The recent hype about artificial intelligence has sparked renewed interest in applying the successful deep learning (DL) methods for image recognition, speech recognition, robotics, strategic games and other application areas to the field of meteorology. There is some evidence that better weather forecasts can be produced by introducing big data mining and neural networks into the weather prediction workflow. Here, we discuss the question of whether it is possible to completely replace the current numerical weather models and data assimilation systems with DL approaches. This discussion entails a review of state-of-the-art machine learning concepts and their applicability to weather data with its pertinent statistical properties. We think that it is not inconceivable that numerical weather models may one day become obsolete, but a number of fundamental breakthroughs are needed before this goal comes into reach. This article is part of the theme issue ‘Machine learning for weather and climate modelling’.

Recent advances in technology have empowered the widespread application of cyber–physical systems in manufacturing and fostered the Industry 4.0 paradigm. In the factories of the future, it is possible that all items, including operators, will be equipped with integrated communication and data processing capabilities. Operators can become part of the smart manufacturing systems, and this fosters a paradigm shift from independent automated and human activities to human–cyber–physical systems (HCPSs). In this context, a Healthy Operator 4.0 (HO4.0) concept was proposed, based on a systemic view of the Industrial Internet of Things (IIoT) and wearable technology. For the implementation of this relatively new concept, we constructed a unified architecture to support the integration of different enabling technologies. We designed an implementation model to facilitate the practical application of this concept in industry. The main enabling technologies of the model are introduced afterward. In addition, a prototype system was developed, and relevant experiments were conducted to demonstrate the feasibility of the proposed system architecture and the implementation framework, as well as some of the derived benefits.

Numerical weather prediction (NWP) models solve a system of partial differential equations based on physical laws to forecast the future state of the atmosphere. These models are deployed operationally, but they are computationally very expensive. Recently, the potential of deep neural networks to generate bespoke weather forecasts has been explored in a couple of scientific studies inspired by the success of video frame prediction models in computer vision. In this study, a simple recurrent neural network with convolutional filters, called ConvLSTM, and an advanced generative network, the Stochastic Adversarial Video Prediction (SAVP) model, are applied to create hourly forecasts of the 2 m temperature for the next 12 h over Europe. We make use of 13 years of data from the ERA5 reanalysis, of which 11 years are utilized for training and 1 year each is used for validating and testing. We choose the 2 m temperature, total cloud cover, and the 850 hPa temperature as predictors and show that both models attain predictive skill by outperforming persistence forecasts. SAVP is superior to ConvLSTM in terms of several evaluation metrics, confirming previous results from computer vision that larger, more complex networks are better suited to learn complex features and to generate better predictions. The 12 h forecasts of SAVP attain a mean squared error (MSE) of about 2.3 K², an anomaly correlation coefficient (ACC) larger than 0.85, a structural similarity index (SSIM) of around 0.72, and a gradient ratio (r_G) of about 0.82. The ConvLSTM yields a higher MSE (3.6 K²), a smaller ACC (0.80) and SSIM (0.65), and a slightly larger r_G (0.84). The superior performance of SAVP in terms of MSE, ACC, and SSIM can be largely attributed to the generator. A sensitivity study shows that a larger weight of the generative adversarial network (GAN) component in the SAVP loss leads to even better preservation of spatial variability at the cost of a somewhat increased MSE (2.5 K²). Including the 850 hPa temperature as an additional predictor enhances the forecast quality, and the model also benefits from a larger spatial domain. By contrast, adding the total cloud cover as predictor or reducing the amount of training data to 8 years has only small effects. Although the temperature forecasts obtained in this way are still less powerful than contemporary NWP models, this study demonstrates that sophisticated deep neural networks may achieve considerable forecast quality beyond the nowcasting range in a purely data-driven way.

As knowledge about the cirrus clouds in the lower stratosphere is limited, reliable long-term measurements are needed to assess their characteristics, radiative impact and important role in upper troposphere and lower stratosphere (UTLS) chemistry. We used 6 years (2006–2012) of Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) measurements to investigate the global and seasonal distribution of stratospheric cirrus clouds and compared the MIPAS results with results derived from the latest version (V4.x) of the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) data. For the identification of stratospheric cirrus clouds, precise information on both the cloud top height (CTH) and the tropopause height is crucial. Here, we used lapse rate tropopause heights estimated from the ERA-Interim global reanalysis. Considering the uncertainties of the tropopause heights and the vertical sampling grid, we define CTHs more than 0.5 km above the tropopause as stratospheric for CALIPSO data. For MIPAS data, we took into account the coarser vertical sampling grid and the broad field of view so that we considered cirrus CTHs detected more than 0.75 km above the tropopause as stratospheric. Further sensitivity tests were conducted to rule out sampling artefacts in MIPAS data. The global distribution of stratospheric cirrus clouds was derived from night-time measurements because of the higher detection sensitivity of CALIPSO. In both data sets, MIPAS and CALIPSO, the stratospheric cirrus cloud occurrence frequencies are significantly higher in the tropics than in the extra-tropics. Tropical hotspots of stratospheric cirrus clouds associated with deep convection are located over equatorial Africa, South and Southeast Asia, the western Pacific, and South America. Stratospheric cirrus clouds were more often detected in December–February (15 %) than June–August (8 %) in the tropics (±20 ^∘). At northern and southern middle latitudes (40–60^∘), MIPAS observed about twice as many stratospheric cirrus clouds (occurrence frequencies of 4 %–5 % for MIPAS rather than about 2 % for CALIPSO). We attribute more frequent observations of stratospheric cirrus clouds with MIPAS to the higher detection sensitivity of the instrument to optically thin clouds. In contrast to the difference between daytime and night-time occurrence frequencies of stratospheric cirrus clouds by a factor of about 2 in zonal means in the tropics (4 % and 10 %, respectively) and at middle latitudes for CALIPSO data, there is little diurnal cycle in MIPAS data, in which the difference of occurrence frequencies in the tropics is about 1 percentage point in zonal mean and about 0.5 percentage point at middle latitudes. The difference between CALIPSO day and night measurements can also be attributed to their differences in detection sensitivity. Future work should focus on better understanding the origin of the stratospheric cirrus clouds and their impact on radiative forcing and climate.

Narrowing surface ozone disparities between urban and nonurban areas escalate health risks in densely populated urban zones. A comprehensive understanding of the impact of ozone photochemistry on this transition remains constrained by current knowledge of aerosol effects and the availability of surface monitoring. Here we reconstructed spatiotemporal gapless air quality concentrations using a novel transformer deep learning (DL) framework capable of perceiving spatiotemporal dynamics to analyze ozone urban–nonurban differences. Subsequently, the photochemical effect on these discrepancies was analyzed by elucidating shifts in ozone regimes inferred from an interpretable machine learning method. The evaluations of the model exhibited an average out-of-sample cross-validation coefficient of determination of 0.96, 0.92, and 0.95 for ozone, nitrogen dioxide, and fine particulate matter (PM_2.5), respectively. The ozone sensitivity in nonurban areas, dominated by a nitrogen-oxide-limited (NO_x-limited) regime, was observed to shift towards increased sensitivity to volatile organic compounds (VOCs) when extended to urban areas. A third “aerosol-inhibited” regime was identified in the Jiaodong Peninsula, where the uptake of hydroperoxyl radicals onto aerosols suppressed ozone production under low NO_x levels during summertime. The reduction of PM_2.5 could increase the sensitivity of ozone to VOCs, necessitating more stringent VOC emission abatement for urban ozone mitigation. In 2020, urban ozone levels in Shandong surpassed those in nonurban areas, primarily due to a more pronounced decrease in the latter resulting from stronger aerosol suppression effects and less reduction in PM_2.5. This case study demonstrates the critical need for advanced spatially resolved models and interpretable analysis in tackling ozone pollution challenges.

The prediction of precipitation patterns up to 2 h ahead, also known as precipitation nowcasting, at high spatiotemporal resolutions is of great relevance in weather-dependent decision-making and early warning systems. In this study, we are aiming to provide an efficient and easy-to-understand deep neural network – CLGAN (convolutional long short-term memory generative adversarial network) – to improve the nowcasting skills of heavy precipitation events. The model constitutes a generative adversarial network (GAN) architecture, whose generator is built upon a u-shaped encoder–decoder network (U-Net) and is equipped with recurrent long short-term memory (LSTM) cells to capture spatiotemporal features. The optical flow model DenseRotation and the competitive video prediction models ConvLSTM (convolutional LSTM) and PredRNN-v2 (predictive recurrent neural network version 2) are used as the competitors. A series of evaluation metrics, including the root mean square error, the critical success index, the fractions skill score, and object-based diagnostic evaluation, are utilized for a comprehensive comparison against competing baseline models. We show that CLGAN outperforms the competitors in terms of scores for dichotomous events and object-based diagnostics. A sensitivity analysis on the weight of the GAN component indicates that the GAN-based architecture helps to capture heavy precipitation events. The results encourage future work based on the proposed CLGAN architecture to improve the precipitation nowcasting and early warning systems.