Climate change exerts physical influence on estuarine and open coastal morphology through sea level rise and changes to wave energy, precipitation, and sediment supply, resulting in flooding and erosion hazards to coastal communities. Along open coasts, the distribution of wave energy in the nearshore is of critical importance for assessing local vulnerability to climate change. Two methods are frequently used to predict wave conditions based on global climate model outputs of atmospheric circulation: dynamical downscaling and statistical downscaling. Statistical downscaling relies on empirical relationships between waves and atmospheric conditions. Statistical downscaling has been applied with success in relatively small ocean basins (e.g., Mediterranean, Atlantic) where waves are generated over a small area and arrive at the coast within a few days of generation. However, in the Pacific Basin, waves are generated over large and distant regions of both the North and South Pacific. Furthermore, waves can travel up to 3 weeks before arriving at the coast (e.g., Southern Ocean-generated waves arriving in Southern California). These challenges have resulted in statistical downscaling studies with limited success. Chapter 2 of this dissertation addresses these challenges by 1) partitioning wave spectra into families that have unique, discrete generation areas and 2) accounting for the time lag between wave generation and wave arrival at the coast. The success of this work in Southern California bodes well for the proliferation of wave climate projections in large ocean basins.

To project future coastal hazards, deep-water waves predicted using the methods described above must be transformed over shelf bathymetry to the nearshore. In complex coastal regions, offshore canyons, shoals, and islands complicate the linkage of nearshore waves to deep-water waves and the atmospheric conditions that generated them. In Chapter 3 of this dissertation, a hybrid statistical-dynamical approach is taken to explore significant spatial variability in nearshore wave conditions of the Southern California Bight, a complex coastal region. It is found that variability is driven by not only static bathymetric controls, but also dynamic large-scale atmospheric patterns. Climate change effects on these atmospheric patterns will lead to new distributions of wave energy along the Southern California Bight coastline and other coastlines around the world.

Along estuarine coasts, the distribution of tracers, such as salt, sediment, and pollutants, is a key factor in determining vulnerability to climate change and development. Extensive scientific effort has yielded a comprehensive understanding of sediment and salt transport in varied estuarine systems. However, tidal dynamics in shallow embayments, which are commonly found flanking deep estuarine channels, have not been described thoroughly. Chapter 4 of this dissertation examines the momentum and salt forcing associated with a shallow, estuarine embayment. This work illuminates mechanisms likely responsible for trapping of sediments in shallow bays and the supply of sediment to estuarine marshes.

The suite of studies presented in this dissertation seeks to contribute to our scientific understanding of open and estuarine coastal response to climate change and to provide information that can readily be applied to coastal policy and engineering.

Mesoscale eddies in the Gulf of Alaska are thought to contribute to

the shelf-slope exchange of nutrients and plankton, enhancing biological production. We report on a study of two anticyclonic mesoscale eddies in this region observed through in situ sampling during August and September 2007. Both eddies exhibited in their cores theta-S profiles with warmer, fresher water relative to the properties of the ambient basin water between 150 and 300 m depth.

Hydrographic properties and satellite altimetry data were analyzed

to identify likely formation regions for each feature. One eddy, sampled near Yakutat, Alaska, originated in the Sitka formation region (221-223° E); the second eddy, sampled south of Kodiak Island, originated near the Kenai Peninsula, southeast of the Kennedy and Stevenson entrances to Cook Inlet — an area not previously studied as a formation region. Subsequent analysis of 16 years of satellite altimeter data (from 1992 to 2008) with an algorithm designed to identify and track eddies revealed approximately 6 Kenai eddies that have formed in this region. Although this number constitutes only 3.2% of the 188 eddies identified by the algorithm during this period, it represents 15.4% of the 39 eddies that formed in or propagated westward into the Alaskan Stream.

Once emerged from the gravel after being spawned in natal streams, Chinook salmon spend many months rearing and growing in freshwater before undergoing smoltification and out-migrating to the ocean. This relatively short period of time is considered to be the most vulnerable and dangerous phase in the life cycle of a Pacific salmon. It is during this phase when smolts navigate around many anthropogenic structures and experience environmental stressors while making their way to the ocean. In California’s Central Valley, the few remaining wild populations of Chinook salmon (Oncorhynchus tshawytscha) out-migrate through a highly modified riverine and estuary landscape characterized by leveed banks, altered flow and temperature regimes, transformed food webs, and limited floodplain and rearing habitat. Juvenile salmon smolts migrate through these landscapes within a relatively short period of time, requiring them to quickly adapt to changing water conditions and habitat types. Understanding the survival rates of wild smolts from source tributaries to the Pacific Ocean is essential in protecting and restoring these populations from the low abundances currently observed. When faced with drought conditions out-migrating smolts experience low flows, elevated water temperatures and high densities of predators while out-migrating to sea. In order to assess smolt survival during drought conditions in late spring (April-May), 304 wild smolts were acoustically tagged and tracked from Mill Creek (Tehama County) to the Pacific Ocean between 2013 and 2016. Total outmigration survival to the ocean was 0.3% during these years, with only one fish making it to the Golden Gate and the Pacific Ocean. These survival estimates are some of the lowest ever recorded for salmon out-migrating to the Pacific Ocean, with much of the mortality occurring within Mill Creek and the Sacramento River. Cumulative survival through Mill Creek (rkm 452-441) was 68% (±12 S.E.), and cumulative survival through the Sacramento River (rkm 441-203) was 7.6% (± 16 S.E.) These low survival rates are likely attributed to low flows in Mill Creek and the Sacramento River resulting from critically dry winters between 2013 and 2015, which were reduced even further by water diversions for agriculture in both Mill Creek and the Sacramento River. During periods of higher flow in 2016 survival rates dramatically increased, suggesting that more water in Mill Creek and the Sacramento River is necessary to improve in-river smolt migration survival during the late spring.

Eastern Boundary Current systems, running along the west coasts of Africa and the Americas, are among the most biologically productive oceanic ecosystems. Their disproportionately large contributions to global marine primary productivity (photosynthesis) and fish catch are supported by upwelling of deep, nutrient rich water, a process driven by the interaction of surface winds and Earth's rotation. Upwelling in these systems may be forced by two mechanisms: equatorward winds at the coastal boundary (coastal divergence), or a cross-shore gradient in the magnitude of winds (wind stress curl). Though bulk estimates of upwelled volume and the individual contributions of coastal divergence and wind stress curl have been estimated, their roles in regulating productivity remain unclear. Similarly, while nutrient supply from upwelling has been measured or modeled in specific regions, its dependence on variability in physical factors including topography, stratification, and latitude has not been adequately addressed. The measurement of primary productivity itself is laborious and complicated in situ, and satellite-borne ocean color sensors currently represent our best option for obtaining the large-scale estimates needed to constrain global carbon budgets. However, after over a quarter century of satellite primary productivity model development, performance has improved little.

The research in this dissertation employs a suite of cutting edge oceanographic tools - computer models, satellite sensors, and autonomous underwater platforms - to elucidate controls on primary productivity and its measurement in coastal upwelling systems. First, an idealized numerical model is used to evaluate the respective roles of stratification, continental shelf topography, latitude, and wind stress magnitude on upwelling source depth and nutrient delivery to the sunlit surface layer (Chapter 1). Next, this analysis is extended to investigate the impact of nearshore reduction in surface wind stress, a poorly constrained process with significant implications for bottom-up control of phytoplankton growth (Chapter 2). Finally, an extensive record of shipboard observations off the California Coast is analyzed to evaluate limitations on primary productivity estimation, and synergistic use of satellites with autonomous underwater gliders is identified as fertile ground for improvement (Chapter 3).

The enhancement of coastal upwelling on the downwind (i.e., equatorial) side of large headlands is studied using a three-dimensional, high-resolution numerical model in an idealized domain. Early simulations show that positive (negative) wind stress curl in the northern (southern) hemisphere does not enhance upwelling compared to having steady winds with zero wind stress curl. Multiple idealized grids are examined to determine the main topographic variations driving enhanced upwelling. The grid with a cape and varying shelf width (widening downstream of the tip of the cape) exhibits the most upwelling enhancement. The widening shelf width contributes to upwelling through a fundamental process derived from the steady-state continuity equation in the shallow water equations. Depth-averaged flow divergence must be accompanied by depth-averaged onshore cross-isobath flow. We find the largest amplitude cross-shore transport occurs near the shelf break, where the shelf is widening downstream. The cape geometry still further enhances upwelling. Examination of the along-isobath momentum balance clearly shows a large pressure gradient just south of the tip of the cape, where the shelf is widening and just south of localized low pressure resulting from flow around the cape. Except for very near the coast, the flow remains largely geostrophic, with flow largely following lines of constant pressure. This results in enhanced onshore cross-isobath flow just south of the low pressure, in the lee of the cape. These combined effects are much larger in the model with both capes and changing shelf width than in models with either no capes or no changing shelf width, leading to the conclusion that the topographic variations work together to create environments particularly favorable to enhanced upwelling as well as those particularly unfavorable.

A three stream irradiance model is implemented and tested to determine its utility for bio-geochemical data assimilation of remote sensing reflectance, Rrs(λ), data in the California Current System. Two different numerical methods were tested to solve the model, but the shooting method using a 4th order classical Runge Kutta scheme was implemented. The optical model solves for Rrs(λ) from the concentrations of variable optical constituents and their respective wavelength dependent light scattering and absorption properties. Using single phytoplankton type absorption and scattering estimates, the optical model produces Rrs values that correlate to satellite observed Rrs more closely for certain phytoplankton types than others. The model produced accurate downward irradiance fields when using observed absorption and scattering profiles obtained from the Ocean Observatories Initiative’s Oregon Shelf Surface Piercing Profiler Mooring. Through this forward modeling based comparison to observations it was found that the optical model can produce accurate profiles under certain conditions, making it promising for data assimilation of Rrs, but outstanding issues remain to be addressed, including accurately resolving community structure, improvements to the parametrization of constituents other than phytoplankton, and atmospheric influence on surface boundary conditions.

The Monterey Peninsula is an ecologically important area, highlighted by numerous marine protected areas (MPAs), but little is known about the specific circulation processes that support its species-rich ecosystems. This suite of research used the Regional Ocean Modeling System (ROMS) to simulate the circulation during the 2014 and 2015 spring/summer upwelling seasons through a series of nested grids to resolve the circulation on the central California coast at approximately 120 m resolution. A particle tracking model, OpenDrift, calculated Lagrangian trajectories to identify source water and simulate near-surface larval transport. We examine the circulation patterns and temperature structure in Carmel Bay, a small embayment in the lee of the Monterey Peninsula. We explore mechanisms driving local population connectivity: self-recruitment and connectivity between populations ~10 km apart by simulating transport of kelp rockfish (Sebastes atrovirens) larvae from populations in southwest Monterey Bay and Carmel Bay. We characterize submesoscale sea surface temperature (SST) fronts off the southern edge of the Monterey Peninsula by strain, that generate different Lagrangian transport patterns. Collectively, these results demonstrate that nearby populations (and thus MPAs) do not have homologous recruitment or exchange, which has important implications for MPA management, and how this heterogeneity results from dynamic circulation processes.

This thesis presents a descriptive analysis of chlorophyll, surface currents, and sea surface temperature within the Gulf of the Farallones, occupying regions of three NOAA National Marine Sanctuaries along the central California coast. The seasonal cycles of each are described from a 24-year chlorophyll-a record, 10-year record of surface currents, and 20-year SST record. EOF analysis of chlorophyll revealed four distinct modes of variability: a single-signed pattern, a north-south split, an onshore-offshore structure, and a three-way split in which the Gulf of the Farallones is opposite to the regions north, south, and seaward. Surface currents via HF Radar and SST are also characterized through EOF analysis, to ultimately investigate their relationship to chlorophyll. Both surface currents and SST are significantly correlated to said patterns in chlorophyll, with the strongest correlation being between the north-south split in chlorophyll and an alongshore current pattern. A 19-year time series of normalized water-leaving radiance at the 555 nm band (nLw555) is used to calculate and describe the seasonal cycle of the San Francisco Bay Plume, which is compared to that of chlorophyll. Monthly and daily anomalies in chlorophyll are both significantly positively correlated to the magnitude of the SF Bay plume. The ecological structure of the area of interest to this study is of considerable importance to understand, particularly in order to implement effective conservation and management strategies.

The base of the marine ecosystem is supported by marine microalgae known as phytoplankton. Phytoplankton rely on the combination of light and nutrients to survive. Plankton and nutrients move with water parcels, and fluid motion dictates the light environment essential to most primary production. In geophysical systems, fluid motion is often described as barotropic or baroclinic. Barotropic motion corresponds to the depth-averaged flow, and is largely independent of stratification. When lateral changes in density occur in a stratified fluid, motion deviates from the depth depth-averaged flow and is referred to as baroclinic motion. Many different types of baroclinic processes exist in the ocean, each one affecting light and nutrient availability for primary production. This dissertation uses idealized numerical models to explore three different baroclinic systems and how each affects light and nutrient availability for primary production in the ocean.

Chapter One investigates the primary production response to the generation of internal tides by fluctuating tidal flow over varying bathymetry. This process concentrates baroclinic wave energy into a coherent structure known as a tidal beam. The tidal beam leads to a large displacement of phytoplankton through a light field that varies exponentially with depth, leading to more light available for primary production. At the same time, the tidal beam elevates the average position of isopycnal surfaces, carrying nutrients and phytoplankton into the euphotic zone. Analysis of Lagrangian parcels that move with ocean currents and representing phytoplankton shows that the effect on nutrient availability enhances primary production more than the increase in light.

Chapter Two turns to a different type of baroclinic wave known as an Island Trapped Wave (ITW). ITWs are analogous to coastal trapped waves where an island acts as a waveguide boundary. In the case considered here, an ITW that travels around the island in a 24-hour period is excited as a resonant response to a land-sea breeze with a 24-hour period. As part of the resonant response, ITWs affect light and nutrient availability for primary production. Nonlinear processes associated with the ITW increase nutrient concentrations in the euphotic zone through elevated advective and diffusive flux divergences. With regard to light availability, a diel light cycle causes a dipole structure of primary production across the island with the largest enhancement of primary production occurring where the upwelling phase of the ITW arrives at noon. To quantify this effect, phytoplankton biomass, nutrient, and light are decomposed into their mean states and fluctuations around their means. Primary production associated with the mean state is the larger, indicating that the enhancement of primary production occurs due to state adjustment of the nutrient field rather than light fluctuations. However, the correlation of light and nutrient fluctuations further increases this enhancement by an additional 30\%.

Chapter Three considers yet another baroclinic process, wind-driven coastal upwelling, and compares the primary production response in two- and three-dimensional idealized numerical models. Coastal upwelling is driven by offshore surface Ekman transport which removes surface water from the coast and replaces it with nutrient-rich waters from below, often causing large phytoplankton blooms. The depth of the upwelled water determines the nutrient content, with deeper waters having higher nutrient concentrations. The geometry of the continental shelf affects the upwelling source depth and thus magnitude of the primary production response. Two-dimensional upwelling theory and numerical modeling studies predict that steeper shelves source water from deeper depths compared to wider shelves. In three-dimensions, changes in alongshore shelf width also adjust upwelling source depth. Deeper, higher nutrient water is transported cross-shelf with the bottom boundary when a shelf widens in the downwind direction. Results from this chapter show that the primary production response is laterally displaced downstream from bottom boundary layer transport by alongshore transport. Primary production is elevated up to 100 km downwind of the change in shelf width after ten days of upwelling. Conversely, where the shelf narrows in the direction of the wind, the reduced transport along the bottom boundary causes less nutrients to be delivered to the euphotic zone and lower levels of primary production. In both cases, changes in shelf width in the alongshore direction locally affect upwelling source depth, and the resulting elevated or diminished nutrient concentrations are then delivered to the euphotic zone downwind of the change in shelf width.

In summary, this dissertation examines how baroclinic processes affect light and nutrient availability differently depending on the context of the generating mechanism. These baroclinic processes primarily affect phytoplankton growth through vertical nutrient transport; however, some processes can affect light availability as well, highlighting the diversity of biological responses to baroclinic motion.