1

I am trying to calculate the blue sky albedo, which I think is the same as land surface albedo, using Google Earth Engine and the 'MCD43A3.006 MODIS Albedo Daily 500m' data.

I've managed to find various formulas, user guides and papers (example 1, example 2) describing this. However, maybe I'm a bit too new to GIS to understand these and how to implement them.

How would one go about estimating the blue sky albedo from the various black sky and white sky albedo bands present in the MODIS dataset?

Maybe something similar to what is produced on the Remote Sensing Lab albedo tool.

My code is here:

var albedo: "ImageCollection MCD43A3.006 MODIS Albedo Daily 500m" // this is in the 'imports' section.  

var albedo_filtered = albedo
                      .filterDate('2019-01-01', '2019-01-31');

var blackSkyAlbedo = albedo_filtered.select('Albedo_BSA_Band1');

var alb_val_range = {
  min: 0.0,
  max: 600.0,
};
var albedo_median = blackSkyAlbedo.median();

Map.addLayer(albedo_filtered, alb_val_range,'all layers', 1, 0.5);
Map.addLayer(albedo_median, alb_val_range, 'Median Black-Sky Albedo', 1, 0.5);

// Map.addLayer(blackSkyAlbedo, alb_val_range, 'Black-Sky Albedo', 1, 0.5);

1 Answer 1

1

There's no unique way to calculate blue sky albedo from the MCD43 product, so no code snippet here. Basically, the MCD43 product gives you black and white sky albedo (directional and diffuse illumination albedo), whereas blue sky can be thought of as a linear combination of these two values, so you need to calculate the ratio of direct-to-diffuse radition.

The main processes that affect this ratio are solar geometry, itself a function of time of day, year and latitude and aerosol optical depth (broadly, aerosols in the visible). So if you know the solar zenith angle and have an estimate of AOD, you can run a radiative transfer code such as 6S, libradtran or similar.

A simple way to do this is to run the models and store the values in a look-up table for each wavelength. If you need blue sky albedo for broad bands, you can find some LUTs already calculated here (that's for NIR, but there's one for SWIR and one for VIS). You open an AOT dataset in GEE (e.g. this one), extract the white sky and black sky albedos (you may need to calculate this using the formulas from here, work out your sun angle, and read in the relevant value from the LUT.

You can read more about the theory in the original paper of Lewis & Barnsley, 1994, or in the Schaaf et al, 2002 paper.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.