Guest commentary

Here in Atlanta, at the annual meeting of the American Meteorological Society, much of the talk is about last week—the snowstorm that paralyzed Atlanta, produced epic traffic tie-ups, and stranded countless school children and office workers.

In this aerial photo, traffic is snarled along the I-285 perimeter north of the metro area after a winter snow storm, Wednesday, Jan. 29, 2014, in Atlanta.  (AP Photo/David Tulis)

Much of the post-storm analysis has raised questions about the failure of state and local officials to better protect the city, given that forecasters predicted hazardous winter weather. But I think we should also draw lessons about improving our forecasts.

Weather forecasting is one of the great success stories of our time. Thanks to decades of investment by the government and private companies, we can often predict major storms three to five days in advance. The fatal westward turn of Superstorm Sandy into New York and New Jersey in 2012 was correctly simulated by a computer model almost six days out—a feat that would have been unattainable just a few years earlier.

Clearly, however, these forecasting capabilities are not sufficient to properly protect our nation. Atlanta is just the latest example in a long series of events (including Sandy, Hurricane Katrina in New Orleans, and the Joplin tornado) that have proved disastrous in some way despite advance warning from meteorologists in each case. As society has become increasingly interconnected, the toll of these weather disasters is staggering. Major weather and climate events have cost the United States more than $1 trillion since 1980, according to the National Oceanic and Atmospheric Administration’s National Climatic Data Center. Moreover, the annual losses appear to be mounting.

What, then, is to be done?

While the factors behind the high cost of weather disasters are complex, at least part of the solution must include longer-range and more detailed forecasts. Such forecasts, when clearly communicated to decision makers who then act on that information, will go a long way toward better safeguarding society.

In the case of Atlanta, forecasters began warning of the possibility of snow days in advance. The National Weather Service, seven hours before the storm, urged motorists to stay off the road except for emergency travel.

It will take time to determine why the response of local and state officials seemed out of sync with the potential for massive disruption. But what if, three or more days before the storm, forecasters could say that the city would be hit by a particularly dangerous storm that would cause roads to become icy and treacherous right about noon on Tuesday? Would schools and businesses have made plans well in advance, thereby avoiding the massive exodus that took place midday and snarled the roads so disastrously?

Through partnerships that bring together universities, private companies, and national labs, researchers are working to improve weather forecasting models. To do so, they are drawing on more detailed observations from satellites and other technologies such as innovative crowd-sourcing tools, as well as on new insights into the physics of storms.

These models, when run on increasingly powerful supercomputers, provide hope of generating predictions that will extend further out and provide more detail about the potential timing, location, and severity of weather events.

To improve communication of these forecasts, teams of social scientists are studying how to better convey weather information using both words and graphics. For example, can a winter weather advisory be more clearly phrased to distinguish between events that could turn roads into skating rinks vs. those likely to have little effect on traffic?

Related: Are meteorologists to blame for snow disasters in Atlanta and Birmingham?

Not only must we strengthen our forecasts and communicate them better, we must also target our technologies to help where society is most vulnerable. This means, in part, focusing on vital transportation pathways. For example, this winter scientists and engineers are equipping snowplows in several states with environmental sensors and translating that data into a near-real-time portrait of weather and road conditions.

When Georgia officials opened Interstate 285 around Atlanta in 1969, forecasts went out just 36 to 48 hours and contained little information about the timing or intensity of approaching storms. We’ve come a long way since then.

Here in the conference hall in Atlanta, the consensus among professionals from private companies, government agencies, and the university community is that the synergies gained when these sectors work together can bring further, essential improvements. Even if we cannot guarantee a city’s response to a winter weather warning, we know that the insights to be gained by investing in efforts like those described here are needed to improve our forecasting capabilities and enhance the safety and economy of our nation.

We will never be able to predict weather perfectly. But, with the proper investments in research and technology, our predictions can keep getting better and better.

Thomas Bogdan is president of the University Corporation for Atmospheric Research, a nonprofit consortium of more than 100 research universities based in Boulder, Colorado.