Daily Albedo from MODIS BRDF/Albedo Parameters

A comprehensive workflow for calculating black, white, and blue sky albedos from the MODIS BRDF model parameters product (MCD43A1) using custom solar zenith angles and optical depth inputs.

Explore the docs » · Interactive Tutorial · Report a Bug

Figure 1. Time series blue, black, and white albedo for 2018 for one pixel in central Florida.

Overview

Use BRDF/albedo model parameters to get black sky albedo at different illumination angles or to combine the black sky and white sky albedo as a function of optical depth to get the blue sky albedo. This series of notebooks covers that process.

Figure 2. Time series difference between black sky albedo computed by the ORNL DAAC MODIS Global Tool and this workflow as a way to validate my implementation of the albedo algorithm. Solar zenith angles are shown by the shaded region.

Notebooks

The entire process to implement the albedo algorithm in Python is documented in these notebooks:

Notebook	Description
0_Introduction.ipynb	Interactive tutorial covers core concepts. Launch it in MyBinder:
1_Workflow.ipynb	Process albedo model parameters to black, white, blue albedos.
2_Batch.ipynb	Bundle workflow into a few functions and loop.
3_Validate.ipynb	Check results against MCD43A3 albedos.
4_Results.ipynb	Calculate stats and plot. Also write outputs for Florida USGS evapotranspiration model. (WIP)
5_HDFs.ipynb	AppEEARS alternative, processing HDFs. (WIP)
Some of these I've barely started (WIP). Stay tuned.

Inputs

You have two options:

1. Input MCD43A1 time series in netCDF format via AppEEARS. Or, if you don't have convenient Python environment
2. you can batch download the HDFs from LP DAAC data pool by following the steps in 5_HDFs.ipynb.

Go to the first notebook (1_Workflow.ipynb) for more details about data access.

Outputs

The figure and text below explain a little bit about how I validated the workflow against MCD43A1 albedos computed by the ORNL DAAC's MODIS Global Tool (Figure 2) and the standard black and white sky albedos from the MCD43A3 product (Figure 4).

The difference between our time series and the one computed by the ORNL DAAC (see Figure 2 above) is negligible. I believe the disparity (see y-axis precision) comes from difference in decimal precision used in the solar zenith angle calculators for this workflow and at the DAAC.

Figure 4. Difference of black sky albedo rasters (2) for January 1, 2018: a. released in MCD43A3 standard product, and b. computed in this workflow.
Figure 4 depicts the differenced black sky albedo rasters for January 1, 2018 computed by (1) the SIPS for the standard black and white sky albedo product (MCD43A3) and (2) this workflow.

License

Distributed under the MIT License. See LICENSE for more information.

Contact

Jack McNelis - jjmcnelis@outlook.com