Skip to main content
Dryad logo

Seasonal growth potential of Oncorhynchus mykiss in streams with contrasting prey phenology and streamflow

Citation

Rossi, Gabriel (2022), Seasonal growth potential of Oncorhynchus mykiss in streams with contrasting prey phenology and streamflow , Dryad, Dataset, https://doi.org/10.6078/D1TM66

Abstract

The growth of any organism depends on habitat conditions, food availability, and their seasonal interactions. Yet in vast literature on Pacific salmon (Oncorhynchus), the seasonal interaction between habitat conditions and food availability has received relatively little attention. We examined juvenile O. mykiss rearing, physical habitat, and resource phenologies in two Mediterranean coastal streams – one perennial, cool, and shaded; the other intermittent, seasonally warm, and sunny. We used a bioenergetics model to investigate the timing and magnitude of growth potential for drift foraging O. mykiss during the spring and summer in both systems. Growth potential peaked at least two months earlier in the intermittent stream than in the perennial stream. By early summer (June), growth potential had declined in the intermittent stream, whereas growth rates were peaking in the perennial stream. However, the mid-July lipid content of juvenile O. mykiss in the intermittent stream was nearly twice that of fish in the perennial stream. By late summer (August), foraging profitability declined in both streams, as abiotic conditions in the intermittent stream approached lethal. In contrast, the perennial stream maintained suitable abiotic conditions even though growth rate was low. We suggest that the divergent resource phenologies and seasonal mortality risks experienced by anadromous O. mykiss rearing in these streams could drive diversification of traits governing size, age, and timing of outmigration.

Methods

Study sites

Elder and Porter Creeks both flow through the Northern California Coast Range (Fig. 2). Elder Creek is a 16.7 km2 tributary of the upper South Fork Eel River in Mendocino County, California within the University of California Angelo Coast Range Reserve (39.7181° N, 123.6527° W, Fig. 2). Elder Creek’s channel is dominated by cobbles and boulders, with streambed gradients ranging from 2% - 5%  (McBain and Trush 2000). The stream is shaded by a dense riparian canopy, primarily comprised of white alder (Alnus rhombifolia), with some bigleaf maple (Acer macrophyllum) and Oregon ash (Fraxinus latifolia). Elder Creek experiences a characteristic Mediterranean-climate flow recession between April and October (Dralle et al. 2017). However, the basin is underlain by deep water-holding argillite shales of the Northern California Coastal belt, which sustain perennial flow even during prolonged drought (Lovill et al. 2018, Rempe and Dietrich 2018), so dry season low flows rarely drop below 0.015 to 0.03 m3/s. 

Porter Creek is a 19.4 km2 tributary to the Russian River in Sonoma County, California (38.5279° N, 122.8992° W, Fig. 2). Porter Creek’s channel (0.5% to 3%) is lower gradient then Elder Creek with a pebble and gravel dominated bed.  The riparian community is a mix of alders (Alnus spp.), buckeye (Aesculus californica), willow (Salix spp.), Oregon ash, and coast redwood (Sequoia sempervirens) as well as invasive shrubs such as Himalayan blackberry (Rubus armeniacus). Except in the wettest years, Porter Creek becomes intermittent during the dry season for much of its length, although the headwaters are typically perennial.  The Franciscan mélange geology of the Porter Creek basin (Jennings 1977) has much lower hydraulic infiltration capacity and a shallower Critical Zone than Elder Creek, leading to low storage of winter precipitation and greater flow intermittency during the dry summer (Hahm et al. 2019).

Sampling regime

Four riffle-pool habitat units were selected in each stream as study units (Fig. 3). The riffle-pool unit is a dominant geomorphic feature in most alluvial streams. During low-flow, they can be partially or completely isolated from each other, so provide discrete habitats for evaluating juvenile salmonid rearing and foraging (Naman et al. 2018, Rossi et al. 2021a). We selected riffle-pool units that supported multiple age classes of foraging salmonids and were separated from each other by at least two pools.

We measured the following variables each month between late-April and August: (1) streamflow, dissolved oxygen, and water temperature; (2) hydraulics in riffles and pools (depth, velocity, width); (3) epibenthic primary productivity and algal standing crop; and (4) seasonal invertebrate drift, standing crop, and infall (these data were only collected in three of the four pools in each stream due to sampling constraints). We also captured juvenile O. mykiss to estimate growth and lipid content. With these data, we developed a drift-foraging bioenergetics model to predict the seasonal change of O. mykiss growth potential in three pools in each stream.  Data collection and statistical methods are described below and in Appendices S1-S5.

Seasonal gradients in water quality and stream hydraulics

Streamflow, dissolved oxygen, and water temperature — Streamflow (Q, Table 1) for Elder Creek was retrieved from USGS gaging station 11475560 “Elder Creek near Branscomb CA,” which is 0.6 km upstream from the Elder Creek’s confluence with the SF Eel River. All of our Elder Creek study sites were within 0.4 km of this gage, with no intervening tributaries. In Porter Creek, streamflow was retrieved from gaging data collected by Trout Unlimited, following methods of Rantz (1982) and (CDFW 2013).

Dissolved oxygen and water temperature are primary abiotic factors controlling habitat quality for salmonids (Bjornn and Reiser. 1991). We measured dissolved oxygen and stream temperature using continuous HOBO U26 data loggers in one Elder Creek and one Porter Creek pool. The dissolved oxygen loggers were calibrated prior to deployment, and the output data were corrected using HOBOware Pro's Dissolved Oxygen Assistant software. Calibration measurements were taken using a handheld YSI Pro20 in each pool. In addition, we collected monthly, manual water temperature and dissolved oxygen measurements in each pool when we measured primary productivity using the handheld YSI Pro20. Hand measurements were taken between 11:00 am and 2:00 pm.

Riffle-pool hydraulics —Measurements of velocity, riffle depth and width were used to differentiate the seasonal patterns in hydraulics between streams. We installed a cross-stream transect midway through the upstream riffle (XS1), and three cross-stream transects in the downstream pool of each study site. Pool transects were placed where the riffle enters the pool (XS2), a second in the pool head patch were salmonids were observed drift foraging (XS3), and a third over the maximum depth of the pool (XS4), Fig. 3. Depth and velocity (Table 1) were measured at 0.25-m increments along each cross section. Velocity was measured with a Marsh-McBirney Flo-Mate 2000 Electro-magnetic Flow Meter at 0.6 depth from the water surface.

The phenology of O. mykiss prey

Primary production and algal standing crop — In California coastal streams, benthic algae and periphyton are the primary source of carbon for many benthic invertebrates (Finlay et al. 2002). Thus, seasonal changes in the standing crop of benthic algae affect energy flow to salmon and the phenology of salmon prey abundance. To track the seasonality of primary production, we estimated primary productivity in pools, by measuring net primary productivity (NPP) and respiration (R) from attached algae on individual cobbles using the light/dark incubation method of Hall and Moll (1975). We estimated benthic algal standing crop using two methods, ash free dry mass (AFDM) and Chlorophyll-a (Chl-a) analysis. We also quantified the presence of filamentous green algae by presence/absence counts on 0.25-m increments along each cross section. Detailed methods for primary production and algal standing crop are described in Appendix S1 and S2, respectively.

Invertebrate drift, standing crop, and infall — To quantify the phenology of invertebrate salmonid prey, we measured invertebrate drift from riffles entering pools, the standing crop of benthic invertebrates on riffle and pool cobbles, and the infall of terrestrial and adult aquatic invertebrates onto the surface of pools. Invertebrates were sampled at six total sites (3 Elder Creek pools and 3 Porter Creek pools) on five dates over the summer. Detailed methods for invertebrate data collection are included in Appendix S3.

All invertebrate samples (drift, standing crop, and infall) were preserved in the field in labeled 50ml centrifuge tubes filled with 90% ethanol. In the laboratory, drift and benthic samples were sorted and invertebrates identified to family or genus (Merritt et al. 2008) under 10X magnification. Each invertebrate was measured to the nearest 0.5mm under a dissecting scope and biomass (mg dry mass) was estimated from family or order-specific length-weight regressions (Benke et al. 1999, Sabo et al. 2002).  These data were used to define energy content (joules/mg) in the bioenergetic model following Rosenfeld and Taylor (2009).

Bioenergetics — The seasonal change in growth potential for drift foraging O. mykiss was estimated using a drift foraging bioenergetic model (Caldwell et al. 2018, Rosenfeld and Taylor 2009). The model was developed in R version 3.5.1. and is based on equations from Rosenfeld and Taylor (2009) and Hayes et al. (2000) and is the same model used in Rossi et al. (2021b). Growth potential was inferred from the modeled Net Rate of Energetic Intake (NREI, joules time-1), estimating the energy acquired by a juvenile fish for growth (gross energy intake – swimming and other metabolic costs). Gross energetic intake is a function of drift concentration (mg/m3), discharge rate through the foraging volume, fish size, prey size and capture probability (Caldwell et al. 2018, Rosenfeld and Taylor 2009). Swimming costs are a function of fish size and focal point flow velocity and water temperature (Caldwell et al. 2018, Rosenfeld and Taylor 2009). Water temperature was incorporated into the swimming costs (Rosenfeld and Taylor 2009); however, we assumed a constant energy assimilation efficiency of 0.6 (Tucker and Rasmussen 1999). The fish focal point velocity (measured at 6/10ths depth) and drift concentration were measured at the thalweg location on cross section 2 at the head of each pool (Fig. 3). This depth was chosen based on our qualitative observations of O. mykiss foraging focal point depth which was primarily in the lower half of the water column in both streams. We maintained this depth for consistency across the seasons.

We modeled NREI for a 100mm (fork length) drift-foraging O. mykiss at a single foraging location in the channel thalweg. We considered the head of the pool the best indicator of changing drift foraging profitability for seasonal comparisons of drift foraging growth potential across streams (Harvey et al. 2006, Smith and Li 1983 Van Leeuwen et al. 2011, Rossi et al. 2021b) and 100mm was representative of the 1+ fish that most commonly occupied the head of the pool. We only modeled one size class of trout since our objective was to model seasonal changes in NREI rather than size-specific growth potential. Caldwell et al. (2018) also noted that NREI for different size classes of drift foraging O. mykiss and foraging locations downstream from the head of the pool locations followed a similar seasonal pattern over the hydrograph recession, although the magnitudes changed. The model was run for n = 3 pools in each stream, and n = 5 dates for a total of 30 model runs (3 pools × 2 streams × 5 dates).  Other factors affecting growth potential (e.g. non-drifting prey and impacts from dissolved oxygen) were described qualitatively (see Discussion).

Growth, and lipid allocation — Since growth potential (modeled NREI) may differ from realized growth (Hughes et al. 2003, Piccolo et al. 2014), we caught pool-dwelling steelhead to evaluate seasonal changes in the size distributions and fish density, and to collect individual growth and lipid content data. Fish were collected using three-pass backpack electrofishing from 3 study pools in each stream (Smith-Root backpack electrofisher model LR24). All recaptured fish were remeasured for length and mass, allowing us to estimate June to July growth.  We did not recapture enough fish during  the April - May interval or the July - September  interval  to estimate growth. In addition to growth, lipid content was estimated using a non-invasive handheld device, the Distell Model 992 Fish Fatmeter (Distellc Inc., West Lothian, Scotland—https://www.distell.com). Lipid allocation in salmonids is used as indicator of stored energy availability. In O. mykiss, lipid levels have been correlated to survival during periods of resource scarcity (Biro et al. 2004). A detailed description of fish mark and recapture for growth, density, and estimation and lipid estimation is included in Appendix S4.

Data Analysis

We used descriptive statistics to quantify the habitat, food web, and resultant growth potential phenologies for O. mykiss in Elder and Porter Creek and to compare seasonal trends in both streams against our directional predictions. We also compared these seasonal patterns against our graphical model (Fig. 1).  In addition we used a post-hoc bioenergetic analysis to evaluate our predictions about the trends and drivers modeled seasonal of growth potential(1-4). We extracted the modeled swimming costs and energy intake to determine their seasonal variance and relative influence on patterns of growth potential.

Usage Notes

Files are .csv of tabulated data. For details, see Methods.