NSF Convergence Accelerator Track J Phase 2: CropSmart - a digital twin for making wiser cropping decisions nationwide

George Mason University, Fairfax VA

Investigators

Abstract

Healthy crop production in the U.S. is critical for not only the food and nutrition security of the U.S. and the world but also the prosperity of the U.S. economy. The USDA Agricultural Innovation Agenda calls for increasing U.S. agricultural production by 40% by 2050. Sound crop management decision-making is a key to achieving this ambitious goal. An example of such decision-making is “should I irrigate my cornfield today? If so, by how many inches of water?” Traditionally, such decisions are made by individuals based on their empirical judgment, which is often subjective and less optimal. Science-based, data-driven approaches for cropping decision-making rely on timely and accurate information on current and predicted future conditions to make optimal decisions. However, it remains a challenge for stakeholders to adopt a data-driven approach because they do not have full and effective access to the timely and accurate information and lack facilities or capacity to process the information. This project will meet the challenge by offering the data-driven optimal cropping decision-making services nationwide up to field scales through developing and operating the CropSmart digital twin. The services will be available to users through both web portals and smartphone Apps. This project will help the nation to achieve its innovation goal, enhance food and nutrition security throughout the U.S., and bring hundred-million-dollar economic returns to the U.S. economy and American communities annually. CropSmart, to be built and operated by this project, is a digital replica of real-world cropping systems over the contiguous U.S. with up to 10-m spatial resolution. It will not only accurately represent the current crop and environment conditions, but also predict, with acceptable confidence levels, future conditions with hypothetical “what if” scenarios, resulting in actionable predictions. CropSmart will provide three services to users: 1) user-specific decision ready information on which the user can make data-driven decision; 2) “what if” tradeoff service which will generate consequences (e.g., yield, economic return) of different user decision options so that the user can find the optimal decision; and 3) decision advice service which will automatically generate optimal decision based on a user’s decision goal. CropSmart will be built by integrating the advanced remote sensing, crop and conditions modeling, artificial intelligence and machine learning (AI/ML), agro-geoinformatics, and digital twin technologies through a multi-disciplinary approach. The major project activities will include: 1) implementing CropSmart to support at least 6 types of top-priority decision-making use-cases specified by the user community; (2) deploying CropSmart operationally to cultivate its user community and show its game-changing impacts; 3) expanding adoption, participation, and impact through a comprehensive engagement program; and (4) establishing a community-based CropSmart.org website and implementing a plan to enable CropSmart to endure after the project expires to maximize long-term impacts. At the end of the performance period, this project will deliver the CropSmart software package, the operational CropSmart services, and a community of at least 6,000 users. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.