Mapping Coral Reefs from Space

A. Mapping Coral Reefs From Space
B. The Case for Remote Sensing of Coral Reefs
C. Coral Reef Satellite Mission

A. Mapping Coral Reefs From Space

To further its mission, PCRF has assembled an interdisciplinary team of advisors and scientists from a variety of institutions including: the College of Charleston, Linnaean Society (U.K.), MIT, Scripps Institution of Oceanography, the Stevens Institute of Technology and USC.  Over the past decade, at a series of workshops and conferences, this team has explored the question of whether it is possible to map coral reefs from space, detect change, and then monitor changes in their health and vitality conditions over time. 

Using radiative transfer theory, PCRF's team has investigated the upwelling optical properties of reef organisms, and results indicate coral reef signals can indeed be observed from space. (Lubin et al, 2001). In 1991, Dr. Phil Dustan, PCRF's Principal Investigator for the study of coral reefs, was the first to map coral reefs using satellite imagery, thereby proving it possible. Dr. Dustan's studies were conducted off the Florida coast, and he used Spot satellite imagery (Dustan et al, 2000). Since then, the team on board PCRF's research vessel has conducted ground-truthing operations on coral reef communities in Southeast Asia comparing actual surveys to Spot imagery. Further research is now underway which demonstrates how satellite imagery can be used to detect change in coral reef health through remote sensing (Dustan et al, 2002).

Although existing satellite technology, such as Spot, Landsat and Ikonos, has provided imagery proving the possibility of mapping and monitoring coral reefs from space, these images are far too gross to provide the level of quantitative and qualitative data necessary for creating a comprehensive map of living coral reefs or detecting and monitoring changes in their health on a global basis. These existing satellite systems do not possess the spatial, spectral or orbital specifications required for this purpose. Therefore, a specially designed coral reef sensor needs to be placed on a dedicated satellite to achieve this objective.

Additionally, for remote locations, satellites must be specifically programmed to acquire imagery, and until recently, reefs have not been considered important targets. For this reason, while single, gross images of coral reefs exist, these images are not acquired routinely but rather selectively and by chance and largely at the discretion of government agencies who happen to deem a tropical country or coastline of political or military significance. Thus, it is rare to find multiple useable images of individual reef systems ñ which are essential for mapping and monitoring them over time. Finally, even if it were possible to acquire all the existing satellite images necessary to cover the coral reef global area, these images would still be very imprecise for coral reef mapping and monitoring, and the cost to purchase these images ñ not including analysis and presentation ñ is estimated to exceed $50 million.

B. The Case for Remote Sensing of Coral Reefs

Remote sensing technology is the only means to supply the data necessary to map and monitor reefs on a global scale in a cost and time effective manner. Traditionally, reef health has been estimated using expensive and tedious underwater survey techniques that by definition cannot cover large areas. Remote sensing by satellite offers the potential to survey coral reef ecosystem health on a geographic scale not previously possible. This becomes even more important when one considers the remoteness of most reefs and the expense of expeditionary travel. However, it is not a simple task because coral reef environments are optically, spatially, and temporally complex. To extract meaningful information from satellite imagery, techniques must be developed to relate the electronic signals received by a spacecraft to the optical properties of the reef community and its associated biological processes.

Specifically, the myriad of beautiful colors on a coral reef are a mixture of the optical properties of plant and animal pigments, including the symbiotic zooxanthellae of corals, substrate characteristics, and the overlying water column (Dustan et al, 2000). Individual colors blend together with increasing scale, generating a larger scale collage that can be identified by spacecraft imagery. This signal becomes degraded as it passes through the atmosphere and water due to the wavelength-specific selective absorption of light. This degradation can be partially mitigated if the properties of the air and water column are known, and accurate depth measurements can be correlated to the precise geographic coordinates of a geo-registered satellite image.

It is important to note that remote sensing in tropical environments is further complicated by the high probability of cloud cover, and clear scenes can be difficult to obtain. For this reason, multiple orbital passes by a satellite dedicated to the study of coral reefs are critical and an imperative for obtaining the data necessary to map and monitor them over time.

Time-series analysis ñ comparing multiple images of a single reef over time ñ shows the variability of the upwelling signal of the reef. While a single image provides only a snapshot of the reef, a time series analysis can provide the data needed to validate the mapping of the reef and its communities, track changes in coral community structure and health and make predictions about the future health and composition of the reef. Such an analysis can distinguish daily and seasonal variability from larger community scale ecological degradation. In order to accurately interpret the images, we need to understand the effects of both natural and human-induced change on the upwelling signal from a reef.

Presently, the most complete coral reef time series exists for the Florida Keys. Using twenty-two Landsat images of the northern Keys from 1982 to 1996, a team led by Dr. Dustan has performed an analysis of pixel-scale variation through time, termed temporal texture. Using both Landsat satellite images and in situ observations, this team has shown that the process of reef degradation has altered both the spatial patterning and variability of pixel brightness, which can be identified in unclassified Landsat imagery (Dustan et al, 2002).

References:

Dustan, P., S. Chakrabarti, and A. Alling. 2000. Mapping and Monitoring the Health and Vitality of Coral Reefs from Satellite: A Biospheric Approach. Life Support and Biosphere Science, Vol 7: 149-159.

 Dustan, P., E. Dobson, and G. Nelson. 2002.  Remote Sensing of Coral Reefs: Detection of Shifts in Community Composition of Coral Reefs Using the Landsat Thematic Mapper. (In revision, Conservation Biology)

Lubin, D., W. Li, P. Dustan, C. H. Mazel, and K. Stamnes. 2001. Spectral Signatures of Coral Reefs: Features from Space. Remote Sensing of Environment, 75, 127-137.