Modelling Coral Bleaching Events Using a Fuzzy Logic Technique

Abstract

A fuzzy logic algorithm was used to model coral bleaching events around Magnetic Island, North Queensland, Australia. Most events during the 80’s and 90’s did not coincide with the strong ENSO (El-Niño Southern Oscillation) phenomenon. The model uses monthly Air Temperature (AT) data obtained from the Townsville Bureau of Meteorology and monthly IGOSS-NMC and Geoffrey Bay Sea Surface Temperature (SST) data sets. Four separate model inputs were considered. They were: (a) raw temperature, (b) a temperature for a particular year is taken by subtracting the value of one year before from the value of that year, (c) a temperature for a particular year is taken by subtracting the 2-year running average from the value of that year, and (d) a temperature for a particular year is taken by subtracting the 3-year running average from the value of that year. Using the difference in seawater temperature with that of the previous year as a model input (case b) gave a better fit to bleaching events than using seawater temperature alone. This indicates that some form of acclimatisation may influence the response of corals to some form of bleaching event. It was also found that when the maximum sea temperature in a particular year exceeds 0.37 °C of the previous one year, coral bleaching occurs.KeywordsFuzzy logicCoral bleachingENSOModelAdaptation