Modelling the distribution of cetaceans using opportunistic presence-only data ...

• Author(s): Fernandez, Marc ; Azevedo, José M. N.
• Date: 2011
• Persistent ID: http://hdl.handle.net/10400.3/3041
• Origin: Repositório da Universidade dos Açores
• Subject(s): Base de Dados; Monitorização de Cetáceos; MONICET database
• Document Type: Conference Object
• Language: English

Description

25th Annual Conference of the European Cetacean Society, Cadiz, Spain 21-23 March 2011. The distribution of cetaceans has always been a difficult and polemic issue, and more so in open waters and pelagic habitats. Due to the high economic cost of surveying and the low detectability of certain cetacean species, the distribution of many cetacean species remains poorly known in many areas. In this work we explore the potential of Spatial Distribution Modelling using presence-only models to obtain a potential distribution map of cetacean populations from opportunistic data. We used opportunistic data from the MONICET database, provided by commercial whale-watching operations in the Azores using a standard, validated, methodology. Two different presence-only models were applied, the Genetic Algorithm for Rule-Set Production (GARP) and the Maximum Entropy Modelling (MaxEnt). Using different GIS programs we prepared a set of different bathimetric derived ecogeographical variables: Depth, Slope, Rugosity and Bathimetric Position Index, using the Bathimetric Terrain Modeller Extension for ArcGIS to calculate the last two. A comparison of the performance of the two different models was carried out, using the Area Under the Curve (AUC) of the Receiver Operator Curve (ROC) and expert knowledge as key factors to see which one of the models produce more realistic results. The relevance of each variable in the models was evaluated using a Jacknife test. The results of this study could be used in a future for a conservation and management purposes in the Azorean waters.