Change detection of satellite imagery for reconnaissance of stressed tropical corals

Considering the vast area of coral reefs, and the remote nature of many of them, a simple but effective procedure to quickly identify the effect of ecosystem stress from satellite imagery is needed. The stresses may be related to El Nino induced temperature and radiation changes, pollution and situation, or unsustainable fishing practices. The procedure must not be dependent upon in situ measurements that typically have been used to correct beam attenuation through the water column, yet must identify signature changes at a variety of depths. The authors have selected Savusavu Bay of Fiji as a test region. A gold mine was recently recommissioned on the west side of the bay. During an anomalous winter storm, the retaining walls of some effluent storage facilities failed and a variety of chemicals were released through the adjacent river system and dispersed over a large coral reef. The authors have SPOT satellite imagery from before and after the event. Preliminary rsconnaissance indicates severe damage to the corals with colonization by algae. The authors assess the change detected from the multidate SPOT data. The Getis statistic, which is based solely on image spatial characteristics, is evaluated as a tool for change detection. Preliminary examination suggests that it meets the requirements for rapid assessment for environmental change without the need for individual image calibration based upon in situ information