GIS Gets Down to Business

ack in 1983, the Bureau of Land Management began an ambitious program to develop an information system to track the land and mineral resources BLM manages and to automate paper-intensive activities such as leasing U.S. properties to oil and gas producers. Today, 15 years and half a billion dollars later, the Interior Department agency has not deployed an operational version of its new system for its 200 offices nationwide.

The history of BLM's notorious project, the subject of many General Accounting Office reports over the years, in some ways mirrors the history of its core technology, geographic information systems. GIS projects often have turned out to be more resource-intensive, difficult to complete and peripheral to agency activities than they were expected to be.

The promise of GIS is a seductive one: to deliver insight by linking up information now maintained in stovepipe systems. The thread that ties together information in a GIS is location. It's a truism among GIS advocates that 85 percent of all data has a location component. Some examples of how this information can help federal managers:

• Many agencies deliver services through regional organizations. A GIS can give a regional breakdown of the allocation of resources, revenues, agency customers and so on. This kind of analysis can support strategic realignment of activities and resources. The Agriculture Department is using GIS technology in its consolidation of bureau offices in virtually every U.S. county.
• Agencies like the Federal Emergency Management Agency and the Centers for Disease Control that respond to crises nationwide can visualize trouble spots and dispatch response teams more effectively with the help of map-based systems.
• An Army National Guard system helps recruiters fill vacant personnel slots. The World Wide Web-based system displays which of the 2,700 Guard units have the most urgent need for new personnel, and recruiters can click on individual map sections to get specifics about the openings. They also can identify local advertising media and obtain ad rates and other details. This information was available to recruiters previously, but it could not be accessed as quickly and easily.

But these uses just scratch the surface of the technology's potential. Let's look at how the Bureau of Land Management will use its Automated Land and Mineral Record System (ALMRS) once it is deployed.

At its 200 offices nationwide, BLM manages 265 million acres of federal land-one-eighth of the nation's area-and the mineral resources underlying 570 million acres. The agency, once known as the General Land Office, keeps records of surveys, tract divisions, title transfers, mining claims, oil and gas leases, grazing permits, natural resources, homesteads and more.

Most of this information is on 1 billion pages of paper in the agency's files. Some of the documents are more than 200 years old and deteriorating, but they are in great demand. Each year, the bureau gets 6 million requests from the public for information in its files, and the files grow substantially as permits and leases are reissued.

ALMRS will be a repository for all the information pertaining to each parcel of land that the bureau manages. Instead of querying several databases to find out about oil leases and pipeline rights of way for a specific tract, for example, there will be one unified set of records. Moreover, the agency plans to make the data available over the Internet so that researchers need not leave their offices to get the information.

Data Discrepancies

The messy nature of the current records systems is causing delays in getting ALMRS up and running, however. The project's long history is a saga of budget snafus, missed opportunities, changing procurement strategies and disputes over technical and architectural matters.

As far back as 1987, according to an official history of the project, a prototype demonstration found "significant problems with data consistency and accuracy" in BLM's older systems. The testers encountered conflicts and mismatches when they imported old data into the prototype of the new system. The bureau started a data management project to cure the defects, but the same kinds of problems still stymie ALMRS implementation today.

Difficulties in marrying up data from disparate systems are one of the reasons GIS technology has not been adopted at the vigorous rate that analysts and marketeers had predicted. Today, however, data warehousing technologies are available to pull information from various sources so that common elements can be identified and logical connections made. Location often is one of the common data elements that link the data sets.

Meanwhile, mainstream database software packages, such as those offered by Informix Corp. and Oracle Corp., are more able to handle location as an extra dimension of data. Specialized software may not be mandatory for many applications. It's common now to display geographic data with a World Wide Web browser and click on map spots to drill down for more data. "If someone can run a Web browser, they can now run a GIS," says George Korte, a manager in Intergraph Corp.'s federal unit and author of The GIS Book (OnWord Press, 1997).

There were other reasons why GIS was slow to take off. It requires powerful computers to process data and display maps. Today's desktop systems are fully capable of handling the load, but until recently a full-featured GIS ran best on an expensive mid-range computer or technical workstation. In fact, BLM's needs spurred the development of IBM Corp.'s first portable computer that had the guts of a high-end workstation.

To see the power of today's PCs for processing geographic information, you need look no further than the Web, if you have a Pentium system and fast modem or local network connected to the Internet. You can view aerial photos of much of the United States and other places worldwide on the Terraserver, a Web site maintained by Microsoft Corp. It is powered by what Microsoft calls "the world's largest PC"-a Compaq Computer Corp. system with eight powerful processors and massive storage capacity.

Although the computer is by no means an ordinary PC, it runs the same Microsoft Windows NT software used on many federal and corporate networks. It displays not only satellite photos supplied by the U.S. Geological Survey and the Russian space agency, but also maps of the same areas. If the information were printed, it would fill 2,000 volumes of atlases, at 500 pages per volume, Microsoft officials say. The company created the Terraserver to show that its software could handle the tough tasks once reserved for mainframe systems.

Incompatibilities

Another hindrance to GIS adoption has been a lack of standards. Historically, GIS software from one vendor would not work with software from another. Once an organization had committed to a particular supplier, there was no turning back, which made such decisions quite momentous. Also, the inability to move data easily across system boundaries limited the return on investment from GIS systems.

As part of its reinventing government initiative, the Clinton administration mandated creation of a National Spatial Data Infrastructure (NSDI) that would incorporate standards and data exchanges. Progress on its development has been steady but slow. "Many useful steps have been taken to further an NSDI since 1994," a National Academy of Public Administration (NAPA) panel reported in January. But it cited resource shortages and policy gaps as major barriers to full realization of the goal.

The panel recommended creation of a private, nonprofit National Spatial Data Council to coordinate development of the infrastructure. There already are a number of coordinating bodies, both within government and outside, including the high-level Federal Geographic Data Committee, an interagency organization headed by Interior Secretary Bruce Babbitt. The committee has the usual handicaps, including lack of visibility and clout, but has completed work on about 10 of 30 proposed standards.

The NAPA panel's assessment of the kinds of coordination that are needed laid bare the extraordinary dimensions of the issues. The public sector needs to work with the private sector, it said, and within the public sector better intragovernmental coordination is important. State and local government agencies are major generators of geographic information and consumers of federal geographic data.

Even at the federal level, the NAPA report says, civilian and defense GIS activities are insufficiently coordinated. For example, the National Imagery and Mapping Agency, the Defense Department's consolidated mapping and GIS agency, plays only a small role in the Federal Geographic Data Committee.

Given the astounding amount of geographic information in government files and the diversity of that information, the lack of a simple scheme for pulling it together into a unified database is understandable. There's information about air quality, soil types, road conditions, the incidence of diseases, population characteristics, vegetation, water resources, land ownership and much more.

Which brings us back to the Bureau of Land Management. In his April response to a scathing GAO report, BLM Director Pat Shea pointed out that ALMRS is "one of the largest systems yet undertaken by a civilian agency" and said delays were caused primarily by the bureau's desire to get the system right. But the total schedule slippage now exceeds two years. Although GAO reported in May that deployment was scheduled to begin before July, it had not begun by August of this year.

In the end, there's no evidence that the geographic character of ALMRS caused most of the project's problems. GAO's specific criticisms in May focused on purported deficiencies in project management, system security, user training and other aspects of ALMRS not related to GIS. It's more likely that the massive scope of the BLM project lies at the core of its problems.

Looking Ahead

In the future, says Steven Cooperman, a spatial data expert with Oracle Corp.'s government division, the government will continue to squeeze more value from existing data by using the geographic connection. But spatial data will be better integrated with other kinds of information. He predicts that spatial data will become even more central to data systems as global positioning systems (GPS) become more widely used.

For example, GPS could help an equipment manager track the locations of vehicles or portable computers. Tiny computers, perhaps in the form of smart cards, will pick up satellite signals, determine their own locations, and report the locations to a central system. The military applications of such systems obviously are attractive to DoD officials, but civilian uses will be plentiful as well.

As a result, the notion of a geographic information system as a distinct type of system will fade, Cooperman predicts. "I see the term GIS beginning to go away," he says. But location will be an ever more important element in analyses and decision-making, he adds.