Machine Learning Algorithm Reveals Oxygen Loss in the Agulhas Current

ABSTRACT:

A recent study conducted by a team of researchers, led by OP–DCC African Seas Regional Hub Ocean Forecasting Biogeochemistry Lead Dr. Thulwaneng Mashifane, has uncovered a significant decline in dissolved oxygen levels in the Agulhas Current, one of the strongest ocean currents in the Southern Hemisphere. Using a Machine Learning algorithm, the researchers linked ocean warming and decreasing oxygen solubility to the deoxygenation trend in the Agulhas Current, raising concerns about the broader impact on biogeochemistry and marine ecosystems in this region.

MAIN DESCRIPTION: 

The Agulhas Current, through the Agulhas Leakage, transports warm, saline waters from the Indian Ocean to the South Atlantic, playing a critical role in global ocean circulation. This new study applied a Machine Learning algorithm to predict and analyze oxygen levels in the Agulhas Current from 2000 to 2023, generating a 24–year data product. The findings reveal a deoxygenation trend of –2.29 ± 0.61 μmol kg-1 yr-1 across the Agulhas Current, representing a 1.47% deoxygenation rate per decade, which exceeds global averages from Machine Learning products.

The primary driver of this oxygen loss is rising sea surface temperatures, which reduce the ocean’s ability to hold oxygen. Seasonal changes in wind patterns and productivity further influence oxygen dynamics in this region. The study suggests that deoxygenation in the Agulhas Current may have implications for the South Atlantic Ocean and the Atlantic Meridional Overturning Circulation (AMOC) through the leakage of these warm, saline and deoxygenated which are eventually transported to the North Atlantic. These findings highlight the importance of ongoing monitoring and the availability of observations to understand, predict and anticipate the impacts of climate change on ocean health.

The dataset generated using the Machine Learning algorithm is available for download and use at the following link: https://zenodo.org/records/11001433.

USEFUL LINKS:

Article link: https://www.sciencedirect.com/science/article/pii/S0079661124002131

PARTNERS:

Council for Scientific and Industrial Research – Southern Ocean Carbon-Climate Observatory (CSIR–SOCCO)

Nansen-Tutu Centre for Marine Environmental Research

South African Environmental Observation Network (SAEON)

University of Cape Town

South African Weather Service

University of the Free State