which can be used to predict anomalous SST several months in advance. CRW’s
current Bleaching Outlook products are based on the predictions of two forecast
systems: (1) the statistical Linear Inverse Modeling (LIM) system provides
deterministic weekly forecasts of SST at 2° resolution (Liu et al. 2009) and, (2) the
more-recently employed dynamical Climate Forecast System (CFS) uses a 28member ensemble to provide a range of SST predictions for each location and at
each forecast time-step. The Bleaching Outlook is determined from these models
by incorporating forecasts out to 18 weeks, with the potential risk level for stress
updated each week. The LIM-based outlook indicates the forecast level of stress,
while the CFS-based outlook provides the probability that identified stress
thresholds will be exceeded. Interpretation of the Bleaching Outlook is provided as
necessary on an approximately monthly basis during bleaching risk periods.
The Australian Bureau of Meteorology has implemented operational dynamical
seasonal forecasts of SST anomalies to aid in the management of the GBR Marine
Park using the Predictive Ocean Atmosphere Model for Australia (POAMA;
Spillman and Alves 2009). POAMA is a global coupled ocean–atmosphere
ensemble seasonal forecast system, where the resolution of the ocean model is 2°
in the zonal direction and 0.5–1.5° in the meridional direction. This is the first
application of a dynamical seasonal model for coral bleaching, providing valuable
forecast tools for reef managers at useful time scales (Spillman and Alves 2009).
Real-time forecasts of SST anomaly out to six months are produced daily based on
the ensemble mean of the last 30 daily forecasts. The system also calculates a GBR
index, the areal mean of monthly SST anomalies within the GBR study region, as
an indicator of the average regional conditions. In addition, thermal stress referenced to the warmest month SST climatology and accumulated for the subsequent
3 months of predictions is presented as Degree Heating Months (Spillman et al.
2010). The use of a forecast ensemble allows the assessment of the likelihood of a
particular SST anomaly or thermal stress threshold being exceeded (Spillman and
Alves 2009). The model has shown good predictive skill in reproducing past
monthly SST anomalies and indices with 0–2 months lead-time in the GBR.
However, uncertainty in the onset of the monsoon currently limits accuracy of
summer predictions in the region at longer lead times (Spillman et al. 2010;
Spillman 2011). Although the model has the capability to predict the Madden
Julian Oscillation (MJO), and hence the onset of the monsoon, its effectiveness is
ultimately limited by overall predictability of the system (Rashid et al. 2010).
Future improvements in model resolution, initialization and ensemble creation,
together with improved understanding of the predictability of these primary
drivers, will lead to more skillful forecasts at longer lead-times.
Reef forecasting systems require highly accurate instrumentation to provide
climate quality SST over shallow near-shore waters for long periods (Gramer et al.
2009; Liu et al. 2009). Higher spatial and temporal resolution is needed than the
SST products currently used by NOAA’s CRW in order to effectively model the
coral bleaching response in distinct sub-regions of a reef system (Beggs 2010).
SST observations at 1 km resolution from AVHRR and MODIS sensors are
12 Thermal Applications
325
current Bleaching Outlook products are based on the predictions of two forecast
systems: (1) the statistical Linear Inverse Modeling (LIM) system provides
deterministic weekly forecasts of SST at 2° resolution (Liu et al. 2009) and, (2) the
more-recently employed dynamical Climate Forecast System (CFS) uses a 28member ensemble to provide a range of SST predictions for each location and at
each forecast time-step. The Bleaching Outlook is determined from these models
by incorporating forecasts out to 18 weeks, with the potential risk level for stress
updated each week. The LIM-based outlook indicates the forecast level of stress,
while the CFS-based outlook provides the probability that identified stress
thresholds will be exceeded. Interpretation of the Bleaching Outlook is provided as
necessary on an approximately monthly basis during bleaching risk periods.
The Australian Bureau of Meteorology has implemented operational dynamical
seasonal forecasts of SST anomalies to aid in the management of the GBR Marine
Park using the Predictive Ocean Atmosphere Model for Australia (POAMA;
Spillman and Alves 2009). POAMA is a global coupled ocean–atmosphere
ensemble seasonal forecast system, where the resolution of the ocean model is 2°
in the zonal direction and 0.5–1.5° in the meridional direction. This is the first
application of a dynamical seasonal model for coral bleaching, providing valuable
forecast tools for reef managers at useful time scales (Spillman and Alves 2009).
Real-time forecasts of SST anomaly out to six months are produced daily based on
the ensemble mean of the last 30 daily forecasts. The system also calculates a GBR
index, the areal mean of monthly SST anomalies within the GBR study region, as
an indicator of the average regional conditions. In addition, thermal stress referenced to the warmest month SST climatology and accumulated for the subsequent
3 months of predictions is presented as Degree Heating Months (Spillman et al.
2010). The use of a forecast ensemble allows the assessment of the likelihood of a
particular SST anomaly or thermal stress threshold being exceeded (Spillman and
Alves 2009). The model has shown good predictive skill in reproducing past
monthly SST anomalies and indices with 0–2 months lead-time in the GBR.
However, uncertainty in the onset of the monsoon currently limits accuracy of
summer predictions in the region at longer lead times (Spillman et al. 2010;
Spillman 2011). Although the model has the capability to predict the Madden
Julian Oscillation (MJO), and hence the onset of the monsoon, its effectiveness is
ultimately limited by overall predictability of the system (Rashid et al. 2010).
Future improvements in model resolution, initialization and ensemble creation,
together with improved understanding of the predictability of these primary
drivers, will lead to more skillful forecasts at longer lead-times.
Reef forecasting systems require highly accurate instrumentation to provide
climate quality SST over shallow near-shore waters for long periods (Gramer et al.
2009; Liu et al. 2009). Higher spatial and temporal resolution is needed than the
SST products currently used by NOAA’s CRW in order to effectively model the
coral bleaching response in distinct sub-regions of a reef system (Beggs 2010).
SST observations at 1 km resolution from AVHRR and MODIS sensors are
12 Thermal Applications
325