A Multiscale Framework for Airflow-Canopy Interaction with Jaeyoung Jung

Abstract: The airflow dynamics within canopies significantly influence their microclimate across a broad continuum of scales by regulating the exchange of mass, momentum, and energy between the earth and atmosphere. These dynamics are key in managing temperature and humidity, and they impact nutrient uptake, gas exchange, along with a variety of biological and ecological processes. Yet, despite more than six decades of intensive efforts to deepen our understanding of canopy flow physics, the complexity of the multiscale feature still poses a challenge to meeting practical demands. Existing models often assume that canopy distributions are locally homogeneous or isotropic, assumptions that have proven effective for macroscale simulations but may not hold at smaller scales where canopy heterogeneity is pronounced. This study introduces a multiscale framework aimed at providing high-resolution simulations of airflows within canopies, including areas of heterogeneous vegetation. The framework consists of macroscopic and microscopic models to capture the multiscale nature of canopy airflows. For the macroscopic component, we formulate the macroscopic model by upscaling the incompressible Navier-Stokes equations via volume averaging and develop exact and approximate Riemann solvers for macroscale problems with discontinuous/heterogeneous porous fields. For the microscopic parts, the study constructs the unit cell problem using idealized microstructures of the canopy and integrates it with machine learning to train surrogate models for unresolvable scale effects on the flow. Through a series of numerical examples, we verify the proposed model against reference solutions, showing good agreements with each other. We believe that by recursively extending our model, it can cover a vast range of scales from a few meters to hundreds of kilometers, offering a comprehensive tool for studying atmosphere-canopy interaction.

Bio: Jaeyoung Jung is a postdoctoral researcher at LEAP and the Department of Civil Engineering and Engineering Mechanics (CEEM) at Columbia University. He specializes in mathematical and numerical modeling of flow physics. Jaeyoung earned his Ph.D. in Civil & Environmental Engineering from Seoul National University, South Korea, where he developed advanced numerical schemes for urban flood modeling through a comprehensive study of shallow flow theory. Recently, his primary focus is on integrating multiscale modeling with machine learning techniques to enhance our understanding of the interactions between the atmosphere and canopies.