A novel experimental solution to a very severe erosion and inundation hazard complex at Washdyke beach is presented. Erosion and inundation at Washdyke are the most severe coastal hazards in New Zealand whether measured in physical terms or in terms of the assets threatened by rapid barrier beach regression and eventual breach. The Washdyke shore was naturally eroding prior to harbour construction at Timaru but has been divorced from the (also eroding) updrift feeder coast to the south since 1879 when construction began. Thus Washdyke Beach has a zero coarse sediment input and is progressively exhausting itself as it abrades coarse sands and gravels to finer sizes and transports sediments northward and/or offshore. Reduced grain size lowers berm heights and increases washover to the backshore, so “rolling over” the beach volume and accelerating the rate of shoreline retreat. An experiment with combined beach reconstruction and renourishment was commissioned in June 1980 after extensive analysis of erosion hazards in 1978 and 1979. The experiment was monitored for 5 years and ran for 6 years to provide interim protection for an ocean sewage outfall, and to evaluate the method for the full 3km of Washdyke barrier. Beach crest heights were raised 2.0–2.5 m to cut down overtopping and washover sediment from the backshore. Washover sediment was used to fill the body of the 300 m long site before a coarse capping of river gravels was added to the beach crest. Design specifications called for 4–7% volume loss per year and a fill “life” of 16–25 years. These were met and a considerably longer fill life has been obtained (53.6 years indicated). The overall reduction in erosion rates was about 55% over 5 years with no crest retreat, no overtopping and nil washover of sediment. The neighbouring untreated coasts retreated between 11.5 and 22.5 m during the experiment. The project presents new methods of performance assessment and fill “lifetime” estimation. These use repeated profile survey data subjected to normalised excursion distance analysis. NED plots were fitted with regression models to distinguish short term “noise” in the beach related to changing sea-states from underlying erosional trends. This was done with high degrees of precision both for the treated section of shore and the “control” sites surveyed on two untreated areas, one updrift and the other downdrift of the site. Linear models reliably estimate the time trends of erosion on both the treated and untreated shores, rather than negative exponential models thought to apply to sand beach fill bodies.
Read full abstract