Because of the formidable challenge of observing the full-depth global ocean circulation in its spatial detail and the many time scales of oceanic motions, numerical simulations play an essential role in quantifying patterns of climate variability and change. For the same reason, predictive capabilities are confounded by the high-dimensional space of uncertain inputs required to perform such simulations. Inverse methods optimally extract and blend information from observations and models. Formal parameter calibration and state estimation, in particular, enables rigorously calibrated and initialized predictive models to optimally learn from sparse, heterogeneous data while satisfying fundamental equations of motion.

The CRIOS group is a core member of the NASA-funded Estimating the Circulation and Climate of the Ocean (ECCO) consortium. The project seeks to produce best possible estimates of the time-evolving global ocean and sea ice circulation over the last several decades through combining all major satellite and in-situ observational networks that form the Global Ocean Observing System (GOOS) into a coherent dynamical framework provided by an ocean model, the Massachusetts Institute of Technology general circulation model (MITgcm). ECCO’s assimilation framework is based on the adjoint method, one of the arguably most advanced data assimilation methods. The formal process of achieving this data-model synthesis is called optimal estimation, or, more loosely, data assimilation. ECCO covers the period 1992 onward, the beginning of which coincides with the first dedicated ocean satellite altimetric mission (TOPEX/POSEIDON) and the decade of the World Ocean Circulation Experiment (WOCE).

• Related CRIOS publications:

• ECCO Consortium et al., Synopsis of the ECCO Central Production Global Ocean and Sea-Ice State Estimate (2020);
• Heimbach et al., Frontiers in Marine Science (2019);
• Fukumori et al., Bull. Amer. Met. Soc. (2018);
• Stammer et al., Ann. Rev. Mar. Sci. (2016);
• Forget et al., Geosci. Model Dev. (2015).

Observations made since the turn of the 21st century indicate that the Arctic physical environment and associated ecosystems are undergoing remarkable changes. To understand these ongoing and predict future changes we need to be able to quantify how much of Arctic climate variability and change is linked to the ocean circulation, how the distribution of sea ice responds to the amplified Arctic warming, and how primary productivity, species distributions and associated ecosystem structures will change with the evolving Arctic state. Our understanding of the driving mechanisms for these processes depends on our knowledge of the Arctic system and our ability to simulate its time-evolving state in a way that is consistent with observations.

The NSF-supported project Arctic Sub-polar gyre sTate Estimate (ASTE) project employs an advanced data assimilation approach for synthesizing oceanographic observations to develop a tool that establishes a baseline understanding of the coupled Arctic ocean-sea ice system. Through this computational tool, we are developing a better quantification of the Arctic system’s changes over the 21st century, and an improved understanding of the dynamical causes for these changes.

This project brings together hydrographic observations obtained from ship, aircraft and autonomous instruments, sea ice data from in situ buoys, satellites and manned vessels, and boundary-layer meteorology within a state estimation framework. The joint interpretation of these heterogeneous data streams enables us to develop a more coherent physical and dynamical description of the evolving Arctic Ocean during the 21st century. Taking advantage of a growing observational database, the project’s two main objectives are: the creation of a dynamically-consistent and comprehensive synthesis of all available Arctic hydrographic and sea ice observations using the adjoint-based Arctic Sub-polar gyre sTate Estimate for the period 2002-present; and scientific investigations, using the synthesis products, of physical mechanisms driving the evolving Arctic system.

• Related CRIOS publications:

• Nguyen et al., J. Adv. Model. Earth Syst. (2021);
• Nguyen et al., J. Atmos. Ocean Technol. (2020);
• Nguyen et al., J. Geophys. Res. (2020);
• Lee et al., Front. Mar. Sci. (2019);
• Bigdeli et al., J. Geophys. Res. (2018);
• Nguyen et al., Oceanography (2017).

To understand the Earth’s changing climate, it is important to estimate how much heat the ocean takes up from the atmosphere and how the ocean recirculates the heat around the globe. Directly obtaining these estimates from measurements is complicated because oceanographers cannot measure the ocean everywhere and at all times. Ocean measurements taken from ships or freely drifting instruments are expensive and difficult to obtain, especially in regions with ice cover or rough weather conditions. Given the high cost and logistical challenges of ocean instrument deployment, ocean observing systems have to be designed carefully. Key questions are: What information is contained in already existing observation networks? Is there any redundancy in current configurations? What do existing systems, such as the Atlantic mooring arrays, tell us about hydrographic and circulation quantities in remote oceanic regions with few observations?

To tackle these questions, our group performs quantitative observing system design, through a combination of adjoint modeling, Bayesian inverse methods, and Hessian uncertainty quantification. By means of these computational tools, quantitative observing system design suggests an optimal observing strategy and supports effective instrument placements in the future.

• Related CRIOS publications:

• Loose and Heimbach, J. Adv. Model. Earth Syst. (2021);
• Loose et al., J. Geophys. Res (2020);
• Fujii et al., Front. Mar. Sci. (2019); Kalmikov and Heimbach (2018).
• “Sensitive Spots … and How to Find Them,” ArcGis Storymap based on Loose et al., ECCO group website, uploaded August 2020.

The ocean’s overturning circulation redistributes heat, salt and dissolved chemicals around the globe, from the surface to great depths. Variations in overturning have been linked to significant climate change on a range of timescales, but the underlying mechanisms remain under investigation. Exploring the cause of observed variations and quantifying the relative importance of various external (atmospheric) forcings versus internal (eddy induced) processes is complex but essential for understanding past and present climate states and anticipating future change.

Most of our group’s research on this topic has aimed to unambiguously attribute variations in net overturning and its constituent currents to past changes in external forcing. Our main tool for this task is the ECCO state estimate and inbuilt adjoint infrastructure, enabling efficient mapping of the timescales and pathways of overturning sensitivity to prior atmospheric changes. These sensitivity distributions provide insights into the underlying adjustment physics and - via convolution with historic atmospheric forcing estimates - enable contributions from wind, heat and freshwater forcing to be both seperately quantified and pinpointed in space and time.

• Related CRIOS publications:

• Kostov et al., Nature Geosci. (2021);
• Frajka-Williams et al., Frontiers in Marine Science (2019);
• Smith and Heimbach, J. Clim. (2019);
• Pillar et al., J. Clim. (2018);
• Lozier et al., Bull. Amer. Met. Soc. (2017);
• Pillar et al., J. Clim. (2016);
• Buckley et al., J. Clim. (2015).

The Antarctic ice sheet gains and loses mass respectively via snowfall and gravity driven ice streams which flow from land to the sea. Ice shelves form where these streams reach the sea, floating over the ocean due to the density difference between ice and seawater. These formations are critical to the climate system since, via lateral buttressing, they hold back continental land ice which can contribute to sea level rise when melted or discharged into the ocean. Ice shelves in the Amundsen Sea Embayment region of the West Antarctic Ice Sheet have some of the highest melt rates in Antarctica, as they are susceptible to relatively warm ocean currents at depth, and hold back enough ice to raise global sea levels by an estimated 1.2 meters. However, it is difficult and expensive to observe the behavior of these ice shelves because the Amundsen Sea Embayment is remote and hostile to human activity. It is therefore common to infer the relationship between nearby oceanic conditions and ice shelf melting qualitatively from sparse observations.

Our research group is focused on quantifying the degree to which the ocean circulation influences melt rates, through the use of the adjoint capabilities of the Massachusetts Insitute of Technology General Circulation Model (MITgcm). Additionally, we perform Bayesian calibration of key parameters and quantities which influence the ice-ocean interactions, such that our models of sub-iceshelf circulation fit observations and follow the governing theory of the general circulation models. This allows us to quantify the degree to which these sparse and difficult to obtain observations constrain our underlying model, and suggest an optimal observing strategy.

• Related CRIOS publications: Antarctica - Goldberg et al., Earth and Space Science Open Archive (2020); Goldberg et al., Ocean Modelling (2018), Jordan et al., J. Geophys. Res. (2018); Goldberg et al., The Cryosphere (2015). Greenland - Straneo et al., Front. Mar. Sci. (2019); Wilson et al., The Cryosphere (2017); Schulz et al., GRL (2022).

The science of sea level is surprisingly complex as it connects a wide range of phenomena of the coupled Earth system. Processes contributing to sea level change include ocean dynamics and thermodynamics driven by atmospheric, hydrologic, and cryospheric forcings. They also include gravitational self-attraction, solid Earth deformation, and Earth rotation response to ice mass and terrestrial water storage changes, and vertical land motion resulting from local processes such as tectonics and groundwater extraction.

Understanding and predicting global and regional sea level rise and coastal flooding has enormous societal implications, with an estimated 230 million people today living near the coast at less than 1 m above current high tide lines. Among others, CRIOS members are part of the NASA Sea Level Change Science Team (N-SLCT), which seeks to address one of the Science and Applications Priority Areas of the U.S. National Academies’ Decadal Strategy for Earth Observation from Space (2018). Regional Sea-Level Change and Coastal Impacts has also been identified as a Grand Challenge of the World Climate Research Program (WCRP).

• Related CRIOS publications: Hamlington et al., Rev. Geophys. (2020); Moon et al., Environment Magazine (2020); Trossman et al., J. Geophys. Res. (2019); Zanna et al., Proc. Natl. Acad. Sci. USA (2019); Sonnewald et al., J. Clim. (2018); Storto et al., Clim. Dyn. (2017); Ocana et al., J. Geophys. Res. (2016).

The Greenland ice sheet has emerged as a significant contributor to present-day,and possibly future global sea level rise. Predictions of mass loss are hampered by the lack of comprehensive observational constraints for model calibration, rigorous computational approaches to perform such calibration, transient simulations that evolve into a realistic present state of an ice sheet that may serve as initial condition for predictions, and accurate histories of past surface climates that may serve as surface forcing boundary conditions in ice sheet model simulations. A rigorous approach for treating the issues in combination is through Bayesian inverse methods based on a time-evolving dynamical model and sparse observations. The advent of observational data consisting of stratigraphic horizons of constant age (also known as radar layer or isochrone data) that have been imaged by numerous ice-penetrating radar surveys serve as a powerful paleo proxy data constraint on the past flow and boundary conditions.

The inverse problem poses a fundamental challenge: can we infer initial and surface boundary conditions, such that simulations based on these inputs match the available observations or proxy data? Solving this inverse problem will enable us to make several key scientific advances. Firstly, we will develop a time-evolving state of the Greenland ice sheet that is consistent with the time history encapsulated in the isochrone data. Secondly, we will infer a set of optimal model parameters (i.e., model calibration) and time-varying optimal surface forcings that are consistent with the underlying ice sheet evolution (i.e., reconstruction of a consistent climate forcing history). And, thirdly, we will develop a realistic transient simulation of the ice sheet over the Holocene period, leading into a present state that may serve as well-balanced initial condition for prediction, and for which initialization shocks or artificial model drift are expected to be minimal, by construction. A key computational ingredient that enables our proposed work is the development of a time-dependent adjoint model of an ice sheet model that is able to evolve internal ice age structures.

• Related CRIOS publications: Logan et al., Geosci. Model Dev. (2020).