Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Complex hybrid models combining deterministic and machine learning components for numerical climate modeling and

Vladimir M Krasnopolsky1, Michael S Fox-Rabinovitz

  • 1Earth System Science Interdisciplinary Center, University of Maryland, College Park, MD, USA. vladimir.krasnopolsky@noaa.gov

Neural Networks : the Official Journal of the International Neural Network Society
|March 11, 2006
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Neural network approach for robust and fast calculation of physical processes in numerical environmental models: compound parameterization with a quality control of larger errors.

Neural networks : the official journal of the International Neural Network Society·2008
Same author

Reducing uncertainties in neural network Jacobians and improving accuracy of neural network emulations with NN ensemble approaches.

Neural networks : the official journal of the International Neural Network Society·2007
Same author

Some neural network applications in environmental sciences. Part I: forward and inverse problems in geophysical remote measurements.

Neural networks : the official journal of the International Neural Network Society·2003
Same author

Some neural network applications in environmental sciences. Part II: advancing computational efficiency of environmental numerical models.

Neural networks : the official journal of the International Neural Network Society·2003

A novel hybrid environmental model combines deterministic and machine learning components for climate modeling and weather prediction. This approach significantly speeds up complex simulations while maintaining accuracy, enabling model improvements.

Area of Science:

  • Environmental numerical modeling
  • Climate science
  • Machine learning applications

Background:

  • Traditional environmental models are computationally intensive.
  • Accurate parameterizations of physical processes are crucial for climate and weather models.
  • Existing models face limitations in speed and complexity.

Purpose of the Study:

  • To introduce a novel hybrid environmental model.
  • To explore the integration of neural networks (NNs) with deterministic models.
  • To enhance the speed and efficiency of climate and weather predictions.

Main Methods:

  • Developed a hybrid model by combining NN-based emulations of physics parameterizations with deterministic model dynamics.
  • Used NNs to emulate time-consuming components like radiation parameterizations.

Related Experiment Videos

  • Integrated these NN emulations into a general circulation model (GCM) to create a hybrid GCM (HGCM).

    • Main Results:

    • Achieved highly accurate and significantly faster emulations of complex model physics.
    • HGCM simulations produced results comparable to traditional GCMs.
    • Demonstrated substantial speed-up in model calculations compared to conventional GCMs.

    Conclusions:

    • The developed hybrid GCM approach is feasible and efficient for modeling complex systems.
    • NN emulations offer a viable method to accelerate environmental numerical modeling.
    • This advancement opens new avenues for improving climate and weather prediction models.