When winter storms head for North Carolina, the folks at Raleigh's National Weather Service office head for Kermit Keeter's green note-binder.
In it they find the extra guidance they often need to give local relevance to the data churned out by the weather service's big computers in suburban Washington.The binder, containing the result of Keeter's study of 10 years worth of big winter storms in the state, attempts to tackle one of North Carolina's most ticklish forecast problems - determining the location and width of the boundary between rain and snow.
That line often crosses the state and can be as narrow as a few miles or as wide as 100. An icy mix usually falls along the line, so identifying it accurately is crucial to school systems and road crews.
Keeter plotted incidents of winter rain, snow, sleet and freezing rain for various locations around the state on graphs that took into account temperature variations in the first 10,000 feet of the atmosphere. By matching computer data with historical information on the graphs, forecasters have been able to fine-tune predictions.
``This is good if you've got somebody with experience using it,' said Keeter, a weather service meteorologist.
A forecaster familiar with North Carolina's quirky weather patterns may pick up on variables that call for an entirely different prediction, Keeter explained. For that reason, Keeter said, he often consults more experienced forecasters at the Weather Service office.
Keeter's study has already led to major forecasting successes.
The binder tipped forecasters off that a mid-February 1987 storm headed for North Carolina was likely to be unusual. A day before it hit, they predicted significant amounts of ice for the Raleigh area.
They were right. The storm dumped a record five inches of sleet on the city Feb. 16, 1987.
``These techniques have allowed us to be more specific,' Keeter said.
Keeter, 43, a weather buff since grade school, learned early how important temperature variations in the atmosphere are in determining precipitation.
As an 8-year-old living in Shelby he monitored radio weather reports from atop Mt. Mitchell 90 miles away.
``I learned quickly that on Mt. Mitchell if the temperature was above freezing, it wasn't going to snow,' Keeter said. ``It may freeze a little, but it wasn't going to snow.'
He did forecasts in the third and fourth grades, and by the time he got to high school friends were calling up before storms to find out if they needed to do their homework.
``I always took a keen interest in forecasting winter storms,' Keeter said. ``I think a lot of people can understand that, because snow in North Carolina is an event. To the average native North Carolinian it ranks somewhere between Christmas and the ACC tournament.'
In attempting to define the rain-snow line, forecasters face situations peculiar to North Carolina and the Southeast.
For one thing, a storm-producing low pressure system sometimes dies out over the region and a new low forms to the south. Forecasters know that such a scenario is generally an indicator that the band of frozen mix separating rain and snow will be very wide.
``We're still trying to figure out all the things we need to key on in order to predict whether storms will come through with a wide or narrow band of mix,' Keeter said.
Another phenomenon that North Carolina forecasters pay close attention to is the damming of cold, dry Canadian air against the Appalachians. That generally means snow for the state.
Warm, moist air rises over the cold air mass, and precipitation begins to fall. Initially, the precipitation evaporates in the dry air, lowering the temperature even further. Eventually it reaches the ground - often as snow.
The situation forecasters find most trying is when temperatures remain close to 32 degrees through the critical layers of the atmosphere. Such was the situation at Greensboro Jan. 22, 1987, when a storm left 12 inches of snow. A degree warmer, and 1.2 inches of rain would have fallen instead.
``Do you understand now why you can have a forecast of one inch of rain and get 10 inches of snow?' Keeter asked.
In North Carolina, forecasters are hampered by a lack of information on temperatures at the various levels of the atmosphere because only one inland point in the state - Greensboro - takes such readings. And those readings come only every 12 hours, although they can be taken more frequently if conditions warrant.
That's particularly critical since the state's climate varies from near tropical in the southeast to near Canadian around Boone.
``You have major potential for forecast busts,' Keeter said.