The Grell-Freitas (GF) deep convection scheme is implemented in the Community Atmosphere Model physics with the Model for Prediction Across Scales dynamical core

In aquaplanet simulations, GF alleviates the too frequent drizzle which is characteristic of the default Zhang-McFarlane (ZM) scheme

In variable-resolution simulations, GF exhibits less (more) resolution sensitivity for precipitation (cloud fraction) compared to ZM

Eddies are nonlinear in the circumpolar current and transition from nonlinear to linear regimes from upper to deep ocean in the gyres

Mixing in the nonlinear regime is well-represented by the rms eddy velocity times a depth-independent energy-containing scale

Mixing in the linear regime follows the vertical structure of the eddy kinetic energy times a depth-independent decay time scale

An large-eddy simulation model implemented with a spectral bin microphysics scheme is validated via Arctic mixed-phase cloud simulations

A refined Eulerian advection scheme for the non-mass characteristic variables of ice particle is developed to improve their conservation

Sensitivity tests reveal that the habit and aspect ratio of precipitating ice particles are sensitive to immersion freezing rates

Development of a new regional eddy-permitting ocean carbon cycle model is documented

The model output is validated against a suite of observations, including ship-based, satellite, and autonomous float data

The model captures the seasonality and trend of partial pressure of carbon dioxide in the surface water

An efficient and practical method to estimate time-varying model parameters is proposed

Particle filtering can be stabilized by combining offline batch optimization and online data assimilation

The new method is successfully applied to the Lorenz 63 model and a conceptual hydrological model

The co-variability of the atmospheric water and energy imbalances is examined in 3 different model configurations

We explain the connection of this co-variability to both the large-scale circulation and cloud properties

We demonstrate how this co-variability can be used to better understand clouds-circulations coupling

A simple two-mode aerosol scheme allows exploration of Aitken effects on aerosol and cloud evolution

Aitken aerosols supplement the accumulation mode and can prevent boundary-layer collapse in some cases

Scavenging of unactivated aerosol particles by cloud droplets has a strong effect on aerosol evolution

Temperature retrievals from radio occultation measurements can be used as ground truth to measure stratospheric model biases

3D convolutional neural networks are suitable for model bias estimation but do not outperform weak-constraint four-dimensional variational

Transfer learning can help to mitigate data limitations when the atmospheric model is upgraded

The discrete element method with bonded particles and physics-based fracture models can qualitatively capture the behavior of sea ice flowing through a channel

Fracture is captured with a non-local stress calculation and Mohr-Coulumb failure model to determine when inter-particle bonds fail

We use spatio-temporal scaling analyses to quantitatively assess the model's ability to capture key properties of sea ice deformation

Application of machine learning (ML) to solve the kinetic equations in global 3-D atmospheric chemistry models improves with online training of the ML solver

Stable yearlong global simulation of chemistry can be achieved with seasonal ML solvers applied to the Super-Fast mechanism in Goddard Earth Observing System (GEOS)-Chem

Computational speedup is five-fold relative to the reference Rosenbrock solver in GEOS-Chem

A new configuration of the Community Atmosphere Model with full chemistry supporting horizontal mesh refinement is developed

This configuration is the beginning of the Multi-Scale Infrastructure for Chemistry and Aerosols

Updating chemistry and horizontal resolution improves simulated ozone and ozone precursors compared to aircraft observations

We investigated the impact of the scale-aware convective parameterization scheme (CPS) across the gray-zone using the Weather Research and Forecasting model

The scale-aware CPS improved simulated convective cells related to rainfall by properly removing atmospheric instability in the gray-zone

Convective available potential energy timescale and entrainment rate modulated in the Multiscale Kain–Fritsch are the key parameters for the improved rainfall simulation in the gray-zone

Coupled atmosphere-ocean model reproduces observed Madden-Julian Oscillation (MJO) eastward propagation whereas uncoupled atmospheric model fails

MJO-induced upper-ocean cooling is essential in reducing prolonged rainfall over the Indian Ocean and aiding in MJO's eastward propagation

Improved air-sea flux algorithm reduces model biases in rainfall, surface wind, and upper ocean temperature

A single category, four moment scheme simulates autoconversion and accretion far better than a two category, two moment scheme

The rain mode forms at diameters that are much smaller than are traditionally considered to be rain

Using one versus two liquid categories is more important than assumptions about drop size distributions

Model simulations of Madden-Julian Oscillation (MJO) wind and rain are very sensitive to model grid resolution and moist physics

Cumulus parameterization and low model resolution affect atmospheric boundary layer processes that can hinder MJO simulations

Cloud-permitting resolution (4 km) and explicit moist physics significantly improve the model prediction of MJO precipitation, wind, and eastward propagation

A complete real-time workflow for 30-s-update 500-m-mesh numerical weather prediction for convective precipitation was developed

The system assimilates data from a phased array weather radar every 30 s and performs 30-min extended forecasts in real time

The forecast accuracy is improved by applying a new thinning method to the dense radar data

The anelastic System for Atmospheric Modeling is extended to a global latitude-longitude grid

The model uses a novel treatment of topography forcing the flow to instantaneously stagnate in the cells inside a prescribed terrain

The model performs well in several standard tests that evaluate the accuracy of global models and in real Earth simulations

In this study we illustrate how the Atmosphere-Ocean Single-Column Model (AOSCM) can be set up and utilized to explore model sensitivity to various uncertain parameters

We use observations from a warm air Arctic intrusion over melting sea ice to showcase how the AOSCM can be applied

A linear transformation for observations and state variables uniquely diagonalizes the Kalman filter update equation

A set of positive real numbers rank order the importance of transformed observations and state variables in the update

In the new variables, optimal covariance localization consists of setting covariances to zero and retaining variances

Controls of high and mid-level clouds are independent of whether a model employs parameterized or explicit convection

The response of anvil cloud fraction to a warming surface and more stable troposphere is highly controlled by radiatively-driven divergence

Mid-level clouds robustly decrease with warming due to integrated cooling increasing slower than convective heat flux

Demonstrate the GEOS-CF system is capable of supporting NASA science missions and applications which observe stratospheric composition

The GEOS-CF model produces realistic stratospheric ozone forecasts, a new capability during anomalous polar vortex conditions

Spatial patterns of the GEOS-CF simulated concentrations of stratospheric composition agree well with independent observations