Where have all the people gone? Enhancing global conservation using night lights and social media.

Abstract

Conservation prioritization at large scales is complex, combining biological, environmental, and social factors. While conservation scientists now more often aim to incorporate human-related factors, a critical yet unquantified challenge remains: to identify which areas people use for recreation outside urban centers. To address this gap in applied ecology and conservation, we developed a novel approach for quantifying human presence beyond populated areas by combining social media "big data" and remote sensing tools. We used data from the Flickr photo-sharing website as a surrogate for identifying spatial variation in visitation globally, and complemented this estimate with spatially explicit information on stable night lights between 2004 and 2012, used as a proxy for identifying urban and industrial centers. Natural and seminatural areas attracting visitors were defined as areas both highly photographed and non-lit. The number of Flickr photographers within protected areas was found to be a reliable surrogate for estimating visitor numbers as confirmed by local authority censuses (r = 0.8). Half of all visitors' photos taken in protected areas originated from under 1% of all protected areas on Earth (250 of -27 000). The most photographed protected areas globally included Yosemite and Yellowstone National Parks (USA), and the Lake and Peak Districts (UK). Factors explaining the spatial variation in protected areas Flickr photo coverage included their type (e.g., UNESCO World Heritage sites have higher visitation) and accessibility to roads and trails. Using this approach, we identified photography hotspots, which draw many visitors and are also unlit (i.e., are located outside urban centers), but currently remain largely unprotected, such as Brazil's Pantanal and Bolivia's Salar de Uyuni. The integrated big data approach developed here demonstrates the benefits of combining remote sensing sources and novel geo-tagged and crowd-sourced information from social media in future efforts to identify spatial conservation gaps and pressures in real time, and their spatial and temporal variation globally.