Update albedo layers

[trey-stafford]

Can we find newer albedo data? Currently we have monthly fields for July of 2018 and 2019 and is derived from Sentinel-3A.

There is a MODIS derived dataset that has monthly data through 2019. That would at least allow us to provide albedo estimates for more than just July, but it wouldn't provide us with newer data.

[twilamoon-science]

1. I sent an email to Jason Box at GEUS to ask about an operational product for the Sentinel-3 albedo. I'm pretty sure he was funded to do that but it doesn't appear done.
2. I think it would make sense to create alternative albedo layers until a Sentinel-3 product is operational. What this looks like is open for discussion. Could be something like a 10-year monthly mean albedo and then a monthly albedo for 1-2 most recent data years. Happy to discuss (and dependent on data size/work effort).

[trey-stafford]

Jason Box provided us with an update via email:

Here's a good opportunity to use our THREDDS server that has almost all of the S3A SICE data for Greenland, ~7 years, 2017 to present.

do a git clone of https://github.com/GEUS-SICE/SICE_gather

code in the src folder

data from other Arctic regions is in development

to raise issues please use the issues function at GitHub

[MattF-NSIDC]

I think this is the key line? Pretty cool Xarray can read right from THREDDS... didn't know that.

https://github.com/GEUS-SICE/SICE_gather/blob/main/src/SICE_gather_to_tif.py#L167

[MattF-NSIDC]

We pulled down a NC file and opened in QGIS. It didn't pick up the CRS correctly, but it seems to be in 3413 already so displayed as expected.

[trey-stafford]

The THREDDSs data linked to in the GH repo above has multiple variables representing albedo:

• albedo_bb_planar_sw: "shortwave broadband planar albedo"
• albedo_bb_spherical_sw: "shortwave broadband spherical albedo"
• albedo_spectral_planar_{01..21}: spectral planar albedo for OLCI bands 1-21.

There are also variables for the effective absorption length, aersol optical depth at 550nm, snow covered fraction, grain diameter, total ozone product, and the Angstrom exponent at AOD550nm and AOD670 nm.

Our current albedo layers are also derived from OLCI (and also from GEUS, different dataset) and show planar broadband albedo. If we want to stick with that, we would choose to use the albedo_bb_planar_sw variable.

I think we could fetch multiple days worth of this data and use gdal (or some other nc tool?) to do an average, but I haven't experimented with that yet.

[MattF-NSIDC]

https://glossary.ametsoc.org/wiki/Spherical_albedo

https://glossary.ametsoc.org/wiki/Plane_albedo

This is an excellent glossary site!

[MattF-NSIDC]

If we want to stick with that, we would choose to use the albedo_bb_planar_sw variable.

Is there value in providing variety here?

[twilamoon-science]

I don't know! The first person that comes to mind to ask is in the field right now. Searching my papers library also doesn't help with this question. Perhaps asking Jason Box at GEUS directly (jeb@geus.dk) is the best bet.