Information gain and loss between masterlists and intermediary-level protocols for the sampling of beached macrolitter

The compared performance of citizen scientists with respect to professional scientists is still an open debate. In particular, marine litter was found to be a topic where relevant inputs proceed from citizen scientists. Here we compared two protocols for marine litter items and material identification, i.e. the OSPAR masterlist and its intermediary version adapted to citizen actions by NGOs, in terms of information structuring (Principal Component Analysis and Normalized Entropy Criterion) and correlation with beach ecology variables (Random Forest regression). Differences between results from different protocols were found related to two main causes: a) the sporadic occurrence of less common items only present in the OSPAR masterlist, and b) the engagement of citizens in specific campaigns proposed by NGOs. Models from both protocols returned high percentages of variability explained. In particular, those referring to the general categories of “materials” resulted best performing than those referring to the more detailed “items” categories. Among ecological variables, beach width was confirmed the most informative one, outperforming exposure and season (at least for the case study considered). This matches with the vision backing the Coast Cleanliness Index, proposed to relate marine litter to beach available area.