|dc.description.abstract||Poverty Bay, on the New Zealand's northeast coastline, is a high-energy coastal embayment that receives significant quantities of fine, fluvial sediment from two river systems. Examining seabed geomechanical characteristics, sediment fluxes and morphological features in northern Poverty Bay provided the opportunity to study interactive processes, including sediment entrainment rates over the cohesively bound mixed sand/mud seabeds, and the dispersion of a turbid river plume. Three field experiments and an extensive seabed-sampling programme were undertaken, and a 2-cell numerical model was developed and validated against infilling rates of the port of Gisborne navigation channel. The implications for existing and future operations at the port of Gisborne, including the suitability of capital and maintenance dredged material for use in reclamation, have been evaluated.
Within the study site, near-bed wave orbital velocities and fluvial inputs control the spatial variability in the seabed sediment textural characteristics, with the seabed grain size decreasing with depth and distance from the mouth of the Turanganui River. A hand-held shear vane was used to quantify the in-situ shear strength of the cohesively bonded mud/sand surficial sediment, and the in-situ shear strength has been correlated (R² = 0.83) with the cumulative effect of several of the sediment textural and environmental parameters.
Near-bed sediment diffusion and entrainment was examined using an array of micro-scale inline pumps. The effect that cohesive bonding of the mixed sand/mud seabed has on retarding the entrainment rate was examined by considering both the textural and geomechanical characteristics of the seabed. No correlation was found between the ratios of estimated to predicted near bed reference concentration (ζ) and either the median grain size (d₅₀) or the ratio of percent sand to mud of the seabed. However, a good correlation was found between ζ and the normalised in-situ shear strength (𝑥) of the seabed (adjusted R² = 0.99), with the relationship of the form,
𝑙n(ζ) = -0.51-1.30 𝑥³
The conversion of the downward flux from sediment traps to a time-averaged suspended sediment concentration (SSC), and near bed reference concentration (C₀) was found to be critically dependant on the fall velocity statistic (w∫) of the grain size distribution. Due to flocculation of the near-bed suspended sediment, the dispersed equivalent w∫ distribution and median w∫ of the distribution resulted in poor predictions of C₀. Instead, an in-situ ‘effective’ w∫ distribution of the suspended sediment was developed using the measured 𝑙ₛ from individual dispersed grain size bins within the SSC profile and the predicted sediment diffusivity (εₛ). C₀ predicted using the effective w∫ distribution agreed with pumped samples and theory predicted C₀, and provided a robust method of determining the time-averaged SSC of mixed sand/mud sediment using sediment traps.
The diffusion of sediment within the near-bed SSC profile varied as a function of w∫, with relatively finer sediment exhibiting less diffusion. The mixing length (𝑙ₛ) profile was well approximated by a linear profile, with 𝑙ₛ varying as a function of elevation (z) and the von Karman constant (κ), i.e. 𝑙ₛ = κ𝚣. Within 4-8 cm of the bed 𝑙ₛ is constant and equal to 0.06 m.
β is defined as the ratio of sediment diffusivity (εₛ) to eddy viscosity (ε∫), and is commonly used as a correction factor for predictive SSC profile equations. By examining the diffusion of sediment within the near-bed SSC profile a new equation for β was obtained,
𝑙n( β ) = -0. 95 - 0.4ψ / 𝑙n(ψ)
The horizontal diffusion of sediment within the Turanganui River plume was found to be a function of the distance (x) from the mouth of the Turanganui River and a decay rate factor (Dₓ), where
Cₓ = Cₒbe ⁽⁻ ͯ/ᴰͯ ⁾
The form of the equation is similar to the Nielsen time-averaged gradient diffusion equation. Dₓ was correlated (R² = 0.92) with river discharge (φᵣᵢᵥₑᵣ), and found to exponentially decrease with increasing φᵣᵢᵥₑᵣ, where
Dₓ = 591+2850 exp(⁻ᵠʳⁱᵛᵉʳ/2.1)
The relationship suggests that the effective w∫ of the suspended sediment within the plume increases with increasing SSC, presumably due to increased flocculated particle size. Based on the developed empirical formulae the Turanganui River is predicted to discharge 7.6x10⁷ kg of fluvial derived sediment into Poverty Bay on an annual basis.
The developed relationships and formulae that describe the hydrodynamical processes within northern Poverty Bay have been implemented in a 2-cell numerical model that considers both downward and horizontal sediment fluxes. The port of Gisborne shipping channel infilling rate, as determined from hydrographic surveys and dredging records, was used to validate the model. The 2-cell numerical model accurately predicts channel infilling rates, and provides an insight into the governing hydrodynamical processes in northern Poverty Bay. The developed model represents an alternative to the implementation of a full 3-dimensional hydrodynamic model.
Hydrographic surveys and dredging records show the channel area passes through the beach nearshore zone, and during the monitoring period maintenance dredging only maintained the channel design depth along the channel leads, while significant infilling occurred near the channel edges (Δh = -1 m). Extrapolation of the results suggests an annual channel infilling rate of 171,400 m³ per year, corresponding to a maintenance dredging cost of approximately $857,000 to $1,200,000 per annum. Wave and river discharge conditions during the monitoring period were smaller than average, and more infilling is expected during stormy years.
Relocating the channel and port entrance to a site in deeper water, further from the mouth of the Turanganui River, with relatively strong seabed in-situ shear strength is identified as a means of reducing dredging costs. The material dredged from the shipping channel is suitable for use in land-based reclamation, particularly if the soil structure is improved by addition of a coarser material such as of the underlying mudstone unit. The bearing capacity of the reclamation can be improved by mixing the mud/sand with concrete to create mudcrete.||