Ocean temperature – long-range forecasts
How to use these forecasts
Seasonal forecasts provide insight into conditions several months into the future. We provide a range of forecast maps. The ones most useful to you will depend on your needs. These forecasts can help guide decisions such as:
- Where do we survey this summer in the marine park?
- Will our fish stocks be further south this year?
- Do we need extra staff to manage fish farm operations this summer?
- Should we harvest from our aquaculture business to avoid a heatwave?
- Are special conservation actions needed?
Ocean temperature is an input to many decisions. Often the difference from what is typical for a particular place and time of year is even more important. We call the differences in temperature "anomalies". Marine heatwave categories highlight prolonged warm anomaly events.
The forecast maps come from the Bureau's long-range climate forecast system ACCESS-S2. Our forecast system works best for the open ocean (~25 km) and doesn't capture fine details right at the coast. The maps reflect temperatures across the top 1-metre of the water column.
We update the forecasts 3 times a week.
Typical sea surface temperature patterns
The ocean is always much warmer near the tropics and much colder towards Antarctica. On top of that, temperature patterns vary quite a lot in both space and time. Each year, the typical sea surface temperature goes up and down. The month when the highest temperatures occur depends on where you are. Around Australia, the warmest month could be anywhere between November and April.
Changes from typical temperatures can be caused by winds, clouds, ocean currents and much more. Our numerical models take many of these into account and help us make long-range forecasts.
The image below shows typical sea surface temperatures for each calendar month. This type of information is called a "climatology". Climatology uses a baseline period and data source. The baseline period is important because it provides a reference point. This can help when comparing information from different sources.
For example, the monthly climatology below uses a baseline period of 1981–2018. The observational data source is a particular collection of satellite observations. You can search for more details on that dataset with "Reynolds OISSTv2.1 (2022)".
Understanding each map
Sea Surface Temperature Anomaly (SSTA)
These maps show if ocean temperatures are forecast to be cooler or warmer than what is typical.
Warmer-than-normal temperatures are positive anomalies. The size of an anomaly depends on the baseline period. That is why we state the baseline period on each map. Checking the period is important if you are comparing different SSTA sources.
These maps reduce the influence of bias by using an SST climatology from the model itself. That means each month ahead in the SSTA forecasts uses its own reference climatology.
Sea Surface Temperature (SST)
These maps show ocean temperature forecasts across the months ahead.
Each temperature value is a forecast for the top 1 meter of the open ocean. Remember that our climate system is best for patterns larger than about 25 km. Temperatures in very shallow water and small bays may be quite different.
Marine Heatwave Category
These maps show if any marine heatwave events are forecast for your region.
The category indicates if any events are forecast and how extreme they could be. Marine heatwave events are periods of especially high temperatures for a particular place. The technical definition requires a period of at least 5 days with the high temperature. What counts as "high" depends on the climatology for that time of year (top 10% across the baseline period).
Probability of Marine Heatwave
These maps show how likely a marine heatwave event could be in the months ahead.
The maps show likelihood as probabilities, also known as a probabilistic forecast. Probabilistic forecasts are calculated by running the climate model many times. This means making many viable forecasts for each place using an 'ensemble' method. If more ensemble members at any location show a marine heatwave, then the chance of a real event is more likely.
Hotspots
These maps show especially warm temperatures relative to annual maximums.
The measurement used is called a 'hotspot' in line with other international services. The maps are comparable to the information provided by the USA service Coral Reef Watch. The technical details behind hotspots depends on a particular climatology. In particular, the reference is the mean temperature for the warmest month. The exact value we use for a reference is the Maximum Monthly Mean (MMM) using a baseline period of 1985 to 2006.
The warmest month is different for each place around the Australian coast.
Degree Heating Months (DHM)
These maps show a particular pattern of prolonged heat important to coral reefs.
The maps highlight 3-month periods of prolonged warm temperatures. This measurement is particularly relevant to thermal stress for coral. The Degree Heating Month values come from the 'hotspot' forecasts. Each map adds together the hotspot values for each 3-month period across the forecast.
Probability of DHM > 1
These maps show if a high Degree Heating Month (DHM) value is likely to occur across the forecast. The maps show the likelihood of a DHM value greater than 1.
This measurement is particularly relevant to thermal stress planning for coral reefs. The probability comes from the same 'ensemble' method described above for marine heatwaves. If more ensemble members at any place shows a DHM event, then the probability of a real event is higher.
SST Forecast Skill Maps
These maps show how well past forecasts have matched real ocean conditions for a specific place and time. This measure is called forecast skill.
The maps show statistical values that summarise this historical skill. Our statistics come from comparing many forecasts ("hindcasts") against actual ocean observations.
While weather forecasts can tell you how much rain to expect tomorrow, long-range forecasts cannot be this specific. This is because the further we look ahead the more likely it is that small variations will grow and shift the overall weather pattern. Skill also depends on location and time of year. The complex patterns in our maps reflect the complex influences in ocean temperature.
The maps show two complementary measurements of skill. A higher Pearson Correlation Coefficient (R) indicates better skill for time patterns. That is, a higher R value indicates skill at forecasting the ups and downs in temperature. A lower Root Mean Squared Error (RMSE) indicates better skill for temperatures overall. The best skill combination is a high 'R" and a low 'RMSE'.
Product Code: IDCK000071
The operational marine heatwave forecast capability is supported by funding from CSIRO, the Fisheries Research and Development Corporation and the Australian Department of Climate Change, Energy, the Environment and Water.
