bottom soil layer was replaced with a layer that represented atmospheric conditions. For
every hour the temperature of this bottom layer was set to the ambient atmospheric
temperature. The modified version of SNTHERM, referred to as SNTHERM-RT,
assumes the road moisture content is zero. Therefore, the latent heat exchange between
the road and the overlying atmosphere is not part of the road/atmosphere energy balance.
But, rain and snow are permitted to accumulate on the road surface and the exchange of
latent heat between either the layer of water or snow on the road surface and the
atmosphere is modeled.
SNTHERM-RT requires a number of initialization parameters including a time series of
meteorological information. In order to provide Iowa DOT with a prognostic capability,
the road surface temperatures were based on an ensemble forecast of the meteorological
conditions rather than observed meteorological conditions. Even though the model
exhibits considerable skill, there were still periods when the predicted road surface
temperature differed from the observed values by 4oC, especially during periods of high
solar flux (cloud free mid day periods). In order to investigate the potential underlying
cause(s) of these differences, we reran SNTHERM-RT for selected MDSS periods using
the observed meteorological conditions and compared the predicted road surface
temperatures with the observed temperatures and the predicted temperatures based on the
ensemble forecast meteorological conditions. We employed the following approach and
obtained the following results.
12.3 Analysis Approach
Two sites were selected for reanalysis based on the proximity of these sites to observed
meteorological data and road surface temperature measurements. The sites selected were
the Ames RWIS site on I-35N and the Ames bridge site. The ambient temperature, and
relative humidity from the Ames RWIS site; the solar flux from the instruments installed
at the nearby Ames garage; wind and precipitation information from the Ames municipal
airport; and cloud information from either the observations from the Ames municipal
airport or the ensemble forecast model were used in the reanalysis. The observed
meteorological information from the three (RWIS, Ames garage, and the Ames municipal
airport) sites, each with different temporal resolutions, was merged to provide an hourly
meteorological information database used to drive SNTHERM-RT. Since the solar flux
was recorded every minute at the Ames garage we decided to use two different
approaches for computing the hourly solar flux for the SNTHERM-RT reanalysis. In the
first approach we used the solar flux recorded on the hour. In the second approach we
average 60 minutes of solar flux measurements centered on the hour. That is, the solar
flux used in the second approach was the average of the 60 measurements that spanned
the period from 30 minutes before the hour to 29 minutes after the hour. The infrared flux
used in the reanalysis was computed from cloud information obtained from the hourly
Ames municipal airport surface observations or from the ensemble forecast model cloud
information. The following table summarizes the meteorological data combinations used
as boundary conditions for the reanalysis of the MDSS road and bridge temperatures. For
all scenarios based on the observed meteorological conditions the ambient temperature,
142
Previous Page
