Persistent Climate Model Biases in the Atlantic Ocean's Freshwater Transport

• Main Authors: Westen, René M., Dijkstra, Henk A.
• Format: Text
• Language: English
• Published: 2024
• Subjects: Indian; North Atlantic Deep Water; North Atlantic
• Online Access: https://doi.org/10.5194/egusphere-2023-1502; https://egusphere.copernicus.org/preprints/2023/egusphere-2023-1502/

The Atlantic Meridional Overturning Circulation (AMOC) is considered to be one of the most dangerous climate tipping elements. The salt–advection feedback plays an important role in AMOC tipping behaviour, and its strength is strongly connected to the freshwater transport carried by the AMOC at 34° S, below indicated by F ovS . Available observations have indicated that F ovS has a negative sign for the present-day AMOC. However, most climate models of the Coupled Model Intercomparison Project (CMIP, phase 3 and phase 5) have an incorrect F ovS sign. Here, we analyse a high-resolution and a low-resolution version of the Community Earth System Model (CESM) to identify the origin of these F ovS biases. Both CESM versions are initialised from an observed ocean state, and F ovS biases quickly develop under fixed pre-industrial forcing conditions. The most important model bias is a too fresh Atlantic Surface Water, which arises from deficiencies in the surface freshwater flux over the Indian Ocean. The second largest bias is a too saline North Atlantic Deep Water and arises through deficiencies in the freshwater flux over the Atlantic Subpolar Gyre region. Climate change scenarios branched from the pre-industrial simulations have an incorrect F ovS upon initialisation. Most CMIP phase 6 models have similar biases to those in the CESM. Due to the biases, the value of F ovS is not in agreement with available observations, and the strength of the salt advection feedback is underestimated. Values of F ovS are projected to decrease under climate change, and their response is also dependent on the various model biases. To better project future AMOC behaviour, an urgent effort is needed to reduce biases in the atmospheric components of current climate models.