It was sometime around 1950 that a hydrographic survey ship last plied these waters taking water depth measurements along its path using a single-beam echo sounder and visual navigation. These data points were laboriously merged with shoreline and hazard information to create this chart, Alaska-Arctic Coast, Scott Pt. to Tangent Pt. Given the low ship traffic in the region, updating this chart was lower priority than other high-traffic areas. But things change—fishing and water-commuting traffic have risen in the area, as has marine tourism; but that’s not all: bottom depths have changed too as currents, erosion, and sediments have worked together to sculpt the seafloor.
In NOAA’s Office of Coast Survey, the Marine Chart Division is responsible for updating the suite of over 1000 nautical charts that keep mariners in U.S. waters safe. Their mandate covers all U.S. territorial waters in the U.S. Exclusive Economic Zone (EEZ), a combined area of 3.4 million square nautical miles that extends 200 nautical miles offshore from the nation’s coastline. The U.S. has the largest EEZ of all nations in the world, but it ranks behind 18 other nations in the number of vessels with hydrographic surveying capabilities. Their job is sizable and expensive. While the Army Corps of Engineers is responsible for maintaining shipping channel depths, providing bathymetry everywhere else in U.S. waters is NOAA’s duty.
Keeping waterways safe is a massive undertaking
The responsibilities of NOAA’s Marine Chart Division are immense. Charged with providing accurate charts for mariners, NOAA cartographers need to know when existing charts are out-of-date. To determine if charts are current, they employ lots of tools. They monitor navigation hazard reports submitted by mariners; they watch ship traffic patterns using vessel positioning information (via the Automatic Identification System); and more-and-more they are turning to satellite information, especially Landsat data.
The field of Satellite Derived Bathymetry (SDB), has been around for nearly a half-century now, but the advent of free Landsat data in 2008 together with the 2013 launch of the more-advanced Landsat 8 satellite and a shift in thinking about SDB products, have led to a reinvigorated use of satellite data in NOAA’s Marine Chart Division.
The concept of SDB is that different wavelengths of light penetrate water to differing degrees. The smaller the wavelengths (e.g. blue and green light) penetrate water more than longer-wavelengths (e.g. near infrared, shortwave infrared). When water is clear and the seafloor bottom is bright (sandy for example) estimates of depth can be made by modeling the depth of light penetration based on the amount of reflectance measured by the satellite. And when multiple visible-wavelength spectral bands are used together, the effects of seafloor reflectance variability and water turbidity are lessened. These modeled depth measurements typically do not meet hydrographic accuracy standards, so in the past SDB measurements were eschewed.
“There’s been a shift in the way we think,” Lieutenant Anthony Klemm, a NOAA Corps Officer in the Office of Coast Survey’s Marine Chart Division, explains, “In the past, if a measurement wasn’t made by the Army Corps or a NOAA survey ship, we didn’t want to use it, but now we are opening up to other technologies to evaluate the health of our current chart suite.”
Because of this sea change in thinking and faced with the daunting job of deciding which charts were most in need of updating, NOAA hydrographers revisited the use of SDB using freely available Landsat data as a viable tool to help them do their jobs.
“NOAA has now been using Landsat imagery for chart adequacy assessment and mission planning,” Shachak Pe’eri, a Research Professor at the Joint Hydrographic Center at the University of New Hampshire, says.
The Joint Hydrographic Center, a think-tank of researchers investigating technology and mapping challenges in NOAA’s Office of Coast Survey, realized that Landsat SDB could be an important reconnaissance tool. A single Landsat image is about 100 nautical miles across and affords a wide overview of a coastal area. Maps of SDB can be compared with existing nautical charts. Places where depth patterns do not match are more closely examined. Has the seafloor changed in this area? If an area looks shallower than what is presented in the chart and if there is a reasonable amount of vessel traffic or corroborating mariners’ reports in the area, the chart location is tagged as a higher-priority candidate for hydrographic mapping—i.e. sending out a hydrographic ship to make depth measurements using sonar (multi-beam or single-beam).
Multi-beam sonar provides very accurate and comprehensive bathymetry, but for the amount of water NOAA is responsible for charting, these expensive ships are in short supply.
Klemm has been out on hydrographic voyages, and knows well the amount of time and effort that goes into gathering bathymetry information. He is excited about the prospect of formally incorporating Landsat SDB into his workflow.
“SDB products to evaluate the current state of existing bathymetry representation is pretty amazing because of the temporal resolution of the satellite data—a little over every two weeks and you get a new shot of an area,” Klemm describes. Landsat 8’s orbit puts it back over a given location every sixteen days.
Because satellites like Landsat can provide “quantifiable information related to the amount of change since the last hydrographic survey,” as Pe’eri wrote, SDB information can figure prominently into the determination of where new hydrographic surveys are most needed.
Pe’eri and Klemm have been working on a NOAA policy about the use of SDB. They are outlining how to use SDB to prioritize hydrographic surveys using a chart adequacy assessment procedure they have developed. They are also working on a policy of how to update a chart with features found using satellite imagery.
“These charts are considered intermediary, but they can be made publicly available and used until a proper hydrographic survey can be performed,” Pe’eri explains.
Landsat is good at identifying new shoals, like that big 12 nm thumb-shaped shoal off of Alaska’s North Slope. And NOAA thinking is that it is better to amend charts to tell mariners that satellites indicated a shoal, even though exact depths cannot be provided until the next hydrographic survey.
Deriving bathymetry with Landsat for 43 years
Uncharted shoals have sunk many ships. In the late 1960s, research groups began to experiment with remote bathymetry using multispectral airborne data in an effort to make measurements over large tracts of coastal waters in search of navigational hazards and shifting bathymetry. With the launch of Landsat 1 in 1972, these newly developed methods could be used with data collected by the satellite’s Multispectral Scanner System and its wide 100 nm-wide images—satellite derived bathymetry was born.
In 1975, NASA teamed with famed oceanographer Jacques Cousteau to conduct an ocean bathymetry experiment using Landsat data to measure water depth in the Bahamas and off of Florida’s eastern coast. Cousteau’s ship, the Calypso, anchored over a study site as Landsats 1 and 2 collected data from overhead, while they simultaneously took depth measurements using the ship’s sonic depth finder. In this pre-GPS timeframe, LORAN-C radio measurements were used for locating the boat position. They also dove to the seafloor to take in situ reflectance measurements with a submarine photometer. This early experiment proved the feasibility of mapping shoals in clear water to depths equal to or greater than those needed for safe shipping.
The International Hydrographic Office, an inter-government organization concerned with making the seas navigable, had once classified shoals as navigational hazards between 0 and 17 meters (56 feet) below the surface, but with the advent of supertankers with drafts of over 20 meters (65 feet) and the capacity to carry massive amounts of oil, shoal definitions had to be broadened.
A year later, a Landsat 2 image acquired on March 29, 1976 revealed a major uncharted 8-km long reef in the Indian Ocean’s Chagos Archipelago: “There was a major reef or bank where the chart showed safe, deep water and some banks appeared to be out of position by more than 15 km relative to the nearest land,” wrote James Hammack, a participant in the NASA/Cousteau experiment and a cartographer with the Defense Mapping Agency’s Hydrographic Center (now part of the National Geospatial-Intelligence Agency). Within a few months, the newly found reef, named Colvocoresses Reef after the USGS cartographer who identified the feature on the Landsat image, was added to DMA nautical chart 61610. In the interim, Notice to Mariners were sent out to warn sailors in the region.
Based on the success of the NASA/Cousteau and Chagos Archipelago experiments, DMA requested that Landsat data be collected globally over coastal areas. This data was used to “augment the completeness” of its nautical chart products. DMA also used Landsat data to visually verify ship-reported navigational hazards.
Some other documented cases of Landsat data providing critical information to navigation include a safe deep passage through Papua New Guinea’s Star Reefs, which was first discovered using Landsat imagery. The Australian Royal Navy ship Flinders confirmed this passageway, which enabled ships to more quickly travel from Australian ports to East Asian ones. Likewise, British Admiralty Chart 322 of the Red Sea near Al Qunfidha had to be completely revised after it was compared with Landsat data.
In 2006, 75 shallow-water features such as reefs, shoals, and seamounts where discovered or found mislocated with the use of Landsat 7.
Landsat aids hydrographic offices around the world
The International Hydrographic Organization and the United Nations’ Intergovernmental Oceanographic Commission jointly create an authoritative, publicly available, global bathymetry map known as the General Bathymetric Chart of the Oceans, or GEBCO. GEBCO charts have been published since 1903. Despite this heritage, only about a tenth of the ocean floor has been mapped.
GEBCO is no stranger to SDB. They have been aware of its capabilities for decades. But now that Landsat data are publically and freely available it is getting more and more use—as no doubt the European Space Agency’s Sentinel-2A, with spectral bands similar to Landsat 8, will as well.
The GEBCO companion how-to guide for creating bathymetric charts, called The GEBCO Cookbook, includes a chapter on using Landsat to derive bathymetry. For cash-strapped national hydrographic offices, using free Landsat data to assess the adequacy of existing charts is essential, allowing them to allocate scarce resources with maximum impact to mariner safety. SDB alone does not meet IHO accuracy standards, but its use as a complimentary prioritization and planning tool is key.
SDB measurements can also “be used to infill regions in remote or inaccessible areas where no (or poor) bathymetry data exists,” shares Stephen Sagar, an Aquatic Remote Sensing Scientist with Australia’s National Earth and Marine Observation Group.
NOAA, as a major Landsat user, has been sponsoring international GEBCO students from around the world (Kenya, Sri Lanka, Ecuador, Philippines, etc.) and teaching them how to use SDB to update charts in their home offices. This summer from July 14–16, NOAA hosted a workshop to share this knowledge in the confidences that using SDB will make mariners worldwide more safe. Hydrographers from 11 countries attended. The workshop was a big success and more workshops are planned for 2016.
NOAA: thinking big about SDB
Water clarity has been a limiting factor when it comes to SDB. If waters are too turbid (full of sediments that obscure light reflectance from the seafloor), then bathymetric measurements cannot be made.
The inability of longer wavelengths, such as shortwave infrared light, to deeply penetrate water allows hydrographers to map shoreline change. But when concentrations of suspended sediments are great enough to thwart penetration by shorter wavelengths, SDB by definition suffers. But in NOAA’s Marine Chart Division, researchers are thinking outside of the SDB-box.
Pe’eri, in a collaborative study with NOAA and the U.S. Coast Guard, has pioneered turbidity mapping as a proxy for bathymetric measurements. In enclosed waterbodies with strong currents, such as bays and sounds, turbid channels show up on Landsat imagery—and these turbid channels illuminate where currents are carving deeper channels that are safe for boat passage.
Back in the arctic, where near-shore changes occur rapidly because of seasonal sedimentation and erosion, new SDB techniques like turbidity mapping are preventing maritime mishaps. In Bechevin Bay, where the easternmost passageway between the Gulf of Alaska and the Bering Sea provides fisherman with a shortcut for three ice-free months a year, the location of sand bars can shift significantly because of melting ice in this narrow passage. With the help of Landsat SDB turbidity maps, the new locations of these sandbars can be estimated. Recently this has led to the discovery of a new, straighter, and more geologically stable channel.
“SDB estimated from Landsat turbidity maps can help guide NOAA charting craft when they are mapping the channel each year and placing channel marking buoys. This saves time and it makes the process safer,” Pe’eri says. “With insufficient knowledge of sandbar locations, the NOAA craft risk running aground and crew can be thrown overboard when that happens.”
Pe’eri’s team has also developed a multi-image method to help separate clear and turbid waters using Landsat data. Techniques such as turbidity mapping will grow increasingly important for navigation planning as warming waters enable more industrial development of the Arctic and set the stage for international shipping routes.
NOAA’s Marine Chart Division has made Landsat a prominent tool in their charting toolbox—especially Landsat 8 with its new deep blue band, improved signal-to-noise and greater dynamic range (12-bit).
“Landsat 8 is overwhelmingly better,” Pe’eri says citing the new satellite’s additional cirrus band which helps him better account for atmospheric noise that can counter accurate SDB and Landsat 8’s better radiometric resolution (which means more signal, less noise, and more measurement fidelity). But it’s not just SDB that this innovative office is utilizing. They are also watching traffic patterns using the Automatic Identification System (AIS) and even light communication from recreational boaters, fishermen, tugboats, and larger vessels, and together with bathymetry measurements are prioritizing which charts are in perilous need of revision.
“We’re making charts safer up there,” Klemm says talking about the recent Beaufort Sea chart revisions, “and that’s so exciting.”
+ “What does the age of the survey mean for nautical charts?” NOAA Coast Survey blog post
+ “Reconnaissance surveying of Bechevin Bay, AK using SDB” presentation by Pe’eri et al. (via hypack.com)
Australian Department of Defense. 1987. Report of the Hydrographic Service Royal Australian Navy (for the year ended 30th June 1987).
Colvocoresses, Alden P. 1997. “Landsat Application To Nautical Charting.” Photogrammetric Engineering and Remote Sensing 63:868.
General Bathymetric Chart of the Oceans. 2015. IHO-IOC GEBCO Cook Book. Edited by IHO-IOC.
Hammack, James C. 1977. “Landsat Goes to Sea.” Photogrammetric Engineering and Remote Sensing 43 (6):683–691.
Harper, Brian, and George Sharman. 2003. “Leadline to Multibeam, Sextant to GPS & Crow Quill to Computer: Bathymetric Data Collection, Compilation, Archiving, and Distribution in The Past Century.” In Charting the Secret Floor of the Ocean Floor: The GEBCO Project 1903-2003.
Huntington, Henry P., Raychelle Daniel, Andrew Hartsig, Kevin Harun, Marilyn Heiman, Rosa Meehan, George Noongwook, Leslie Pearson, Melissa Prior-Parks, Martin Robards, and George Stetson. 2015. “Vessels, risks, and rules: Planning for safe shipping in Bering Strait.” Marine Policy 51 (0):119-127. doi: dx.doi.org/10.1016/j.marpol.2014.07.027
James, Elizabeth B. 1983. Use of Landsat for Navigation Products at DMA. Defense Mapping Agency Hydrographic/Topographic Center.
Joy, Richard Thomas. 1984. “An Assessment of the Potential Role of Multispectral Imagery in Bathymetric Charting.” Master’s Degree, Naval Postgraduate School.
Kilbride, A.R. Nautical Charting Applications of Photogrammetry and Remote Sensing Hydrographic Office, U.K.
Klemm, Anthony, Shachak Pe’eri, Ricardo Freire, John Nyberg, and Shepard Smith. 2015. “Nautical Chart Adequacy Evaluation using Publicly-Available Data.” US Hydro 2015, National Harbor, MD, March 16–19, 2015.
Pacheco, A., J. Horta, C. Loureiro, and Ó Ferreira. 2015. “Retrieval of nearshore bathymetry from Landsat 8 images: A tool for coastal monitoring in shallow waters.” Remote Sensing of Environment 159 (0):102-116.Doi:http://dx.doi.org/10.1016/j.rse.2014.12.004
Pe’eri, Shachak, Chukwama Azuike, Lee Alexander, Christopher Parrish, and Andrew Armstrong. 2012. “Beyond the Chart: the Use of Satellite Remote Sensing for Assessing Chart Adequacy and Completeness Information.” Canadian Hydrographic Conference 2012, Niagara Falls, Canada, May 15–17, 2012.
Pe’eri, Shachak, Chukwuma, Azuike, and Christopher Parrish. 2013. “Satellite-derived Bathymetry: A Reconnaissance Tool for Hydrography.” Hydro International: 16–19.
Pe’eri, Shachak, Patrick Keown, Leland Snyder, Michael Gonsalves, and John Nyberg. 2015. “Reconnaissance surveying of Bechevin Bay, AK using satellite-derived bathymetry.” US Hydro 2015, National Harbor, MD March 16–19, 2015.
Pe’eri, Shachak, Christopher Parrish, Chukwuma Azuike, Lee Alexander, and Andrew Armstrong. 2014. “Satellite Remote Sensing as a Reconnaissance Tool for Assessing Nautical Chart Adequacy and Completeness.” Marine Geodesy 37 (3):293-314. doi: 10.1080/01490419.2014.902880.
Pe’eri, Shahack, Shepard Smith, Leland Synder, and Brian Madore. 2014. “Satellite-Derived Bathymetry Using Multiple Images: The Alaskan North Slope Case Study.” Canadian Hydrographic Conference 2014, St. John’s, Newfoundland and Labrador, Apr. 14–17, 2014.
Polcyn, Fabian C. 1976. NASA/Cousteau Ocean Bathymetry Experiment: Remote Pe’eri, Shachak, Patrick Keown, Leland Snyder, Michael Gonsalves, and John Nyberg. 2015. “Reconnaissance surveying of Bechevin Bay, AK using satellite-derived bathymetry.” US Hydro 2015, National Harbor, MD March 16–19, 2015.Bathymetry Using High Gain Landsat Data. Greenbelt, MD: NASA.
Polcyn, Fabian C., and David R. Lyzenga. 1973. “Updating Coastal and Navigational Charts Using ERTS-1 Data.” Third Earth Resources Technology Satellite-1 Symposium, Washington, D.C., Dec. 10–14, 1973.
Stutz, Matthew L., and Orrin H. Pilkey. 2011. “Open-Ocean Barrier Islands: Global Influence of Climatic, Oceanographic, and Depositional Settings.” Journal of Coastal Research:207-222. doi: 10.2112/09-1190.1.