Complementarity of radar and visible-infrared sensors in assessing rangeland condition

Abstract

Ground investigations of rangeland use and condition are restricted by resource and travel limitations. Management of these vast, multiuser resources could be enhanced by information from remote sensing. Microwave radar imagery is becoming readily available and its all weather capability provides greater reliability than visible-infrared (VIR) sensors. The objective of this study was to evaluate a combination of radar and visible-infrared sensors as tools in rangeland monitoring. Data from Landsat Thematic Mapper (TM) and SPOT VIR sensors were compared with data from airborne and ERS-1 synthetic aperture radar (SAR) sensors to determine similarities and contrasts with a view to exploiting any synergism. The study site was the Agriculture and Agri-Food Canada Onefour Research Substation in southeastern Alberta. The TM and SPOT VIR sensor bands were highly correlated. The radar sensors were correlated with each other to a lesser degree. Correlations between VIR and radar were not high. Vegetation type influenced VIR reflectance and radar backscatter. Russian wildrye pastures had high radar backscatter as well as high VIR reflectance. Native range had low VIR reflectance and low radar backscatter. Crested wheatgrass pastures had low VIR reflectance and high radar backscatter. Other features, such as exposed sedimentary Cretaceous softrock, had high values for both sensor groups, while intermittent water bodies or shallow depressions characterized by high clay content and strong microtopography had high radar backscatter but low VIR reflectance. More information can be obtained from the combination of both types of sensor than from either alone.