Investigation of ecologically relevant wind patterns on Marion Island using Computational Fluid Dynamics and measured data

A need was identified to have reliable quantitative wind pattern data for Marion Island (MI) in the study of terrestrial ecology (fauna and flora) on this well-preserved and remote territory in the sub-Antarctic Indian Ocean. The present work aims to estimate a high-resolution map of wind speed, direction, and turbulence on MI through the use of Computational Fluid Dynamics (CFD); an engineering-based discipline for modelling fluid physics. This paper presents the comparison between experimental measurements and numerical simulation results of the full-scale atmospheric flow around MI and its closely neighbouring Prince Edward Island (PEI). In particular, a Reynolds Averaged Navier-Stokes (RANS) approach was used with a k-epsilon turbulence closure model adapted to atmospheric boundary layer flows. This was implemented in the ANSYS FLUENT 2019R3 software and validated against wind measurements recorded over the span of two years. A 26.9% mean absolute error was seen when predicting wind speed, a 14.8 degrees error when predicting wind direction, and a 32.6% error when predicting turbulent kinetic energy. The Baseline CFD model provides future researchers on MI with a reliable means of estimating wind properties near ground level. Quantities such as average wind speed, turbulence, gusting and annual wind behaviour can now be recreated at any location on the island; something that was not previously possible without environmentally intrusive surveillance methods. This research has refined CFD modelling for islands and highlights the potential for the application of this methodology to estimate spatial variation in wind parameters at biologically-relevant scales on other islands.