of chemicals. In either scenario SNTHERM-RT does not model the impact of traffic on
the road surface conditions.
Fig. 12.4. SNTHERM-RT predicted road surface temperature using the average solar
flux, observed meteorological conditions, and the forecast cloud conditions. Also plotted
is the maximum and minimum predicted road surface temperatures based on the
ensemble forecast model meteorological conditions.
An analysis of the Root Mean Square Error (RMSE) for the SNTHERM-RT predicted
road surface temperatures using the observed meteorological conditions, average solar
flux, and the forecast cloud conditions from the ensemble forecast model (denoted as
"Based on Obs" in Fig. 12.5) and the predicted maximum (based on fcst red line) and
minimum ("Based on Fcst" blue line) road surface temperature based on the ensemble
forecast meteorological conditions is given in Fig. 12.5. Prior to 06:00L on DOY 75 all
three curves in Fig. 12.5 have similar RMSE values. In the periods after 12:00L on DOY
76 when there was no snow on the road surface, the RMSE values based on the observed
weather conditions tend to be higher than the RMSE based on the ensemble forecast
conditions. This is an artifact of how SNTHERM-RT was run using the observed weather
conditions. As indicated earlier, snow was not artificially removed from the road surface
in these model runs. The snow was allowed to melt naturally. In the SNTHERM-RT runs
based on the ensemble forecast model meteorological conditions the snow was removed
from the road surface if the Ames DOT garage indicated plowing or chemicals had been
applied to the road. The effect of not removing the snow is evident in Fig. 12.4 in the
time period from 21:00L DOY 75 to 12:00L DOY 76. During the afternoon of DOY 75
the solar flux values were on the order of 400 W/m2. Since there was snow on the road
the energy associated with the solar flux melted the snow and was not available to heat
146
Previous Page
