org/10.3996/JFWM-22-041.S9 (9.644 MB PDF)

We extend thanks to Bill Rice and Jim Magee with the U.S. Fish and Wildlife Service for their help in establishing this project. Jacqueline Knutson (Natural Resources Conservation Service), Mike Roberts (Montana Department of Natural Resources and Conservation), and Austin McCullough (Montana Fish, Wildlife and Parks) were integral in developing the objectives of the project. Our thanks to Erin Ryan, Jason Ilgen, and Matt Toner and all the BFTC staff for the instruction, assistance, and guidance. Undergraduate assistance from Maddie Moxness, Audrey Jones, and Brianna Bowman during trial data collection was essential in the completion of this study. Funding for this project was provided by the USFWS. We would also like to thank the reviewers and the Associate Editor at the Journal of Fish and Wildlife Management for improving this manuscript through their constructive feedback and encouragement.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.