Interesting Article On U.S. ModelsJanuary 10th, 2013 at 11:48 pm by Chad Evans under Chad's WLFI Weather Blog
Weather models converged well to predict Hurricane Sandy’s arrival, but meteorologists complained of problems with U.S. forecasting ability shown up by a European model.
By Dan Vergano, USA Today
A European forecast that closely predicted Hurricane Sandy’s onslaught days ahead of U.S. and other models is raising complaints in the meteorological community.
The European Center for Medium-Range Weather Forecasts on Oct. 23 warned that Hurricane Sandy would hit the East Coast on Monday, days ahead of other major hurricane models, which mostly saw the storm heading out into the Atlantic.
Weather models take data from satellites, weather balloons and other observations and plug them into atmospheric physics equations that computers use to march ahead in time to make storm track forecasts.
Although weather models converged remarkably well by the weekend to predict Hurricane Sandy’s arrival with great accuracy, meteorologists complained of weaknesses in U.S. forecasting capability that were exposed by the better performance of the European model.
“It’s embarrassing, we should have the best forecasts on the planet. And it has an economic cost,” says meteorologist Cliff Mass of the University of Washington in Seattle. Hurricane Sandy’s losses are now estimated at $10 billion by catastrophe-estimating firm EQECAT, in Oakland, with more than 7 million people without power in mid-Atlantic states in the storm’s aftermath.
The European center’s prediction was made on more powerful computers, and ran on higher-resolution models of the weather that simulated the future over longer time periods, beyond eight days, than the one employed by the federal National Weather Service. The European model is widely seen as the best at predicting hurricanes, Mass and others say, as demonstrated with Hurricane Isaac in late August.
“The U.S. does not lead the world; we are not No. 1 in weather forecasting, I’m very sorry to say that,” says AccuWeather’s Mike Smith, author of Warnings: The True Story of How Science Tamed the Weather. “But the bottom line is that forecasters nailed this storm days ahead of its arrival. The people behind Europe’s model should receive a Nobel Prize in physics, this was that powerful a moment in weather science.”
“In general, there are reasons to believe the European model does perform better,” says meteorologist Chris Davis of the National Center for Atmospheric Research in Boulder, Colo. Forecasters typically rely on collections, or ensembles, of model forecasts to make storm warnings, weighing the strengths and weaknesses of each one, Davis added. That makes decisions tricky and reliant on human expertise. “This was an unusual storm, with very sensitive conditions, so perhaps in this case a model with (higher-resolution) saw things developing sooner than others.”
The U.S. Global Forecast System model run by the U.S. National Weather Service switches to its highest-resolution in forecasts only when looking seven days ahead, unlike the European one, which looks ahead further.
“All the models have their strength and weaknesses,” said Louis Uccellini, head of the National Weather Service’s National Centers for Environmental Prediction. He acknowledged that the European model made its East Coast prediction ahead of the U.S. and United Kingdom’s meteorological service. He noted the European model was off in some of its predictions, as well, regarding where it would make landfall and low-pressure measures.
Uccellini pointed to improvements coming to the National Weather Service, which is installing more powerful computers to run fine-grained weather models around August 2013. He does say, however, that extending forecasts beyond the eight-day ones produced by Europe’s model is still “up for debate” at the National Weather Service.
Mass criticized the National Oceanic and Atmospheric Administration, which runs the weather service, for its handling of weather satellites. The Government Accountability Office this summer noted cost overruns and failures in NOAA satellite programs designed to replace a polar weather orbiter that was killed because of cost overruns by the Obama administration in 2010. The real problem, though, is that other models use satellite data more effectively, Mass said.
While he criticizes the National Weather Service as slow-moving, Smith notes that the $991 million agency has suffered budget cuts in the past two years, with more cuts proposed this year by the White House and Congress. “Does it make any sense to cut the weather forecasting service’s budget given how much weather affects the economy?” he asks.
Uccellini says that National Weather Service models have begun a move to use more satellite data in the past year. “We don’t shy away from criticism or what people recommend,” he says. “We shouldn’t lose sight of the predictions that we nailed, for a record storm surge, record low-pressure and record snowfall that our forecasters called correctly based on how they read the models.”