Quantifying Sand Transport Sensitivity to Dune Shape: Field-Validated CFD with AirSketcher

Abstract

Coastal dune management often alters crest and lee geometry, yet quantifying the transport impact of small shape changes is difficult without heavy models. This paper presents a streamlined, field-anchored CFD workflow (AirSketcher) that ingests a side-profile image, auto-detects the dune outline, and computes near-surface flow. A neutral ABL powerlaw inlet (? = 0.16) is applied consistently across scenarios; turbulence is closed with a one-equation eddy-viscosity model. Model skill is established against multi-height mast measurements at Tomahawk Beach (Dunedin, NZ): height-matched profile correlations are strong (?2 ≈ 0.985, 0.968, 0.840, 0.951), and the solver reproduces stoss speed-up, crest amplification, and lee-side recovery. For design comparison, a geometry-aware transport proxy is formed as a cubic line-integral of speed along a near-surface polyline (Bagnold-type scaling). Three shapes are evaluated under the validated wind: baseline, top-cut, and back-cut. Integrated proxies (∫ ? 3 d?) are 804,446, 667,430, and 484,779 m4 s−3, implying ≈ −17% (top-cut) and ≈ −40% (back-cut) vs. baseline. Sensitivity tests show the ranking is stable to threshold choice, probe height, and roughness. Flow patterns explain the reductions crestpeak attenuation, muted lee jets, earlier reattachment—concentrating deposition nearer the crest, especially for the back-cut. The approach offers a rapid, interpretable metric for screening dune modifications before committing to fully coupled morphodynamic modeling.