Our colleagues at Stanford University, including Jillian Deines, Sherrie Wang, and Harvest partner David Lobell, continue their research in the U.S. Corn Belt, recently publishing an article in Environmental Research Letters titled “Satellites reveal a small positive yield effect from conservation tillage across the US Corn Belt.” Conservation tillage aims to reduce soil erosion and improve soil health by minimizing plowing and mechanical disturbances in agricultural fields. Although reduced tillage is good for long-term soil health and cost effective, the direct impacts on crop yields is less clear. Their work expands upon and aims to clarify the often ambiguous and diverse results found in existing field trial research of conservation tillage practices, which have typically not been able to capture real-world field conditions.
The group uses machine learning algorithms along with satellite datasets of tillage practices and crop production in the U.S. Corn Belt region from 2005 to 2017 in order to evaluate yield impacts at the field-scale over thousands of acres of farmlands. During their study of these areas across this long term time span, the average yield increase for maize was found to be 3.3% and the soybean average yield increase was 0.74%. However, it does take several years for these benefits to be seen after initially switching to lower impact tillage practices. Notably, the research also documents variability in these yield effects due to both soil and weather events, which can directly affect the production responses in the areas studied.
Our partner's are advancing the study of soil health and agricultural practices in the U.S. Corn Belt by providing science-based information on conservation tillage and other soil conservation approaches, specifically regarding their relationship to increased crop yield. Furthermore, conservation tillage may also be linked to lowered production costs and reduction of supplemental water demands. By using machine learning causal inference methods in combination with Earth observation data, the effects of long term conservation practices will continue to be evaluated over time. This data-derived information is encouraging to farmers and other stakeholders when adopting conservation practices because they can implement these methods with more confidence that soil health and production increases will be realized in the long-term.
While results showed moderate improvements to crop yield using conservation tillage systems over the area and time span studied, the group cautions that further research is required in order to specifically define the management methods that contribute most to yield gains. Moving forward, research will need to be expanded to large-scale fields and evaluated in combination with specific data on management practices.