Stated that he would like to see the research program continue given the rapid
progress being made. He indicated that he appreciates the outreach that is being
done on road weather issues.
The interface is intuitive and easy to navigate.
They never just "plow". If they are out there, they will always drop something.
The only time that they just plow is if the snow is falling too fast.
He really likes the addition of the blowing snow alert. Bridge frost forecasts
would be helpful. He would like to have a program running in the background
that would pop up an alert window if an element crosses a threshold (e.g. blowing
snow).
The MDSS would best be used in an organization with a dedicated "snow desk"
where someone could help in coordinating the different garage operations.
Currently, the garages use a "waterfall effect" for coordination. As the weather
affects one route area, the supervisor calls the next downstream office that passes
the information on.
16 LESSONS LEARNED OR CONFIRMED
This second MDSS field season provided a wealth of information to the development
team. Lessons learned or confirmed are listed below.
Road Weather Forecast System (RWFS)
There were many cases were the human could see light snow in the raw model
data, but the RWFS filtered it out. The weights used for QPF drifted due to
poor observational data and the regression process added negative biases to
the data, which filtered weak cases. Corrective actions included:
a. Fixing the weights using expert opinion until good winter
precipitation observations become available
b. Negating automatic bias correction on QPF data due to poor winter
precipitation observations
The data fusion used in the RWFS performed well for verifiable parameters.
The forward error correction scheme worked well.
The blowing snow potential product was beneficial. A more sophisticated
blowing snow product is desired. The algorithm should also include a road
direction factor to account for crosswinds vs. along-road winds. The users
161
Previous Page
