For more than 20 years, the Naval Postgraduate School’s home-grown, innovative Autonomous Ocean Flux Buoy has braved some of the Earth’s harshest environments in support of student research and scientific discovery.
In September of 2019, NPS Emeritus Research Professor of Oceanography Tim Stanton will travel to the Arctic as part of the largest expedition in human history to study drastic changes in sea ice throughout the region. While there, Stanton and his NPS colleague Bill Shaw will deploy four autonomous ocean flux buoys (AOFBs), joining more than 600 participants from 17 countries in the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) in “the first year-round expedition into the central Arctic exploring the Arctic climate system.”
While a mission like MOSAiC would be considered a once in a lifetime opportunity for any scientist or researcher, the journey is not the first of its kind for Stanton, who has been conducting scientific expeditions and research in the region for over three decades. In fact, the trip will bring his work back to the Arctic ice he dangled a prototype instrument pack through 22 years prior — a prototype that would evolve, over the next two decades, into the current design of the AOFBs that will be deployed as part of MOSAiC.
At its heart, Stanton’s AOFBs are the quintessential NPS story — a story of innovation, research and development occurring in answer to real-world necessity. It is a story of the continuous improvement and evolution of specially-designed tools, and the experimentation and knowledge those tools enabled, the student fieldwork and research they created and contributed to, the data they provided, and the myriad scientific and academic projects they supported. To track the AOFB through its many incarnations and uses is to see science in the making, and NPS functioning in accordance with its highest purposes and directives.
THE AUTONOMOUS OCEAN FLUX BUOY (AOFB): AN OVERVIEW
The AOFB is a system that was (and continues to be) custom designed and constructed by the Ocean Turbulence Group at NPS for the purpose of measuring turbulent fluxes in the upper ocean below sea ice. These measurements are key in determining processes controlling energy exchange between the ocean, ice cover and atmosphere via ever-increasing thinning and melt-out of the sea ice pack — information crucial to providing more accurate climate modeling.
The buoys answer an important need in Arctic research, as detailed by Stanton: “Our access to the Arctic, even in this century, is so limited by severe conditions … The AOFB was an inevitable and really important evolution from the whole concept of manned ice camps. These autonomous systems fill in a whole bunch of physical process questions that you just can’t really address any other way, because you need to sample multiple locations, hopefully concurrent and inter-seasonal.”
AOFBs are composed of two main components: a surface buoy and an attached instrument frame tailor-made to be suspended into the upper ocean. The surface buoy is home to the system’s processing electronics, an Iridium satellite modem, Global Positioning System (GPS) electronics, GPS and Iridium antennae and batteries. The instrument frame, which is attached to the surface buoy via a series of rigid poles, is equipped with a custom-built flux package and a downward looking 300 kilohertz Acoustic Doppler Current Profiler (ACDP).
When the Autonomous Ocean Flux Buoy is lowered beneath the ice shelf, a suite of sensors provide a comprehensive set of data points and calculations detailing the turbulent environment it resides it. Over the years, nearly 40 AOFBs have deployed in the Arctic and Antarctic regions, transmitting 22GB of data back to the university supporting faculty and student research. (Courtesy NPS Research Assistant Professor William Shaw)
The buoy’s flux package is comprised of a suite of sensors that includes an inductive conductivity cell and platinum resistance thermometer, along with a fast response thermistor. A custom designed acoustic travel time sensor measures the three velocity components down to 0.1mm/s—levels accurately resolving small turbulent motions in the mean flow. These sensors, located together within a sample volume Stanton describes as “about the size of a coffee cup,” use an eddy correlation technique to directly measure the turbulent fluxes of salt, heat, and momentum via burst-sampling over approximately 20-minute long Reynolds averaging periods. The buoy’s measurements and calculations are then transmitted back to NPS for analysis and storage.
The AOFB, however, was not birthed fully formed, and Stanton is the first to point out that its evolution has been both gradual and continuous in nature. However, he also acknowledges that there were milestones within that evolution where the AOFB took major steps forward in the march toward its current iteration, and its deployment in MOSAiC’s history-making expedition.
MARK 1: BIRTH OF THE PROTOTYPE FLUX PACKAGE
By 1997, Stanton and other Arctic researchers had amassed enough experience doing ice camp work to understand that how they conducted their research was going to have to significantly change if they had any hope of accurately studying and representing the physical processes at work in the dramatically changing Arctic waters and ice. Although Stanton could clearly see the need for an autonomous system capable of conducting longer-term process studies than the 4-6-week ice-camps that the Arctic climate allowed researchers, the technology necessary to design such a system did not yet exist.
Stanton used this lag to continue his research, keeping abreast of new developments in applicable technology even as he designed an AOFB that was not yet possible to build. He continued to lever his broad electronics design experience—a passion he had pursued since childhood. It wasn’t long before innovations in electronics reached a point that allowed him to develop a micro-powered system that could be integrated into the flux package that would become the heart of the AOFB.
Stanton procured the necessary electronics and built a rough prototype flux package with plans to test it during his participation as a researcher in the Surface Heat Budget of the Arctic Ocean (SHEBA) program, where he joined other scientists conducting studies on an icebreaker moored to an ice floe in the Beaufort Sea from October 1997 to October 1998. Designed as a coordinated effort to better elaborate how the interaction of clouds, ice cover, the atmosphere and the upper ocean determine the energy balance in the Arctic, SHEBA — a $20 million, yearlong, NSF-led experiment — offered Stanton the perfect testing ground for the instrument pack that he now considers the “Mark 1” prototype of the AOFB.
In between his duties as a SHEBA researcher, Stanton drilled a 14-inch hole through the ice and deployed his prototype through it, using a simple tripod to place the sensor one to two meters below the ice and stabilize it as best as possible. Stanton covered the hole with plywood, and set up a small pup tent next to it to protect the laptop that functioned as the sensor’s data logger and the car battery that powered the apparatus. Since the generator he had only lasted a few hours in the extreme Arctic climate, Stanton had little choice but to return to the site by snowmobile every 24 hours to change or charge the battery.
That instrument pack — lowered down a hole in the ice at the end of a melting runway that marked the scientists’ only exit from the ice floe they were studying — would become, with a few tweaks and the addition of a protective shell, the AOFB. It was, in his own words, “a very primitive, hybrid-y thing.”
Primitive though it may have been, it worked, and Stanton left SHEBA with several days’ worth of data sets to study. This data, when combined with knowledge gained during SHEBA and his many years’ experience conducting Arctic research in ice camps and on ice breakers, gave Stanton the information necessary to begin building the AOFB in earnest.
2002: THE BUOY COMES TOGETHER
In April 2002, after three years of development — during which Stanton applied for funding, conducted other research, and continued teaching — the AOFB was ready for the ultimate field test: deployment at the North Pole Environmental Observatory (NPEO). Stanton recalls the initial buoy was “technologically pretty intact,” in that the buoy’s infrastructure was “very close, in the first actual field effort” to what would eventually be its final form.
Initial buoys used commercial acoustic analog travel time sensors to send acoustic pulses between transducer pairs; when the pulses were received by one pair, the other transmitted a pulse back, and the reciprocal travel time was measured. The difference in the travel time represented current going along that path, and having three such paths allowed the flux package to resolve three components of velocity: one vertical and two horizontal.
Along with velocity, additional sensors within the flux package concurrently measured temperature and conductivity. Salinity was calculated via the temperature and conductivity measurements, and the fluctuations of this suite of measurements comprised the primary flux measurement.
“We correlate the vertical wobbles in the fluid going by with wobbles in temperature or salinity, using time series of these variables sampled four times per second” Stanton explained. “Then you simply correlate them together, and the mean of that is by definition the flux, or transport, … and the trick is to be able to resolve the very small turbulent fluctuations accurately, without a lot of sensor noise.”
In early iterations of the AOFB, these correlations were calculated within the buoy, which sent the data back to NPS for examination. With the buoy’s next big overhaul in 2010, however, buoys would be programmed to return not only the finished calculations to NPS, but the raw data, as well, which Stanton describes as “much more informative” for researchers working with AOFB data sets.
The amount of data generated by the buoys is not insignificant: Stanton estimates that AOFBs transfer about a megabyte of data per day to NPS through an Iridium modem that “calls home” to NPS twice a day. Aside from its use in sponsored research projects, this data has been instrumental to student thesis work over the nearly two decades that AOFBs have been deployed and returning data.
Stanton credits his work with NPS senior oceanographer/programmer Jim Stockel, and their development of protocols to make use of cutting-edge technology like Iridium satellite data transmission more robust, as key to the early success of the AOFB — especially when it came to the transfer of data. Stanton posits the robustness of these protocols as owing to the duo’s programming of block transfers and handshakes in their transfer protocols to overcome the frequent connection drop-offs during data transfer (drop-offs require the system to redial and reconnect). Expending the time and energy on developing these protocols resulted in a crucial success: the data from the early buoys was transmitted to NPS intact, without the gaps that often accompanied early Iridium satellite data transmissions (which made data very difficult to interpret).
The early buoys were not without their problems, though, and Stanton’s solution to one of these demonstrates NPS’ increasing role in the development and production of the AOFB. One such problem was Stanton’s use of original equipment manufacturing (OEM) sensors, which he would purchase separately and then integrate into the AOFB instrument package. While initially successful, over time it became evident that the practice was not only prohibitively expensive, but also introduced problems of drift and noise that limited resolution of the underlying turbulent fluxes.
This “drift” — essentially a low frequency change in sensors with time — interfered with the ability of the AOFB to gather accurate readings, and Stanton took it upon himself to redesign the sensors to address the issue. One by one, over a succession of about five years, he designed and implemented his own electronics, replacing those in the physical sensors he had ordered, thus correcting the problems they came with — addressing issues with drift and offering significantly improved noise floors.
This pattern continued over the years as Stanton took on an increasing amount of the electronic design and assembly for the buoys. The result is a creation intrinsically tied to NPS: presently, Stanton estimates that eighty percent of the AOFB is designed at NPS, and its reach extends far beyond this, as the buoys continue to be deployed in projects that range from the marginal ice zones and ice floes of the Arctic to the ocean boundary layers of Antarctic glacier ice shelves.
Watch then Research Professor Tim Stanton discuss his team’s history-making project at the Pine Island Glacier in this 2013 edition of Inside NPS.
2010 CHANGES: “SORT OF A MID-LIFE THINK THROUGH”
By 2010, Research Associate Professor Bill Shaw had transitioned from a post-doc position to a co-researcher in the Ocean Turbulence Group, taking on many aspects of analysis and logistics for the group’s projects. The timing was fortuitous, as Stanton was ready to implement another set of changes to the AOFB as a result of what he terms “sort of a mid-life think through” for the buoys. This “think through” was spurred by both changes in the overall nature of the research the AOFB was being engaged to complete and the availability of new technology that would allow the buoy to more accurately engage in these scientific processes.
Perhaps one of the most dramatic of these changes was spurred by Stanton’s desire to “take advantage of sampling the acoustic signals straight away and do all of the processing chain as a digital function.” Advances in radio electronics for cell phones also opened opportunities to improve the AOFB, as innovations in digital radio technologies became available in chipsets small and power-efficient enough for use in the buoys.
This meant Stanton could make the move to digital down converters capable of measuring incoming signals, digitizing them at 20 million times per second, and, as he describes, “squirting them out into another processing chip that sifts the frequency down to baseband.” The resulting increase in sensitivity allows for a much more stable and lower noise measurement, and enables the resolution of fractions of a degree of phase at the acoustic frequency in these reciprocal travel times, thus increasing the accuracy of the AOFB’s data sets.
Additional changes to the buoy include a move from buoy hulls recycled from abandoned thermistor string buoys scavenged from the NPS oceanography department to the custom-designed instrument frame/buoy combination currently in use. Stanton also installed very high resolution lower-powered analog to digital converters to replace OEM temperature sensors and an inertial measurement unit (IMU) to account for tilt and package motion in the correlation processing, after which Stockel stepped in to complete the software overhauls necessary when any aspect of the AOFB is changed.
These changes enable the buoy to complete its measurements with more precision, and via very low noise electronics. Stanton is quick to point out that the lower the velocity noise, the better resolved the correlations that represent the fluxes measured by the AOFB, and he notes that the 2010 upgrades resulted in “more than an order of magnitude better noise performance” due to the use of the direct sampling system.
THE CONTINUING QUEST FOR POWER
With each set of changes, Stanton is forced to revisit one of the buoy’s most enduring challenges: providing AOFBs with a strong enough power supply to weather their expected one to two-year deployment. Stanton freely admits that, when it comes to the AOFB, he’s spent much of his time “chasing power” — a pursuit that has led he and his colleagues to engage in creative solutions when it comes to extending the lifespan of the buoys.
While all of the on-board electronics are micro-powered, an autonomous system designed to take multiple readings, record and process data, and engage in two-way communications multiple times daily for a two-year period requires a tremendous amount of power. In order to address this, each buoy is outfitted with large lithium chloride primary cells as its primary power, which are supplemented in the summer with the use of solar cells.
The addition of these solar cells points to another evolution of the AOFB, in this case inspired by changes in research focus (from the seasonal focus of early Arctic research to researchers’ more recent acknowledgment of how many details of the summer melt are, as Stanton puts it, “very important in determining the extent of the ice pack.” He credits this realization as key to a shift in Arctic research, noting “Our whole emphasis towards the upper part of the ocean water column and its structure and the way it’s trapping heat became a scientific issue that we investigated more and more.”
As an additional power saving feature, the buoy’s controller allows sampling setups to be changed remotely with each 12-hour call-in to the Iridium server at NPS. This allows Stanton to keep his power usage down, an important factor in an instrument designed to operate remotely for years. Turbulence is sampled in 40-minute increments that can be scheduled at 1 to 8-hour intervals in response to the available power.
MOSAiC will also bring Stanton an opportunity to revisit an earlier attempt to add wind power to the AOFB’s arsenal: while past attempts were hampered by the loss of equipment to harsh Arctic gusts, Stanton is currently exploring ways to provide a more sustainable solution for the buoy’s upcoming deployment.
CONTINUING COLLABORATIONS: MOSAiC AND BEYOND
Other MOSAiC changes will see the deployment of four AOFBs recently modified via the addition of a mechanical track that allows for the movement of the flux package in order to study the upper three meters of the water column below the ice, a CTD (conductivity, temperature, and depth) device, and an optics package designed to measure turbidity and upwelling / downwelling solar radiation in the water column. The optical measurements include chlorophyll and dissolved organic matter to study the role of increasing biological activity in the late summer to increased upper ocean heat trapping.
Stanton is quick to point to the importance of collaboration when it comes to the development, evolution, and use of the AOFB over the years. A recent example of this is the addition of the mechanical track, the importance of which was illustrated by research conducted during the 2014 Marginal Ice Zone (MIZ) experiment and later published by Stanton and former NPS PhD student Shawn Gallaher, along with researchers from three other institutions.
MIZ revealed that having the flux package fixed in two and a half to three meters meant that researchers were missing the fact that there were changes in the way that turbulence is injected into the ocean during the summer months. Since the mapping of small-scale surface heat trapping processes are not accounted for in current climate models, any modification that will help capture these will move MOSAiC that much closer to its goal of providing parametrizations of sub-grid scale processes that will increase the accuracy of coupled regional and climate-scale models. Our current models significantly under-predict Arctic summer ice retreat compared with observations.
MOSAiC is far from the only collaboration Stanton and the AOFB have engaged in. Since 2005, all AOFBs have been deployed concurrently with two other pieces of profiling instrumentation in a cooperative effort that includes the complete sharing of all data produced: The Ice-Tethered Profiler (ITP; out of Woods Hole Oceanographic Institute, or WHOI), and the Ice Mass Balance Buoy (IMB; courtesy of the US Army Corps of Engineers).
Having this trio of complimentary measurements available from so many locations greatly increases their utility in understanding the coupled ocean-ice-atmosphere system. Stanton stresses that “having several concurrent measurements at different points around the Arctic, year in and year out, through different programs is of enormous value for understanding this rapidly evolving coupled system.”
Over the years, AOFBs have been deployed as part of funded projects as diverse as the ONR funded MIZ, the Stratified Ocean Dynamics in the Arctic Research Initiative (SODA), as well as several NSF funded projects. This funding has not only helped to advance the evolution and widespread use of the AOFB in Arctic and Antarctic research; it has also supported research opportunities for NPS students like Gallaher.
While Stanton is passionate about getting students out into the field whenever possible, the high logistic cost and attendance restrictions of Arctic research often hamper his ability to do so. In the case of MIZ and SODA, however, Stanton was able to take three NPS students into the field. He was also able to have an additional four pairs in the field to conduct experiments with the AOFB during the Navy’s Ice Exercises (ICEX), which take place about one hundred miles north of the Alaska coast every two years, and offer scientists like Stanton and his colleagues an excellent testing ground for their latest innovations.
Participating in every aspect of the process has paid off for Stanton’s students, with many of the students who have worked on various aspects of the AOFB continuing to work in oceanography, meteorology or applied physics after graduating NPS.
“It’s very gratifying that most of our students do go and use these problem-solving experiences,” Stanton said. “They go to [work in] the forecasting centers, including the National Ice Center, out in the fleet, and a good percentage of them go on up the ladder and [end up] at senior levels.”