The USDA’s Conservation Reserve Program (CRP) works with private landowners to implement conservation management practices via the transition of agricultural land to grasslands or forests on lands participating in the program. This program provides long term investments for landowners and supports ecosystem services like soil health, clean water, wildlife habitat and carbon sequestration.
However, a lack of information on vegetation within existing CRP lands presents challenges for USDA managers and landowners to characterize current landscape conditions and assess success of grassland and forest conservation practices. This project continues CBI’s remote sensing work on CRP forests by enhancing workflows with additional satellite imagery and cloud-computing, and expands predictive modeling to CRP grassland holdings in the Western United States.
This project leverages the powerful cloud-computing platform, Google Earth Engine
, machine learning, and an extensive suite of satellite-derived imagery and environmental datasets to map vegetative cover on CRP grasslands in Washington, Colorado and Kansas. Given the variation represented in grassland ecosystems, this project relies on multiple satellite imagery sources (including Sentinel and Landsat platforms) and robust modeling techniques (such as Random Forest) to map and predict the vegetative cover of CRP sites. These outputs lend insight into the condition of grasslands and contribute to efforts aimed at monitoring the condition and success of the conservation activities on land within the CRP Program going forward.
Vegetation maps and analysis results are displayed in an online decision support tool, which presents and summarizes important metrics for USDA’s Conservation Reserve Program lands. This tool will allow USDA staff and landowners to view pertinent spatial information and guide decision making, to ensure conservation goals are monitored and met, thus ensuring a bright future for CRP lands.