relative humidity, wind speed, and precipitation were the same, only the solar flux and
the cloud information used to compute the IR fluxes varied.
Table 12.1. Meteorological combinations used for the different SNTHERM-RT model
runs based on the observed meteorological conditions
Solar Flux
Cloud information
Hourly average values
Ames municipal airport surface observations
Hourly average values
Based on the forecast model
Values on the hour
Ames municipal airport surface observations
Values on the hour
Based on the forecast model
12.4 Results
We also decided to look at the spatial variability in the road surface temperature using the
CRREL infrared camera. Infrared camera imagery of the road temperature sensor located
on I-35N just north of the Ames DOT garage was obtained for 3/17/2004 at
approximately 16:25 local on a sunny day (see Fig. 12.3). Four Areas Of Interest (AOI)
surrounding the location of the road temperature sensor were selected. Histograms of the
temperature information associated with the four AOIs were generated. These AOIs
cover a fairly small aerial extent. The approximate diameter of the road temperature
sensor is less than one foot. The striations in the AOIs are associated with the grooves in
the concrete road surface and we believe are due to differential heating associated with
shadowing. The temperature distributions associated with the AOIs is not symmetrical.
The temperature range for the distributions is on the order of 1.5 to 2.0 degrees
centigrade. The surface road temperature sensor is on the order of 5 degrees centigrade
hotter than the road surface temperature in the AOIs. We believe this is because the
material that is used to embed the sensor (referred to as the `hockey puck') in the road
has thermal properties that differ significantly from the road. Additionally, there may be
thermal contact resistance on the underside of the temperature sensor puck. The
temperatures as observed in the IR camera imagery may not reflect the temperature
measured by the internal sensor in the puck. SNTHERN-RT is a one-dimensional energy-
mass balance model and is not capable of predicting the temperature variability evident in
the IR imagery. The road surface temperature predicted by SNTHERM-RT will depend
on the physical, thermal, and optical properties of the road layers used in the model; the
hourly meteorological information used to drive the model; and the model physics.
Because of uncertainties and spatial variability in the physical, thermal, and optical
properties of the road layers, we should not be surprised to see differences of several
degrees centigrade between the SNTHERM-RT predicted road surface temperatures and
the measured road temperatures.
To provide Iowa DOT with a prognostic capability, the meteorological conditions used to
drive SNTHERM-RT were obtained from an ensemble forecast model. During the
reanalysis, observed meteorological data was used to determine if there was any
significant difference between the predicted road surface temperatures using the two
different sources of meteorological information, the observed meteorological conditions
143
Previous Page
