The data in the provided files are the ETOPO5 gridded elevation data for the Earth, subsetted for the Arctic area above 50 degrees North. The data were provided to NCAR for this project by Michael S. Loughridge and George F. Sharman at the National Geophysical Data Center (NGDC), Boulder, Colorado. For further information about the ETOPO5 data set, and for access to the full data set and other subsetting and data formatting options, please see: http://www.ngdc.noaa.gov/mgg/global/seltopo.html The elevation data in the files whole_arctic.bin, east_arctic.bin, and west_arctic.bin are 2-byte, signed, big-endian integer values representing meters above sea level (i.e. negative values are below sea level). They are 1-band image files with no header. The data are also available as ArcInfo (vs. 8.0.1) grids, which have been tarred as directories to include the requisite info directory, or as ASCII files in row-major order (with ArcInfo headers, which may be stripped if unnecessary for your needs). The following table gives the information and specifications necessary to display and use these data. These files should be easily read by the image processing system of your choice. File Columns Rows Upper Left Lower Right Name Coordinate Coordinate --------------------------------------------------------------- whole_arctic 4321 481 90 degrees N 50 degrees N 0 degrees E 360 degrees E west_arctic 2161 481 90 degrees N 50 degrees N 180 degrees W 0 degrees E east_arctic 2161 481 90 degrees N 50 degrees N 0 degrees E 180 degrees E The following is a discussion of the nature and limitations of the gridded topography data contained in the ETOPO5 data set north of 50 degrees North, some educated discussion (guesses) as to the source and processing of the data, and the inferences to be drawn from the observable artifacts in the data set: ETOPO5 is actually a collection of several of gridded elevation data sets, combined into a single, global coverage. All of the submarine topography is derived from the UNCLASSIFIED Navy data set known as "Digital Bathymetric Data Base at 5 minute resolution" (DBDB5). North of 50 degrees North, the land areas are derived from two data sources: the U.S. Navy Fleet Numerical Oceanography Center (FNOC) global 10-minute data set of modal elevations, and the Defense Mapping Agency(DMA) 5-minute gridded elevation data for North America, Europe, and parts of Asia. Little is formally documented about the processing that went into constructing these data sets, although observable artifacts and knowledge of techniques that would result in such artifacts can partially complete the picture. All of the submarine data were derived from a profile-fitting program applied to digitized contours from Naval Oceanographic Office bathymetric maps and other sources. Clustering of depth values about even depth values of 1000, 500, and 100 meters suggests profile fitting to digitized contours. Overshoot at sharp changes in slope, such as the continental shelf-slope break, likewise suggests curve- fitting with an incorrect tension component of the fitting program. Many continental slopes show striations down the slope gradient as would be generated by second order fitting to profiles drawn perpendicular to the contours of the slope. And finally, north of 74 degrees North latitude, the data appear unnaturally smooth, in sharp contrast to the submarine topography further south. This would indicate a significant change in the tension component of the fitting algorithm, or perhaps even a change in algorithm altogether. In a visual presentation of the data, as in "Surface of the Earth," World Data Center-A for Marine Geology and Geophysics Data Report MGG-5, this gives the appearance of very smooth, or "out of focus," topography north of 74 degrees North. Much of this work was done in the early 1970s and thus represents a relatively primitive understanding of the sea floor in remote regions, such as the Arctic. Land processing is no more straightforward. The FNOC 10-minute global gridded data set was originally created in the 1960s and 1970s by visually extracting elevations from small-scale 1:1,000,000 and 1:2,000,000 scale aeronautical and jet naviga- tional charts. Values were estimated for 10 x 10 minute cells of latitude and longitude in terms of maximum, minimum, and modal elevations, all rounded to the nearest 100 feet. Data were hand encoded and later scanned by optical character reader for digital storage. Though expressed here in meters, the artifacts of quantizing at 100 foot intervals and clustering at 1000 foot (contour) intervals are evident in the data. Additionally, quantizing tends to produce a checkerboard pattern of topography in regions of low relief or poorly mapped terrain. The DMA 5-minute grids have little or no documentation of source and processing, although they are believed to be derivatives of more detailed data developed by the former U.S. Army Map Service. The quality of these data appears to be significantly better than the FNOC 10-minute data, especially in well mapped regions such as the conterminous United States or in Europe, although there remains evidence of clustering of values about contour intervals. The inference to be drawn from these observations is that these data, the best at hand, are relatively poor. Outside of Canada and Europe, the land data have a vertical resolution of 30 m (the 100 ft estimation process) or more, and horizontal resolutions that frequently degrade to as much as 1 degree, producing a checkerboard pattern of topography. Likewise, Arctic submarine features portrayed in this data set are poorly resolved and are portrayed only as accurately as was known in the UNCLASSIFIED world by the early 1970s. Resolution is defined here as the distance of separation between two features at which they become detectable as two and not a single feature. Additionally, the data north of 74 degrees North appear to have been subjected to a long wavelength (~1/2) smoothing filter which gives the data a smooth character. The conclusions to be drawn are that these polar topography data sets are of low resolution and accuracy, and useful for only the most general descriptions of polar topography. Vertical resolution usually worse than 30 meters and horizontal resolution is no better than 10 minutes and frequently as poor as a full degree. They are based on published maps that are now almost 20 years old, which in turn were based on earlier observations. The authors of the sea floor maps almost certainly did not avail themselves of the paradigm of plate tectonics to guide their contouring of the very sparse data in the Arctic basin and thus the topography will lack the linear and orthogonal character that might be inferred from the same data today. However, for the moment, these are the best at hand. CONTACT INFORMATION: Further questions and requests for information about this data set may by forwarded to George Sharman, National Geophysical Data Center/NOAA, Code E/GC3, 325 Broadway, Boulder, Colorado, 80303- 3328. Phone: (303) 497-6345. FAX: (303) 497-6513. Internet: gfs@mail.ngdc.noaa.gov.