Coral Reef Calcification and Production After the 2016 Bleaching Event at Lizard Island, Great Barrier Reef

Severe coral bleaching events have affected the GBR causing massive losses of hard coral cover. Here we use flow respirometry approaches to assess coral reef net ecosystem calcification (NEC) and net ecosystem production following the 2015/2016 bleaching event at Lizard Island in the northern Great Barrier Reef, a heavily impacted area. Previous studies conducted in 2008 and 2009 (Silverman et al., 2014, http://10.1016/j.gca.2014.09.011) were used as preimpact data. Lagrangian and Eulerian approaches provided varied results. Estimated NEC (29.1 to 137.7 mmol m−2 day−1) and NEP (−876.7 to 50.5 mmol m−2 day−1) rates in 2016 were highly sensitive to assumptions about reef water residence times and oceanic end‐member concentrations. Replicating the methodology used for the 2008 and 2009 study resulted in postbleaching NEC in 2016 of 32 ± 10.8 mmol m−2 day−1, 40%–46% lower than prebleaching estimates in 2008 (61 ± 12 mmol m−2 day−1) and 2009 (54 ± 13 mmol m−2 day−1). The slopes of the total alkalinity versus dissolved inorganic carbon plot decreased from ~ 0.3 in 2008 and 2009 to 0.1 in 2016, indicating elevated organic production and a shift in community function. Changes in NEC relative to the previous study were not driven by changing Ωarag. Coral cover shifted from 8.3% and 7.1% in 2008 and 2009 to 3.0% in 2016. We demonstrate a clear decrease in coral reef NEC following bleaching and highlight that subtle assumptions/methodological differences may create bias in the interpretation of results. Therefore, comparing coral reef metabolism data sets and predicting long‐term coral reef calcification based on existing short‐term data sets needs to be done with care. Key Points Net ecosystem calcification varied from 32 to 138 mmol m‐1 d‐1 using different flow respirometry approaches and end‐members Using assumptions that match prebleaching estimates, bleaching reduced NEC by 40%‐46% off Lizard Island Subtle assumptions/methodological differences may create bias in the interpretation of results