Forecasting daily foot traffic in recreational trails using machine learning

Abstract

This paper discusses weather factors that may affect the level of visitation at recreational walking trails and provides insights into how specific factors (wind, rain etc.) can influence visitation. The quantity of visitors received affects trail management strategies, as there are often damaging effects attributed to the excessive visitation of natural areas. Therefore, accurate forecasting can inform trail management plans. Trail partners have expressed a demand for a system that can deliver qualitative insights to inform trail management while also providing accurate visitor forecasts. This study applied the approach, utilising Machine Learning and historic footfall data from electronic people-counting sensors alongside weather data; our model is a first in the introduction of Tourism Climate Indexes into forecasting models. Factors influencing visitation levels at three walking trails across the Atlantic Area of Europe were discussed. The results highlight that the model predicts trail use with satisfactory accuracy to inform adaptive management frameworks measuring visitor experience indicators.

Management implications

• •

  Environmental monitoring can gather insights into the situational factors that affect visitation levels on their trails, or if there are other contributing factors aside from weather data that could be investigated.

• •

  Trail-related recreation operators can formulate and develop strategies and plans to prevent the occurrence of tourist crowding or congestion in periods of high demand and increase trail visitor arrivals in low demand.

• •

  Trail managers can develop new service that will attract visitors under different weather conditions such as shelters, indoor museums, tents that hosts visitors during rainy or sunny days.

• •

  Trail managers can prepare for a lower trail visitation demand through marketing and offering alternative recreational activities.