NOTE: The galeon mailing list is no longer active. The list archives are made available for historical reasons.
Hi Simon, My original contention was that the primary use case for WCS in our community is to communicate metocean data to *other* communities. The metocean community already has tools and protocols for working with 4D data. I do not see metocean scientists adopting commercial GIS tools for doing their everyday science work, just as you say that "serious" geoscience users don't use them for solid earth applications. I don't see much value in adapting the existing metocean tools to use WCS when they already work very well with OPeNDAP. (Although I'm slowly coming around to the idea that WCS might be a useful way to share standard *metadata* about metocean coverages for cataloguing and discovery purposes.) I can, however, see GIS tools and protocols being used to access metocean data by non-metocean specialists. Some of these users are desperate to get hold of metocean data but don't want to get to grips with the technical complexity, and they probably won't use tools designed for metocean scientists. I'll give two examples, taken from real work done in (or with) my department: 1) A volcanologist needs a column-integrated water vapour field to correct Synthetic Aperture Radar retrievals over a deforming volcano. 2) A coastguard needs forecasts of surface winds and surface ocean currents to drive an oil spill model, or to direct search-and-rescue operations. This is a limited view only of course, but this is why I'm concerned about adoption by commercial GIS developers, rather than the metocean community. The user in case (1) uses a well-known GIS package. In case (2) I'm not 100% sure if the GIS package is COTS or developed in-house. Neither of these users wants or needs metocean data in its full complexity: (a) they only need 2D fields (in fact I see very little demand for 3D metocean fields outside science and aviation); (b) they don't need to know the original grid of the data: a reprojection into lat-lon space would do fine. In other words, we can satisfy these use cases (and many others probably) with a simplified version of a WCS (maybe this is what WCS1.2 "core" will do for us). (I also note that the time dimension is very important in both cases, which is a current limitation of many GIS systems.) This is of course only my view from where I'm sitting. Cheers, Jon On Thu, Sep 25, 2008 at 3:17 PM, <Simon.Cox@xxxxxxxx> wrote:
As you are probably aware, I'm not a metocean person so am a lurker on these conversations. But are "most GIS users" the target audience for metocean data? My science background is solid earth. We work in 3-D a lot. But we don't make a lot of use of GIS systems - usually only when demanded by project sponsors. Mostly we are loading and unloading from specialized 3-D software. The benefit of standard interfaces like the OGC ones is cross-domain data availability, and emerging software libraries. Traditional GIS is pretty hopeless in 3-D, and not great on time-series either. So to me it would seem to be of quesionable relevance to serious geoscience applications, particularly simulation and modelling. Now you may want GIS users to have access to some 2-D sections, so I guess they might be clients of a 2-D coverage service. But I wouldn't expect the "major GIS vendors" to have much interest in anything more elaborate. Is there a "2-D geospatial profile" of netCDF? Best - Simon
galeon archives: