Skip to main content
Pixel-level frequency of low-intensity tillage, estimated as the number of times each pixel was classified as such divided by the total number of times that pixel was flagged as soybean in CDL

Stanford partners publish report on US tillage practices

Harvest partners Dr. David Lobell and George Azzari, from Stanford's Department of Earth System Science, Center on Food Security and the Environment, released a report titled Satellite mapping of tillage practices in the North Central US region from 2005 to 2016, to be published in the journal Remote Sensing of Environment in February 2019. In this article, they explore how tillage practices have shifted, noting that in many areas around the world, including the United States, soil management has transitioned over the past decades from annual intensive tilling of the soil to reduced or no-till methods that minimize soil disturbance. This kind of reduced tillage can be environmentally beneficial, as it may reduce soil erosion, increase soil moisture, and reduce fuel use. Scarce data on these practices, however, makes it difficult to track changes in these practices, therefore making it hard to evaluate the possible impacts of these changes on the environment and production. 

For this study, the authors used composites of satellite imagery from Landsat 5, 7, and 8, and Sentinel-1 in combination with producer data from about 5900 georeferenced fields to generate annual large-scale maps of tillage intensity from 2005 to 2016. Using different parameter combinations, they were able to test data to determine how useful the classification model was at determining tillage practice areas, finding the best model to have 75-79% accuracy at a 30m resolution. Landsat alone was almost as accurate at classifying tillage areas as when including data from Sentinel, surprising due to its higher resolution. The authors used a USDA survey of tillage practices to determine the accuracy of the model at the state level. While the study was able to highlight the significant increase in low-tillage practices in counties of North Central US from 2005-2016, the accuracy of this model makes it a potentially very useful tool for further study of tillage practices. 

News Date
Dec 7, 2018