Near Reconnaissance Level Zone of Influence Interpretations for Vegetated Space 'Source Title' Asian-Pacific Remote Sensing Journal Edition Vol.4, No.2, January 1992
Bibliographic details
Summary
The northwestern boundary of Pakistan's Cholistan desert was selected for a study of the spectral composition of vegetation profiles with multiband SPOT HRV and Landsat TM data. The objectives encompassed regression fits on zones of active and passive change within vegetatin across areal sampling of approximately 3,460 km2. Measures of radiance were converted to indices of change that explained the homogeneity of land use over subsets of vegetation overlap, with classified themes like open cover/fallow land, silt and sand dunes. The matrixed composition of these indices of change, through a graphical model, could be used not only to infer the nature of variable space along best-fit projections of zones of influence but also the banded contribution of each sensor to the variable space. For the SPOT HRV XS results, variable aggregation along a zone of influence is indicative of a trend wherein changes in dense vegetation herald subsequent changes in its affiliation with other classified variables. Its multispectral band width is such as to facilitate large but definitive groupings in the variable space with a minimum of overlap. Landsat TM data sets have been particularly revealing in regard to many interactions within the variable space that had not been encumbered within the coarser spectral resolutin of SPOT data. The relatively complex interactive mix is also inclusive of a trend that emphasizes increasing homogeneity of land use away from the site of origin. The model invokes a tertiary level of interpretation over vegetation space, which itself is a useful paramter when assessing biomass for areal and yield estimations. Against successive validations through multiseasonal extractions across varied-zones of coverage, the modelled renderings could prove supplemental to rec nnaissance level attributes of irrigation suitability land classification by successively highlighting those areas most receptive to emerging land-use dynamics.