Mapping polarised sky views to the behaviour of oriented animals under clouds and trees

Abstract

Many navigating animals must be able to extract reliable directional information from visual scenes cluttered by vegetation and clouds. Building on a computational model of the insect polarised light compass, I combine behavioural experiments, polarisation and colour imaging, anatomical reconstruction and modelling to uncover how ball rolling dung beetles cope with such environmental noise. Using µCT based 3D reconstructions of individual ommatidia, I quantify spatial resolution across the visual field and simulate how eye design limits access to celestial compass cues. In parallel, I challenge the orientation of Kheper lamarcki under natural clouds and foliage in South Africa, while simultaneously recording the skylight and surrounding scene with a polarisation camera, providing a matched description of the beetles’ available visual input in the field. By integrating these skylight statistics with my anatomical model and trajectory data, I test how beetles reweight allothetic celestial information and idiothetic self motion cues when the sun and polarised light pattern are degraded. Preliminary results show that foliage has little impact on compass performance, whereas clouds strongly disrupt celestial guidance. When allothetic and idiothetic cues are in conflict, the beetles climb onto their dung ball and rotate, remapping their internal compass to the current skylight. The beetles seem to rely more heavily on celestial cues under a clear sky and more heavily on idiothetic cues under overcast skies to maintain straight paths. This work provides the first mechanistic framework that links natural skylight structure, eye anatomy and behaviour in dung beetles, and outlines general principles for robust celestial compasses that could inspire autonomous systems operating in visually variable outdoor environments.