This README_DRG.txt is the same for all zipped DRG files created at about this time at this site. DRG (Digital Raster Graphic) files are scanned georeferenced USGS topo maps. Details at http://gis.ess.washington.edu/data/raster/drg/docs/usgsinfo.txt (Note: For historical reasons, the ess (formerly known as geology) web site was named duff. We are now calling it gis, though the name duff.ess (but not duff.geology) will still work. The original USGS product covers the entire mapsheet, including information in the margins. The files are geotiff files, meaning that they contain a tag that describes the projection (UTM, NAD27) and the transformation from pixels to UTM coordinates. The transformation is redundantly defined by a six-line .tfw file. Different software will see this differently. Photoshop will ignore the geotiff tag, and will not copy it out after you edit an image. Workstation arcinfo will read the geometry part of the internal tag, but will allow the .tfw file to override it. ArcGIS and ENVI will read the entire geotiff tag. The UTM coordinate system is essentially 60 zones in each (N&S) hemisphere. Everything from 126 degrees west to 120 degrees west (the Pacific to Chelan) is UTM zone 10, while everything from Chelan east to Missoula, Montana is in zone 11. Coordinates are further complicated by the spheroid and datum. The paper maps are based on the North American Datum of 1927, which incorporates the the Clarke 1866 spheroid. Many users prefer data in NAD83, or in HPGN/HARN, which is a subtle improvement over NAD83. In the Seattle 1:250K quad, difference between NAD83 and HPGN/HARN range from about 10 centimeters to 40 centimeters, far less than the map accuracy or cell size. Because NAD83 uses the GRS 1980 spheroid (assuming a slightly different shape of the earth), a given point on the earth has a different latitude in NAD27 and NAD83. This fact--and the fact that all the little brass survey markers have been resurveyed--results in different UTM coordinates depending on the datum. Although a precise transformation from one datum to another is extremely complicated, it can be closely (within centimeters) approximated simply by redefining the corner of the map. For workstation arcinfo, you can shift an image from NAD27 to NAD83 by replacing the .tfw file with the tfw83 file. However, this will just confuse ArcGIS. A better solution is to use the geotifcp.exe program (See http://gis.ess.washington.edu/data/raster/geotiff/ or http://geotiff.osgeo.org/) to create a duplicate tif file with an alternate tag. For example, enter the following at the DOS command line: geotifcp.exe -g o47122e6_harn_geotag.txt o47122e6.tif o47122e6_harn.tif (Note: this example wrong wrong in the 2004 version.) Non-DOS/windows users can download the file from remotesensing.org and build the executable for your operating system, though I had trouble with libraries when I did this. So, what did we do here? First, we took the standard DRG files (still available at http://gis.ess.washington.edu/data/raster/drg/index.html) and clipped the images to the quadrangle boundaries. The resultant images were not perfect rectangles, so they were mosaicked and reclipped to the smallest rectangle that included the quadrangle. The result was slightly overlapping rectangles that could be combined into a seamless map. The file that you got with this README file is a proper geotiff file with an internal projection and positional information. This paragraph applies to the 7.5' quads (beginning with "o") and the 7.5' x 15' 1:25000 quads (beginning with "k") which substitute for the 1:24000 quads in the Seattle area. The 1 x 2-degree quads, based on 1:250,000-scale maps were done in a similar manner. Each pixel represents 25.4 meters. These files have packbits compression, with four bits per pixel. You can use geotifcp to achieve higher compression at the cost of total portability. Zipped files would be about the same size regardless of the tiff file's internal compression. The tfw file for Okanogan has been fixed. The internal positioning information had always been correct. The c46124a1 map is two degrees high, covering the sliver of coast on both the Cape Disappointment and Copalis Beach quads. Most of the tiles began with the original USGS tiff files, but we got the quads south of 46 degrees (mostly in Oregon) from http://www.ncgc.nrcs.usda.gov/products/datasets.html Their DRG files apparently are shifted by resampling, and stop at UTM zone boundaries, but they cover the whole country, and are accompanied with the fgdc metadata files that all responsible GIS people produce. This paragraph applies to the 1 x 2-degree quads whose names begin with "c". The 30' x 60' quads, based on 1:100,000-scale maps, were treated in a similar manner, but the colors were more of a problem. The original scanned maps have many colors which do not match from map to map, and which are often quite bad. We performed digital gymnastics to build common color tables so that we could combine adjacent maps, and we fixed some of the most egregious color problems. The results are not so pretty as the 7.5' maps. They are all stored as 8-bit files. We developed a new product: 1:24000-scale maps mosaicked in one-degree blocks. These filenames begin with "z". These files are huge, 400-600 megabytes each. Tell us if you find such large files to be useful. Notes on compression: The 7.5' quads have 13 colors, and could have been stored a 4-bit files by using the "-d" switch in geotiffcp. This was not done for the following reasons: 1) listgeocp -d maps color 255 (white) to 15 (black). I should have taken care of this two steps back. 2) On clean files (i.e. files with contours that are brown, not brown with black and red spots), 4-bit files are mysteriously larger. 3) When files are zipped, they are the same size anyway. I was tempted to use zlib compression, but not all software can read such files. (The previous sentence was written a long time ago.) SO EACH ZIP FILE CONTAINS: 1) This file 2) a geotif image of a map, readable by lots of software. Some software will recognize the map projection and coordinates. Other software will complain that it cannot read a TIFF tag, but will otherwise read the image. 3) a .tfw file, describing the transformation from row/column to UTM NAD27 coordinates. 4) a .tfw83 file. If this file is renamed to .tfw, SOME software will display the image in NAD83 coordinates. 5) a .tfwharn file. This is the same as the .tfw83 file, but with small adjustments. 6) a _geotag.txt file. This is the ASCII equivalent of the projection information which is encoded in the geotiff file. 7) a _harn_geotag.txt file. If this file is used with the geotifcp program, you can make a NAD83/HARN copy of this file. 8) a _nad83_geotag.txt file. If this file is used with the geotifcp program, you can make a NAD83 copy of this file. 9*) Some of these zipfiles contain .fgd metadata files from the National Resource Conservation Service. Harvey Greenberg hgreen@uw.edu Wed Nov 23 13:50:27 PST 2005 (Small changes 2018)