OOS 52-1 - Integrating field ecology, remote-sensing, and data science to build a living database for biocontrol

Predicting pest outbreaks and how to reliably control them has eluded ecologists for decades. We need novel techniques, leveraging ecoinformatics, satellite data, machine learning and other innovations in data science to better understand the ecology of agricultural systems, specifically what drives pest dynamics over time. As the introduction to this session, we present a cyber-infrastructure platform and coordinated innovation network we are building to connect ecologists, remote-sensing specialists, and data scientists, funded by the USDA. Our Living Database for Biocontrol is expanding a pest control database to vastly increase information on georeferenced pest and natural enemy distributions globally; acquiring relevant life history traits data matching major pests and natural enemies represented in the pest control database; acquiring Earth observations (EO) data of vegetation and climate for georeferenced locations of insect data; and developing the software infrastructure to automate the continued acquisition and processing of insect, traits, and EO data necessary for analysis. We highlight the use of this platform to bring new insights to pest control modeling, demonstrating in one case how vegetation productivity and phenology drive pest infestations and that applying remote sensed variables to pest abundance data offers much stronger predictive ability than typical approaches considering the amount of non-crop habitat in the landscape. This presentation will serve as the public launch of our beta Living Database, which aims to equip researchers who work with growers, land managers, and other decision-makers, with the data integration and analytical tools necessary to build the science and knowledge for confronting pest control challenges.