SBIR Phase I: Automatically Configurable Graphical Processing Unit Optimized Ocean Carbon Modeling Platform

About This Grant

The broader/commercial impact of this Small Business Innovation Research (SBIR) Phase I project is to enhance trust, transparency, and accuracy in the emerging marine Carbon Dioxide Removal (mCDR) industry by developing a scientifically rigorous ocean modeling platform that is easy to use. Accurate verification of carbon removal is critical for generating high-integrity carbon credits, which are essential for building a credible carbon market. This project will help standardize and automate the verification process, reducing costs for project developers and lowering barriers to entry across the mCDR sector. By ensuring that carbon credit reflects true, measurable carbon removal, the platform will foster confidence among investors, verifiers, and policymakers, encouraging greater market participation and sustainable growth. As the mCDR sector scales, this project has the potential to drive job creation in environmental science, data technology, and sustainable marine industries, contributing to U.S. tax revenue and economic resilience. In addition to supporting the development of a trustworthy and effective carbon removal industry, this project will help protect marine ecosystems, and strengthen coastal economies, contributing to long-term societal and environmental well-being. This project proposes the development of an advanced ocean modeling platform that integrates Graphical Processing Unit (GPU)-optimized modeling software with artificial intelligence (AI) tools to automate the configuration, validation, and analysis of regional ocean models for marine Carbon Dioxide Removal (mCDR). The core innovation lies in leveraging Oceananigans, a cutting-edge ocean simulation framework, and integrating it with AI-driven automation to streamline traditionally complex modeling workflows. This approach aims to make high-resolution, scientifically rigorous ocean modeling easy to use by non-expert users. Key research objectives include automating the retrieval and processing of ocean data, developing tools for automated model validation, and testing the feasibility of real-time adaptive domain boundaries to enhance computational efficiency. The project will also explore using large language models (LLMs) to simplify model setup and gradient-based methods for optimizing mCDR project parameters and uncertainty quantification. If successful, this work will significantly reduce the cost and complexity of ocean modeling, providing a scalable solution to support the verification and reporting of carbon removal in the emerging mCDR sector. The outcome will be a more efficient, user-friendly modeling platform that can be applied to a wide range of marine environmental challenges. 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.