Best Practices for Road Weather Management Version 2.0
Title:
Improving Forecasting Performance by Combining Forecasts: The Example of Road-Surface Temperature Forecasts
Abstract:
A recent paper by J.E. Thornes reported the results of a comparative evaluation of location-specific, road-surface temperature forecasts formulated by two organizations, the UK Meteorological Office (MO) and Oceanroutes (OR). This paper broadens the scope by considering the possibility of combining the MO and OR forecasts. Combing forecasts can improve forecasting performance when one set of forecasts contains information ? in this case, information about observed road-surface temperatures ? not contained in the other set of forecasts. Both simple averaging and regression modeling are examined in this study as methods of combining these road-surface temperature forecasts. The simple average forecasts outperform the MO or OR forecasts, both when these road-surface temperature forecasts are evaluated as forecasts of a continuous variable and when they are evaluated as forecasts of a binary variable. Combining via regression modeling leads to further modest improvements in forecasting performance. Overall, simple averaging captures a substantial fraction of the benefits that can be realized by combining the MO and OR forecasts for this small sample of road surface temperature data. Several issues and implications of the results are briefly discussed.
Source(s):
Meteorological Applications of the Royal Meteorological Society Vol. 3 (United Kingdom), National Center for Atmospheric Research and Prediction and Evaluation Systems (United States). For an electronic copy of this resource, please direct your request to WeatherFeedback@dot.gov.
Date: 1996
Author:
Brown, Murphy
Keywords:
Pavement temperature
PDF files can be viewed with the Acrobat® Reader®.
