Browse
Abstract
Meteorological satellite imagery is a critical asset for observing and forecasting weather phenomena. The Joint Polar Satellite System (JPSS) Visible Infrared Imaging Radiometer Suite (VIIRS) Day–Night Band (DNB) sensor collects measurements from moonlight, airglow, and artificial lights. DNB radiances are then manipulated and scaled with a focus on digital display. DNB imagery performance is tied to the lunar cycle, with the best performance during the full moon and the worst with the new moon. We propose using feed-forward neural network models to transform brightness temperatures and wavelength differences in the infrared spectrum to a pseudo-lunar reflectance value based on lunar reflectance values derived from observed DNB radiances. JPSS NOAA-20 and Suomi National Polar-Orbiting Partnership (SNPP) satellite data over the North Pacific Ocean at night for full moon periods from December 2018 to November 2020 were used to design the models. The pseudo-lunar reflectance values are quantitatively compared to DNB lunar reflectance, providing the first-ever lunar reflectance baseline metrics. The resulting imagery product, Machine Learning Nighttime Visible Imagery (ML-NVI), is qualitatively compared to DNB lunar reflectance and infrared imagery across the lunar cycle. The imagery goal is not only to improve upon the consistent performance of DNB imagery products across the lunar cycle, but ultimately to lay the foundation for transitioning the algorithm to geostationary sensors, making global continuous nighttime imagery possible. ML-NVI demonstrates its ability to provide DNB-derived imagery with consistent contrast and representation of clouds across the full lunar cycle for nighttime cloud detection.
Significance Statement
This study explores the creation and evaluation of a feed-forward neural network to generate synthetic lunar reflectance values and imagery from VIIRS infrared channels. The model creates lunar reflectance values typical of full moon scenes, enabling quantifiable comparisons for nighttime imagery evaluations. Additionally, it creates imagery that highlights low clouds better than its infrared counterparts. Results indicate the ability to create visually consistent nighttime visible imagery across the full lunar cycle for the improved nighttime detection of low clouds. Wavelengths chosen are available on both polar and geostationary satellite sensors to support the utilization of the algorithm on multiple sensor platforms for improved temporal resolution and greater simultaneous geographic coverage over polar orbiters alone.
Abstract
Meteorological satellite imagery is a critical asset for observing and forecasting weather phenomena. The Joint Polar Satellite System (JPSS) Visible Infrared Imaging Radiometer Suite (VIIRS) Day–Night Band (DNB) sensor collects measurements from moonlight, airglow, and artificial lights. DNB radiances are then manipulated and scaled with a focus on digital display. DNB imagery performance is tied to the lunar cycle, with the best performance during the full moon and the worst with the new moon. We propose using feed-forward neural network models to transform brightness temperatures and wavelength differences in the infrared spectrum to a pseudo-lunar reflectance value based on lunar reflectance values derived from observed DNB radiances. JPSS NOAA-20 and Suomi National Polar-Orbiting Partnership (SNPP) satellite data over the North Pacific Ocean at night for full moon periods from December 2018 to November 2020 were used to design the models. The pseudo-lunar reflectance values are quantitatively compared to DNB lunar reflectance, providing the first-ever lunar reflectance baseline metrics. The resulting imagery product, Machine Learning Nighttime Visible Imagery (ML-NVI), is qualitatively compared to DNB lunar reflectance and infrared imagery across the lunar cycle. The imagery goal is not only to improve upon the consistent performance of DNB imagery products across the lunar cycle, but ultimately to lay the foundation for transitioning the algorithm to geostationary sensors, making global continuous nighttime imagery possible. ML-NVI demonstrates its ability to provide DNB-derived imagery with consistent contrast and representation of clouds across the full lunar cycle for nighttime cloud detection.
Significance Statement
This study explores the creation and evaluation of a feed-forward neural network to generate synthetic lunar reflectance values and imagery from VIIRS infrared channels. The model creates lunar reflectance values typical of full moon scenes, enabling quantifiable comparisons for nighttime imagery evaluations. Additionally, it creates imagery that highlights low clouds better than its infrared counterparts. Results indicate the ability to create visually consistent nighttime visible imagery across the full lunar cycle for the improved nighttime detection of low clouds. Wavelengths chosen are available on both polar and geostationary satellite sensors to support the utilization of the algorithm on multiple sensor platforms for improved temporal resolution and greater simultaneous geographic coverage over polar orbiters alone.
Abstract
The purpose of this research is to build an operational model for predicting wildfire occurrence for the contiguous United States (CONUS) in the 1–10-day range using the U-Net 3+ machine learning model. This paper illustrates the range of model performance resulting from choices made in the modeling process, such as how labels are defined for the model and how input variables are codified for the model. By combining the capabilities of the U-Net 3+ model with a neighborhood loss function, fractions skill score (FSS), we can quantify model success by predictions made both in and around the location of the original fire occurrence label. The model is trained on weather, weather-derived fuel, and topography observational inputs and labels representing fire occurrence. Observational weather, weather-derived fuel, and topography data are sourced from the gridded surface meteorological (gridMET) dataset, a daily, CONUS-wide, high-spatial-resolution dataset of surface meteorological variables. Fire occurrence labels are sourced from the U.S. Department of Agriculture’s Fire Program Analysis Fire-Occurrence Database (FPA-FOD), which contains spatial wildfire occurrence data for CONUS, combining data sourced from the reporting systems of federal, state, and local organizations. By exploring the many aspects of the modeling process with the added context of model performance, this work builds understanding around the use of deep learning to predict fire occurrence in CONUS.
Significance Statement
Our work seeks to explore the limits to which deep learning can predict wildfire occurrence in CONUS with the ultimate goal of providing decision support to those allocating fire resources during high fire seasons. By exploring with what accuracy and lead time we can provide insights to these persons, we hope to reduce loss of life, reduce damage to property, and improve future event preparedness. We compare two models, one trained on all fires in the continental United States and the other on only large lightning fires. We found that a model trained on all fires produced a higher probability of fire.
Abstract
The purpose of this research is to build an operational model for predicting wildfire occurrence for the contiguous United States (CONUS) in the 1–10-day range using the U-Net 3+ machine learning model. This paper illustrates the range of model performance resulting from choices made in the modeling process, such as how labels are defined for the model and how input variables are codified for the model. By combining the capabilities of the U-Net 3+ model with a neighborhood loss function, fractions skill score (FSS), we can quantify model success by predictions made both in and around the location of the original fire occurrence label. The model is trained on weather, weather-derived fuel, and topography observational inputs and labels representing fire occurrence. Observational weather, weather-derived fuel, and topography data are sourced from the gridded surface meteorological (gridMET) dataset, a daily, CONUS-wide, high-spatial-resolution dataset of surface meteorological variables. Fire occurrence labels are sourced from the U.S. Department of Agriculture’s Fire Program Analysis Fire-Occurrence Database (FPA-FOD), which contains spatial wildfire occurrence data for CONUS, combining data sourced from the reporting systems of federal, state, and local organizations. By exploring the many aspects of the modeling process with the added context of model performance, this work builds understanding around the use of deep learning to predict fire occurrence in CONUS.
Significance Statement
Our work seeks to explore the limits to which deep learning can predict wildfire occurrence in CONUS with the ultimate goal of providing decision support to those allocating fire resources during high fire seasons. By exploring with what accuracy and lead time we can provide insights to these persons, we hope to reduce loss of life, reduce damage to property, and improve future event preparedness. We compare two models, one trained on all fires in the continental United States and the other on only large lightning fires. We found that a model trained on all fires produced a higher probability of fire.
Abstract
This paper illustrates the lessons learned as we applied the U-Net3+ deep learning model to the task of building an operational model for predicting wildfire occurrence for the contiguous United States (CONUS) in the 1–10-day range. Through the lens of model performance, we explore the reasons for performance improvements made possible by the model. Lessons include the importance of labeling, the impact of information loss in input variables, and the role of operational considerations in the modeling process. This work offers lessons learned for other interdisciplinary researchers working at the intersection of deep learning and fire occurrence prediction with an eye toward operationalization.
Abstract
This paper illustrates the lessons learned as we applied the U-Net3+ deep learning model to the task of building an operational model for predicting wildfire occurrence for the contiguous United States (CONUS) in the 1–10-day range. Through the lens of model performance, we explore the reasons for performance improvements made possible by the model. Lessons include the importance of labeling, the impact of information loss in input variables, and the role of operational considerations in the modeling process. This work offers lessons learned for other interdisciplinary researchers working at the intersection of deep learning and fire occurrence prediction with an eye toward operationalization.
Abstract
Global deep learning weather prediction models have recently been shown to produce forecasts that rival those from physics-based models run at operational centers. It is unclear whether these models have encoded atmospheric dynamics or simply pattern matching that produces the smallest forecast error. Answering this question is crucial to establishing the utility of these models as tools for basic science. Here, we subject one such model, Pangu-Weather, to a set of four classical dynamical experiments that do not resemble the model training data. Localized perturbations to the model output and the initial conditions are added to steady time-averaged conditions, to assess the propagation speed and structural evolution of signals away from the local source. Perturbing the model physics by adding a steady tropical heat source results in a classical Matsuno–Gill response near the heating and planetary waves that radiate into the extratropics. A localized disturbance on the winter-averaged North Pacific jet stream produces realistic extratropical cyclones and fronts, including the spontaneous emergence of polar lows. Perturbing the 500-hPa height field alone yields adjustment from a state of rest to one of wind–pressure balance over ∼6 h. Localized subtropical low pressure systems produce Atlantic hurricanes, provided the initial amplitude exceeds about 4 hPa, and setting the initial humidity to zero eliminates hurricane development. We conclude that the model encodes realistic physics in all experiments and suggest that it can be used as a tool for rapidly testing a wide range of hypotheses.
Abstract
Global deep learning weather prediction models have recently been shown to produce forecasts that rival those from physics-based models run at operational centers. It is unclear whether these models have encoded atmospheric dynamics or simply pattern matching that produces the smallest forecast error. Answering this question is crucial to establishing the utility of these models as tools for basic science. Here, we subject one such model, Pangu-Weather, to a set of four classical dynamical experiments that do not resemble the model training data. Localized perturbations to the model output and the initial conditions are added to steady time-averaged conditions, to assess the propagation speed and structural evolution of signals away from the local source. Perturbing the model physics by adding a steady tropical heat source results in a classical Matsuno–Gill response near the heating and planetary waves that radiate into the extratropics. A localized disturbance on the winter-averaged North Pacific jet stream produces realistic extratropical cyclones and fronts, including the spontaneous emergence of polar lows. Perturbing the 500-hPa height field alone yields adjustment from a state of rest to one of wind–pressure balance over ∼6 h. Localized subtropical low pressure systems produce Atlantic hurricanes, provided the initial amplitude exceeds about 4 hPa, and setting the initial humidity to zero eliminates hurricane development. We conclude that the model encodes realistic physics in all experiments and suggest that it can be used as a tool for rapidly testing a wide range of hypotheses.
Abstract
Increases in wildfire activity and the resulting impacts have prompted the development of high-resolution wildfire behavior models for forecasting fire spread. Recent progress in using satellites to detect fire locations further provides the opportunity to use measurements toward improving fire spread forecasts from numerical models through data assimilation. This work develops a physics-informed approach for inferring the history of a wildfire from satellite measurements, providing the necessary information to initialize coupled atmosphere–wildfire models from a measured wildfire state. The fire arrival time, which is the time the fire reaches a given spatial location, acts as a succinct representation of the history of a wildfire. In this work, a conditional Wasserstein generative adversarial network (cWGAN), trained with WRF–SFIRE simulations, is used to infer the fire arrival time from satellite active fire data. The cWGAN is used to produce samples of likely fire arrival times from the conditional distribution of arrival times given satellite active fire detections. Samples produced by the cWGAN are further used to assess the uncertainty of predictions. The cWGAN is tested on four California wildfires occurring between 2020 and 2022, and predictions for fire extent are compared against high-resolution airborne infrared measurements. Further, the predicted ignition times are compared with reported ignition times. An average Sørensen’s coefficient of 0.81 for the fire perimeters and an average ignition time difference of 32 min suggest that the method is highly accurate.
Significance Statement
To initialize coupled atmosphere–wildfire simulations in a physically consistent way based on satellite measurements of active fire locations, it is critical to ensure the state of the fire and atmosphere aligns at the start of the forecast. If known, the history of a wildfire may be used to develop an atmospheric state matching the wildfire state determined from satellite data in a process known as spinup. In this paper, we present a novel method for inferring the early stage history of a wildfire based on satellite active fire measurements. Here, inference of the fire history is performed in a probabilistic sense and physics is further incorporated through the use of training data derived from a coupled atmosphere–wildfire model.
Abstract
Increases in wildfire activity and the resulting impacts have prompted the development of high-resolution wildfire behavior models for forecasting fire spread. Recent progress in using satellites to detect fire locations further provides the opportunity to use measurements toward improving fire spread forecasts from numerical models through data assimilation. This work develops a physics-informed approach for inferring the history of a wildfire from satellite measurements, providing the necessary information to initialize coupled atmosphere–wildfire models from a measured wildfire state. The fire arrival time, which is the time the fire reaches a given spatial location, acts as a succinct representation of the history of a wildfire. In this work, a conditional Wasserstein generative adversarial network (cWGAN), trained with WRF–SFIRE simulations, is used to infer the fire arrival time from satellite active fire data. The cWGAN is used to produce samples of likely fire arrival times from the conditional distribution of arrival times given satellite active fire detections. Samples produced by the cWGAN are further used to assess the uncertainty of predictions. The cWGAN is tested on four California wildfires occurring between 2020 and 2022, and predictions for fire extent are compared against high-resolution airborne infrared measurements. Further, the predicted ignition times are compared with reported ignition times. An average Sørensen’s coefficient of 0.81 for the fire perimeters and an average ignition time difference of 32 min suggest that the method is highly accurate.
Significance Statement
To initialize coupled atmosphere–wildfire simulations in a physically consistent way based on satellite measurements of active fire locations, it is critical to ensure the state of the fire and atmosphere aligns at the start of the forecast. If known, the history of a wildfire may be used to develop an atmospheric state matching the wildfire state determined from satellite data in a process known as spinup. In this paper, we present a novel method for inferring the early stage history of a wildfire based on satellite active fire measurements. Here, inference of the fire history is performed in a probabilistic sense and physics is further incorporated through the use of training data derived from a coupled atmosphere–wildfire model.
Abstract
The prediction of tropical rain rates from atmospheric profiles poses significant challenges, mainly due to the heavy-tailed distribution exhibited by tropical rainfall. This study introduces overparameterized neural networks not only to forecast tropical rain rates but also to explain their heavy-tailed distribution. The investigation is separately conducted for three rain types (stratiform, deep convective, and shallow convective) observed by the Global Precipitation Measurement satellite radar over the west and east Pacific regions. Atmospheric profiles of humidity, temperature, and zonal and meridional winds from the MERRA-2 reanalysis are considered as features. Although overparameterized neural networks are well known for their “double descent phenomenon,” little has been explored about their applicability to climate data and capability of capturing the tail behavior of data. In our results, overparameterized neural networks accurately estimate the rain-rate distributions and outperform other machine learning methods. Spatial maps show that overparameterized neural networks also successfully describe the spatial patterns of each rain type across the tropical Pacific. In addition, we assess the feature importance for each overparameterized neural network to provide insight into the key factors driving the predictions, with low-level humidity and temperature variables being the overall most important. These findings highlight the capability of overparameterized neural networks in predicting the distribution of the rain rate and explaining extreme values.
Significance Statement
This study aims to introduce the capability of overparameterized neural networks, a type of neural network with more parameters than data points, in predicting the distribution of tropical rain rates from gridscale environmental variables and explaining their tail behavior. Rainfall prediction has been a topic of importance, yet it remains a challenging problem for its heavy-tailed nature. Overparameterized neural networks correctly captured rain-rate distributions and the spatial patterns and heterogeneity of the observed rain rates for multiple rain types, which could not be achieved by any other previous statistical or machine learning frameworks. We find that overparameterized neural networks can play a key role in general prediction tasks, with potential expanded applicability to other domains with heavy-tailed data distribution.
Abstract
The prediction of tropical rain rates from atmospheric profiles poses significant challenges, mainly due to the heavy-tailed distribution exhibited by tropical rainfall. This study introduces overparameterized neural networks not only to forecast tropical rain rates but also to explain their heavy-tailed distribution. The investigation is separately conducted for three rain types (stratiform, deep convective, and shallow convective) observed by the Global Precipitation Measurement satellite radar over the west and east Pacific regions. Atmospheric profiles of humidity, temperature, and zonal and meridional winds from the MERRA-2 reanalysis are considered as features. Although overparameterized neural networks are well known for their “double descent phenomenon,” little has been explored about their applicability to climate data and capability of capturing the tail behavior of data. In our results, overparameterized neural networks accurately estimate the rain-rate distributions and outperform other machine learning methods. Spatial maps show that overparameterized neural networks also successfully describe the spatial patterns of each rain type across the tropical Pacific. In addition, we assess the feature importance for each overparameterized neural network to provide insight into the key factors driving the predictions, with low-level humidity and temperature variables being the overall most important. These findings highlight the capability of overparameterized neural networks in predicting the distribution of the rain rate and explaining extreme values.
Significance Statement
This study aims to introduce the capability of overparameterized neural networks, a type of neural network with more parameters than data points, in predicting the distribution of tropical rain rates from gridscale environmental variables and explaining their tail behavior. Rainfall prediction has been a topic of importance, yet it remains a challenging problem for its heavy-tailed nature. Overparameterized neural networks correctly captured rain-rate distributions and the spatial patterns and heterogeneity of the observed rain rates for multiple rain types, which could not be achieved by any other previous statistical or machine learning frameworks. We find that overparameterized neural networks can play a key role in general prediction tasks, with potential expanded applicability to other domains with heavy-tailed data distribution.
Abstract
We developed and applied a machine-learned discretization for one-dimensional (1D) horizontal passive scalar advection, which is an operator component common to all chemical transport models (CTMs). Our learned advection scheme resembles a second-order accurate, three-stencil numerical solver but differs from a traditional solver in that coefficients for each equation term are output by a neural network rather than being theoretically derived constants. We subsampled higher-resolution simulation results—resulting in up to 16× larger grid size and 64× larger time step—and trained our neural-network-based scheme to match the subsampled integration data. In this way, we created an operator that has low resolution (in time or space) but can reproduce the behavior of a high-resolution traditional solver. Our model shows high fidelity in reproducing its training dataset (a single 10-day 1D simulation) and is similarly accurate in simulations with unseen initial conditions, wind fields, and grid spacing. In many cases, our learned solver is more accurate than a low-resolution version of the reference solver, but the low-resolution reference solver achieves greater computational speedup (500× acceleration) over the high-resolution simulation than the learned solver is able to (18× acceleration). Surprisingly, our learned 1D scheme—when combined with a splitting technique—can be used to predict 2D advection and is in some cases more stable and accurate than the low-resolution reference solver in 2D. Overall, our results suggest that learned advection operators may offer a higher-accuracy method for accelerating CTM simulations as compared to simply running a traditional integrator at low resolution.
Significance Statement
Chemical transport modeling (CTM) is an essential tool for studying air pollution. CTM simulations take a long computing time. Modeling pollutant transport (advection) is the second most computationally intensive part of the model. Decreasing the resolution not only reduces the advection computing time but also decreases accuracy. We employed machine learning to reduce the resolution of advection while keeping the accuracy. We verified the robustness of our solver with several generalization testing scenarios. In our 2D simulation, our solver showed up to 100 times faster simulation with fair accuracy. Integrating our approach to existing CTMs will allow broadened participation in the study of air pollution and related solutions.
Abstract
We developed and applied a machine-learned discretization for one-dimensional (1D) horizontal passive scalar advection, which is an operator component common to all chemical transport models (CTMs). Our learned advection scheme resembles a second-order accurate, three-stencil numerical solver but differs from a traditional solver in that coefficients for each equation term are output by a neural network rather than being theoretically derived constants. We subsampled higher-resolution simulation results—resulting in up to 16× larger grid size and 64× larger time step—and trained our neural-network-based scheme to match the subsampled integration data. In this way, we created an operator that has low resolution (in time or space) but can reproduce the behavior of a high-resolution traditional solver. Our model shows high fidelity in reproducing its training dataset (a single 10-day 1D simulation) and is similarly accurate in simulations with unseen initial conditions, wind fields, and grid spacing. In many cases, our learned solver is more accurate than a low-resolution version of the reference solver, but the low-resolution reference solver achieves greater computational speedup (500× acceleration) over the high-resolution simulation than the learned solver is able to (18× acceleration). Surprisingly, our learned 1D scheme—when combined with a splitting technique—can be used to predict 2D advection and is in some cases more stable and accurate than the low-resolution reference solver in 2D. Overall, our results suggest that learned advection operators may offer a higher-accuracy method for accelerating CTM simulations as compared to simply running a traditional integrator at low resolution.
Significance Statement
Chemical transport modeling (CTM) is an essential tool for studying air pollution. CTM simulations take a long computing time. Modeling pollutant transport (advection) is the second most computationally intensive part of the model. Decreasing the resolution not only reduces the advection computing time but also decreases accuracy. We employed machine learning to reduce the resolution of advection while keeping the accuracy. We verified the robustness of our solver with several generalization testing scenarios. In our 2D simulation, our solver showed up to 100 times faster simulation with fair accuracy. Integrating our approach to existing CTMs will allow broadened participation in the study of air pollution and related solutions.
Abstract
Forecasters routinely calibrate their confidence in model forecasts. Ensembles inherently estimate forecast confidence but are often underdispersive, and ensemble spread does not strongly correlate with ensemble-mean error. The misalignment between ensemble spread and skill motivates new methods for “forecasting forecast skill” so that forecasters can better utilize ensemble guidance. We have trained logistic regression and random forest models to predict the skill of composite reflectivity forecasts from the NSSL Warn-on-Forecast System (WoFS), a 3-km ensemble that generates rapidly updating forecast guidance for 0–6-h lead times. The forecast skill predictions are valid at 1-, 2-, or 3-h lead times within localized regions determined by the observed storm locations at analysis time. We use WoFS analysis and forecast output and NSSL Multi-Radar/Multi-Sensor composite reflectivity for 106 cases from the 2017 to 2021 NOAA Hazardous Weather Testbed Spring Forecasting Experiments. We frame the prediction task as a multiclassification problem, where the forecast skill labels are determined by averaging the extended fraction skill scores (eFSSs) for several reflectivity thresholds and verification neighborhoods and then converting to one of three classes based on where the average eFSS ranks within the entire dataset: POOR (bottom 20%), FAIR (middle 60%), or GOOD (top 20%). Initial machine learning (ML) models are trained on 323 predictors; reducing to 10 or 15 predictors in the final models only modestly reduces skill. The final models substantially outperform carefully developed persistence- and spread-based models and are reasonably explainable. The results suggest that ML can be a valuable tool for guiding user confidence in convection-allowing (and larger-scale) ensemble forecasts.
Significance Statement
Some numerical weather prediction (NWP) forecasts are more likely to verify than others. Forecasters often recognize situations where NWP output should be trusted more or less than usual, but objective methods for “forecasting forecast skill” are notably lacking for thunderstorm-scale models. Better estimates of forecast skill can benefit society through more accurate forecasts of high-impact weather. Machine learning (ML) provides a powerful framework for relating forecast skill to the characteristics of model forecasts and available observations over many previous cases. ML models can leverage these relationships to predict forecast skill for new cases in real time. We demonstrate the effectiveness of this approach to forecasting forecast skill using a cutting-edge thunderstorm prediction system and logistic regression and random forest models. Based on this success, we recommend the adoption of similar ML-based methods for other prediction models.
Abstract
Forecasters routinely calibrate their confidence in model forecasts. Ensembles inherently estimate forecast confidence but are often underdispersive, and ensemble spread does not strongly correlate with ensemble-mean error. The misalignment between ensemble spread and skill motivates new methods for “forecasting forecast skill” so that forecasters can better utilize ensemble guidance. We have trained logistic regression and random forest models to predict the skill of composite reflectivity forecasts from the NSSL Warn-on-Forecast System (WoFS), a 3-km ensemble that generates rapidly updating forecast guidance for 0–6-h lead times. The forecast skill predictions are valid at 1-, 2-, or 3-h lead times within localized regions determined by the observed storm locations at analysis time. We use WoFS analysis and forecast output and NSSL Multi-Radar/Multi-Sensor composite reflectivity for 106 cases from the 2017 to 2021 NOAA Hazardous Weather Testbed Spring Forecasting Experiments. We frame the prediction task as a multiclassification problem, where the forecast skill labels are determined by averaging the extended fraction skill scores (eFSSs) for several reflectivity thresholds and verification neighborhoods and then converting to one of three classes based on where the average eFSS ranks within the entire dataset: POOR (bottom 20%), FAIR (middle 60%), or GOOD (top 20%). Initial machine learning (ML) models are trained on 323 predictors; reducing to 10 or 15 predictors in the final models only modestly reduces skill. The final models substantially outperform carefully developed persistence- and spread-based models and are reasonably explainable. The results suggest that ML can be a valuable tool for guiding user confidence in convection-allowing (and larger-scale) ensemble forecasts.
Significance Statement
Some numerical weather prediction (NWP) forecasts are more likely to verify than others. Forecasters often recognize situations where NWP output should be trusted more or less than usual, but objective methods for “forecasting forecast skill” are notably lacking for thunderstorm-scale models. Better estimates of forecast skill can benefit society through more accurate forecasts of high-impact weather. Machine learning (ML) provides a powerful framework for relating forecast skill to the characteristics of model forecasts and available observations over many previous cases. ML models can leverage these relationships to predict forecast skill for new cases in real time. We demonstrate the effectiveness of this approach to forecasting forecast skill using a cutting-edge thunderstorm prediction system and logistic regression and random forest models. Based on this success, we recommend the adoption of similar ML-based methods for other prediction models.
Abstract
Precipitation values produced by climate models are biased due to the parameterization of physical processes and limited spatial resolution. Current bias-correction approaches usually focus on correcting lower-order statistics (mean and standard deviation), which make it difficult to capture precipitation extremes. However, accurate modeling of extremes is critical for policymaking to mitigate and adapt to the effects of climate change. We develop a deep learning framework, leveraging information from key dynamical variables impacting precipitation to also match higher-order statistics (skewness and kurtosis) for the entire precipitation distribution, including extremes. The deep learning framework consists of a two-part architecture: a U-Net convolutional network to capture the spatiotemporal distribution of precipitation and a fully connected network to capture the distribution of higher-order statistics. The joint network, termed UFNet, can simultaneously improve the spatial structure of the modeled precipitation and capture the distribution of extreme precipitation values. Using climate model simulation data and observations that are climatologically similar but not strictly paired, the UFNet identifies and corrects the climate model biases, significantly improving the estimation of daily precipitation as measured by a broad range of spatiotemporal statistics. In particular, UFNet significantly improves the underestimation of extreme precipitation values seen with current bias-correction methods. Our approach constitutes a generalized framework for correcting other climate model variables which improves the accuracy of the climate model predictions, while utilizing a simpler and more stable training process.
Abstract
Precipitation values produced by climate models are biased due to the parameterization of physical processes and limited spatial resolution. Current bias-correction approaches usually focus on correcting lower-order statistics (mean and standard deviation), which make it difficult to capture precipitation extremes. However, accurate modeling of extremes is critical for policymaking to mitigate and adapt to the effects of climate change. We develop a deep learning framework, leveraging information from key dynamical variables impacting precipitation to also match higher-order statistics (skewness and kurtosis) for the entire precipitation distribution, including extremes. The deep learning framework consists of a two-part architecture: a U-Net convolutional network to capture the spatiotemporal distribution of precipitation and a fully connected network to capture the distribution of higher-order statistics. The joint network, termed UFNet, can simultaneously improve the spatial structure of the modeled precipitation and capture the distribution of extreme precipitation values. Using climate model simulation data and observations that are climatologically similar but not strictly paired, the UFNet identifies and corrects the climate model biases, significantly improving the estimation of daily precipitation as measured by a broad range of spatiotemporal statistics. In particular, UFNet significantly improves the underestimation of extreme precipitation values seen with current bias-correction methods. Our approach constitutes a generalized framework for correcting other climate model variables which improves the accuracy of the climate model predictions, while utilizing a simpler and more stable training process.
