Techniques for computer-assisted mapping of rangeland change

Stephen Yool, Mary Jane Makaio, Joseph M. Watts

Research output: Contribution to journalArticle

28 Citations (Scopus)

Abstract

Management of grasslands subject to replacement by woody species requires an understanding of the scales and patterns of change and how to detect and express them. We used to Jornada del Muerto Basin of southern New Mexico as a case study, testing the suitability of Landsat Multi-Spectral Scanner (MSS) data for detecting vegetation changes. Cycles of drought and heavy grazing have apparently changed the once extensive grasslands of the upland Jornada and surrounding areas gradually into a patchwork of shrublands and relict grasslands. Integrated remote sensing and geographic information system (GIS) techniques can facilitate automated detection of these rangeland changes. A GIS was used to store and process two 4-band, co-registered multi-temporal Landsat MSS scenes collected in July 1983 and August 1992. Scene-to-scene radiometric calibration was performed using a regression technique. The data were then evaluated for changes 3 different ways using algorithms based on 1) differences between the 'red' (chlorophyll absorption) bands for the 2 scenes; 2) the Euclidean distances between the 'red and 'near-infrared' bands for the scenes; and 3) a standardized principal components analysis using all 8 MSS bands. A threshold of 3 standard deviations above the mean was applied to each of the 3 resulting 'change' images to represent areas of extreme change. Correlations among these images ranged between 0.83 and 0.95. We conclude these techniques can identify successfully the patterns and extent of extreme change, and thus have potential value for management of our rangelands resources.

Original languageEnglish (US)
Pages (from-to)307-314
Number of pages8
JournalJournal of Range Management
Volume50
Issue number3
StatePublished - May 1997

Fingerprint

computer techniques
scanners
rangeland
rangelands
Landsat multispectral scanner
range management
grassland
Landsat
geographic information systems
grasslands
shrubland
shrublands
remote sensing
principal component analysis
near infrared
chlorophyll
highlands
calibration
replacement
grazing

Keywords

  • change detection
  • GIS
  • range management
  • remote sensing

ASJC Scopus subject areas

  • Animal Science and Zoology
  • Ecology

Cite this

Techniques for computer-assisted mapping of rangeland change. / Yool, Stephen; Makaio, Mary Jane; Watts, Joseph M.

In: Journal of Range Management, Vol. 50, No. 3, 05.1997, p. 307-314.

Research output: Contribution to journalArticle

Yool, S, Makaio, MJ & Watts, JM 1997, 'Techniques for computer-assisted mapping of rangeland change', Journal of Range Management, vol. 50, no. 3, pp. 307-314.
Yool, Stephen ; Makaio, Mary Jane ; Watts, Joseph M. / Techniques for computer-assisted mapping of rangeland change. In: Journal of Range Management. 1997 ; Vol. 50, No. 3. pp. 307-314.
@article{2a1716886e12408e80f6232fa6812f00,
title = "Techniques for computer-assisted mapping of rangeland change",
abstract = "Management of grasslands subject to replacement by woody species requires an understanding of the scales and patterns of change and how to detect and express them. We used to Jornada del Muerto Basin of southern New Mexico as a case study, testing the suitability of Landsat Multi-Spectral Scanner (MSS) data for detecting vegetation changes. Cycles of drought and heavy grazing have apparently changed the once extensive grasslands of the upland Jornada and surrounding areas gradually into a patchwork of shrublands and relict grasslands. Integrated remote sensing and geographic information system (GIS) techniques can facilitate automated detection of these rangeland changes. A GIS was used to store and process two 4-band, co-registered multi-temporal Landsat MSS scenes collected in July 1983 and August 1992. Scene-to-scene radiometric calibration was performed using a regression technique. The data were then evaluated for changes 3 different ways using algorithms based on 1) differences between the 'red' (chlorophyll absorption) bands for the 2 scenes; 2) the Euclidean distances between the 'red and 'near-infrared' bands for the scenes; and 3) a standardized principal components analysis using all 8 MSS bands. A threshold of 3 standard deviations above the mean was applied to each of the 3 resulting 'change' images to represent areas of extreme change. Correlations among these images ranged between 0.83 and 0.95. We conclude these techniques can identify successfully the patterns and extent of extreme change, and thus have potential value for management of our rangelands resources.",
keywords = "change detection, GIS, range management, remote sensing",
author = "Stephen Yool and Makaio, {Mary Jane} and Watts, {Joseph M.}",
year = "1997",
month = "5",
language = "English (US)",
volume = "50",
pages = "307--314",
journal = "Rangeland Ecology and Management",
issn = "1550-7424",
publisher = "Society for Range Management",
number = "3",

}

TY - JOUR

T1 - Techniques for computer-assisted mapping of rangeland change

AU - Yool, Stephen

AU - Makaio, Mary Jane

AU - Watts, Joseph M.

PY - 1997/5

Y1 - 1997/5

N2 - Management of grasslands subject to replacement by woody species requires an understanding of the scales and patterns of change and how to detect and express them. We used to Jornada del Muerto Basin of southern New Mexico as a case study, testing the suitability of Landsat Multi-Spectral Scanner (MSS) data for detecting vegetation changes. Cycles of drought and heavy grazing have apparently changed the once extensive grasslands of the upland Jornada and surrounding areas gradually into a patchwork of shrublands and relict grasslands. Integrated remote sensing and geographic information system (GIS) techniques can facilitate automated detection of these rangeland changes. A GIS was used to store and process two 4-band, co-registered multi-temporal Landsat MSS scenes collected in July 1983 and August 1992. Scene-to-scene radiometric calibration was performed using a regression technique. The data were then evaluated for changes 3 different ways using algorithms based on 1) differences between the 'red' (chlorophyll absorption) bands for the 2 scenes; 2) the Euclidean distances between the 'red and 'near-infrared' bands for the scenes; and 3) a standardized principal components analysis using all 8 MSS bands. A threshold of 3 standard deviations above the mean was applied to each of the 3 resulting 'change' images to represent areas of extreme change. Correlations among these images ranged between 0.83 and 0.95. We conclude these techniques can identify successfully the patterns and extent of extreme change, and thus have potential value for management of our rangelands resources.

AB - Management of grasslands subject to replacement by woody species requires an understanding of the scales and patterns of change and how to detect and express them. We used to Jornada del Muerto Basin of southern New Mexico as a case study, testing the suitability of Landsat Multi-Spectral Scanner (MSS) data for detecting vegetation changes. Cycles of drought and heavy grazing have apparently changed the once extensive grasslands of the upland Jornada and surrounding areas gradually into a patchwork of shrublands and relict grasslands. Integrated remote sensing and geographic information system (GIS) techniques can facilitate automated detection of these rangeland changes. A GIS was used to store and process two 4-band, co-registered multi-temporal Landsat MSS scenes collected in July 1983 and August 1992. Scene-to-scene radiometric calibration was performed using a regression technique. The data were then evaluated for changes 3 different ways using algorithms based on 1) differences between the 'red' (chlorophyll absorption) bands for the 2 scenes; 2) the Euclidean distances between the 'red and 'near-infrared' bands for the scenes; and 3) a standardized principal components analysis using all 8 MSS bands. A threshold of 3 standard deviations above the mean was applied to each of the 3 resulting 'change' images to represent areas of extreme change. Correlations among these images ranged between 0.83 and 0.95. We conclude these techniques can identify successfully the patterns and extent of extreme change, and thus have potential value for management of our rangelands resources.

KW - change detection

KW - GIS

KW - range management

KW - remote sensing

UR - http://www.scopus.com/inward/record.url?scp=0030619097&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030619097&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0030619097

VL - 50

SP - 307

EP - 314

JO - Rangeland Ecology and Management

JF - Rangeland Ecology and Management

SN - 1550-7424

IS - 3

ER -