Characterizing Wetland Landscapes: A Spatiotemporal Analysis of Remotely Sensed Data at Cheyenne Bottoms, Kansas

Abstract

Generalized wetland landscape characterization efforts using remotely sensed satellite data have typically employed land-use and cover-classification routines with mixed results. This study argues that such classification routines are, at times, limited in their ability to capture the fluctuating spatial configuration of water and terrestrial features unique to wetland environments. Rather, we recommend the use of a multitemporal imaging approach, which records land-cover variability based directly on land-surface feature reflectivity signals. Using Landsat Thematic Mapper data and small-format aerial photography, this study monitors wetland change over a period of 15 yrs. at the internationally recognized Cheyenne Bottoms wetland site in central Kansas. Noticeable fluctuations in wetland cover resulting from both variable precipitation patterns and changing stewardship activities are observed. In addition to documenting generalized spatial trends, the multitemporal method effectively monitors plant vigor, while high spatial resolution small format aerial photography aids in monitoring cattail growth trends and is used to coordinate targeted management efforts at controlling this invasive plant. Our results suggest that multitemporal remote sensing methods are well suited to capturing qualitative information on change in wetland and other similar environments observing fuzzy and often shifting transitions between land and water.