UCAR > Communications > Staff Notes > November 1999 Search


November 1999
Science
One of Colorado's most memorable fall snowstorms struck only two years ago, on 24-25 October 1997. Although this event pounded much of the state, not all residents suffered equally. Doug Wesley (COMET) analyzed some of the small-scale variations in a paper presented on 15 September in Denver at the American Meteorological Society's 17th Conference on Weather Analysis and Forecasting. Doug collaborated with Greg Poulos (Colorado Research Associates), John Snook (NOAA), and Michael Meyers (National Weather Service) on "Meso-Gamma Snowfall and Wind Variability in a Front Range Blizzard."

As is often the case, the heaviest snow on 24-25 October fell across the Front Range foothills, along with the western portion of the Palmer Divide. Even in this region, Doug found variations in total snowfall of as much as 25 inches (64 centimeters) across a span of only 10 miles (16 kilometers) in the foothills west of Boulder. Wind speeds differed by twofold or more in some places across the plains. "Some areas--Boulder, for example--experienced only 10- to 25-knot wind speeds [12-29 mph] during heavy snowfall, while others--Denver International Airport--were crippled by 30- to 55-knot gusts [35-63 mph] that generated large snowdrifts and near-zero visibility," says Doug. Because local forecasts tend to cover several counties and geographic regions at a time, he adds, "These variations create serious problems for operational forecasters, both in the forecasting and nowcasting modes."

The storm's snow developed through a combination of upslope wind and the dynamics induced by a large, cold upper-level storm that moved through the central Rockies. Colorado snowstorms often feature streaks of intensified snow, a few miles wide and tens of miles long, that are caused by distinct bands of instability. On 24-25 October, though, "Detailed examination of radar features . . . did not reveal significant banded structure in the precipitation." Instead, the regions of snow were fairly nebulous and were stationary for long periods.

What, then, caused the strong variations in snowfall accumulation? Upslope is one clear factor, although some of the heaviest totals did not occur over the highest elevations east of the Continental Divide. Another noteworthy feature is that the character of the snow itself varied greatly. Snow-to-liquid ratios ranged from about 12:1 (relatively wet snow for Colorado) at low levels to nearly 25:1 later in the storm (a more typical ratio). The ratio also varied distinctly with elevation.

As for the computer models used during the storm, Doug found that the major forecast tools pegged the overall likelihood of heavy snow. "However," he notes, "it is immediately evident that much finer detail is required before forecasters can successfully apply model data in improving predictions of extreme events such as this storm." Large gradients in wind and snowfall occurred on scales narrower than that of even the most precise models currently being used.

Doug has been working with the Colorado State University Regional Atmospheric Modeling System (RAMS) model (along with all available observations) to recreate the storm's dynamics and snowfall production. One preliminary result: better horizontal resolution in the model could aid forecasters in predicting wind and snowfall gradients in real time. Better parameterizations of microphysics--as featured in this version of RAMS--may also make a difference, says Doug: "Model snowfall accumulations are inherently dependent on the degree of sophistication of the model microphysics."


In this issue...
Other issues of Staff Notes Monthly


UCAR | NCAR | UOP

Edited by Bob Henson, bhenson@ucar.edu
Prepared for the Web by Jacque Marshall