[Imports] Antarctica coastline/shelf-ice import
Stephen V. Mather
svm at clevelandmetroparks.com
Tue Mar 5 17:17:16 UTC 2013
As far as licensing, RAMP is public domain data, so that won't be an issue. RAMP coastline data was manually interpreted, so there are no processing artefacts. If history/memory serves me, the original was to be automatically derived, but in the end the 12 months of algorithm development were replaced by 160 hours of manual digitizing... . I know both MOA and RAMP well; I worked on RAMP in it's final years.
Overall, I'm suggesting that long term it would be less work to dump the RAMP coastline in and conflate the ice shelf changes from MOA, and the end product would be more precise, smoother, etc., but less work depends on how much work you've already put into preparing MOA.
I suppose if I'm going to be critical, I need to put skin in the game, and I have none to offer for a few months, so MOA away!
Stephen V. Mather
(216) 635-3243 (Work)
From: Christoph Hormann [chris_hormann at gmx.de]
Sent: Tuesday, March 05, 2013 11:38 AM
To: imports at openstreetmap.org
Subject: Re: [Imports] Antarctica coastline/shelf-ice import
On Tuesday 05 March 2013, Stephen V. Mather wrote:
> Hi Jochen,
> You might contact Byrd Polar Research Center
> (http://bprc.osu.edu/) for detailed vector information of the
> Antarctic coast-- the resolution would be better than 125m, more in
> the 10-30m range. They should have ice shelf and grounded ice info
> from the Radarsat Antarctic Mapping Project (RAMP). I would contact
> their librarian about such data. MOA may be more up-to-date for some
> ice shelf changes particularly in the Antarctic peninsula and similar
> places, so it would require some conflation of datasets, but would
> give an order of magnitude better data that MOA alone.
We have looked into alternative data sources and i have contacted both
USGS and British ADD concerning use of their data - without success,
(if anyone with direct contact to either of these wants to try go ahead)
I am aware of the RAMP data products although i have not looked in
detail into the coastline data. From what they write about it though:
it seems deriving the coastline from SAR data is very difficult and they
accomplish it with a lot of filtering - possibly loosing a lot of small
detail in the process. From a quick look it seems for example some of
the islands in MOA are missing in the RAMP data.
The great advantage of the MOA data is that there is matching coastline
and shelf ice data and it seems sufficiently reliable to import for the
whole continent without detailed manual verification. We are pretty
sure it is better than the existing data everywhere we plan to replace
it. We do not regard this as universally the best available data but
as a good starting point for further improvements. So if the RAMP data
should prove to be better (i.e. more detailed and sufficiently
accurate) in some places it would make sense to do further imports (in
smaller chunks after sufficient verification of the accuracy) - always
assuming it can be used license-wise.
Imports mailing list
Imports at openstreetmap.org
More information about the Imports