40-5 Comparison of UAS-Derived Streambed Grain-Size Distributions and In-Situ Wolman Pebble Counts

Session: Geomorphology (Posters)

Presenting Author:

Authors:

Abstract:

Traditional fluvial geomorphic assessments rely on in-situ field techniques such as Wolman pebble counts to characterize streambed substrate. While effective for basic assessments, these methods are labor-intensive, spatially limited, and subject to human sampling bias. This study evaluates the efficacy of uncrewed aerial system (UAS) imagery combined with object-based image analysis (OBIA) as an alternative approach for quantifying substrate grain-size distributions in low-gradient streams.

A 116-meter riffle–pool–run sequence along the East Fork of the Clarks River in Calloway County, Kentucky, was surveyed using both traditional Wolman pebble counts and UAS-derived measurements. Four geomorphically discrete sites (riffle, run, pool, and point bar) were sampled in situ using 100-count Wolman methods, using a gravelometer. The same reach was imaged using a DJI Phantom 4 UAS flown at approximately 6-meter altitude, producing approximately 800 images with sub-2 cm pixel resolution.  A 3D Orthomosaic image was created using Pix4D and OBIA was conducted in ArcGIS Pro.  The image was segmented into polygons, particle diameters and areas were calculated, and 100 clasts were randomly sampled per site to replicate Wolman sampling procedures.

Mean grain-size bin counts were statistically similar between methods; however, the cumulative distributions differ significantly (Kolmogorov-Smirnov, D = 0.51–0.69; p < 0.001). Wolman D50 values ranged from 16–22.6 mm, whereas UAS-derived D50 values ranged from 8–64 mm. The largest discrepancies occurred in wetted riffle and pool environments due to water depth and surface glare, causing polygon merging.  UAS-derived results closely matched or exceeded field precision on the point bar site.

These results demonstrate that UAS-based OBIA can effectively characterize streambed substrates in dry, shallow, or low-turbidity environments.  Potential biases in traditional Wolman sampling are suggested, based on apparent in-situ sampling bias in this project. With continued refinement of segmentation parameters and environmental constraints, UAS-based methods offer a scalable and reproducible tool for geomorphic assessments of streams.

Geological Society of America Abstracts with Programs. Vol. 58, No. 1, 2026

© Copyright 2026 The Geological Society of America (GSA), all rights reserved.