The coastal ocean is an extraordinarily energetic place where water and sediments are always in motion. More than a third of the human population lives near coastline globally, and we are collectively dependent on the coastal ocean for subsistence, commerce, and recreation. Rising sea levels and increasing intensity of storms are just two consequences of climate change that are influencing and will continue to influence the dynamics of coastal ecosystems. It is these dynamic physical characteristics and important mutual influences that make the coastal ocean critical to study but equally challenging to observe.
Oceanographers who study coastal ocean processes face a cost-benefit trade-off when planning sampling efforts. Surveys from oceanographic vessels provide opportunities to take measurements over broad areas, but time constraints, ship costs, and vessel drafts limit surveys. In contrast, instruments mounted on moorings can measure long-term time series, but only at discrete strategic locations. Autonomous underwater vehicles (AUVs) offer a mobile and continuous sampling approach, but AUVs are expensive to deploy and maintain, and strong currents, waves, and salinity gradients can reduce maneuverability or prohibit sampling through exceptionally dynamic regions.
Biologging—attaching miniature sensors to animals—is an emerging method for making long-term, low cost, and widely distributed autonomous measurements of the environment [Biuw et al., 2007; Harcourt et al., 2019]. Marine animals like seabirds and seals often access hard to sample locations, and they do so under their own power. Advances in data transmission and sensor technologies are facilitating the development of miniaturized biologging devices that can make oceanographic measurements and are suitable for small diving marine animals like seabirds.
Oceanographic Measurements from Cormorants
The Cormorant Oceanography Project, initiated in 2013, is advancing biologging tag technologies for use with cormorants to measure in situ oceanographic conditions. Cormorants and shags make up a family (Phalacrocoracidae) of about 40 species of birds that inhabit coastal oceans and inland waterways from the tropics to high latitudes. Marine cormorants typically forage along the seafloor at depths up to 80 meters, and they can make more than 100 dives each day. Between dives, cormorants rest on the sea surface, so their movements allow both water column and surface conditions to be measured with biologging.
The biologging tags we currently use are equipped with small, low-power, fast-response sensors to measure water temperature, conductivity (for water salinity levels), and pressure (for water depth). Each tag also features an inertial measurement unit (IMU) to monitor acceleration and orientation. A GPS unit, triggered when a bird surfaces, provides locations for georeferencing measurements, and solar cells recharge the tags’ batteries (at the time of writing, some tags have been transmitting continually for more than 2 years). The sensors collect large volumes of data that are transmitted and retrieved using two-way cellular communications. Cellular communications also allow us to transmit new sampling programs to the tags (Figure 1).
The data provide measurements from unsampled dynamic coastal marine environments, allowing us to improve model predictions.
We are processing tag sensor data to obtain fundamental information about vertical temperature and salinity profiles, bottom soundings (which measure bathymetry), surface currents, surface gravity wave statistics (which characterize wave motions at the water-air interface), and air-sea temperature contrasts (which help us to understand ocean-atmospheric coupling). The data provide measurements from unsampled dynamic coastal marine environments, allowing us to correct uncertainties in boundary conditions and parameters of ocean models and thus to improve model predictions in a process known as data assimilation.
Processing bottom soundings obtained from pressure records gathered during cormorant dives requires disentangling bird behavior from the data. For example, we use dive shape to distinguish benthic (seafloor) dives from dives to intermediate water depths, which reduces uncertainty in the bottom sounding data.
For information about surface currents and surface gravity wave statistics, we use consecutive GPS fixes and high-frequency IMU measurements. Compiling this environmental information requires using the IMU data to distinguish active bird behavior (e.g., flying and paddling) from drifting passively on the ocean surface.
Measuring well-resolved temperature and salinity profiles is theoretically straightforward with data from diving birds, although engineering challenges remain. These challenges include designing a sensor housing that produces temperature measurements with a short response time and developing a small conductivity sensor that produces stable measurements for the duration of tag deployments. We are working with tag manufacturers to iteratively develop and test tag and sensor prototypes to improve profile measurement capabilities.
Finally, contrasts in air and sea temperatures can theoretically be measured at the beginning of dives when birds first submerge and at the end of dives when they surface. Precisely measuring air temperature is more challenging than measuring water temperatures, however, so improving determination of these contrasts is a long-term goal of the project.
Outfitting Cormorants in the Columbia River Estuary
Insights into Marine Bird Ecology
Cormorants, which forage in biologically rich nearshore areas, can be used as indicators of ecosystem health. In particular, cormorants tend to follow boom-bust cycles that track the availability of the fish they eat. Yet the specific ecological role of many cormorant species is unclear.
Like other predators, cormorants are often viewed as being in direct competition with humans, and they are vilified, persecuted, or simply ignored. The animal movement data collected through the Cormorant Oceanography Project, in tandem with oceanographic data, provide important basic information on the birds’ foraging ecology, distributions, and migrations. This information, in turn, is valuable for efforts such as marine spatial planning, in which human activities are coordinated to balance demands for development with the need to protect the environment.
In summer 2019, we fit 22 Brandt’s cormorants (Phalacrocorax penicillatus) captured from roosting sites near the mouth of the Columbia River, on the Oregon-Washington border, with biologging tags using backpack-style harnesses. At about 40 grams, these tags weighed less than 3% of a cormorant’s body mass, minimizing effects on the birds’ normal activities [Fair et al., 2010].
Brandt’s cormorants are fish-eating foot-propelled pursuit divers—meaning they chase prey—and are endemic to the California Current, a coastal ocean current flowing between British Columbia and Baja California. We found that Brandt’s cormorants are generally loyal to their roosting sites and foraging areas. The Columbia River estuary was their core habitat during the summer, but individual birds moved both north and south. Thus, our tagged birds collected concentrated data near the mouth of the river as well as along much of the Pacific coast of North America (Figure 2a), diving as far as 79 meters below the sea surface. This study allowed us to try out various tag types and collect oceanographic data to use in an assimilative model within a well-studied and highly dynamic estuary system (Figure 2b).
Because of the birds’ autonomy in where and when they dive, the data they collect are heterogeneously distributed, making it difficult to interpret oceanographic information with analysis methods that require regular sampling intervals (e.g., averaging data over a long time at one location or performing a spectral analysis on a time series of data). Instead, we are applying techniques from inverse modeling and data assimilation and are using a numerical ocean model to fill gaps between data points and to infer ocean properties not directly observed.
The ability to estimate bathymetry from frequent, autonomous biologging measurements may have a practical utility for safe ship navigation and channel maintenance.
For example, we are inferring seafloor bathymetry from our biologging data. Coastal bathymetry is often poorly known, and it is always changing. The mouth of the Columbia River is continually being filled in with sand, which forms unpredictable shoals and channels [Stevens et al., 2020], and annual dredging operations remove at least 1.5 meters of sediment from navigational channels to keep them safe for commercial shipping. The ability to estimate bathymetry from frequent, autonomous biologging measurements thus may have a practical utility for safe ship navigation and channel maintenance.
Instead of trying to determine the shape of the seafloor directly from the scattered data coming from the cormorants, we apply data inversion. First, we consider various model seafloor profiles using a method developed by Evensen [2009], in which a sample of randomly generated candidate bathymetries is run through a numerical model to obtain a least squares–based statistical relationship between these bathymetries and the observational surface current data [Wilson et al., 2010]. Then this relationship can be inverted (back-calculated) to estimate the bathymetry that best fits the real data.
We are currently testing this inverse approach for use with distributed biologged measurements of surface currents. In the future, we may use similar techniques to determine parameters other than bathymetry, such as the strengths of cold-water currents that upwell from the depths of coastal oceans, for which we could make use of temperature data collected by the cormorants.
Upgrading Tag Technology
Although the tags we have used to date have proven effective, continued advances in tag attachment methods, targeted sampling (recording data when birds are foraging or resting on the sea surface), and battery miniaturization, as well as in tag solar panels, sensors, electronics, and communications, would help to optimize biologging devices for improved data collection and for use with different species.
Biologging tags should be as small as possible relative to the mass of the animals carrying them, and they should be positioned to have negligible impacts on the animals’ energy expenditure as they fly or dive. Whereas high-latitude cormorants, including Brandt’s, tend to have larger bodies (2.5 kilograms), tropical cormorants can be as small as 360 grams, necessitating further tag miniaturization.
Furthermore, although the use of cell phone technology allows tags to transmit large amounts of data, data transmission is possible only in locations with cell phone coverage. The tags also require occasional electronic updates to keep up with consumer-driven advances in cellular technologies (e.g., 5G).
Finally, considering the birds’ autonomy, the biologging data they collect pose challenges to coordinating near-real-time data processing, archiving, and distribution. Maintaining data provenance, including measures of uncertainty associated with behavioral biologging data (as distinct from uncertainties in data obtained by conductivity-temperature-depth instruments), requires flexibility that has yet to be built into many oceanographic data repositories.
A Work in Progress
Since 2019, we have collected more than half a million dive profiles from three species of cormorants foraging in a range of near-shore habitats (Figure 3): In addition to the Brandt’s cormorants from the Columbia River estuary, we have fit biologging tags to pelagic cormorants to study the water near Middleton Island, Alaska, and to Socotra cormorants in the Arabian Gulf off the United Arab Emirates. We are now evaluating these data and comparing them to numerical models.
Among other findings, these comparisons have revealed errors in models of temperatures for deep, cold ocean water that upwells to the surface, a common source of uncertainty in regional-scale ocean models on the U.S. West Coast. Future work will investigate other uncertainties about how coastal ocean environments function. For instance, we will use tag data from Socotra cormorants in the Arabian Gulf to diagnose sea surface temperature biases that occur in models and that are commonly associated with uncertain atmospheric forcing (e.g., the influence that dust storms exert on incoming shortwave radiation) [Lorenz et al., 2020].
Although we have made much progress in tag development, our near-term goals are to improve the response times of the temperature sensors we use, to test and improve our conductivity sensors, and to put smaller versions of these tags through trials. Over the next couple of years, we also aim to scale up our tag deployments through international collaborations and through the development of a global cormorant oceanography network. Furthermore, we are building an automated data pipeline through the Animal Telemetry Network to provide our biologging data to the oceanographic research community in near-real time.
With these efforts, we are continuing to expand the range of techniques and data that scientists have at their disposal to better understand highly dynamic—and highly important—coastal ocean environments.
Acknowledgments
The Cormorant Oceanography Project is sponsored by the U.S. Office of Naval Research. Dylan S. Winters (Oregon State University) processed the data for and produced Figures 1, 2, and 3. H. Tuba Özkan-Haller and Donald E. Lyons (Oregon State University), Reginald Beach (Office of Naval Research), and Christopher Wackerman (Naval Research Laboratory, Stennis, Miss.) provide project oversight and guidance. Sabir Bin Muzaffar (United Arab Emirates University) leads the tagging efforts for Socotra cormorants. The biologging tags we use were developed by Ornitela, Ornithology and Telemetry Applications, Vilnius, Lithuania. The 2014 data collection in the Columbia River was supported by Bird Research Northwest field crews, with special thanks owed to Daniel Roby, Yasuko Suzuki, Alexa Piggott, Peter Loschl, Kirsten Bixler, John Mulligan, and Anna Laws and with logistical assistance from Real Time Research. In 2019, efforts in the Columbia River were supported by Stephanie Loredo, Jason Piasecki, Emily Scott, Daniel Battaglia, Margaret Conley, Sam Stark, Tim Lawes, and Olivia Bailey (all at Oregon State University). Middleton Island data collection in 2020 was facilitated by Scott Hatch (Institute for Seabird Research and Conservation) and Jenna Schlener (McGill University), and field efforts were supported by Jillian Soller and Brendan Higgins (both at Oregon State University). Work with birds was approved by the Animal Care and Use Committee of Oregon State University, United Arab Emirates University, and the Office of Naval Research Bureau of Medicine and Surgery and by permits from the U.S. Geological Survey, Oregon Department of Fish and Wildlife, Washington Department of Fish and Wildlife, and Alaska Department of Fish and Game.
References
Biuw, M., et al. (2007), Variations in behavior and condition of a Southern Ocean top predator in relation to in situ oceanographic conditions, Proc. Natl. Acad. Sci. U. S. A., 104, 13,705–13,710, https://doi.org/10.1073/pnas.0701121104.
Evensen, G. (2009), Data Assimilation: The Ensemble Kalman Filter, 307 pp., Springer, Berlin, https://doi.org/10.1007/978-3-642-03711-5.
Fair, J., E. Paul, and J. Jones (Eds.) (2010), Guidelines to the Use of Wild Birds in Research, 3rd ed., Ornithological Counc., Washington, D.C., birdnet.org/info-for-ornithologists/guidelines-english-3rd-edition-2010/.
Harcourt, R., et al. (2019), Animal-borne telemetry: An integral component of the ocean observing toolkit, Front. Mar. Sci., 6, 1–21, https://doi.org/10.3389/fmars.2019.00326.
Lorenz, M., K. Klingbeil, and H. Burchard (2020), Numerical study of the exchange flow of the Persian Gulf using an extended total exchange flow analysis framework, J. Geophys. Res. Oceans, 125, e2019JC015527, https://doi.org/10.1029/2019JC015527.
Stevens, A. W., et al. (2020), Observations of coastal change and numerical modeling of sediment-transport pathways at the mouth of the Columbia River and its adjacent littoral cell, U.S. Geol. Surv. Open File Rep., 2020-1045, 82 pp., https://doi.org//10.3133/ofr20201045.
Wilson, G. W., H. T. Özkan-Haller, and R. A. Holman (2010), Data assimilation and bathymetric inversion in a two-dimensional horizontal surf zone model, J. Geophys. Res., 115, C12057, https://doi.org/10.1029/2010JC006286.
Author Information
Rachael A. Orben (rachael.orben@oregonstate.edu) and Adam G. Peck-Richardson, Department of Fisheries, Wildlife, and Conservation Sciences, Hatfield Marine Science Center, Oregon State University, Newport; and Greg Wilson, Dorukhan Ardağ, and James A. Lerczak, College of Earth, Ocean, and Atmospheric Sciences, Oregon State University, Corvallis
Citation:
Orben, R. A., A. G. Peck-Richardson, G. Wilson, D. Ardağ, and J. A. Lerczak (2021), Cormorants are helping characterize coastal ocean environments, Eos, 102, https://doi.org/10.1029/2021EO163427. Published on 23 September 2021.
This news article is included in our ENGAGE resource for educators seeking science news for their classroom lessons. Browse all ENGAGE articles, and share with your fellow educators how you integrated the article into an activity in the comments section below.
Text © 2021. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.