Award: OCE-2038422

The Atlantic Meridional Overturning Circulation (AMOC) is a metric that summarizes the zonally integrated and vertically accumulated northward flow of water in the Atlantic Ocean. While integrating a range of dynamically distinct regimes, a useful summary is that the AMOC is characterized by northward flow of warm, saline waters of subtropical origin, and returns cold water at depth. Because the associated circulation contributes to the northward transport of heat in the global climate system, the AMOC is thought to have an appreciable impact on the climate over adjacent land masses, in particular the European and North African continents. Inflow of warm, saline water into the Arctic, either through the Barents Sea Opening or Fram Strait also determine hydrographic properties in the Arctic.

Comprehensive, spatial and temporally dense measurements of hydrographic properties that constitute the AMOC in its entirety, i.e., capturing its latitudinal and temporal variations, are virtually impossible to conduct, requiring the need to measure elements of the AMOC at a particular latitude or across logistically viable trans-basin sections. Two major such sections have been in place, the RAPID-MOCHA array in the subtropical Atlantic since 2004, and the Overturning in the Subpolar North Atlantic Program (OSNAP) array in the subpolar North Atlantic. These are complemented by the observational backbone of the Global Ocean Observing System, notably the Argo profiling program, shipborne full-depth hydrographic sections, and satellite altimetry. 

In moving to a logistically and financially viable observing system that covers time scales representing major climatic fluctuations in the Atlantic, enables detailed dynamical understanding of AMOC variability, and detects potential secular changes , two major questions that this project addresses are: (1) To what extent are non-observed “quantities of interest” or “climate indices” (e.g., regions of major oceanic heat convergence/divergence and their role in shaping the atmospheric state, heat transport across specific sections in the North Atlantic) captured by the existing (or to be designed) observing system? (2) Can redundancy and complementarity of observational assets be quantified in a way that may inform future observational sampling.

The project addressed these questions within the context of a comprehensive ocean state estimation (data assimilation) framework. A feature of the specific framework used is the availability of the derivatives of the underlying ocean model, i.e., codes for the model’s tangent linear and adjoint model. This gives access to algorithmic methods of uncertainty quantification based on Hessian (i.e., second derivative) information of the model-data misfit with respect to the uncertain control variables. Using the concept of “dynamical proxy potential” introduced by Loose et al. (JGR, 2020) and Loose and Heimbach (JAMES, 2021) enabled us to quantify the degree of shared adjustment processes (as mediated by oceanic Rossby and Kelvin waves) between unobserved quantities of interest (QoI) and given observational assets. This in turn provides a metric of uncertainty reduction in the QoI based on the information contained in the observational asset. For a larger number of observations, a measure of observational complementarity versus redundancy is also provided. The project extends previous work by considering the full ensemble of mooring locations underlying the OSNAP array, and the corresponding ensemble of adjoint sensitivities. Calculations are idealized in the sense that mooring data are vertically averaged over the water column in order to maintain computational feasibility. The results reveal a complex picture of complementarity versus redundancy among different mooring pairs. They are to be regarded as a pilot for what is a complex, multilevel optimization problem (also known as Optimal Experimental Design) and have been shared within the International CLIVAR AMOC Task Team for further analysis.

 

Last Modified: 04/09/2025
Modified by: Patrick Heimbach