Shark movements in the Revillagigedo Archipelago and connectivity with the Eastern Tropical Pacific

Long-distance movements of sharks within and between islands pose substantial challenges for resource managers working with highly migratory species. When no-take zones do not cover the critical areas that sharks use as part of their lifecycle, exposure to fishing activities can be significant. Shark movements between the Marine Protected Areas (MPAs) of the Eastern Tropical Pacific (ETP) have been studied for several years, however little is known about the strength of connectivity between these islands. We analyzed the extensive MigraMar ultrasonic telemetry dataset to assess how Galapagos sharks (Carcharhinus galapagensis) and silky sharks (Carcharhinus falciformis) use different islands as stepping-stones during their migrations within the Revillagigedo National Park and other ETP islands. Of the 66 sharks monitored, 63.5% moved within the same island, 25.4% between two islands or more and only 10.1% across different MPAs. A C. falciformis tagged in Roca Partida Island, Revillagigedo, travelled to Clipperton Atoll and another one tagged in Darwin Island travelled to the atoll on two different years. The largest movement of C. galapagensis was accomplished by a shark tagged at Socorro Island, Revillagigedo, later detected at Clipperton and finally recorded in Darwin Island, Galapagos. This last path was in fact, one of the longest movements ever recorded for the species. Although long-distance dispersion was not common, our results highlight the need for co-operation between different countries to ensure adequate protection for sharks in the form of swimways and other conservation tools in the ETP.


117
The definition of the extent and occurrence of long-range movements and 118 population connectivity are necessary for a full understanding of the ecology of a 119 species and hence for designing effective conservation action (18). By assessing 120 movement frequency, Network Analysis (NA) can be used to identify important 121 movement corridors between core habitats of a species (6,19). NA provides a new 122 insight into the connectivity of specific habitats and the animals moving between them.

123
It also proves valuable in revealing important information on distinct spatial and 124 temporal changes in animal movements (6,19). For example, an area with a high 125 degree of centrality would suggest strong site fidelity by wide-ranging animals, hence 126 the animals may return to the same location from many different areas (6).

127
Movements and residency patterns of key marine animals are still poorly    Fig 2). Some sharks were observed for over a period of 275 five years. We plotted the wet (stormy) and dry seasons to observe the relationship of 276 shark detections with changes in water temperature. During the stormy season, the 277 number of movements were reduced, which also shows drastic changes in the sea 278 surface temperature of almost 10°C in a single day (Fig 3). Although both species 279 showed similar distribution patterns, when we compared their records according to the were not random and that some hours are more important than others (Fig 4). 284

312
Of the overall detections analyzed, 63.5% of individuals undertook movements 313 within the same island, 25.4% between islands and 10.1% across MPAs (Fig 7 and 8). , which is one of the longest movements ever recorded for the species (Fig 11).

343
Circles represent the nodes and the arrows indicate the edges or movement paths. The 344 size of the circles represents the degree, that is, the number of links for each receiver.

346
Circles represent the nodes and the arrows indicate the edges or movement paths. The 347 size of the circles represents the degree, that is, the number of links for each receiver.

Network analysis and metrics 350
According to the NA metrics, C. falciformis has a more complex network with 351 significantly higher values than C. galapagensis in terms of the number connections 352 between the nodes or "edges" (X 2 = 44.714, df=1, p<0.5), the complexity of the 353 network or "degree of centrality" (X 2 = 40.164, df=1, p<0.5) and the proportion of 354 nodes used of the total options or "density" (X 2 = 14.238, df=1, p<0.5), whereas the number of recorded stations or "nodes" did not show a significant difference between galapagensis in terms of the number edges, the centralization and the density (p<0.5),

359
whereas the number of nodes did not show a significant difference (p= 0.75

370
According to this study, C. falciformis moved more frequently to nearby sites, but 371 also showed high fidelity to their tagging location. Some sharks may gain more 372 protection because of the location of their highest movement site and/or the 373 distribution of management zones in the system (2,40). Consequently, targeting 374 specific sites based on prior knowledge and increasing the level of protection to 375 include closely spaced habitats (20 km) may perform better for species like C.

376
falciformis than having a single reserve.

377
During this study, one of the longest movements for C. galapagensis has been 378 recorded. A subadult female of 181 cm TL was tagged in Socorro Island the 26 th of (2,200 km to the north) in two different years, showing that this species present very 385 long movement patterns as it was expected.

386
Shark populations are not homogenously distributed in different habitats of the 387 ecosystem that can support a higher diversity and abundance (2,5). Many shark species 388 are known to aggregate on outer parts of reef slopes that are generally exposed to 389 stronger current flow (3,4,41), where productive foraging grounds are present (40).

390
Hence, currents probably shape the shark community and define spatial and temporal 391 patterns of habitat use. sharks are not just highly residential, but they also start long dispersal from these sites 397 to other islands and marine reserves (more than 100 km).

398
We determined that these stepping-stones are sites where earlier studies have California, e.g. 14) and Southern ETP (Malpelo, Cocos and Galapagos, e.g. 8).

450
The ideal MPA design provides protection for all life stages of the species of 451 concern, which is impractical for most shark species because they are wide ranging. territorial waters of Clipperton Atoll should be continued with an extension of the 458 protection area to reach this critical threshold of 70,000 km 2 (e.g. 12). The observed 459 movements between marine protected areas suggest that these species are vulnerable to 460 domestic fisheries as well as multinational fisheries on the high seas, as these species

484
In Revillagigedo, the center of activity showed that sharks tend to be present 485 during the wet months and they do move between sites during these months. The shark 486 movement seasonality is related to current exposure, storms and temperature.

487
According to the network analysis there are movements between MPAs but are not