Title: Dynamic Mode Decomposition with Control for Climate Analyses – Opportunities for Attribution

Abstract: Understanding the complex dynamics of climate patterns under varying anthropogenic emission scenarios is crucial for predicting future environmental conditions and formulating sustainable policies. Using Dynamic Mode Decomposition with control (DMDc), we analyze surface air temperature patterns, along with sea surface temperature, geopotential height, and precipitation data, from climate simulations to elucidate the effects of various climate-forcing agents. This method, enhanced by incorporating forcing information as a control variable, identifies both common modes of variability—such as the North Atlantic Oscillation and El Niño Southern Oscillation—and distinct impacts of aerosols and carbon emissions. Notably, the DMDc approach, employed during the ForceSMIP hackathon, reveals how these emissions’ effects differ across climate scenarios, particularly under higher radiative forcing conditions. Our findings underscore DMDc’s effectiveness in detecting forced responses and variability modes, providing valuable insights into changing spatial patterns under different forcing scenarios