THE MAMMOTH SITE OF HOT SPRINGS SD, THE MAPPING OF A LONG TERM EXCAVATION SITE

Discovered in 1974, the Mammoth Site of Hot Springs, South Dakota is a Late Pleistocene faunal trap that claimed the lives of at least 58 Columbian and Wooly mammoths, and preserved their remains along with those of 83 other plant and animal taxa. Careful mapping of this site over the last 36 years has provided researchers with a wealth of information curation of specimens and scientific study.

Initially, the Mammoth Site used string-grid mapping and hand-drawn cartography. The maps were very accurate and the best to be had at the time, but did not take advantage of emerging technologies. Early efforts to computerize the mapping process used AutoCAD, early GIS software, and a Nikon laser transit. However, the software was not well suited to specimen mapping and the Nikon transit proved difficult to use in the bonebed. The system automated the catalog but proved to be less accurate and less detailed than the original string-grid system. In 2007, The Mammoth Site embarked on a complete re-map of the bonebed using the latest technology. The site upgraded to ESRI ArcGIS 9 mapping software, supported by a Trimble 5200 robotic transit and a crane-based camera (CraneCam).

The remap project was completed in 2009, and the new map now shows the precise locations of over 1,100 bones and hundreds of fragments and geological features to within a few millimeters. The new map also contains 28 fields of ancillary data for each specimen including species, orientation, quality of preservation, associations with other specimens, and the names and thoughts of the excavators. The Mammoth Site is taking the mapping project in two new directions. The first is a project to add legacy data to the ArcGIS 9 map, where, using previously existing anchor points, data from the hand-drawn maps will be incorporated into the current electronic map. The second will give the maps another dimension. Using photographs from the CraneCam and special software, a preliminary three dimensional point cloud has already been generated that compares favorably with LIDAR-style 3-D maps.