Air pollution is a major health hazard, and chemical transport models (CTMs) are valuable tools that aid in our understanding of the risks of air pollution at both local and regional scales. In this study, the Polair3D CTM of the Polyphemus air quality modeling platform was set up over Quebec, Canada, to assess the model’s capability in predicting key air pollutant species over the region, at seasonal temporal scales and at regional spatial scales.

Global Navigation Satellite Systems (GNSS) provides moisture observations through its densely distributed ground station network. In this research, we assimilate a new type of observation called tropospheric gradient observations, which has never been incorporated into a weather model. We develop a forward operator for gradient-based observations and conduct an assimilation impact study. The study shows significant improvements in the model's humidity fields.

Numerical models are widely used in air pollution modeling but suffer from significant biases. The machine learning model designed in this study shows high efficiency in identifying such biases. Meteorology (relative humidity and cloud cover), chemical composition (secondary organic components and dust aerosols), and emission sources (residential activities) are diagnosed as the main drivers of bias in modeling PM_2.5, a typical air pollutant. The results will help to improve numerical models.

An ensemble emission inventory is created with the aim of monitoring the status and progress made with the development of EU-wide inventories. This emission ensemble serves as a common benchmark for the screening and allows for the comparison of more than two inventories at a time. Because the emission “truth” is unknown, the approach does not tell which inventory is the closest to reality, but it identifies inconsistencies that require special attention.

This study is about the modelling of the atmospheric composition in Europe during the summer of 2022, when massive wildfires were observed. It is a sensitivity study dedicated to the relative impacts of two modelling processes that are able to modify the meteorology used for the calculation of the atmospheric chemistry and transport of pollutants.

In order to forecast localized warm-sector rainfall in the south China region, numerical weather prediction models are being run with finer grid spacing. The conventional convection parameterization (CP) performs poorly in the gray zone, necessitating the development of a scale-aware scheme. We propose a machine learning (ML) model to replace the scale-aware CP scheme. Evaluation against the original CP scheme has shown that the ML-based CP scheme can provide accurate and reliable predictions.

We developed a regionally refined climate model that allows resolved convection and performed a 20-year projection to the end of the century. The model has a resolution of 3.25 km in California, which allows us to predict climate with unprecedented accuracy, and a resolution of 100 km for the rest of the globe to achieve efficient, self-consistent simulations. The model produces superior results in reproducing climate patterns over California that typical modern climate models cannot resolve.

We have developed the Terrestrial Ecosystem Model in R (TEMIR), which simulates plant carbon and pollutant uptake and predicts their response to varying atmospheric conditions. This model is designed to couple with an atmospheric chemistry model so that questions related to plant–atmosphere interactions, such as the effects of climate change, rising CO₂, and ozone pollution on forest carbon uptake, can be addressed. The model has been well validated with both ground and satellite observations.

HARMONIE WINS50 reanalysis data with 0.025° × 0.025° resolution from 2019 to 2021 were coupled with the LOTOS-EUROS Chemical Transport Model. HARMONIE and ECMWF meteorology configurations against Cabauw observations (52.0° N, 4.9° W) were evaluated as simulated NO₂ concentrations with ground-level sensors. Differences in crucial meteorological input parameters (boundary layer height, vertical diffusion coefficient) between the hydrostatic and non-hydrostatic models were analysed.

This paper presents a method for calculating balloon drift from historical radiosonde ascent data. The drift can reach distances of several hundred kilometres and is often neglected. Verification shows the beneficial impact of the more accurate balloon position on model assimilation. The method is not limited to radiosondes but would also work for dropsondes, ozonesondes, or any other in situ sonde carried by the wind in the pre-GNSS era, provided the necessary information is available.

All accurate climate models use equations with poorly defined parameters, where knobs for the parameters are turned to fit the observations. This process is called tuning. In this article, we use another paradigm. We use a thermodynamic hypothesis, the maximum entropy production, to compute temperatures, energy fluxes, and precipitation, where tuning is impossible. For now, the  1D vertical model is used for a tropical atmosphere. The correct order of magnitude of precipitation is computed.

This paper describes a coupled atmosphere–mixed-layer ocean simulation setup that will be used to study weather processes in Southeast Asia. The set-up has been used to compare high-resolution simulations, which are able to partially resolve storms, to coarser simulations, which cannot. We compare the model performance at representing variability of rainfall and sea surface temperatures across length scales between the coarse and fine models.

Probabilistic precipitation nowcasting (local forecasting for 0–6 h) is crucial for reducing damage from events like flash floods. For this goal, we propose the DEUCE neural-network-based model which uses data and model uncertainties to generate an ensemble of potential precipitation development scenarios for the next hour. Trained and evaluated with Finnish precipitation composites, DEUCE was found to produce more skillful and reliable nowcasts than established models.

For modeling atmospheric chemistry, it is necessary to provide data on emissions of pollutants. These can come from various sources and in various forms, and preprocessing of the data to be ingestible by chemistry models can be quite challenging. We developed the FUME processor to use a database layer that internally transforms all input data into a rigid structure, facilitating further processing to allow for emission processing from the continental to the street scale.

We describe the multivariate static background error covariance (B) for the JEDI-MPAS 3D-Var data assimilation system. With tuned B parameters, the multivariate B gives physically balanced analysis increment fields in the single-observation test framework. In the month-long cycling experiment with a global 60 km mesh, 3D-Var with static B performs stably. Due to its simple workflow and minimal computational requirements, JEDI-MPAS 3D-Var can be useful for the research community.

This study uses the CAM5–SIT coupled model to investigate the effects of SST feedback frequency on the MJO simulations with intervals at 30 min, 1, 3, 6, 12, 18, 24, and 30 d. The simulations become increasingly unrealistic as the frequency of the SST feedback decreases. Our results suggest that more spontaneous air--sea interaction (e.g., ocean response within 3 d in this study) with high vertical resolution in the ocean model is key to the realistic simulation of the MJO.

We introduce an open-source software, the PCMDI Metrics Package (PMP), developed for a comprehensive comparison of Earth system models (ESMs) with real-world observations. Using diverse metrics evaluating climatology, variability, and extremes simulated in thousands of simulations from the Coupled Model Intercomparison Project (CMIP), PMP aids in benchmarking model improvements across generations. PMP also enables efficient tracking of performance evolutions during ESM developments.

Compared to the silicate weathering fluxes measured at large river basins, the current models tend to systematically overestimate the fluxes over the tropical region, which leads to an overestimation of the global total weathering flux. The most possible cause of such bias is found to be the overestimation of tropical surface erosion, which indicates that the tropical vegetation likely slows down physical erosion significantly. We propose a way of taking this effect into account in models.

To enhance the efficiency of experiments using SCAM, we train a learning-based surrogate model to facilitate large-scale sensitivity analysis and tuning of combinations of multiple parameters. Employing a hybrid method, we investigate the joint sensitivity of multi-parameter combinations across typical cases, identifying the most sensitive three-parameter combination out of 11. Subsequently, we conduct a tuning process aimed at reducing output errors in these cases.

Wildfire is often presented in the media as a danger to human life. Yet globally, millions of people’s livelihoods depend on using fire as a tool. So, patterns of fire emerge from interactions between humans, land use, and climate. This complexity means scientists cannot yet reliably say how fire will be impacted by climate change. So, we developed a new model that represents globally how people use and manage fire. The model reveals the extent and diversity of how humans live with and use fire.

In atmospheric models, rain formation is simplified to be computationally efficient. We trained a machine learning model, SuperdropNet, to emulate warm-rain formation based on super-droplet simulations. Here, we couple SuperdropNet with an atmospheric model in a warm-bubble experiment and find that the coupled simulation runs stable and produces reasonable results, making SuperdropNet a viable ML proxy for droplet simulations. We also present a comprehensive benchmark for coupling architectures.

We introduce SolFinder 1.0, a decision-making tool to select trade-offs between different objective functions for optimal aircraft trajectories, including fuel use, flight time, NO_x emissions, contrail distance, and climate impact. The module is included in the AirTraf 3.0 submodel and uses weather conditions simulated by the EMAC atmospheric model. This paper focuses on the ability of SolFinder to identify eco-efficient trajectories, reducing a flight's climate impact at limited cost penalties.

Traditionally, high-resolution atmospheric models employ periodic boundary conditions, which limit simulations to domains without horizontal variations. In this research open boundary conditions are developed to replace the periodic boundary conditions. The implementation is tested in a controlled setup, and the results show minimal disturbances. Using these boundary conditions, high-resolution models can be forced by a coarser model to study atmospheric phenomena in realistic background states.

Lagrangian particle dispersion models are key for studying atmospheric transport but can be computationally intensive. To speed up simulations, the MPTRAC model was ported to graphics processing units (GPUs). Performance optimization of data structures and memory alignment resulted in runtime improvements of up to 75 % on NVIDIA A100 GPUs for ERA5-based simulations with 100 million particles. These optimizations make the MPTRAC model well suited for future high-performance computing systems.