A modelling framework for integrating Scitizen Science data and professional surveys in Ecology: A case study on bird mortality hotspots caused by powerlines

The number of data sources related to the spatial distribution of species and events affecting biodiversity is wide. Unfortunately, not all these sources of information we have access to are collected in standardized ways. Hence, proposing models that make use of these simultaneously becomes a challenge that needs to be addressed carefully. The goal of our case study is to find hotspots of bird mortality caused by powerlines in Trøndelag County, central Norway. There are two types of information available: professional (presence/absence) surveys collected by expert scientists where carcasses were searched for with a dog under powerlines and opportunistic records (presence only) of Citizen Scientists across the region uploaded to the national portal ‘artsobservasjoner’. We propose a modeling framework that integrates both sources of information considering their different properties. This framework assumes that the different types of information available have a common underlying process, represented as a Gaussian Random Field. The framework also accounts for the spatial and systematic biases characteristic to Citizen Science data by modeling the observed point pattern as a thinned version of the true point pattern. Our modeling framework lies within the group of Latent Gaussian Models. Hence, it can be easily fitted using the INLA-SPDE approach. We test the models we propose through simulations and comparison criteria against simpler models.