|
If you operate near, or on the water, having accurate information about present and future ocean conditions is crucial—the more you can anticipate, the better. At Calypso Science, we specialise in providing accurate data on ocean currents, tides, and waves. We achieve this by developing hyperlocal-resolution models tailored to our clients’ areas of interest, adapting the level of detail based on the complexity of each site. Hyperlocal resolution means our model pixels are up to ~100 times finer than standard models. With over 40 custom domains implemented worldwide—and more on the way—we may very well be the global leader in this field! Our partnership with Predictwind and Oceanum Science was pivotal in optimizing our workflow and has enabled us to work on tens’ of new hyperlocal-resolution domains around the world and improve ocean data/ tidal current and elevation, information on complex and interesting parts of the ocean. Our tidal predictions power the Predictwind recreational product suite that is trusted by over a million users, ranging from recreational users to world-class sailors (Predictwind was an official weather provider for the famous Vendee Globe race!). For commercial and industrial applications, we provide direct access to our domain catalog, via a flexible API for seamless integration and can implement new hyperlocal-resolution domains on demand, anywhere in the world. Our data supports a wide range of use cases, including improving vessel routing accuracy and ensuring safe navigation through complex coastal areas and passages (e.g., channels, islands, and strong currents). Other applications include pilot training in navigation simulators, ocean condition monitoring for aquaculture operations, and drift predictions for vessels, oil spills, or search and rescue efforts. Hyperlocal-resolution ? As we discussed in a previous post, generating accurate ocean data and forecasts requires downscaling predictions from global ocean models. This involves creating smaller hyperlocal-resolution models of specific regions, nested within broader but less detailed global models like Copernicus’s GLORYS. A crucial step for this process is sourcing hyperlocal-resolution bathymetry, as seafloor topography significantly influences ocean currents. Additionally, information on seabed composition (e.g., sand, mud, rock) can be used to adjust friction parameters, further refining flow predictions. Higher spatial resolution enables models to capture essential seafloor features—such as islands, headlands, shoals, and narrow channels—leading to more precise flow simulations. This is comparable to the process of “zooming in” which we are all familiar with these days. In GoogleMaps for example, you get more details as you zoom in, seeing more and more details of the land and seascapes. If you zoom in on global models, you don’t get that update, in fact some of your favourite islands, and bays may not even exist in some of these models if they are smaller than the pixel size (e.g. 7km for Copernicus’s global domain). For more details, you need to implement smaller, higher resolution models, and that’s what we do. A case study for the coast of France To illustrate the process, we present a case study of our workflow for recent domain implementations on the coast of France. A first step is to review available literature on the study area, and look for what data is already available. In well surveyed regions with strong research centers (e.g. Europe, US), there may already be a lot of good information on bathymetry or even regional tidal models that we can use as a starting point. In some others, data might be scarce, in that case, we start from the GEBCO bathymetric dataset and Oceanum Science ‘s global tidal solution at 400m (i.e. matching GEBCO spatial resolution). To get to the appropriate level of details at scale finer than 400m, we may have to digitize bathymetry from nautical charts (e.g. for the Panama Canal Bay), or use satellite-derived information. The French coast is already well studied and modelled, and past research projects by IFREMER led to the generation of useful tidal atlases. Tidal atlases allow you to generate tidal elevations and currents for any time, past, present or future (yes tides are that predictable!). There were 3 nesting levels starting at 2km down to 250m resolution for the entire coast. Compared to many regions worldwide, the available open-access tidal data is already of high quality—though somewhat hidden away in obscure FTP servers. Our first step was to ingest this data into OCEANUM.IO Datamesh cloud-platform, making it more accessible, reusable, and efficiently served (see Figure 1, top).  Figure 1. IFREMER’s tidal atlases ingested in the OCEANUM.IO Datamesh cloud-platform. However, with a resolution of 250m by 250m, the existing data isn’t detailed enough to fully capture the complexity of tidal currents in intricate coastal areas. To address this, we introduced an additional nesting level, refining resolution to 50–100m in high-energy regions with strong tidal currents and significant maritime activity—such as the (in)famous Raz de Sein, Ushant, the Bay of Quiberon, and the Gulf of Morbihan. This was achieved using a flexible mesh approach (triangular elements instead of fixed rectangular grids) to enhance detail where needed (Figure 2). Model predictions were then validated against tide gauge and current meter observations to ensure accuracy.  Figure 2. Hyperlocal-resolution domains in Brittany (France) with complex tidal currents, such the (in)famous Raz de Sein, and Fromveur Passage near Ushant , Bay of Quiberon and Gulf of Morbihan. These new domains, along with the ingested IFREMER atlases*, have been integrated into our existing catalog, which already includes a 400m-resolution global tidal solution and over 40 hyperlocal-resolution (<100m) domains. All data is accessible via OCEANUM.IO Datamesh cloud-platform with flexible API access, offering a variety of export options—including time series and gridded data in formats like NetCDF, Zarr, GRIB, and S-100—for seamless integration with your preferred tools, third-party applications, and navigation systems. If your area isn’t covered, we can quickly set up new domains on demand, typically within 1–2 weeks. (* we make the open-source domains available at no extra costs via same interface) We have generated sample files for the Raz de Sein area in GRIB, NetCDF, and S100 formats, allowing you to test them in your preferred navigation software. We’d love to hear your feedback! One of the nice features of the area, besides the very strong currents (up to 6 knots!) is how the tidal flow splits at the rocky reefs where the Tevennec lighthouse sits, creating a counter current in its lee. Some other eddy and counter currents develop near both the Finistere headland and Sein Island.  Figure 3. Tidal current animation in the Raz de Sein Our goal is to support the development of digital twins for the ocean areas you navigate or operate in. Depending on the application, we can create tailored apps that combine model predictions with real-time data from tide gauges, current meters, and wave buoys—potentially integrating AI-assisted data assimilation. These apps can also provide additional insights on wave and wind conditions. We currently have several trials underway in Australia and New Zealand, but that’s a topic for another article.
Finally, it's important to note that tides are just one factor influencing ocean movement. In some areas, like the Gulf of Morbihan, tides dominate because the region is sheltered from Atlantic swells and currents, with little room for strong wind effects. However, in other locations, additional forces such as wind, waves, large-scale ocean currents (like the Gulf Stream), river inflows, and storm surges play a significant role. In these cases, models become more complex, incorporating 3D circulation driven by temperature and salinity differences (baroclinic forces) and interactions with the atmosphere and waves. We would be happy to provide you with demo access to our catalog and platform, allowing you to evaluate its capabilities and explore how it can enhance your operations. We’re also eager to hear your feedback on the sample files in GRIB, NetCDF, and S100 formats! Comments are closed.
|
RSS Feed