Browse
Abstract
We present and evaluate a deep learning first-guess front-identification system that identifies cold, warm, stationary, and occluded fronts. Frontal boundaries play a key role in the daily weather around the world. Human-drawn fronts provided by the National Weather Service’s Weather Prediction Center, Ocean Prediction Center, Tropical Analysis and Forecast Branch, and Honolulu Forecast Office are treated as ground-truth labels for training the deep learning models. The models are trained using ERA5 data with variables known to be important for distinguishing frontal boundaries, including temperature, equivalent potential temperature, and wind velocity and direction at multiple heights. Using a 250-km neighborhood over the contiguous U.S. domain, our best models achieve critical success index scores of 0.60 for cold fronts, 0.43 for warm fronts, 0.48 for stationary fronts, 0.45 for occluded fronts, and 0.71 using a binary classification system (front/no front), whereas scores over the full unified surface analysis domain were lower. For cold and warm fronts and binary classification, these scores significantly outperform prior baseline methods that utilize 250-km neighborhoods. These first-guess deep learning algorithms can be used by forecasters to locate frontal boundaries more effectively and expedite the frontal analysis process.
Significance Statement
Fronts are boundaries that affect the weather that people experience daily. Currently, forecasters must identify these boundaries through manual analysis. We have developed an automated machine learning method for detecting cold, warm, stationary, and occluded fronts. Our automated method provides forecasters with an additional tool to expedite the frontal analysis process.
Abstract
We present and evaluate a deep learning first-guess front-identification system that identifies cold, warm, stationary, and occluded fronts. Frontal boundaries play a key role in the daily weather around the world. Human-drawn fronts provided by the National Weather Service’s Weather Prediction Center, Ocean Prediction Center, Tropical Analysis and Forecast Branch, and Honolulu Forecast Office are treated as ground-truth labels for training the deep learning models. The models are trained using ERA5 data with variables known to be important for distinguishing frontal boundaries, including temperature, equivalent potential temperature, and wind velocity and direction at multiple heights. Using a 250-km neighborhood over the contiguous U.S. domain, our best models achieve critical success index scores of 0.60 for cold fronts, 0.43 for warm fronts, 0.48 for stationary fronts, 0.45 for occluded fronts, and 0.71 using a binary classification system (front/no front), whereas scores over the full unified surface analysis domain were lower. For cold and warm fronts and binary classification, these scores significantly outperform prior baseline methods that utilize 250-km neighborhoods. These first-guess deep learning algorithms can be used by forecasters to locate frontal boundaries more effectively and expedite the frontal analysis process.
Significance Statement
Fronts are boundaries that affect the weather that people experience daily. Currently, forecasters must identify these boundaries through manual analysis. We have developed an automated machine learning method for detecting cold, warm, stationary, and occluded fronts. Our automated method provides forecasters with an additional tool to expedite the frontal analysis process.
Abstract
The two-step U-Net model (TU-Net) contains a western North Pacific subtropical high (WNPSH) prediction model and a precipitation prediction model fed by the WNPSH predictions, oceanic heat content, and surface temperature. The data-driven forecast model provides improved 4-month lead predictions of the WNPSH and precipitation in the middle and lower reaches of the Yangtze River (MLYR), which has important implications for water resources management and precipitation-related disaster prevention in China. When compared with five state-of-the-art dynamical climate models including the Climate Forecast System of Nanjing University of Information Science and Technology (NUIST-CFS1.0) and four models participating in the North American Multi-Model Ensemble (NMME) project, the TU-Net produces comparable skills in forecasting 4-month lead geopotential height and winds at the 500- and 850-hPa levels. For the 4-month lead prediction of precipitation over the MLYR region, the TU-Net has the best correlation scores and mean latitude-weighted RMSE in each summer month and in boreal summer [June–August (JJA)], and pattern correlation coefficient scores are slightly lower than the dynamical models only in June and JJA. In addition, the results show that the constructed TU-Net is also superior to most of the dynamical models in predicting 2-m air temperature in the MLYR region at a 4-month lead. Thus, the deep learning-based TU-Net model can provide a rapid and inexpensive way to improve the seasonal prediction of summer precipitation and 2-m air temperature over the MLYR region.
Significance Statement
The purpose of this study is to examine the seasonal predictive skill of the western North Pacific subtropical high anomalies and summer rainfall anomalies over the middle and lower reaches of the Yangtze River region by means of deep learning methods. Our deep learning model provides a rapid and inexpensive way to improve the seasonal prediction of summer precipitation as well as 2-m air temperature. The work has important implications for water resources management and precipitation-related disaster prevention in China and can be extended in the future to predict other climate variables as well.
Abstract
The two-step U-Net model (TU-Net) contains a western North Pacific subtropical high (WNPSH) prediction model and a precipitation prediction model fed by the WNPSH predictions, oceanic heat content, and surface temperature. The data-driven forecast model provides improved 4-month lead predictions of the WNPSH and precipitation in the middle and lower reaches of the Yangtze River (MLYR), which has important implications for water resources management and precipitation-related disaster prevention in China. When compared with five state-of-the-art dynamical climate models including the Climate Forecast System of Nanjing University of Information Science and Technology (NUIST-CFS1.0) and four models participating in the North American Multi-Model Ensemble (NMME) project, the TU-Net produces comparable skills in forecasting 4-month lead geopotential height and winds at the 500- and 850-hPa levels. For the 4-month lead prediction of precipitation over the MLYR region, the TU-Net has the best correlation scores and mean latitude-weighted RMSE in each summer month and in boreal summer [June–August (JJA)], and pattern correlation coefficient scores are slightly lower than the dynamical models only in June and JJA. In addition, the results show that the constructed TU-Net is also superior to most of the dynamical models in predicting 2-m air temperature in the MLYR region at a 4-month lead. Thus, the deep learning-based TU-Net model can provide a rapid and inexpensive way to improve the seasonal prediction of summer precipitation and 2-m air temperature over the MLYR region.
Significance Statement
The purpose of this study is to examine the seasonal predictive skill of the western North Pacific subtropical high anomalies and summer rainfall anomalies over the middle and lower reaches of the Yangtze River region by means of deep learning methods. Our deep learning model provides a rapid and inexpensive way to improve the seasonal prediction of summer precipitation as well as 2-m air temperature. The work has important implications for water resources management and precipitation-related disaster prevention in China and can be extended in the future to predict other climate variables as well.
Abstract
Globally available environmental observations (EOs), specifically from satellites and coupled Earth system models, represent some of the largest datasets of the digital age. As the volume of global EOs continues to grow, so does the potential of these data to help Earth scientists discover trends and patterns in Earth systems at large spatial scales. To leverage global EOs for scientific insight, Earth scientists need targeted and accessible exposure to skills in reproducible scientific computing and spatiotemporal data science, and to be empowered to apply their domain understanding to interpret data-driven models for knowledge discovery. The Generalizable, Reproducible, Robust, and Interpreted Environmental (GRRIEn) analysis framework was developed to prepare Earth scientists with an introductory statistics background and limited/no understanding of programming and computational methods to use global EOs to successfully generalize insights from local/regional field measurements across unsampled times and locations. GRRIEn analysis is generalizable, meaning results from a sample are translated to landscape scales by combining direct environmental measurements with global EOs using supervised machine learning; robust, meaning that the model shows good performance on data with scale-dependent feature and observation dependence; reproducible, based on a standard repository structure so that other scientists can quickly and easily replicate the analysis with a few computational tools; and interpreted, meaning that Earth scientists apply domain expertise to ensure that model parameters reflect a physically plausible diagnosis of the environmental system. This tutorial presents standard steps for achieving GRRIEn analysis by combining conventions of rigor in traditional experimental design with the open-science movement.
Significance Statement
Earth science researchers in the digital age are often tasked with pioneering big data analyses, yet have limited formal training in statistics and computational methods such as databasing or computer programming. Earth science researchers often spend tremendous amounts of time learning core computational skills, and making core analytical mistakes, in the process of bridging this training gap, at risk to the reputability of observational geostatistical research. The GRRIEn analytical framework is a practical guide introducing community standards for each phase of the computational research pipeline (dataset engineering, model training, and model diagnostics) to promote rigorous, accessible use of global EOs in Earth systems research.
Abstract
Globally available environmental observations (EOs), specifically from satellites and coupled Earth system models, represent some of the largest datasets of the digital age. As the volume of global EOs continues to grow, so does the potential of these data to help Earth scientists discover trends and patterns in Earth systems at large spatial scales. To leverage global EOs for scientific insight, Earth scientists need targeted and accessible exposure to skills in reproducible scientific computing and spatiotemporal data science, and to be empowered to apply their domain understanding to interpret data-driven models for knowledge discovery. The Generalizable, Reproducible, Robust, and Interpreted Environmental (GRRIEn) analysis framework was developed to prepare Earth scientists with an introductory statistics background and limited/no understanding of programming and computational methods to use global EOs to successfully generalize insights from local/regional field measurements across unsampled times and locations. GRRIEn analysis is generalizable, meaning results from a sample are translated to landscape scales by combining direct environmental measurements with global EOs using supervised machine learning; robust, meaning that the model shows good performance on data with scale-dependent feature and observation dependence; reproducible, based on a standard repository structure so that other scientists can quickly and easily replicate the analysis with a few computational tools; and interpreted, meaning that Earth scientists apply domain expertise to ensure that model parameters reflect a physically plausible diagnosis of the environmental system. This tutorial presents standard steps for achieving GRRIEn analysis by combining conventions of rigor in traditional experimental design with the open-science movement.
Significance Statement
Earth science researchers in the digital age are often tasked with pioneering big data analyses, yet have limited formal training in statistics and computational methods such as databasing or computer programming. Earth science researchers often spend tremendous amounts of time learning core computational skills, and making core analytical mistakes, in the process of bridging this training gap, at risk to the reputability of observational geostatistical research. The GRRIEn analytical framework is a practical guide introducing community standards for each phase of the computational research pipeline (dataset engineering, model training, and model diagnostics) to promote rigorous, accessible use of global EOs in Earth systems research.
Abstract
Two distinct features of anthropogenic climate change, warming in the tropical upper troposphere and warming at the Arctic surface, have competing effects on the midlatitude jet stream’s latitudinal position, often referred to as a “tug-of-war.” Studies that investigate the jet’s response to these thermal forcings show that it is sensitive to model type, season, initial atmospheric conditions, and the shape and magnitude of the forcing. Much of this past work focuses on studying a simulation’s response to external manipulation. In contrast, we explore the potential to train a convolutional neural network (CNN) on internal variability alone and then use it to examine possible nonlinear responses of the jet to tropospheric thermal forcing that more closely resemble anthropogenic climate change. Our approach leverages the idea behind the fluctuation–dissipation theorem, which relates the internal variability of a system to its forced response but so far has been only used to quantify linear responses. We train a CNN on data from a long control run of the CESM dry dynamical core and show that it is able to skillfully predict the nonlinear response of the jet to sustained external forcing. The trained CNN provides a quick method for exploring the jet stream sensitivity to a wide range of tropospheric temperature tendencies and, considering that this method can likely be applied to any model with a long control run, could be useful for early-stage experiment design.
Abstract
Two distinct features of anthropogenic climate change, warming in the tropical upper troposphere and warming at the Arctic surface, have competing effects on the midlatitude jet stream’s latitudinal position, often referred to as a “tug-of-war.” Studies that investigate the jet’s response to these thermal forcings show that it is sensitive to model type, season, initial atmospheric conditions, and the shape and magnitude of the forcing. Much of this past work focuses on studying a simulation’s response to external manipulation. In contrast, we explore the potential to train a convolutional neural network (CNN) on internal variability alone and then use it to examine possible nonlinear responses of the jet to tropospheric thermal forcing that more closely resemble anthropogenic climate change. Our approach leverages the idea behind the fluctuation–dissipation theorem, which relates the internal variability of a system to its forced response but so far has been only used to quantify linear responses. We train a CNN on data from a long control run of the CESM dry dynamical core and show that it is able to skillfully predict the nonlinear response of the jet to sustained external forcing. The trained CNN provides a quick method for exploring the jet stream sensitivity to a wide range of tropospheric temperature tendencies and, considering that this method can likely be applied to any model with a long control run, could be useful for early-stage experiment design.
Abstract
The objective of this paper is to employ machine learning (ML) and deep learning (DL) techniques to obtain, from input data (storm features) available in or derived from the HURDAT2 database, models capable of simulating important hurricane properties (e.g., landfall location and wind speed) consistent with historical records. In pursuit of this objective, a trajectory model providing the storm center in terms of longitude and latitude and intensity models providing the central pressure and maximum 1-min wind speed at 10-m elevation were created. The trajectory and intensity models are coupled and must be advanced together, 6 h at a time, as the features that serve as inputs to the models at any given step depend on predictions at the previous time steps. Once a synthetic storm database is generated, properties of interest, such as the frequencies of large wind speeds, may be extracted from any part of the simulation domain. The coupling of the trajectory and intensity models obviates the need for an intensity decay model inland of the coastline. Prediction results are compared with historical data, and the efficacy of the storm simulation models is evaluated at four sites: New Orleans, Louisiana; Miami, Florida; Cape Hatteras, North Carolina; and Boston, Massachusetts.
Abstract
The objective of this paper is to employ machine learning (ML) and deep learning (DL) techniques to obtain, from input data (storm features) available in or derived from the HURDAT2 database, models capable of simulating important hurricane properties (e.g., landfall location and wind speed) consistent with historical records. In pursuit of this objective, a trajectory model providing the storm center in terms of longitude and latitude and intensity models providing the central pressure and maximum 1-min wind speed at 10-m elevation were created. The trajectory and intensity models are coupled and must be advanced together, 6 h at a time, as the features that serve as inputs to the models at any given step depend on predictions at the previous time steps. Once a synthetic storm database is generated, properties of interest, such as the frequencies of large wind speeds, may be extracted from any part of the simulation domain. The coupling of the trajectory and intensity models obviates the need for an intensity decay model inland of the coastline. Prediction results are compared with historical data, and the efficacy of the storm simulation models is evaluated at four sites: New Orleans, Louisiana; Miami, Florida; Cape Hatteras, North Carolina; and Boston, Massachusetts.
Abstract
Many urban coastal communities are experiencing more profound flood impacts due to accelerated sea level rise that sometimes exceed their capacity to protect the built environment. In such cases, relocation may serve as a more effective hazard mitigation and adaptation strategy. However, it is unclear how urban residents living in flood-prone locations perceive the possibility of relocation and under what circumstances they would consider moving. Understanding the factors affecting an individual’s willingness to relocate because of coastal flooding is vital for developing accessible and equitable relocation policies. The main objective of this study is to identify the key considerations that would prompt urban coastal residents to consider permanent relocation because of coastal flooding. We leverage survey data collected from urban areas along the East Coast, assessing attitudes toward relocation, and design an artificial neural network (ANN) and a random forest (RF) model to find patterns in the survey data and indicate which considerations impact the decision to consider relocation. We trained the models to predict whether respondents would relocate because of socioeconomic factors, past exposure and experiences with flooding, and their flood-related concerns. Analyses performed on the models highlight the importance of flood-related concerns that accurately predict relocation behavior. Some common factors among the model analyses are concerns with increasing crime, the possibility of experiencing one more flood per year in the future, and more frequent business closures resulting from flooding.
Abstract
Many urban coastal communities are experiencing more profound flood impacts due to accelerated sea level rise that sometimes exceed their capacity to protect the built environment. In such cases, relocation may serve as a more effective hazard mitigation and adaptation strategy. However, it is unclear how urban residents living in flood-prone locations perceive the possibility of relocation and under what circumstances they would consider moving. Understanding the factors affecting an individual’s willingness to relocate because of coastal flooding is vital for developing accessible and equitable relocation policies. The main objective of this study is to identify the key considerations that would prompt urban coastal residents to consider permanent relocation because of coastal flooding. We leverage survey data collected from urban areas along the East Coast, assessing attitudes toward relocation, and design an artificial neural network (ANN) and a random forest (RF) model to find patterns in the survey data and indicate which considerations impact the decision to consider relocation. We trained the models to predict whether respondents would relocate because of socioeconomic factors, past exposure and experiences with flooding, and their flood-related concerns. Analyses performed on the models highlight the importance of flood-related concerns that accurately predict relocation behavior. Some common factors among the model analyses are concerns with increasing crime, the possibility of experiencing one more flood per year in the future, and more frequent business closures resulting from flooding.
Abstract
Accurate cloud-type identification and coverage analysis are crucial in understanding Earth’s radiative budget. Traditional computer vision methods rely on low-level visual features of clouds for estimating cloud coverage or sky conditions. Several handcrafted approaches have been proposed; however, scope for improvement still exists. Newer deep neural networks (DNNs) have demonstrated superior performance for cloud segmentation and categorization. These methods, however, need expert engineering intervention in the preprocessing steps—in the traditional methods—or human assistance in assigning cloud or clear-sky labels to a pixel for training DNNs. Such human mediation imposes considerable time and labor costs. We present the application of a new self-supervised learning approach to autonomously extract relevant features from sky images captured by ground-based cameras, for the classification and segmentation of clouds. We evaluate a joint embedding architecture that uses self-knowledge distillation plus regularization. We use two datasets to demonstrate the network’s ability to classify and segment sky images—one with ∼85 000 images collected from our ground-based camera and another with 400 labeled images from the WSISEG-Database. We find that this approach can discriminate full-sky images based on cloud coverage, diurnal variation, and cloud-base height. Furthermore, it semantically segments the cloud areas without labels. The approach shows competitive performance in all tested tasks, suggesting a new alternative for cloud characterization.
Significance Statement
Cloud macrophysical properties such as cloud-base height and coverage determine the amount of incoming radiation, mostly solar, and outgoing radiation, partly reflected from the sun and partly emitted from the Earth system, including the atmosphere. When this radiative budget is out of balance, it can affect our climate. Reporting sky conditions or cloud coverage from ground-based sky-imaging equipment is crucial in understanding Earth’s radiative budget. We present the application of a novel artificial intelligence approach to autonomously extract relevant features from sky images, for the characterization of atmospheric conditions. Unlike previous strategies, this novel approach requires reduced human intervention, suggesting a new path for cloud characterization.
Abstract
Accurate cloud-type identification and coverage analysis are crucial in understanding Earth’s radiative budget. Traditional computer vision methods rely on low-level visual features of clouds for estimating cloud coverage or sky conditions. Several handcrafted approaches have been proposed; however, scope for improvement still exists. Newer deep neural networks (DNNs) have demonstrated superior performance for cloud segmentation and categorization. These methods, however, need expert engineering intervention in the preprocessing steps—in the traditional methods—or human assistance in assigning cloud or clear-sky labels to a pixel for training DNNs. Such human mediation imposes considerable time and labor costs. We present the application of a new self-supervised learning approach to autonomously extract relevant features from sky images captured by ground-based cameras, for the classification and segmentation of clouds. We evaluate a joint embedding architecture that uses self-knowledge distillation plus regularization. We use two datasets to demonstrate the network’s ability to classify and segment sky images—one with ∼85 000 images collected from our ground-based camera and another with 400 labeled images from the WSISEG-Database. We find that this approach can discriminate full-sky images based on cloud coverage, diurnal variation, and cloud-base height. Furthermore, it semantically segments the cloud areas without labels. The approach shows competitive performance in all tested tasks, suggesting a new alternative for cloud characterization.
Significance Statement
Cloud macrophysical properties such as cloud-base height and coverage determine the amount of incoming radiation, mostly solar, and outgoing radiation, partly reflected from the sun and partly emitted from the Earth system, including the atmosphere. When this radiative budget is out of balance, it can affect our climate. Reporting sky conditions or cloud coverage from ground-based sky-imaging equipment is crucial in understanding Earth’s radiative budget. We present the application of a novel artificial intelligence approach to autonomously extract relevant features from sky images, for the characterization of atmospheric conditions. Unlike previous strategies, this novel approach requires reduced human intervention, suggesting a new path for cloud characterization.
Abstract
Earth system models struggle to simulate clouds and their radiative effects over the Southern Ocean, partly due to a lack of measurements and targeted cloud microphysics knowledge. We have evaluated biases of downwelling shortwave radiation in the ERA5 climate reanalysis using 25 years (1995–2019) of summertime surface measurements, collected on the Research and Supply Vessel (RSV) Aurora Australis, the Research Vessel (R/V) Investigator, and at Macquarie Island. During October–March daylight hours, the ERA5 simulation of SWdown exhibited large errors (mean bias = 54 W m−2, mean absolute error = 82 W m−2, root-mean-square error = 132 W m−2, and R 2 = 0.71). To determine whether we could improve these statistics, we bypassed ERA5’s radiative transfer model for SWdown with machine learning–based models using a number of ERA5’s gridscale meteorological variables as predictors. These models were trained and tested with the surface measurements of SWdown using a 10-fold shuffle split. An extreme gradient boosting (XGBoost) and a random forest–based model setup had the best performance relative to ERA5, both with a near complete reduction of the mean bias error, a decrease in the mean absolute error and root-mean-square error by 25% ± 3%, and an increase in the R 2 value of 5% ± 1% over the 10 splits. Large improvements occurred at higher latitudes and cyclone cold sectors, where ERA5 performed most poorly. We further interpret our methods using Shapley additive explanations. Our results indicate that data-driven techniques could have an important role in simulating surface radiation fluxes and in improving reanalysis products.
Significance Statement
Simulating the amount of sunlight reaching Earth’s surface is difficult because it relies on a good understanding of how much clouds absorb and scatter sunlight. Relative to summertime surface observations, the ERA5 reanalysis still overestimates the amount of sunlight entering the Southern Ocean. We taught some models how to predict the amount of sunlight entering the Southern Ocean using 25 years of surface observations and a small set of meteorological variables from ERA5. By bypassing the ERA5’s internal simulation of the absorption and scattering of sunlight, we can drastically reduce biases in the predicted surface shortwave radiation. Large improvements in cold sectors of cyclones and closer to Antarctica were observed in regions where many numerical models struggle to simulate the amount of incoming sunlight correctly.
Abstract
Earth system models struggle to simulate clouds and their radiative effects over the Southern Ocean, partly due to a lack of measurements and targeted cloud microphysics knowledge. We have evaluated biases of downwelling shortwave radiation in the ERA5 climate reanalysis using 25 years (1995–2019) of summertime surface measurements, collected on the Research and Supply Vessel (RSV) Aurora Australis, the Research Vessel (R/V) Investigator, and at Macquarie Island. During October–March daylight hours, the ERA5 simulation of SWdown exhibited large errors (mean bias = 54 W m−2, mean absolute error = 82 W m−2, root-mean-square error = 132 W m−2, and R 2 = 0.71). To determine whether we could improve these statistics, we bypassed ERA5’s radiative transfer model for SWdown with machine learning–based models using a number of ERA5’s gridscale meteorological variables as predictors. These models were trained and tested with the surface measurements of SWdown using a 10-fold shuffle split. An extreme gradient boosting (XGBoost) and a random forest–based model setup had the best performance relative to ERA5, both with a near complete reduction of the mean bias error, a decrease in the mean absolute error and root-mean-square error by 25% ± 3%, and an increase in the R 2 value of 5% ± 1% over the 10 splits. Large improvements occurred at higher latitudes and cyclone cold sectors, where ERA5 performed most poorly. We further interpret our methods using Shapley additive explanations. Our results indicate that data-driven techniques could have an important role in simulating surface radiation fluxes and in improving reanalysis products.
Significance Statement
Simulating the amount of sunlight reaching Earth’s surface is difficult because it relies on a good understanding of how much clouds absorb and scatter sunlight. Relative to summertime surface observations, the ERA5 reanalysis still overestimates the amount of sunlight entering the Southern Ocean. We taught some models how to predict the amount of sunlight entering the Southern Ocean using 25 years of surface observations and a small set of meteorological variables from ERA5. By bypassing the ERA5’s internal simulation of the absorption and scattering of sunlight, we can drastically reduce biases in the predicted surface shortwave radiation. Large improvements in cold sectors of cyclones and closer to Antarctica were observed in regions where many numerical models struggle to simulate the amount of incoming sunlight correctly.
Abstract
We assess the suitability of unpaired image-to-image translation networks for bias correcting data simulated by global atmospheric circulation models. We use the Unsupervised Image-to-Image Translation (UNIT) neural network architecture to map between data from the HadGEM3-A-N216 model and ERA5 reanalysis data in a geographical area centered on the South Asian monsoon, which has well-documented serious biases in this model. The UNIT network corrects cross-variable correlations and spatial structures but creates bias corrections with less extreme values than the target distribution. By combining the UNIT neural network with the classical technique of quantile mapping (QM), we can produce bias corrections that are better than either alone. The UNIT+QM scheme is shown to correct cross-variable correlations, spatial patterns, and all marginal distributions of single variables. The careful correction of such joint distributions is of high importance for compound extremes research.
Abstract
We assess the suitability of unpaired image-to-image translation networks for bias correcting data simulated by global atmospheric circulation models. We use the Unsupervised Image-to-Image Translation (UNIT) neural network architecture to map between data from the HadGEM3-A-N216 model and ERA5 reanalysis data in a geographical area centered on the South Asian monsoon, which has well-documented serious biases in this model. The UNIT network corrects cross-variable correlations and spatial structures but creates bias corrections with less extreme values than the target distribution. By combining the UNIT neural network with the classical technique of quantile mapping (QM), we can produce bias corrections that are better than either alone. The UNIT+QM scheme is shown to correct cross-variable correlations, spatial patterns, and all marginal distributions of single variables. The careful correction of such joint distributions is of high importance for compound extremes research.
