Nonlinear modelling of the association between winter weather severity and maintenance expenditures

Abstract

The winter weather severity plays a crucial role in determining the resources required for winter maintenance activities. This study utilized nonlinear models to examine the relationship between winter weather variables (temperature, wind speed, and snowfall) and winter maintenance expenditures (labor, material, and equipment) based on data from the state of Illinois. The data were collected and aggregated by year and district to align with the expenditure data, and the state was divided into three climatic zones, with separate models developed for each. The ROUT method was employed to identify and eliminate outliers before conducting nonlinear modelling. The best-fitting model was selected using cross-validation and R2 evaluation. The results demonstrate that the chosen nonlinear models effectively depict the connection between winter weather and winter maintenance expenses. These findings can aid agencies in efficiently allocating resources for winter maintenance.