Effects of Turbidity on Predation Vulnerability of Juvenile Humpback Chub to Rainbow Trout and Brown Trout

Closure of Glen Canyon dam in 1963 resulted in significant changes to the physical environment of the Colorado River in Grand Canyon (Sabo et al. 2012), including decreased water temperature (Vernieu et al. 2005) and reduced turbidity (Voichick and Topping 2014). These changes allowed introduced fish such as rainbow trout Oncorhynchus mykiss and brown trout Salmo trutta to flourish, whereas native species such as humpback chub Gila cypha declined in abundance and distribution (Minckley 1991; Mueller and Marsh 2002). Humpback chub was listed by the U.S. Fish and Wildlife Service (USFWS) in 1967 and was given full protection as an endangered species pursuant to the U.S. Endangered Species Act (ESA 1973, as amended). As Lake Powell filled with water, an estimated 95% of the sediment supply to the Colorado River through Grand Canyon was cut off and deposited into Lake Powell (Topping et al. 2000). Today, most of the sediment that enters into the Colorado River in Grand Canyon comes from two main tributaries, the Paria River (river kilometer [RKM] 0.32) and the Little Colorado River (RKM 99.3), although smaller tributaries also contribute (Griffiths et al. 2012). On average, the turbidity of the Colorado River in Grand Canyon at Phantom Ranch (RKM 140) is now less than 24 formazin nephlometric units (FNUs) for half of the year (Voichick and Topping 2014). In contrast, the predam Colorado River at the same location had an average turbidity greater than 907 FNUs 50% of the time, and almost always exceeded 250 FNUs (Voichick and Topping 2014). These changes in turbidity likely contributed to changes in the fish community (Minckley 1991; Yard et al. 2011).

There are currently six remaining populations of humpback chub, with the largest population inhabiting the Colorado and Little Colorado rivers within Grand Canyon National Park, 122 km downstream from Glen Canyon Dam (USFWS 2002). Adult humpback chub living in the mainstem Colorado River in Grand Canyon enter the Little Colorado River each spring to spawn (Kaeding and Zimmerman 1983; Gorman and Stone 1999). Larval and juvenile humpback chub rear in the Little Colorado River until late July when monsoon flooding causes large numbers of juvenile humpback chub (40–60-mm total length [TL]), to leave the Little Colorado River and enter the mainstem Colorado River (Clarkson and Childs 2000; Yackulic et al. 2014).

Predation on juvenile humpback chub by rainbow trout and brown trout within the mainstem Colorado River has been identified as a potential cause for humpback chub population declines in Grand Canyon (Coggins et al. 2011). Investigations of trout diets within the Colorado River indicate that these species do eat juvenile native fish (Blinn et al. 1993; Marsh and Douglas 1997; Yard et al. 2011), but impacts of trout predation on native fish populations are difficult to predict because predation vulnerability changes with environmental conditions. Turbidity changes the way visual predators such as trout detect prey (Barrett et al. 1992; Utne-Palm 2001; Stuart-Smith et al. 2004). For example, the ability of trout to use contrast to successfully identify and react to prey decreases as water becomes more turbid and light decreases (Davies-Cooley and Smith 2001; Utne-Palm 2001). Even small increases from 4 nephlometric turbidity units (NTUs) to 14 NTUs reduced reactive distance of rainbow trout to mealworms in artificial streams by 20% and turbidities of 30 NTUs reduced reactive distance by 55% compared to clear water (Barrett et al. 1992).

Turbidity varies on both a seasonal and annual basis (Figure 1), leading to highly variable incidence of piscivory for trout inhabiting this area of the Colorado River (Yard et al. 2011). It has been hypothesized that management actions that increase turbidity within the Colorado River could lead to increases in recruitment of native Colorado River fishes by reducing predation mortality (Albrecht et al. 2010). Understanding the relationship between water clarity and predation vulnerability is therefore critical when evaluating management actions designed to benefit native fish species such as humpback chub within the Colorado River. We conducted laboratory experiments to evaluate how short-term predation vulnerability of humpback chub changes in response to turbidity by using juvenile captive-reared humpback chub and bonytail Gila elegans as prey and wild-caught adult rainbow trout and brown trout as predators. Evaluating predation interactions in a laboratory has advantages in that individual factors can be isolated and controlled without the high logistical costs and difficulties of conducting studies in remote field locations.

Captive-reared juvenile humpback chub were obtained from the U.S. Fish and Wildlife Service (USFWS) Southwestern Native Aquatic Research and Recovery Center in Dexter, New Mexico, and bonytail were obtained from the Wahweap Utah State Fish Hatchery in Bigwater, Utah. Bonytail were included in the experiment because limited numbers of humpback chub are available for research purposes. Bonytail are closely related to humpback chub, native to the Colorado River and its tributaries, and make good surrogates for humpback chub (Kappenman et al. 2012) because of similarities in morphology and life history, especially as juveniles (Douglas et al. 1989; Karp and Tyus 1990). All humpback chub and bonytail were held in indoor holding tanks at 20°C until predation trials began. Wild adult rainbow trout and brown trout were collected from Lees Ferry and Canyon Creek, Arizona, respectively, by using electrofishing techniques. Trout were maintained in captivity for a minimum of 2 wk at (18 ± 2°C) and 3.5 ppt salinity before predation trials to acclimate them to being in captivity and remove any external protozoan parasites. Trout were fed live fathead minnows Pimephales promelas ad libitum during this 2-wk period, but they were not fed for the 48-h period before predation tests to motivate the trout to behave as predators. All fish were maintained in separate holding tanks in a temperature-controlled greenhouse at the U.S. Forest Service Rocky Mountain Research Station in Flagstaff, Arizona.

Predation trials were conducted in 12 replicate 568-L fiberglass tanks that measured 1.2 × 0.9 m (length × width), with a water depth of 30 cm. Each tank was outfitted with an Aqua Logic® 1/3-hp drop-in chiller and a 44-L/min mag-drive pump and air stone to provide water circulation and aeration. Each tank was covered with 6.4-mm mesh netting to prevent fish from escaping. An initial set of trials was executed over a broad range of turbidities (0, 50, 200, 500, and 1,000 FNUs; Figure 4), with subsequent trials focused on a narrower range of lower turbidities (0, 25, 50, 75, 100, and 150 FNUs) where the greatest change in prey survival seemed to occur. Four replicates of each turbidity level were conducted simultaneously per trial, with temperatures held at ±1°C from the target temperature of 15°C. An Onset® HOBO temperature logger monitored water temperature during each predation trial in each tank and was downloaded after each trial. In the event that water temperature fluctuated more than ±1°C from 15°C during the course of the 24-h trial, the data from that trial were excluded from the analysis because water temperature is also known to impact predation vulnerability for these species (Ward and Morton-Starner 2015). Ammonia and nitrite levels in each tank were tested with an API® aquarium test kit (Nessler's reagent) during the initial trial and were found to be undetectable. Amquel Plus® ammonia remover was added to each tank at the beginning of each subsequent trial to keep ammonia and nitrite levels low during the experiments and prevent adverse effects on fish behavior that could impact predation results. Four trout were placed into each tank, and 12 humpback chub or bonytail were placed into a cylinder-shaped mesh basket within each tank at the beginning of the trial. All fish were allowed to recover from capture and handling and to acclimate to the predation tanks for a period of 48 h.

Fine sediments were gathered from the lower Paria River near the confluence with the Colorado and the Little Colorado rivers directly above Grand Falls. This mud was sifted through a 63-μm sieve and blended using a standard kitchen blender to obtain a suspension of fine silt and clay particles. This suspension was added gradually at the beginning of the acclimation period and tested using a Hach® 201 turbidimeter until the turbidity stabilized at each target level. A 44 L/min mag-drive pump located in each tank provided water circulation to keep sediment suspended. Once desired turbidity levels were reached, further turbidity measurements were made using a variety of water clarity– and light-measuring instruments so that turbidity could be compared more easily between this study and other published research (Figure 2; Data S1). These instruments included an YSI® 6136 turbidity probe measuring in FNUs, a Hach 201 turbidimeter measuring in NTUs, and a Secchi disk measuring depth in millimeters.

After the 48-h acclimation period, turbidity was measured again and additional sediment was added if needed to reach target turbidity levels. To initiate the predation experiment, we placed baskets containing prey fish into each tank and then carefully tipped the baskets over to release chub into each tank. The baskets were then removed. Tanks were left undisturbed for 24 h under natural light. After 24 h, water in each tank was drained and all surviving fish were captured, counted, and measured for TL. Experiments took place from October 2013 to May 2014 and typically began between 0800 and 1000 hours. In total, 124 individual overnight trials were conducted using 256 rainbow trout (223–330-mm TL) and 208 brown trout (193–399-mm TL). Prey fish consisted of two sizes of juvenile chub: 120 individual humpback chub (41–64-mm TL) and 1,008 bonytail (60–73-mm TL, Table 1; Data S2 and S3). Logistic regression was used to calculate parameter estimates and to evaluate the effects of prey size and turbidity on predation vulnerability. The Prediction Profiler tool in JMP® 10 statistics was used to calculate the probability of chub survival and 95% confidence intervals (CIs) under a range of turbidities commonly found within the Colorado River in Grand Canyon (Figure 1). The Prediction Profiler in JMP statistics displays a predicted profile trace of a single variable, whereas others are held constant, and allows individual variables within the logistic regression model to be evaluated separately.

Initial trials with a wide range of turbidities (0–1,000 FNUs; Figure 4) and humpback chub prey revealed that the greatest changes in predation vulnerability occurred at turbidities less than 200 NTUs (Figure 3). Subsequent trials using bonytail, conducted at 0–150 FNUs (Figure 6), indicated that increased turbidity significantly reduced predation vulnerability of bonytail to both rainbow trout and brown trout (P < 0.001, likelihood ratio tests; Table 2), although increases in survival with increasing turbidity were much more pronounced for rainbow trout than for brown trout (Figure 5). Prey and predator sizes were tightly controlled in our studies, but prey size still had a significant effect on predation vulnerability for both rainbow trout and brown trout ((P = 0.0069 and P = 0.0193, respectively; likelihood ratio tests; Table 2; Figure 5). The effects of fish size and water temperature on predation relationships are well established and have been described in detail for these species previously (Ward and Morton-Starner 2015). The type of sediment used in our trials (Paria River or Grand Falls origin) did not have a significant effect on predation vulnerability for rainbow trout or brown trout (P > 0.05; Table 2), despite visually apparent differences in water color with different sediment origins (Figures 4 and 6). An increase in turbidity from 0 FNU to 25 FNUs (with all other factors held constant) resulted in a 36% increase in survival of 65-mm TL bonytail to a 285-mm TL rainbow trout (24–60%; 95% CI; Figure 5) and a 25% increase in survival of a 65-mm bonytail to a 285-mm TL brown trout (17–33%; 95% CI). An increase in turbidity from clear water to 150 FNUs (with all other factors held constant) resulted in a more than fourfold increase in juvenile chub survival to predation by both rainbow trout (21–91%) and brown trout (7–35%; Figure 5).

Abiotic factors such as turbidity can have large impacts on biotic interactions and play an important role in structuring animal communities (Dunison and Travis 1991). Our study suggests that rainbow trout and brown trout are less effective predators on native chub species as turbidity increases (Figure 2), with turbidity as low as 25 FNUs reducing predation vulnerability of bonytail to rainbow trout and brown trout by 36 and 25%, respectively. Our observed decrease in predation vulnerability associated with increases in turbidity is consistent with other published studies conducted in both laboratory and natural environments.

Relatively low levels of turbidity (≤25 NTUs) have been shown to negatively impact the ability of trout to feed effectively. Shaw and Richardson (2001) studied rainbow trout exposed to fine sediment in constructed channels along a natural stream system and found that sediment pulses of only 23 NTUs led to food deprivation. Gregory and Levings (1998) observed Pacific salmon Oncorhynchus spp. in clear water to be more than twice as likely to have recently consumed prey than salmon migrating in a turbid river. In addition, diet studies on brown trout in a Tasmanian lake showed that as turbidity increased from 26 to 141 NTUs, the proportion of fish with food in their stomachs decreased from 98 to 20% (Stuart-Smith et al. 2004).

At moderate levels of turbidity, cutthroat trout Oncorhynchus clarkii feeding on live oligochaetes in a laboratory stream consumed virtually all available prey in clear water, but as turbidity increased consumption decreased: minimal feeding was observed at 200 NTUs and no feeding was observed at 400 NTUs (Harvey and White 2008). We observed a similar pattern in our trials: in clear water rainbow trout consumed all available prey, but at 150 FNUs only 9% of bonytail were consumed.

Elevated turbidity levels that impair the foraging ability of salmonids can affect fish health, growth, distribution, and population size. For example, high concentrations of suspended sediment can result in stress, physiological damage, and reduced growth in salmonids (Newcombe and Jensen 1996; Shaw and Richardson 2001). The effect that elevated turbidity has varies depending on the life stage of the fish (Shaw and Richardson 2001; Utne-Palm 2001) and on the type and particle size of the sediment (Servizi and Martens 1987; Newcombe and Jensen 1996). Larger particles, high concentrations, and long exposures are the most damaging (Servizi and Martens 1987; Newcombe and Jensen 1996).

Behavioral changes that occur in response to increased turbidity for both predator and prey can affect fish distributions, prey availabilities, and encounter rates (Yard et al. 2011), making interpretation of the effects of altered turbidity on fish populations challenging. High turbidity can affect the behavior of juvenile and small-bodied fishes that are preyed upon by larger fish. Increases in turbidity can provide cover and reduce predation risk for juvenile fish, often leading to an increase in overall activity at higher turbidities (Gregory 1993; Stone 2010). Many species prefer turbid conditions over clear water and will use more turbid environments when there is a perceived predation risk (Chiu and Abrahams 2010).

Turbidity also changes activity patterns in predators. In clear water, trout are most active during twilight, but with increases in turbidity these patterns are disrupted and trout become more active in general (Gregory and Levings 1998; Harvey and White 2008). At high turbidities, trout may need to spend more energy seeking out and finding prey that are visually difficult to detect (Sweka and Hartman 2001), but this increased activity does not necessarily lead to increased food consumption (Shaw and Richardson 2001; Harvey and White 2008).

Our laboratory study used naïve hatchery fish as prey which may have minimized predator avoidance behavior and increased predation vulnerability compared to that of wild fish. Although the results of our laboratory studies do not quantify the survival of wild juvenile humpback chub within the Colorado River, they do give information on how survival of chub may change in the Colorado River as a result of turbidity alteration. This information may be useful in evaluating the ecological implications of turbidity changes caused by Glen Canyon Dam and in evaluating potential management actions aimed at benefitting humpback chub within Grand Canyon. Relatively small changes in turbidity may be sufficient to alter predation dynamics of rainbow trout on humpback chub in the Grand Canyon. Management actions aimed at augmenting sediment below Glen Canyon Dam have been considered (Randle et al. 2007), but not implemented because of high cost. Evaluations of low level silt and clay augmentation only, for the purpose of producing turbidity in the 25–50-FNU range, to reduce predation vulnerability of native fishes may warrant further evaluation and field testing.

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. Raw data used to compare turbidity measures and to evaluate effects of turbidity on predation vulnerability in laboratory trials. Turbidity measurement tools include a YSI 6136 turbidity probe measuring in formazin nephlometric units (FNUs), a Hach 201 turbidimeter measuring in nephlometric turbidity units (NTUs), and a Secchi disk measuring depth in millimeters.

Found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S1 (23 KB XLXS).

Data S2. Archived data used to calculate the probability that a juvenile bonytail Gila elegans will survive predation by rainbow trout Oncorhynchus mykiss as turbidity increases from 0 to 150 formazin nephlometric units (FNUs) at 15°C in laboratory trials.

Found at DOI: Found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S2 (26 KB XLXS).

Data S3. Archived data used to calculate the probability that a juvenile bonytail Gila elegans will survive predation by brown trout Salmo trutta as turbidity increases from 0 to 150 formazin nephlometric units (FNUs) at 15°C in laboratory trials.

Found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S3 (22 KB XLXS).

Reference S1. Griffiths RE, Topping DJ, Andrews T, Bennett GE, Sabol TA, Melis TS. 2012. Design and maintenance of a network for collecting high-resolution suspended-sediment data at remote locations on rivers, with examples from the Colorado River. U.S. Geological Survey Techniques and Methods, Book 8, chapter C2.

Found at DOI: http://pubs.usgs.gov/tm/tm8c2/tm8c2_text.pdf (13.96 MB pdf). Also found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S4 (14305 KB pdf).

Reference S2. Mueller GA, Marsh PC. 2002. Lost, A desert river and its native fishes: a historical perspective of the lower Colorado River. Information and Technology Report USGS/DRD/ITR-2002-0010:U.S. Government Printing Office, Denver, Colorado.

Found at DOI: https://www.fort.usgs.gov/sites/default/files/products/publications/10026/10026.pdf (3.26 MB pdf). Also found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S5 (3339 KB PDF).

Reference S3. Randle TJ, Lyons JK, Christensen RJ, Stephen RD. 2007. Colorado River ecosystem sediment augmentation appraisal engineering report. U.S. Bureau of Reclamation, Technical Service Center, Sedimentation and River Hydraulics Group, Denver, Colorado.

Found at DOI: http://www.gcmrc.gov/library/reports/physical/Fine_Sed/Randle2007.pdf (4.73 MB pdf). Also found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S6 (4847 KB PDF).

Reference S4. U.S. Fish and Wildlife Service. 2002. Humpback chub (Gila cypha) recovery goals: amendment and supplement to the humpback chub recovery plan. U.S. Fish and Wildlife Service, Denver, Colorado.

Found at DOI: http://www.fwspubs.org/doi/suppl/10.3996/082012-JFWM-071/suppl_file/10.3996_082012-jfwm-071.s12.pdf (1.3 MB pdf). Also found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S7 (1336 KB PDF).

Reference S5. Vernieu WS, Hueftle SJ, Gloss SP. 2005. Water quality in Lake Powell and the Colorado River. Pages 69–685 in Gloss SP, Lovich JE, Melis TS, editors. The state of the Colorado River Ecosystem in Grand Canyon. Volume 1282. U.S. Geological Survey Circular; Reston, Virginia.

Found at DOI: http://pubs.usgs.gov/circ/1282/c1282.pdf (6.73 MB pdf). Also found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S8 (6898 KB PDF).

Reference S6. Voichick N, Topping DJ. 2014. Extending the turbidity record—making additional use of continuous data from turbidity, acoustic-Doppler, and Laser diffraction instruments and suspended-sediment samples in the Colorado River in Grand Canyon. U.S. Department of the Interior, U.S. Geological Survey, Scientific Investigations Report 2014-5097.

Found at DOI: http://pubs.usgs.gov/sir/2014/5097/pdf/sir2014-5097.pdf (1.88 MB pdf). Also found at DOI: http://dx.doi.org/10.3996/102015-JFWM-101.S9 (1932 KB PDF).

We thank Zane Olsen at the Wahweap Fish Hatchery in Bigwater, Utah, and Manuel Ulibarri at the Southwestern Native Aquatic Resources and Recovery Center in Dexter, New Mexico, for providing research specimens. The U.S. Forest Service, Rocky Mountain Research Station in Flagstaff, Arizona, provided greenhouses to conduct this research. Scott Vanderkooi, Kimberly L. Dibble, two anonymous reviewers, and the Associate Editor provided valuable reviews of this manuscript.

This work was conducted under Federal Endangered Species permit TE821356-2. The use of trade, firm, product or corporation names in this publication is for informational use only and does not constitute an official endorsement or approval by the U.S. Government.