 |
 |
 |
 |
 |
 |
 |
 |
 |
|
|||||||||||
|
Site Flash 1998Peter McCartney, CAP Data Manager The past year has been a learning experience at CAP LTER, and no less so for the data management team. Our primary goals for the first year have been:
ResourcesComputing activities at CAP LTER are carried out over a distributed network of resources. At the core is the computing facility maintained jointly by the CAP LTER and the Archaeological Research Institute (a collections and data repository within the Department of Anthropology). In January, Ted Oliver was hired as staff to assist McCartney in database management and system administration. The lab currently has approximately 1500 square feet devoted to computing activities. The LTER server is a PII 300mhz equipped with redundant power, cooling and storage (RAID 5). Current capacity is 27 gigabytes. Backups are made to tape following a 3 week rotation. This server is used for the LTER data archive, web and FTP services and general user file services. In addition to personal computers used by the data manager and staff, the lab offers 6 Windows NT workstations for research and several Windows 95 machines used for data entry. We also received two Macintosh clones that were part of a larger corporate gift from Motorola. The database server for CAP LTER is Microsoft SQL Server. Applications programming is done using Microsoft Access and Visual Basic. Databases are accessed through the web service using server-side scripting (MS Active Server Pages). GIS software includes ESRI ArcView,MapObjects, and ERDAS Imagine. X-windows software is provided for running remote sessions on the university Unix systems. The Information Technology GIS lab is a partnering facility directed by Jana Hutchins (co-leader of the Data/GIS/RS core team). It provides Unix and NT access to ArcInfo, ArcView, and Imagine. The lab also provides high-quality input and output devices. Because of pre-existing working relationships with many LTER community partners, the GIS lab is a primary channel for data-sharing between these sources. Finally, the Geological Remote Sensing Lab directed by Phil Christensen receives and processes all remote sensing data used by CAP LTER. This lab is a mixed Unix and NT environment using ER Mapper and ERDAS Imagine as the primary software tools. This lab is currently developing a remote sensing server which may be used as the primary distribution medium for LTER image data. ProceduresThe general procedure for designing, implementing and documenting project database was developed and has been applied to six pilot projects. After consulting with the PIs, a data modeling tool (ER/Studio from Embarcadero) was used to produce a generalized schema of the database complete with entity and attribute metadata. One advantage to this type of software is the ability to deploy a database schema on several physical platforms. For CAP LTER, most databases are built in MS Access, then migrated to SQL Server after design becomes stable. Interfaces are developed with Access which features very rapid development time, rich options for quality control, and easy redirection of forms when tables are migrated from Access to SQL Server. Completed databases and entry applications are delivered in less than a week after the initial meeting with PI. After initial development, draft metadata in the form of data dictionaries are posted on the data management section of the intranet site. Data management has also been involved in projects that seek to acquire and import existing data from external sources. Research Assistants employed by such projects frequently work with these datasets at the LTER lab so that we may provide support for these activities. Considerable effort has gone into designing a comprehensive database to hold metadata for CAP LTER datasets. Three sources were considered: Michner et al. paper on metadata, the Federal Geographic Data Committee, and an informal survey of existing metadata documents on other LTER sites. Our strategy has been to use the Michner paper as a conceptual guide to the different levels of information required, while adapting the specificity of the FGDC standard wherever applicable (e.g., handling keywords, projection information, entity and attribute descriptions). Once a draft format was developed, it was compared to existing LTER metadata to ensure maximum completeness. The draft structure of the database has been posted in the form of an Entity-Relationship diagram and a data dictionary. Based on the LTER data managers' draft, a data access policy was developed for CAP and is awaiting final approval. It will be published on the data access page of the web site, along with a mandatory login form for visitors wishing to download data. ProductsTwo immediate tasks were to define the core databases needed to manage LTER data and to set up the CAP LTER website to provide a medium for accessing information. A series of integrated databases were defined and implemented as web applications: a bibliography based on a previously compiled list of environmental literature for central Arizona, a projects listing personnel directory, and a current events calendar. The dataset catalog and associated web application is nearing completion and should be online within a few months. This application will rely on database calls for both the data tables and the metadata. The plan is to design metatada output to follow several formats on demand including the LTER exchange format introduced by Porter last summer, FGDC, and any future LTER standard. Because the participants in CAP LTER are distributed, the web site has played in important role in local, as well as external, communication. An Intranet site was created for posting internal announcements, meeting minutes, and extensive computer support pages. A series of NNTP discussion groups were created and participants have been encouraged to use them so that a record can be made of our electronic dialogs. Two projects, one directed by McCartney and Fry, the other by Phil Christensen and Mike Ramsey compiled existing GIS framework data and remote sensing imagery to provide a spatial framework for initiating LTER research. Available datasets have been posted for web and ftp access; in addition, websites were created to summarize and preview GIS and remote sensing data. LTER data management was active in external grant efforts. McCartney was coPI on three proposals that involve collaboration between the ARI/LTER lab and various departments. One, a proposal to NSF for vBNS connectivity was funded in November. Two other NSF proposals are pending: one to the Archaeology program to prepare metadata and internet publication of a 30+ year research project on the archaeological site of Teotihuacan, Mexico, and the other to the URI program to study social factors associated with relative success of several water-based community development projects. McCartney and Fry also co-authored the data management portion of a proposal to the EPA Empact program to bring environmental information to the public. McCartney offered a course in GIS in Archaeology in the Spring semester, which was attended by several students from the LTER program. In response to requests from PI's, a modified version of this course focussing on GIS and Database management for ecological research may be offered in the following spring. DirectionsTop priority for next year is getting the dataset catalog online and to publish the datasets that are ready to go online. We will also be adding a web-based GIS application developed with Visual Basic and MapObjects to provide more interactive display of GIS layers from the study area. In response to PI feedback, a new feature to be added in the fall will be a regular series of training workshops designed to familiarize LTER scientists with the resources and data available at the LTER lab. These will include primers on accessing the LTER, basic SQL database query, and using simple GIS tools to produce data-rich maps. There will also be training on using data entry applications developed in MS Access. With a supplemental grant for curation from the LTER program, we will expand the current catalog databases to include inventories for specimen collections and for reference collections from the Biology department. These will parallel similar curation catalogs currently maintained for the archaeological collections. Finally, new solutions for getting more complete and rapid information on project research designs and methods will be sought. Both web-based forms for entering project description and more aggressive follow-up with PIs will be considered. The existing project descriptions which are database-derived, will be augmented with static web pages for projects that provide more visual and narrative information. |
||
|
Contact CAP LTER | Webmaster |
||
|
Central
Arizona - Phoenix Long-Term Ecological Research |
||
