Can deep learning beat numerical weather prediction?

Improving the operational forecasts of outdoor Universal Thermal Climate Index with post-processing

The Universal Thermal Climate Index (UTCI) is a thermal comfort index that describes how the human body experiences ambient conditions. It has units of temperature and considers physiological aspects of the human body. It takes into account the effect of air temperature, humidity, wind, radiation, and clothes. It is increasingly used in many countries as a measure of thermal comfort for outdoor conditions, and its value is calculated as part of the operational meteorological forecast. At the same time, forecasts of outdoor UTCI tend to have a relatively large error caused by the error of meteorological forecasts. In Slovenia, there is a relatively dense network of meteorological stations. Crucially, at these stations, global solar radiation measurements are performed continuously, which makes estimating the actual value of the UTCI more accurate compared to the situation where no radiation measurements are available. We used seven years of measurements in hourly resolution from 42 stations to first verify the operational UTCI forecast for the first forecast day and, secondly, to try to improve the forecast via post-processing. We used two machine-learning methods, linear regression, and neural networks. Both methods have successfully reduced the error in the operational UTCI forecasts. Both methods reduced the daily mean error from about 2.6∘ C to almost zero, while the daily mean absolute error decreased from 5∘ C to 3∘ C for the neural network and 3.5∘ C for linear regression. Both methods, especially the neural network, also substantially reduced the dependence of the error on the time of the day.

Physics-informed shape optimization using coordinate projection

The rapid growth of artificial intelligence is revolutionizing classical engineering society, offering novel approaches to material and structural design and analysis. Among various scientific machine learning techniques, physics-informed neural network (PINN) has been one of the most researched subjects, for its ability to incorporate physics prior knowledge into model training. However, the intrinsic continuity requirement of PINN demands the adoption of domain decomposition when multiple materials with distinct properties exist. This greatly complicates the gradient computation of design features, restricting the application of PINN to structural shape optimization. To address this, we present a novel framework that employs neural network coordinate projection for shape optimization within PINN. This technique allows for direct mapping from a standard shape to its optimal counterpart, optimizing the design objective without the need for traditional transition functions or the definition of intermediate material properties. Our method demonstrates a high degree of adaptability, allowing the incorporation of diverse constraints and objectives directly as training penalties. The proposed approach is tested on magnetostatic problems for iron core shape optimization, a scenario typically plagued by the high permeability contrast between materials. Validation with finite-element analysis confirms the accuracy and efficiency of our approach. The results highlight the framework's capability as a viable tool for shape optimization in complex material design tasks.

Hybrid AI-enhanced lightning flash prediction in the medium-range forecast horizon

Traditional fully-deterministic algorithms, which rely on physical equations and mathematical models, are the backbone of many scientific disciplines for decades. These algorithms are based on well-established principles and laws of physics, enabling a systematic and predictable approach to problem-solving. On the other hand, AI-based strategies emerge as a powerful tool for handling vast amounts of data and extracting patterns and relationships that might be challenging to identify through traditional algorithms. Here, we bridge these two realms by using AI to find an optimal mapping of meteorological features predicted two days ahead by the state-of-the-art numerical weather prediction model by the European Centre for Medium-range Weather Forecasts (ECMWF) into lightning flash occurrence. The prediction capability of the resulting AI-enhanced algorithm turns out to be significantly higher than that of the fully-deterministic algorithm employed in the ECMWF model. A remarkable Recall peak of about 95% within the 0-24 h forecast interval is obtained. This performance surpasses the 85% achieved by the ECMWF model at the same Precision of the AI algorithm.

Accurate Machine Learning Prediction in Psychiatry Needs the Right Kind of Information

This Viewpoint discusses the type and amount of data needed for machine learning models to accurately predict diagnoses and treatment outcomes at the individual patient level.

Connecting Agent-Based Models with High-Dimensional Parameter Spaces to Multidimensional Data Using SMoRe ParS: A Surrogate Modeling Approach

Across a broad range of disciplines, agent-based models (ABMs) are increasingly utilized for replicating, predicting, and understanding complex systems and their emergent behavior. In the biological and biomedical sciences, researchers employ ABMs to elucidate complex cellular and molecular interactions across multiple scales under varying conditions. Data generated at these multiple scales, however, presents a computational challenge for robust analysis with ABMs. Indeed, calibrating ABMs remains an open topic of research due to their own high-dimensional parameter spaces. In response to these challenges, we extend and validate our novel methodology, Surrogate Modeling for Reconstructing Parameter Surfaces (SMoRe ParS), arriving at a computationally efficient framework for connecting high dimensional ABM parameter spaces with multidimensional data. Specifically, we modify SMoRe ParS to initially confine high dimensional ABM parameter spaces using unidimensional data, namely, single time-course information of in vitro cancer cell growth assays. Subsequently, we broaden the scope of our approach to encompass more complex ABMs and constrain parameter spaces using multidimensional data. We explore this extension with in vitro cancer cell inhibition assays involving the chemotherapeutic agent oxaliplatin. For each scenario, we validate and evaluate the effectiveness of our approach by comparing how well ABM simulations match the experimental data when using SMoRe ParS-inferred parameters versus parameters inferred by a commonly used direct method. In so doing, we show that our approach of using an explicitly formulated surrogate model as an interlocutor between the ABM and the experimental data effectively calibrates the ABM parameter space to multidimensional data. Our method thus provides a robust and scalable strategy for leveraging multidimensional data to inform multiscale ABMs and explore the uncertainty in their parameters.

Informed Random Forest to Model Associations of Epidemiological Priors, Government Policies, and Public Mobility

Background. Infectious diseases constitute a significant concern worldwide due to their increasing prevalence, associated health risks, and the socioeconomic costs. Machine learning (ML) models and epidemic models formulated using deterministic differential equations are the most dominant tools for analyzing and modeling the transmission of infectious diseases. However, ML models can be inconsistent in extracting the dynamics of a disease in the presence of data drifts. Likewise, the capability of epidemic models is constrained to parameter dimensions and estimation. We aimed at creating a framework of informed ML that integrates a random forest (RF) with an adapted susceptible infectious recovered (SIR) model to account for accuracy and consistency in stochasticity within the dynamics of coronavirus disease 2019 (COVID-19). Methods. An adapted SIR model was used to inform a default RF on predicting new COVID-19 cases (NCCs) at given intervals. We validated the performance of the informed RF (IRF) using real data. We used Botswana's pharmaceutical interventions (PIs) and non-PIs (NPIs) adopted between February 2020 and August 2022. The discrepancy between predictions and observations is modeled using loss functions, which are minimized, interpreted, and used to assess the IRF. Results. The findings on the real data have revealed the effectiveness of the default RF in modeling and predicting NCCs. The use of the effective reproductive rate to inform the RF yielded an excellent predictive power (84%) compared with 75% by the default RF. Conclusion. This research has potential to inform policy and decision makers in developing systems to evaluate interventions for infectious diseases.

Highlights

This framework is initiated by incorporating model outputs from an epidemic model to a machine learning model.An informed random forest (RF) is instantiated to model government and public responses to the COVID-19 pandemic.This framework does not require data transformations, and the epidemic model is shown to boost the RF's performance.This is a baseline knowledge-informed learning framework for assessing public health interventions in Botswana.

Accurate medium-range global weather forecasting with 3D neural networks

Weather forecasting is important for science and society. At present, the most accurate forecast system is the numerical weather prediction (NWP) method, which represents atmospheric states as discretized grids and numerically solves partial differential equations that describe the transition between those states1. However, this procedure is computationally expensive. Recently, artificial-intelligence-based methods2 have shown potential in accelerating weather forecasting by orders of magnitude, but the forecast accuracy is still significantly lower than that of NWP methods. Here we introduce an artificial-intelligence-based method for accurate, medium-range global weather forecasting. We show that three-dimensional deep networks equipped with Earth-specific priors are effective at dealing with complex patterns in weather data, and that a hierarchical temporal aggregation strategy reduces accumulation errors in medium-range forecasting. Trained on 39 years of global data, our program, Pangu-Weather, obtains stronger deterministic forecast results on reanalysis data in all tested variables when compared with the world's best NWP system, the operational integrated forecasting system of the European Centre for Medium-Range Weather Forecasts (ECMWF)3. Our method also works well with extreme weather forecasts and ensemble forecasts. When initialized with reanalysis data, the accuracy of tracking tropical cyclones is also higher than that of ECMWF-HRES.

Explainable sequence-to-sequence GRU neural network for pollution forecasting

The goal of pollution forecasting models is to allow the prediction and control of the air quality. Non-linear data-driven approaches based on deep neural networks have been increasingly used in such contexts showing significant improvements w.r.t. more conventional approaches like regression models and mechanistic approaches. While such deep learning models were deemed for a long time as black boxes, recent advances in eXplainable AI (XAI) allow to look through the model's decision-making process, providing insights into decisive input features responsible for the model's prediction. One XAI technique to explain the predictions of neural networks which was proven useful in various domains is Layer-wise Relevance Propagation (LRP). In this work, we extend the LRP technique to a sequence-to-sequence neural network model with GRU layers. The explanation heatmaps provided by LRP allow us to identify important meteorological and temporal features responsible for the accumulation of four major pollutants in the air ([Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text]), and our findings can be backed up with prior knowledge in environmental and pollution research. This illustrates the appropriateness of XAI for understanding pollution forecastings and opens up new avenues for controlling and mitigating the pollutants' load in the air.

Taking advantage of quasi-periodic signals for S2S operational forecast from a perspective of deep learning

The quasi-periodic signals in the earth system could be the predictability source for sub-seasonal to seasonal (S2S) climate prediction because of the connections among the lead-lag time of those signals. The Madden-Julian Oscillation (MJO) is a typical quasi-periodic signal, which is the dominant S2S variability in the tropics. Besides, significantly periodic features in terms of both intensity and location are identified in 10-40 days for the concurrent variation of the subtropical and polar jet streams over Asia in this study. So far, those signals contribute less and are not fully applied to the S2S prediction. The deep learning (DL) approach, especially the long-short term memory (LSTM) networks, has the ability to take advantage of the information at the previous time to improve the prediction after then. This study presents the application of the DL in the postprocessing of S2S prediction using quasi-periodic signals predicted by the operational model to improve the prediction of minimum 2-m air temperature over Asia. With the help of deep learning, it finds the best weights for the ensemble predictions, and the quasi-periodic signals in the atmosphere can further benefit the S2S operational prediction.

SMoRe ParS: A novel methodology for bridging modeling modalities and experimental data applied to 3D vascular tumor growth

Multiscale systems biology is having an increasingly powerful impact on our understanding of the interconnected molecular, cellular, and microenvironmental drivers of tumor growth and the effects of novel drugs and drug combinations for cancer therapy. Agent-based models (ABMs) that treat cells as autonomous decision-makers, each with their own intrinsic characteristics, are a natural platform for capturing intratumoral heterogeneity. Agent-based models are also useful for integrating the multiple time and spatial scales associated with vascular tumor growth and response to treatment. Despite all their benefits, the computational costs of solving agent-based models escalate and become prohibitive when simulating millions of cells, making parameter exploration and model parameterization from experimental data very challenging. Moreover, such data are typically limited, coarse-grained and may lack any spatial resolution, compounding these challenges. We address these issues by developing a first-of-its-kind method that leverages explicitly formulated surrogate models (SMs) to bridge the current computational divide between agent-based models and experimental data. In our approach, Surrogate Modeling for Reconstructing Parameter Surfaces (SMoRe ParS), we quantify the uncertainty in the relationship between agent-based model inputs and surrogate model parameters, and between surrogate model parameters and experimental data. In this way, surrogate model parameters serve as intermediaries between agent-based model input and data, making it possible to use them for calibration and uncertainty quantification of agent-based model parameters that map directly onto an experimental data set. We illustrate the functionality and novelty of Surrogate Modeling for Reconstructing Parameter Surfaces by applying it to an agent-based model of 3D vascular tumor growth, and experimental data in the form of tumor volume time-courses. Our method is broadly applicable to situations where preserving underlying mechanistic information is of interest, and where computational complexity and sparse, noisy calibration data hinder model parameterization.

Climate Change Effects on Pathogen Emergence: Artificial Intelligence to Translate Big Data for Mitigation

Plant pathology has developed a wide range of concepts and tools for improving plant disease management, including models for understanding and responding to new risks from climate change. Most of these tools can be improved using new advances in artificial intelligence (AI), such as machine learning to integrate massive data sets in predictive models. There is the potential to develop automated analyses of risk that alert decision-makers, from farm managers to national plant protection organizations, to the likely need for action and provide decision support for targeting responses. We review machine-learning applications in plant pathology and synthesize ideas for the next steps to make the most of these tools in digital agriculture. Global projects, such as the proposed global surveillance system for plant disease, will be strengthened by the integration of the wide range of new data, including data from tools like remote sensors, that are used to evaluate the risk ofplant disease. There is exciting potential for the use of AI to strengthen global capacity building as well, from image analysis for disease diagnostics and associated management recommendations on farmers' phones to future training methodologies for plant pathologists that are customized in real-time for management needs in response to the current risks. International cooperation in integrating data and models will help develop the most effective responses to new challenges from climate change.

A framework for benchmarking uncertainty in deep regression

We propose a framework for the assessment of uncertainty quantification in deep regression. The framework is based on regression problems where the regression function is a linear combination of nonlinear functions. Basically, any level of complexity can be realized through the choice of the nonlinear functions and the dimensionality of their domain. Results of an uncertainty quantification for deep regression are compared against those obtained by a statistical reference method. The reference method utilizes knowledge about the underlying nonlinear functions and is based on Bayesian linear regression using a prior reference. The flexibility, together with the availability of a reference solution, makes the framework suitable for defining benchmark sets for uncertainty quantification. Reliability of uncertainty quantification is assessed in terms of coverage probabilities, and accuracy through the size of calculated uncertainties. We illustrate the proposed framework by applying it to current approaches for uncertainty quantification in deep regression. In addition, results for three real-world regression tasks are presented.

A Change Management Approach with the Support of the Balanced Scorecard and the Utilization of Artificial Neural Networks

Artificial Intelligence (AI) has revolutionized the way organizations face decision-making issues. One of these crucial elements is the implementation of organizational changes. There has been a wide-spread adoption of AI techniques in the private sector, whereas in the public sector their use has been recently extended. One of the greatest challenges that European governments have to face is the implementation of a wide variety of European Union (EU) funding programs which have evolved in the context of the EU long-term budget. In the current study, the Balanced Scorecard (BSC) and Artificial Neural Networks (ANNs) are intertwined with forecasting the outcomes of a co-financed EU program by means of its impact on the non-financial measures of the government body that materialized it. The predictive accuracy of the present model advanced in this research study takes into account all the complexities of the business environment, within which the provided dataset is produced. The outcomes of the study showed that the measures taken to enhance customer satisfaction allows for further improvement. The utilization of the proposed model could facilitate the decision-making process and initiate changes to the administrational issues of the available funding programs.

ECMWF short-term prediction accuracy improvement by deep learning

This paper aims to describe and evaluate the proposed calibration model based on a neural network for post-processing of two essential meteorological parameters, namely near-surface air temperature (2 m) and 24 h accumulated precipitation. The main idea behind this work is to improve short-term (up to 3 days) forecasts delivered by a global numerical weather prediction (NWP) model called ECMWF (European Centre for Medium-Range Weather Forecasts). In comparison to the existing local weather models that typically provide weather forecasts for limited geographic areas (e.g., within one country but they are more accurate), ECMWF offers a prediction of the weather phenomena across the world. Another significant benefit of this global NWP model includes the fact, that by using it in several well-known online applications, forecasts are freely available while local models outputs are often paid. Our proposed ECMWF-enhancing model uses a combination of raw ECMWF data and additional input parameters we have identified as useful for ECMWF error estimation and its subsequent correction. The ground truth data used for the training phase of our model consists of real observations from weather stations located in 10 cities across two European countries. The results obtained from cross-validation indicate that our parametric model outperforms the accuracy of a standard ECMWF prediction and gets closer to the forecast precision of the local NWP models.

Importance of ozone precursors information in modelling urban surface ozone variability using machine learning algorithm

Surface ozone (O\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$_3$$\end{document}3) is primarily formed through complex photo-chemical reactions in the atmosphere, which are non-linearly dependent on precursors. Even though, there have been many recent studies exploring the potential of machine learning (ML) in modeling surface ozone, the inclusion of limited available ozone precursors information has received little attention. The ML algorithm with in-situ NO information and meteorology explains 87% (R\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{2}$$\end{document}2 = 0.87) of the ozone variability over Munich, a German metropolitan area, which is 15% higher than a ML algorithm that considers only meteorology. The ML algorithm trained for the urban measurement station in Munich can also explain the ozone variability of the other three stations in the same city, with R\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{2}$$\end{document}2 = 0.88, 0.91, 0.63. While the same model robustly explains the ozone variability of two other German cities’ (Berlin and Hamburg) measurement stations, with R\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{2}$$\end{document}2 ranges from 0.72 to 0.84, giving confidence to use the ML algorithm trained for one location to other locations with sparse ozone measurements. The inclusion of satellite O\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$_3$$\end{document}3 precursors information has little effect on the ML model’s performance.

Prediction of Harvest Time of Apple Trees: An RNN-Based Approach

In the field of agricultural research, Machine Learning (ML) has been used to increase agricultural productivity and minimize its environmental impact, proving to be an essential technique to support decision making. Accurate harvest time prediction is a challenge for fruit production in a sustainable manner, which could eventually reduce food waste. Linear models have been used to estimate period duration; however, they present variability when used to estimate the chronological time of apple tree stages. This study proposes the PredHarv model, which is a machine learning model that uses Recurrent Neural Networks (RNN) to predict the start date of the apple harvest, given the weather conditions related to the temperature expected for the period. Predictions are made from the phenological phase of the beginning of flowering, using a multivariate approach, based on the time series of phenology and meteorological data. The computational model contributes to anticipating information about the harvest date, enabling the grower to better plan activities, avoiding costs, and consequently improving productivity. We developed a prototype of the model and performed experiments with real datasets from agricultural institutions. We evaluated the metrics, and the results obtained in evaluation scenarios demonstrate that the model is efficient, has good generalizability, and is capable of improving the accuracy of the prediction results.

Super-Resolving Ocean Dynamics from Space with Computer Vision Algorithms

Surface ocean dynamics play a key role in the Earth system, contributing to regulate its climate and affecting the marine ecosystem functioning. Dynamical processes occur and interact in the upper ocean at multiple scales, down to, or even less than, few kilometres. These scales are not adequately resolved by present observing systems, and, in the last decades, global monitoring of surface currents has been based on the application of geostrophic balance to absolute dynamic topography maps obtained through the statistical interpolation of along-track satellite altimeter data. Due to the cross-track distance and repetitiveness of satellite acquisitions, the effective resolution of interpolated data is limited to several tens of kilometres. At the kilometre scale, sea surface temperature pattern evolution is dominated by advection, providing indirect information on upper ocean currents. Computer vision techniques are perfect candidates to infer this dynamical information from the combination of altimeter data, surface temperature images and observing-system geometry. Here, we exploit one class of image processing techniques, super-resolution, to develop an original neural-network architecture specifically designed to improve absolute dynamic topography reconstruction. Our model is first trained on synthetic observations built from a numerical general-circulation model and then tested on real satellite products. Provided concurrent clear-sky thermal observations are available, it proves able to compensate for altimeter sampling/interpolation limitations by learning from primitive equation data. The algorithm can be adapted to learn directly from future surface topography, and eventual surface currents, high-resolution satellite observations.

A Framework for Four-Dimensional Variational Data Assimilation Based on Machine Learning

The initial field has a crucial influence on numerical weather prediction (NWP). Data assimilation (DA) is a reliable method to obtain the initial field of the forecast model. At the same time, data are the carriers of information. Observational data are a concrete representation of information. DA is also the process of sorting observation data, during which entropy gradually decreases. Four-dimensional variational assimilation (4D-Var) is the most popular approach. However, due to the complexity of the physical model, the tangent linear and adjoint models, and other processes, the realization of a 4D-Var system is complicated, and the computational efficiency is expensive. Machine learning (ML) is a method of gaining simulation results by training a large amount of data. It achieves remarkable success in various applications, and operational NWP and DA are no exception. In this work, we synthesize insights and techniques from previous studies to design a pure data-driven 4D-Var implementation framework named ML-4DVAR based on the bilinear neural network (BNN). The framework replaces the traditional physical model with the BNN model for prediction. Moreover, it directly makes use of the ML model obtained from the simulation data to implement the primary process of 4D-Var, including the realization of the short-term forecast process and the tangent linear and adjoint models. We test a strong-constraint 4D-Var system with the Lorenz-96 model, and we compared the traditional 4D-Var system with ML-4DVAR. The experimental results demonstrate that the ML-4DVAR framework can achieve better assimilation results and significantly improve computational efficiency.

OpenForecast: An Assessment of the Operational Run in 2020–2021

OpenForecast is the first openly available national-scale operational runoff forecasting system in Russia. Launched in March 2020, it routinely provides 7-day ahead predictions for 834 gauges across the country. Here, we provide an assessment of the OpenForecast performance on the long-term evaluation period from 14 March 2020 to 31 October 2021 (597 days) for 252 gauges for which operational data are available and quality-controlled. Results show that OpenForecast is a robust system based on reliable data and solid computational routines that secures efficient runoff forecasts for a diverse set of gauges.

South America Seasonal Precipitation Prediction by Gradient-Boosting Machine-Learning Approach

Machine learning has experienced great success in many applications. Precipitation is a hard meteorological variable to predict, but it has a strong impact on society. Here, a machine-learning technique—a formulation of gradient-boosted trees—is applied to climate seasonal precipitation prediction over South America. The Optuna framework, based on Bayesian optimization, was employed to determine the optimal hyperparameters for the gradient-boosting scheme. A comparison between seasonal precipitation forecasting among the numerical atmospheric models used by the National Institute for Space Research (INPE, Brazil) as an operational procedure for weather/climate forecasting, gradient boosting, and deep-learning techniques is made regarding observation, with some showing better performance for the boosting scheme.

Revealing drivers and risks for power grid frequency stability with explainable AI

Stable operation of an electric power system requires strict operational limits for the grid frequency. Fluctuations and external impacts can cause large frequency deviations and increased control efforts. Although these complex interdependencies can be modeled using machine learning algorithms, the black box character of many models limits insights and applicability. In this article, we introduce an explainable machine learning model that accurately predicts frequency stability indicators for three European synchronous areas. Using Shapley additive explanations, we identify key features and risk factors for frequency stability. We show how load and generation ramps determine frequency gradients, and we identify three classes of generation technologies with converse impacts. Control efforts vary strongly depending on the grid and time of day and are driven by ramps as well as electricity prices. Notably, renewable power generation is central only in the British grid, while forecasting errors play a major role in the Nordic grid.

Two-Stage Spatiotemporal Context Refinement Network for Precipitation Nowcasting

Precipitation nowcasting by radar echo extrapolation using machine learning algorithms is a field worthy of further study, since rainfall prediction is essential in work and life. Current methods of predicting the radar echo images need further improvement in prediction accuracy as well as in presenting the predicted details of the radar echo images. In this paper, we propose a two-stage spatiotemporal context refinement network (2S-STRef) to predict future pixel-level radar echo maps (deterministic output) more accurately and with more distinct details. The first stage is an efficient and concise spatiotemporal prediction network, which uses the spatiotemporal RNN module embedded in an encoder and decoder structure to give a first-stage prediction. The second stage is a proposed detail refinement net, which can preserve the high-frequency detailed feature of the radar echo images by using the multi-scale feature extraction and fusion residual block. We used a real-world radar echo map dataset of South China to evaluate the proposed 2S-STRef model. The experiments showed that compared with the PredRNN++ and ConvLSTM method, our 2S-STRef model performs better on the precipitation nowcasting, as well as at the image quality evaluating index and the forecasting indices. At a given 45 dBZ echo threshold (heavy precipitation) and with a 2 h lead time, the widely used CSI, HSS, and SSIM indices of the proposed 2S-STRef model are found equal to 0.195, 0.312, and 0.665, respectively. In this case, the proposed model outperforms the OpticalFlow method and PredRNN++ model.

Artificial Intelligence Revolutionises Weather Forecast, Climate Monitoring and Decadal Prediction

Artificial Intelligence (AI) is an explosively growing field of computer technology, which is expected to transform many aspects of our society in a profound way. AI techniques are used to analyse large amounts of unstructured and heterogeneous data and discover and exploit complex and intricate relations among these data, without recourse to an explicit analytical treatment of those relations. These AI techniques are unavoidable to make sense of the rapidly increasing data deluge and to respond to the challenging new demands in Weather Forecast (WF), Climate Monitoring (CM) and Decadal Prediction (DP). The use of AI techniques can lead simultaneously to: (1) a reduction of human development effort, (2) a more efficient use of computing resources and (3) an increased forecast quality. To realise this potential, a new generation of scientists combining atmospheric science domain knowledge and state-of-the-art AI skills needs to be trained. AI should become a cornerstone of future weather and climate observation and modelling systems.

Tracking droplets in soft granular flows with deep learning techniques

The state-of-the-art deep learning-based object recognition YOLO algorithm and object tracking DeepSORT algorithm are combined to analyze digital images from fluid dynamic simulations of multi-core emulsions and soft flowing crystals and to track moving droplets within these complex flows. The YOLO network was trained to recognize the droplets with synthetically prepared data, thereby bypassing the labor-intensive data acquisition process. In both applications, the trained YOLO + DeepSORT procedure performs with high accuracy on the real data from the fluid simulations, with low error levels in the inferred trajectories of the droplets and independently computed ground truth. Moreover, using commonly used desktop GPUs, the developed application is capable of analyzing data at speeds that exceed the typical image acquisition rates of digital cameras (30 fps), opening the interesting prospect of realizing a low-cost and practical tool to study systems with many moving objects, mostly but not exclusively, biological ones. Besides its practical applications, the procedure presented here marks the first step towards the automatic extraction of effective equations of motion of many-body soft flowing systems.