Location: Watershed Physical Processes Research
Project Number: 6060-13000-029-032-S
Project Type: Non-Assistance Cooperative Agreement
Start Date: Sep 4, 2023
End Date: Sep 3, 2026
Objective:
Develop remote sensing and machine learning tools to locate, identify, and characterize existing structural conservation practices to support soil erosion modeling efforts.
Approach:
1. Develop methodology using remote sensing, machine learning, and geoprocessing algorithms to identify existing terraces, waterways, and grassed strips and buffers in agricultural fields.
2. Support the integration of the developed tools into existing soil erosion modeling systems based on RUSLE2 and EphGEE models.
3. Investigate the applicability of machine learning feature extraction methods to determine extent, depth, and width of ephemeral gully channels using imagery and LiDAR elevation data.
4. Extend toolset to include retrieval of vegetation, crop residue, and tillage practices from remotely sensed data.
The long-term, post-project goal is to extend this work into an automated system of conservation practice detection that can improve soil erosion modeling for better, sustainable agricultural practices.
