Present-day North Atlantic salinity constrains future warming of the Northern Hemisphere

The Atlantic Meridional Overturning Circulation (AMOC) plays a crucial role in regulating global surface temperature by transporting heat and carbon from the tropics to the North Atlantic (NA). Future global warming is expected to weaken the AMOC, impacting ocean heat transport and increasing freshwater flux into the NA, leading to changes in atmospheric and oceanic circulation. The NA is a significant carbon sink, responsible for about 25% of the global ocean's anthropogenic carbon inventory. Carbon uptake in this region is closely tied to AMOC intensity, influencing the solubility and subduction of carbon. Greater carbon absorption by the NA could mitigate global warming. However, considerable inter-model spread exists in projections of NA carbon uptake among Earth System Models (ESMs), partly due to variations in simulated AMOC intensity. This uncertainty hinders accurate projections of future global warming and related impacts like sea level rise and extreme weather events. To address this, the study uses present-day sea surface salinity (SSS) as an emergent constraint to reduce the uncertainty in NA carbon uptake and its impact on Northern Hemisphere warming.

Previous research highlights the significant role of the AMOC in regulating global climate and the expected decline in its intensity under future warming. Studies have explored the impacts of AMOC weakening on various climate phenomena, including shifts in the Intertropical Convergence Zone and the emergence of a "warming hole" in the North Atlantic subpolar region. The NA's role as a major carbon sink and its connection to AMOC intensity have also been extensively studied, showing that greater carbon uptake in the NA could potentially delay global warming. However, substantial uncertainties remain in the projections of future NA carbon uptake due to inter-model discrepancies in the simulation of AMOC intensity. This has motivated research into finding constraints to better predict future climate scenarios, highlighting the importance of reducing uncertainties in carbon uptake.

This study utilized data from 31 Earth System Models (ESMs) from CMIP6 (18 models) and CMIP5 (13 models), including those with coupled ocean biogeochemistry schemes. The analysis focused on the period from 1850 to 2100, covering historical and future climate projections under the SSP5-8.5 scenario. The researchers employed a total least squares (TLS) method for linear regression, minimizing the perpendicular distance between data points and the regression line to account for correlated errors between variables. A multivariate probability distribution function (PDF) was used to estimate the uncertainty range, considering the relationships between all variables. The global heat-carbon coupling parameter (α) was calculated by regressing anthropogenic ocean carbon storage against ocean heat storage. An emergent constraint was applied to reduce uncertainty in future projections. This involved combining a linear regression between present-day SSS and future climate variables (NH surface temperature and cumulative NA carbon uptake) with observational data of SSS from the World Ocean Atlas 2018 (WOA18) using conditional probability density functions. The AMOC stream function was calculated from model outputs, using the maximum value at 26°N below 50m as an index. The North Atlantic subpolar region was defined as 55°W-15°W and 45°N-65°N, and the Northern Hemisphere as 0-360°E and 10°N-90°N. Inter-model uncertainty was defined as one standard deviation of changes among the ESMs.

The study revealed a strong negative correlation between changes in Northern Hemisphere surface temperature and cumulative carbon uptake in the North Atlantic across CMIP6 ESMs. Models with greater NA carbon uptake projected weaker future warming, attributable to a reduction in the greenhouse effect. Present-day SSS in the NA subpolar region was identified as a reliable indicator of AMOC strength and a strong predictor of future NA carbon uptake and NH surface warming. Models with higher present-day SSS tended to simulate larger future carbon uptake. The observed SSS in the NA subpolar region (34.75 ± 0.04 psu from WOA18) was used as an emergent constraint, reducing the uncertainty in projected NH surface temperature warming from 5.9 ± 1.4 °C to 5.5 ± 1.0 °C (a 30% reduction) and increasing the cumulative NA carbon uptake from 50.1 ± 10.7 PgC to 54.7 ± 5.1 PgC (a 53% reduction) under the SSP5-8.5 scenario. Similarly, uncertainty in future global mean surface temperature was reduced by 23%. Out-of-sample testing with CMIP5 models supported these findings. The study also found that the relationship between present-day SSS and future NH warming was stronger in ESMs that included an ocean biogeochemical model component. Using the present-day AMOC strength as a constraint yielded a less robust relationship compared to using SSS.

The findings indicate that current ESMs may overestimate future NH warming and underestimate NA carbon uptake. The strong relationship between present-day SSS in the NA subpolar region and future warming suggests that this salinity serves as a valuable constraint for improving climate projections. The utilization of observed SSS significantly reduces the uncertainty in predicting future warming, highlighting the importance of sustained SSS observations for accurate climate projections. The stronger relationship in ESMs with ocean biogeochemistry suggests that it is crucial to use ESMs that accurately model the carbon cycle. This study contributes to improved understanding and improved prediction of future climate change, emphasizing the significant influence of NA oceanographic processes on global warming.

This research demonstrates a robust relationship between present-day North Atlantic subpolar SSS and future Northern Hemisphere warming, mediated by variations in anthropogenic carbon uptake in the North Atlantic. This relationship provides a valuable emergent constraint, significantly reducing uncertainty in future climate projections. Continued monitoring of SSS in this crucial region is critical for refining climate models and more accurately predicting future warming trends. Future research could investigate the underlying mechanisms driving this relationship in greater detail and explore other potential emergent constraints for improving climate projections.

The study relies on the use of Earth System Models (ESMs), which inherently have limitations and uncertainties. The accuracy of the emergent constraint depends on the reliability of both the model simulations and the observational data. Although out-of-sample testing was conducted, the findings may not be universally applicable across all climate scenarios or regions. Future research should explore the robustness of this constraint under various climate scenarios and emission pathways.