Capabilities of Lidar- and Satellite Data in Assessing the Drivers of Avian Diversity in a Fragmented Landscape

In modern landscapes, small habitat patches such as woodlands isolated in an agricultural matrix, can be important refuges for wildlife. However, their value as habitat may be compromised by their size and thus knowledge of how habitat structure influences habitat quality is vital to maximize species diversity. This study examined the factors driving avian diversity in four small woods in an agricultural landscape, and how accurately remote sensing (RS) metrics were able to quantify this. Linear mixed-effect models were used to combine annual breeding bird census data with data of habitat structure from satellite images and airborne lidar. The aims were firstly to examine the drivers of bird diversity, and secondly to reveal the strengths and weaknesses of the compared RS datasets in quantifying them. The results showed that, at first, bird diversity increased significantly towards the edges, being driven in part by vegetation structure. The amount of understorey vegetation was the most significant driver of diversity, due to which lidar-based models outperformed satellite-based ones. In general, lidar metrics correlated strongly with bird diversity, but such relationships were not discovered with satellite image metrics. The results indicate that the drivers of diversity, especially in fragmented woodlands may be too fine-scaled to be studied without sufficient consideration of the structural component of vegetation, which was proven to be attainable from lidar data.