Wetlands—estuaries, inland wetlands, and freshwater lakes

Two major disasters, the Indian Ocean Tsunami of December 2004 and the flooding of New Orleans after Hurricane Katrina in August 2005, have heightened global awareness of the importance of wetlands for reducing wave energies and negative impacts of floods on coastal communities (Danielsen et al. 2005). Both situations have also led to research that uses remote sensing to help understand changes in coastal wetlands over regional scales. These types of studies would be difficult to complete with classic field methods because of the breadth of their spatio-temporal scopes. Remote sensing helps scientists to identify the most beneficial approaches to reduce wetland losses, and to target restoration programs. Remote sensing can increase understanding of wetland change and provide an evidence base for policy makers. We will start with an example of a major analysis of the historical conversion of mangrove habitats prior to the Indian Ocean Tsunami, seeking insights into whether intact coastal wetlands provide protection. We will have a related example for the Louisiana coast and hurricane vulnerability later in the chapter. Giri et al. (2008) used more than 750 Landsat images to map tsunami-prone coastal areas of Indonesia, Malaysia, Thailand, Burma (Myanmar), Bangladesh, India, and Sri Lanka. Imagery was centered on four different time periods (as close as possible to the central calendar year, given cloud cover in many images): mid-1970s, 1990, around 2000, and 2005. Because of the size of the study area, they resampled data to the Albers equal area map projection, normalized for solar irradiance, and produced maps for each time period. The authors used supervised classification to map the water bodies and unsupervised classification (isodata cluster analysis) to classify the remaining images as mangrove, non-mangrove, or barren lands. Field data and high-resolution satellite images (QuickBird, IKONOS) were the source of map validation. They then produced post-classification change maps by subtracting the classifications of pairs of wetland maps, comparing 1975–90, 1975–2000, 1975–2005, 1990–2000, 1990–2005 and 2000–5.