A Sunspot Maximum Prediction Using a Neural Network

Abstract

Based on the rapid increase in the sunspot numbers for the current solar cycle, the current cycle (number 22) is expected to produce the largest number of major solar flares of any cycle in recent history. Since a number of different types of spacecraft anomalies are related maxima in the solar activity, it is important to predict both the amplitude of the solar cycle and the time at which the maximum is expected to occur. We have taken a novel approach to making such a prediction. We have used Neural Network Simulation Software from California Scientific Software to predict the maximum 13-month- smoothed sunspot number for cycle 22 and the month in which this maximum will occur. The neural network predicts the maximum sunspot number for cycle 22 to be 194 + or - 26, to occur 42 months (March 1990) after the minimum. The uncertainty in the prediction is taken to be the standard deviation of the difference between the predicted sunspot maximum and the observed sunspot maximum for 15 test cases taken from previous cycles. the prediction is in line with the predictions by NASA, 195 + or - 40, to occur in February 1990, and by NOAA, Space Environment Laboratory, 203 + or - 30, to occur in March 1990.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 1990
Accession Number
ADA225866

Entities

People

  • D. J. Gorney
  • Harry C. Koons

Organizations

  • The Aerospace Corporation

Tags

Communities of Interest

  • Advanced Electronics
  • Space

DTIC Thesaurus Topics

  • Artificial Satellites
  • Chemical Reactions
  • Chemistry
  • Classification
  • Compound Semiconductors
  • Corporations
  • Cosmic Rays
  • Detectors
  • Laser Spectroscopy
  • Materials
  • Materials Science
  • Physics Laboratories
  • Security
  • Semiconductor Devices
  • Solar Activity
  • Space Sciences
  • Space Systems

Readers

  • Computational Modeling and Simulation
  • Space/Atmospheric Physics.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks
  • Space