Use of fourier transforms to define landscape scales of analysis for disturbances: a case study of thinned and unthinned forest stands

Abstract

Various disturbances such as disease and management practices causecanopy gaps that change patterns of forest stand structure. This study examinedthe usefulness of digital image analysis using aerial photos, FourierTranforms,and cluster analysis to investigate how different spatial statistics areaffected by spatial scale. The specific aims were to: 1) evaluate how a Fourierfilter could be used to classify canopy gap sizes objectively, 2) determinewhich statistics might be useful for detecting and measuring disturbanceimpacts, and 3) examine the potential for this method to determine spatialdomains in a pair of ponderosa pine (Pinusponderosa) stands in the Black Hills of South Dakota,USA.The eventual goal is to develop an operational method of assessing the impactsof natural disturbances such as disease. Results indicated that several spatialmetrics discriminated between harvested and unharvested stands. We hypothesizethat these metrics will be useful as spatial measures of disease impact if theanalyses are performed on specific size classes of forest gaps.