Forest Service puts its diverse databases in orderBY PATRICIA DAUKANTAS
| GCN STAFF
Standardizing Forest Service data is as much a cultural challenge as a technological one.
The widely dispersed agency keeps ecological and geographic data in hundreds of databases and filing cabinets. Researchers for separate forests often have different ways to count the same things.
Now managers from coast to coast are coordinating and standardizing their records and database formats throughout the service. Even though work on the Natural Resource Information System is distributed among many offices, the coordination effort itself represents a big change in the way the 95-year-old agency manages information.
Paul Bradford, assistant director of corporate resource databases and a leader in the project, said he hopes the NRIS modules'organized around ecological themes'will all be up and running by October 2002.
The service manages 192 million acres of public land within 155 national forests and 20 grasslands'about 8.5 percent of U.S. land area. Local line managers enjoy a good deal of autonomy and over the years have built information systems without consistency.
'We have 95 years of people going their own ways,' said Michael A. Cummings, a Forest Service resource information specialist. Information has been tucked away in nonstandard database fields as well as in desk drawers.
When researchers began to study national forests as part of larger ecosystems, the haphazard recording methods became a hindrance to analysis and decision-making.
Even where forest lands abut, there has been no way to analyze data for both areas together, Cummings said.Falling behind
The Forest Service got its first big computer in 1983, an Eclipse MV minicomputer from Data General Corp., Cummings said. But decentralized responsibility meant that agency efforts still proceeded at the local level. The service lagged behind other agencies in getting Web access and external e-mail.
Around 1989, Oracle Corp. databases were added to the Data General platform, and local and regional officers started building their own databases. But they didn't always collect the same types of data or record them consistently between forests or regions.
In 1997, the Forest Service's first chief information officer started managing computer applications, and around the same time, inventorying databases and other applications for the year 2000 transition, Bradford said.
The date change issues 'got management's attention as to what was out there,' Cummings said.
Bradford credited Christopher Risbrudt, the director of ecosystem management coordination, with advocating NRIS to other senior executives.
NRIS consists of seven modules, five of which are organized around forest components: vegetation, soil, air, water and animal life.
A sixth module, called Human Dimensions, will aid socioeconomic assessments of national forest use, and a Tools module has analytical applications for data from the other modules.
size="2" color="#FF0000">The Forest Service's Paul Bradford says the project will be complete by October of next year.
The NRIS team works with program representatives to make sure they fulfill the Forest Service's needs at national as well as local levels.
For 'more consistent, more supportable decisions,' Cummings said, the Forest Service is establishing protocols for data collection.
As an example, Bradford noted, millions of dollars each year go into stand exams, or tree censuses based on statistical sampling. Yet workers in different forests were not coding identical species or measuring tree growth in the same way.
Now the programmers of the Field Sampled Vegetation module are developing common standards for coding and measurement to make data consistent.
NRIS will also link tabular data with images and maps in geographic information systems, Bradford said. Researchers can designate a part of a forest and link to its fish, trees, soil, and air and water quality data to make land management decisions.
When completed, NRIS will consolidate hundreds of databases into six interoperable data warehouses that all the researchers can access.
Programmers are now transferring the data, but less than half of the databases have migrated so far, Cummings said.
NRIS development has cost about $10 million per year over the last couple of years, Bradford said.
The modules consist of Oracle7 7.3 databases running on more than 120 IBM RS/6000 servers distributed around the country, Cummings said. A couple of modules still use Microsoft Access because they were created for the smaller database manager. Eventually Bradford's team wants to upgrade all the modules to Oracle8i.Building a bridge
To handle GIS data, NRIS uses several products from Environmental Systems Research Institute of Redlands, Calif., and will turn to ESRI's ArcSDE spatial data engine to bridge the gap between the GIS data and the Oracle databases.
Though NRIS is designed for field-level employees, the Forest Service is considering copying the data into a publicly accessible data warehouse.
'Some of our specialists aren't comfortable with that because they feel they own that data,' Bradford said.
Line officers use their data to determine how to manage the lands under their jurisdiction.