Earthdata Forum

Dear colleagues,
I have a question in regard to the diffuse attenuation coefficient product Kd_490. I am using this product for my study of the coastal water turbidity around La Réunion. The Kd_490 values vary with the water depth and no data of that coefficient is provided for nearshore areas and regions of relatively shallow water depths (like reefs, lagunes). Does anybody know which underlying factors in the procedure of the image precessing are considered and what determines the exclusion of areas (apart from land and clouds)? Is depth actually the crucial factor in that regard as I assume or what else is considered?
I would be very happy about a source I can cite clarifying this question. Thanks a lot!

You don't mention which level data you are using.   In general, land, turbid, and shallow clear water are much brighter than deeper water.   Ocean pixels that are close to land can seem overly bright due to the Gaussian shape of the sensor response function.   Missing data are usually "flagged" in the level-2 files, so it may be helpful to find a nice cloud-free level-2 files and see which flags are set in the shallow water areas.

I don't think there is a depth flag/mask per se that would result in the missing pixels you see nearshore. If you take a look at: https://oceancolor.gsfc.nasa.gov/atbd/ocl2flags/ you'll see that the standard masks applied during level-2 processing are LAND, HILT, STRAYLIGHT, and CLDICE. If you use SeaDAS you can easily visualise which l2 flags are active or not for a given pixel. I suspect that if your reef pixels are nearshore (i.e. within 1 km), they will likely be masked as LAND or STRAYLIGHT. You could learn more by using l2gen to re-process the scenes and toggling on and off the default masks.

Remember:
(i) The quality of the derived product could be contaminated by straylight emanating from the adjacent land pixel.

(ii) For optically shallow waters, light reflected from the seafloor can confound standard ocean colour algorithms designed for optically deep waters. So standard Kd_490 may not be appropriate for your application.

With regards, to shallow water remote sensing, the following may be useful:
Barnes et al. (2017): https://doi.org/10.1016/j.rse.2017.10.013
McKinna et al. (2015): https://doi.org/10.1002/2014JC010224

Dear both,
thank you very much for the answers! Those publications are very useful. In fact, I realized that only a little fraction of my area of interest consists of shallow waters, sandy sea floor and reefs. La Réunion is characterized by a very steep bathymetry, the water is deep just off the coast and as it is a volcanic island the source rock is very dark. Therefore, I would be happy about any recommendations of turbidity detection algorithms around volcanic oceanic islands and/or open ocean waters that can be applied to MODIS ocean color data. Any suggestions?
Thanks a lot!