The proposed project has two overlapping scopes. The first is livestock epidemiological modeling of infectious transboundary animal diseases (TADs)—a major threat to the agricultural system, food security and the economy. Efforts to understand spatiotemporal disease dynamics and evaluating the effect of response actions to control outbreaks are greatly aided by mathematical modeling. For this we have developed the Animal Movement Model (AMM) and the Disease Outbreak Simulation (DOS). AMM is a Bayesian hierarchical model that uses incomplete and patchy samples of livestock shipments in a Markov Chain Monte Carlo algorithm to predict complete cattle shipment networks in space and time. DOS uses a density-dependent kernel parameterized from published disease outbreak data together with shipments from AMM to simulate transmission of TADs in agricultural systems. Additionally, we have developed a specialized Bayesian hierarchical state-space model for bovine tuberculosis (bTB) that fits parameters of a within-herd bTB model to multi-year test data while taking long-term herd population dynamics into account. The model is informed by data from an ongoing bTB-eradication program in Michigan, USA, where bTB is a large problem. The second scope of the project is the modeling of Swedish ungulate wildlife populations such as wild boar, fallow deer, and roe deer. Using hunting reports and ungulate-vehicle collision data, we build biologically-informed population models of these species to infer their population size at the current moment of time, as well as to predict their future population growth. The estimation is carried out using Bayesian state-space models across all of Sweden, taking spatiotemporal autocorrelation into account to borrow statistical strength across county borders. All models are continuously refined and expanded on to include new features, specific scenarios, and additional data in order to facilitate more detailed assessments, predictions, and application to a wider range of diseases and species. Because of a recent outbreak of Highly Pathogenic Avian Influenza in the US, we will primarily focus our work with AMM and DOS on this emergent disease. The models we work with are very computationally demanding—either through the sheer number of simulation replicates that is needed for reliable results (DOS), or because of the models’ inherent complexity (AMM, bTB). Our primary focus for the TAD modeling have been the USA with a very large livestock population, and with both AMM and DOS working at the resolution of the individual herd the computational tasks can easily become massive. The same holds for wildlife population modeling: the model’s spatiotemporally autocorrelated Bayesian state-space structure means that a single model fit takes around 20 hours. Additional species will be included next, which will extend the dimensionality and computation time further. Having access to the supercomputer resources that are provided by NAISS is absolutely critical for the continuation of these research projects. Because the group and scope continues to grow, we ask that computational needs for future projects will not be solely based on previous utilization to avoid bottlenecks and challenges to meet tight deadlines.