Search

Many coastal areas currently at risk of flooding in New Zealand will in the not-too-distant future suffer increased intensity and frequency of Climate Change-Related Flooding Hazards (CCRFH) (Church et al., 2013; Woodruff et al., 2013). STRAND is a (Royal Society of New Zealand) Marsden-funded project (2021-2024) that aims to forecast effects of CCRFH on residential property values. The components of the project include financial modelling, multicriteria decision analysis (Coastal Asset Risk Index: CARI) and physical modelling. The latter applies projections of sea level rise relative to terrain data and the situated properties, incorporating outputs from hydrogeological (groundwater) models, tidal (efficiency) data and precipitation forecasts. This abstract updates on progress with a GIS-based physical model of future flooding in South Dunedin (Nguyen et al, 2019). A pilot version of the model has been implemented (Moore et al, 2022), addressing flooding due to CCRFH through bathtub (SLR applied directly to the DEM relative to MHWS-MSL) and geometric (as above, but relative to a groundwater surface model and tidal efficiency) modes. For both, Monte Carlo simulation was applied to model uncertainty in the DEM, SLR scenario and groundwater surface. In addressing limitations of the pilot model, we will document progress on a thorough analysis of sources of uncertainty stated as before, and architecture choice (climate change scenarios, inundation models, zero baselines and spatial resolutions), prior to a comprehensive model run using Nesi high-performance computing facilities. This phase is applied to the South Dunedin scenario only, with the optimal parameters for uncertainty modelling a architectures tested being used for Christchurch and Tauranga scenarios, along with precipitation model output.

Have financial businesses changed their behaviour in the aftermath of global financial crisis? We address this question by introducing a new and more parsimonious method to quantify the level of financial misconduct and apply this to financial offences between 2004 and 2016. This exercise allows us to investigate whether Capture-Recapture methods can be deployed to handle problems of partial observability and how they compare to previous methods set out to achieve the same goal. In our two stage approach, first, we estimate the rate at which offending businesses are detected, then we look at how the number of detected offenders changed after 2010, and use these two layers of information to make inferences on the deterrent effect of financial regulation. Our results offer evidence that a drop in the number of detected offences post-global financial crisis was driven largely by improved deterrence.