This data set represents the results of calculations carried out on the NOAA data and also contains the source code of the programs that made the calculations. The objective was to derive the average sea-surface temperature of each month and week throughout the whole 10-year series, meaning, for example, that data from January of each year would be averaged together. The result is 12 monthly and 52 weekly images for each of the oceanic regions. Averaging the images in this way tends to reduce the number of grid cells that lack valid data and to suppress interannual variability.
The images are provided in three formats: a modified form of run-length encoding (MRLE), Graphics Interchange Format (GIF), and Macintosh PICT format.
Also included in the data set are programs that can retrieve seasonal temperature profiles at user-specified locations and that can decompress the data files. These nongraphical SST retrieval programs are provided in versions for UNIX, MS-DOS, and Macintosh computers. Graphical browse utilities are included for users of UNIX with the X Window System, 80386- based PC's, and Macintosh computers.
The Bourne shell script do_avgs describes the manner in which the program average was used to calculate the monthly and composite monthly average sea ice concentration files. The program average is given a list of sea ice concentration data files in HDF format. For each grid cell in the images, it calculates the arithmetic average of the corresponding cell in the input files, and writes the average image into a new file. This program is run once for each month of each year represented in the original data to yield monthly averages. Then the composite monthly averages are calculated using the monthly averages for each year.
The Bourne shell script do_imgs converts the monthly average and composite monthly average files into GIF and PICT images. This script uses two programs from the freely available PbmPlus toolkit developed by Jeff Poskanzer.
The ETOPO5 global gridded elevation and bathymetry data set describes world topography on a grid whose cells measure 5 minutes of latitude by 5 minutes of longitude. Since the ice concentration data are given in an azimuthal equidistant projection, the elevation of cells in the ice concentration grid must be determined by appropriately sampling or interpolating the ETOPO5 data. Because the ETOPO5 grid is finer than the ice concentration grid, a sampling algorithm was combined with a spatial averaging scheme to derive the desired elevation data. The program that implements this algorithm is called topo.c; its strategy is to find the ice grid cell coordinates of the center of each ETOPO5 cell, and calculate the average depth of all of the ETOPO5 cells that fall within an ice grid cell. The results are two 16-bit images with big- endian byte ordering in which each signed 16-bit word describes the average depth or elevation in meters (sea level is zero) of a cell in the ice concentration grid.
Along with the 16-bit depth data files, displayable images are created using depthppm.c in portable pixmap format, and from that intermediate stage, GIF and PICT images are created using the PbmPlus toolkit. Depthppm reduces the complexity of the depth data by assigning one color to a range of depths. The relationship between depth ranges and colors is given in the source code file depthppm.c.
The Bourne shell script do_topo carries out these operations.