Abstract
Assessments of growth can provide information needed to understand how fish populations respond to changing environmental conditions and management actions, including ecosystem experimentation. We estimated growth rates and parameter uncertainty from otoliths of endangered Humpback Chub Gila cypha from the Colorado River in Grand Canyon, Arizona. We then compared growth of Humpback Chub < age 2 that were 1) occupying the mainstem Colorado River during a period of variable discharge and cooler water temperatures (1980–1998; epoch 1), 2) occupying the Colorado River during a period of moderate discharge variability and warmer water (2001–2011; epoch 2), and 3) occupying the unregulated Little Colorado River. Because growth rates of juvenile Humpback Chub (< age 2) may be more sensitive to changes in environmental conditions than adult fish, we used analysis of covariance and linear models to compare growth of juvenile fish (slopes) between epochs and capture sites (mainstem Colorado River vs. Little Colorado River). Our analysis of covariance results were ambiguous (age × epoch × site interaction; P = 0.06). However, individual linear regressions of size and age by epoch and site suggest biologically important differences in growth, as evidenced by slower growth in the Colorado River in epoch 1 than in epoch 2, and slower growth in the Colorado River compared with the Little Colorado River for all time periods. Overall our results 1) provide information on growth and growth variability useful for parameterizing models to assess population viability and 2) provide empirical information on how growth of juvenile and adult Humpback Chub growth may respond to changing environmental conditions.
Introduction
The Humpback Chub Gila cypha is a large, morphologically distinct minnow endemic to the canyons within the Colorado River basin. The largest population occupies the Colorado River about 100 km below Glen Canyon Dam in Grand Canyon, Arizona, where adults are potamodromous and migrate to the unregulated Little Colorado River to spawn (Kaeding and Zimmerman 1983; Gorman and Stone 1999; Coggins et al. 2006b). As a federally endangered species (United States Fish and Wildlife Service [USFWS] 1967; United States Endangered Species Act [ESA 1973, as amended]), Humpback Chub population status is of significant interest to resource managers, including hydropower operators at Glen Canyon Dam. The Grand Canyon population of Humpback Chub has therefore been studied extensively as part of the Glen Canyon Dam Adaptive Management Program, where research has focused on Humpback Chub population response to planned management experiments and unexpected environmental conditions (Coggins et al. 2011; Finch et al. 2014; Gerig et al. 2014). Determining which environmental factors influence growth of Humpback Chub is essential to informing population assessment models and designing future experimental water release scenarios (Coggins et al. 2006a, 2006b; Finch et al. 2014).
In this paper we provide updated age and growth information derived from otoliths for juvenile and adult Humpback Chub occupying the Colorado and Little Colorado rivers. We use this information to improve model parameter estimates for otolith samples collected from 1989 to 1993 (Hendrickson 1993, 1997) and from 2001 to 2011 to characterize uncertainty of these parameter estimates. Because the oldest fish collected from the second epoch (2001–2011) lived 50% or more of their lives in the first epoch (1980–1993), we combined otolith samples from both time periods to characterize growth. We then compared growth of juvenile Humpback Chub during a period of cooler, more hydrologically variable conditions in the Colorado River (1980–1998, epoch 1; see figure 2 in Pine et al. 2017 [this issue]) to growth during a period of warmer water and less variable Colorado River flows (2001–2011, epoch 2; see figure 2 in Pine et al. 2017), as well as comparing growth from individuals occupying the unregulated Little Colorado River (see figure 3 in Pine et al. 2017 [this issue]). These analyses should provide insights into potential effects of environmental extremes expected with future climate change, including serendipitous temperature regimes that may be beneficial for Humpback Chub growth and survival (Clarkson and Childs 2000; Melis et al. 2016). Our results should also be useful for informing models used to determine status and trends of Humpback Chub, as well as population response to management actions designed to promote recovery.
Methods
We graphically compared mean annual daily discharge and mean annual water temperature for the Colorado River and Little Colorado River between epochs 1 and 2 by calculating summary statistics and 95% confidence intervals (CIs). For humpback collected from 1989 to 1993 (epoch 1), detailed collection and otolith preparation methods and age estimates are available in Hendrickson (1993, 1997). Otolith samples suitable for age analyses from epoch 1 (1980–1993) were collected by Hendrickson in 1989–1993 (n = 173; Hendrickson 1997). Of these, 165 had known sampling locations (n = 69 from the mainstem Colorado River and n = 96 from the Little Colorado River). During this study, we were not able to collect large numbers of Humpback Chub for otolith analyses due to their protected status under the Endangered Species Act. Humpback Chub otolith samples from epoch 2, 2001–2011 (n = 103; n = 54 from the mainstem Colorado River and n = 49 from the Little Colorado River) primarily came from incidental mortalities that occurred as part of standard fish monitoring activities. For detailed information on environmental conditions and fish collection procedures see companion paper by Pine et al. (2017 [this issue]).
We used generally similar analytical methods for estimating age samples for the two epochs. We dissected lapillar otoliths from the fish, and cleaned, dried, and embedded them in epoxy (EpoFix or Epoxicure). Once embedded, we cut frontal plane sections with a Buehler Isomet diamond saw and polished them down to 3 μm with successively finer lapping paper (3M). To enhance growth bands and make measurements, we photographed otoliths with transmitted light microscopy (40–630×) using ImageJ (Ferreira and Rasband 2012). We identified annual growth bands (annuli) by examining opaque zones (the annuli) separated by translucent spaces and counted them on optical images or directly at the microscope. Similarly, we counted daily increments from digital images and often double-checked these at the microscope. We read each otolith at least twice. We determined ages of as many fish as possible. Comparisons of ages were made by the same reader on a subset of fish from both epochs and found these comparisons to be in close agreement. We used daily ages of age-0 fish to estimate hatch-date distributions for Humpback Chub in each epoch. We then converted daily ages converted to fractions of a year for inclusion in growth models with older fish. We accounted for differences in hatch date of older fish (spring or fall) by assigning a birth date to all fish based on hatch-date distributions for each epoch. We then calculated the hatch date as a fraction of a year and added it to the annulus counts on each fish where daily rings were not previously enumerated.
We plotted age at capture (in years or fraction of a year) against total length at capture; we estimated total length from standard length as necessary using equations from Hendrickson (1997) or derived from our own samples. We fit a von Bertalanffy growth model equation,
to these data, where Lt = length at age t, L∞ = asymptotic length, k = metabolic rate, and t0 is a constant of integration (Text S1, Supplemental Material; Data S1, Supplemental Material). Because the oldest (adult) fish collected in epoch 2 had lived more than half of their lives during epoch 1, we tested differences in growth between the two epochs only for juvenile fish < age 2.
For fish < age 2, we fit linear models of fish age and length using analysis of covariance to determine if length (response variable) differs as a function of epoch. We also used linear models to determine if length varied by capture location (Colorado River vs. Little Colorado River) by testing for differences in slopes (e.g., growth rate; Text S2, Supplemental Material; Data S2, Supplemental Material). We then used simple linear regression to examine the slopes and intercepts for each epoch and capture location to see whether any patterns in growth were apparent for corresponding samples of fish. We examined assumptions related to homogeneity in variance for both the analysis of covariance and linear regression models using the Bartlett test, and considered variances significantly different between epochs if P ≤ 0.05.
Results
Discharge and water temperature differed between epochs in the mainstem Colorado River. Mean daily discharge (by convention, cubic feet per second [CFS]) was higher and more variable during epoch 1 than epoch 2 (epoch 1 mean = 16,001 CFS, 95% CI = 15,805–16,199 CFS; epoch 2 mean = 12,347 CFS, 95% CI = 12,240–12,454 CFS; Pine et al. 2017a [this issue]: figures 2 and 3). Mean daily discharge was highest in the mid-1980s and lowest in the early 1990s. Mean daily water temperature was cooler and less variable during epoch 1 than epoch 2 (epoch 1 mean = 9.26°C, 95% CI = 9.23–9.29°C; epoch 2 mean = 10.22°C, 95% CI = 10.16–10.28°C; Pine et al. 2017 [this issue]: figure 2). In the Little Colorado River, mean daily discharge was also higher and more variable during epoch 1 than epoch 2 (epoch 1 mean = 263 CFS, 95% CI = 229–296 CFS; epoch 2 mean = 145 CFS, 95% CI = 120–170 CFS; Pine et al. 2017a [this issue]: figure 3). Little Colorado River water temperature data during Epoch 1 are only available for 1992-1993 and 1995-1998. For these years water temperature was cooler in the Little Colorado River in Epoch 1 than observed in Epoch 2 (Epoch 1 mean = 17.5°C, 95% CI = 17.3-17.8 °C; Epoch 2 mean = 18.1°C, 95% CI = 17.9-18.3°C; Pine et al. 2017a [this issue]: figure 3).
We found that Humpback Chub size and age frequency distributions (Figures 1A and 1B) and hatch-date distributions (Figure 1C) were generally similar between epochs. The majority of samples from both epochs are small fish (< 200 mm TL) and < age 3 (Figure 1A). Otolith samples from larger, older fish were rare (Figure 1A) particularly for fish > age 10 and were especially rare in the oldest age classes (> age 25; epoch 1, n = 1; epoch 2, n = 3). As a result, these sample sizes were small and the oldest fish in epoch 2 lived more than 50% of their lives in epoch 1, so we combined samples from both epochs and fit a single von Bertalanffy growth curve (Table 1; Figure 2) to these data.
For juvenile Humpback Chub, we found that the growth rate was statistically significantly different between epochs. Overall, the fully interactive model (age × epoch × site) better explained the data than the additive model (age + epoch + site) or either the age × site or age × epoch models based upon Akaike information criterion (AIC) comparison (Δ AIC = 17.02). Results from the model with the three-way interaction show that the age and epoch terms were individually significant (P < 0.001) as well as the age × epoch interaction (P = 0.03) and capture location (age × site interaction, P < 0.001; Table 2). To further explore the age × site and age × epoch interactions, we fit simple linear regression models (variances between epochs were homogenous; Bartlett test, P = 0.31) to the age and length data for each epoch (Figure 3), and then individually to each epoch and location (Figure 4). These results suggest that juvenile fish collected from the mainstem Colorado River in epoch 1 had slower growth than did Humpback Chub from epoch 1 collected in the Little Colorado River, and from either location in epoch 2 (Table 3; Figure 4). Overall, this suggests that juvenile Humpback Chub living in the Little Colorado River during Epoch 1 had higher growth rates than fish living in the mainstem Colorado River at the same time, but juvenile growth in the mainstem Colorado River appears to have improved in epoch 2 over epoch 1.
Discussion
Our results suggest that growth patterns of Humpback Chub have changed over the past 30 y, which may be related to changes in flow and temperature regimes. In a companion manuscript on Humpback Chub length–weight relationships, the b parameter (shape) in a standard nonlinear length–weight relationship was also found to be higher in years with warmer water temperatures (Hayes et al. 2017 [this issue]). Despite small sample sizes for adult fish, our results provide one of the only estimates of growth and uncertainty around growth parameters for native cyprinids in the Colorado River across a range of environmental conditions. This information is useful in assessing population status and possible responses to future climate scenarios or management actions.
We found on average a 1.5–2°C increase in temperature for more than 90% of the time in the Colorado River in epoch 2 compared to epoch 1. While these incrementally warmer conditions observed in epoch 2 may have allowed modest increases in growth for juvenile Humpback Chub (leading to the faster growth observed for juvenile humpback collected in the mainstem in epoch 2), the same response is not as clear for adults. Coggins and Pine (2010) used incremental growth information from tag recaptures of adult Humpback Chub to estimate growth and bioenergetics parameters under different water temperature regimes in the Colorado River. Their results suggest that maximum growth improvements for adult Humpback Chub are predicted for changes in temperature of 10°C or more (water temperatures from 10 to 20°C; Coggins and Pine 2010), substantially more than the differences observed between epochs 1 and 2 in this study. Adult Humpback Chub growth would also be expected to slow as adult fish approach asymptotic length (Figure 2).
Until recently, juvenile Humpback Chub that emigrated from the Little Colorado River to the Colorado River were thought to have low survival due to a combination of low water temperatures impeding swimming ability and high predation rates from nonnative trout (Marsh and Douglas 1997; Clarkson and Childs 2000). In this case, lower growth rates for juvenile Humpback Chub collected in the mainstem Colorado River during epoch 1 may be reflecting growth rates of fish that survived the transition from the Little Colorado River to the mainstem Colorado River. Recent results suggest that survival and movements between the Little Colorado and Colorado rivers may be much more complex for juvenile Humpback Chub than previously thought (Hayden et al. 2012; Limburg et al. 2013; Finch et al. 2014), such that fish may be moving between systems to maximize access to prey resources and foraging arenas while minimizing energetic costs and predation risk. These movements would lead to growth patterns that would reflect conditions in both river systems and may help explain why growth patterns in epoch 2 in the mainstem Colorado River were more similar to growth rates observed in the Little Colorado River for both epochs.
Lower, more stable river flows may have altered invertebrate food-base community composition and production in epoch 2, which could affect fish growth (Kennedy et al. 2013). While data on Humpback Chub diet through time is limited, there is some evidence that the occurrence of certain prey resources may have shifted over recent decades in the area around the Little Colorado River (Kaeding and Zimmerman 1983; Kubly 1990; Valdez and Ryel 1995; Valdez and Hoffnagle 1999; Cross et al. 2013) for unknown reasons (Pinney 1991; Blinn et al. 1995), potentially altering Humpback Chub growth rates. Finch et al. (2014) assessed juvenile Humpback Chub growth response to a 60-d steady flow experiment in the Colorado River (in comparison to the extant daily fluctuating flows) and found that growth rates during the steady flow periods were actually lower than growth rates during fluctuating flows. Greater flow variation likely increases short-term availability of invertebrate prey items in drift below dams (Kennedy et al. 2013; Miller and Judson 2014) due to scouring and dislodgement of invertebrates.
Following the 2008 high-flow experiment in Grand Canyon, annual invertebrate biomass and production in the Lee's Ferry reach of the Colorado River below Glen Canyon Dam declined by more than 50% (Cross et al. 2011). These changes were primarily driven by reduced New Zealand mudsnail Potamopyrgus antipodarum and amphipod Gammarus lacustris production and biomass. Coincident with decreased invertebrate production, Cross et al. (2011) documented increases in drifting invertebrate species (Chironomidae and Simuliidae). Along with changes in the invertebrate community composition, juvenile rainbow trout Oncorhynchus mykiss populations increased in this same area and remained high for over a year following the high-flow event (Cross et al. 2011; Korman et al. 2012). While these flow effects on invertebrate drift and responses in age-0 fish have been observed at Lee's Ferry in the Colorado River for nonnative rainbow trout, it is uncertain whether a similar response persists 100 km downstream of the dam. This is a key area of future investigation.
Our analysis for adult fish could be improved if there were additional Humpback Chub age samples available in the ≥ 250-mm size range (likely fish age 10 to age 30). Additional samples of older fish would better characterize the variability in size at age for older fish as well as possibly identify older maximum age. Age samples of fish > age 10 were primarily available from epoch 1 and the oldest aged fishes (> age 25) were from one fish in epoch 1 and three fish from epoch 2. Sparse samples of the oldest age classes of fish are common in growth analyses because older age classes are less common in the population due to natural mortality. Additionally, older fish use different habitats than younger age classes and may be less vulnerable to capture in fisheries monitoring programs. In our study area, large, old Humpback Chub are routinely sampled, and age reconstruction based on tag recaptures suggests these fish have been at liberty for long time periods (Coggins et al. 2006a; Coggins and Pine 2010). We included the otolith samples of old fish from both epochs in our growth estimates because they provide the only information available on size-at-age for large Humpback Chub. Excluding this information would cause a greater bias in growth parameter estimation (and parameter estimates derived from growth information such as mortality) because this would basically be assuming Humpback Chub do not reach large sizes and older ages, when we know that they do. As a federally listed endangered species, the directed take of adult Humpback Chub for age analyses has not been permitted since the Hendrickson (1993, 1997) samples were collected in 1989–1993. We encourage careful cataloging and preserving of fish or otoliths from natural or incidental mortalities to improve estimates of growth in future years.
Given the highly dynamic nature of large desert watersheds such as the Colorado and Little Colorado rivers, it is likely that variations in growth simply reflect selection for phenotypic plasticity that has allowed Humpback Chub to survive. Future work should focus on elucidating the mechanisms responsible for driving growth through integrative studies of both food web and fish population dynamics in relation to physical factors such as flow releases and water temperatures. This type of work, as shown by Cross et al. (2011) and Korman et al. (2012), has benefited our understanding of nonnative fish population responses to flow alterations and is applicable to native fish populations farther downstream. A complementary area of work could be completed at a larger scale by comparing growth patterns of Humpback Chub from Grand Canyon with individuals from less-regulated areas of the upper Colorado River basin (Muth et al. 2000; Stanford and Ward 2001). Ultimately these results may help inform hypotheses related to restoring “natural flows” below Glen Canyon Dam or determining other potential management actions to minimize extinction risk and promote population recovery of Humpback Chub and other native species in regulated rivers similar to the Colorado River.
Supplemental Material
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 authors for the article.
Text S1. The nonlinear growth model fitting for Humpback Chub growth. The code is annotated line by line.
Found at DOI: 10.3996/062014-JFWM-046.s1.
Text S2. The analysis of covariance model-fitting of length–age relationships for Humpback Chub. The code is annotated line by line.
Found at DOI: 10.3996/062014-JFWM-046.s2.
Data S1. CSV file contains the data used for all analyses. This data file contains headers that are described as follows: ID = textual description of each individual fish collected based on information in U.S. Geological Survey Grand Canyon Monitoring and Research Center (USGS-GCMRC) database; site = location fish was obtained either in the Little Colorado River (LCR) or the Colorado River mainstem (CO); year = year of collection; month = month of collection; date = date of collection in a MM/DD/YYYY format; TL= total length of Humpback Chub collected; study = 1 for epoch 1 and 2 for epoch 2; study2 = “Hendrickson” for epoch 1 and “NSE (Nearshore Ecology)” for epoch 2; birthdate = calculated birthdate; age_whole = age in whole numbers; age_frac = age in fraction of years based on collection date; rand_yr_frac_NSE = year fraction assigned to fish from NSE era (epoch 2); rand_yr_frac_Hend = year fraction assigned to fish from Hendrickson era epoch 1).
Found at DOI: 10.3996/062014-JFWM-046.s3.
Data S2. CSV file contains the data used for all analyses. This data file contains the following headers: ID = textual description of each individual fish collected based on information in USGS-GCMRC database; site = location fish was obtained either in the Little Colorado River (LCR) or the Colorado River mainstem (CO); year = year of collection; month = month of collection; date = date of collection in a MM/DD/YYYY format; TL= total length of Humpback Chub collected; study = 1 for epoch 1 and 2 for epoch 2; birthdate = calculated birthdate; age = age of fish in fractions of year.
Found at DOI: 10.3996/062014-JFWM-046.s4.
Reference S1. Ferreira T, Rasband W. 2012. ImageJ User Guide: IJ 1.46r. Revised edition. Bethesda, Maryland: U.S. National Institutes of Health.
Found at DOI: 10.3996/062014-JFWM-046.s5; also available at: http://rsb.info.nih.gov/ij/docs/guide/user-guide.pdf.
Reference S2. Hendrickson DA. 1993. Interim progress report on a study of the utility of data obtainable from otoliths to management of Humpback Chub (Gila cypha) in the Grand Canyon. Final report. Phoenix, Arizona: Arizona Game and Fish Department.
Found at DOI: 10.3996/062014-JFWM-046.s6; also available at: http://www.gcmrc.gov/library/reports/biological/fish_studies/GCES/Hendrickson1993.pdf.
Reference S3. Hendrickson DA. 1997. A preliminary study of utility of data obtainable from otoliths to management of Humpback Chub in the Grand Canyon. Final report. Phoenix, Arizona: Arizona Game and Fish Department.
Found at DOI: 10.3996/062014-JFWM-046.s7; also available at: http://www.gcmrc.gov/library/reports/biological/fish_studies/GCES/Hendrickson1997.pdf.
Reference S4. Kubly, DM. 1990. The endangered Humpback Chub (Gila cypha) in Arizona. A review of past studies and suggestions for future research. Report to Bureau of Reclamation, Glen Canyon Environmental Studies. Phoenix, Arizona: Arizona Game and Fish Department.
Found at DOI: 10.3996/062014-JFWM-046.s8; also available at: http://www.nativefishlab.net/library/textpdf/12469.pdf.
Reference S5. Muth, RT, Crist LW, LaGory KE, Hayse J, Bestgen KR, Ryan TP, Lyons JKA, Valdez RA. 2000. Flow and temperature recommendations for endangered fishes in the Green River downstream of Flaming Gorge Dam. Final Report. Upper Colorado River Endangered Species Recovery Program Project FG-53. Argonne, Illinois: Argonne National Laboratory/U.S. Department of Energy.
Found at DOI: 10.3996/062014-JFWM-046.s9; also available at: http://warnercnr.colostate.edu/docs/fwcb/lfl/PDF/LFL-120-Muth_et_al-2000-Rpt.pdf.
Reference S6. Valdez RA, Ryel RJ. 1995. Life history and ecology of Humpback Chub (Gila cypha) in the Colorado River, Grand Canyon, Arizona. Final report to the Bureau of Reclamation, Salt Lake City, Utah, contract no. 0-CS-40-09110. Logan, Utah: BIO/WEST Report, Inc.
Found at DOI: 10.3996/062014-JFWM-046.s10; also available at: http://www.gcmrc.gov/library/reports/biological/Fish_studies/Biowest/ Valdez1995f.pdf.
Acknowledgments
This paper and two companion papers in this issue were developed as part of the “Nearshore Ecology Project” funded by U.S. Bureau of Reclamation to the GCMRC and the University of Florida. First we would like to thank a key member of the NSE team Mike Dodrill for his work throughout the NSE project. We would also like to thank Jessie Pierson and Jake Hall for their extraordinary assistance in the field during this project. We would like to thank our many cooperators, including Navajo Nation Department of Fish and Wildlife, USFWS, Arizona Game and Fish Department, and U.S. National Park Service for permitting, technical, and field assistance. We are very appreciative of our many boatmen and volunteers who assisted with this project and thank Humphrey Summit Support and GCMRC logistics for their many hours of hard work to make this project possible. We thank the State University of New York, University of Florida, and the Florida Cooperative Wildlife Research Unit for administrative and technical support. R. Van Haverbeke, D. Stone, D. Ward, W. Persons, S. Vanderkooi, and C. Yackulic all provided comments on earlier versions of this work. We acknowledge the helpful efforts of journal reviewers P. Budy, O. Gorman, and the Associate Editor. Collecting permits for this project were obtained from Arizona Game and Fish Department (Scientific Collecting Permit SP790940), USFWS (Federal Fish and Wildlife Permit TE212896-0), Navajo Nation Department of Fish and Wildlife (Scientific Collecting Permit 586), and the U.S. National Park Service (Scientific Research and Collecting Permit GRCA-2011-SC1-0041). Animals were handled in accordance with animal welfare protocols at the University of Florida (IFAS ARC Permit 001-09FAS).
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.
References
Author notes
Citation: Pine WE III, Limburg K, Gerig B, Finch C, Chagaris D, Coggins L, Speas D, Hendrickson DA. 2017. Growth of endangered humpback chub in relation to temperature and discharge in the lower Colorado River. Journal of Fish and Wildlife Management 8(1):322-332; e1944-687X. doi:10.3996/062014-JFWM-046
The findings and conclusions in this article are those of the author(s) and do not necessarily represent the views of the U.S. Fish and Wildlife Service.