The ocean plays a significant role in absorbing carbon dioxide from the atmosphere; however, determining how much carbon the ocean absorbs—and when, and where—is challenging since ocean observations to measure this are limited. To address this challenge researchers have used cutting edge machine learning algorithms to fill in observational pCO2 gaps, or used models to represent ocean processes.
A new study, funded in part by the Climate Observations and Monitoring program, and led by a team from Lamont Doherty Earth Observatory (LDEO) Columbia University (PI: McKinley), recently published a study that utilizes machine learning to merge these observation-based and model-based methods to improve the quantification of ocean carbon uptake. Lead author L. Gloege highlights this new LDEO-HPD pCO2 product and demonstrates that it agrees better with independent data than currently available products. The new product can be used as a valuable diagnostic and visualization tool to evaluate spatio-temporal model fields.
For more information, contact Virginia Selz.