Ocean Temperature Outlooks | Coral Bleaching Risk

These pages provide seasonal sea surface temperature forecasts for regions around Australia. The forecasts provide information about likely ocean conditions. They can be used for various purposes including to monitor coral bleaching risk in the Great Barrier Reef and other reefs around Australia. See summaries on the 2016 and 2017 marine heatwaves on the Reef.

Product information

Sea surface temperature (SST) anomalies are calculated by subtracting model SST climatology from model SST values. The climatology is the monthly mean SST over the period 1990–2012, computed relative to the start month and lead time for the model. Removing the climatological SST from SST forecast values reduces the effects of any model bias.

Sea surface temperature (SST) is calculated by adding a monthly SST climatology, derived from Reynolds OISSTv2 (Reynolds et al. 2002) for 1990–2012, to the model SST anomalies. This corrects for model drift with lead time.

Hotspots indicate sea surface temperatures that exceed the mean temperatures of the warmest summer month on average . They are calculated as the positive differences between monthly SST values and the Maximum Monthly Mean (MMM) climatology. The MMM is defined as the average of the hottest month of each year from 1985 to 2012, and de-trended back to November 1988, and is provided by NOAA Coral Reef Watch (Heron et al. 2014).

Degree Heating Months (DHM) gives an indication of both the magnitude and duration of thermal stress. They are defined as the summation of Hotspots over a three month window. For the model this translates to the summation of three consecutive lead times for a forecast start date.

Probability of exceedance maps show the proportion of the 99 ensemble members that exceed the defined SST anomaly (0.6 ℃, 1 ℃) and DHM thresholds (1).

Index forecasts are calculated by taking the areal average of monthly model SST or SST anomaly values (as defined above) in a certain region and are plotted against lead time. All 99 ensemble forecast members are plotted together with the ensemble mean (average of 99 forecast ensembles) to indicate the forecast spread.

Lead time is defined as the time elapsed between the model run date and the forecast date. For example if the model run date is 1 December 2018, lead times of 0, 1 and 2 months correspond to the forecasts for December 2018, January 2019 and February 2019 respectively.

Forecast accuracy

Forecast accuracy is assessed to provide a measure of the skill of the climate forecast system ACCESS-S. Retrospective forecasts (hindcasts) of SST anomalies are compared to observed SST anomalies for the period 1990–2012.

The Pearson correlation coefficient (R) and root mean square error (RMSE) are calculated between the model ensemble mean and observed SST anomalies, and are available as maps on the spatial forecast page.

Monthly climatologies

Climatology is defined as the long term monthly mean SST for each month over a number of years. The climatology figure below shows the monthly climatology calculated at each pixel for each month of 1990–2012 using Reynolds OISSTv2 (Reynolds et al. 2002).



Further reading