Regional method to quantify coastal anthropogenic carbon changes

The global ocean plays a critical role in mitigating climate change by sequestering atmospheric CO₂, removing approximately 26% of anthropogenic carbon emissions since the Industrial Revolution. While significant progress has been made in estimating open-ocean anthropogenic carbon (C_anthro), the coastal ocean remains less understood due to its dynamic nature and complex processes and shortage of long-term high-quality datasets. Hence it is challenging to quantify the coastal anthropogenic carbon from the observation data. In this study, we propose a regional empirical regression-based anthropogenic carbon estimation approach (RECA) tailored for coastal regions. Using synthetic data from six different global ocean biogeochemical models, we evaluate the uncertainties in C_anthro estimation and assess the contributions of non-steady-state natural and anthropogenic components to estimation biases in the four North American coast oceans. We also compare RECA with established regression-based methods (CAREER and eMLR(C*)) that are widely used in open-ocean regions to determine their applicability in coastal settings. Our results demonstrate that RECA effectively captures overall C_anthro with minimal large-scale biases. However, subregional analyses reveal challenges in separating anthropogenic and natural CO₂ signals, emphasizing the influence of natural variability. This study provides a unified framework for high-resolution C_anthro estimation in coastal waters, evaluates its uncertainties, and paves the way for improved coastal carbon monitoring and climate action.