[Coral-List] Identifying bleaching resistant reef location: Should we have expected the far north (remote) GBR to have been so vulnerable to coral bleaching?

[James Hendee]

Hi, Scott,

    In your post, below, you say, "Such coherence between predicted and
observed responses is far and away the best test of the reliability of a
model – independent of the choice of predictive error score and the like.”

    I'm assuming when you say "predictive error score and the like" you
are referring specifically or generally to the post of Ruben van Hooidonk,

    http://coral.aoml.noaa.gov/pipermail/coral-list/2016-April/016815.html

    I'm wondering why you dismiss his critique.  What specifically is
wrong or misleading with what he says?  Also, I don't think he used the
term "predictive error score," a term which I'm not familiar with, so
I'm wondering if you have a citation or clarification for that term (I'm
guessing "and the like" refers to van Hooidonk's discussion of skill
score accuracy).

    Thanks!

    Cheers,
    Jim

On 4/10/16 10:14 PM, Scott Wooldridge wrote:
> Hi All,
>
>
> I have received numerous emails asking whether the ‘BleachRisk’ model
> (attached papers below) correctly identified the remote far northern reefs
> of the GBR as sensitive to thermal stress – given the extensive bleaching
> that has occurred this 2016 summer. The answer is YES – most definitely.
>
>
> https://www.researchgate.net/publication/299534164_Excess_seawater_nutrients_enlarged_algal_symbiont_densities_and_bleaching_sensitive_reef_locations_1_Identifying_thresholds_of_concern_for_the_Great_Barrier_Reef_Australia
>
>
>
> https://www.researchgate.net/publication/299939163_Excess_seawater_nutrients_enlarged_algal_symbiont_densities_and_bleaching_sensitive_reef_locations_2_A_regional-scale_predictive_model_for_the_Great_Barrier_Reef_Australia
>
>
>
> I am in the process of writing a short report that details the
> environmental reasons why, and I will make this report available to
> CoralList in the next day or so. In the interim it is worthwhile to think
> on (and rationalise against current opinions) the physical reef attributes
> that the BleachRisk framework predicts to be indicative of reefs with ‘low
> bleaching resistance’ (i.e. high bleaching sensitivity).
>
>
> Quoting directly from the manuscript, ‘Statistically significant linear
> regression relationships linked ‘low’ *bleaching resistance* (i.e., high
> bleaching sensitivity) reef areas with *Porites *colonies that have reduced
> tissue layer thickness (Fig. 8a) and multi-decade sclerochronological
> histories characterised by reduced skeletal density (Fig. 8b), and enhanced
> skeletal extension rates (Fig. 8c).’
>
>
> Therefore contrary to most people’s opinion, the BleachRisk model predicts
> that a bleaching sensitive reef will have fast-growing corals (based on
> regional comparisons). That is, fast skeletal growth at ambient/optimal
> temperatures is actually a very good proxy for bleaching sensitivity under
> thermal stress.
>
>
>
> I have recently written a ‘Think again paper’ on this issue, which explains
> why a fast growing reef with high coral cover may be judged healthy under
> present conditions, whilst at the same time being extremely vulnerable to
> future thermal stress levels.
>
>
>
> https://www.researchgate.net/publication/265297066_Assessing_coral_health_and_resilience_in_a_warming_ocean_Why_looks_can_be_deceptive
>
>
> Indeed, this explanation fits perfectly with the situation in the far
> northern GBR, which for example, has Porites colonies with the highest
> annual skeletal extension rates recorded across the entire GBR (Lough and
> Barnes, 2000).  To highlight this fact, I have redrawn Fig. 8 to include
> the skeletal growth attributes of Porites in the Far North.
>
>
> https://www.researchgate.net/publication/300890755_Bleaching_sensitivity_and_skeletal_growth_characteristics_of_far_northern_GBR
>
>
>
> This identified spatial linkage between an ‘inferred’ model response and an
> independent ‘observable’ biological responses (not included within the
> model calibration process), strengthens our confidence in the model,
> including the suggestion that bleaching resistant (versus bleaching
> sensitive) reefs are readily identifiable and explainable by characteristic
> coral health signatures that exist outside of thermal stress conditions.
>
>
>
> Such coherence between predicted and observed responses is far and away the
> best test of the reliability of a model – independent of the choice of
> predictive error score and the like.
>
>
>
> Scott Wooldridge
>
>
> Catchment to Reef Management Solutions, Newcastle, Australia (2280)
>
>
> https://www.researchgate.net/profile/Scott_Wooldridge
> _______________________________________________
> Coral-List mailing list
> Coral-List at coral.aoml.noaa.gov
> http://coral.aoml.noaa.gov/mailman/listinfo/coral-list