Dataset: Coral bleaching and mortality data from a field experiment manipulating herbivore pressure and nutrient availability on a coral reef in Moorea, French Polynesia from 2018 to 2022

Experimental Setup

In June 2018, we established a factorial experiment manipulating consumer pressure and nutrient availability in areas that mimicked disturbance from a cyclone. At 12m depth on the north shore forereef of Moorea, French Polynesia (17.47° S, 149.82° W) we established 8 ~30m² plots. Within each plot, we then nested 4 different consumer exclosures (~1.25 m² each) with different size openings that allowed different size fishes (herbivores and corallivores, hereafter referred to as ‘consumers’) access to the benthos. The exclosure frames consisted of 0.5cm stainless steel all-thread drilled into the reef matrix and epoxied into place. These frames were then wrapped with plastic-coated, galvanized wire to create the following levels of consumer pressure: 1) Very Low (2.5cm × 2.5cm (1inx1in) openings); 2) Low (5cm × 5cm (2inx2in) openings); 3) Medium (7.5cm × 7.5cm (3inx3in) openings); 4) High (4 sides of 2.5cm × 2.5cm (1inx1in) openings but no top). Others have used a similar design to create a gradient of consumer pressure to mimic the effects of different levels of fishing (Holbrook et al. 2016, Schmitt et al. 2019). We included sides, but not tops, on the High consumer pressure treatment to control for potential artifacts on water flow, although we have shown these are minimal (Zaneveld et al. 2016). Exclosures were scrubbed every 12-16 weeks to remove fouling organisms. 

We also included a second treatment of Ambient or Enriched nutrient conditions with each plot (which included four nested consumer exclosures) was then assigned to either Ambient or Enriched nutrient conditions.  Thus, each combination of consumer pressure (Very Low, Low, Medium, High) and nutrients (Ambient or Enriched) had n=4 for replication. For the enrichment we placed 175 g of Osmocote^® (19-6-12, N-P-K) slow-release garden fertilizer into 5 cm diameter PVC tubes with 10, 1 cm holes drilled into them. These tubes were wrapped in fine plastic mesh to retain the fertilizer. This method is similar to our previous work (e.g., Zaneveld et al. 2016). PVC enrichment tubes were attached to the corners of each exclosure and onto a piece of stainless steel all-thread in the center of each plot (5 enrichment tubes per exclosure). We replaced enrichment tubes every 12-16 weeks except for two periods during the COVID-19 pandemic when travel to Moorea was not possible and enrichment tubes were deployed for longer than usual before replacement (deployed from 01-30-2020 to 08-31-2020, and from 08-31-2020 to 02-08-2021). We analyzed water samples from the experimental plots to evaluate the effect of nutrient enrichment. We showed that the enrichment treatment effectively enriched nutrients throughout the experiment, but that the effect of the enrichment decreased with time since replacement of enrichment tubes. This suggests that the enrichment may have been negligible by the end of each deployment of the enrichment tubes, and during the two periods during the pandemic when tubes were deployed longer than usual. 

The treatments were arranged as in this example:

• Plot A1 was Ambient nutrients and contained one replicate of each of the exclosure treatments (Very Low, Low, Medium, and High consumer pressure).
• Plot A2 was Enriched nutrients and contained one replicate of each of the exclosure treatments (Very Low, Low, Medium, and High consumer pressure).

n=4 for each nutrient-consumer pressure combination e.g., - Enriched/High consumer pressure; Ambient/Very Low consumer pressure etc. 

Data Collection

During three periods each year from 2018-2022 (April, August, November - except for two time points missed due to the COVID-19 pandemic), we quantified benthic cover in each exclosure via point contacts on orthorectified photomosaics. Using Olympus TG cameras, we took ~64 individual photographs of each exclosure which were then stitched together using Agisoft Metashape software. The creation of these photomosaics generates high-resolution imagery that allows identification of benthic space holders with high taxonomic resolution. At the first time point in our time series (Aug 2018), we gave each individual Pocillopora and Acropora a unique identifying number. During the peak of coral bleaching, in May 2019, we took additional top-down photographs of each exclosure, allowing us to quantify whether each individual coral had bleached and, if so, what proportion of the colony surface area was bleached or dead.

Working in ImageJ, investigators estimated the surface area of each individual coral colony to be bleached or dead to the nearest 5%. Because corals undergo natural, seasonal variation in Symbiodiniaceae density that can affect their coloration, we defined bleached tissue only as tissue that had lost all pigmentation (i.e., completely white). Bleaching severity was bimodal, with most corals either 0 or 100 percent bleached. However, approximately 36% of corals with some bleaching had less than 100% of their tissue bleached.  For analyses, we defined 50% bleached as a severe bleaching threshold. Using our photomosaics, we then tracked the annual survival or mortality of each coral in August of each year (Aug 2019, Aug 2020, Aug 2021, Aug 2022) by visually estimating the mortality of each individual coral colony to the nearest 5% as we had done for the initial bleaching and mortality quantification. Corals frequently had partial mortality, with colonies sometimes having small amounts of remnant tissue which would remain for many months. Therefore, we used a threshold of the loss of 90% of live tissue to define a coral as being dead since we never observed recovery from this level of tissue loss.