Rebecca Gjini
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Rebecca Gjini has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
Andrew I. L. Williams
Scripps Institution of Oceanography, University of California San Diego, La Jolla, USA
7 shared publications- Duncan Watson-Parris
Scripps Institution of Oceanography and Halıcıoğlu Data Science Institute, University of California San Diego, La Jolla, CA, USA
7 shared publications - Ellen H. Davenport
Scripps Institution of Oceanography, University of California San Diego, La Jolla, USA
7 shared publications Eric Pham
Halıcıoğlu Data Science Institute, University of California San Diego, La Jolla, USA
7 shared publicationsJ. Varan Madan
Scripps Institution of Oceanography, University of California San Diego, La Jolla, USA
7 shared publicationsJared Brzenski
Scripps Institution of Oceanography, University of California San Diego, La Jolla, USA
7 shared publications
Latest Publications
- Open Access
- 22 September 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-31
Reply on RC2
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- 22 September 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-31
Open AccessReply on RC2
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- Open Access
- 22 September 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-31
Reply on RC1
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- 22 September 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-31
Open AccessReply on RC1
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- Open Access
- 1 September 2026
- Published by American Geophysical Union (AGU) in Journal of Advances in Modeling Earth Systems
- Vol. 18 (9)
The Ensemble Kalman Inversion Race
Ensemble Kalman methods were initially developed to solve nonlinear data assimilation problems in oceanography but are now popular in applications far beyond their original use cases. Of particular interest is climate model calibration. As hybrid physics and machine‐learning models advance, the number of parameters and complexity of parameterizations in climate models will continue to grow. To fully realize these advances, we must move from laborious hand‐tuning to calibration‐driven model development in rapid iteration cycles. Thus, robust calibration of these parameters plays an increasingly important role. We focus on learning climate model parameters by minimizing the misfit between modeled and observed climate statistics in an idealized setting. Ensemble Kalman methods are a natural choice for this problem because they are derivative‐free, scalable to high dimensions, and robust to noise caused by statistical observations. Given the many variants of ensemble methods proposed, an important question is: Which ensemble Kalman method should be used for climate model calibration ? To answer this question, we perform systematic numerical experiments to explore the relative computational efficiencies of several ensemble Kalman methods. The numerical experiments involve statistical observations of Lorenz‐type models of increasing complexity, frequently used to represent simplified atmospheric systems, with some featuring neural network parameterizations. For each test problem, several ensemble Kalman methods and a derivative‐based method “race” to reach a specified accuracy, and we measure the computational cost required to achieve the desired accuracy. We investigate how prior information and the parameter or data dimensions define the computational costs of the various methods.
- 1 September 2026
- Published by American Geophysical Union (AGU) in Journal of Advances in Modeling Earth Systems
- Vol. 18 (9)
Open AccessThe Ensemble Kalman Inversion Race
Ensemble Kalman methods were initially developed to solve nonlinear data assimilation problems in oceanography but are now popular in applications far beyond their original use cases. Of particular interest is climate model calibration. As hybrid physics and machine‐learning models advance, the number of parameters and complexity of parameterizations in climate models will continue to grow. To fully realize these advances, we must move from laborious hand‐tuning to calibration‐driven model development in rapid iteration cycles. Thus, robust calibration of these parameters plays an increasingly important role. We focus on learning climate model parameters by minimizing the misfit between modeled and observed climate statistics in an idealized setting. Ensemble Kalman methods are a natural choice for this problem because they are derivative‐free, scalable to high dimensions, and robust to noise caused by statistical observations. Given the many variants of ensemble methods proposed, an important question is: Which ensemble Kalman method should be used for climate model calibration ? To answer this question, we perform systematic numerical experiments to explore the relative computational efficiencies of several ensemble Kalman methods. The numerical experiments involve statistical observations of Lorenz‐type models of increasing complexity, frequently used to represent simplified atmospheric systems, with some featuring neural network parameterizations. For each test problem, several ensemble Kalman methods and a derivative‐based method “race” to reach a specified accuracy, and we measure the computational cost required to achieve the desired accuracy. We investigate how prior information and the parameter or data dimensions define the computational costs of the various methods.
- Open Access
- 16 August 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-31
Comment on egusphere-2026-2871
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- 16 August 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-31
Open AccessComment on egusphere-2026-2871
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- Open Access
- 24 July 2026
- Published by Copernicus GmbH
- p. 1-31
Comment on egusphere-2026-2871
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- 24 July 2026
- Published by Copernicus GmbH
- p. 1-31
Open AccessComment on egusphere-2026-2871
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- Open Access
- 17 July 2026
- Published by Copernicus GmbH in Geoscientific Model Development
- Vol. 19 (14) , 6451-6466
JCM v1.1: a differentiable, intermediate-complexity atmospheric model
In this paper we present version 1.1 of the JAX Circulation Model (JCM). JCM is a differentiable atmospheric model that is built in Python with the JAX numerical library and is released under a free and open-source software (FOSS) license. Earth system modeling is rapidly evolving, particularly through hybrid approaches that combine known dynamics with data-driven components. However, the training and validation of hybrid methods in traditional models remain difficult due to the absence of gradients and the complexity of legacy code. Differentiable models written in modern frameworks offer a path forward. JCM couples physics parameterizations to the Dinosaur dynamical core through a flexible interface that makes substitution of other schemes easy. The default parameterization scheme uses the SPEEDY (Simplified Parameterizations, primitivE-Equation DYnamics) intermediate-complexity physics scheme. This modularity supports benchmarking across physical and machine-learned schemes, with direct access to gradients for sensitivity analysis, calibration, and online learning. We show validation of JCM against the original Fortran SPEEDY code at T31 resolution. We also highlight JCM's differentiability and efficiency and outline plans for extending the framework to a differentiable Earth system model. JCM provides a lightweight yet expressive platform for accelerating research in climate modeling.
- 17 July 2026
- Published by Copernicus GmbH in Geoscientific Model Development
- Vol. 19 (14) , 6451-6466
Open AccessJCM v1.1: a differentiable, intermediate-complexity atmospheric model
In this paper we present version 1.1 of the JAX Circulation Model (JCM). JCM is a differentiable atmospheric model that is built in Python with the JAX numerical library and is released under a free and open-source software (FOSS) license. Earth system modeling is rapidly evolving, particularly through hybrid approaches that combine known dynamics with data-driven components. However, the training and validation of hybrid methods in traditional models remain difficult due to the absence of gradients and the complexity of legacy code. Differentiable models written in modern frameworks offer a path forward. JCM couples physics parameterizations to the Dinosaur dynamical core through a flexible interface that makes substitution of other schemes easy. The default parameterization scheme uses the SPEEDY (Simplified Parameterizations, primitivE-Equation DYnamics) intermediate-complexity physics scheme. This modularity supports benchmarking across physical and machine-learned schemes, with direct access to gradients for sensitivity analysis, calibration, and online learning. We show validation of JCM against the original Fortran SPEEDY code at T31 resolution. We also highlight JCM's differentiability and efficiency and outline plans for extending the framework to a differentiable Earth system model. JCM provides a lightweight yet expressive platform for accelerating research in climate modeling.
- Open Access
- 18 June 2026
- Published by Copernicus GmbH in EGUsphere
- Vol. 2026
Technical note: Evaluation of conceptual predator-prey models for the quantitative modeling of precipitating open-cell stratocumulus via feature-based Bayesian inversion of a suite of Large eddy simulations
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- 18 June 2026
- Published by Copernicus GmbH in EGUsphere
- Vol. 2026
Open AccessTechnical note: Evaluation of conceptual predator-prey models for the quantitative modeling of precipitating open-cell stratocumulus via feature-based Bayesian inversion of a suite of Large eddy simulations
We consider two very different types of models of precipitating open-cell stratocumulus clouds. The first model type is a computationally expensive large eddy simulation (LES), that resolves convection and clouds at high temporal and spatial resolutions. The second model type is the nonlinear cloud and rain (C&R) equation, a scalar delay differential equation (DDE) that interprets interactions of precipitation and clouds phenomenologically by predator (rain) and prey (cloud) dynamics. We evaluate the extent to which one may use the C&R equation as a quantitative tool for representing selected aspects of an LES. Specifically, we estimate parameters of the C&R equation from a suite of LES via feature-based Bayesian inversions and track the evolution of posterior distributions over C&R model parameters under changing meteorological conditions in the LES. Our inversions show that the C&R equation can be calibrated to generate limit cycles that are quantitatively compatible with cycles of cloud growth and decay across a wide spectrum of meteorological conditions. The successful inversions reiterate the robustness of the predator-prey analogy to the dynamics of precipitating open-cell stratocumulus. When we interpret the inversions jointly, however, we observe counterintuitive and partially nonphysical shifts and changes in the posterior distributions over the C&R model parameters. Our evaluation study thus highlights the challenges one faces when mapping LES dynamics to a scalar DDE, which can stem either from structural inadequacies in the DDE model, or from the specific feature-based inversion framework, or a mixture of both.
- Open Access
- 23 March 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-20
Comment on egusphere-2025-6266
In this paper we present version 1.0 of the JAX Circulation Model (JCM). JCM is an open-source, differentiable atmospheric model built in Python using the JAX numerical library. Earth system modeling is rapidly evolving, particularly through hybrid approaches that combine known dynamics with data-driven components. However, the training and validation of hybrid methods in traditional models remain difficult due to the absence of gradients and the complexity of legacy code. Differentiable models written in modern frameworks offer a path forward. JCM couples physics parameterizations to the Dinosaur dynamical core through a flexible interface that makes substitution of other schemes easy. The default parameterization scheme uses the SPEEDY (Simplified Parameterizations, primitivE-Equation DYnamics) intermediate-complexity physics scheme. This modularity supports benchmarking across physical and machine-learned schemes, with direct access to gradients for sensitivity analysis, calibration, and online learning. We show validation of JCM against the original Fortran SPEEDY code at T31 resolution. We also highlight JCM's differentiability and efficiency and outline plans for extending the framework to a differentiable Earth system model. JCM provides a lightweight yet expressive platform for accelerating research in climate modeling.
- 23 March 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-20
Open AccessComment on egusphere-2025-6266
In this paper we present version 1.0 of the JAX Circulation Model (JCM). JCM is an open-source, differentiable atmospheric model built in Python using the JAX numerical library. Earth system modeling is rapidly evolving, particularly through hybrid approaches that combine known dynamics with data-driven components. However, the training and validation of hybrid methods in traditional models remain difficult due to the absence of gradients and the complexity of legacy code. Differentiable models written in modern frameworks offer a path forward. JCM couples physics parameterizations to the Dinosaur dynamical core through a flexible interface that makes substitution of other schemes easy. The default parameterization scheme uses the SPEEDY (Simplified Parameterizations, primitivE-Equation DYnamics) intermediate-complexity physics scheme. This modularity supports benchmarking across physical and machine-learned schemes, with direct access to gradients for sensitivity analysis, calibration, and online learning. We show validation of JCM against the original Fortran SPEEDY code at T31 resolution. We also highlight JCM's differentiability and efficiency and outline plans for extending the framework to a differentiable Earth system model. JCM provides a lightweight yet expressive platform for accelerating research in climate modeling.
- Open Access
- 20 March 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-20
Comment on egusphere-2025-6266
In this paper we present version 1.0 of the JAX Circulation Model (JCM). JCM is an open-source, differentiable atmospheric model built in Python using the JAX numerical library. Earth system modeling is rapidly evolving, particularly through hybrid approaches that combine known dynamics with data-driven components. However, the training and validation of hybrid methods in traditional models remain difficult due to the absence of gradients and the complexity of legacy code. Differentiable models written in modern frameworks offer a path forward. JCM couples physics parameterizations to the Dinosaur dynamical core through a flexible interface that makes substitution of other schemes easy. The default parameterization scheme uses the SPEEDY (Simplified Parameterizations, primitivE-Equation DYnamics) intermediate-complexity physics scheme. This modularity supports benchmarking across physical and machine-learned schemes, with direct access to gradients for sensitivity analysis, calibration, and online learning. We show validation of JCM against the original Fortran SPEEDY code at T31 resolution. We also highlight JCM's differentiability and efficiency and outline plans for extending the framework to a differentiable Earth system model. JCM provides a lightweight yet expressive platform for accelerating research in climate modeling.
- 20 March 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-20
Open AccessComment on egusphere-2025-6266
In this paper we present version 1.0 of the JAX Circulation Model (JCM). JCM is an open-source, differentiable atmospheric model built in Python using the JAX numerical library. Earth system modeling is rapidly evolving, particularly through hybrid approaches that combine known dynamics with data-driven components. However, the training and validation of hybrid methods in traditional models remain difficult due to the absence of gradients and the complexity of legacy code. Differentiable models written in modern frameworks offer a path forward. JCM couples physics parameterizations to the Dinosaur dynamical core through a flexible interface that makes substitution of other schemes easy. The default parameterization scheme uses the SPEEDY (Simplified Parameterizations, primitivE-Equation DYnamics) intermediate-complexity physics scheme. This modularity supports benchmarking across physical and machine-learned schemes, with direct access to gradients for sensitivity analysis, calibration, and online learning. We show validation of JCM against the original Fortran SPEEDY code at T31 resolution. We also highlight JCM's differentiability and efficiency and outline plans for extending the framework to a differentiable Earth system model. JCM provides a lightweight yet expressive platform for accelerating research in climate modeling.
- Open Access
- 11 March 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-20
Comment on egusphere-2025-6266
In this paper we present version 1.0 of the JAX Circulation Model (JCM). JCM is an open-source, differentiable atmospheric model built in Python using the JAX numerical library. Earth system modeling is rapidly evolving, particularly through hybrid approaches that combine known dynamics with data-driven components. However, the training and validation of hybrid methods in traditional models remain difficult due to the absence of gradients and the complexity of legacy code. Differentiable models written in modern frameworks offer a path forward. JCM couples physics parameterizations to the Dinosaur dynamical core through a flexible interface that makes substitution of other schemes easy. The default parameterization scheme uses the SPEEDY (Simplified Parameterizations, primitivE-Equation DYnamics) intermediate-complexity physics scheme. This modularity supports benchmarking across physical and machine-learned schemes, with direct access to gradients for sensitivity analysis, calibration, and online learning. We show validation of JCM against the original Fortran SPEEDY code at T31 resolution. We also highlight JCM's differentiability and efficiency and outline plans for extending the framework to a differentiable Earth system model. JCM provides a lightweight yet expressive platform for accelerating research in climate modeling.
- 11 March 2026
- Published by Copernicus GmbH
- Vol. 2026, 1-20
Open AccessComment on egusphere-2025-6266
In this paper we present version 1.0 of the JAX Circulation Model (JCM). JCM is an open-source, differentiable atmospheric model built in Python using the JAX numerical library. Earth system modeling is rapidly evolving, particularly through hybrid approaches that combine known dynamics with data-driven components. However, the training and validation of hybrid methods in traditional models remain difficult due to the absence of gradients and the complexity of legacy code. Differentiable models written in modern frameworks offer a path forward. JCM couples physics parameterizations to the Dinosaur dynamical core through a flexible interface that makes substitution of other schemes easy. The default parameterization scheme uses the SPEEDY (Simplified Parameterizations, primitivE-Equation DYnamics) intermediate-complexity physics scheme. This modularity supports benchmarking across physical and machine-learned schemes, with direct access to gradients for sensitivity analysis, calibration, and online learning. We show validation of JCM against the original Fortran SPEEDY code at T31 resolution. We also highlight JCM's differentiability and efficiency and outline plans for extending the framework to a differentiable Earth system model. JCM provides a lightweight yet expressive platform for accelerating research in climate modeling.