RapportValidation BioScore 2.0 distribution maps for breeding birds

The aim of the Bioscore 2.0 project is to design a tool for assessing the impacts of European Community policies on biodiversity in Europe (van Hinsberg et al. 2014).

In the course of developing BioScore 2.0 the Planbureau voor de Leefomgeving (PBL) asked for a response to the following tasks and topics:

1. Determine a full multivariate model (FMM) version for all BioScore 2.0 species, that is, with all pressure factors in one single multivariate analysis. Make a validation of the FMM's, preferably on independent data if available or by a cross-validation. Provide a possibility to examine the model for non-linearities and interactions in the relationship between predictors and predictions.
2. Describe provenance and quality of the distribution data that are used for model calibration and validation.
3. Compare the results of BioScore 2.0 of the current situation with the FMM of the current situation.
4. Compare the univariate dose-response relations that were used for BioScore 2.0 with the respective partial dependence plots of the FMM for ten sample species.
5. For each of the four ecosystem types (urban, forest, other natural and agrarian), specify which of the BioScore 2.0 species are typical and characteristic for them (may be more than one type).
6. Complement the earlier delivered univariate dose-response relationships for BioScore 2.0 for ten sample species. Display the quadratic as well as the linear relationshipes, analyse and describe the differences.
7. Determine the effect of cut-off value choice for ten sample species. Describe the effects of the different choices.

In this report we pursue these questions and present the results of the analyses.