Pablo F. Méndez, Luis Santamaría (EBD-CSIC)

The overall goal is of the modelling exercise is to better understand and predict the influence of multiple environmental and management/policy factors over the waterbird community of the protected marsh of the Doñana PA. The main interests are twofold. First, the interest lies in the ability of BNs to capture the perceptions and beliefs of different stakeholders (e.g., decision-makers vs experts) before and after existing and new information is incorporated into the BN for the modelling of new scenarios. In addition, the interest lies in the use and overall performance of BNs as a tool for ecological prediction and decision support, able to incorporate and update information from different sources, including extra expert knowledge and empirical sub-models, in a modular fashion. In this case, we considered their twofold diagnostic (bottom-up reasoning) and explanatory/predictive (top down reasoning) capability, which allows the modelling process to progress in situations of incomplete information (e.g., missing or still incipient data on key system variables) (Marcot et al. 2006, Castelletti and Soncini-Sessa 2007). We were also interested in their ability to integrate both qualitative and quantitative information (Chen and Pollino 2012). In both cases, we chose BNs due to their appeal for collaborative model development, visualization of causal relations, filling of data gaps and identification of uncertainties (Castelletti and Soncini-Sessa 2007, Kleemann et al. 2017).

For details on the case study, see Section 6.3 in ECOPOTENTIAL Deliverable 7.3