Answers are grounded in this paper and cite their sources.
Rapid growth in demand for multiuse trails, together with constraints on public funding, has increased the need for robust demand analysis to inform effective and equitable planning decisions. Existing studies often overlook spatial heterogeneity in demand relationships and the temporal characteristics of prediction error, which can reduce the reliability of investment choices, hinder the development of policies tailored to local conditions, and amplify uncertainty in long-term decision-making contexts. To address these limitations, this study develops direct demand models using 2022 Annual Average Daily Trail Traffic (AADTT) derived from monitoring data collected using a coordinated monitoring program in the Cincinnati metropolitan area. Ordinary least squares, geographically weighted regression, and multiscale geographically weighted regression (MGWR) are applied to examine the relationships between trail use and its correlates. Model performance is evaluated using spatial and temporal validation. The results reveal pronounced spatial heterogeneity and scale effects in the influences of the built environment, natural landscape, and sociodemographic characteristics on AADTT. By capturing these features, MGWR achieves superior model fit and higher predictive accuracy than the alternative approaches. Temporal validation further shows that prediction errors increase as the time gap between model calibration and application widens, and that disruptive events, such as the COVID-19 pandemic, substantially intensify this pattern. These findings highlight the importance of applying local regression techniques in direct demand modeling and of adaptively updating models to better reflect future scenarios, supporting effective and resource-efficient trail planning and management under fiscal constraints.
No takes yet. Share an insight, caveat, or question.
Chen et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: