Climate change has led to a gradual increase of sea surface temperatures in the Northern Gulf of Alaska (NGA) interspersed with marine heatwaves (MHW) that impose a rapid but temporary perturbation of sea surface temperature. MHWs have the potential to alter marine microbial community structure, which may impact the production and transfer of carbon to higher trophic levels. The year 2019 was characterized as an MHW in the North Pacific, with sea surface temperatures in the NGA reaching ~2.5 º C above average during 2019 and ~1 º C above average during 2020, while 2021 had near-average sea surface temperatures. To characterize shifts in the NGA's microbial community, samples for DNA and flow cytometry were collected on NGA Long Term Ecological Research cruises in summers 2018-2021. Flow cytometry sample analysis revealed higher abundances of picoeukaryotes, Synechococcus, and nanoeukaryotes in the summer of 2019 relative to 2020 and 2021 on the continental shelf. The diversity of eukaryotic microbes was lower during 2018-19 than 2020-2021, with similar patterns observed within the diversity of individual eukaryotic taxa. Conversely, the diversity of prokaryotic microbes was higher during 2019 than 2020. Different environmental conditions were correlated with small cell abundance and microbial diversity. Elevated picoeukaryote abundance was associated with higher temperature and inversely correlated with chlorophyll a concentration, while Synechococcus abundance was anti-correlated with the concentration of nitrate and phosphate. Shannon diversity of 18S reads correlated with lower salinity measurements while Shannon diversity of 16S reads was not significantly correlated with any tested biological or environmental variables. These correlations indicate that increases in sea surface temperature, along with associated changes in nutrient concentrations and salinity, act as environmental drivers with the potential to shifts the NGA's microbial community structure. Such a community shift towards pico-nanophytoplankton may reduce trophic transfer efficiency and decrease the production of fisheries and other higher trophic levels in a warmer NGA.

Monitoring the status of Arctic marine ecosystems is aided by oceanographic moorings that autonomously collect data year-round. Near Hanna Shoal in the northeast Chukchi Sea, the Chukchi Ecosystem Observatory moorings include an ASL Environmental Sciences Acoustic Zooplankton Fish Profiler (AZFP) datalogger, a multi-frequency upward-looking sonar that is programmed to collect data from across the upper 30 m of the water column every 10-20 seconds. Using six years of nearly continuous data, here we describe a statistical analysis of the datalogger's 455 kHz acoustic backscatter return signal. When used in conjunction with a selforganizing map machine learning algorithm, these data allow us to accurately differentiate between the presence of sea ice and open water and characterize surface waves. The approach detects short-duration (e.g., 15 minutes or longer) sea ice leads that pass over the mooring in winter, and sparse ice floes that pass over in summer. The ability to algorithmically identify small-scale features within the information-dense acoustic dataset enables rich characterizations of sea ice conditions and the ocean surface wave environment. Example applications include quantifying the recurrence of leads during ice-covered seasons, sparse ice in otherwise open water, statistics of ice keels and level ice, and wave height statistics. By automating the acoustic data processing and alleviating labor- and time-intensive analyses, we can maximize the use of these year-round acoustic data. Beyond applications to newly produced datasets, the approach opens possibilities for the efficient extraction of new information from existing upward-looking sonar records from recent decades.

Surface tides, when obstructed by bottom relief, give rise to periodic oscillations within the stratified oceanic interior. Such transformation of the depth independent (barotropic) tide into internally propagating (baroclinic) waves comprises 1/3 of the global energy losses from the surface tide. Internal waves of tidal period known as internal tides tend to have low vertical shear and hence are very stable and long lived. They have been observed to propagate essentially unchanged across ocean basins. Details of the internal tide wave life-cycle are not well known, yet turbulent dissipation powered by the slow decay of these waves is one of the key processes shaping deep ocean water properties. The Tasman Sea stands out as a natural laboratory to investigate the internal tide life cycle. In this dissertation, the generation and propagation of internal tides were examined by means of realistic simulations of ocean circulation under varying conditions, and were compared to observations obtained during the Tasman Tidal Dissipation Experiment (TTIDE). The simulations reveal that the barotropic-to-baroclinic conversion is intensified at the Macquarie Ridge near New Zealand by coupling with secondary, nonlocally produced internal tides. Because of this complexity, regionally varying hydrographic conditions drive remarkable temporal and spatial variability of internal tide generation. The internal tides that are created at the ridge constructively superpose into a spatially confined, beam-like feature (Tasman beam) that radiates across the Tasman Sea over 1000 kilometers from its generation region and reaches the Tasman shelf. The beam is described well at first order by simple plane wave propagation theory, but also exhibits non-plane wave characteristics associated with diffraction. Additional intricacy arises from development of a standing wave, the result of the beam's reflection near Tasmania. Temporal changes include hydrography-induced refraction and strong perturbations from interactions with eddies. It is concluded that in-situ mooring measurements and ship surveys of internal tides exhibit a great deal of apparent spatial and temporal variability that can be difficult to interpret. This variability can largely be eliminated in the analysis of numerical models which allow the underlying wave field energy life cycle to be quantified.

The Northern Gulf of Alaska (NGA) shelf is a productive high-latitude environment where nutrient dynamics are greatly impacted by the seasonal variability in freshwater input and water column mixing. Iron is a key nutrient on the NGA shelf that directly modulates primary production, but inputs are difficult to quantify due to high biological uptake and control exerted by Fe-binding organic ligands. Other lithogenic elements such as aluminum and manganese have the same sources as iron (rivers and sediment) and similar abiotic removal via particle scavenging, but exhibit quasi- conservative behavior in seawater allowing for their use as tracers of these sources. Thus, Al and Mn distributions can help provide insight into iron inputs and the relative importance of various mechanisms influencing nutrient dynamics in the NGA. The data are derived from spring, summer, and fall NGA LTER (long term ecological research) cruises from 2018 and 2019 that included a focused five-day Copper River plume study, several surface transects from Kayak Island to Kodiak Island, and vertical profiles at several locations sparsely distributed throughout the shelf. We find that seasonal patterns in the surface concentrations of dMn and dAl mirrored annual glacial melt cycles, with the lowest values observed in spring and higher values in summer and fall. Spatial patterns were also apparent as both metals tended to be lower offshore than inshore, and were also lower overall (by 1-2 orders of magnitude) on transects further from the outflow of the Copper River, a major source of freshwater to the NGA. Extremely high concentrations in the Copper River plume (≤1395 nM dAl, ≤128 nM dMn) and strong correlations with salinity (p < 0.0001) highlight their quasi-conservative nature, and their usefulness as tracers of freshwater input, which helps inform iron inputs from this source. Enhanced dAl and dMn concentrations within nepheloid layers in subsurface waters indicate regions where a sedimentary source of iron is likely to be important. Residence times for dAl and dMn in surface waters over the NGA shelf were estimated to be 31 days (dAl) and 42 days (dMn) on average based on summer and fall data from both years.

Egg production and copepodite growth rates were measured for the calanoid copepods Pseudocalanus spp., Calanus marshallae/glacialis, and Metridia pacifica in the northern Bering and southern Chukchi Seas during June of 2017 and 2018. For all taxa, instantaneous growth rates generally decreased with increasing copepodite stage, though the differences between most stages was not significant. The growth rates for Pseudocalanus spp. averaged 0.03 ± 0.002 day⁻¹, Calanus spp. 0.09 ± 0.004 day⁻¹, and M. pacifica 0.05 ± 0.03 day⁻¹. Egg production rates increased with prosome length for all species, but when standardized to body weight this trend reversed. All Pseudocalanus species had similar weight-specific egg production (SEP): 0.18 ± 0.01 for P. acuspes, 0.15 ± 0.00 for P. newmani, and 0.11 ± 0.02 for P. minutus. The SEP for Calanus was considerably lower, 0.09 ± 0.01, while for M. pacifica it was 0.11 ± 0.01. These rates suggest considerable discrepancies between growth rates and egg production weights that we propose are due to differences in life history strategies. Pseudocalanus reproduce nearly year round, they appear to invest less in somatic growth, preferring to quickly reach their adult stage where they invest heavily into reproduction. Calanus spp. have 1 or possibly 2 generations per year in this region, they invest more into somatic growth in order to ensure their population is ready for a reproductive season timed to the spring phytoplankton bloom. The more omnivorous M. pacifica is also likely limited to 1 or 2 generations, although their ability to thrive on a wider range of food sources than Calanus seems to allow for relatively higher investment in reproduction and perhaps lower investment in somatic growth. Consistent with other studies, global growth models do not match our observations particularly well, likely because they are dominated by egg production estimates at lower latitudes.

Small pelagic fish abundances can vary widely over space and time making them difficult to forecast, partially due to large changes in the number of individuals that annually recruit to the spawning population. Recruitment fluctuations are largely driven by variable early life stage survival, particularly through the first winter for cold temperate fishes. Winter survival may be influenced by juvenile fish size, energy stores, and other factors that are often poorly documented, which may hamper understanding recruitment processes for economically and ecologically important marine species. The goal of this research was to improve understanding of recruitment of Pacific herring (Clupea pallasii) within Prince William Sound (PWS) through recruitment modeling and by identifying factors influencing winter survival of young-of-the-year (YOY) herring. Towards this end, my dissertation addresses three specific objectives: 1) incorporate oceanographic and biological variables into a herring recruitment model, 2) describe patterns in growth and condition of PWS YOY herring and their relationship to winter mortality risks, and 3) compare the growth, condition, swimming performance, and mortality of YOY herring that experience different winter feeding levels. In the recruitment modeling study, annual mean numbers of PWS herring recruits-per-spawner were positively correlated with YOY walleye pollock (Gadus chalcogrammus) abundance in the Gulf of Alaska, hence including a YOY pollock index within a standard Ricker model improved herring recruitment estimates. Synchrony of juvenile herring and pollock survival persisted through the three-decade study period, including the herring stock collapse in the early 1990s. While the specific mechanism determining survival is speculative, size-based tradeoffs in growth and energy storage in PWS YOY herring indicated herring must reach a critical size before winter, presumably to reduce size-dependent predation. Large herring switched from growth to storing energy, and ate more high-quality euphausiid prey, which would delay the depletion of lipid stores that compelled lean herring to forage. Lipid stores were highest in the coldest year of the seven-year field study, rather than the year with the best diets. With diets controlled in a laboratory setting, spring re-feeding following restricted winter diets promoted maintenance of size and swimming ability, but had little effect on mortality rates compared to fish continued on restricted rations. Declines in gut mass, even among fully fed herring, and low growth potential suggest limited benefits to winter feeding. Mortalities due to food restriction compounded by disease were highest among herring that fasted through winter months, and among small herring regardless of feeding level. Taken together, these findings illustrate the importance of achieving a critical size and high lipid stores in the critical period before winter to promote YOY herring winter survival and ultimately recruitment.

It has been hypothesized that climate change will reduce the strength and episodic nature of vernal phytoplankton blooms, increase heterotrophy of microbes and zooplankton, and weaken the tight coupling between pelagic and benthic production that is characteristic of Arctic continental shelves. As a part of the Arctic Shelf Growth, Advection, Respiration, and Deposition rates measurement (ASGARD) project, I quantified sinking particle fluxes and incubated sinking particles to measure the rate of microbial respiration associated with those particles. These measurements were used to characterize the strength of the pelagic-benthic connection. After a record-breaking year of warm temperatures and low-ice conditions in the northern Bering and southern Chukchi Seas, we observed massive vernal fluxes of sinking particulate organic carbon, ranking amongst the highest observed in the global oceans. Moreover, low rates of particle-associated microbial respiration indicate negligible recycling of sinking organic matter within the water column. These results suggest that the strength of the biological carbon pump may be maintained or enhanced in a warming Arctic, supporting strong benthic and upper trophic level productivity and carbon export.

The magnitude and spatio-temporal patterns of particulate material flux from the surface ocean through mesopelagic and bathypelagic depths determines sequestration of atmospheric carbon and the food supplied to deep-dwelling ocean life. The factors that influence how and where this organic material is exported from euphotic depths are poorly understood. Zooplankton are thought to play a key role in modulating the transport of surface-produced particles to depths through consumption, fragmentation, active diel vertical migration, and fecal pellet production, thus it is important to study both particulate matter and zooplankton in tandem. In this study, I use an in-situ optical instrument, the Underwater Video Profiler 5 (UVP5), to describe broad scale patterns of large (> 100 μm) particles and zooplankton across a longitudinal transect of the Pacific Ocean during April to June 2015. Satellite-derived surface chlorophyll-a was employed to describe the timescales over which particles arrive in meso- and bathypelagic depths after a productivity peak. High abundances and volumes of particles are noticeable beyond the euphotic zone across the Equator, transition zone, and the sub-arctic Pacific, indicating increased export in these high-nutrient low-chlorophyll (HNLC) areas. In two of these areas, the Equator and transition zone, large abundances and volumes of particles extend into bathypelagic depths. High abundances of zooplankton were seen in all areas where high abundances of particles are seen in bathypelagic waters. Rhizaria were revealed to be pervasive across all biogeographic regions, and appear to play a role in particle attenuation in the sub-arctic Pacific. The insight into patterns between particles, zooplankton, and productivity identify HNLC regions as deserving more detailed examination in future studies of biological pump efficiency.

Trace metals in the marine environment are found in trace amounts, but are important tracers of oceanographic processes, and bioactive trace metals can impact ocean biogeochemistry through their nutrient or toxic influence of microbial populations. Sea ice is an intrinsic feature of the Arctic Ocean that likely plays a key role in the cycling of trace metals, given that this substrate can concentrate, alter, and transport these elements. Warming conditions in the Arctic have decreased sea ice cover over the past decades and the loss of sea ice threatens to drastically change the Arctic ecosystem, but the implications are not entirely understood. The scarcity of studies on Arctic sea ice entrained trace metals is due in part to the lack of commercially available sampling equipment capable of collecting sea ice without introducing contamination, and in part to the logistic and economic difficulties in accessing remote Arctic sea ice sites. Natural heterogeneity related to large sediment loads incorporated in uneven patches across Arctic fast ice poses a challenge when designing observational studies of trace metals in sea ice. The scope of this thesis is on the study of trace metals in Alaskan Beaufort Sea fast ice environment. The study includes snow, sea ice and seawater under the ice. Analysis of dissolved (Mn, Fe, Cu and Zn) and particulate (Al, Mn, Fe, Cu and Zn) phases was carried out from 50 ice cores collected with a trace metal clean ice corer developed at the University of Alaska Fairbanks. The results of this study indicated that the ice corer developed at UAF was able to collect uncontaminated samples. Highly variable and elevated concentrations of particulate (> 0.2 μm) trace elements were observed due to the notable variability in the amount of sediment incorporated within ice cores, but surprisingly dissolved (< 0.2 μm) metal concentrations were relatively low and consistent. The observed low dissolved metal concentrations, along with low bulk salinity and low percent leachable particulate trace metal fractions, suggest that desalination removed reactive metals from the ice matrix prior to sampling. Spatial variability of dissolved and particulate trace metals was statistically analyzed and indicated generally negligible variability on the meter scale, but significant variability on the kilometer scale, for both size classes. These results emphasize that future studies of trace metals in sea ice should include temporal and spatial considerations.

The study examined adaptation in the seagrass genus Phyllospadix to rocky substrates, habitats not generally exploited by seagrasses. One hypothesis tested whether the genus exhibits anatomical features distinguishing it from other seagrasses. A corollary predicted that individual Phyllospadix species show additional specialization, based on observations that three species are distinctly zoned where they occur together. A second hypothesis tested a model of carbon assimilation that predicts that submerged aquatic plants growing on hard substrates, such as Phyllospadix species and most marine algae, experience less transport resistances to inorganic carbon uptake than rooted and rhizoidal plants. As a consequence, it was predicted that Phyllospadix species would show enzymatic discrimination against carbon-13 similar to marine algae and dissimilar to other seagrasses. Carbon isotopic variability in Phyllospadix serrulatus and Phyllospadix torreyi was compared with that of the algae Egregia menziesii and Halosaccion amerlcanum growing at the same location. Carbon isotopic variability in eelgrass. Zostera marina, was also examined to provide a basis of comparison to sediment rooted seagrasses. Comparison with Z. marina was useful in defining anatomical features in Phyllospadix that are adaptations to rocky littoral environments. These features include greater hypodermal fiber and roothair development, thickened rhizomes, and smaller lacunae. Comparison among Phyllospadix spp. for microhabitat adaptations was less fruitful. Phyllospadix spp. show carbon isotopic discriminatory patterns distinct from Z. marina and marine algae. Although marine algae and Phyllospadix spp. overlapped isotopically, only the seagrasses became isotopically lighter with increasing intertidal height, probably through atmospheric carbon dioxide incorporation. Carbon isotope ratios in submerged seagrasses did not appear to be affected by water motion, as predicted by boundary layer considerations. An observed correlation between leaf thickness and leaf isotopic ratios also indicated complications to simple models of carbon assimilation in submerged aquatic plants.