High resolution survey and modelling techniques to assess the sustainability of gravel extraction on the wild Waiapu River, East Cape, New Zealand

Abstract

The Waiapu River, located on New Zealand's East Cape, is situated within a highly dynamic landscape with a combination of weak geological substrate, active tectonics and high rates of rainfall. This active setting has been primed for geomorphic change following early 20th century deforestation. Gullying and other mass wasting processes have led to the Waiapu having the highest suspended sediment yields among New Zealand rivers. The river is perceived as aggrading and unstable, but closer examination shows different parts of the river in different states of surplus, stability or deficit. With newly accelerating growth in gravel extraction activity, there is strong interest in establishing background rates of bedlaod transfer in the upper reaches of the Waiapu River. In a survey period from 15th February 2019 to 13th February 2020, four reaches within the upper Waiapu River were examined, with the uppermost and lowermost reaches under gravel extraction operations that are consented to remove a total of 110,000 m³ of gravel annually. Digital Elevation Models (DEM) derived from aerial LiDAR and UAV surveys were used to construct DEMs of Difference (DoD) in order to evaluate changes in morphology induced both by floods and by gravel extraction. While the suspended sediment load for the Waiapu River is well documented, this study looks to assess rates of bedload transfer. For this, the morphological method was used to determine bedload transport rates using the DoDs. To supplement this methodology a 1D, cross-section based model (BedloadWeb) and a 2D morphodynamic model (Delft3D) were used to estimate bedload transport potential for the inter-survey period. Short term morphological changes are found to be directly influenced from extraction pits on the active plain of the river following flood events capable of geomorphic work. These long extraction pits features redirect ow and may induce further lateral erosion. The change model suggests that an average of 34,000 m³ (1700) of material was eroded/deposited from any of the 250-m long discretised cells along the river. By routing the net change within each cell downstream to the next cell, it was found that a strong deficit in the longitudinal gradient emerges at the downstream-most extraction site, suggesting that gravel extraction is in excess of recharge. Modelled bedload transport rates suggest yields of 15,150 m³ (750) and are a reasonable minimum in equilibrium, non-aggrading conditions. Maximum estimates are notably higher (78,000 m³ 3900), but such concentrations are not the norm, and represent non-equilibrium hysteresis conditions that might apply at the onset of flooding. The bedload transport rates assessed here via the morphological method are likely to be anomalously low, given that it was an especially dry year for the Waiapu. This helps to emphasise that fluctuations in yield are probably high (+/- 150%, by the modelled estimates) and thus the gravel extraction consenting process should take this into account, perhaps offering a fixed base removal rate, with adjustment to the following year in the event of high flows in the previous year.