continue to request information on areas that may be prone to drifting snow.
Snowdrift models should be investigated for this application.
Probabilistic/Confidence Products: Work should continue to provide
probabilistic information for key parameters (e.g., road temperature,
precipitation amount, etc.). The conditional probability of precipitation type
graphical product was well received by the users.
Road Condition & Treatment Module RCTM)
Road Temperature Model
SNTHERM Initialization: Using surface and subsurface data from RWIS
helps predictions, particularly when the sun angles are high and the
temperatures are not driven as much by air temperatures.
Use of insolation data from the mesoscale models (instead of calculated it
from modeled cloud layer data) improves the road temperature predictions.
Rules of Practice
Addition of storm characterization logic significantly improved the product.
Adjustment of chemical amounts upward to be less idealistic about amount of
scatter and splatter was the correct approach.
Rules need to account for multiple treatments being performed at the same
time (e.g., more than one truck on a route).
Inclusion of logic to incorporate blowing snow effects would be beneficial.
The rules of practice should be enhanced to automatically recommend
different chemicals based on the predicted road conditions.
Weather Modeling
The "hot start" model approach improved the prediction of precipitation.
Forecast Period: Because winter maintenance decisions are often made 24 hrs
in advance, it would be best to run the "best" models out to at least 24 hrs.
Overall, predicting weather at plow route scales is extremely challenging and
more research is required to improve weather observations and numerical
162
Previous Page
