Monitoring the Spatial Distribution of Cover Crops and Tillage Practices Using Machine Learning and Environmental...

Area of Science:

• Agricultural Science
• Environmental Science
• Remote Sensing

Background:

• Conventional farming leads to soil degradation and carbon loss.
• Conservation agriculture (tillage, cover cropping) offers sustainable alternatives.
• Limited research exists on spatial distribution of these practices influenced by environmental factors.

Purpose of the Study:

• To map conservation agriculture practices in eastern South Dakota using remote sensing.
• To identify the spatial distribution of cover crops and tillage intensity.
• To assess the influence of pedoclimatic and topographic factors on practice adoption.

Main Methods:

• Utilized machine learning classifiers trained with in situ field data.
• Employed satellite-derived spectral indices and environmental drivers (precipitation, growing degree days, surface texture).
• Conducted analysis for the 2022 and 2023 growing seasons.

Main Results:

• Achieved classification accuracies exceeding 80% for detecting cover crops and tillage intensity.
• Revealed that 4% of corn and soybean fields in eastern South Dakota utilized cover crops.
• Demonstrated significant impact of seasonal precipitation, growing degree days, and surface texture on conservation practice adoption.

Conclusions:

• Satellite-derived indices and environmental data effectively map conservation agriculture practices.
• Environmental factors play a crucial role in the adoption of these sustainable farming methods.
• Developed methods support monitoring of climate-smart agriculture and contribute to Measurement, Reporting, and Verification (MRV) solutions.