Analyzing Spatial and Temporal Traffic Crash Dynamics Across Rural and Urban Areas of North Dakota

Abstract: Understanding traffic crash seasonal and spatial variations is crucial for improving road safety. Several studies have investigated crash seasonality and emphasized its importance in implementing effective traffic safety measures. However, there is a notable lack of comprehensive research focused on cold regions like North Dakota, which face unique traffic safety challenges due to extreme weather conditions, varying road surface conditions, and differences in traffic and geographical patterns across different locations. This study analyzes how traffic crash variables, such as crash severity, posted speed limits (PSLs), collision types, intersection types, and surface conditions change with seasonal and spatial variations. A time-stamped dataset of North Dakota traffic crashes from 2013 to 2023 was used. The Additive Seasonal Decomposition Technique (ASDT) was employed to separate the crash data into its seasonal, trend, and residual components, distinguishing patterns over time and space. The proposed technique was effective in discovering the seasonal patterns across different settings. The analysis identified distinct seasonal patterns, with urban areas showing more pronounced seasonal peaks than rural areas and a steeper decline in crash frequency, reflecting the effectiveness of safety interventions. Moreover, intersections and low PSL areas demonstrated more pronounced seasonal variation and a sharper decline in crashes compared to roadway segments, highlighting the impact of safety improvements at these locations. The findings underscore the importance of identifying seasonal and spatial crash patterns to inform targeted traffic safety strategies. By using the proposed technique, transportation agencies can implement data-driven measures to improve road safety, particularly during high-risk periods.