Clustering to Characterize Extreme Marine Conditions for the Benthic Region of the Northeastern Pacific Continental Margin

Anthropogenic ${\text{CO}}_2$ emissions lead to ocean warming, deoxygenation and acidification. Superimposed on long-term trends are episodic extremes of temperature, oxygen, and acidity. Here we present an innovative method for assessing single and compound extremes using a high-resolution regional model of the Northeastern Pacific Ocean. We use an unsupervised clustering approach to identify regions with similar habitat characteristics near the seafloor, define extreme thresholds seasonally using a fixed baseline (1996–2020) within each cluster, and quantify the fraction of ocean waters that exceed these thresholds for both single and compound stressors. Compound extremes (most commonly of $O_2$ and acidification) are rare but show an increasing trend in some clusters. Potential predictability of occurrence of extremes is demonstrated by correlation with basin-scale climate variability.