Land classification is so central to geography that its use, and the use of its derivative and corresponding products, is seldom even questioned. Since its earliest implementations land classification has adapted to changes in geographic scale and in the nature of the categorical systematics upon which it is based. Land classification has changed in its techniques and in how it adapts to technological changes, particularly those related to remote sensing and geographic information systems. The adaptation of land classification to digital pixel-based classification spawned a wide range of land classification error analysis techniques. These techniques do not easily transfer to non-pixel based classification error analysis as recent research on rapid land assessment methodologies and land change error analysis has shown. This disparity suggests a need to reevaluate the very nature of land classification research. To begin such an evaluation, this lecture provides a retrospective on the roots of land classification research, examines some of the milestones of that research, and describes the divergent paths such research has taken. It examines the importance of land classification in these times of ever decreasing global resources, and reviews its potential legal, social, and economic implications. Based on this retrospective, this paper advances the need for geographic researchers to envision land classification not only as a set of techniques, but more generally to focus on systematic geography in all its facets as a research agenda in its own right.
Estimating Aboveground Biomass in Tropical Forests: Field Methods and Error Analysis for the Calibration of Remote Sensing Observations
