Assessing and monitoring semi-arid shrublands using object-based image analysis and multiple endmember spectral mixture analysis

Yuki Hamada, Douglas A. Stow, Dar A. Roberts, Janet Franklin, Phaedon C. Kyriakidis

Research output: Contribution to journalArticlepeer-review

23 Scopus citations

Abstract

Arid and semi-arid shrublands have significant biological and economical values and have been experiencing dramatic changes due to human activities. In California, California sage scrub (CSS) is one of the most endangered plant communities in the US and requires close monitoring in order to conserve this important biological resource. We investigate the utility of remote-sensing approaches - object-based image analysis applied to pansharpened QuickBird imagery (QBPS/OBIA) and multiple endmember spectral mixture analysis (MESMA) applied to SPOT imagery (SPOT/MESMA) - for estimating fractional cover of true shrub, subshrub, herb, and bare ground within CSS communities of southern California. We also explore the effectiveness of life-form cover maps for assessing CSS conditions. Overall and combined shrub cover (i.e.; true shrub and subshrub) were estimated more accurately using QBPS/OBIA (mean absolute error or MAE, 8.9 %) than SPOT/MESMA (MAE, 11.4 %). Life-form cover from QBPS/OBIA at a 25 × 25 m grid cell size seems most desirable for assessing CSS because of its higher accuracy and spatial detail in cover estimates and amenability to extracting other vegetation information (e.g.; size, shape, and density of shrub patches). Maps derived from SPOT/MESMA at a 50 × 50 m scale are effective for retrospective analysis of life-form cover change because their comparable accuracies to QBPS/OBIA and availability of SPOT archives data dating back to the mid-1980s. The framework in this study can be applied to other physiognomically comparable shrubland communities.

Original languageEnglish (US)
Pages (from-to)3173-3190
Number of pages18
JournalEnvironmental Monitoring and Assessment
Volume185
Issue number4
DOIs
StatePublished - Apr 2013

Keywords

  • Habitat monitoring
  • Mixture models
  • Object-based image analysis
  • Remote sensing
  • Shrublands

ASJC Scopus subject areas

  • General Environmental Science
  • Pollution
  • Management, Monitoring, Policy and Law

Fingerprint

Dive into the research topics of 'Assessing and monitoring semi-arid shrublands using object-based image analysis and multiple endmember spectral mixture analysis'. Together they form a unique fingerprint.

Cite this