Temporal and weather-specific analysis of injury severity in three-wheeled rickshaw crashes: a heterogeneity-aware modeling approach

Three-wheeled motorized rickshaws (3-WMRs) present serious safety risks in developing countries, yet the joint effects of weather and temporal changes on crash severity remain underexplored. Using six years of police-reported 3-WMR crash data from Rawalpindi, Pakistan (2017-2022), we estimate a partially constrained temporal random-parameters logit model with heterogeneity in means to capture unobserved factors and time-varying effects across weather types and biennial periods. Severe or fatal injuries are more likely in crashes involving motorcycles, pedestrians, or trucks; single-vehicle events; older drivers; peak-hour traffic; wrong U-turns; speeding; and distraction. Importantly, the magnitudes and directions of several effects vary by weather and over time. Rainy conditions elevate severity risk (e.g., reduced friction), whereas sunny weather is linked to riskier behaviors (e.g., speeding). The results support adaptive, weather-specific countermeasures, including stricter speed enforcement in rain, enhanced signing/markings for low visibility, and targeted training for high-risk groups such as older riders.