Evaluation of GNSS-based Volunteered Geographic Information for assessing visitor spatial distribution within protected areas: A case study of the Bavarian Forest National Park, Germany

Benefit
Use
Region
Year
 Back to library  Link to study

How to cite this study

Horst, L., Taczanowska, K., Porst, F. and Arnberger, A. 2023. Evaluation of GNSS-based Volunteered Geographic Information for assessing visitor spatial distribution within protected areas: A case study of the Bavarian Forest National Park, Germany. Applied Geography 150(2023), 102825.

Overview

This study analyzes the reliability of Global Navigation Satellite System-based (GNSS) Volunteered Geographic Information (VGI) using three platforms: GPSies, Outdooractive, and Komoot. These data sources are compared with automatic visitor counters to evaluate the spatial distribution of hikers in Germany’s Bavarian Forest National Park. In a comparison of all three platforms, GPSies and Outdooractive had the strongest correlation of spatial distribution and length of hiking activities.

Relevance

This study is relevant to researchers using novel data sources to estimate visitation in natural areas. The comparison of correlations between all three VGI platforms and automatic trail counts provides valuable insight into the comparability and reliability of each platform for future researchers interested in using these programs. This study also emphasizes that VGI data should be used as a complementary resource to on-site data to improve overall visitation estimates.

Location

This study is located in Bavarian Forest National Park in southeastern Germany.

Trail Type

The 24,250-ha Bavarian Forest National Park has 350 km of hiking trails, 200 km of cycling trails, and 80 km of cross-country ski trails. The park receives nearly 1.4 million visits each year.

Purpose

The purpose of this study was to assess the comparability and reliability of each VGI platform and the automatic counter data in Bavarian Forest National Park. The research was conducted in cooperation with the Bavarian Forest National Park, which offered support, organizational help, and provision of VGI data sources.

Findings

  • Across all platforms, the average hiking length was 11.2 km. The average hiking length on GPSies was 13.9 km, 12.9 km on Outdooractive, and 9.8 km on Komoot 
  • The three highest peaks of the park were found to be highly frequented in the spatial analysis in all three data platforms. 
  • The researchers were unable to make a final conclusion on VGI data’s overall reliability since the sample size of the VGI data was too small when compared to the automatic trail count sensor data which was only available from 2013-2014. 
  • In a comparison analysis of the three platforms, the VGI data (available from 2013-2018) had the strongest correlation in number of activities between GPSies and Outdooractive. The next strongest was GPSies and Komoot followed by Komoot and Outdooractive. 
  • The strongest correlation of the spatial distribution and length of hiking tours was between GPSies and Outdooractive.
  • Using the automatic field data available from 2013-2014, there was a significant positive correlation between the field counts and the number of activities for GPSies. The correlation between automatic count data and Outdooractive was not significant, and the calculation of correlation between count data and Komoot data was not possible.

Methods

Three Global Navigation Satellite System-based (GNSS) Volunteered Geographic Information (VGI) platforms were used: GPSies, Outdooractive, and Komoot. GPSies is now a part of AllTrails, but because the study was conducted before the transition, GPSies is listed under its old name in this study. The data for each platform was downloaded and converted into point shapefiles. Data uploaded from January 1, 2013 – December 31, 2018 was used in the comparability analysis between platforms, and only public activities labeled “hiking” were selected. 402 data records were available for GPSies, 250 for Outdooractive, and 1,090 for Komoot. Activity lengths were compared among platforms by uploading the data into ArcGIS. To analyze a spatial distribution of visitors in the park, a kernel density analysis, hotspot analysis, and tracks count per raster cell was applied.

Automatic counter data was only available from April 28, 2013, to April 27, 2014, so the comparison between VGI platforms’ number of activities and counter data was only for this period. The total number of entries from all 11 counters was 250,552 and ranged from about 5,300 to 79,000 entries per site. Statistical analysis was done using SPSS while spatial analysis was done using ArcGIS. 

 

Added to library on November 27, 2023