Search

The differentiable programming paradigm is a cornerstone of modern scientific computing. It refers to numerical methods for computing the gradient of a numerical model's output. Many scientific models are based on differential equations, where differentiable programming plays a crucial role in calculating model sensitivities, inverting model parameters, and training hybrid models that combine differential equations with data-driven approaches. Furthermore, recognizing the strong synergies between inverse methods and machine learning offers the opportunity to establish a coherent framework applicable to both fields. Differentiating functions based on the numerical solution of differential equations is non-trivial. Numerous methods based on a wide variety of paradigms have been proposed in the literature, each with pros and cons specific to the type of problem investigated. Here, we provide a comprehensive review of existing techniques to compute derivatives of numerical solutions of differential equations. We first discuss the importance of gradients of solutions of differential equations in a variety of scientific domains. Second, we lay out the mathematical foundations of the various approaches and compare them with each other. Third, we cover the computational considerations and explore the solutions available in modern scientific software. Last but not least, we provide best-practices and recommendations for practitioners. We hope that this work accelerates the fusion of scientific models and data, and fosters a modern approach to scientific modelling.

Numerical models of the ocean and ice sheets are crucial for understanding and simulating the impact of greenhouse gases on the global climate. Oceanic processes affect phenomena such as hurricanes, extreme precipitation, and droughts. Ocean models rely on subgrid-scale parameterizations that require calibration and often significantly affect model skill. When model sensitivities to parameters can be computed by using approaches such as automatic differentiation, they can be used for such calibration toward reducing the misfit between model output and data. Because the SOMA model code is challenging to differentiate, we have created neural network-based surrogates for estimating the sensitivity of the ocean model to model parameters. We first generated perturbed parameter ensemble data for an idealized ocean model and trained three surrogate neural network models. The neural surrogates accurately predicted the one-step forward ocean dynamics, of which we then computed the parametric sensitivity.

The Massachusetts Institute of Technology General Circulation Model (MITgcm) is widely used by the climate science community to simulate planetary atmosphere and ocean circulations. A defining feature of the MITgcm is that it has been developed to be compatible with an algorithmic differentiation (AD) tool, TAF, enabling the generation of tangent-linear and adjoint models. These provide gradient information which enables dynamics-based sensitivity and attribution studies, state and parameter estimation, and rigorous uncertainty quantification. Importantly, gradient information is essential for computing comprehensive sensitivities and performing efficient large-scale data assimilation, ensuring that observations collected from satellites and in-situ measuring instruments can be effectively used to optimize a large uncertain control space. As a result, the MITgcm forms the dynamical core of a key data assimilation product employed by the physical oceanography research community: Estimating the Circulation and Climate of the Ocean (ECCO) state estimate. Although MITgcm and ECCO are used extensively within the research community, the AD tool TAF is proprietary and hence inaccessible to a large proportion of these users. The new version 2 (MITgcm-AD v2) framework introduced here is based on the source-to-source AD tool Tapenade, which has recently been open-sourced. Another feature of Tapenade is that it stores required variables by default (instead of recomputing them) which simplifies the implementation of efficient, AD-compatible code. The framework has been integrated with the MITgcm model main branch and is now freely available., Comment: 12 pages, 2 figures, 4 tables, submitted to Joint Laboratory on Extreme Scale Computing Future Generation Computer Systems (JLESC-FGCS)

Uncertainty in estimation of Drake Passage transport is analyzed in a Hessian-based uncertainty quantification (UQ) framework. The approach extends the adjoint-based ocean state estimation method to provide formal error bounds functionality. Mechanisms of uncertainty propagation in an idealized barotropic model of the Antarctic Circumpolar Current are identified by analysis of Hessian and Jacobian derivative operators, generated via algorithmic differentiation (AD) of the MIT ocean general circulation model (MITgcm). Inverse and forward uncertainty propagation mechanisms are identified, projecting uncertainty between observation, control and state variable domains. Time resolving analysis of uncertainty propagation captures the dynamics of uncertainty evolution and reveals transient and stationary uncertainty regimes. The UQ system resolves also the dynamical coupling of uncertainty across different physical fields, as represented by the off-diagonal posterior covariance structure. The spatial patterns of posterior uncertainty reduction and their temporal evolution are explained in terms of barotropic ocean dynamics. Global uncertainty teleconnection mechanisms are associated with barotropic wave propagation. Uncertainty coupling via data assimilation is demonstrated to dominate the reduction of Drake Passage transport uncertainty, highlighting the importance of correlation between different oceanic variables on the large scale., Comment: Submitted to Monthly Weather Review on Sep-28-2015

This paper discusses the application of an open source data analytics framework NASA is developing through its Advanced Information Systems Technology (AIST) program that is the platform to enable Sea Level research and estimating ocean circulation modeling.

Developments in ocean data assimilation (DA) and observing system technologies are intertwined. New observation types lead to new DA methods, and new DA methods such as Coupled Data Assimilation can change the value of existing observations or indicate where new observations can have greater utility for monitoring and prediction. Practitioners are encouraged to make better use of observations that are already available, for example in strongly coupled data assimilation where ocean observations can be used to improve atmospheric analyses and vice versa. Ocean reanalyses are useful for the analysis of climate,as well as initializing operational long-range prediction models. There are remaining challenges for ocean reanalyses due to biases and abrupt changes in the ocean observing system throughout its history, the presence of biases and drifts in models, and simplifying assumptions made in the DA methods. From a governance point of view, more support is needed to interface the observing community and the ocean DA community. For prediction applications, the ocean DA community must work with the ocean observing community to establish protocols for rapid communication of ocean observing data on NWP timescales. There is potential for new observations to enhance the observing system by supporting prediction on multiple timescales, ranging from the typical timescale of numerical weather prediction covering hours to weeks, out to multiple decades. It is highly encouraged that communication be fostered between thesecommunities to allow operational prediction centers the ability to provide guidance to the design of a sustained and adaptive observing network.

During the past few years the Goddard Earth Observing System (GEOS) and Massachusetts Institute of Technology (MIT) modeling groups have produced, respectively, global atmosphere-only and ocean-only simulations with km-scale grid spacing. These simulations have proved invaluable for process studies and for the development of satellite and in-situ sampling strategies. Nevertheless, a key limitation of these "nature" simulations is the lack of interactivity between the ocean and the atmosphere, which limits their usefulness for studying air-sea interactions and for designing observing missions to study these interactions. To remove this limitation, we aim to couple the km-scale GEOS atmosphere simulation to the km-scale MIT ocean simulation.

During the past few years the Goddard Earth Observing System (GEOS) and Massachusetts Institute of Technology (MIT) modeling groups have produced, respectively, global atmosphere-only and ocean-only simulations with km-scale grid spacing. These simulations have proved invaluable for process studies and for the development of satellite and in-situ sampling strategies. Nevertheless, a key limitation of these "nature" simulations is the lack of interaction between the ocean and the atmosphere, which limits their usefulness for studying air-sea interactions and for designing observing missions to study these interactions. We present here results from a coupled GEOS-MIT "nature run" simulation, wherein we have coupled a cubed-sphere-720 (~ 1/8) configuration of the GEOS atmosphere to a lat-lon-cap-1080 (~ 1/12) configuration of the MIT ocean. We compare near-surface diagnostics of this fully coupled ocean-atmosphere simulation to equivalent atmosphere-only and ocean-only simulations. A particular focus of the comparisons is the coupled versus uncoupled differences in interactions between Sea Surface Temperature (SST) and ocean surface wind. We discuss, in particular, a several-day mode of temporal variability in the SST-wind cycle and how it is represented in the different model simulations and in observationally-based products. A mechanism for the cycle, which is driven by SST-wind feedback, is proposed.

During the last two plus decades, The Goddard Earth Observing System (GEOS) and Massachusetts Institute of Technology (MIT) modeling groups have developed, respectively, atmosphere-only and ocean-only global general circulation models. These two models (GEOS and MITgcm) have demonstrated their data assimilation capabilities with the recent releases of the Modern Era Reanalysis for Research Applications, Version 2 (MERRA-2) atmospheric reanalysis and the Estimating the Circulation and Climate of the Ocean, Version 4 (ECCO-v4) ocean (and sea ice) state estimate. Independently, the two modeling groups have also produced global atmosphere-only and ocean-only simulations with km-scale grid spacing which proved invaluable for process studies and for the development of satellite and in-situ sampling strategies.Recently, a new effort has been made to couple these two models and to leverage their data-assimilation and high resolution capabilities (i.e., eddy-permitting ocean, cloud-permitting atmosphere). The focus in the model development is put on sub-seasonal to decadal time scales. In this talk, I discuss the new coupled model and present some first coupled simulation results. This will include a high-resolution coupled GEOS-MIT simulation, whereby we have coupled a cubed-sphere-720 (~ 1/8 deg) configuration of the GEOS atmosphere to a lat-lon-cap-1080 (~ 1/12 deg) configuration of the MIT ocean. We compare near-surface diagnostics of this fully coupled ocean-atmosphere set-up to equivalent atmosphere-only and ocean-only simulations. In the comparisons we focus in particular on the differences in air-sea interactions between sea surface temperature (SST) and wind for the coupled and uncoupled simulations.

During the last two plus decades, The Goddard Earth Observing System (GEOS) and Massachusetts Institute of Technology (MIT) modeling groups have developed, respectively, atmosphere-only and ocean-only global general circulation models. These two models (GEOS and MITgcm) have demonstrated their data assimilation capabilities with the recent releases of the Modern Era Reanalysis for Research Applications, Version 2 (MERRA-2) atmospheric reanalysis and the Estimating the Circulation and Climate of the Ocean, Version 4 (ECCO-v4) ocean (and sea ice) state estimate. Independently, the two modeling groups have also produced global atmosphere-only and ocean-only simulations with km-scale grid spacing which proved invaluable for process studies and for the development of satellite and in-situ sampling strategies.Recently, a new effort has been made to couple these two models and to leverage their data-assimilation and high resolution capabilities (i.e., eddy-permitting ocean, cloud-permitting atmosphere). The focus in the model development is put on sub-seasonal to decadal time scales. In this talk, I discuss the new coupled model and present some first coupled simulation results. This will include a high-resolution coupled GEOS-MIT simulation, whereby we have coupled a cubed-sphere-720 (~ 1/8°) configuration of the GEOS atmosphere to a lat-lon-cap-1080 (~ 1/12°) configuration of the MIT ocean. We compare near-surface diagnostics of this fully coupled ocean-atmosphere set-up to equivalent atmosphere-only and ocean-only simulations. In the comparisons we focus in particular on the differences in air-sea interactions between sea surface temperature (SST) and wind for the coupled and uncoupled simulations.

During the last two plus decades, The Goddard Earth Observing System (GEOS) and Massachusetts Institute of Technology (MIT) modeling groups have developed, respectively, atmosphere-only and ocean-only global general circulation models. These two models (GEOS and MIT-GCM (General Circulation Model)) have demonstrated their data assimilation capabilities with the recent releases of the Modern Era Reanalysis for Research Applications, Version 2 (MERRA-2) atmospheric reanalysis and the Estimating the Circulation and Climate of the Ocean, Version 4 (ECCO-v4) ocean (and sea ice) state estimate. Independently, the two modeling groups have also produced global atmosphere-only and ocean-only simulations with km-scale grid spacing which proved invaluable for process studies and for the development of satellite and in-situ sampling strategies. Recently, a new effort has been made to couple these two models and to leverage their data-assimilation and high resolution capabilities (i.e., eddy-permitting ocean, cloud-permitting atmosphere). The focus in the model development is put on sub-seasonal to decadal time scales. In this talk, I discuss the new coupled model and present some first coupled simulation results. This will include a high-resolution coupled GEOS-MIT simulation, whereby we have coupled a cubed-sphere-720 (approximately 1/8 degrees) configuration of the GEOS atmosphere to a latitude-longitude-cap-1080 (approximately 1/12 degrees) configuration of the MIT ocean. We compare near-surface diagnostics of this fully coupled ocean-atmosphere set-up to equivalent atmosphere-only and ocean-only simulations. In the comparisons we focus in particular on the differences in air-sea interactions between sea surface temperature (SST) and wind for the coupled and uncoupled simulations.