5
METHODS
This document provides objective and subjective verification of winter 2003-2004 MDSS
weather forecasts and road treatment recommendations, respectively. In addition,
analyses of the road temperature model, rules of practice and mesoscale weather model
components are provided. Where possible, results and recommendations are summarized
to provide guidance to those planning to implement MDSS components or similar
technologies.
Obtaining sufficient high quality verification data continues to be a challenge for this
project and similar efforts. Issues related to verification data limitations are described
throughout this document.
Surface weather observation quality was assessed in 2003 via coincident observations of
state and road parameters. Differences apparent in the observations, themselves, set a
lower bound for the accuracy one can expect from the MDSS forecasts, for if the
observations can only be measured within a certain tolerance, then differences between
such observations and the MDSS forecasts are attributable to the uncertainty in the
observations themselves.
Objective verification is achieved via direct comparisons of MDSS forecasts and
observations from National Weather Service and Road Weather Information system
(RWIS) weather stations, including diagrams of root-mean squared error (RMSE) and
bias for state parameter fields (e.g. air temperature, wind speed) as well as road
temperature. The complexity and subjective nature of the verification of road treatment
recommendations lends itself well to a case-study approach. This approach places the
recommendations into the necessary context of the forecast itself, as well as the actual
conditions that occurred. Contrast between the forecast and reality are often at the root of
differences between the recommended and actual treatments. Several case studies are
presented here, including light snow, heavy snow, mixed precipitation, cold rain and
blowing snow events.
6
MDSS VERIFICATION ROUTES
A total of 16 routes were configured in the MDSS for the 2003-2004 field demonstration.
The selected routes are shown in Fig. 6.1 and a corresponding description of the routes is
provided in Table 6.1. Separate treatment plans were generated by the MDSS prototype
for each of the routes. Much of the verification will focus on Route ID #3 (I-35, north of
Ames), due to the proximity of an RWIS, the Ames airport METAR and the Ames
Garage, where efforts were focused during the demonstration.
10
Previous Page
