12.2.5 Limitations
The current limitations and concerns when using satellite-derived SST and modelderived products stem from a number of sources:
• First, and possibly most importantly, the accuracy of the output is entirely
dependent on how well the ecology and physical processes of reefs are understood. Invariably, the reality is that we have at best only a limited understanding of
these processes. Testing and validation procedures, which are a vital component
of any model or product, should ensure that model outputs correlate strongly with
observed patterns. However, correlation, no matter how strong, does not imply
causation (Aldrich 1995). Interpretation of results, therefore, needs to be cautious
with respect to the science, and tailored with respect to the application.
• The complexity of most models is often well beyond the understanding of the
non-specialist; therefore applying products and interpreting outputs in the
management environment must be done with specialists involved at every step.
• Most modeling applications encompass a number of diverse disciplines, each
requiring their own specialist skills to ensure that the source data and the output
products are appropriately used and applied. The key to successful modeling studies
is close collaboration with experts in appropriate fields of science and management.
• Satellite thermal radiometers inherently map the surface properties of the water,
not that of the subsurface, which is of predominant interest in most science and
management applications. While considerable work has gone into calibrating
thermal SST data to ‘match’ the bulk water temperature, bias errors do occur,
especially during periods of extreme stratification of the water column and high
atmospheric water vapor content. These conditions inevitably occur in the lead-up
to coral bleaching events when accuracy is most important. Caution in interpretation and prudent validation of bleaching nowcasts are therefore in order.
• The spatial and temporal mismatch between different thermal products can lead
to significant problems in the application and interpretation of data. A common
source of error in the literature is the application of relatively coarse satellite data
to ecological data and processes at specific locations. Spatial heterogeneity on the
micro-scale may not be detected using coarse spatial resolution, and conversely,
patterns on the macro-scale may not be detected using fine spatial resolution.
While the application of remotely-sensed thermal data has greatly advanced our
understanding of coral reef ecosystems, caution should be applied in selecting the
optimal resolution for the application and in the interpretation of results.
12.3 Example Thermal Applications
While the direct application of remotely sensed temperature products has been
highly beneficial to reef management and reef studies in general, the largest
potential gain is the integration of these empirical data with other datasets,
12 Thermal Applications
327
The current limitations and concerns when using satellite-derived SST and modelderived products stem from a number of sources:
• First, and possibly most importantly, the accuracy of the output is entirely
dependent on how well the ecology and physical processes of reefs are understood. Invariably, the reality is that we have at best only a limited understanding of
these processes. Testing and validation procedures, which are a vital component
of any model or product, should ensure that model outputs correlate strongly with
observed patterns. However, correlation, no matter how strong, does not imply
causation (Aldrich 1995). Interpretation of results, therefore, needs to be cautious
with respect to the science, and tailored with respect to the application.
• The complexity of most models is often well beyond the understanding of the
non-specialist; therefore applying products and interpreting outputs in the
management environment must be done with specialists involved at every step.
• Most modeling applications encompass a number of diverse disciplines, each
requiring their own specialist skills to ensure that the source data and the output
products are appropriately used and applied. The key to successful modeling studies
is close collaboration with experts in appropriate fields of science and management.
• Satellite thermal radiometers inherently map the surface properties of the water,
not that of the subsurface, which is of predominant interest in most science and
management applications. While considerable work has gone into calibrating
thermal SST data to ‘match’ the bulk water temperature, bias errors do occur,
especially during periods of extreme stratification of the water column and high
atmospheric water vapor content. These conditions inevitably occur in the lead-up
to coral bleaching events when accuracy is most important. Caution in interpretation and prudent validation of bleaching nowcasts are therefore in order.
• The spatial and temporal mismatch between different thermal products can lead
to significant problems in the application and interpretation of data. A common
source of error in the literature is the application of relatively coarse satellite data
to ecological data and processes at specific locations. Spatial heterogeneity on the
micro-scale may not be detected using coarse spatial resolution, and conversely,
patterns on the macro-scale may not be detected using fine spatial resolution.
While the application of remotely-sensed thermal data has greatly advanced our
understanding of coral reef ecosystems, caution should be applied in selecting the
optimal resolution for the application and in the interpretation of results.
12.3 Example Thermal Applications
While the direct application of remotely sensed temperature products has been
highly beneficial to reef management and reef studies in general, the largest
potential gain is the integration of these empirical data with other datasets,
12 Thermal Applications
327