Coral Reef Ecosystems of the Pacific Remote Islands Marine National Monument: a 2000-2016 Overview

This booklet provides an overview of key findings about spatial patterns and temporal trends of the coral reef ecosystems observed during NOAA’s Pacific Reef Assessment and Monitoring Program (Pacific RAMP) research surveys conducted in the U.S. Pacific Remote Islands Areas from 2000 to 2016 by the Coral Reef Ecosystem Program (CREP) of the NOAA Pacific Islands Fisheries Science Center (PIFSC) . This booklet was developed by collaborative efforts among CREP, Bren School of Environmental Science and Management at University of California, Santa Barbara (UCSB) and Pacific Island Regional Office (PIRO).

This booklet and a more extensive monitoring report (in prep) for the Pacific Remote Islands Marine National Monument (PRIMNM) was initially requested by the NOAA PIRO, Marine National Monument Program to inform the development of the fishery management plan for the PRIMNM. The financial support was provided by the NOAA Coral Reef Conservation Program and PIRO.

This document may be referenced in the following manner:

Boyle, S., V. De Anda, K. Koenig, E. O’Reilly, M. Schafer, T. Acoba, A. Dillon, A. Heenan, T. Oliver, D. Swanson, B. Vargas-Ángel, M. Weijerman, I. Williams, L. Wegley Kelly, R. Brainard (2017). Coral reef ecosystems of the Pacific Remote Islands Marine National Monument: a 2000–2016 overview. NOAA Pacific Islands Fisheries Science Center, PIFSC Special Publication, SP-17-003, 62 p. http://doi.org/10.7289/V5/SP-PIFSC-17-003

A multi-metric composite condition index that integrates multiple coral reef metrics encompassing both the biological community and oceanographic and climatological indicators was developed to describe the overall condition or status of the coral reefs within the PRIMNM following an approach developed for a suite of NOAA National Coral Reef Monitoring Program (NCRMP) Coral Reef Condition Report Cards. The Ecosystem Sciences Division (ESD) at the Pacific Islands Fisheries Science Center created a Benthic Condition Index, a Reef Fish Condition Index, and a Climate Condition Index that use various Pacific RAMP data sets collected in recent years (listed in related items section). The overall Coral Reef Condition Index is composed of equally weighted Benthic Condition, Fish Condition, and Climate Condition Indices. ESD conducted several analyses to score metrics used in the Coral Reef Condition Index.

STEP 1: Temperature stress was calculated for the Climate Condition Index.

1. Source degree heating week data from NOAA’s Coral Reef Watch 50 km Virtual Stations from 11/28/2000 to 10/13/2016.

2. For each year in the reporting period, the maximum Degree Heating Week (DHW) value was determined.

3. The maximum DHW values were then tallied in an event frequency chart developed by the Coral Reef Watch Program. The different thresholds for the grading chart were based on Coral Reef Watch’s standard bleaching alert threshold levels. If the tally for a column fell between two cut-off values, values were rounded up.

4. The final temperature stress score was assigned based on the lowest grade received for the entire reporting period.

STEP 2: Aragonite saturation state scores were calculated as a measure of the ocean acidification for the Climate Condition Index.

1. Source aragonite saturation state calculated from data from 2010-2015.

2. The grading scheme developed was based on pCO2, which was used to model past and future climatic conditions. The grading scheme was based on even splits between preindustrial CO2 levels and 2 times this level, at 70µatm. These levels were then converted to aragonite saturation state, which is key variable for ocean acidification.

3. A linear regression was calculated based on these thresholds. The resulting equation was then used to calculate a score for each aragonite saturation state. For aragonite saturation states higher than 3.3, the equation y=30.8x-41.5 was used. For values less than 3.63, the equation y=200x-600 was used. The aragonite saturation state values from 2010-2015 were then entered into these equations to determine a final score.

4. The resulting scores were then averaged together for 2010-2015 to determine a final ocean acidification score for each island.

STEP 3: Reef material growth were scored for the Climate Condition Index.

1. Source reef material growth calculated from calcification accretion units (CAUs) from 2012-2015.

2. To determine grading thresholds, a pacific-wide distribution of reef material growth was log-transformed to make the distribution approximately normal. Even quantiles were then set at 0, 20, 40, 60, 80, 100.

3. A logistic regression equation based on these thresholds was used to calculate a score for each reef material growth value. The logistic equation corresponding with the distribution was y=12.753*ln(x)+117.55. Carbonate accretion rates were then plugged into this equation to determine the final score.

4. The resulting scores were then averaged together for X to X year to determine a final reef material growth score for each island.

STEP 4: The final scores derived from the steps 1-3 were equally weighted and averaged to determine the final Climate Condition Index score.

STEP 5: Benthic cover were scored for the Benthic Condition Index.

1. To score benthic cover, Pacific-wide percent cover data for coral, CCA, and macroalgal cover was divided into 20% quantiles to determine threshold reference points. Reference points are as follows:

Percent Coral Cover: 0, 10, 20, 30, 40, 100

Percent CCA Cover: 0, 2, 5, 10, 20, 100

Percent Macroalgae Cover: 100, 40, 30, 10, 5, 0

2. The weighted mean percent for each island and each functional group obtained from REA StRS data, was then scored on a standard 0-100 scale accordingly based on the reference points

3. A final benthic cover score was calculated for each island as a composite score of the three major components, each weighted equally.

STEP 6: Generic richness were scored for the Benthic Condition Index.

1. The mean weighted generic richness for each island, was divided by the Pacific-wide observed maximum of 30 used to standardize the scoring.

2. Standardized scores were placed into 20% bins and scored accordingly. Reference points for standardized generic richness were 0, 15, 40, 60, 80, 100.

STEP 7: Adult and juvenile coral colony densities were scored for the Benthic Condition Index.

1. To score adult and juvenile density estimates, genera and species were selected for each island based on their abundance, importance, and consistent identification across the islands.

2. Observed maximums for each selected genera and species across islands were used to standardize scores for each taxon. Observed maximum was unique to each taxon.

3. Standardized scores were then assigned into bins, based on 20% quantiles, which were truncated at a low range as a conservative approach, accounting for habitat specific differences in density for adults and juveniles. Reference points were as follows: 0, 10, 20, 40, 60, 100. Scores were then calculated accordingly for each island

STEP 8: Partial mortality of the adult coral was calculated for the Benthic Condition Index.

1. The adult coral partial mortality indicator was calculated as mean ‘old dead’ percent of selected genera and species. Old dead mortality is defined as the non-living portion of a colony where corallite structures are either gone or covered over by organisms that are not easily removed.

2. Partial mortality scoring used the same selection of genera and species as the adult and juvenile density coral population estimates.

3. Ranges in mortality were evaluated across the Pacific to determine a maximum threshold.

4. A maximum of 30% mortality was used to standardize the scores

5. Standardized scores were based placed in 20% quantile bins. Reference points were as follows 100, 80, 60, 40, 20, 0. Scores were then calculated accordingly for each island.

STEP 9: The final scores derived from steps 5-8 were equally weighted and averaged to determine the final Benthic Condition Index score.

STEP 10: Instantaneous fish biomass was scored for the Fish Condition Index.

1. First, for every Pacific island, the proportion of present biomass relative to modeled biomass was calculated (Williams et al., 2015). The first metric of the fish community index was calculated using these data on present day, mean fish biomass and predicted mean fish biomass.

2. Based on the proportions of inhabited islands, arbitrary scores were assigned to the highest and lowest proportions found at these islands. For example, Oahu in the Main Hawaiian Island Archipelago was found to have the lowest proportion of biomass relative to the modeled estimates. The proportion at Oahu, 0.24 (i.e., reef fish biomass at Oahu is 24.6% of the modeled fish biomass estimates determined from Williams et al., 2015) was used to assign the lowest possible score for all islands in the Pacific, a “failing,” “very poor” score, or a 50% grade.

3. From the failing score at Oahu, scoring was then assigned based to assign a passing grade to the inhabited island with the highest score and a “fair” grade to a proportion of 0.33. The inhabited island with the highest proportion, Swains in American Samoa with a proportion of 0.77, was assigned a “good” score.

4. These designations allowed us to determine scoring equations based on a regression for the lower and higher scoring islands. The scoring equations were arbitrarily determined with the intention of having the best populated islands receiving a “fair” or “good” score.

5. Ultimately, if an island had a proportion of modeled biomass of 0.8, the associated score was “good” or a grade of 90%. Anything higher than 0.9 was given a score of “excellent” or a 100% grade. Anything between 0.33 and 0.77 (not inclusive) received a score of “fair” and a grade between 70% and 80% (not inclusive). A “poor” score (between 60% and 70%, not inclusive) was assigned to proportions between 0.33 and the lowest proportion, 0.24. The equations are as follows:

a. Proportion greater than 0.9: Score = 100

b. Proportion between 0.33 and 0.9: Score = 55.7 + (Proportion * 42.8)

c. Proportion less than 0.33: Score = Proportion * 210

STEP 11: Predator biomass was scored for the Fish Condition Index.

1. Scoring for piscivore biomass was done in the same way as instantaneous fish biomass, the first metric included in the fish index that was described in the previous section. The scoring equations were similarly determined from a range of relative piscivore biomass found at inhabited islands in the Pacific. The equations were arbitrarily determined to ensure the lowest scores “failed” and the best inhabited islands received a “fair” or “good” score. This scoring method results in most all remote islands received an “excellent” score (i.e., 100%). The scoring equations are as follows:

d. Proportion greater than 0.9: Score = 100

e. Proportion between 0.1 and 0.9: Score = 60 + ((Proportion - 10) * 0.5)

f. Proportion less than 0.1: Score = 60 * (Proportion/10)

STEP 12: Shark abundance was scored for the Fish Condition Index.

1. The second metric included in the predator average index was shark abundance. Shark abundance was scored similarly to both instantaneous fish biomass and piscivore biomass however, the modeled baseline of shark abundance was obtained from a different study. Nadon et al., (2012) modeled shark abundance, measured as sightings from towed-diver surveys between 2010 and 2015, in the absence of human presence. These island-level estimates of shark density offered a reconstructed baseline of biomass and abundance of key groups which incorporated natural pressures and drivers, as well as anthropogenic forcings that contribute to differences in the fish community across the Pacific (Nadon et al., 2012). With the data on modeled “pristine” (i.e., in the absence of human presence or influence) shark abundance, this metric was scored in the same way as both piscivore and instantaneous fish biomass. The scoring equations are as follows:

g. Proportion greater than 0.6: Score = 100

h. Proportion between 0.1 and 0.6: Score = 60 + ((Proportion - 10) * 0.8)

i. Proportion between 0.02 and 0.1: Score = 50 * ((Proportion -2) * 1.25)

j. Proportion less than 0.02: Score = 50 * (Proportion/2)

2. Finally, the predator average score was determined by averaging the scores for the two predator metrics, piscivore biomass and shark abundance, at each island. This resulted in a specific predator average score for each island.

STEP 13: Mean size of target families was scored for the Fish Condition Index.

Due to the absence of fishing in this monument, we used key families that were present at all islands and represent important ecological functions (Heenan and Williams, 2013). The families included were Acanthuridae, Serranidae, Scaridae, Lethrinidae, and Lutjanidae. Fishes of these families represent important ecological roles on coral reefs. These fishes were present at all island with one exception, at Johnston Atoll Serranids have never been sighted.

To account for differences in fish community as a result of biogeographical and large-scale oceanographic gradients, we first had to determine meaningful regions to group the Pacific islands. Varying environmental conditions such as nutrient availability and temperature significantly influence the naturally occurring community at each island. In particular, highly productive waters of the equatorial upwelling islands tend to have higher biomass of sharks, other piscivores, and planktivores (Williams et al., 2015). Grouping islands by bioregion allowed for standardized comparison across islands that is not confounded by natural variation in oceanographic characteristics.

The islands in the Pacific were divided into three unique bioregions: tropical Western Pacific (Mariana Archipelago and Wake Atoll), central Polynesia (American Samoa, Palmyra Atoll, Baker and Howland Islands, and Kingman Reef), and Hawai‘i (Main and NW archipelagos). Once the biogeographic regions were determined, the island-level, mean size of each family was standardized within each region and subsequently scored. This method ensured that the range of scores (i.e., excellent, good, fair, poor, and very poor) were specific to each region and accounted for the effect of large-scale oceanographic differences.

To score this metric, for all the islands within each of the three regions, the mean and standard deviation was calculated for each fish family.

Second, the standard Z-score was calculated based the equation: Z =(X - µ) / s where X is the island-level mean size of each family, µ is the region-level mean of each family, and s is the standard deviation.

After the Z-scores for each island and family were calculated, the Z-scores were averaged across all families for each island. This resulted in an island-level Z-score that accounts for broad biogeographical differences in the mean size of five key fish families.

Then, the mean Z-scores across all 40 islands included were ranked from highest (largest positive deviation from the region specific mean) to lowest (largest negative deviation from the region specific mean).

Finally, the final score was similarly arbitrarily determined. If the mean Z-score was greater than 1, the island received a score of 100% (i.e., “excellent”). Conversely, a Z-score of -1 was assigned a 60% (i.e., “poor”) score. The linear relationship between the lowest (-1) and highest (1) Z-scores was then calculated to determine the scoring equation. This linear equation: Score =0.2x + 0.8 was subsequently used to score the other islands.

STEP 14: The final scores derived from steps 10-13 were equally weighted and averaged to determine the final Fish Condition Index score.

STEP 15: The overall Coral Reef Condition Index were derived for each island in the Pacific Remote Islands Marine National Monument by averaging the Climate, Benthic and Fish Condition indices.

Pacific Remote Island Areas (PRIA) including Baker, Howland, Jarvis, and Wake islands, Kingman Reef, and Johnston and Palmyra atolls.

Not applicable

Not applicable

Coral Reef Ecosystem Program (CREP) Data Sharing Recommendations, version 9.0 updated August 12, 2015:

CREP welcomes the opportunity to collaborate on research issues contributing to the scientific basis for better management of marine ecosystems. CREP has a very diverse set of field activities that generates large volumes of data using an array of data collection protocols.

The following recommendations are for your consideration as you use this data:

1) Data analyses should take all field exigencies into account. The most effective way to do this would be active collaboration with CREP principal investigators.

2) In all presentations, product releases, or publications using data generated by CREP, proper acknowledgement of both CREP and the individuals responsible for data collection is expected. Citing the DOI (if available) is preferred, a non-DOI example is listed below.

3) If you collect or generate data for the same study areas, CREP requests that you share relevant information on complimentary data collections.

4) Those receiving data are strongly urged to inform the CREP Data Management Team of any errors and discrepancies that are discovered during the course of using these data. They are further urged to bring to the attention of the Team all problems and difficulties encountered in using these data. This information is necessary in order to improve the collections and to facilitate more efficient and economical data processing and retrieval. The users are asked to supply copies of any missing data that may be located, and to provide information as to significant subsets and special aggregations of data that are developed in using the material provided.

All data sets used in this booklet are archived at the NOAA National Centers for Environmental Information (NCEI) ocean archive. See the list of "Related Items" to link to the documentation for the related datasets. For more in-depth information, consult the scientific papers referenced throughout this booklet.

None

This document may be referenced in the following manner:

Boyle, S., V. De Anda, K. Koenig, E. O’Reilly, M. Schafer, T. Acoba, A. Dillon, A. Heenan, T. Oliver, D. Swanson, B. Vargas-Ángel, M. Weijerman, I. Williams, L. Wegley Kelly, R. Brainard (2017). Coral reef ecosystems of the Pacific Remote Islands Marine National Monument: a 2000–2016 overview. NOAA Pacific Islands Fisheries Science Center, PIFSC Special Publication, SP-17-003, 62 p. http://doi.org/10.7289/V5/SP-PIFSC-17-003

None

None

10 datasets collected by NOAA Coral Reef Ecosystem Program (CREP) for the Pacific Remote Islands and all NOAA monitored regions within the Pacific were used for the creation of the index. An additional dataset was collected by NOAA’s Coral Reef Watch Program through the POES (Polar Operational Environmental Satellites) AVHRR (Advanced Very High Resolution Radiometer) to determine degree heating weeks for the PRIMNM.

Temperature stress data, in the form of degree heating weeks, was compiled from Coral Reef Watch 50km virtual stations from 2000-2016. The aragonite saturation states were calculated based on data collected by CREP from 2010-2015. Reef material growth was determined based on Calcification Accretion Unit (CAU) data collected by CREP from 2012-2015.

Benthic community data was collected through Rapid Ecological Assessment depth-stratified random sampling surveys conducted by CREP from 2014-2015. Fish community data was collected from 2010-2015 by CREP through both stationary point count and towed-diver survey methods. Condition scores were developed based on metric-specific thresholds and were then combined together to create a multi-metric Coral Reef Condition Index.

The coral reef condition index is a calculation based on existing data and can be reproduced

Researchers are trained to identify and measure metrics within the benthic and fish communities. Both classroom and field training and tests must be completed for all researchers before each round of surveys begins. This training ensures that the surveys are conducted consistently by all researchers within survey efforts and also across periodic survey efforts. Oceanographic data is collected according to consistent and systematic methodology including standardized settlement plates as well as remotely sensed satellite data. These methods are standardized across the entire Pacific.

For a majority of the metrics, data was collected on cruise between 2010 and 2015. The target precision is large enough that inter-annual change would likely only be detectable on a decade timescale. Therefore, all survey years are likely comparable. This could easily be addressed by running the analysis separately for the different years.

Due to the remoteness of many of the Pacific Islands and the high cost of surveying these areas, the biological surveys are limited to short time periods within bi- or tri-annual survey years. This may offer gaps in intra-annual variability at each island especially in terms of highly migratory fishes. The oceanographic and climatological data has the highest level of completeness due to the data being collected more systematically.

The accuracy of the results is inherent of the raw data. Please see the metadata for each metric to determine the accuracy of each metric.

The overview includes data from 2000 to 2016.

This Coral Reef Condition Index is the first iteration of the PRIMNM report card. Due to the change in data collection methods, there were no long-term datasets available, except for degree heating weeks. Therefore, this analysis can be made robust through time as more data is collected with the current NOAA methods.

Data used in this report went through quality assurance and control checks upon the data entry into the database. Reported values were reviewed by multiple researchers to ensure the analyses conducted.

To ensure quality of the coral reef condition index, outliers that confound the data are accounted for across datasets. In addition, environmental drivers of biological communities are considered by accounting for regional differences in the biological datasets. Please see metadata for individual metrics for details.