UBC Theses and Dissertations
Forest cultural ecosystem services assessment using social media data Lingua, Federico
To achieve comprehensive sustainable forest management the entire suite of ecosystem services that forests provide must be considered. Centuries of research have equipped forest managers with numerous tools to assess the value of timber, the main focus of most management plans in many countries, while cultural ecosystem services (CES) are often neglected. This is partly due to the challenges that CES pose given the nature of non-marketable goods. The ubiquitous approach to study CES is through the use of in-situ surveys, which are expensive and time-consuming. The overarching objective of this thesis is to explore new ways to characterize CES and develop methodologies for forest managers to integrate CES into forest management plans. To this end, a novel data source is exploited throughout this dissertation: geotagged social media images, acquired by visitors to British Columbia’s forested areas and provincial parks. These images and their metadata were crowdsourced from Flickr, analyzed with machine learning and deep learning techniques, and combined with various traditional economic methods such as travel cost and benefit transfer analysis. In doing so, it was possible to characterize, map, and estimate the value of the CES provided by forested areas across British Columbia. The number of recreational visits in BC’s forests was estimated to be over 44 million per year, and the most frequently identified activities were hiking (30.2%), skiing (16.1%), and water related activities such as fishing and kayaking (13.1%). Findings suggest that crowdsourced social media images are a useful, versatile, and unique data source. If used appropriately, they have the potential to tackle the mapping and valuing issues that historically limited the inclusion of CES in forest management plans.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International