Imaging the Oceans and Their Coasts
NASA helps expand specialized remote sensing applications
By Jan Svejkovsky, Ph.D.

Ever since Captain Ahab searched the world's seas for the Great White Whale, viewing ocean regions from space offers obvious advantages. The oceans' vast size and often inhospitable environment make direct field sampling very difficult and expensive. For that reason, the oceanography research community embraced satellite remote sensing in the 70s, and its use became commonplace in many research programs during the late 80s and 90s. Satellite imagery is most commonly used to supplement field studies by obtaining a large-scale synoptic view of surface temperature and plankton pigment patterns, wave distribution, surface height, and other variables. A very powerful technique also involves the analysis of image time series from instruments with high revisit frequencies such as the AVHRR, CZCS, and its successor, SeaWiFS. The time series help extend field sampling efforts in both time and space and allow the study of environmental changes on various time and space scales-from days to decades.

A tropical reef area near Saint Thomas, Virgin Islands, as imaged with Ocean Imaging's DMSV aerial sensor at 1m resolution. The 4 channel instrument allows accurate classification of depth and bottom type in support of shallow water surveying.  

   The state of constant, rapid change inherent in the marine environment also poses unique challenges for its remote sensing. Currents and winds can displace water masses at speeds of up to 10 miles per hour. Plankton blooms can change large areas from crystal clear to pea soup in 1 or 2 days. Such dynamics impose high requirements on a useful marine remote sensing platform: the orbit and image swath should allow very high revisit frequency-preferably daily. Data delivery to the end-user should be in real-time or close to it.
    Then, there is the problem of resolution and spatial scale. Many ocean phenomena are extremely large, thus requiring very wide data swaths to image comprehensively. On present-day instruments, a wide scan swath comes at the expense of spatial resolution. While the 1 kilometer resolution of AVHRR and SeaWiFS may be adequate for large scale ocean studies, it is too coarse for monitoring many localized coastal processes.
    Finally, atmospheric attenuation poses a much larger problem for visible and infrared ocean images than it does for images of terrestrial targets. Thousands of multispectral images from Landsat, Spot, and IRS satellites are successfully being used for terrestrial applications without the need for atmospheric corrections. However, since the brightness of the ocean is generally much less than the land, atmospheric effects can drastically obscure vital information or cause unacceptable error. For example, up to 75% of the signal received by the SeaWiFS sensor is due to reflections from molecules and aerosols in the atmosphere, which must be removed to enhance the 25% of the signal actually originating at or below the ocean surface. Successful atmospheric corrections require specific wavelength bands, accurate on-board calibration systems, and an overall high signal-to-noise ratio.
    Many of the above constraints have not been adequately met by past and present satellite sensors. It is for this reason that the progression of marine remote sensing from the research arena to routine operational and commercial use has been much slower than for terrestrial applications. As the result, many as-yet untapped uses of remote sensing exist in the marine and coastal regions that will especially benefit from instruments aboard NASA's new satellites, as well as soon-to-be-launched commercial sensors.
    Ocean Imaging Corporation is one of a handful of companies worldwide that specialize in the application of remote sensing technology to the marine markets. In addition to commercial services, the company provides satellite data acquisition and processing support for numerous academic research groups and is co-investigator on research grants from the National Science Foundation and the U.S. Navy. With assistance from NASA's Commercial Remote Sensing Program (CRSP), Ocean Imaging is also working on several remote sensing techniques applicable to marine and coastal regions that can be made operational or be commercialized. Most will utilize the enhanced capabilities of the upcoming sensors and improved data distribution infrastructures such as EOSDIS.
    To rise to the challenge of remotely sensing the marine environment, and to satisfy the needs of their clients, Ocean Imaging emphasizes a multi-sensor approach for most monitoring applications. A good example is Ocean Imaging's work with coastal sanitation districts and water quality authorities to integrate remote sensing technology into monitoring of sewage outfalls and storm drains. In early work, Ocean Imaging found that surfacing sewage and runoff plumes can be discerned in specially processed SeaWiFS, AVHRR imagery, and, when available, Landsat TM and Spot visible band imagery. This data could be combined with surface current estimates derived from tracking thermal pattern displacements in AVHRR infrared image series. Together, the information can be used by local authorities to monitor the surfacing of the effluent, estimate its movement, and plan appropriate action to reduce public health risks. The increased spatial resolution in some of NASA's MODIS and MISR bands, coupled with more frequent revisit coverage than is presently possible with Landsat, Spot, and IRS, will greatly aid this application. Future high resolution commercial sensors (e.g., IKONOS) will also increase monitoring efficiency.
    Sewage effluent is not only the nose-wrenching stuff we usually think of. One of its most problematic components is cooking grease from restaurants and homes that was improperly disposed of down the drain. Under abnormal circumstances, some of it remains in the effluent let out to sea. Similarly, in regions such as southern California where it seldom rains, motor oil and other pollutants accumulate on roadways and are washed out to sea through runoff drains following a storm. In anticipation of NASA's LightSAR satellite, Ocean Imaging is investigating ways to add SAR data to the effluent plume monitoring, which could discern ocean surface areas made smooth by grease, oil and other surfactants. SAR's ability to image through the clouds could be especially complementary, since sewage and runoff problems are usually greatest immediately following rain storms, when cloud cover prevents the collection of useful optical and thermal imagery.

AVHRR data showing a turbidity plume over the Orange County Sanitation District outfall. Coliform bacteria measurements made by the district in response to thisimage are shown. Counts over 20 MPN are considered a health hazard.   

   The unique imaging capabilities planned for LightSAR-data collection in multiple polarizations-will be especially useful for developing new marine and coastal applications. Since the wave-caused "roughness" of the ocean surface is recorded most in vertically polarized data, multiple polarizations could be used to more clearly identify non-ocean targets. This capability can greatly increase satellite sensing's usefulness for marine law enforcement applications-detecting vessels, illegal fishing gear (e.g., high seas drift nets), or the illegal dumping of oil sludge and garbage.
    The combined use of optical, thermal, and SAR sensor data is also proving useful in studies of coastal circulation and its effects on biological processes. Under funding from NSF's GLOBEC and the Navy's National Oceanographic Partnership Program, Ocean Imaging is developing techniques to merge the different data types into time series that would allow fine-scale monitoring of near-shore currents under all weather conditions. In cooperation with scientists from Oregon State University, these efforts will aid development of circulation prediction models on much finer scales than previously possible. The results will also be applied to investigations of how localized physical forcing and its variability affect the distributions of fish, invertebrate larvae, and other organisms.
    Fisheries have especially benefited from marine remote sensing technology. Ocean Imaging maintains several subscription services that provide commercial and sport fishing fleets with satellite-derived fishing condition advisories in near-real-time. The company is also active in fisheries research, studying the effects of environmental variability on fish distribution and migration patterns. Water temperature and current patterns are the most valuable satellite-sensed variables, with ocean color also being useful. The multi- sensor approach has again proved to be the most effective. Temperature is derived from AVHRR, color or turbidity information is obtained from SeaWiFS and AVHRR. Currents are computed from sequential AVHRR images, but for large offshore areas TOPEX altimeter data is used to derive large-scale flow patterns. The high number of spectral bands available on MODIS will provide new benefits for fisheries research, since they will allow better separation of sediment and plankton types-parameters important for studies of habitat and fish larvae survival.
    Historically, the most widespread use of satellite oceanography has been over large, off-shore areas. From an applications perspective, however, the near-shore and coastal zones provide the most growth potential. The accurate mapping of shoreline and its change due to erosion readily suit themselves to remote sensing. The mapping of shallow water bathymetry and bottom substrate classification is another application that Ocean Imaging is working on with NASA's CRSP assistance. Submarine substrate classification is made difficult by the water column's wavelength-dependent attenuation of the substrate's reflectance. A coral reef or eel grass bed renders a different multispectral signature when 15 feet deep than at 5 feet, and a different one still when exposed at low tide.
    Experimental data from hyperspectral airborne instruments such as NASA's AVIRIS aid the development of algorithms to adjust for the water attenuation effects. The results can be applied to derive useful habitat maps with only a limited number of bands usually available on other aerial sensors and future high resolution satellites.
    The next few years promise to be an exciting time in satellite oceanography. The availability of new data types and development of improved algorithms will continue to increase the utilization and value of remote sensing in oceanographic research. Concurrently, increasing commercial application of remote sensing for routine coastal mapping and monitoring will introduce novel "eye-in-the-sky" technology to a multitude of new users.

About the Author:
Dr. Jan Svejkovsky is the president and founder of Ocean Imaging in Salona Beach, California.

Back