next up previous
Next: Scientific Datasets and Up: Collections Previous: Collection Loading Strategy

Loading of Datasets

The following dataset loading activities occurred over the preceding year.

The Earth Data Analysis Center (EDAC), located in New Mexico, cooperated with Alexandria to load 66 images of satellite images and aerial photographs. Alexandria and EDAC at first shared their respective basic data structure to find the common ground.
The Sierra Nevada Ecosystem Project (SNEP) worked closely with Project Alexandria to see their cataloged items be included in Alexandria and accessible from the Alexandria catalog. SNEP is a congressionally mandated 3-year study of the entire Sierra Nevada range. The Project produced a full report of their study as well as 354 GIS coverages available in Alexandria.
In relation to Historic Aerial Photographs of Santa Barbara County, California Fairchild Aerial Surveys flight C-4950 (1938) produced air photos that are still used very hearvily, and was therefore selected as a test case for scanning and metadata creation. 1994 Aerial Photographs of Southern California NAPP-2d cycle (National Aerial Photography Program of the U.S. Geological Survey) was a second flight selected for scanning, at the request and with the support of the University of Arizona's Dr. Hsinchun Chen, for a research project he is working on with ADL and with NCSA. This one-month project (January 6-mid-February 1997) has resulted in approximately 400 frames (ca. 15 GB total) being scanned and metadata being created, including obtaining coordinates from indexes by the use of GIS software. We are working on having this project continue.

A small Database of DEMs/DOQs/DRGs was loaded in conjunction with the activities of the ADL Operational Library Team working with Microsoft (San Francisco Office) to construct a 30GB pilot test for a proposed one-terabyte dataset. This database is intended to serve as a proving ground for serving data out over the Web, using Microsoft database-management software. Datasets experimented with were the DEMs (U.S. Geological Survey Digital Elevation Models, derived from 1:24,000-scale map sheets), DOQs (USGS Digital Orthophotoquads) and DRGs (USGS Digital Raster Graphics; scans of 1:24,000-scale map sheets). Microsoft felt that the individual DEM/DOQ/DRGs were sufficiently large (around 30 MB) that they had to be "sliced and diced" in order to be transportable over the Web in some short time frame (e.g. 10 seconds) that would be acceptable to Internet users. In addition, metadata had to be either created or derived - at a parent level for each of the major types of data, and at the child level (and thus tied back to the appropriate parent record) for each file. The team was able to complete the 30GB pilot-test database in time for a January 1997 demo to Microsoft management.



next up previous
Next: Scientific Datasets and Up: Collections Previous: Collection Loading Strategy



Terence R. Smith
Thu Feb 20 13:50:53 PST 1997