Road temperatures were also compared between the RWIS and the forecasts for the
RWIS site and segment 3, which is along I-35 from mile marker 112 128, north of
Ames (Fig. 10.23). For the time periods where RWIS data was available for verification
of road temperature during this event, forecasts are generally within about one degree
Celsius.
Road Temperature Comparison for Jan 26, 2004
4
AMW OB
2
0
-2
-4
-6
-8
-10
-12
AMW RWIS OB
AMW RWIS wFSL
AMW RWIS woFSL
-14
SEG 3 wFSL
SEG3 woFSL
-16
18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18
Time (UTC)
Jan 26, 2004
Jan 27, 2004
Fig. 10.23. Road Temperature (C) time-series plot comparing the Ames
RWIS observations to the RWFS forecasts (both with and without the FSL
supplemental models) for the RWIS site and segment 3. The vertical lines
represent the time period that the Ames METAR was reporting falling
Cumulative verification results calculated for the entire state of Iowa for this event
indicate that many of the errors observed at Ames were also evident throughout the
domain, in general. In particular, forecast air temperatures and dew points were
reasonably good throughout the event, but errors grew to be quite large for forecasts
beyond 30 hours (Fig. 10.24). Relatively large errors in dew point (up to 7C) air
temperature (up to 3C) noted at Ames before the event appeared to be outliers compared
to the entire domain, which experienced only slightly larger than normal errors though
hour 10. The relatively small amplitude to the all-station error may partially be a matter
of larger, short-lived errors at different stations occurring at different times as the cold
front moved across Iowa, resulting in relatively low overall RMSE values at any one time
across the entire state. This is true for all fields. Cloud cover errors before and well after
the event appeared to be more significant at Ames than overall. Still, there was evidence
49
Previous Page
