Climate Change: Processed HWSD Linked With Landsat

Carbon dioxide (CO2) is the most commonly produced greenhouse gas and is a major contributor to climate change.
Carbon dioxide is removed, or sequestered, from the atmosphere when it is absorbed by plants as part of the biological carbon cycle, and it can be stored as soil organic carbon (SOC). SOC sequestration is beneficial in fighting climate change, but it is also crucial to soil health, fertility, and ecosystem services. Regenerative Agriculture practices may be a solution to sequestering large amounts of CO2.

There is a need for an efficient, accurate, and inexpensive method of quantifying the amount of SOC at any given location. Current methods have drawbacks. For example, dry combustion with automated analyzers is expensive and involves pretreatment with acid, which may destroy organic matter.

Machine learning can help. Large soil attribute datasets such as the Harmonized World Soil Database (https://www.fao.org/soils-portal/soil-survey/soil-maps-and-databases/harmonized-world-soil-database-v12/en/) and remote sensing imagery from Landsat 8 (https://www.usgs.gov/landsat-missions/landsat-8) can be used to train machine learning models to quantify SOC at any given location. However, processing the Harmonized World Soil Database is time-consuming. When I processed it, the program took several days to complete running, and that was after optimizing the code to run quickly. That's why I created this Kaggle dataset — to provide an easy and accessible way to get started with the Harmonized World Soil Database and also link it to Landsat 8 imagery. For each grid cell from the Harmonized World Soil Database, its identifier (MU_GLOBAL) is provided, along with its latitude and longitude coordinates. The MU_GLOBAL ID is used to link to its corresponding Landsat image and its soil properties in the Harmonized World Soil Database.