Modeling Impacts of Weather Conditions in Agent-Based Transport Microsimulations

Abstract

Despite all mitigation efforts, climate change will impact transport systems. As the exact effects are still uncertain, a range of possible climate change outcomes has to be considered. For that purpose, it is important to have an instrument modeling a transport system and its behavior in different weather conditions. This paper shows that agent-based micro-simulations represent a promising approach for comprehensively modeling the impacts on transport systems. Moreover, based on evidence from previous literature, a new comprehensive classification of the various transport system aspects affected by climate change is proposed: Transport infrastructure, safety, travel behavior and socio-economic circumstances. Existing weather-sensitive models are restricted to impacts on transport infrastructure and driving behavior or include very simplistic activity-travel behavior models. Different approaches are sketched for the simulation of regular and unexpected weather conditions within the Multi-Agent Transport Simulation (MATSim). For regular weather conditions, it is proposed to search for tipping points where the performance of the transport system starts to significantly deteriorate due to the weather condition. For certain impacts categories, in particular for activity-travel behavior, the degree of impact and sometimes direction varies within the literature, calling for further analysis. Modeling unexpected weather conditions requires the application of the within-day replanning module within MATSim. The key challenge here is the definition of the appropriate replanning strategies when an unexpected event occurs.