Cycling speed and weather

Abstract

Weather significantly influences cycling speed, with important influencing factors being physical effort, safety concerns and comfort. This influence can affect cycling behavior and ultimately cycling attractiveness. However, little evidence exists on how weather influences cycling speed. This study examines this influence, considering varied weather-sensitivity across people and the self-selection bias that some people avoid cycling during bad weather. It also explores cyclists and geographic heterogeneity, i.e. across cyclists and places. Using GPS-tracked cycling data and detailed weather data, a factor mixture model revealed different weather-sensitivity groups, and multilevel linear models with an autoregressive residual structure estimated the impact of weather components, weather-sensitivity groups and bicycle infrastructure on cycling speed. The results reveal three groups: weather-sensitive cyclists, who are sensitive to bad weather; less-weather-sensitive cyclists, who, in contrast, are less influenced by weather overall; and less-rain-sensitive cyclists, who are similarly less affected by weather, but are especially less sensitive to rain. These three groups do not differ in cycling speed. In addition, the influence of rain on cycling speed does not vary across groups after accounting for bicycle infrastructure, indicating no cyclist heterogeneity regarding the impact of rain on cycling speed. Snow, ice, twilight and darkness decrease speed, while fog increases it. Tailwinds increase speed, and headwinds decrease speed; wind shelter mitigates these effects by reducing wind exposure, indicating geographic heterogeneity. The findings are relevant for urban planning and infrastructure design to reduce the negative impacts of weather.