Bayesian design methods for improving the effectiveness of ecosystem monitoring

• Published in: Environmental and ecological statistics 2024-12, Vol.31 (4), p.893-919
• Main Authors: Thilan, A. W. L. Pubudu, Peterson, Erin, Menéndez, Patricia, Caley, Julian, Drovandi, Christopher, Mellin, Camille, McGree, James
• Format: Article
• Language: English
• Subjects: Adaptive sampling; Australia; Barrier reefs; Bayesian analysis; Bayesian theory; Biomedical and Life Sciences; Chemistry and Earth Sciences; Computer Science; corals; cost effectiveness; Data collection; Design; Design techniques; Ecology; Ecosystem disturbance; ecosystems; Effectiveness; Environmental monitoring; Great Barrier Reef; Health Sciences; Life Sciences; Math. Appl. in Environmental Science; Medicine; Monitoring; Physics; Probability theory; Spatial data; Spatial discrimination; Spatial resolution; Spatial variations; Statistical models; Statistics for Engineering; Statistics for Life Sciences; Surveys; Theoretical Ecology/Statistics
• ISSN: 1352-8505; 1573-3009
• DOI: 10.1007/s10651-024-00623-9

Adaptive design methods can be used to make changes to survey designs in ecosystem monitoring to ensure that informative data are collected in an ongoing, cost-effective, and flexible manner. Such methods are of particular benefit in environmental monitoring as such monitoring is often very costly and in many cases consists of only a few sampling sites from which inference about a larger geographical region is needed. In addition, ecological processes are continuously changing, and monitoring programs must account for both known and unknown drivers, so making changes to data collection plans over time may be needed based on the current state and understanding of the process of interest. Through considering a Long-term Monitoring Program of Australia’s Great Barrier Reef, this paper aims to develop adaptive design approaches to efficiently monitor coral health through the consideration of a statistical model that accounts for both spatial variability and time-varying disturbance patterns. In particular, to develop this model, we considered time-varying disturbance data that have been reproduced at a fine spatial resolution for uniform representation over the study region. By adopting our proposed approach, we show that adaptive designs are able to significantly reduce survey effort while still remaining effective in, for example, quantifying the effects of different environmental disturbances.