ABSTRACT
Wetlands are highly diverse ecosystems that play a strong role in global climate projections; acting as both sources and sinks for carbon. Detection of wetlands by remote sensing means can be challenging due extreme variation in size, shape, and spectral response as well as a tendency for intra-annual fluctuations in water extent and vegetative components in temperate latitudes. In this chapter, a case study is presented demonstrating an effective approach for detecting inland freshwater wetlands of varying size, shape, and composition using high spatial resolution GeoEye-1 imagery, additional derivative inputs, and an object-based multiresolution segmentation approach. This method was evaluated on two freshwater wetlands that undergo significant change in water extent and spatial distribution of macrophytic communities within a single growing season. Case study results yielded an overall accuracy of 90% during the early spring window and 85% during the late season summer window. Higher early season accuracy was attributed to lower vegetation presence allowing clearer delineation of aquatic boundaries compared to late season vegetation growth, which partly obscured the water signal beneath. While multisensor approaches have also shown success, this case study demonstrates that a single-sensor optical imagery can also provide an accessible option for accurate wetland mapping and monitoring.
