Biodiversity conservation has been slow to benefit from recent technological advances such as large scale ecological databases, citizen science observations, and machine learning algorithms. Meanwhile the effects of climate change are accelerating changes in biodiversity and dramatically affecting human health and wealth. New technologies can be used to address critical conservation needs, providing more accurate, faster and less expensive ways to measure changes in populations and threats, monitor assets, and enforce regulations if used in ways that promote human-computer partnerships.Of particular concern is sampling bias and the impacts of incomplete sampling in remote areas on species distribution model estimates. Here we created a program through Adventure Scientists Conserving Pollinator Biodiversity where citizen scientists were trained, managed, and incentivized to observe butterflies and their nectar plants in remote areas of the western United States. Observations were recorded and vetted through the iNaturalist smartphone application. We evaluated if this targeted sampling approach which changed the training and incentives for participants and provided a different framework for high value sites improved the human-computer partnership and our confidence in the species distribution models important to conservation decision-making and management.
Results/Conclusions
Adventure Scientists participants recorded 2,049 butterfly observations representing 122 verified butterfly species and 3,962 ecological associations with 645 verified nectar plant species across 203 sites in five states. We observed improvements in species distribution models for butterflies by collaborating with Adventure Scientists with a target sample approach in remote areas especially for rarer species. We also found the Adventure Scientists project approach incentivized participants for looking, not finding organisms in space and time; implemented a framework for suggesting high value sampling sites to participants; and provided participants with incentive to contribute to a community and shared problem, not an individual competition. We feel this approach improved the human-computer partnership and enhanced observation and monitoring data quality. Targeted sampling in a relatively small number of critical areas is one way to enhance data collection for biodiversity conservation using human-computer partnerships.