Summary: After 7 years of development and expenditures of $100 million, the National Weather Service has implemented an Automation of Field Operations and Services (AFOS) in two of its four principal regions. GAO examined the AFOS project's justification, technical adequacy, and management.
The Weather Service should halt implementation of AFOS, which is its automated data processing and telecommunications system, until it more completely resolves the system's problems and clearly establishes that the benefits of full operation are worth the substantial costs. GAO found substantial problems in the system's design, operation, maintenance, and management. GAO found that several of the design problems are inherent in the system and cannot be resolved short of a complete redesign. Because of system limitations, the Weather Service has had to freeze the development of AFOS before functions initially planned could be added. To perform these and other added functions, the Weather Service is designing a totally new system which it expects to have in service by 1989 or 1990. Despite the limitations of AFOS, the Weather Service plans to complete national implementation and to use AFOS on an as is basis from 1982 to 1990. AFOS is currently not scheduled to operate until 1984 without backup from the system it replaces, which GAO believes is an unusually long trial period. If AFOS were to be abandoned, it would cost the Weather Service only $28 million to continue using the present system for 8 years. To date, cost overruns of approximately $22 million have been incurred in the development phase. GAO and the Weather Service agree that a new system must be developed. However, GAO doubts that the Weather Service has the staff necessary to simultaneously operate and maintain AFOS and develop the new system.
