A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics
Abstract Due to the rapidly changing climate, the frequency and severity of extreme weather is expected to increase over the coming decades. As fully‐resolved climate simulations remain computationally intractable, policy makers must rely on coarse‐models to quantify risk for extremes. However, coar...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
American Geophysical Union (AGU)
2024-03-01
|
| Series: | Journal of Advances in Modeling Earth Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1029/2023MS004122 |
| _version_ | 1827292855365795840 |
|---|---|
| author | B. Barthel Sorensen A. Charalampopoulos S. Zhang B. E. Harrop L. R. Leung T. P. Sapsis |
| author_facet | B. Barthel Sorensen A. Charalampopoulos S. Zhang B. E. Harrop L. R. Leung T. P. Sapsis |
| author_sort | B. Barthel Sorensen |
| collection | DOAJ |
| description | Abstract Due to the rapidly changing climate, the frequency and severity of extreme weather is expected to increase over the coming decades. As fully‐resolved climate simulations remain computationally intractable, policy makers must rely on coarse‐models to quantify risk for extremes. However, coarse models suffer from inherent bias due to the ignored “sub‐grid” scales. We propose a framework to non‐intrusively debias coarse‐resolution climate predictions using neural‐network (NN) correction operators. Previous efforts have attempted to train such operators using loss functions that match statistics. However, this approach falls short with events that have longer return period than that of the training data, since the reference statistics have not converged. Here, the scope is to formulate a learning method that allows for correction of dynamics and quantification of extreme events with longer return period than the training data. The key obstacle is the chaotic nature of the underlying dynamics. To overcome this challenge, we introduce a dynamical systems approach where the correction operator is trained using reference data and a coarse model simulation nudged toward that reference. The method is demonstrated on debiasing an under‐resolved quasi‐geostrophic model and the Energy Exascale Earth System Model (E3SM). For the former, our method enables the quantification of events that have return period two orders longer than the training data. For the latter, when trained on 8 years of ERA5 data, our approach is able to correct the coarse E3SM output to closely reflect the 36‐year ERA5 statistics for all prognostic variables and significantly reduce their spatial biases. |
| first_indexed | 2024-04-24T13:15:07Z |
| format | Article |
| id | doaj.art-857a65e9d48e46ba857b4da0245ef22d |
| institution | Directory Open Access Journal |
| issn | 1942-2466 |
| language | English |
| last_indexed | 2024-04-24T13:15:07Z |
| publishDate | 2024-03-01 |
| publisher | American Geophysical Union (AGU) |
| record_format | Article |
| series | Journal of Advances in Modeling Earth Systems |
| spelling | doaj.art-857a65e9d48e46ba857b4da0245ef22d2024-04-04T21:25:34ZengAmerican Geophysical Union (AGU)Journal of Advances in Modeling Earth Systems1942-24662024-03-01163n/an/a10.1029/2023MS004122A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events StatisticsB. Barthel Sorensen0A. Charalampopoulos1S. Zhang2B. E. Harrop3L. R. Leung4T. P. Sapsis5Department of Mechanical Engineering Massachusetts Institute of Technology Cambridge MA USADepartment of Mechanical Engineering Massachusetts Institute of Technology Cambridge MA USAPacific Northwest National Laboratory Richland WA USAPacific Northwest National Laboratory Richland WA USAPacific Northwest National Laboratory Richland WA USADepartment of Mechanical Engineering Massachusetts Institute of Technology Cambridge MA USAAbstract Due to the rapidly changing climate, the frequency and severity of extreme weather is expected to increase over the coming decades. As fully‐resolved climate simulations remain computationally intractable, policy makers must rely on coarse‐models to quantify risk for extremes. However, coarse models suffer from inherent bias due to the ignored “sub‐grid” scales. We propose a framework to non‐intrusively debias coarse‐resolution climate predictions using neural‐network (NN) correction operators. Previous efforts have attempted to train such operators using loss functions that match statistics. However, this approach falls short with events that have longer return period than that of the training data, since the reference statistics have not converged. Here, the scope is to formulate a learning method that allows for correction of dynamics and quantification of extreme events with longer return period than the training data. The key obstacle is the chaotic nature of the underlying dynamics. To overcome this challenge, we introduce a dynamical systems approach where the correction operator is trained using reference data and a coarse model simulation nudged toward that reference. The method is demonstrated on debiasing an under‐resolved quasi‐geostrophic model and the Energy Exascale Earth System Model (E3SM). For the former, our method enables the quantification of events that have return period two orders longer than the training data. For the latter, when trained on 8 years of ERA5 data, our approach is able to correct the coarse E3SM output to closely reflect the 36‐year ERA5 statistics for all prognostic variables and significantly reduce their spatial biases.https://doi.org/10.1029/2023MS004122climate modelingextreme event statisticsdebiasingnudgingmachine learning |
| spellingShingle | B. Barthel Sorensen A. Charalampopoulos S. Zhang B. E. Harrop L. R. Leung T. P. Sapsis A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics Journal of Advances in Modeling Earth Systems climate modeling extreme event statistics debiasing nudging machine learning |
| title | A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics |
| title_full | A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics |
| title_fullStr | A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics |
| title_full_unstemmed | A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics |
| title_short | A Non‐Intrusive Machine Learning Framework for Debiasing Long‐Time Coarse Resolution Climate Simulations and Quantifying Rare Events Statistics |
| title_sort | non intrusive machine learning framework for debiasing long time coarse resolution climate simulations and quantifying rare events statistics |
| topic | climate modeling extreme event statistics debiasing nudging machine learning |
| url | https://doi.org/10.1029/2023MS004122 |
| work_keys_str_mv | AT bbarthelsorensen anonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT acharalampopoulos anonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT szhang anonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT beharrop anonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT lrleung anonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT tpsapsis anonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT bbarthelsorensen nonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT acharalampopoulos nonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT szhang nonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT beharrop nonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT lrleung nonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics AT tpsapsis nonintrusivemachinelearningframeworkfordebiasinglongtimecoarseresolutionclimatesimulationsandquantifyingrareeventsstatistics |