Modelling spatially sampled proportion processes. (English) Zbl 1462.62705
Summary: Many ecological processes are measured as proportions and are spatially sampled. In all these cases the standard procedure has long been the transformation of propor tional data with the arcsine square root or logit transformation, without considering the spatial correlation in any way. This paper presents a robust regression model to analyse this kind of data using a beta regression and including a spatially correlated term within the Bayesian framework. As a practical example, we apply the proposed approach to a spatio-temporally sampled fishery discard dataset.
MSC:
62P12 | Applications of statistics to environmental and related topics |
62J12 | Generalized linear models (logistic models) |