As the impacts of climate change continue to mount, it is crucial for ecologists and policymakers to understand how climate influences the links between ecosystems and their social and management context. Surprisingly, despite significant effort to quantify the impacts of changing phenology on ecosystems, there has been little effort to simultaneously measure how changes in climate drive changes in the phenology of natural resources and the people that use those resources. At least in part, this gap reflects the difficulty of collecting social and ecological data at comparable temporal and spatial scales as well as the difficulty of reconciling research methodologies between the natural and social sciences. We overcame those challenges by tracking both human visitor behavior and wildflower phenology using a large, field-validated dataset derived from the Flickr photo sharing service and volunteer citizen scientists. We use this unique dataset to measure how climate influences the spatial and temporal match between human visitors and seasonal displays of subalpine wildflowers at Mt. Rainier National Park (Washington, USA) from 2009 to 2015.
Results/Conclusions
We show that the phenological match between these ecological and social systems was highly sensitive to the date of seasonal snow disappearance during our study. Early snow melts, comparable to predicted mean conditions in the late 21st century, caused dramatically reduced temporal overlap between wildflower blooms and park visitors. We also show that climate-driven phenological mismatch was greatest in regions of Mt. Rainier National Park where vehicle access is restricted seasonally, and in regions where wildflower phenology was most sensitive to snow melt. This indicates that variations in both the ecological response to climate, and constraints on visitor behavior have strong and direct influences on the potential for social-ecological mismatch in phenology under climate change in this system. In-line with ecological and social theory, we expect social-ecological mismatches in phenology to be common in systems where users of natural resources have limited information about climatic drivers of their resource, and where seasonal shifts in their behavior is constrained by non-climatic factors. Recent dramatic growth in the volume of publicly-available georeferenced photographs coupled with recent advances in computer vision and machine learning, may soon make it possible to to test this hypothesis at very large spatial scales.