The long-range weather forecasts that warned of where Hurricane Katrina would strike depended on data from polar satellites. They capture not only details over the Arctic and Antarctic but also virtually every point on the planet's surface as the world turns under them. Now the replacements for the aging U.S. military and civilian fleet are in jeopardy. The program is as much as $3 billion over budget, and the launch of the first replacement satellite is as many as three years behind schedule.

Today two agencies each maintain two polar satellites at a time: the National Oceanic and Atmospheric Administration and the U.S. Department of Defense. In 1994 the U.S. decided to create the National Polar-orbiting Operational Environmental Satellite System to serve the combined needs of NOAA and the DOD. The joint program's six orbiters would pack in 14 instruments, roughly twice as many as either military or civilian polar orbiters, and save $1.8 billion.

After the U.S. awarded the satellite contract to Northrop-Grumman in 2002, a test satellite with four sensors was supposed to fly this May, and the first fully functioning replacement was set to roll out in June 2008. Now the launch of the test orbiter (overseen by NASA) is delayed until April 2009 and the first replacement is pushed to 2012, NOAA and Defense Department officials testified before the House of Representatives science committee in November. "It is now clear that almost from the outset decisions were made with too little analysis of the technical challenges involved," said committee chairman Representative Sherwood Boehlert of New York at the hearing.

Northrop-Grumman explained that most of the delays and cost growth resulted from problems with the orbiters' sensors. One in particular concerned an infrared camera dubbed VIIRS, which is supposed to collect images of clouds and probe sea-surface temperatures. Northrop-Grumman now says that it assumed VIIRS development would be simpler than it turned out to be because aspects of VIIRS were based on existing instruments.

Both the government and contractors failed to recognize problems in time to fix them, stated David Powner of the Government Accountability Office at the hearing. Powner also noted that the satellite program's executive leadership issued few major decisions, often opting instead to conduct further analysis and review.

To get the program back on track, NOAA, the DOD and NASA hired a new joint program manager. Moreover, Raytheon, the subcontractor building VIIRS, replaced the entire team working on the instrument, and Northrup-Grumman has increased oversight of its subcontractors. To save costs and time, NOAA administrator Conrad C. Lautenbacher said the program might eliminate some sensors from the orbiters. Still, an independent review conducted by the Aerospace Corporation suggests any changes will still leave it at least $1 billion over budget. The House science committee also worries that the joint agency nature of the program hampers problem solving.

Workarounds are possible, if not ideal. The military satellites can probably survive well beyond 2012, although they do not provide all the data NOAA needs for its weather-forecasting models. NASA research satellites or long-duration high-altitude flights might also help. Lautenbacher noted that engineers could bolster the test orbiter--it can already obtain 93 percent of the data that fully functioning satellites can--and could keep existing NOAA polar satellites flying past their average four-year useful life span. European satellites might be an option--if they are available and produce data in a format compatible with U.S. weather-forecasting models.

A revised plan for the program should fall into place this spring. In the meantime, NOAA and the DOD will launch the last of their original polar satellites in December 2007 and October 2011, respectively. If NOAA's fails, as roughly one out of every 10 do during or right after launch, U.S. civilian polar orbiter coverage might suffer a gap of four years or more. (That satellite might be especially prone to failure because the contractor, Lockheed Martin, accidentally dropped it, causing significant damage.)

A gap could be serious, explains Richard A. Anthes, president of the University Corporation for Atmospheric Research in Boulder, Colo. It could, he remarks, "make the difference on whether a forecast of where a hurricane makes landfall has an error of a few tens of miles or a few hundred miles."