Assessing Icing Conditions in the High Elevations of the Northeast
The goals of this project were 1) to understand the meteorological conditions that produce icing conditions in the Northeast and 2) to develop numerical computer models that can predict the timing, location, and severity of icing conditions.
The goals of this project were addressed through the observation and analysis of atmospheric conditions that produce icing by using icing observations to improve forecasts from a high-resolution numerical computer model.
Icing detectors measured icing accretion rates for the 2011-12 and 2012-13 winters at the Mt. Washington summit (6288'), Cog Railway Halfway house (4500'), Cannon Mountain (4032'), and Mount Mansfield (3950'), Vermont. A graduate student at Plymouth State University cataloged significant icing events (≥24 hours) and analyzed the regional-scale weather conditions that produce these icing events. A NASA satellite icing product was verified for accuracy with the icing observations. The icing detector and wind speed at Mt. Washington were used to calculate cloud liquid water content (LWC). During the 2012-13 winter, weather balloons were launched during intensive observation periods to gather vertical profiles of meteorological conditions through the lower troposphere during icing events. The LWC data was used with radiosonde data to verify supercooled LWC measurements from two radiometers located at the base of Mt. Washington.
The Weather and Research Forecasting (WRF) model was run at high resolution (~1 km) for several icing events to determine its accuracy in predicting icing around the high elevations of northern Vermont and New Hampshire. Five WRF model microphysics parameterization schemes were tested for icing prediction. Icing observations at the four sites were used to verify the model and determine skill at forecasting icing and non-icing conditions. The WRF model was coupled with a high resolution digital terrain model and the FASST Land-Surface Model to incorporate atmospheric-surface interactions (exchanges of heat and moisture) that should help improve icing forecasts. Results indicate that simulated cloud LWC is most sensitive to microphysics scheme choice, rather than cloud number concentration. Single-moment schemes produced results as good or better than the computationally-expensive double-moment schemes, suggesting single-moment schemes may be best for icing-related applications for both efficiency and forecast accuracy.
Mt. Washington rotating multicylinder icing data from this project and other past projects were analyzed to determine characteristic drop number densities and droplet size distributions. The results revealed that at Mt. Washington drop number densities are generally larger and best-fit size distributions are generally narrower than is typically assumed in the WRF model (details in Jones et al. 2014). Further research is needed to determine if these results are common in other mountain ranges globally.
Jones, K.F., G. Thompson, K.J. Claffey, and E.P. Kelsey, 2014: Gamma distribution parameters for cloud drop distributions from multicylinder measurements. J. Appl. Meteor. and Climatol., 53, 1606-1617.