Storm Surge Forecast
With an increase in extreme storm events and sea level rise, events of coastal flooding and erosion are becoming more frequent.
There are many factors that contribute to and come as a result of storm surges – and we can’t fix all of them. However thanks to funding from MBIE Smart Ideas we have the opportunity to create a forecast system that people can learn and use, placing the power back in the hands of those who are directly affected by these events.
What is storm surge
Storm surge is a critical component of coastal flooding. This is an event where a low pressure system creates an additional water elevation on top of the tide. In the case where such an event coincides with high tidal elevation it can result in inundation. This flooding can cause damage to coastal assets such as houses, natural features, urupā, marae, roads and more.
Our goals
Our project aims to support communities and emergency managers to take specific and appropriate actions to minimise risks to personal safety around storm surge.
Outputs
The Storm Surge forecast has been developed and made available through SwellMap. In addition to the forecast, all scientific findings will be published in leading international journals. All new code will be made open source, allowing groups like local councils, private consultancies and universities to leverage in house modelling and observational assets. This will allow users to develop bespoke forecasting capabilities at scales far beyond what can be achieved at a national level.
Up until this project, Aotearoa New Zealand didn’t have a publicly accessible storm surge forecast. The previous options were limited to paying customers.
The storm surge forecast that has been made freely available through this research is updated every 6 hours thanks to machine learning algorithms developed by the team. It is also free, easily accessible and doesn’t require any input from the user.
Using the forecast
Currently the forecast is hosted on SwellMap, and can be found under the ‘Map’ > ‘Ocean Layers' > ‘Storm surge / sea level height (without tide)’.
The forecast shows a national model of Aotearoa New Zealand, and a scale showing blue which means the surge is negative (either there is no surge or the water level is decreased compared to tidal elevation) and red at the other end for a higher sea level height (and a higher risk of storm surge).
(Without tide) means the model does not factor in tides. The astronomical part of the tide is solely related to the motion of the earth, the moon, the sun and other astral bodies and is predictable accurately over very long time scale using fast models based on harmonic analysis.. The model shows the user the risk of storm surge, regardless of what the tides are.
How accurate is the forecast?
The forecast has been developed over the past year and a half, with a team of ocean data and modelling experts. To help visualise the accuracy of the forecast, the graphic below shows some model validation methods that help to verify the forecast.
Historical Storm Surges are shown in the maps on the far left. These are real events that have happened in the past and are used as a ‘test’ to see whether the model outputs match what was actually observed at the time.
Time Series are the graphs in the middle of the graphic below. The two lines represent the model predictions (green dotted line) compared to the numerical model data - observed data (black line). This gives us an insight on the performance of the model.
Q-Q plots are shown below on the far right of the graphic. Each point of the scatter plot represents a single event (value of the storm surge on the horizontal axis) and the correspond prediction (vertical axis). The dashed diagonal line on the graph represents when the predicted values are equal to the truth meaning that any point below the line represents an event for which the model over-estimates the event and conversely.
The closer the coloured dots are to the line the more accurate the forecast is. The various plots shown below are examples of the forecast data plotted against actual historical storm surge events in Aotearoa New Zealand.
Historical Storm Surges
Time Series
Q-Q PLot
What’s next?
The addition of a ‘beach safety forecast’
Currently when viewing the storm surge forecast, the user can click on any area and see a hyperlocal ocean forecast. To be added to this option is a ‘beach safety forecast’ specific to coastal hazard including the risks of storm surge.
Visual edits to the enhance the accessibility of the forecast
Changes to the forecast page will be implemented to ensure it is accessible to the forecast viewers. This includes things like changing the colour scale to be more visually appropriate and adding instructional messaging.
Model code to be made available online
All new code will be made open source, allowing groups like local councils, private consultancies and universities to leverage in house modelling and observational assets. This will allow users to develop bespoke forecasting capabilities at scales far beyond what can be achieved at a national level.
