Project Grant 2603941

The National Science Foundation Division of Mathematical Sciences awarded Iowa State University of Science and Technology $200,000 on August 1, 2026, under the Mathematical and Physical Sciences program (CFDA 47.049) to develop matrix-free computational methods for large-scale spatial analysis.

The project will create scalable computational tools for spatial prediction, inference, uncertainty quantification, and model evaluation that avoid forming large dense matrices, addressing the computational prohibitiveness of existing statistical approaches as spatial and environmental data resolution increases. Core computational tasks—including kriging, restricted maximum likelihood estimation, likelihood evaluation, conditional simulation, and spatial inference—currently require computational effort cubic in sample size; the project targets methods that reduce this burden. Applications include high-dimensional genotype-by-environment experiments in maize to understand phenotypic stability and plasticity, with anticipated advances supporting agricultural productivity and data-driven decision-making in food systems. The project will also produce educational resources, software tools, and training in statistics, data science, and computational methods for the broader scientific community.

Work is performed at Iowa State University in Ames, Iowa, with a period of performance from August 1, 2026, through July 31, 2029.