We undertake detailed statistical analysis to support our projects in flooding, hydrology, coastal, meteorology, climatology, engineering and probabilistic modelling.
High-dimensional joint probability analysis
Advanced univariate and multivariate extreme value models
Multivariate regression analysis
Quantification of model uncertainty
Monte Carlo simulation for complicated processes.
These statistical assessments can be used to understand the likelihood of multiple weather and environmental events occurring at the same time. For example, what is the likelihood of rainfall occurring within 24 hours of flooding? Or the chance of extreme waves will coincide with a coastal surge and a high tide?
When used in conjunction with our probabilistic modelling and engineering assessments they allow the structural response and risk-based performance to be assessed. For example, using structural fragility we can estimate the likelihood of extreme wave overtopping causing a defence failure and inundating a community.
Detailed multivariate probability assessments are now becoming best-practice for the design of structures. Rather than design a flood levee on a 100-year rainfall event or a breakwater on a 200-year wave event, they commonly allow up to six different variables to be considered. For example, a design flood event would consider the likelihood of rainfall, flows and tailwater levels coinciding to have a 1 in 100-year probability, whilst a design coastal event would consider the combined probability of waves (height, period, direction), wind (speed, direction), tide and surge levels. These assessments give greater confidence when designing the defence Standard of Protection, ensuring structures are not over- or under-designed.
Multivariate assessment of design coastal conditions and wave tracks for breakwater design.