Orthogonal decomposition is a cornerstone signal-processing technique widely adopted across sciences and engineering. The two most influential algorithms—the Proper Orthogonal Decomposition (POD) and Dynamic Mode Decomposition (DMD)—now result in a research domain with more than 100,000 publications, including more than 7,300 in 2025 alone. The popularity, relevance, and novelty of orthogonal decomposition research make it particularly attractive for wind engineering methods decompose complex fields into mutually independent modes, isolate dominant structures, and enable reduced-order modeling.

Our research has demonstrated that orthogonal decomposition extends far beyond conventional utilities. By synchronizing multi-field datasets, we revealed previously inaccessible physical insights. Flow–pressure–field coupling exposed the phenomenological excitation sources driving structural responses, identifying the coherent structures responsible for specific surface-pressure (wind load) patterns. Flow– concentration synchronization disclosed real-time ventilation and pollutant-removal pathways in urban settings, revealing removal flux and quantifying instantaneous contributions. Algorithmic advances established input-convergence criteria and new modal-classification strategies based on both energy content and evolutionary significance. Most recently, our developments in spectral POD resolved key limitations of cross-spectral POD, yielding higher-accuracy wind-signal simulations and enabling high-dimensional inputs.

Leveraging the powerful orthogonal decompositions, our research ahead aims to address larger, more consequential, and multi-faceted challenges involving wind, extreme heat, pollution, and human interaction. Microclimate, wind energy, extreme weather, vehicle aerodynamics & aeroacoustics, and the increasingly complex behavior of wind under climate change continue to challenge our international wind engineering community.