Box 3-C--Lessons Learned From the GOES Experience

The GOES system has been widely praised for its abilities to track both slow moving weather fronts and rapidly developing violent storms. GOES is credited with saving many lives since the first satellite was launched in 1975. For example, GOES images have contributed to improved early warning of violent storms, resulting in a global 50 percent decrease in storm-related deaths. Yet the development of the GOES follow-on, called GOES-Next, has met anything but calm weather. GOES-Next has been beset by management and technical problems that have resulted in a large cost overrun.[1]

NASA and NOM have a long history of cooperation in developing spacecraft. An agreement between the two agencies, originally signed in 1973, gives the Department of Commerce and NOAA responsibility for operating the environmental systems and requires NASA to fund development of new systems, and fund and manage research satellites. This NASA line item is known as the Operational Satellite Improvement Program, and was usually funded at an average level of about $15 million per year.[2] Prior to initiating GOES-Next development, this division of labor seemed to work well. NASA had developed the TIROS and Nimbus research satellites, which carry instruments that were eventually transferred to NOM operational satellite systems. NASA and NOM budgets and organizational structure were based to an extent on the agreed-upon division of responsibility.

NASA and NOM cooperation became less effective over time. During the transition to the Reagan Administration in 1981, NASA faced cost overruns with ongoing programs and began to spend more of the available resources, including the line item that was used for NOM development, on the Space Shuttle. In addition, the Reagan Administration was slow to appoint senior agency management in NASA. As internal pressures mounted, NASA decided not to fund development of NOM operational sensors and spacecraft. With the concurrence of the Office of Management and Budget, NASA eliminated the budget line used to fund development of new sensors for NOAA systems.

The GOES satellites operating at the time had life expectancies that would carry the program through the late 1 980s. NOM decided to build a GOES follow-on by 1989 that included a major design change. The system requirements led to a very sophisticated design. NOM wanted to improve the sensor's visible and infrared resolution and to operate the sounder simultaneously with the imager. In responding to a GAO investigation of the GOES program, NASA officials agreed that NOM's requirements would be hard to meet.[3] In an effort to shave costs, NOM eliminated the Phase B engineering review, an evaluation of satellite design and design changes.[4] What was not clear to NOM program managers at the time was how great a departure from the original design was required. NOM was confronted with the following in deciding how to replace the GOES-D satellite:

• NASA established a policy that future NOM satellites should be designed to be launched by the Shuttle. The existing GOES design, optimized for launch on an expendable launch vehicle, was not. NASA's policy was subsequently revised after the Challenger disaster in 1986, but the new design (GOES-Next) had already been locked in.

• Several of the early GOES satellites had not demonstrated adequate reliability, failing earlier than expected. This forced a decision to advance the procurement schedule.[5]

• Since NOM was traditionally an operational entity, it had little hope of receiving approval for satellite R&D funding, yet was pressed to proceed with a follow-on NASA procurement.

• Satellite manufacturers, though aware of the problems with the original GOES design, stated that providing simultaneous imaging and sounding could be incorporated with only modest risk. NOM and NASA managers were skeptical of these claims, but they also needed to proceed quickly with the new design.

• A detailed interim engineering review for the GOES-Next plan was canceled for budget reasons. This review might have revealed some of the problems contained in the original design.[6]

These factors complicated the decision to proceed with the improvement to GOES, which became known as GOES-Next. The design change dictated by launch capabilities was unavoidable, given NASA's launch policy. NOM proceeded with an ambitious effort that carnouflaged some of the risk involved with developing GOES-Next. Nothing in the history of either of the contractors involved (Ford Aerospace[7] and ITT) indicated they were less than qualified for the task.

The experience with GOES-Next highlights the problems of interagency cooperation within the U.S. government. When NASA stopped funding development of operational satellites, agency responsibilities were no longer clear. Funding authority for development of future operational satellites needs to be clarified.

____________________

1 GOES-Next was originally bid at about $650 million; estimated total costs are now over $1.7 billion, including launch costs, which should average nearly $100 million per launch.

2 This figure was significantly higher during the early 1970s.

3 U.S. General Accounting office, "Weather Satellites: Action Needed to Resolve Status of the U.S. Geostationary Satellite Program," Report to the Chairman, Committee on Science, Space, and Technology, House of Representatives, July 1991.

4 Eric J. Lerner, "Goes-Next Goes Astray," Aerospace America, May 1992.

5 The early GOES satellites (D-H) were plagued with the same problem--a small component that was essential to determining the direction of the field of view of the VAS sensor prematurely failed. The problem was eventually overcome, but not before NOAA was faced with an early replacement for two of its operational satellites.

6 Lerner, op. cit., footnote 15.

7 Now part of Loral Corporation.

SOURCE: Office of Technology Assessment, 1993.