Excessive loading of fine sediments into western rivers has degraded spawning and rearing habitat for salmonids, and contributed substantially to their declines. Impacts on salmon redds have been studied extensively, but effects on juvenile rearing are less well documented. In a field experiment in the South Fork Eel River, we investigated the impacts of deposited fine sediment on juvenile steelhead. Our experimental design allowed us to isolate the effects of fine bed sediments from other covarying factors and to reveal the mechanisms of their effects. Increasing levels of embeddedness with deposited fine sediment (from zero to 100%) decreased growth and survival of juvenile steelhead trout. The nearly linear decreases in growth resulted from decreased food availability and metabolic costs of increased activity and intraspecific aggression. The invertebrate community changed from one of more available prey to one of unavailable burrowing taxa with higher levels of deposited fine sediment. Steelhead in more heavily embedded channels showed more continuous movement and aggression and higher incidence of injury. This study shows a direct impact of riverbed composition on salmonid rearing success, which has been identified as a life history bottleneckby models informing efforts to recover these populations.

As organisms undergo life history transitions, track resources, avoid stress, and evade death, they distribute themselves across landscapes. Organismal co-occurrence sets the stage for biotic interactions, which can feed back to control the distribution and abundance of interacting species in ecological communities. Classically, competition and bottom-up forces have been thought to be the most important drivers of community structure, however, examples of predation, parasitism, mutualism, and facilitation highlight the ubiquity and importance of these other interactions. In freshwaters, anthropogenic impacts, especially species introductions and climate warming, have resulted in novel species assemblages, with altered webs of interactions compared to historic conditions. Human management often seeks to provide conditions that favor native species and inhibit non-natives. Success requires an understanding species interactions and their roles in community dynamics. In my first chapter, I describe the distributional dynamics of the assemblage of aquatic fishes, reptiles and amphibians in the South Fork Eel River. In Chapter 2, I describe the seasonal migration of an introduced predatory fish, Sacramento Pikeminnow (Ptychocheilus grandis), and how climate warming and water withdrawals could increase their negative impact on rearing native salmonids and other fauna. In Chapter 3, my colleagues and I explore the positive interactions between Pacific Lamprey (Entosphenus tridentatus), juvenile Steelhead trout (Oncorhynchus mykiss), and Foothill Yellow-legged Frogs (Rana boylii). These case studies emphasize the need to consider ecological interactions, and in general, community ecology thinking, as we try to restore and manage ecosystems.

Abstract

From Cobble to Canyon: inferring the effects of discharge, climate and trophic interactions

on primary productivity in a Northern California river, over reach to watershed to

coastal ocean and annual to multi-decadal scales.

by

John Blackburn Sculley

Doctor of Philosophy in Integrative Biology

University of California, Berkeley

Professor Mary E. Power, Chair

Climate change along the California North Coast is expected to have significant impacts on primary production by riverine algae in coastal watersheds through increased precipitation intensity and duration (Snyder and Sloan 2005), and regional changes in fog frequency (Johnstone and Dawson 2010, Lebassi et al. 2009), potentially altering solar irradiance and extending higher flows and spates later into the summer low flow period when algae accrue. Changes in flow velocity and irradiance may alter relationships between primary producers and affect biomass accrual (Dodds 1991a, Stevenson 1983). Primary production in the Eel River has been established by isotopic studies to feed local aquatic and terrestrial food webs (Finlay et al. 2002) before being exported by winter high flows to the continental shelf and submarine canyons by the following summer (Drexler et al. 2006).

I examined the combined and separate effects of two different radiation levels and flow velocity ranges on the growth and interaction of algae that dominate mainstem primary producer guilds in the food web of a northwestern California river during the biologically active summer baseflow season: Cladophora glomerata (L.) Kutz, (the dominant benthic macroalga) and Cocconeis placentula (Her), C. pediculus (Ehrenb.) Kütz., and Epithemia turgida (Ehrenb.) Kütz. and E. sorex Kütz. (the dominant genera of epiphytic diatoms on Cladophora during the middle and later summer months, respectively). Cladophora-epiphyte assemblages attached to cobbles were collected from the South Fork Eel River (39°43'47''N, 123°38'34''W), and placed in in-stream channels that were gated and shaded to provide two velocity regimes (slow (0-9 cm/s) and faster (10-20 cm/s)) and two light treatments (shaded (peak irradiance 150 µE/m2/s) and full sunlight (peak irradiance 1500 µE/m2/s)). Cladophora filaments grew 68% longer under the higher light level (26.5 vs 15.8 cm in high vs low light), but did not differ under different water velocities. Epiphyte densities on Cladophora were 65% higher (6141 vs 3726 cells/mm2 of host area) in slow velocity treatments. However, epiphyte biovolume was 67% higher (4.26 vs. 2.55 x 105 µm3/mm2) in faster velocity treatments. Total epiphyte biovolume per host filament increased 88% in higher light.

I identified and counted freshwater diatom frustules in datable annual layers (varves) from a sedimentary core recovered from the submarine Eel canyon, which receives up to 50% of discharged sediment from the nearby Eel River. These freshwater diatoms, mainly Epithemia, Cocconeis, and Rhoicosphenia spp., appear to provide an accurate paleo-productivity (bloom size) proxy of algae and algae-based food web dynamics over an 84 year period. With more than 47% of the variation in algal bloom height being captured by Epithemia spp. in offshore canyon cores, and more than 33% of the variation captured by total freshwater diatoms, I have discovered a useful proxy for algal biomass throughout the Eel River basin. I can thus extend temporal inferences about hydrologically mediated food web responses, e.g. the two-state algae bloom response to bankfull discharge regimes, back to 1917, and upscale these inferences to the entire 9540 km2 watershed.

Using this proxy record of freshwater diatom frustules for basin-wide algal productivity, I discovered a strong and significant correlation between Rhopalodiaceae and total freshwater frustules and two satellite records of marine chlorophyll in the coastal ocean adjacent to the Eel mouth. Records of mean annual chlorophyll from the SeaWiFS sensor over the period 1997-2001 (the period where the core and satellite records overlap) and from the CZCS sensor over the entire mission period of 1978-1986 are positively correlated with Rhopalodiaceae and total freshwater frustules, accounting for between 62 and 72% of the variation in mean annual chlorophyll a in a ~2400 km2 region of the continental shelf directly offshore of the Eel River mouth. I also found a significant and strongly positive correlation of mean annual marine chlorophyll with peak averaged modal Cladophora bloom height from upstream algal surveys during the period of overlap with the satellite record (1988-2011). River plumes have been posited as a possible source of nutrients for marine phytoplankton blooms off the California and Oregon coasts (Bruland et al. 2001, Wetz et al. 2006). These results imply a connection between riverine algae and the offshore marine ecosystem. While upwelling is clearly the major driver of marine phytoplankton blooms, my data support a role for freshwater-derived nutrients in winter/spring marine productivity.

Taken together, these results provide a more detailed picture of how environmental conditions can affect the biomass of hosts and epiphytes in riverine primary producer assemblages, and they enable the tracking of this biomass from production in river channels to burial in submarine canyons. Better knowledge of how environmental controls exert different effects on the relative growth and colonization rates of hosts and epiphytes will improve my understanding of how the energy flow and biogeochemical fluxes that they mediate may change with climate change. The ability to track primary production using the direct correlation between freshwater diatoms and bloom height allows us insight into algal food web dynamics far earlier than contemporary observations permit. These records have the potential to greatly increase the spatial and temporal scales of inferences linking food webs to environments, enhancing our understanding of how river webs may respond to future changes in climate, land cover, and other factors affecting riverine runoff and conditions, as well as their linkages to marine systems worldwide.

Scouring winter floods strongly influence the structure and dynamics of food webs in rivers with winter flood, summer drought hydrographs. Reduction or elimination of scouring floods, in addtition to altering physical conditions within rivers, may negatively affect salmonid populations by reducing energy flow to them from lower trophic levels. We compared food webs of two northern California rivers with drastically different flow regimes to assess the effects of differences in food web structure on the distribution and growth of juvenile steelhead trout. The upper Mad River has a highly regulated flow regime and rarely experiences scouring winter floods, while the upper Van Duzen River is free-flowing and experiences frequent scouring floods. Thrroughout spring and summer 1994 densities of the large, grazing caddisfly Dicosmoecus gilvipes exceeded 801m2 in the Mad River, but were < 21m2 in the Van Duzen. Consequently, filamentous green algae was nearly absent in the Mad from June through September, but was relatively abundant in the Van Duzen. Densities of other stream-dwelling insects (primarily Chironomids and mayflies, which are the preferred prey of juvenile salmonids) and juvenile steelhead were consistently mlich lower in the Mad than the Van Duzen. At the end of the summer, Dicosmoecus pupated, and thus became inactive, which resulted in a large bloom of filamentous green algae (primarily Oedogonium and Cladophora) and a several-fold increase in the densities of Chironomids and mayflies in the Mad but not the Van Duzen. River flows in 1995 were much higher than in 1994 and both rivers experienced a number of scouring floods. As a consequence, Dicosmoecus densities were reduced to < 21m2 in both rivers throughout the spring and summer. The Mad experienced a large Cladophora bloom, and densities of Chironomids and mayflies were several times higher in 1995 than 1994. These observations support our hypothesis that eliminating scouring floods favors large, slow-growing benthic insect taxa over smaller, faster-growing taxa whose populations build up rapidly after floods. Since the former are invulnerable to predation by juvenile salmonids, energy flow is reduced and juvenile salmonid populations decline.

Experimental manipulation of juvenile steelhead in artificial channels with and without Dicosmoecus allowed us to test this hypothesis more directly and without potential influences from factors other than food web structure that may have varied between regulated and unregulated rivers. Experimental results revealed that Dicosmoecus significantly reduced the availability of small prey, which resulted in negative juvenile steelhead growth. Together, our surveys and experimental results show that elimination of scouring floods alters energy pathways in river food webs resulting in reduced biomass available to fish populations. Modifying flow regimes regulated by dams, so they more closely resemble natural hydrographs, may be an important step in restoring salmonid populations in some rivers.

Pacific salmon (Oncorhynchus) are an iconic genus of anadromous fish whose decline over the last century has linked indigenous peoples, government agencies, fishers, advocates, and scientists in the common, albeit messy, cause of salmon recovery. While the history of ecology, and the earliest research on Pacific salmon, is grounded in the understanding that food web and prey dynamics control the abundance and distribution of predators, the last half-century of salmon management has focused primarily on hatchery production and physical habitat suitability. The continued decline of Pacific salmon, after decades of efforts to restore them, has prompted salmon ecologists to remember what Charles Elton (1927) understood nearly a century ago: “Food is the burning question in animal society, and the whole structure and activities of the community are dependent upon questions of food-supply.” My dissertation proceeds from the observation that both salmon management and salmon ecology have placed too little emphasis on aquatic food webs, animal behavior, and biotic interactions as key drivers of salmon production and resilience—and thus as chief mechanisms of salmon recovery.

This dissertation is comprised of a series of studies and experiments from 2016 to 2019, which explore how the seasonal interaction among physical habitat, phenology of salmon prey, and foraging behavior affects juvenile salmon growth and performance. The experiments described here take place near the southern end of the range of Pacific salmon—in Mediterranean streams along the Pacific Coast of California. These streams experience a characteristic recession of streamflow between spring and fall, which is an ideal natural gradient to investigate the dynamic interaction between physical habitat and prey phenology. The studies described here also focus on one species of Pacific salmon—Oncorhynchus mykiss (the “steelhead” or “rainbow trout”).

My first chapter starts from the observation that O. mykiss are typically considered drift foragers that occupy higher velocity habitat than other salmonids. Yet their persistence in Mediterranean coastal streams with seasonally low velocity and drift suggests an underappreciated flexibility in foraging behavior. Using novel 3-D videogrammetry along with measurements of invertebrate drift and stream hydraulics, I quantified the seasonality of foraging behaviors that O. mykiss employ during the spring-summer flow recession. I also explored the factors that trigger of foraging mode shifts, which are important for both ecological interactions and salmon management. I found that 80% of O. mykiss were drift foraging in May, 2018; while over 70% used alternative foraging modes by July. Foraging mode was significantly correlated to maximum water velocity and to riffle crest depth. While drift concentration was a poor univariate predictor, the top-ranked statistical models of foraging mode included both hydraulic variables and drift concentration. This study revealed that O. mykiss in Mediterranean streams express more diverse foraging behavior than previously suggested in the literature, and that changes in foraging mode could be predicted by simple hydraulic indicators. This study also lead us to speculate that non-drifting prey fluxes, as well as risk aversion, may influence late-summer foraging decisions in Mediterranean coastal streams.

In my second chapter I took advantage of an existing streamflow augmentation project to conduct a novel manipulative experiment to elevate dry-season flow in an intermittent, salmon-bearing stream. Using a Before-After-Control-Impact (BACI) study design, I directly measured the responses of O. mykiss behavior, their invertebrate prey, physical habitat, and water quality to flow enhancement. I found that elevating mid-summer baseflow for one week caused significant increases in dissolved oxygen, riffle width, depth, velocity, invertebrate drift, and benthic invertebrate standing crop – and these changes were evident at sites more than 1 km downstream from the point of flow augmentation. I also documented consistent behavioral changes indicating that pool-dwelling salmonids were able to exploit the increased prey abundance and improved metabolic environment in the treatment reach after flow augmentation. Counts of foraging salmonids near riffle-pool transition zones doubled, foraging movement remained steady (relative to decreasing movement in the control reach), and the proportion of drift foraging salmonids increased from an average of 26% to an average of 77% after augmentation. Collectively, these results suggest that dry season flow enhancement can significantly improve habitat profitability for salmonids, but that the timing and location of flow augmentation, relative to downstream points of interest, are critical to meeting ecological objectives.

In my third chapter I propose a conceptual model for how stream hydraulics and secondary production interact to affect foraging profitability for juvenile salmon during the streamflow recession in Mediterranean coastal streams. I evaluated the model by comparing two distinct salmonid rearing streams—one perennial, cool, and shaded; the other intermittent, seasonally warm, and sunny. In both streams, I conducted a photon-to-salmon food web study to document the seasonality of hydraulic habitat relative to the phenology of prey biomass for O. mykiss. I also estimated the seasonal change in O. mykiss growth potential using a drift foraging bioenergetics model. Both streams offered profitable foraging opportunities for O. mykiss, but modeled growth potential peaked at least two months earlier in the intermittent stream, where lipid storage measured in mid-July was nearly twice as high. In contrast, modeled and measured early-summer growth was higher in the perennial stream. By late summer, foraging profitability declined in both streams. However, abiotic conditions approached lethal tolerance levels in the intermittent stream, whereas the perennial stream maintained suitable abiotic conditions all summer. Seasonal changes in growth potential were explained by the timing of streamflow recession relative to the phenology of prey fluxes. But ecological interactions (especially intraspecific competition) and flexible foraging behavior both appear to have mediated realized foraging opportunity. These observations lead me to speculate that juvenile O. mykiss life histories should differ among streams, due to the timing of their spring-summer growth potential and to differing late summer survival rates. Collectively, these stream types could contribute to a stabilizing portfolio of juvenile salmonid life histories.

Disturbance by flooding can dramatically disrupt population and community structure in stream ecosystems. My dissertation research focused on two species that are vulnerable to high flow events: the caddisfly Dicosmoecus gilvipes and the western pearlshell mussel Margaritifera falcata. I investigated their influence on energy and nutrient dynamics in a Northern California coastal river.

In Chapter 1, I examined the impact of the limnephilid caddisfly Dicosmoecus gilvipes on periphyton structure and ecosystem processes. Dicosmoecus larvae reduced periphyton accrual, chlorophyll a, gross primary productivity, and ammonium uptake in experimental channels, and their impact persisted 46 days after the larvae were removed. Given Dicosmoecus vulnerability to high flow events, any change in flood timing, frequency, and/or magnitude due to river regulation or climate conditions may significantly alter ecosystem processes in Northern California streams.

In Chapter 2, I investigated whether flood timing would have a differential impact on Dicosmoecus gilvipes populations. Specifically, I measured how critical flow thresholds and habitat use varies with larval size. Critical flow velocity and dimensionless flow threshold indices increased with larval size, as did their flow velocity preference. The results suggest early flood events will have a greater impact on Dicosmoecus populations than later flood events of a similar magnitude, and during low-flow periods the interaction between Dicosmoecus distribution, periphyton composition and productivity, and flow velocity may significantly impact ecosystem processes on smaller scales.

In Chapter 3, I investigated the functional role of cases built by Dicosmoecus gilvipes. The larvae collect thin plant material and Douglas-Fir needles and build arrow-shaped lateral extensions on their case. Larvae with lateral extensions experienced fewer revolutions and regained their footing faster in experimental trials than those without. The results suggest lateral extensions provide stability against overturning in fast flow and may improve their ability to forage efficiently in turbulent flow conditions.

In Chapter 4, I manipulated the presence and absence of the mussel Margaritifera falcata in stream mesocosms. I measured their impact on organic matter accrual, microbial activity in the sediment, and the growth of larval Pacific lamprey, Lampetra tridentata. Margaritifera presence increased microbial activity in the sediment and larval lamprey growth. Organic matter accrual was not significantly affected. The results suggest that lamprey larvae benefit from native mussels, and that lamprey populations may decrease with the rapid decline of native freshwater mussels.

In summary, the presence of both Dicosmoecus gilvipes and Margaritifera falcata had significant affects on ecosystem processes. Knowledge of species impacts on energy and nutrient dynamics and the physical conditions that control species abundance and distribution is essential to predicting both small- and large-scale consequences of an altered hydrograph, whether due to river regulation or climate change.

Project Objectives

Spatial scales of the food webs that support the growth of juvenile salmonids in California rivers are largely unknown, but such information is essential for management of flows, species, and other ecosystem factors influencing salmonid production. Our project objectives were:

1. to determine whether predictable variation in algal d13C with water velocity, and measurements of consumer d13C and d15N could be used to examine energy flow and trophic structure in food webs of the South Fork Eel River, and other salmon-bearing streams in Northern California, and

2. to provide this information as a basis for examining how changes in access to energy sources following river regulation or invasions of exotic species might influence the access of various web members, including salmonids, to these sources.

Cyanobacteria are ubiquitous in aquatic ecosystems across the earth. In many environments they are present at low abundances, however under certain environmental conditions cyanobacteria bloom and become one of the dominant organisms in an waterbody, degrading aquatic food webs and water quality. Cyanobacteria evolved over 2 billion years ago, and cyanobacterial harmful algal blooms (cyanoHABs) have been documented for decades. Of particular concern is the production of cyanotoxins, secondary metabolites toxic to humans and other organisms, by certain strains of cyanobacteria. Most research of cyanoHABs has been of planktonic blooms in lakes or estuaries, and cyanotoxin production by benthic cyanobacteria in rivers has been more recent, but in many rivers benthic cyanobacteria are the primary source of cyanotoxins. With field surveys and monitoring, manipulative field experiments, and genome-resolved metagenomics, this dissertation investigated the ecology of benthic cyanobacteria in the Eel River, California.

Like most coastal rivers in Northern California, the Eel River is in a Mediterranean climate. During the seasonal summer drought rivers become shallow, slow flowing, and warm, excellent habitat for benthic algal production. Non-toxic benthic algae are consumed by vertebrate and invertebrate grazers and are foundational to the aquatic summer food web. However, these benthic algal assemblages can tip towards toxicity when cyanobacteria begin to dominate the assemblage. Over a dozen dogs have died in the Eel River since the year 2000 due to ingesting cyanobacteria. Neighboring watersheds have also experienced dog deaths from cyanotoxin poisoning, and benthic cyanobacterial mats are an increasing public health concern in multiple Northern California rivers. Prior to this dissertation there were few data on the distribution and ecology of benthic toxigenic cyanobacteria in California rivers.

During the summers of 2013-2015, I documented spatial and temporal patterns of cyanobacterial occurrence and cyanotoxin concentrations in the watershed, showing widespread distribution of anatoxin-a produced by benthic cyanobacteria. Solid phase adsorption toxin tracking (SPATT) samplers were deployed weekly to record dissolved microcystin and anatoxin-a levels at 10 sites throughout the watershed, and 187 cyanobacterial mat samples were collected from 31 sites to measure intracellular anatoxin-a and microcystins. Anatoxin-a levels were higher than microcystin for both SPATT and cyanobacterial mat samples. Species of benthic Anabaena and Phormidium were frequently found to produce cyanotoxins in the Eel watershed. Of the benthic mats sampled, 50% had detectable anatoxin-a (mean µg g-1 DW= 1.71, max= 70.93), while 24% had detectable microcystins (mean µg g-1 DW= 0.067, max= 2.29). SPATT cyanotoxin levels peaked in mid-summer in warm mainstem reaches of the watershed. This is one of the first documentations of widespread anatoxin-a occurrence and anatoxin-a production by benthic cyanobacterial mats in a North American watershed.

Field experiments were used to study the buoyancy of benthic Anabaena spp. mats to understand implications for Anabaena dispersal. Experiments addressed oxygen bubble production and dissolution on the buoyancy of Anabaena dominated benthic mats in response to light exposure. Samples of Anabaena dominated mats were harvested from the South Fork Eel River and placed in settling columns to measure floating and sinking velocities, or deployed into in situ ambient and low light treatments to measure the effect of light on flotation. Floating and sinking occurred within minutes and were driven by oxygen bubbles produced during photosynthesis, rather than intracellular changes in carbohydrates or gas vesicles. Light experiment results showed that in a natural ambient light regime, mats remained floating for at least 4 days, while in low light mats begin to sink in <24 hours. The ability of Anabaena mats to maintain their buoyancy will markedly increase their downstream dispersal distances. Increased buoyancy also allows toxin-containing mats to collect along shorelines, increasing threats to human and animal public health.

Within cyanobacterial mats are a consortia of microbes interacting together to process and exchange molecules to maintain their growth. Currently, little is known about the diversity of the biosynthetic capacities of cyanobacterial species and associated microbes in freshwater benthic mats in rivers. I sampled 22 Phormidium mats collected across the Eel River network and used genome-resolved metagenomics to 1) investigate cyanobacterial and co-occurring microbial assemblage diversity, 2) probe their metabolic potential, and 3) evaluate their capacities for toxin production. From the genomes of seven strains from one species group we describe the first anatoxin-a operon from the genus Phormidium. Importantly, community composition within the mat appears to be associated with the presence of cyanobacteria capable of producing anatoxin-a. Bacteroidetes, Proteobacteria, and novel Verrucomicrobia dominated the microbial assemblages. Interestingly, some mats also contained Candidate Phylum Kapabacteria and Candidate Phyla Radiation bacteria from Absconditabacteria (SR1), Parcubacteria (OD1) and Doudnabacteria (SM2F11). Although the majority of genomes were unique to a particular sample, metabolic diversity was low across samples. In addition to oxygenic photosynthesis and carbon respiration, metabolic capacities include aerobic anoxygenic photosynthesis, sulfur compound oxidation and breakdown of urea. These results show the importance of organic carbon and nitrogen to energy flow and nutrient cycling within mats and the interactions between potential anatoxin-a production and microbial assemblage composition.

By combining watershed-scale observations, field experiments, and genomic analyses this dissertation provides novel information about the ecology of toxigenic benthic cyanobacteria in California rivers. Comprehensive knowledge of the ecology of benthic cyanobacteria is necessary to understand it’s impacts to public and ecosystem health and to better predict where and when blooms might occur in rivers.

Found in nearly all terrestrial ecosystems, small in size and able to occupy a variety of hunting niches, spiders' consumptive effects on other arthropods can have important impacts for ecosystems. This dissertation describes research into spider populations and their interactions with potential arthropod prey in California's native and non-native grasslands. In meadows found in northern California, native and non-native grassland patches support different functional groups of arthropod predators, sap-feeders, pollinators, and scavengers and arthropod diversity is linked to native plant diversity.

Wandering spiders' ability to forage within the meadow's interior is linked to the distance from the shaded woodland boundary. Native grasses offer a cooler conduit into the meadow interior than non-native annual grasses during midsummer heat. Juvenile spiders in particular, are more abundant in the more structurally complex native dominated areas of the grassland. Potential prey species and abundance differ along the trajectory from woodland boundary to meadow interior and may have consequences for the survival of juvenile spiders that reside primarily near the woodland boundary.

Spiders' ability to influence the eating habits of potential prey species is regulated by temperature and grassland characteristics. In experimental wolf spider enclosures, cooler habitats in which spiders were present evidenced lower herbivory by large chewing invertebrates. Herbivory differed across a gradient of temperature and sunlight but newly restored plants in exposed and warmer soil faced higher rates of herbivory than established plants; herbivory may be better mediated by spider activity in cooler environments.