Application of airborne multispectral digital imagery to quantify riverine habitats at different base flows

Abstract

AbstractWe employed an integrated system of airborne remote sensing and ground surveys for regional mapping of instream habitats under variable flows over a 70 km section of the Lower Yakima River in southern Washington, USA. Airborne multispectral digital imagery was obtained in conjunction with field survey measurements and used to quantify the spatial extent, condition and temporal changes of selected river habitat characteristics under two different flows (14 and 28 m3 s−1). Under each flow, geomorphic measures were quantified (e.g. channel complexity, number and size of habitats). Water depth and velocity were also classified for instream habitats, and temperature and turbidity were recorded. Remote sensing classification accuracies for islands, exposed rocks and water surfaces were greater than 99%, while more detailed depth/flow classifications were less accurate (68% and 72%, for the high and low flows, respectively). While high turbidity (>4 Nephelometric Turbidity Units (NTU)), shadows and bidirectional reflectance factor reduced classification accuracies, the overall effect of these factors was minimal. Under the low‐flow situation, off‐channel habitats were less abundant, more isolated and had shallower depths and warmer temperatures. Our analysis suggests that airborne multispectral imagery, coupled with appropriate ground truth data, can be a viable method for regional mapping of diverse riverine habitats under variable flows. We concluded from this analysis that the higher flow situation provided substantially better habitat than currently exists in the Lower Yakima River. Copyright © 2002 John Wiley & Sons, Ltd.