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. A storm surge is an extreme event where low pressure systems (storms) combine with a high tides to cause coastal flooding. 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. This could reduce risk to vulnerable property and infrastructure; benefit coastal planning and mitigation measures and increase future resilience in the face of projected increases in coastal flooding events.

  • 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 option was limited to those with data science skills through an ocean model called ROMS (Regional Ocean Modelling System). The ROMS method of creating a storm surge forecast was not only inaccessible to the general public, but the forecast is only updated once per day, and is expensive to run.

The storm surge forecast that has been made available through this research is updated every minute 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 for lower sea level height (and therefore less of a storm surge risk) 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 as they are constant and do not need to be forecast. 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 how well the model performs (green dotted line) when compared to the numerical model data - observed data (black line).

  • Q-Q plots are shown below on the far right of the graphic. This is a method where a scatterplot plots two quantiles (a quantile determines how many values in a distribution are above or below a certain limit) against one another. The Q-Q plot shows us visually whether two data sets have a common distribution. The two datasets included in the plots below are the Prediction (our model forecast) on the Y axis, and the Observation (reality) on the X axis.

    The closer the coloured dots are to the line (if the distribution is common) 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 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.

Our Partners