Investigating three-dimensional mesoscale habitat complexity and its ecological implications using low-cost RGB-D sensor technology

Structural complexity, that is, spatial patterns of structural units of interest', is one of the most fundamental drivers in habitat use of animals and other functions of habitats. Assessment of habitat structural complexity (e.g. mangrove root complexity) and its ecological implications (e.g. mangroves as nursery sites) has, however, been hampered by the lack of satisfactory methods for realistically measuring and analysing field habitat structure. We demonstrate a simple low-cost approach comprising an RGB_D scanning sensor (Kinect), free and readily accessible software (Meshlab, Matlab (or Octave), Qubicle Constructor and Voxeliser) and fractal dimension to measure mangrove root structure and complexity at spatial scales relevant to juvenile fish use of the habitat. The cube-counting method was used for three-dimensional characterisation of fractal dimensions of space around aerial roots of mangroves from a fish point of view'. Space-fish relationship under different ecological situations was assessed by simulating a falling tide and using different rectangular prisms to represent varying fish allometries utilizing the space. Fractal dimension (FD) of space around roots decreased with decreasing height of the object' mimicking a falling tide. The same root system may therefore look more complex to aquatic animals as the water level decreases. We also assessed the variation in FD responding to manipulation of the density and arrangement of objects (pneumatophores). Difference in FD was generally small, due attention to replication is necessary for comparisons of habitat structural complexity using this metric. This simple low-cost approach will enable field measurements of habitat structure at spatial scales relevant to its influence on habitat use and animal behaviour, as well as allowing realistic replication of field structure for laboratory-based experiments.