usgs-landsat-collection2.md

Indexing USGS' Landsat Collection 2

Landsat data is available from a public bucket located in the Oregon region of AWS. The bucket is s3://usgs-landsat and requires "requester pays" access.

aws s3 ls --request-payer requester s3://usgs-landsat/collection02/level-2/

An ODC product definition for the Landsat surface reflectance data is inluded in this repo. Add this to datacube using the following command:

datacube product add products/lsX_c2l2_sr.odc-product.yaml

or

datacube product add https://raw.githubusercontent.com/opendatacube/datacube-dataset-config/main/products/lsX_c2l2_sr.odc-product.yaml

Installing the required tools

Install the scripts required by following the instructions here like this:

pip install odc_apps_dc_tools

Note that because of the requester pays requirement, you must have AWS credentials present in the environment you're working in.

Indexing from a STAC API

Indexing for a region of interest can be done using the STAC API, like this:

stac-to-dc \
--catalog-href='https://landsatlook.usgs.gov/stac-server/' \
--rewrite-assets='https://landsatlook.usgs.gov/data/,s3://usgs-landsat/' \
--bbox='25,20,35,30' \
--collections='landsat-c2l2-sr' \
--datetime='2020-01-01/2020-03-31'

Added 707 Datasets, failed 0 Datasets

You can change the bounding box or datetime range to search anywhere in the world.

Note about requester pays

Please note that you must use AWS credentials to access the data from the requester pays bucket. The USGS has information about cloud data access here.

There are some docs here on how to set up the requester pays flag using the ODC.