California Sea Grant College Program

Studies of fish diets can improve the understanding of trophic distributions

and the predatory role of a species in an ecosystem. Identifying the spatial and

temporal variability in the diets of fishes can provide useful information for stock

assessments and management. Yellowtail Rockfish (Sebastes flavidus) are a

commercially and recreationally important fishery species, and abundant along

the central California coast, yet the most recent studies of diets occurred over 20

years ago in Oregon and Washington. To provide current information from

California, I examined the food habits of Yellowtail Rockfish collected near

Cordell Bank, the Farallon Islands, and Half Moon Bay in 2013 and 2014 using

gut content and stable isotope analyses. Yellowtail Rockfish analyzed in this

study were semi-pelagic predators, feeding primarily on tunicates, crustaceans,

and teleosts. Based on δ15N values, fish caught in 2013, on average, were

feeding at lower trophic levels than those caught in 2014, and δ13C values

indicated that fish caught at the northern-most sites were feeding on more

pelagic-influenced carbon sources. Yellowtail Rockfish in central California can

be described as opportunistic feeders because predation patterns were

temporally localized, and diets consisted mostly of transitory prey sources. The

Yellowtail Rockfish diet information presented in this study fills data gaps of a key

life history component, and will be useful for future stock assessments.

Riverine landscapes are shaped by dynamic and complex interactions between streamflow and floodplain landforms, and these physical processes drive productive and diverse freshwater ecosystems. However, human activities have fundamentally altered river-floodplain processes and degraded ecosystems. Flow regime variability has been homogenized and floodplains disconnected from rivers due to dams, diversions, levee building, and land use change. Reconciling competing demands to support ecosystems and resilience to future change is a core scientific and management challenge. This dissertation describes spatiotemporal dynamics of floodplain environments, introducing a method for flood regime classification and establishing a methodological approach for hydrospatial analysis to quantify and evaluate the response of floodplain inundation patterns and related physical habitat to restoration and flow regime change under climate change. It is motivated by the need to develop process-based and landscapescale strategies to better manage flow regimes and landscapes together, such as coordinated levee-removal floodplain restoration and environmental flow allocations. River restoration literature is synthesized herein to examine trajectories from form-based to process-based approaches, recognize that highly modified large rivers may require coordinated physical habitat restoration and environmental flows implementation, and suggest opportunities for improved integration of restoration strategies. A river’s flood regime drives a variety of different physical and ecological functions. Characterizing different floods of a flood regime informs understanding of climate and watershed processes and the management of natural floodplain dynamics. Following cluster analysis approaches used in flow regime classification, a flood regime typology was developed for the Cosumnes River, the only major unregulated river of the west slope Sierra Nevada, California, USA. A primary contribution of this study is the establishment of flood regime classification that moves beyond typical flood frequency analysis to address a range of ecologically-relevant flood characteristics, including duration and timing. Rehabilitating freshwater ecosystems of highly modified rivers under a changing future requires improved understanding and quantification of land-water interactions. Despite ecological implications, quantification of spatiotemporal variability is rare, particularly for management applications. An approach for evaluating spatiotemporal floodplain inundation patterns, or the hydrospatial regime, is presented in several studies. Physical inundation characteristics and associated habitat were quantified in space and time, and responses to restoration and climate change induced flow scenarios were evaluated and compared. The multi-metric approach is demonstrated for a recent levee-removal restoration site along the lower Cosumnes River. The novel hydrospatial analytical approach developed and presented herein applies twodimensional hydrodynamic modeling and spatial analysis to quantitatively summarize, in space and time, a range of ecologically-relevant physical metrics relating to inundation extent, depth, velocity, frequency, -iiiduration, timing, rate of change, connectivity, and heterogeneity. Comparison of metrics before and after levee-removal restoration on the Cosumnes River floodplain showed that while inundation extent greatly increased with restoration, responses varied in space and time and were different for different metrics. Changes in metrics were most substantial at intermediate flood flows. Subsequently, habitat criteria for a native floodplain fish species, Sacramento splittail (Pogonichthys macrolepidotus), were applied to the physical metrics. Findings suggest that restoration nearly doubled overall habitat availability, though benefits varied considerably in space and time. Flow-habitat relationships were nonlinear and not oneto- one, indicating habitat availability mediated by the physical complexity of the floodplain. Finally, floodplain responses to climate change induced streamflow scenarios were compared and the relative impacts of levee-removal restoration across the scenarios were evaluated. Results reflected the balance of increasing extreme winter flooding and declining spring flooding under future climate change scenarios. Magnitude and direction of change depended on the climate change scenario and metric. Levee removal had the general effect of dampening climate change impacts, though the relative impacts of climate change scenarios were greater than that of restoration in some cases. This body of work presents a new methodology to analyze flow-landscape interactions, and in turn contributes to understanding of flow-ecology relationships, susceptibility to anthropogenic change, and improvements to water and land management. Several broad implications emerge from this research. It demonstrates the capacity of a riverine landscape to serve different functions at different times and supports improved management toward variable conditions. Another contribution is advancing the use of hydraulic metrics over hydrologic metrics for better connections between physical processes and ecological functions. Further, the approach allows for ecologically-relevant criteria that are spatially and temporally dependent to be evaluated explicitly (e.g., duration, connectivity, temporal sequence of flood events). Findings show that, for habitat evaluation within complex floodplain environments, habitat availability is not likely to be a simple function of flow. Floodplain hydrospatial regime responses to climate change will be mediated by flow-landscape interactions, with the potential for physical restoration activities to mitigate impacts of climate change. Despite highly modified physical processes, climate change, and freshwater diversity and productivity declines globally, there is great capacity to better balance human and ecosystem requirements. This dissertation expands scientific understanding of and informs management toward dynamic and heterogeneous riverine landscapes that support functional and resilient ecosystems.

The majority of Northern California estuaries are small, flooded, river valleys that are largely unstudied due to their small sizes and remote locations. Yet these estuaries serve as important nursery areas for many marine fish species including rockfish, flatfish, smelt, and herring, and they are vital to anadromous species such as Chinook Salmon (Oncorhynchus tshawytscha) and Steelhead (O. mykiss). I sampled the summer and winter fish and invertebrate communities of the Big, Mad, and Ten Mile river estuaries. Fish were sampled via beach seine or fyke net and invertebrates were sampled via benthic cores, June 2014-June 2016. This research is part of a larger suite of studies establishing baseline conditions in Northern California Marine Protected Areas (MPAs). Big and Ten Mile river estuaries in Mendocino County were designated as MPAs in 2012. The Mad River Estuary in Humboldt County was selected as a non-MPA site to investigate its potential as a reference estuary. In the Mad River Estuary, additional sampling was conducted and a diet study was carried out on the feeding habits of two benthic fishes: Pacific Staghorn Sculpin (Leptocottus armatus) and English Sole (Parophrys vetulus). iii Fish abundance and diversity varied more by season (i.e. summer, winter) than by estuary, while invertebrate diversity varied more by estuary than by season. The Big River Estuary had the strongest ocean connection and the most marine fish and invertebrate species. The Mad River Estuary fish and invertebrate communities were most similar to the Ten Mile River Estuary, which had the least ocean connectivity and species diversity. Additional sampling in the Mad River Estuary showed that fish and invertebrate communities were diverse from spring through fall, and that invertebrate communities within an estuary differed more by upstream distance than by season. Staghorn Sculpin diet in the Mad River Estuary varied by location of capture, but not by season.

Traditionally, the recreational fishery for Pacific halibut has been open in California from 1 May through 31 October. In 2014, however, the Pacific halibut fishery was closed in California during the month of August for the first time in history in an effort to reduce harvest and bring total catch closer to what is allocated to our region by the Pacific Fisheries Management Council (PFMC) Catch Sharing Plan. To determine the effects that the closure had on businesses along the North Coast, I conducted an economic impact survey in 2014. The results of the survey showed that fishing-related businesses lost between zero percent and eight percent of their revenue in 2014, as a result of the closure; lodging and traveler service companies lost between 0.3 percent and one percent of their revenue in the same year. None of the businesses changed the number of employees as a result of the closure. We estimated a decrease in revenue for businesses on the North Coast to be between $189,750 and $222,250. Age and growth are important components in stock assessment models, but biological data in general are scarce on populations of Pacific halibut found in northern California. For this reason, I conducted a study that examined the age and growth of iii Pacific halibut landed in this region, expanding on a previous study to examine possible interannual variation in the age/growth structure, and broadened the study into central Oregon, to compare two distinct bioregions. Results from my study show that mean size-at-age of female Pacific halibut from northern California and central Oregon was larger than those from the IPHC setline surveys in most of Alaska, but similar to those from Oregon and Washington. In addition, fish from this study in northern California and central Oregon were smaller for a given age than those from the 2014 IPHC survey conducted in northern California. Possible reasons for the trend in size-at-age include poor oceanic conditions during my study, the movement of slower-growing halibut into northern Californian waters, and sampling error. The maturity stage of female gonads is also an important component in stock assessment models, but these data are also scarce for Pacific halibut populations in northern California. For this reason, I conducted a study that characterized the maturation of Pacific halibut landed in northern California and central Oregon. I also compared the macroscopic maturity staging method currently utilized by the IPHC against the more rigorous microscopic methods (microscopic staging and measuring oocyte diameter). Results of this study and that of Perkins (2015) indicate that Pacific halibut caught in northern California and central Oregon matured three years earlier than those caught during IPHC setline surveys in waters off of Alaska, and about a year earlier than those caught by the IPHC in Oregon and Washington. The length-at-50%-maturity for Pacific halibut caught in northern California and central Oregon was smaller than that of fish caught in the IPHC setline survey. In addition, for all three stages of maturity observed in iv females (immature, mature, and resting; spawning-stage females were not observed) there was at least 66 percent agreement between macroscopic and histological staging methods, with the highest level of agreement (94 percent) seen in mature ovaries. This study largely validated the macroscopic staging methods because of its high accuracy in identifying mature ovaries; the inaccuracy in distinguishing resting versus immature ovaries had little effect on length- and age-at-maturity analysis.

Marine Protected Areas (MPAs) are a management tool used to protect and sustain many ecologically and economically important fish species from overexploitation by recreational and commercial fishing. Lingcod (Ophiodon elongatus) and some of its prey species, such as rockfish (Sebastes spp.), are species that are protected from fishing in some California MPAs. Lingcod is an apex predator that consumes a variety of fish and invertebrate species. In this study, I sought to assess the effect of an MPA on the abundance, size and diet of Lingcod. I hypothesized that Lingcod in a no-take MPA would be more abundant and larger than Lingcod in an adjacent reference site (REF) that was open to fishing. Furthermore, I hypothesized that diet would differ between Lingcod in caught the MPA and Lingcod in the REF. I collected Lingcod from the Point Buchon State Marine Reserve (MPA) and an adjacent REF site that was open to fishing. I measured, weighed, sexed, and collected stomach contents from Lingcod using the gastric lavage (stomach pumping) technique. Then, I identified prey items from Lingcod stomach contents down to the lowest taxonomic level possible and quantified diet composition by percent by occurrence, percent by number, and percent by mass. Lingcod in the MPA consumed more fish prey items than Lingcod in the REF site. Lingcod in the REF consumed more cephalopod prey items than Lingcod in the MPA. I analyzed the four most common prey items (rockfish, anchovies, flatfish, and octopus) for nutritional content. My data suggest that Lingcod increased in size and abundance in a no-take MPA because they do not suffer from fishing mortality. However, a more nutritious diet could also contribute to a biologically significant advantage for Lingcod in the MPA. To address this would require further research focused on calculating the net energy (gross energy extracted from the prey item minus the energetic costs of handling and digesting the prey item) obtained by Lingcod from consuming different fish and cephalopod prey items. MPAs can be an effective management tool for protecting fish stocks, although, it is important to understand the interspecific interactions between predator and prey species to adaptively mange MPAs and the species that reside within them.

The Scott and Shasta rivers, Klamath River tributaries, experience spatial disparity in habitat quality in spring and summer as a result of historical and current landuse.

Juvenile Coho Salmon (Oncorhynchus kisutch) born in the upper tributary reaches often rear in natal streams before migrating to sea. However, those born in the lower reaches often encounter unsuitable habitat and emigrate during their first spring to seek non-natal rearing habitats. It is assumed that these early outmigrants are population losses. This study evaluated first-summer survival, and contribution to the adult population, of non-natal rearing juveniles in the Klamath River Basin. In the spring of 2014 and 2015 juveniles were tagged using Passive Integrated Transponder (PIT) tags as they were leaving the lower Scott and Shasta Rivers. Movement and survival was subsequently tracked using recapture and detection efforts in potential mainstem summer rearing locations. Strontium microchemistry from otolith samples of returning adult Coho Salmon throughout the basin was analyzed to estimate the contribution of non-natal rearing juveniles to adult returns. Few tagged individuals were detected in non-natal rearing habitats, but those detected in these habitats had survival rates comparable to natal-rearing individuals. Otolith analysis indicated that the proportion of juvenile Coho Salmon rearing in non-natal habitats varied by spawning site. In total, 53% of the 116 adults sampled reared in a natal location as juveniles, while 47% reared in a non-natal location. These results suggest that non-natal rearing can contribute to adult returns and could be a significant population segment with increased restoration.

Phenology is the study of seasonal, biological events and how they are influenced by climate. Climate change has prompted an earlier arrival of spring in numerous ecosystems. It is uncertain whether such changes are occurring in coastal upwelling ecosystems, because these regions are subject to decadal climate oscillations and regional climate models predict later seasonal upwelling. To answer this question, chapter 1 investigated decadal changes in the phenology of 43 larval fish species in southern California. The first principal component of this dataset showed a progression towards the earlier appearance of larvae, although 18% of phenological events exhibited seasonal delays. These changes were best explained by a secular trend towards earlier warming of surface waters. Species with earlier phenology were characterized by an offshore, epipelagic distribution, while fishes with delayed phenology were more likely reside in coastal, demersal habitats. Chapter 2 focused on improving understanding of how oceanic factors affect fish spawning habitat. Spawning habitat models can be applied to examine variations in fish phenology. Using data from spring cruises conducted between 1998-2004, dynamic height was investigated as a variable affecting the spawning habitat of anchovy, sardine, and jack mackerel. The greatest probability of encountering anchovy, sardine, and jack mackerel eggs occurred at dynamic heights of 79–83 cm, 84–89 cm, and 89–99 cm, respectively. Dynamic height explained more variance than any other variable (e.g., temperature, salinity, chlorophyll, zooplankton volume, geographic currents, eddies) in models of sardine and anchovy spawning habitat. Chapter 3 examined variations in phytoplankton phenology across the North Pacific using a hindcast of the Community Earth System Model 1.0 (CESM1) forced with atmospheric observations. Comparisons with SeaWiFS chlorophyll indicated that CESM1 could simulate mean dates of phytoplankton bloom initiation with as much skill as it could predict mean bloom magnitude. The first principal component for each of five phenological metrics (bloom initiation, midpoint, termination, duration, and magnitude) was either correlated with the Multivariate ENSO Index or displayed a long-term trend. Compared to terrestrial ecosystems, long-term trends in phytoplankton phenology were noteworthy due to their rapid rate of change and greater prevalence of delayed phenology.