The Changing Gravel-bed Surface in a Mid-reach of the Aggrading Hapuku River: Using Emerging Technologies to Capture Texture Change During the Passage of a Sediment Wave

Abstract

Large, pulsed inputs of hillslope sediment delivered to gravel-bed rivers are often transported as ‘waves’ of material, modifying transport capacity as they accumulate and evacuate, altering river morphology. In particular, the alteration of bed surface texture is important, as this has direct implications for sedimentary and hydraulic processes, and ultimately in-stream habitat and flood risk. It is difficult to detect and accurately measure bed surface changes in real-world fluvial systems; however, with emerging technologies in the realm of sUAS, SfM, photosieving and roughness estimation it is now possible to rapidly quantify the textural changes associated with the passage of a sedimentary wave. The largest landslide triggered in the 7.8 Mw, 2016 Kaikōura earthquake dumped approximately 13M m³ of sediment into the Hapuku River, inducing dramatic morphological changes within the valley-confined downstream reaches. This study leverages emerging technologies to investigate subsequent changes to the grain-size distribution, effective roughness and facies unit assemblage of a site downstream of a steep canyon, during the initial build-up and migration of a wave of this pulse material. Four repeat sUAS surveys were completed between April and September, 2019, and were used to create sub-centimetre resolution SfM models. A new method for the extraction of surface grain-size distributions was then tested, pairing high-resolution orthomosaics with Purinton & Bookhagen's (2019a) edge-detection tool, PebbleCountsAuto. Effective roughness was extracted from the dense point cloud surfaces, and both measurements were used to quantify texture change and assist surface mapping of the site’s facies assemblage alterations. Changes to bed texture were also related to ongoing processes of aggradation and erosion, based upon bed elevation change revealed by coeval airborne LiDAR surveys, collected in parallel research. Over the survey period, the bed surface grain-size distribution showed a significant change, at the 95% confidence level. D50 and D84 percentiles decreased by 6.7% and 10.1% respectively between July and September 2019, as the proportion of new wave material increased. The standard deviation of the grain-size distribution also lowered by 17%, indicating the improved sorting of the surface sediment mixture as the wave deposits buried and infilled the previously heterogeneous surface gravel units. This was also reflected in the 38% decrease in site’s effective roughness and the 17% increase in the aggradation-linked facies units that characterised the surface of the wave deposit’s downstream end. Overall, the final results revealed the fining, smoothing and homogenising effects of wave arrival upon the site’s bed surface, while also presenting a real-world example for the application of emerging technologies for pulse research at the bed surface scale. The findings also emphasised the necessity of a multi-faceted approach to bed assessment, using facies assemblages to give spatial context to wave migration effects, something that is lost when considering averaged roughness and grain-size values alone.