Low-lying coral reef-lined islands and coasts are vulnerable to marine flooding. These events are a result of a combination of sea surface elevation and wave height, and can cause severe flooding even on windless, sunny days. Increasing resilience of coastal communities will be a key to the continued inhabitance of these islands and coastal areas in the near future.
Early warning systems (EWS) can be a key tool to increasing resilience of coastal communities, but current systems are either too computationally expensive and data hungry to be efficiently set up around the world, or neglect essential physics and underestimate the potential flooding.
On 5-7 February 2018 in Honolulu, Hawaii, scientists from the US National Oceanic and Atmospheric Administration (NOAA), the US Geological Survey (USGS), the Australian Commonwealth Scientific and Industrial Research Organisation (CSIRO) and Deltares (Netherlands) organised an international workshop titled Understanding Flooring on Reef-lined Islands Coasts. A group of 30 experts, representing 15 agencies and research institutes from around the globe, attended to discuss key issues related to the vulnerability of island states to wave-driven flooding due to climate change and sea level rise.
During the workshop a number of quite sophisticated efforts to develop wave-driven flood forecasts were identified. These models are very costly to implement and do not lend themselves to being easily implemented more broadly. Some of these models have been implemented on a small number of locations, e.g. one in the Marshall Islands covering the eastern end of Majuro, one in Fiji and a number of models that have been implemented around the Hawaiian Islands. There are other models either planned or under development. All of these models are very site specific and only provide forecasts for a small strip of coastline.
Following the workshop, it was clear that the science and knowledge already existed to develop a simplified, less computationally expensive, Early Warning System for forecasting wave-driven flood events along coral reef-lined coasts, that is capable of being implemented worldwide.
NOAA STAR, the USGS Pacific Coastal and Marine Science Centre and Deltares ( jointly have the tools and knowledge to develop such a simplified EWS, and have combined their capabilities to form a team to develop and deliver a simplified EWS for flood forecasting. This model is not as complex and possibly not as accurate as the sophisticated site-specific flood models mentioned above, however, it will be possible to provide global coverage for all reef-lined islands and coasts.
Aim of Project
The aim of this project is to provide all nations and people living on a coral reef-lined coast anywhere in the world with forecasts of wave-driven flood events.
The EWS will be developed in close consultation with users. There will be a range of options that will be made available to users for the derivation of forecasts and warnings. For those Nations with the capability to derive and disseminate the forecasts and warnings themselves, the project team will work closely with the relevant agencies to provide them with the ability to run the models themselves. For those users who prefer that NOAA or another agency provide a packaged product then we will work with those users to ensure that the package that is developed and run at NOAA, or another agency, is appropriate and fit for purpose.
The Wave-driven Flood-forecasting on Reef-lined Coasts Early warning system (WaveFoRCE) will provide coastal flood forecasts from the present to 180 hours (7.5 days) into the future at 3-hourly intervals, with the forecasts being updated once every 6 hours. This means that every six hours a new set of forecasts will be generated that provide the user with a set of 61 forecasts for each point along the entire coast at 200 metre intervals.
Rather than using expensive wave-rider buoys for input data to the model, NOAA have the ability to combine satellite data and hydrodynamic models to derive sea surface elevation and wave height, period and direction, globally. Deltares and USGS have developed a simple model that predicts wave-runup given the input of wave and sea surface elevation outside the reef. This simple model successfully propagates the waves across the reef (attenuating as necessary), through the lagoon and up onto the land, where the extent of the runup can be predicted.
The WaveFoRCE System
The sea surface elevation is a combination of: tides (full cycle twice per day), local sea level (which fluctuates on a daily to weekly time scale), storm surge (from near-by storms or cyclones), short term climate change (e.g. ENSO and PDO), and long term climate change (e.g. due to global warming).
Waves can be categorized into locally-derived storm waves that are generated by nearby storms or cyclones (i.e. short wavelength wind waves) and those waves that are generated by distant storms (e.g. long wavelength swell waves from storms in the Arctic and Antarctic or cyclones somewhere else in the ocean basin). Tsunamis are also a source of waves that can cause flooding, however this EWS will not produce forecasts that include tsunamis, these are covered by other forecasting systems.
Current Status of WaveFoRCE
The current version of WaveFoRCE utilizes existing satellite and modeled wave and sea surface height forecasts as inputs. It combines them with bathymetry and remotely-sensed reef properties, and applies simplified assumptions about friction coefficients, shallow water bathymetric profiles, and wave shadowing effects in a Bayesian Network developed to predict wave-driven coastal flooding on coral reef-lined coasts.
The WaveFoRCE methodology has been successfully tested at a case study site at Roi-Namur, Kwajalein Atoll, RMI. The initial simulations showed that hindcasts of wave run-up using the WaveFoRCE methodology and satellite-derived estimates of reef geometry are less accurate by 6-49% than outputs from a more sophisticated high-resolution model called XBeach. However, including an estimate of reef depth (in this case, using local knowledge of MSL of -0.5±0.5m) alongside satellite-derived reef width, increases the skill of the WaveFoRCE methodology such that the error relative to the XBeach high resolution model was reduced to just 7%.
Two further site demonstrations, one in Guam and the other in Oahu, are currently being implemented.
The success of WaveFoRCE at Roi-Namur demonstrates that a relatively simple EWS methodology does work well and is therefore ready to implement as a global forecasting system.
With the confirmation that WaveFoRCE works for the test case, the WaveFoRCE team are ready to roll it out across the world. Once we have secured funding for this phase of the WaveFoRCE development, we will take two to three years to implement it on all coral reef-lined coasts throughout the world.
During these first three years the WaveFoRCE team will engage with users and make heavy use of their input regarding the development of the WaveFoRCE output and delivery mechanisms. We are mindful of the need to ensure that both the way that WaveFoRCE output is presented to the user as well as the method of delivery are suitable for each country covered by WaveFoRCE. We do not expect that a “one size fits all” system will be acceptable and therefore wish to have users engage in this development process.
In the last two years of the project, the WaveFoRCE team expect to be improving WaveFoRCE in locations where we find that it doesn’t perform as well as we expected (most likely by improving the bathymetry for that site). We will also be switching our user interactions to more of an outreach mode where we will be engaging with users to gather validation data on the performance of the model as well as to run instructional workshops to ensure that the use of WaveFoRCE is fully understood by all key users within each country.
WaveFoRCE After Development and Rollout
After the five year development and rollout project, steps will be taken to ensure that WaveFoRCE becomes operational and therefore supported into the future. There are already a number of agencies who have expressed interest in this aspect of WaveFoRCE (e.g. NOAA and the Australian Bureau of Meteorology).
WaveFoRCE can also be run in hind-cast mode, allowing historic flood events to be documented. When combined with topographic maps, this can be useful for planning future development of infrastructure and the design of mitigation strategies for existing infrastructure.
Lastly, throughout the development of WaveFoRCE, the WaveFoRCE team will maintain close cooperative relations with other wave-driven flood modelers via the UFORIC Working Group.
User and Stakeholder Involvement
The WaveFoRCE team are looking forward to building an extensive user and stakeholder network and we will be involving them in the design of the outputs and the collection of information relating to flood events. So, if you are interested to become involved with the WaveFoRCE project or just want to know more about WaveFoRCE, please send us an email and we will be in touch.