Marine litter has recently become a recognized global ecological concern, and its
distribution and impacts on deep-sea habitats are under continuous investigation.
Here we focus on marine litter data collected as a by-product of trawl fishery surveys
regularly conducted at a local scale in the Mediterranean. Litter data are multivariate,
have space-time structure, and are semi-continuous, i.e. they combine information on
occurrence and conditional-to-presence abundance. Data on potential environmental
drivers obtained by remote sensing or GIS technologies are also available with
different spatial support. The modeling strategy is based on a two-part model that
enables handling the zero-inflation problem and the spatial correlation characterizing
the data. In the spirit of multi-species distribution models, we propose to jointly infer
different litter categories in a Hurdle-model framework. The effects of potential
environmental drivers and shared spatial effects linking abundances and probabilities
of occurrences of litter categories are implemented via the SPDE approach in the
computationally efficient INLA context. Results support the possibility of better
understanding the spatio-temporal dynamics of marine litter in the study area.
8 Aprile 2022
Alessio Pollice
Dipartimento di Economia e Finanza, Università degli Studi di Bari Aldo Moro