A wave forecast built for Maldivian reefs

Raalhu is the public face of the Maldives Experimental Wave Forecast System (MEWFS), a research project that runs a nearshore wave model for Maldivian atolls every day and publishes the result as a map you can explore.

Reading the map

Tap or click anywhere on the water to read the forecast at that point. The panel shows the conditions at the selected time and how they change through the forecast day. Drag the timeline, or press play, to move through the hours. All times are Maldives time (UTC+5).

Wave height

Significant wave height (Hs): the average height of the highest third of waves. Individual waves vary, and the occasional wave can be nearly twice as high.

Mean period

The average time between wave crests (Tm01). Short periods are local wind chop. Long periods mean ocean swell, which carries far more energy than chop of the same height.

Depth contours

The thin dark lines mark the shoreline (0 m) and water depths of 2, 5 and 10 m from the model bathymetry.

Wave direction

The direction the most energetic waves come from. Streaks and arrows on the map point the way the waves travel.

Wave power

Energy carried per metre of wave crest, estimated from height and mean period with the deep-water formula. A guide to how hard waves hit a reef or shore.

Spread

How widely wave directions fan out. Low values mean clean, lined-up swell. High values mean a confused sea.

Wind and swell

The wind and open-ocean swell taken from the global forecast and fed into the model at its edges.

The model

The forecast is based on SWAN (Simulating WAves Nearshore), a third-generation spectral wave model widely used for coastal and nearshore environments. It solves the spectral action balance equation and simulates how waves travel and change across the model area.

Processes represented

• Wave propagation, shoaling and refraction
• Wave growth by wind, and whitecapping
• Depth-induced breaking
• Non-linear wave-wave interactions

Computational mesh

The Maldives has reef-fringed coastlines, narrow channels, lagoons and shallow reef platforms. To represent them, MEWFS uses unstructured triangular meshes generated with pyTri, a Python mesh generator developed for coastal modelling. Mesh density follows the bathymetry: fine around reefs, coastlines and areas of interest, coarser offshore.

Bathymetry

Depths come from the high-resolution Maldives bathymetric compilation described by Rasheed et al. (2020).

Forcing

Open-ocean swell at the model edges and the wind over the atoll are taken from the NOAA GFS-Wave global forecast (00 UTC run).

Daily workflow

1. The latest global wave forecast is downloaded.
2. Swell height, period and direction are extracted along each open edge of the atoll domain, with the wind over the atoll.
3. SWAN runs once for each forecast hour on the atoll mesh.
4. The results are resampled onto a regular grid and packed into small image files, one per hour.
5. The files are uploaded to this site, hosted on GitHub Pages. Your browser decodes them to draw the map and read values at any point.

Preparation, model runs, post-processing and publication are automated by the Raalhu modelling workflow.

Limitations

MEWFS currently runs on two Acer Veriton mini-computers. They are enough to produce experimental forecasts, but far from the computing power of an operational forecasting centre, so the model configuration is simplified to keep run times practical.

• The system is under active development.
• Each hour is computed as a separate steady-state run.
• Model resolution is limited by the available hardware, so not every lagoon, reef and shoreline process is resolved.
• Tides and currents are not included.
• Coverage is limited to selected atolls.

Validation

MEWFS is in its research and development phase. No formal verification framework is in place yet, and the forecasts have not been systematically compared with wave buoys, satellite data or long-term hindcasts. Assessment so far rests on internal testing and visual inspection.

Planned work includes wave buoy validation, satellite comparison, forecast skill assessment, error statistics and long-term hindcast evaluation. Until that is done, treat every product on this site as experimental.

Funding

MEWFS was developed through a Small Grant Research Project funded by Maldives University. The grant supported the modelling workflow, mesh generation methods, forecast automation, visualisation tools and public release of the forecasts through this site.