Tracking tropical cyclones to improve risk assessment

Although cyclone seasons are well-known, being able to predict precisely where and how severely these storms are likely to affect a particular location is much harder to calculate. Cyclones are erratic weather phenomena, and the historical data available is often not sufficiently detailed to be able to base future predictions on. This is especially true in parts of the world where the record-keeping of tropical storms is relatively recent. But what if you could access information on thousands of years of cyclones? HR Wallingford scientists have developed a method to expand the dataset for specific sites, allowing for much more precise modelling of likely future cyclone events.

A tropical cyclone is an intense circular storm that originates over warm tropical oceans and is fuelled by water with a temperature above 26.5°C. The characteristic and damaging features of tropical storms, are low atmospheric pressure with high winds and heavy rain, resulting in large waves and destructive surges at coasts.

Low pressure at the centre of the cyclone causes the sea level to bulge, and strong cyclone winds push the water ahead. Sites with a wide continental shelf experience a larger surge than those with a narrow shelf. Shallow water and constrictions such as estuaries and bays also enhance the wind-driven surge.  But the effects can be very localised. Only specific cyclone tracks cause a significant surge, and calculating wind fields is crucial for accurate wave prediction.

Dr Stephen Grey, Principal Marine Scientist in the Coasts and Oceans Group at HR Wallingford, said: “Our aim was to find a way to improve the methodology for predicting the occurrence of severe tropical storms at a given location. Having identified sources of historical cyclone track data, we set out to develop tools to extract, analyse and quality control this data. We then investigated methods to simulate cyclone wind fields, and to investigate and validate the modelling of cyclone induced surge and waves.”

Cyclones were modelled using TELEMAC-2D and SWAN open-source software to predict surge and wave conditions at the site for each cyclone.  

Dr Ye Liu, Principal Statistician in HR Wallingford's Flood and Water Management Group, explained: “One of the main challenges we face when modelling cyclone prediction is a lack of data because the historic cyclone tracks for which data are available are often too few. We needed to devise a method to create synthetic cyclone tracks to increase the number of events from which to predict extreme conditions. Crucially, the simulated tracks had to be realistic and statistically valid, but they also needed to include rare but possible events like Cyclone Gonu which occurred in the Gulf of Oman and headed towards Iran, where these types of storm are virtually unheard of.” 

Dr Grey added: “This new Probabilistic Cyclone Modelling Tool which uses statistical methods, in combination with surge and wave modelling, will improve the assessment of cyclone risk at specific locations. This can then be used to inform the design of coastal infrastructure such as seawalls and breakwaters to match the conditions of a specific site, and thus avoid the danger of either under or overdesign which can be equally risky and costly.”

A paper - ‘A probabilistic approach to tropical cyclone modelling’ – will be presented at the Conference on Ocean, Offshore and Arctic Engineering (OMAE) in Glasgow, Scotland, in June 2019.