Georeferencing
Georeferencing is the process of converting locality information in a catalog record (including feature names, directions, and alternative coordinate systems like section-township-range) to a latitude and longitude. Georeferencing provides the following benefits:
Allows samples to be plotted on a map
Allows users to search by coordinate ranges or polygons in a map interface
Allows samples to be added to data repositories requiring coordinates
The primary drawback is that users of the database may be confused by the addition of site info and coordinates to record summary lines, which among other things may discourage re-use of locality records. Additionally, coordinates determined by georeferencing are necessarily interpretive and should never be used to update or refine the data in a record without careful consideration. For example, you would never back- populate the state/province and district/county names for a record matched only a township.
Georeferencing can be done manually or via script.
Manual georeferencing is labor-intensive, requiring 5-10 minutes per record even with modern tools like GEOLocate and Google Maps. However, using this method allows access to sources that are not available in machine-readable formats (like Mindat’s locality records) and may be necessary for complex records.
Automatic georeferencing can be done fairly quickly and works well with fairly simple records, especially legacy records with limited geographical info. It can struggle with complex records containing multiple place names or directional info and is limited to tools and data sources that are machine-friendly. The primary georeferencing tool in Mineral Sciences at present is a script called situate.py that works from the GeoNames gazetteer.
In many cases, script-based georeferencing can match some but not all information in a record. This is common with records that include mine names, which are highly specific but poorly represented in gazetteers. It is often possible to match feature, town, or county names in these records, in which case a reasonable workflow would be to georeference the record to that level and flag it for manual review at a later date.
Collection-specific georeferencing information
Meteorites collected by the Antarctic Search for Meteorites Program (ANSMET)
Coordinates for meteorites collected by the Antarctic Search for Meteorites (ANSMET) program are taken from NASA’s Complete Listing of Antarctic Meteorites in the U.S. Collection. NASA describes their coordinate data as follows:
Position data has been derived from a variety of surveying techniques (primarily GPS since the early 1990’s) used by ANSMET field teams and reflects a typical uncertainty of around 10 m. The data has been truncated to avoid confusing high precision (many decimal places) with high accuracy (low error). These positions reflect a realistic “best available” accuracy, at ˜50x higher resolution than required by federal regulations.
All meteorites collected by ANSMET are assigned source=ANSMET, radius=10 m, and a determination date marking the year of collection. Meteorites collected before 1993 are assigned method=Surveying and datum=Unknown. Meteorites collected during or after 1993 are assigned method=GPS and datum=WGS 84.