S-102 Data Distribution

Giving the end-user exactly what they need has never more relevant!

The trends in distribution of data in other fields are increasingly becoming more user-driven. Other methods for data distribution, such as streaming services, are taking over from traditional methods.

From our side this is an area that commands much more research – the natural step is to look at how this is currently being solved, such as the media, where streaming is an easy and efficient way to distribute large volumes of data.

Large amounts of data require good capacity for data transfer solutions, with an initial data load requiring high capacity and bandwidth. In order to maintain changes to the data, a solution that does not require a complete data transfer of all the data, would be the more preferable option – instead only changes to the original data would be updated.

In the project, transfers from the PRIMAR database were tested, using the PRIMAR download API, directly into an end user interface (in this case the S-102 Demonstrator).

Access to test data was limited during the project, so focus was mainly on the areas where we planned and carried out operational tests.

It would be of great interest to follow up and carry out further analysis when access to S-102 data increases. Our expectation is that several countries will start producing S-102 data in 2019. Canada and the United States are are already quite ambitious in terms of building the S-102 data coverage in 2019.

Download API

For ease of access to S-102 data from the S-102 Demonstrator application, PRIMAR expanded by using an API where the client application downloads a GeoJSON file, with one feature for each of the S-102 data sets that the user has access to.

Each feature then has the S-102 coverage (Multi) Polygon as its geometry and other attributes including the S-102 name and the URL where the client can download the S-102 file. The client application (the Demonstrator) downloads the GeoJSON file which then presents a list of S-102 files to the user in a drop down menu.

Data structure and formats

When distributing large volume of bathymetric data, there are several items to take into consideration. The data can have different capabilities and different users might have different preferences when it comes to data structures and formats, and for good reason!

  • Grid as used by S-102
    • PRO
      • Efficient data format to use in 3D applications.
      • The grid formats can store several values in each grid cell – elevation/depth + uncertainty.
    • CONS
      • An algorithm needs to be determined to find the grid cell value based on all of the soundings for that particular grid cell – pilots wants the shallowest, whilst the submarines want a weighted average.
      • Problems with neighbouring grids not matching in grid cell origin, projection or grid cell size. Canada has a uniform grid for all of its waters, however, the grid does not extend over borders to neighbouring countries.
  • Tight elevation curves as in bENC or “HD” S-57
    • PRO
      • Can use existing software on producer side and client side, although the software might need some minor adjustments.
      • Efficient storage format (at least for normal resolution and terrain).
      • No problems with projection as S-57 data is in unprojected WGS-84 coordinates.
    • CONS
      • Not able to store different grid values for shallowest v’s weighted average.
      • Might be less efficient than grids for very detailed data such as 1m horizontal and/or the terrain is exceptionally steep.
  • Point Cloud
    • PRO
      • No problem with projection.
      • All values are included – the need to discuss algorithms to pick a grid value is no longer a requirement.
    • CONS
      • Massive data volume. The volume is so colossal that even a normal computer can only handle a small area, and transferring over the internet could prove problematic. Tiling has the potential to make this more dynamic, but it is still a significant problem.
      • Not an efficient way to store “no-data”?
      • Might include noise – if filtered away, some users might want to know the filtering algorithm used (as with the grid cell value algorithm discussion).