Across the globe, floods are the most damaging natural hazard both in terms of the number of people affected and the economic loss. This is also the case within New Zealand due to many towns and cities being located on floodplains that are at risk of inundation (Rouse, 2012). Flood levees (“stopbanks” in New Zealand) aim to prevent inundation within these communities. The maintenance of the stopbanks is generally the responsibility of the local and regional councils. However, along with council- maintained stopbanks, numerous privately-owned or unconsented stopbank structures exist.

The Tasman District Council recognises this and acknowledges that several privately-owned or undocumented stopbanks exist within their jurisdiction. These undocumented stopbanks are not part of the Council’s assets meaning there is limited documentation of their physical properties, such dimensions, age, purpose and flood design capacity. As a result of this, the undocumented stopbanks in the Tasman region currently have an unassessed impact on flood routing. This is a knowledge gap which this study aimed to address.

A condition assessment was conducted and found that the undocumented stopbanks had more woody vegetation growing in and around the stopbanks as well as more voids and surface damage than the Council stopbanks. The undocumented stopbanks were also grazed by livestock. Both woody vegetation and cattle grazing on stopbanks are recommended against in national (and international) guidelines.

To assess the effects of the undocumented stopbanks on inundation extent, a computational flood model was developed using the HEC-RAS hydraulic routing software, with the primary topographic boundary condition obtained from high resolution airborne LiDAR data. The flow boundary conditions were located at gauges (that record stage and flow measurements) maintained by the Tasman District Council. There were several significant ungauged catchments within the model domain, flows from which were estimated using the Ministry of Works TM61 method. The model floodplain and channel roughness values were calibrated and validated against historic flood maps based on comparisons of inundation extent. In the analysis, the model was used in scenarios where the terrain was modified to represent the removal and maintenance of both Council and undocumented stopbanks.

The Council stopbanks were originally designed to protect against a 50-year flood event and maintain 0.60 m of freeboard. However, due to natural processes and gravel extraction, the stream bed has degraded significantly since the stopbanks were built.

Council stopbanks within the model were able to prevent inundation of the adjacent land where they were present, with no occurrences of overtopping. In a 100-year event simulation, the council stopbanks were able to maintain approximately a one metre of freeboard. The modelled Council stopbanks protected the largest area of the stopbanks assessed, approximately 4.2 km² in a 5-year event.