Trails

The three agencies executing the Joint Powers Agreement (Department of Natural Resources, Greater Minnesota Regional Parks and Trails Commission, and the Metropolitan Council) along with their partners (individual units, implementing agencies, etc.) manage over 180 parks and 150 trails across the state of Minnesota. Understanding visitation to parks and trails across the state of Minnesota is essential for planning, programming, investment, and fiscal allocation decisions.

Traditional estimates of park and trail visitation have relied on a combination of intercept surveys, in-field visitation counts, and automated trail counters. Within the last decade, rapid adoption of mobile devices like smartphones allows for visitation to be estimated using aggregated and anonymized location-based services (LBS) data (Figure 1).

This digital report provides an overview of conclusions reached thus far about trail visitation across the state of Minnesota. Subsequent analyses, research products, and detailed documentation will be communicated later.

Using LBS data to understand park visitation is also a component of the project. Conclusions to-date about park visitation were  presented on September 14, 2022  to Legacy Partners. Parks will not be discussed in this update. An updates on parks will occur when LBS data becomes available for 2022 (estimated March/April 2022).

This Project was funded with Legacy Partnership Research Funds from the State of Minnesota Parks and Trails Legacy Fund in collaboration with the Minnesota Department of Natural Resources, the Greater Minnesota Regional Parks and Trails Commission, and the Metropolitan Regional Park Agencies.

Theoretically, the concept for measuring trail use is identical for both traditional trail counters and LBS methods. Both rely on “gates” through which visitors pass. In traditional settings, these "gates" are physical sensors on trails (Figure 2); LBS data detects users passing through geographic gates. This theoretical alignment means that neither traditional nor LBS approaches can identify the number of unique visitors along a trail’s entire length. For instance, visitors who do “out-and-back” trips will be counted twice (passing through the same gate once in each direction).

LBS data provide an opportunity to measure trails more completely than is possible with traditional methods. Each trail typically consists of multiple segments split by roads or other intersecting trails/paths. LBS data can measure use along any of these standardized, mapped trail segments. Conversely, traditional trail counters are typically placed in single strategic locations along a trail.

Trail geography files were obtained from each agency.  DNR trails  and  Metro regional trails  were accessed from the Minnesota Geospatial Data Commons on 18 October 2022. Greater Minnesota files were obtained via ArCGIS online in November 2022.

Trails were decomposed into  OpenStreetMap  (OSM) trail segments (Figure 3). OSM is a widely-used mapping service used by StreetLight (the LBS data provider) and other private companies and government sectors around the world (i.e., Amazon, Apple, Microsoft, ESRI).

Both bike and pedestrian use (“volume”) were measured for each trail segment. Specifically, we use the StreetLight API to run a zone activity analysis for both bike and pedestrian metrics for each trail segment in each month in the study period (January 2019 - December 2021). The resulting data consists of 211,608 observations (2,939 trail segments x 3 years x 12 months x 2 travel modes).

StreetLight captures significantly more data than is feasible using traditional methods. By collecting data continuously each month, StreetLight produces estimates with more temporal detail than methods such as trail counters generally provide.

In addition to temporal detail, StreetLight captures use estimates for trails at the segment level. Each trail is split into several OSM segments. This means that, instead of one trail-level estimate, we can explore data for each of these smaller subsections. Across the three systems, there are nearly 3,000 unique trail segments (Figure 4). Segments are generally shorter in more urban areas.

The Minnesota Department of Transportation (MnDOT) has a number of  trail counters deployed across the state  which provide daily measures of bicycle and pedestrian activity, either by direct measurement or by imputation. Ten MnDOT trail counter sites overlap regional or state trails in the project sample. Average monthly bicycle and pedestrian counts for weekdays and weekends between 2019 and 2021 were compared at these sites.

At most MnDOT sites, LBS data validates well (Figure 5). Some differences in estimates are expected because StreetLight and trail counter data reflect slightly different geographic areas (OSM segment versus physical trail counter location). The overall Pearson's correlation between trail counter and LBS data at all ten validation sites is 0.89, indicating a strong positive relationship.

This validation confirms that LBS data can be used to calculate viable use estimates at a variety of trail types. Analysis and validation continues on many aspects of this project.

In a few instances, trail use is better measured using the methodology developed for park visitation. This is true for trails which are not associated with OSM data and for a few geographically consolidated mountain bike trail systems. These trails will be processed and analyzed with parks in early 2023. In other instances, OSM data may not yet reflect on-the-ground conditions for newly established trails. It is beyond the scope of this project to ensure that current trail conditions are reflected in global mapping data.

Trails connect Minnesotans to a wide range of land uses, essential services, cities, and other attractions. In addition to the recreation value of trails, trails are important components of transportation networks, connect park systems, and advance natural resource conservation efforts across the state. Due to the variety of landscapes and diverse services trail provide, not all segments of trails are used equally (Figure 6).

Portions of trails which are located near cities or densely-populated areas generally have higher use. Trail use is generally highest in the summer with a larger share of bikers. Winter months tend to see lower average use with walkers representing a larger percentage of users. Walkers also represent a larger percentage of users in more urban or residential areas.

This data can be explored for each trail and trail segment. In the interactive figure below (Figure 6), select a trail of interest to see a time series of trail use from January 2019 - December 2021, a map of activity by trail segment, a time series of mode share, and a summary table. Hover over any element and/or zoom in for additional detail.

Trail use is generally highest in the summer and lowest in the winter (Figure 6). During summer months, bikers generally make up the majority of trail users while pedestrians make up a larger percentage of users in the winter. Trails in urban areas are generally less sensitive to seasonal differences (i.e. Bruce Vento versus Gateway), likely due to a combination of trail maintenance (i.e., snow clearing) and trip purposes (i.e., short utilitarian trips versus long recreation trips).

Some trails have clear peaks of activity throughout the day (Figure 7). Notably, some trails see peaks in use around 9am and 5pm on weekdays, reflecting traditional commuting patterns. These patterns are generally more evident for bicycle use, except for in very walkable and/or urban areas. Trails with primarily recreational use have a less prominent peak in use which occurs around midday. Weekend use tends to be more consistent throughout the day with some morning peaks in activity.

Across all three park systems, many trails had increased use in 2020 (Figure 8). This was likely driven by behavior changes caused by the COVID-19 pandemic. Use in 2021 largely returned to pre-pandemic levels.

•  Unit-level park and trail figures 
•  Previous StreetLight updates 
•  Parks StoryMap  (last updated May 2022; results subject to change at next update)