A Case Study Analysis of Clustering Applied to VAS Measurements

Abstract

This study is a joint effort to test clustering of VISSR Atmospheric Sounder measurements in a more operational setting. The operational setting differs from previous tests of clustering. The clustering technique had to handle VAS measurements which were not spatially continuous, due to operational scanning requirements for skipping certain scan lines. This was handled by considering only those VAS channels which were available at every Field of View. The operational test also dealt with cloud-contaminated FOVs. Clustering can detect cloud-contaminated VAS measurements by treating them as a group and eliminating clusters which are either outliers or which produce suspect soundings. A comparison was made between retrievals produced from clustered VAS measurements to retrievals produced from blocked VAS measurements. Differences between retrievals using the two methods are significant, especially considering the small area of concern. Neither set of retrievals can necessarily be shown to be better, due to a lack of conventional measurements for comparison at such high resolution. However, slight improvements in retrievals can be expected due to increased signal-to-noise of the VAS measurement which are clustered as compared to present blocking schemes.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 07, 1990
Accession Number
ADA228926

Entities

People

  • Donald W. Hillger
  • James F. Purdom
  • John S. Snook

Tags

Communities of Interest

  • Space

DTIC Thesaurus Topics

  • Aircrafts
  • Algorithms
  • Artificial Satellites
  • Atmospheres
  • Case Studies
  • Climate Change
  • Clustering
  • Contamination
  • High Resolution
  • Infrared Detectors
  • Measurement
  • Radiosondes
  • Surface Temperature
  • Teamwork
  • Test And Evaluation
  • Water Vapor

Fields of Study

  • Environmental science

Readers

  • Atmospheric Remote Sensing.
  • Quantum Chemistry
  • Systems Analysis and Design