ESA > Join & Share > Forums > Sentinel Forum - Ocean Research > seabed studies with Sentinel-2

Sentinel Forum - Ocean Research


Show posts:
Jump to forum:

seabed studies with Sentinel-2

Hi All,

It's awfully quiet! I thought I'd kick things off with a suggestion for the Sentinel-2 chief scientist, if it's not too late! This came out of discussions at the sen4sci meeting arising from Sam Lavender's work on the Great Barrier Reef, which couldn't be continued with Sentinel-2 data because of the decision to stop recording >20km from the coast. The discussion broadened into coastal water processes and centred around increasing the 20km limit, but to me this misses the original point because interesting shallow areas could be arbitrarily far from the coast. You could draw boxes round features such as the Barrier Reef, but this is clunky and arbitrary - who knows what important features might be missed?
I propose the following:

1. Use existing satellite Kd with depth Z from bathymetry to calculate 2 way transmission t=exp(-2Kd Z). Do this for at least a year's worth of data globally at high resolution (eg 1km).

2. Set a threshold at which you might stand a chance of detecting the bottom, eg t=0.1

3. Set a threshold percentage of cases that exceed this threshold, eg 50%. At a given location, if the percentage exceeds the threshold (ie the bottom is acceptably visible an acceptable proportion of the time), treat this location as if it were land.

The two thresholds allow a certain amount of tweaking to balance the twin evils of missing vital data and overloading the capacity of the system to process the data. It could be argued that you might include a lot of boring muddy bottom, but I would argue that just as we don't distinguish between busy rainforest and featureless desert, we should take a similar stance with the sea floor - if you can see it, it might be interesting! Anyway, I suspect the increase in data volume would be quite small.

Re: seabed studies with Sentinel-2

Dear Peter,

thank you for your contribution and please accept our apologies for the late reply. Currently we are finalizing the documents outlining the preliminary needs for LOC sciences which will be made public for an open web review soon. This will be announced to all WS participants. I look forward to your participation in the open web review. Please add your comments to the relevant document.

Show posts:
Jump to forum: