The human-bear conflict radar: An operational ecological digital twin for forecasting and monitoring conflict between humans and European brown bears (Ursus arctos) in Bulgaria

Ecological digital twins have great potential to support decision-making in dynamic systems due to their continuous updating and focus on end user requirements during creation. However, this approach is new to ecology and must be tested, preferably by application in fields where it would be most beneficial. One such field is human-wildlife conflict where events unfold rapidly and often require adaptive and fast-paced decision-making. We use the case study of human-bear conflict in the Central Balkan National Park, Bulgaria to create the first operational digital twin for monitoring and responding to human-wildlife conflict. The Human-Bear Conflict Radar is a digital twin designed to forecast conflict risk and bear dispersal following conflict events. Continuous data is provided in the form of near real-time records of human-bear conflict events entered into a data collection app which is used to update the digital twin. Monthly conflict risk forecasts are supplemented by local short-term (in the order of hours) conflict risk forecasts informed by a bear dispersal model following records of conflict events. This dispersal model is also used to display nowcasts of bear dispersal following conflict events and to determine whether these events are likely to be caused by a new or known, potentially habituated, bear. We present the data requirements, models and workflow for this digital twin alongside the open source code to encourage further use of this tool and to provide a framework for future digital twin development in this field.