Our atmosphere is a chaotic system, where even a tiny change can have a big impact. This makes it difficult to assess if small changes, such as the move to a new hardware architecture, will significantly affect a weather and climate model. We present a methodology that allows to objectively verify this. The methodology is applied to several test cases, showing a high sensitivity. Results also show that a major system update of the underlying supercomputer did not significantly affect our model.

This paper describes a tool embedded in a global climate model for sampling atmospheric conditions and monitoring physical processes as a numerical simulation is being carried out. The tool facilitates process-level model evaluation by allowing the users to select a wide range of quantities and processes to monitor at run time without having to do tedious ad hoc coding.

We present a new terrestrial carbon cycle data assimilation system, built on the PEcAn model–data eco-informatics system, and its application for the development of a proof-of-concept carbon reanalysis product that harmonizes carbon pools (leaf, wood, soil) and fluxes (GPP, Ra, Rh, NEE) across the contiguous United States from 1986–2019. Here, we build on a decade of work on uncertainty propagation to generate the most complete and robust uncertainty accounting available to date.

In our study we quantify the ability of the future satellite sensor SPEXone, part of the NASA PACE mission, to estimate aerosol emissions. The sensor will be able to retrieve accurate information of aerosol light extinction and most importantly light absorption. We simulate SPEXone spatial coverage and combine it with an aerosol model. We found that SPEXone will be able to estimate species-specific (e.g. dust, sea salt, organic or black carbon, sulfates) aerosol emissions very accurately.

NOAA's National Air Quality Forecast Capability (NAQFC) continues to protect Americans from the harmful effects of air pollution, while saving billions of dollars per year. Here we describe and evaluate the development of the most advanced version of the NAQFC to date, which became operational at NOAA on 20 July 2021. The new NAQFC is based on a coupling of NOAA's operational Global Forecast System (GFS) version 16 with the Community Multiscale Air Quality (CMAQ) model version 5.3.1.

Numerical models are an important tool to assess the air quality in cities, as they can provide near-continouos data in time and space. In this paper, air pollution for an entire city is simulated at a high spatial resolution of 40 m. At this spatial scale, the effects of buildings on the atmosphere, like channeling or blocking of the air flow, are directly represented by diffuse obstacles in the used model CAIRDIO. For model validation, measurements from air-monitoring sites are used.

A single-column version of the global climate model CNRM-CM6-1 has been designed to ease development and validation of the model physics at the air–sea interface in a simplified environment. This model is then used to assess the ability to represent the sea surface temperature diurnal cycle. We conclude that the sea surface temperature diurnal variability is reasonably well represented in CNRM-CM6-1 with a 1 h coupling time step and the upper-ocean model resolution of 1 m.

The Organization of Tropical East Pacific Convection (OTREC) field campaign focuses on studying convection in the eastern Pacific and Caribbean. Observations obtained from dropsondes have been assimilated into the ECMWF model and compared to a model run in which sondes have not been assimilated. The model performs well in both simulations, but the assimilation of sondes helps to reduce the departure for pre-tropical-storm conditions. Variables important to studying convection are also studied.

The region of southern Africa (SAF) is highly vulnerable to the impacts of climate change and is projected to experience severe precipitation shortages in the coming decades. Reliable climatic information is therefore necessary for the optimal adaptation of local communities. In this work we show that regional climate models are reliable tools for the simulation of precipitation over southern Africa. However, there is still a great need for the expansion and maintenance of observational data.

This study aimed to improve runoff simulations and explore deep soil hydrological processes for a highly varying soil depth and complex terrain watershed in the Loess Plateau, China. The actual soil depths and river channels were incorporated into the model to better simulate the runoff in this watershed. The soil evaporation scheme was modified to better describe the evaporation processes. Our results showed that the model significantly improved the runoff simulations.

Large air quality and climate models require vast amounts of computational power. Machine learning tools like neural networks can be used to make these models more efficient, with the downside that their results might not make physical sense or be easy to interpret. This work develops a physically interpretable neural network that obeys scientific laws like conservation of mass and models atmospheric composition more accurately than a traditional neural network.

In numerical weather prediction, data assimilation is frequently utilized to enhance the accuracy of forecasts from equation-based models. In this work we use a machine learning framework that approximates a complex dynamical system given by the geopotential height. Instead of using an equation-based model, we utilize this machine-learned alternative to dramatically accelerate both the forecast and the assimilation of data, thereby reducing need for large computational resources.