The following supplements accompany the article

Movements and foraging of predators associated with mesophotic coral reefs and their potential for linking ecological habitats

Yannis P. Papastamatiou*, Carl G. Meyer, Randall K. Kosaki, Natalie J. Wallsgrove, Brian N. Popp

*Corresponding author: [email protected]

Marine Ecology Progress Series 521:155–170 (2015)

Supplement 1 Evaluation of receiver performance Each pulse train emitted by acoustic transmitters consists of 7 pulses with fixed time intervals between the first two pulses (the synchronization interval). The synch time must be within a very specific range for the transmission to be recognized by the VR2W. The synch number represents the number of pulse trains sent out by tags within range of the VR2W. The intervals between pulses 5 and 6 and 6 and 7 are known as checksums and if changed then the transmission is not recorded as a true detection (and is logged as a rejection, Simpfendorfer et al. 2008). A pulse train must have valid synch and checksum intervals to be logged as a detection. For each receiver we calculated the rejection coefficient (RCF=Rejections/Synchs), and the noise quotient (N=Pulses-(7*Synchs)). A positive N suggests environmental noise, while negative values suggest that acoustic collisions are more significant than noise (Simpfendorfer et al. 2008). RCF values were arcsine transformed to conform to normality. RCF values differed between receivers (One-way ANOVA, d.f.=1124, F=20.14, p<0.001) which was due to the Deep receiver (0.03 ± 0.03) and the West Spur and Groove (0.02 ± 0.03) having higher RCF values than the others. However, RCF values were still low with on average only 3 % of detections being rejected due to invalid checksums. The noise quotients also varied between receivers, with more positive values on the deep receiver (One-way ANOVA, d.f.=2238, F=24.0, p<0.001). The noise quotients varied widely, but overall were centered around 0. The deep receiver showed seasonal changes, with evidence of increased acoustic noise in the fall. These results show that receivers performed similarly, with the exception of the deep receiver which suffered from increased acoustic noise during the fall. However, although there were negative spikes in the noise quotient, there did not appear to be evidence of excessive acoustic collisions from multiple transmitters. Note that the receivers with the highest RCF values were also the receivers with the most detections.

References

Simpfendorfer CA, Heupel MR, Collins AB (2008) Variation in the performance of acoustic receivers and its implication for positioning algorithms in a riverine setting. Can J Fish Aquat Sci 65:482−492



Fig. S1 Seasonal changes in the the noise quotient of the deep receiver. Note increased acoustic noise during the fall.

Supplement 2 Percentage of time individual animals were detected in deep and shallow habitats

Animal TL (cm) Sex Deep (%) Shallow (%) Galapagos 15307 168 M 4 36 Galapagos 15315 148 F 3 10 Galapagos 15319 199 M 63 13 Galapagos 61972 136 F 2 22 Trevally 15309 92 - 16 14 Trevally 15311 109 - 33 11 Trevally 15313 89 - 32 6 Trevally 15317 101 - 7 12 Trevally 61957 94 - 55 11 Trevally 61976 115 - 10 13 Trevally 61959 104 - 27 51 Trevally 61960 107 - 4 19 Table S1 Acoustically tagged predators and percentage number of days detected on the MCE (deep, n=1) and shallow (n=6) receivers, out of the 365 d monitoring period.

 Supplement 3 Spatial partitioning between sharks and trevally

Fig. S2 Non-metric multidimensional scaling ordination showing differences in the percentage of time spent at specific receivers by Galapagos sharks (Carcharhinus galapagensis, green triangles) and giant trevally (Caranx ignobilis, blue triangles) at Pearl and Hermes Atoll. Data is over a year monitoring period.

 Supplement 4 Seasonal cyclical movements in giant trevally

Fig. S3 Continuous wavelet transformations for 3 giant trevally (Caranx ignobilis) acoustically tagged on a mesophotic reef at Pearl and Hermes Atoll. X-axis represents number of days since tagging (September 2011). Y-axis is frequency (in days) of periodicity in numbers of detections. The red dashed line indicates a periodicity of 24 h. Areas circled in black are statistically significant. Patterns outside the Cone of Influence should not be considered. Notice the 8-12 day cyclical signal demonstrated by all animals 150 days after tagging (February).

 Supplement 5 Taylor dyadic influence scores of predator movements

Giant trevally

Deep NW SE SW West CH West SP Deep 0.000 -0.002 0.002 0.019 -0.000 -0.031 NW 0.002 0.000 0.000 0.002 0.000 -0.002 SE -0.002 0.000 0.000 0.000 -0.000 0.002 SW -0.019 -0.002 0.000 0.000 -0.004 0.034 West CH 0.000 0.000 0.000 0.004 0.000 -0.002 West SP 0.031 0.002 -0.002 -0.034 0.002 0.000

Galapagos sharks

Deep NW SW West CH West SP Deep 0.000 0.000 -0.002 0.000 -0.025 NW 0.000 0.000 0.000 0.000 0.000 SW 0.002 0.000 0.000 0.000 0.012 West CH 0.000 0.000 0.000 0.000 0.000 West SP 0.025 0.000 -0.012 0.000 0.000 Table S2 Taylor dyadic influence scores for giant trevally (Caranx ignobilis) and Galapagos sharks (Carcharhinus galapagensis) at Pearl and Hermes atoll. Each index is representative of the net exchange of movements between two nodes (acoustic listening stations). Negative values mean there were more movements out of rather than into the receiver. Note more asymmetric movements by giant trevally.

 Supplement 6 Network analysis of movements for giant trevally tagged in shallow reefs

Fig. S4 Network analysis of movements for giant trevally (Caranx ignobilis) tagged in shallow water (n=5) at Pearl and Hermes Atoll. Each node represents an acoustic receiver and the size of the node is proportional to betweeness. Edge thickness is proportional to the number of actual movements between nodes.

Out deg In deg Betweeness Deep 1.961 1.961 0 WestSP - - - SW 10.784 10.784 8.333 NW - - - WestCH 30.882 31.863 16.667 SE 25.980 25.980 8.333 NE 0.980 0 0 Mean 14.118 14.118 6.667 SD 12.274 12.758 6.236 CV 86.9 90.4 93.5 Net centralization (%) 20.956 22.181 12.5 Table S3 Corresponding network statistics for all giant trevally tagged in shallow waters (n=5) at Pearl and Hermes Atoll. Out and in degree relate to degree centrality statistics.

 Supplement 7 Seasonal changes in swimming depth and body temperature for Galapagos sharks

a b 0 0

20 20

40 40 Depth (m) Depth (m)

60 60

80 80 Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep

c 0

20 Deep West CH SW 40 West SP Depth (m)

60

80 Sep Nov Jan Mar May Jul Sep

Fig. S5 Seasonal changes in swimming depth for 3 Galapagos sharks G1 (a), G2 (b), and G3 (c) at Pearl and Hermes Atoll. Data were obtained from internally implanted acoustic transmitters. Each symbol corresponds to particular acoustic listening station (see Fig. 1 for locations).

 a b 28 28

26 26

24 24

22 22 Body temperature (°C) Body temperature (°C) Body

20 20 Sep Nov Jan Mar May Jul Nov Jan Mar May Jul Sep

c 28 Deep 26 SW West SP

24

22 Body temperature (°C) Body

20 Sep Nov Jan Mar May Jul Sep

Fig. S6 Seasonal changes in body temperature in Galapagos sharks G1 (a), G2 (b), and G3 (c) at Pearl and Hermes atoll. Data were obtained from internally implanted acoustic transmitters. Each symbol corresponds to particular acoustic listening station (see Fig. 1 for locations).

 Supplement 8 Seasonal changes in body temperature for giant trevally

a b 28 28

26 26

24 24

22 22 Body temperature (°C) Body Body temperature (°C) Body

20 20 Oct Dec Feb Apr Jun Aug Oct Sep Nov Jan Mar May Jul Sep

c d 28 28

26 26

24 24

22 22 Body temperature (°C) Body Body temperature (°C) Body

20 20 Sep Nov Jan Mar May Jul Sep Sep Nov Jan Mar May Jul Sep

Deep NW SE SW West SP

Fig. S7 Seasonal changes in body temperature for giant trevally U1 (a), U2 (b), U3 (c), U4 (d) at Pearl and Hermes atoll. Data were obtained using implanted acoustic transmitters. Each symbol corresponds to a particular acoustic listening station. The location of each listening station can be seen in Fig. 1. Shaded rectangles indicate the known period of spawning in giant trevally in Hawaii.

 Supplement 9 Stable isotope compositions for fishes caught in shallow and mesophotic habitats

-10 brown surgeonfish Top predators Multibarred goatfish -12 Old woman wrasse Shallow reef Galapagos shark -14 Giant trevally Milletseed butterfly fish (deep) Soldier fish (deep) -16 Chromis verater (deep) C Milletseed butterfly fish (shallow)

13 Tuna d -18

-20 Pelagic predator

-22 Mesophotic reef -24 6 7 8 9 10 11 12

d15N

Fig. S8 Bulk stable isotope signatures for top predators, pelagic predators (tuna) and shallow (<30 m) and mesophotic (> 55 m) reef communities. Sample sizes were as follows: brown surgeonfish (Acanthurus nigroris, n=8), Multibarred goatfish (Parapeneus multifasciatus) shallow (n=13), Multibarred goatfish deep (n=2), Old woman wrasse (Thalassoma ballieui, n=16), Milletseed butterfly fish shallow (Chaetodon miliaris, n=10), Milletseed butterfly fish deep (n=4), Soldierfish deep (Myripristis chryseres, n=4), Chromis verater deep (n=11). Values for tuna were taken from Hilting et al. 2013.

References

Hilting AK, Currin CA, Kosaki RK (2013) Evidence for benthic primary production support of an apex predator- dominated coral reef food web. Mar Biol 160:1681−1695


Supplement 10 Sensitivity analysis for tissue discrimination factors and mixing model outputs

Study Species Discrimination Discrimination Shark Shark MCE Trevally Trevally δ 15N δ 13C shallow (%) shallow MCE (%) (%) (%)

Kim et al. Leopard 3.7 ± 0.4 ‰ 1.7 ± 0.5 ‰ 59 (77-41) 41 (59-23) 60(67-43) 50(57-40) 2011 shark

Madigan et Bluefin 1.9 ±0.4 ‰ 1.8 ±0.3 ‰ 65 (43-85) 36 (57-16) 57(69-44) 43(56-31) al. 2012 tuna

Caut et al. Fish 2.65 ± 0.08 ‰ -0.03 ± 0.19 73(92-53) 27 (48-8) 60 (48-72) 40 (28-52) 2009* ‰

Hussey et Sharks 2.43 ± 0.27 ‰ 0.86 ± 0.28 ‰ 67 (47-86) 33 (14-53) 63 (50-75) 37 (25-50) al. 2010

Hussey et Sharks 2.43 ± 0.27 ‰ 0.86 ± 0.28 ‰ 86(98-68) 14(32-2) - - al. 2010 after lipid correction**

Table S4 Effect of tissue discrimination factors on mixing model output. The dietary contribution of shallow and mesophotic reefs (MCE) are given for both sharks and trevally (median (95 % Confidence Intervals)).*Discrimination factors are dependent on individual δ15N and δ13C values.**Lipid correction of shark bulk isotope values using equation in Logan et al. 2008.

References

Caut S, Angulo E, Courchamp F (2009) Variation in discrimination factors (Δ15N and Δ13C): the effect of diet isotopic values and applications for diet reconstruction. J Appl Ecol 46: 443−453

Hussey NE, Brush J, McCarthy ID, Fisk AT (2010) δ15N and δ13C diet-tissue discrimination factors for large sharks under semi-controlled conditions. Comp Biochem Physiol A 155:445−453

Kim SL, Casper DR, Galvan-Magana F, Ochoa-Diaz R, Hernandez-Aguilar SB et al (2011) Carbon and nitrogen discrimination factors for elasmobranch soft tissues based on a long term controlled feeding study. Environ Biol Fishes 95: 37-52

Logan JM, Jardine TD, Miller TJ, Bunn SE, Cunjak RA, Lutcavage ME (2008) Lipid corrections in carbon and nitrogen stable isotope analysis: comparison of chemical extraction and modelling methods. J Anim Ecol 77:838-846

Madigan DJ, Litvin SY, Popp BN, Carlisle AB, Farwell CJ, Block BA (2012) Tissue turnover rates and isotopic trophic discrimination factors in the endothermic teleost, Pacific bluefin tuna (Thunnus orientalis). PLoS ONE 7:e49220


Supplement 11 Representative fish abundance on shallow and mesophotic reefs

Fig. S9 Differences in fish abundance between species sampled for the isotope study, from both shallow reefs (n=38) and mesophotic reefs (MCE, n=20) at P&H atoll. Abundance estimates were obtained from dive transects, bars represent means and error bars are standard error. Fish species are Acanthurus nigroris (AN), Chaetodon miliaris (CM), Chromis verater (CV), Myripristis chryseres (MC), Parupeneus multifasciatus (PM), and Thalassoma balleui (TB). Dive surveys consisted of 25x2 m transects, with all conspicuous non-cryptic fishes counted and identified to species level. Data from Kosaki et al. unpublished.