Making a successful weather forecast depends on a variety of factors - accurate computer model guidance, upper air observations from aircraft and weather balloons, forecaster intuition, and luck all come to mind. One of the key pillars of the modern forecasting enterprise consists of data gathered from two polar orbiting satellites, which is fed into computer model runs to help hone their forecast projections.
Now that pillar may be weakened by congressional budget cuts, and the National Oceanic and Atmospheric Administration (NOAA) is sounding the alarm, telling Congress that a failure to restore funding for development of the next generation of polar orbiting satellites, known as the Joint Polar Satellite System (JPSS), would significantly reduce the accuracy of weather forecasts, particularly medium-range forecasts, and may have an outsized impact on forecasts for extreme events, such as blizzards or hurricanes.
The budget dispute centers on funding for the remainder of the current fiscal year, which ends in October 2011. A House-passed bill would keep funding for NOAA satellite programs at fiscal year 2010 levels, or $382 million. In comparison, President Obama’s fiscal year 2012 budget request calls for $1.1 billion to be spent on JPSS, a difference of $689 million.
The satellite program has already been substantially delayed, and with one of the current satellites nearing the end of its expected lifetime, further delays could mean a major gap in coverage which, NOAA says, would significantly erode forecast skill.
“... we are already behind schedule and there will be a gap between the existing satellites that provide that weather forecasting and when the JPSS satellites will come on board,” Secretary of Commerce Gary Locke told a House appropriations subcommittee on March 17.
Locke, whose department includes NOAA, warned lawmakers about the consequences of losing one of the current satellites before launching a replacement. “We are now able to provide forecasting as far as out to seven days, whether it is for hurricanes, major snowstorms, and so forth, and especially over Alaska and other parts of the East Coast. Once that gap occurs, for however long it takes until the satellites are in orbit, our ability to accurately predict with confidence weather patterns, weather hurricanes or major snowstorms, will be reduced down to three to five days.”
There are currently two polar orbiting satellites in daily use by NOAA and European agencies. As their name suggests, polar orbiting environmental satellites have orbits that take them above the poles. To be more specific, they make 360-degree orbits around the poles, from about 517 miles above the surface, with the earth rotating underneath the satellites’ path. “In a 24-hour period, the 14 orbits of each polar satellite provide two complete views of weather around the world. By having imagery of the whole globe, meteorologists are able to develop models to predict the weather out to five to ten days,” a NOAA website states.
Polar orbiting satellites differ from the satellites that provide most of the imagery shown in television weathercasts and on websites like the Capital Weather Gang. Most of those images come from geostationary satellites. NOAA’s geostationary satellites, known as the “GOES” series, orbit at a much higher altitude, and are positioned above the earth’s equator at speeds equal to the planet’s rotation. This allows them to stay in a constant position relative to the planet.
So how would forecasters fare if they lost one of the two polar orbiting satellites, as could happen if there is a delay in developing a new satellite? Not very well, it turns out, at least in two specific high-impact weather situations.
NOAA’s National Weather Service (NWS) recently examined (see Powerpoint presentation) how the loss of data from one of the polar-orbiting satellites would have altered the forecasts for the “Snowmageddon” event in February 2010, and they are completing a similar analysis of predictions of the New York City blizzard in late December 2010.
By running computer model simulations of each storm, while withholding satellite-derived data from one set of simulations, meteorologists have been able to tease out the polar orbiting satellite’s contribution.
What they found is that the absence of satellite data would have underestimated the severity of the Snowmageddon storm, transforming it from an epic event into an ordinary moderate to heavy snowstorm. Specifically, five-day forecasts made without the benefit of satellite data would have under-predicted snowfall amounts by at least ten inches in the D.C. area, and errors in one-day precipitation forecasts for rainfall in the Southeast US would have increased by up to 50 percent.
Dr. Louis Uccellini, director of NOAA’s National Centers for Environmental Prediction, says the goal of these studies are to isolate the role that instruments aboard a particular polar orbiting satellite play in producing an accurate forecast for a high-impact extreme weather event, several days in advance.
“... We wanted to see what the impact would be if the data from an afternoon orbit was eliminated,” he said via email. “For this case study on the February 2010 “Snowmageddon” storm we did not know if and to what extent the impact would be, but wanted to determine if there would be a difference by eliminating data from JPSS instruments that would be on the afternoon orbit and run the model over the week prior to this historic event.”
The results speak for themselves. Without data from the two instruments the heavy rain forecasts over the Gulf region and the snow in the mid-Atlantic region predicted half as much precipitation as these regions actually received in the storm. We are in the process of analyzing precipitation forecasts without data from the two instruments for the December 26-27, 2010 New York City blizzard. Initial results show that if we deny using data from the instruments we get a similar outcome.
Is NOAA Exaggerating the Risks?
Steve Tracton, a CWG contributor who was actively involved while at NCEP in the design, conduct, and evaluation of experiments designed to assess the relative importance of differing observing systems on model forecasts, points out assessing the value of individual observing systems is among the most challenging in development of numerical weather prediction.
Based on his experience and knowledge of the essential elements of the satellite impact experiments NOAA conducted, Steve is confident the results are scientifically credible. “As Louis says, “the results speak for themselves”, but not necessarily any further,” Tracton says. “The sample of one or two cases is far too limited to draw general conclusions about the importance of the satellite data to significant weather over the U.S.”
Tracton agrees with Uccellini, though, that overall polar satellites play an integral role in improving the accuracy of model forecasts.
As Uccellini stated: “Satellites play an irreplaceable role in forecasting with their ability to quickly provide large amounts of information about the atmosphere, land and ocean and in vital areas where other means of observations are more sporadic or aren’t possible. In terms of forecasting, you can’t tell where you’re going if you don’t know where you are. Forecasting begins with the most complete and accurate understanding of current conditions to most accurately predict into the future.”