- Satellite Imagery
- IMaRS Team
- Contact IMaRS
The Millennium Coral Reef Mapping Project: Understanding, Classifiying and Mapping Coral Reef Structures Worldwide Using High Resolution Remote Sensing Spaceborne Images.
The current trend in the development of algorithms for coral reef mapping using satellite images is to try to build analytical procedures to automatically derive the percent cover of various components, or end-members, of the reefs (coral, algae, sediments, etc.). In principle, this should be the most rapid approach for consistently processing large areas. However, such efficiency can only be obtained if the reefs to be mapped have a limited number of relevant classes that are spectrally distinguishable using available wavelengths. If any reef could be mapped in the same way using the right spectral library, then processing is theoretically reproducible in time and space. Recent simulations suggest that this approach will likely become feasible and successful in the future when using very high spatial/spectral resolution imagery and sensors optimized for spectral resolution (Hochberg and Atkinson (2003), Hochberg et al. (2003)).
At the present time, an automated global algorithm cannot completely identify the end members that are desired by the research and management communities. Bathymetric effects, changing optical properties of the water column and highly heterogeneous cover make this task extremely challenging in the real world. For global studies, heterogeneity of cover is a particular problem, because an optimal solution of end members in one location may produce unsatisfactory results when applied to another location. Moreover, this approach is justified only if one is interested in the end-members of the reefs (brown algae, crustose coralline algae, hard and soft corals, sand, pavement, etc.) and if these end-members imply some sort of ecological, biological, chemical, geological or socio-economical properties useful to the user of the map. For instance, coral percent cover is an important parameter because it is linked to such measurement objectives as reef health, reef calcification, or fish population sheltering. These end-members inherently represent and carry properties/functions of the reefs at a given scale, in time and space.
There are an infinite number of combinations of end-members, providing in theory an infinite numbers of reefs. But in reality, the number of reefs on this planet is finite. Even though processes that control reef growth/erosion are complex and have provided a large variety of reef structures, there are still a finite number of basic reef structures (Kennedy and Woodroffe, 2002). In the same way that community structure and cover by corals, algae, and sand can be used as end-members to represent properties and functions of reefs (e.g. resilience, resistance), there are geomorphological end-members that also provide information on reef properties and functions but at a different scales in time and space (Hopley, 1982, Pandolfi, 2002). These properties and functions are longer-term processes, not necessarily those tracked by the coral-algae-sand end-members. Geomorphological end-members are related to reef growth, reef erosion, sea-level variations, ocean chemistry, hydrodynamics, climate, weather and tectonic events (McNutt and Menard, 1978, Hopley, 1982, Purdy and Winterer, 2001, Montaggioni, 2001, Kennedy and Woodroffe, 2002). They reflect energy and matter cycling processes, benthic productivity and settlement of biodiversity at geological scales. These processes cannot necessarily be inferred using current percent cover of dominant coral-algae species (Hatcher, 1997).