Modelling seawater pCO2 and pH in the Canary Islands region based on satellite measurements and machine learning techniques

Abstract

Recent advancements in remote sensing systems, combined with new machine-learning model-fitting algorithms, have enabled the estimation of seawater carbon dioxide partial pressure (pCO(2,sw)) and pH (pH(T,is)) in the waters around the Canary Islands (13-19 degrees W; 27-30 degrees N). Continuous time-series data collected from moored buoys and Voluntary Observing Ships (VOS) between 2019 and 2024 were used to train and validate the models, providing a robust observational basis for satellite-derived estimates. Among all models tested, bootstrap aggregation (bagging) performed best, achieving an RMSE of 2.0 mu atm (R-2>0.99) for pCO(2,sw) and 0.002 for pH(T,is). Multilinear regression (MLR), neural networks (NN) and categorical boosting (CatBoost) also showed good predictive skill, with RMSE values between 5.4 and 10 mu atm for pCO(2,sw) (360-481 mu atm) and 0.004-0.008 for pH(T,is) (7.97-8.07). Using the most reliable model, we identified an increasing trend in pCO(2,sw) of 3.51 +/- 0.31 mu atmyr(-1), exceeding the atmospheric CO2 growth rate (2.3 mu atmyr(-1)), alongside an acidification trend of -0.003 +/- 0.001 yr(-1). Over the 2019-2024 period, rising atmospheric CO2 and increasing sea surface temperatures (reaching up to 0.2 degrees C yr(-1) during the unprecedented 2023 marine heatwave) likely contributed to these trends. The Canary Islands region shifted from a weak CO2 source (0.90 Tg CO2 yr(-1)) in 2019 to 4.5 Tg CO2 yr(-1) in 2024. After 2022, eastern sites that previously acted as annual CO2 sinks became net sources.