Search for: All records

Abstract The rapid intensification (RI) of tropical cyclones (TC), defined here as an intensity increase of ≥ 30 kt in 24 hours, is a difficult but important forecasting problem. Operational RI forecasts have considerably improved since the late 2000s, largely thanks to better statistical models, including machine learning (ML). Most ML applications use scalars from the Statistical Hurricane Intensity Prediction Scheme (SHIPS) development dataset as predictors, describing the TC history, near-TC environment, and satellite presentation of the TC. More recent ML applications use convolutional neural networks (CNN), which can ingest full satellite images (or time series of images) and freely “decide” which spatiotemporal features are important for RI. However, two questions remain unanswered: (1) Does image convolution significantly improve RI skill? (2) What strategies do CNNs use for RI prediction – and can we gain new insights from these strategies? We use an ablation experiment to answer the first question and explainable artificial intelligence (XAI) to answer the second. Convolution leads to only a small performance gain, likely because, as revealed by XAI, the CNN’s main strategy uses image features already well described in scalar predictors used by pre-existing RI models. This work makes three additional contributions to the literature: (1) NNs with SHIPS data outperform pre-existing models in some aspects; (2) NNs provide well calibrated uncertainty quantification (UQ), while pre-existing models have no UQ; (3) the NN without SHIPS data performs surprisingly well and is fairly independent of pre-existing models, suggesting its potential value in an operational ensemble. 

Abstract Many of our generation’s most pressing environmental science problems are wicked problems, which means they cannot be cleanly isolated and solved with a single ‘correct’ answer (e.g., Rittel 1973; Wirz 2021). The NSF AI Institute for Research on Trustworthy AI in Weather, Climate, and Coastal Oceanography (AI2ES) seeks to address such problems by developing synergistic approaches with a team of scientists from three disciplines: environmental science (including atmospheric, ocean, and other physical sciences), AI, and social science including risk communication. As part of our work, we developed a novel approach to summer school, held from June 27-30, 2022. The goal of this summer school was to teach a new generation of environmental scientists how to cross disciplines and develop approaches that integrate all three disciplinary perspectives and approaches in order to solve environmental science problems. In addition to a lecture series that focused on the synthesis of AI, environmental science, and risk communication, this year’s summer school included a unique Trust-a-thon component where participants gained hands-on experience applying both risk communication and explainable AI techniques to pre-trained ML models. We had 677 participants from 63 countries register and attend online. Lecture topics included trust and trustworthiness (Day 1), explainability and interpretability (Day 2), data and workflows (Day 3), and uncertainty quantification (Day 4). For the Trust-a-thon we developed challenge problems for three different application domains: (1) severe storms, (2) tropical cyclones, and (3) space weather. Each domain had associated user persona to guide user-centered development. 

Abstract Predicting the timing and location of thunderstorms (“convection”) allows for preventive actions that can save both lives and property. We have applied U-nets, a deep-learning-based type of neural network, to forecast convection on a grid at lead times up to 120 min. The goal is to make skillful forecasts with only present and past satellite data as predictors. Specifically, predictors are multispectral brightness-temperature images from theHimawari-8satellite, while targets (ground truth) are provided by weather radars in Taiwan. U-nets are becoming popular in atmospheric science due to their advantages for gridded prediction. Furthermore, we use three novel approaches to advance U-nets in atmospheric science. First, we compare three architectures—vanilla, temporal, and U-net++—and find that vanilla U-nets are best for this task. Second, we train U-nets with the fractions skill score, which is spatially aware, as the loss function. Third, because we do not have adequate ground truth over the fullHimawari-8domain, we train the U-nets with small radar-centered patches, then apply trained U-nets to the full domain. Also, we find that the best predictions are given by U-nets trained with satellite data from multiple lag times, not only the present. We evaluate U-nets in detail—by time of day, month, and geographic location—and compare them to persistence models. The U-nets outperform persistence at lead times ≥ 60 min, and at all lead times the U-nets provide a more realistic climatology than persistence. Our code is available publicly. 

Abstract This paper describes the development of U-net++ models, a type of neural network that performs deep learning, to emulate the shortwave Rapid Radiative-transfer Model (RRTM). The goal is to emulate the RRTM accurately in a small fraction of the computing time, creating a U-net++ that could be used as a parameterization in numerical weather prediction (NWP). Target variables are surface downwelling flux, top-of-atmosphere upwelling flux ( ), net flux, and a profile of radiative-heating rates. We have devised several ways to make the U-net++ models knowledge-guided, recently identified as a key priority in machine learning (ML) applications to the geosciences. We conduct two experiments to find the best U-net++ configurations. In Experiment 1, we train on non-tropical sites and test on tropical sites, to assess extreme spatial generalization. In Experiment 2, we train on sites from all regions and test on different sites from all regions, with the goal of creating the best possible model for use in NWP. The selected model from Experiment 1 shows impressive skill on the tropical testing sites, except four notable deficiencies: large bias and error for heating rate in the upper stratosphere, unreliable for profiles with single-layer liquid cloud, large heating-rate bias in the mid-troposphere for profiles with multi-layer liquid cloud, and negative bias at lowzenith angles for all flux components and tropospheric heating rates. The selected model from Experiment 2 corrects all but the first deficiency, and both models run ~10 4 times faster than the RRTM. Our code is available publicly. 

Abstract Radiative transfer (RT) is a crucial but computationally expensive process in numerical weather/climate prediction. We develop neural networks (NN) to emulate a common RT parameterization called the Rapid Radiative Transfer Model (RRTM), with the goal of creating a faster parameterization for the Global Forecast System (GFS) v16. In previous work we emulated a highly simplified version of the shortwave RRTM only—excluding many predictor variables, driven by Rapid Refresh forecasts interpolated to a consistent height grid, using only 30 sites in the Northern Hemisphere. In this work we emulate the full shortwave and longwave RRTM—with all predictor variables, driven by GFSv16 forecasts on the native pressure–sigma grid, using data from around the globe. We experiment with NNs of widely varying complexity, including the U-net++ and U-net3+ architectures and deeply supervised training, designed to ensure realistic and accurate structure in gridded predictions. We evaluate the optimal shortwave NN and optimal longwave NN in great detail—as a function of geographic location, cloud regime, and other weather types. Both NNs produce extremely reliable heating rates and fluxes. The shortwave NN has an overall RMSE/MAE/bias of 0.14/0.08/−0.002 K day⁻¹for heating rate and 6.3/4.3/−0.1 W m⁻²for net flux. Analogous numbers for the longwave NN are 0.22/0.12/−0.0006 K day⁻¹and 1.07/0.76/+0.01 W m⁻². Both NNs perform well in nearly all situations, and the shortwave (longwave) NN is 7510 (90) times faster than the RRTM. Both will soon be tested online in the GFSv16. Significance StatementRadiative transfer is an important process for weather and climate. Accurate radiative transfer models exist, such as the RRTM, but these models are computationally slow. We develop neural networks (NNs), a type of machine learning model that is often computationally fast after training, to mimic the RRTM. We wish to accelerate the RRTM by orders of magnitude without sacrificing much accuracy. We drive both the NNs and RRTM with data from the GFSv16, an operational weather model, using locations around the globe during all seasons. We show that the NNs are highly accurate and much faster than the RRTM, which suggests that the NNs could be used to solve radiative transfer inside the GFSv16. 

Abstract In the last decade, much work in atmospheric science has focused on spatial verification (SV) methods for gridded prediction, which overcome serious disadvantages of pixelwise verification. However, neural networks (NN) in atmospheric science are almost always trained to optimize pixelwise loss functions, even when ultimately assessed with SV methods. This establishes a disconnect between model verification during versus after training. To address this issue, we develop spatially enhanced loss functions (SELF) and demonstrate their use for a real-world problem: predicting the occurrence of thunderstorms (henceforth, “convection”) with NNs. In each SELF we use either a neighborhood filter, which highlights convection at scales larger than a threshold, or a spectral filter (employing Fourier or wavelet decomposition), which is more flexible and highlights convection at scales between two thresholds. We use these filters to spatially enhance common verification scores, such as the Brier score. We train each NN with a different SELF and compare their performance at many scales of convection, from discrete storm cells to tropical cyclones. Among our many findings are that (i) for a low or high risk threshold, the ideal SELF focuses on small or large scales, respectively; (ii) models trained with a pixelwise loss function perform surprisingly well; and (iii) nevertheless, models trained with a spectral filter produce much better-calibrated probabilities than a pixelwise model. We provide a general guide to using SELFs, including technical challenges and the final Python code, as well as demonstrating their use for the convection problem. To our knowledge this is the most in-depth guide to SELFs in the geosciences. Significance StatementGridded predictions, in which a quantity is predicted at every pixel in space, should be verified with spatially aware methods rather than pixel by pixel. Neural networks (NN), which are often used for gridded prediction, are trained to minimize an error value called the loss function. NN loss functions in atmospheric science are almost always pixelwise, which causes the predictions to miss rare events and contain unrealistic spatial patterns. We use spatial filters to enhance NN loss functions, and we test our novel spatially enhanced loss functions (SELF) on thunderstorm prediction. We find that different SELFs work better for different scales (i.e., different-sized thunderstorm complexes) and that spectral filters, one of the two filter types, produce unexpectedly well calibrated thunderstorm probabilities. 

Abstract Neural networks (NN) have become an important tool for prediction tasks—both regression and classification—in environmental science. Since many environmental-science problems involve life-or-death decisions and policy making, it is crucial to provide not only predictions but also an estimate of the uncertainty in the predictions. Until recently, very few tools were available to provide uncertainty quantification (UQ) for NN predictions. However, in recent years the computer-science field has developed numerous UQ approaches, and several research groups are exploring how to apply these approaches in environmental science. We provide an accessible introduction to six of these UQ approaches, then focus on tools for the next step, namely, to answer the question:Once we obtain an uncertainty estimate (using any approach), how do we know whether it is good or bad?To answer this question, we highlight four evaluation graphics and eight evaluation scores that are well suited for evaluating and comparing uncertainty estimates (NN based or otherwise) for environmental-science applications. We demonstrate the UQ approaches and UQ-evaluation methods for two real-world problems: 1) estimating vertical profiles of atmospheric dewpoint (a regression task) and 2) predicting convection over Taiwan based onHimawari-8satellite imagery (a classification task). We also provide Jupyter notebooks with Python code for implementing the UQ approaches and UQ-evaluation methods discussed herein. This article provides the environmental-science community with the knowledge and tools to start incorporating the large number of emerging UQ methods into their research. Significance StatementNeural networks are used for many environmental-science applications, some involving life-or-death decision-making. In recent years new methods have been developed to provide much-needed uncertainty estimates for NN predictions. We seek to accelerate the adoption of these methods in the environmental-science community with an accessible introduction to 1) methods for computing uncertainty estimates in NN predictions and 2) methods for evaluating such estimates. 

This paper synthesizes multiple methods for machine learning (ML) model interpretation and visualization (MIV) focusing on meteorological applications. ML has recently exploded in popularity in many fields, including meteorology. Although ML has been successful in meteorology, it has not been as widely accepted, primarily due to the perception that ML models are “black boxes,” meaning the ML methods are thought to take inputs and provide outputs but not to yield physically interpretable information to the user. This paper introduces and demonstrates multiple MIV techniques for both traditional ML and deep learning, to enable meteorologists to understand what ML models have learned. We discuss permutation-based predictor importance, forward and backward selection, saliency maps, class-activation maps, backward optimization, and novelty detection. We apply these methods at multiple spatiotemporal scales to tornado, hail, winter precipitation type, and convective-storm mode. By analyzing such a wide variety of applications, we intend for this work to demystify the black box of ML, offer insight in applying MIV techniques, and serve as a MIV toolbox for meteorologists and other physical scientists. 

Abstract We present an overview of recent work on using artificial intelligence (AI)/machine learning (ML) techniques for forecasting convective weather and its associated hazards, including tornadoes, hail, wind, and lightning. These high-impact phenomena globally cause both massive property damage and loss of life, yet they are very challenging to forecast. Given the recent explosion in developing ML techniques across the weather spectrum and the fact that the skillful prediction of convective weather has immediate societal benefits, we present a thorough review of the current state of the art in AI and ML techniques for convective hazards. Our review includes both traditional approaches, including support vector machines and decision trees, as well as deep learning approaches. We highlight the challenges in developing ML approaches to forecast these phenomena across a variety of spatial and temporal scales. We end with a discussion of promising areas of future work for ML for convective weather, including a discussion of the need to create trustworthy AI forecasts that can be used for forecasters in real time and the need for active cross-sector collaboration on testbeds to validate ML methods in operational situations. Significance StatementWe provide an overview of recent machine learning research in predicting hazards from thunderstorms, specifically looking at lightning, wind, hail, and tornadoes. These hazards kill people worldwide and also destroy property and livestock. Improving the prediction of these events in both the local space as well as globally can save lives and property. By providing this review, we aim to spur additional research into developing machine learning approaches for convective hazard prediction.