In years past, when polar explorers such as Captain James Cook or Sir John Franklin probed the edges of the icy sea seeking passage, they had little or no idea what conditions were like beyond the view of the lookout. Today, however, a ship’s pilot has access to detailed, up-to-date information on the frozen sea: where the ice lies, how old and how thick it is, and where leads (open channels) have opened. Knowing where a polynya (an open area of sea within the ice field) may be found is especially useful to a submarine commander. All this information comes from the National Ice Center (NIC) in Washington, D.C.
NIC generates this critical and timely information using sophisticated GIS applications and procedures, developed in-house to meet the unique challenges encountered in mapping sea ice around the poles. These applications employ current best practices, and, as software paradigms and capabilities advance, NIC keeps pace with emerging technologies.
The NIC, which serves maritime interests around the world, was established in 1995. It is an agency comprised of three components: the U.S. Navy (Naval Ice Center), the National Oceanic and Atmospheric Administration (NOAA), and the U.S. Coast Guard.
NIC provides worldwide operational sea ice analyses and forecasts of Arctic, Antarctic, Great Lakes and Chesapeake Bay ice conditions to support customers with varying interests. Its primary mission is to provide strategic, tactical and operational ice products and services to meet the requirements of U.S. national interests and U.S. government agencies. NIC products, available on the World Wide Web, are also used by many outside organizations to derive or interpret information of scientific value.
NIC regularly produces over 80 distinct ice information products on a daily, weekly and fortnightly schedule, in addition to supplying special support products to particular missions operating in icy seas.
A steady flow of raw ice data is critical to NIC operations; however, the remoteness of the polar regions limits the amount of direct observation of sea ice that is possible. For this reason, more than 95% of the data used in NIC sea ice analyses are derived from remote sensors on polar-orbiting satellites. These data are supplemented by reconnaissance, buoy reports, forecast model data and analysis information from other ice-tracking agencies.
In order to turn raw inputs into finished information products, a team of skilled and motivated ice analysts must select appropriate inputs and analyze, evaluate and interpret the information at hand in order to delineate and characterize ice conditions wherever sea ice is found.
In order to meet its formidable operational challenges, NIC has come to rely on GIS as a vital technology that facilitates achievement of mission goals. Seven days a week, NIC GIS systems ingest 80 gigabytes of satellite imagery and other raw data; process, prepare and deliver those data to the analyst; provide a desktop working environment for the analysis work; and finally, accept, store and disseminate the fruits of that analysis. NIC has invested considerable thought and effort into streamlining the entire continuous process.
The heart of the GIS solution is the Satellite Image Processing and Analysis System (SIPAS). This extension to ArcGIS is the primary interface between individual ice analysts and the wider NIC GIS system. Within SIPAS, the analyst selects a working area (from 38 Arctic or 13 Antarctic regions, or the Great Lakes or Chesapeake/Delaware Bay areas), selects a range of imagery appropriate to the area, and proceeds to delineate zones of ice concentration, stages of development (age), and ice forms (floe sizes), and assigns attributes to the created polygons. Map topology rules ensure that no splinter or overlapping polygons are created, and other checks ensure that areas delineated at the regional boundaries edge and attributes match polygons already created in neighboring regions.
The attribution of ice polygons allows a great deal of information to be preserved. As it happens, there are two standard formats endorsed by the World Meteorological Organization for these data. The first format is the Egg code, also known as MANICE, and is rooted in manual charting practice. The Egg name derives from the egg-shaped elliptical figure around which the various cipher values of the ice characterization are grouped. The primary figures are located inside the egg and make up four horizontal lines of ciphers. Alone on the top line is a value for the ice concentration of the entire polygon, given in tenths coverage. The second line shows up to three values, also in tenths, of concentrations in sub-areas of the polygon, ranged from left to right, from thickest to thinnest. The third line contains up to three values and indicates the stage of development for the ice concentration. Below this, the fourth line describes the form the ice takes. Other supplemental information can sometimes be placed outside the Egg ellipse. All values are shown as one- or two-figure numeric or numero-glyphic codes, using code domains particular to the code’s position in the Egg. For example, a stage of development code of 4· (four-dot) means first year ice more than 120 cm thick.
The other sea ice attribute coding system is Sea Ice Grid 3 (SIGRID-3), which is composed of a single alpha-numeric string with alphabetic attribute flags and numeric code values, for example, CT90CA3087-9CB2085-9CC3084-9CF-9-9CN-9CD81. The code group in our example flagged CA consists of the numeric pairs 30, 87, and -9. This tells us that the thickest ice in our polygon has a concentration of three-tenths, it is first year ice between 30 and 70 cm thick, and the ice form is not recorded. These values would correspond to the leftmost column of figures in the second, third and fourth lines of an ice Egg.
Because the full SIGRID-3 code is quite flexible and incorporates an extensive domain of optional variable identifiers, character flags are used rather than relying on position in the code string alone. NIC, however, generally does not make use of that additional descriptive potential, which can include codes to describe features such as melt forms or snow lying on the ice.
The Egg code is the system used by the analyst to describe ice conditions and is displayed on graphic products such as the regional ice charts. It is the SIGRID-3 code, however, that is entered and preserved in the geodata attribute tables. This practice requires the editing and map composition software to occasionally translate between the code systems, although handling these translations is relatively easy compared with the more formidable task of building a software tool to interactively set the codes in the first place. As an analyst moves through the attributing process, complex interrelationships between the indicators constrain choices later in the workflow, precluding some options and restricting others. The simplicity and elegance of the SIPAS Egg tool belies the intricate and complicated business logic under the hood.
SIPAS replaced an earlier editing workflow that involved multiple steps and stages and required users to bounce between different software packages to complete their work. The unified .Net-based ArcGIS extension application streamlines the workflow and allows analysts to focus their attention on the factor they know best: ice conditions.
The great majority of NIC ice information products are created using the SIPAS editor. This editing environment handles most of the data shuffling, processing and housekeeping operations for ice analysts, allowing them to concentrate on interpreting ice conditions. Core ice data products are generated out of the enterprise geodatabase archive that absorbs the main output from the analysis, while certain special products that require particular attention are generated directly through SIPAS modules by the analysts.
Other NIC products are generated with the assistance of custom utilities developed at NIC. For example, large Antarctic icebergs are tracked weekly using a custom utility called the Iceberg Database Edit Application (IDEA).