Table of Contents

• Coverage
• Dataset Description
  • Methods & Sampling
  • Data Processing Description
  • BCO-DMO Processing Description
• Data Files
• Supplemental Files
• Related Publications
• Related Datasets
• Parameters
• Project Information
• Funding

Data were collected from 160 data sources including data from established monitoring programs and new data extracted from the literature. To date, we have coral data for 12,266 sites, from 83 countries, from 1977 to 2020. Coral-cover data were extracted from the primary literature using WebPlotDigitizer version 4.6 (Marin, 2017). Sampling points that fell on land or were > 1 km from any coral reef were removed. If sites were not named or given explicit coordinates, the coordinates were estimated and a comment was added to the data table. The coordinates were entered into Google Earth and the location names, distance to land in meters, and exposure were determined and recorded for each site. Exposure to waves was based on a site’s potential exposure to predominant winds, swell, and fetch (i.e., the extent of open ocean). Mean turbidity (Kd490) was added for each site (Sully and Woesik, (2020)). The Marine Ecoregions of the World (MEOW) shapefiles (Spalding et al., 2007) and IUCN’s (International Union for Conservation of Nature) World Database on Protected Areas (2022) were used to determine in which marine realm and protected area each site was located. Veron et al. (2015)'s ecoregions shapefiles were used to determine the ecoregion of each site. Data on the types of reef habitats were extracted from the Allen Coral Atlas (Lyons et al., 2022). The Coral Reef Temperature Anomaly Database (CoRTAD version 6; Saha et al., 2018), which is a collection of sea surface temperature variables, was used to extract temperature metrics for each sampling event (Sully et al., 2019). CoRTAD values were only extracted for a sampling event if sampled data had a clearly defined month and year — where sampling events were missing a date, the 15th day of the month was used. For any data given as a range (i.e., depth or date), the midpoint was taken and a comment was added to the HeatCRD.

See van Woesik and Kratochwill (2024) https://doi.org/10.1038/s41597-024-03221-3 for more information.

WebPlotDigitizer (Marin, 2017; version 4.6) was used to extract coral-cover data from primary literature. Data figures from each publication were uploaded as image files, axes were calibrated based on the chart type, and points added to extract the coral cover values at each location. Those values were then added to the database.

QGIS (QGIS, 2024; version 3.26.3) was used to extract habitat, MPA, ecoregion, and realm. Habitat shapefiles were downloaded from Allen Coral Atlas (Lyons at al. 2022), Ecoregion shapefiles from Veron et al. (2015), Realm shapefiles from Marine Ecoregions of the World (MEOW) (Spalding et al., 2007). A csv with the coordinates of each site was uploaded to QGIS along with the aforementioned shapefiles. I ensured the coordinate reference systems were accurate then used the NNJoin (havatv, 2019; version 3.1.3) plugin to extract the shapefile values that correspond to each site coordinate. The NNJoin plugin gives a distance value for any points that fell outside of a polygon. These values were then added to the database.

R (R Core Team, 2023; version 4.3.0) was used to extract CoRTAD and turbidity. R code (Sully, 2019a) was used to extract the turbidity data for each site as a raster. R code (Sully, 2019b) was used to extract the CoRTAD environmental data for each sampling event. 

-Extract "Sources_tbl" table from published Access database (van Woesik & Kratochwill, 2024) into "sources_without_errors.xlsx".
-Load "Heatwaves and Coral Recovery Database.xlsx" (as primary dataset), "Sources DOIs.csv", and "sources_without_errors.xlsx" into the BCO-DMO system.
-Join "Sources DOIs.csv" and "sources_without_errors.xlsx" on citation field.
-Find and replace special characters and non-standard spaces in all files.
-Find and replace garbled chars in all files.
-Add combined date field to primary dataset in yyyy-mm-dd format for rows with exact date.
-Round numbers to the 100th place in primary dataset.
-Publish final files as "933334_v1_heatwave_coral_recovery_data.csv" and "933334_supplement_heatwave_coral_recovery_data_sources".

NSF Award Abstract:

Coral reefs are the world’s most diverse marine ecosystem that provide invaluable goods and services for millions of people worldwide. Yet, coral reefs are experiencing thermal-stress events worldwide and their communities are changing. While coarse-grained climate models predict that few coral reefs will survive the 3ºC sea-surface temperature rise in the coming century, field studies show localized pockets of coral survival and recovery, even under high temperature conditions. Quantifying recovery from thermal-stress events is central to making accurate predictions of coral reef trajectories into the near future. This study examines the differential rates of coral recovery following thermal-stress events, globally, and determines the extent to which regional and local conditions influence recovery. This research is taking advantage of the recent progress in spatio-temporal analyses. One of the most transformative aspects of this work is determining where coral recovery rates differ from expectations, and how those differences relate to regional and local conditions. The research is of relevance to all persons that live and work near coral reefs. What happens to reef corals has cascading consequences on other reef-associated organisms and influences whether coral reefs can keep up with sea-level rise. This project is increasing scientific capacity by training a post-doctoral scholar and a PhD student in big-data analysis and making these analysis techniques broadly available. High quality and free online tutorials are supporting standards-driven instruction for high school math, science, and computer teachers in R, a programming language and software environment used for statistical computing and graphics. The project is producing large-scale data and computational resources, which are benefitting diverse users such as students, scientists, resource managers and the broader public.

The current rapid rate of climate change threatens coral reefs. Quantifying recovery from thermal-stress events is central to making accurate predictions of coral-reef trajectories into the near future. Coral populations in different geographic regions and under different local conditions vary in their capacity to tolerate or recover from thermal stress. However, how and why coral responses differ remains poorly understood. There is a clear need for accurate predictions of coral trajectories following thermal-stress events and for determining which interacting factors most influence coral recovery. This study is characterizing the relationships between the rates of coral recovery, frequency and intensity of thermal-stress events, geographic location, habitat, and local conditions that slow or enhance coral recovery. Four approaches are being used to analyze coral recovery: (i) a binary approach, (ii) a meta-analysis approach, (iii) an inverse-problem approach, and (iv) a state-space approach. Spatial and temporal differences in rates of coral recovery are being quantified by capitalizing on the latest developments in spatio-temporal analyses within a Bayesian framework. Observed outcomes of coral recovery are being compared with predicted outcomes to identify areas where recoveries are either higher or lower than expected, and to assess context-dependencies of coral recovery in relation to local and regional conditions. The most transformative aspect of the study is the identification of localities with greater than expected recovery rates, which could guide future conservation decisions by enabling managers to target coral reefs with specific characteristics for protection from human disturbances by designating them as potential refuges as the oceans continue to warm.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.