Euring2023 integrated modelling workshop

Video recordings of the integrated modelling workshop at Euring 2023 conference.

Instructors: Marc Kéry, Michael Schaub, Nicolas Strebel

Data integration has been one of the dominant trends in ecological statistics during the last 30 years. By data integration we mean the use of different data types that are informative about the same underlying process in the form of a joint likelihood. Such data integration in an integrated model (IM) makes very much sense ("use all the available information in an estimation task"), typically leads to more precise parameter estimates, and sometimes lets one estimate a parameter that is not estimable from each data set alone. Among the audience of the EURING analytical meetings, integrated population models (IPMs) may be the best-known instance of an IM, but exactly the same data integration principles apply also for a large number of IMs that have been described in the species distribution modeling (SDM) field, e.g., for occupancy and N-mixture models.
We present the principles of data integration that underlie all IMs and then provide illustrations from the SDM and IPM fields using JAGS. Participants will walk away with a good understanding of what IMs are, pointers to further resources, why we build IMs and how, and with a stock of sample R and JAGS code for some typical IMs that they may adapt for their own projects.