Terpios hoshinota, Rutzler and Muzik, 1993

Yomogida, Masashi, Mizuyama, Masaru, Kubomura, Toshiki & Reimer, James Davis, 2017, Disappearance and Return of an Outbreak of the Coral-killing Cyanobacteriosponge in Southern Japan., Zoological Studies 56 (7), pp. 1-10 : 2-4

Survey of Terpios hoshinota

We utilized the line-intercept transect ( LIT) method to investigate benthos. Surveys were carried out ten times (March, August, September, October, December in 2010; June, October in 2011; February and May in 2012; and September 2014). Transects were 10 m length and 1 m in width ( Beenaerts and Berghe 2005). Five transects were conducted for each survey, except for in September 2014 when ten transects were conducted. The surveys from August 2010 to May 2012 were permanent transects (marked with pegs nailed into the carbonate substrate), with four placed to perpendicular to shore, and one parallel, and all at 2-3 m depths. However, initial transects from March 2010 were not permanent but placed within the Yakomo survey area, while in September 2014 the pegs marking transects had been lost, likely from typhoons, and again we performed non-permanent transects within the survey area ( Fig. 1 ).

All transects (permanent and non-permanent) were separated by 10 to 30 m intervals from neighboring transects on the reef, within an area of approximately 150 × 100 m ( Fig. 1 ). Digital camera images for each transect were taken. Each transect consisted of 25-40 images. We used Adobe Illustrator CS 4 to create a continuous connected transect image for each transect that was utilized in subsequent analyses.

Benthic survey image analyses

Benthos captured in survey images were divided into the following categories:

1) T. hoshinota , forming a thin coating on coral or hard sediment, black or light gray in color;

2) Macroalgae, including both red and green algae, not including Ceratodictyon spongiosum (see #3 below);

3) Sponge algae ( C. spongiosum ), easily distinguishable in the field by dark green or black branching form;

4) Cyanobacteria, visible brown mat-like colonies on the seafloor;

5) Hermatypic corals, including scleractinians and the hydrozoan genus Millepora ;

6) Dead coral, completely white or bleached coral, may be covered in algae;

7) Other benthos, including giant clams, zoantharians, soft corals, and all other organisms not covered by categories 1 to 6 above, including mobile sea cucumbers and sea urchins;

8) Sand/carbonate, generally sand, unconsolidated small rubble, or hard carbonate;

9) Unknown, portions of images hard to see due to shading or lack of focus.

The above benthic categories were measured to the nearest mm. Values of each benthic category were then displayed as a percentage the total length (1000 cm). In cases when length of categories could not be confirmed due to twisting of the tape measure or shading, we used ImageJ (National Institutes of Health, USA) to obtain measurements.

Extent of T. hoshinota outbreak

To determine the extent of T. hoshinota on the Yakomo coast, we conducted a simple snorkel survey in September 2014. Teams of snorkelers swam in both directions (roughly southeast and northwest; Fig. 1 ) along the coast from the entry point at the survey area at Yakomo. The survey stopped when no T. hoshinota was noticed for 100 m since the previous last sighting of T. hoshinota . GPS coordinates were taken at the west and east ends of the T. hoshinota outbreak, and we then prepared a simple map based on coordinates.

Typhoons

Typhoon data (windspeed, wave height) were obtained from the Japan Meteorological Agency homepage (http://www.jma.go.jp/jma/index.html). As there was no locally available wave data for Okinoerabu-jima Island, we utilized data from the closest available point off Naha, Okinawajima Island, Okinawa Prefecture (26°12'44.4"N, 127°40'44.8"E; Ministry of Land, Infrastructure and Transport Port Authority, National Harbor Ocean Wave Information Network (Naufasu)).

Statistical analyses

The percentage data of transects were analyzed in R version 3.3.2 ( R Core Team 2016) with the Bray-Curtis similarity index calculated using the percentage data. Subsequently, we also performed a cluster analysis using the Bray-Curtis similarity index. Non-parametric multi-dimensional scaling ( NMDS) was utilized for examining the similarity of each survey’s data utilizing all nine benthic categories.

Figures

Fig. 1. (A) Location of Okinaerabu-jima Island, Kagoshima, Japan in the northwestern Pacific and, (B) map of Yakomo coast on Okinoerabu-jima Island. Red dotted box shows the Terpios hoshinota survey area, white dotted lines show the approximate area of T. hoshinota along the coast, and red solid lines approximate locations of permanent transects. Images from Google Earth.