Tests of an Improved Oceanographic Expert System

Abstract

For 8 years the Naval Research Laboratory's Remote Sensing Applications Branch has been developing an oceanographic expert system (OES) that models some aspects of the kinematics of the Gulf Stream and its associated eddies. The OES uses a rule base to provide ring-motion forecasts and a neural network to forecast Gulf Stream motion. Previous work showed that OES forecasts of ring motion are at least as good as those produced by other methods. Recent work led to improvements in the ring-motion geometry equations, replacement of the Gulf Stream motion logic, and the addition of a natural-language explanation facility. The first two changes required 'retuning' of the OES. The changes were designed to remove linear trends in the mean forecast position errors for noninteracting (with the Gulf Stream) rings. This report presents a comparison of the present system's performance with 'prototype' and improved parameters. For noninteracting rings, the improved system provides ring-motion forecasts that are superior to a no-motion assumption 75% or more of the time for both warm-core rings and cold-core rings, for both 7-day and 14-day forecasts. The forecast ring positions are within 20 km of the true positions. This report provides more complete information on the OES and the test results. Eddies, Gulf Stream.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 04, 1994
Accession Number
ADA280123

Entities

People

  • Matthew Lybanon

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • Space

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Computers
  • Data Sets
  • Expert Systems
  • Geographic Regions
  • Geometry
  • Gulf Stream
  • Intelligent Systems
  • Knowledge Based Systems
  • Language
  • Military Research
  • Models
  • Natural Languages
  • Naval Warfare
  • Neural Networks
  • Remote Sensing
  • Statistics

Readers

  • Atmospheric Science/Meteorology
  • Coastal and Marine Engineering/Sediment Transport/Hydraulic Engineering
  • Oceanography.

Technology Areas

  • AI & ML