Mapping Wetlands Cover Types with Directional Polarization Signatures

We present an approach to mapping surface waters at local to global scales. In prior research we demonstrated that wide angle remotely sensed imagery having high radiance values measured near the direction of the sun glint in the principal plane signaled the presence of either open water or inundated plant communities, in contrast to dry upland plant communities. We demonstrated that regional to global scale wetlands issues that do not involve one meter resolution per se may be addressed with acceptable accuracy by applying spectral mixture analysis (SMA) and atmospheric correction techniques to the 6-10 km pixel imagery from POLDER. Here we investigate the influence of wind speed upon the discrimination process, showing that at very low wind speeds and very high wind speeds we are not able to discriminate the three cover types and that when wind speeds are moderate there is potentially an error associated with estimation of the areal extent of open water areas.