MEPS 331:101-108 (2007)  -  doi:10.3354/meps331101

A robustness metric integrating spatial and temporal information: application to coral reefs exposed to local and regional disturbances

Andrew R. G. Price1,*, Matt J. Keeling1,2, Ian N. Stewart2

1Ecology & Epidemiology Group, Department of Biological Sciences, and 2Mathematics Institute, University of Warwick, Coventry CV4 7AL, UK

ABSTRACT: Capturing complex ecosystem dynamics through aggregate metrics is challenging, especially against a background of climate instability and local disturbances. Robustness, a key determinant of long-term success or failure, is particularly awkward to distil, because it also conveys many different meanings. Hence, metrics tend to be limited or specialised. Herein we develop an index (Q) which captures one aspect of robustness and, unusually, integrates both spatial and temporal information. Q is defined as the probability that the value of a macroscopic system feature remains above a specified fraction of ‘baseline’ condition, at a proportion of spatial localities sampled over time. As a test case we analyse coral cover from a long-term survey in Thailand. Sites were subjected to dredging effects in 1986 an 1987 (Sites A and B), elevated sea temperatures and depressed sea levels especially in 1997 and 1998 (Sites A, B and C) following climate related events in the Indian Ocean. Sites A and C have similar Q values, because they experienced similar levels of overall disturbance, albeit different types of disturbance. We show that this metric can also assess specific (e.g. climate related) disturbances. The cut-off value of Q for declaring a system robust is arbitrary. However, our analyses support recent biodiversity studies that used the same coral data set, which suggest remarkable robustness of these reef flats to adverse environmental conditions. Our metric could potentially also quantify change in composition of process oriented features, such as key functional groups, a better proximal determinant of robustness of reefs to hurricanes and other agents of environmental change than coral cover. Probabilistic measures similar to Q could have even wider application, for example to quantify feedback and self-regulation. In a greenhouse world, this behaviour in ecosystems may become increasingly significant.


KEY WORDS: Robustness · Spatiotemporal information · Global change · Coral reefs


Full text in pdf format