Improved prediction of dimethyl sulfide (DMS) distributions in the northeast subarctic Pacific using machine-learning algorithms

Abstract

Dimethyl sulfide (DMS) is a volatile biogenic gas with the potential to influence regional climate as a source of atmospheric aerosols and cloud condensation nuclei (CCN). The complexity of the oceanic DMS cycle presents a challenge in accurately predicting sea surface concentrations and sea–air fluxes of this gas. In this study, we applied machine-learning methods to model the distribution of DMS in the northeast subarctic Pacific (NESAP), a global DMS hot spot. Using nearly two decades of ship-based DMS observations, combined with satellite-derived oceanographic data, we constructed ensembles of 1000 machine-learning models using two techniques: random forest regression (RFR) and artificial neural networks (ANN). Our models dramatically improve upon existing statistical DMS models, capturing up to 62 % of observed DMS variability in the NESAP and demonstrating notable regional patterns that are associated with mesoscale oceanographic variability. In particular, our results indicate a strong coherence between DMS concentrations, sea surface nitrate (SSN) concentrations, photosynthetically active radiation (PAR), and sea surface height anomalies (SSHA), suggesting that NESAP DMS cycling is primarily influenced by heterogenous nutrient availability, light-dependent processes and physical mixing. Based on our model output, we derive summertime, sea–air flux estimates of 1.16 ± 1.22 Tg S in the NESAP. Our work demonstrates a new approach to capturing spatial and temporal patterns in DMS variability, which is likely applicable to other oceanic regions.