Landscape connectivity analysis: new metrics that account for patch quality, neighbors' attributes and robust connections.

Context: Landscape connectivity quantification is essential to achieve effective conservation of wildlife. Graph theory is a common mathematical framework for representing habitat patch networks and evaluating their connectivity. Objectives: While several graph-related indices have been used for this purpose, there is still need to evaluate landscape connectivity from different perspectives. Methods: Based on classic network indices, we developed five novel landscape connectivity indices that incorporated attributes related to the quality of the patches, the connections among them, the dispersal probabilities and the overall habitat availability. To evaluate their performance, we applied them on three ecological networks developed for three species with different traits, at different scales, and compared them with widely used landscape connectivity indices. Results: The developed indices identified, in all the three networks, critical patches that were well-connected with others, and also aggregated valuable features which could affect network quality. Some critical patches and specific network connectivity properties could be overlooked if not considering the additional dimensions of information that these advanced indices incorporate. Conclusions: We showed that the developed indices exhibited flexibility and wide applicability, offering novel insights on the evaluation of node contribution to the overall network connectivity. By integrating spatial patterns and processes, the solutions provided by the new indices could serve as an effective tool towards supporting landscape conservation planning. [ABSTRACT FROM AUTHOR]