Moray eels are more common on coral reefs subject to higher human pressure in the greater Caribbean

Summary

Proximity and size of the nearest market (‘market gravity’) have been shown to have strong negative effects on coral reef fish communities that can be mitigated by the establishment of closed areas. However, moray eels are functionally unique predators that are generally not subject to targeted fishing and should therefore not directly be affected by these factors. We used baited remote underwater video systems to investigate associations between morays and anthropogenic, habitat, and ecological factors in the Caribbean region. Market gravity had a positive effect on morays, while the opposite pattern was observed in a predator group subject to exploitation (sharks). Environmental DNA analyses corroborated the positive effect of market gravity on morays. We hypothesize that the observed pattern could be the indirect result of the depletion of moray competitors and predators near humans.

Graphical Abstract

Highlights

• •Baited remote underwater videos and environmental DNA were used to assess morays
• •Market gravity had a strong positive effect on moray abundance
• •Morays and sharks were negatively associated
• •Lack of competitors and predators may explain increased morays on reefs near humans

Environmental science; ecology; biological sciences; zoology; animals; ethology

Introduction

The size and proximity of the nearest market has a strong negative effect on predatory coral reef fishes, with some relief provided by closing reef areas to fishing (Cinner et al., 2018; Graham et al., 2017; MacNeil et al., 2020; Valdivia et al., 2017). This pattern has been observed on coral reefs worldwide, affecting commercially exploited predator taxa ranging from groupers to sharks (Cinner et al., 2018; Graham et al., 2017; MacNeil et al., 2020; Valdivia et al., 2017). Moray eels (Family Muraenidae; hereafter referred to as ‘morays’) are not readily detected in conventional reef surveys, and it is unknown how they respond to human impacts on coral reef systems (Chan, 2017; Gilbert et al., 2005; Mehta and Wainwright, 2007; Mouillot et al., 2013). Morays comprise nearly 200 species of functionally unique reef predators, with their elongate bodies and raptorial pharyngeal jaws enabling them to ambush relatively large fish, crustaceans, and octopuses in habitats with complex structure (Chan, 2017; Gilbert et al., 2005; Mehta and Wainwright, 2007; Mouillot et al., 2013). Morays have little to no commercial value because they are generally unpalatable and can contain toxic levels of ciguatera (Chan, 2017). We hypothesize that size and proximity of markets and closing reefs to fishing would therefore have no effect on morays.

The aim of the present study was to explore relationships between morays and two human-related factors that typically affect predatory reef fish: whether a reef is open or closed to fishing (‘protection status’) and the size and proximity of the nearest market (‘market gravity’), as well as the interaction between these two factors. Market gravity is a measurement of human impact on a location that is a function of both accessibility in terms of travel times to markets and the size of these markets (i.e., human population size [Cinner et al., 2018]). We included reef structural complexity as a factor in the model, which generally has a positive effect on reef fish (MacNeil et al., 2020; Valdivia et al., 2017). We also explored relationships between sharks and these factors to enable comparison between morays and a commercially exploited predator taxon.

Results and discussion

Video sampling throughout the greater Caribbean was conducted as part of the Global FinPrint Project (https://globalfinprint.org/) and involved the deployment of 60-minute baited remote underwater video systems (BRUVS) with ∼1 kg of crushed oily fish in a bait cage within the field of view (see Transparent methods in Supplemental information). A total of 3,157 BRUVS were deployed between June 2015 and April 2019 across 67 reefs in 12 Caribbean nations (Table 1). BRUVS were set in randomized locations on each reef, which was defined in this study as a continuous patch of hard-bottom coral reef habitats of ∼10 km² across a depth range of 2–40 m. Sightings of morays and sharks were converted to MaxN, which is the maximum number of individuals of each species observed in one frame per 60-minute BRUVS (Harvey et al., 2018). We fit a generalized linear model (GLM) with a Tweedie compound Poisson error structure, which explained 56% of the deviance in mean reef-level moray MaxN (all species combined) and included the factors: log-transformed market gravity (+), protection status (open [+] or closed [–] to fishing), reef complexity (+), and the interaction between log-transformed market gravity and protection status (+) (Table 2, Figure 1A). Reef complexity had the greatest effect (F = 57.91, p < 0.0001), followed by log-transformed market gravity (F = 33.89, p < 0.0001). Protection status (F = 3.22, p = 0.08) and the interaction between protection status and log-transformed market gravity (F = 0.13, p = 0.72) did not significantly affect reef-level moray MaxN. In a model exploring how the same factors affected shark MaxN log-transformed market gravity had the largest effect ([–]; F = 27.50, p < 0.0001), followed by reef complexity ([+]; F = 8.10, p = 0.01). Protection status (open [–] or closed [+] to fishing; F = 0.03, p = 0.87) and the interaction between protection status and log-transformed market gravity ([–]; F = 0.64, p = 0.43) had no significant effect on reef-level shark MaxN (Figures 1B, Table 3). There was no relationship between reef complexity and market gravity in our sampling (Figure 1C), but there was a negative correlation between morays and sharks at both the reef and national scale (Figures 1D and 2). A model where the factor log-transformed market gravity was replaced with reef-level shark MaxN explained 45% of the deviance in mean reef-level moray MaxN, where mean shark MaxN had the greatest effect ([-]; F = 30.81, p < 0.0001), followed by reef complexity ([+]; F = 20.57, p < 0.0001) (Table 4, Figure 1D). Protection status did not significantly affect moray MaxN here either (F = 2.24, p = 0.14). Overall, the GLM with the factor log-transformed market gravity (AIC = −36.63) predicted moray MaxN better than the GLM with the factor shark MaxN (AIC = −22.84). A spatial co-occurrence probability model for shark and moray species was used to determine positive, negative, or random interactions at the reef-level using the R library ‘co-occur’ (Griffith et al., 2016). Spatial co-occurrence between sharks and morays was dominated by negative probabilities (Veech, 2013; Figure 3).

Table 1.

BRUVS locations and corresponding anthropogenic and habitat factors

Location	Reef	Protection status	Market gravity	Reef complexity	No. BRUVS	Moray MaxN	Shark MaxN
Antigua and Barbuda	Barbuda – North	Open	8	1.53	42	0.05	0.43
	Barbuda – Palasar	Closed	9	2.11	45	0.13	0.71
Bahamas	Abaco – Bight of Old Robinson Bay	Open	19.02	1.54	61	0	0.1
	Abaco – Bight Reef	Open	18.41	1.8	68	0.03	1.04
	Abaco – Elbow Cay Reef	Open	22.2	1.71	57	0.04	0.54
	Abaco – Green Turtle Cay Reef	Open	19.52	1.28	61	0.05	0.79
	Abaco – Guana Cay Reef	Open	18.03	1.41	58	0	1.1
	Abaco – Water Cay Bay	Open	19.67	1.67	60	0	0.1
	Andros – North	Open	24.57	1.13	57	0.11	0.84
	Andros – South	Open	19.95	0.84	56	0	1.09
	Berry Islands – Chub Cay	Open	20.91	0.74	39	0.05	1.03
	Bimini – North	Open	79	1.44	53	0	0.7
	Bimini – South	Open	79	1.43	52	0	0.6
	Conception Island	Closed	4.4	0.23	42	0	0.83
	∗Eleuthera – South	Open	9	N/A	0	N/A	N/A
	∗Exumas – Middle	Closed	8.67	0.68	67	0.01	1.3
	Exumas – North	Open	11.47	1.02	67	0	1.21
	Exumas – South	Open	8	0.49	42	0	1.12
	New Providence – North	Open	1135.84	0.47	32	0.25	0.5
	New Providence – South	Open	134.33	0.29	38	0.08	1.03
	∗San Salvador	Open	4.81	0.41	78	0.04	0.65
Barbados	Northeast	Open	124.75	0.16	29	0	0.24
	Northwest	Open	173.81	1.36	31	0.52	0.06
	Southeast	Open	295.89	2.05	35	1	0.57
	Southwest	Open	18665.35	0.91	35	0.51	0
Belize	Belize City	Open	106.64	1.46	39	0.26	0
	Glover's Reef East	Closed	80.45	2.06	123	0.23	0.92
	Glover's Reef West	Open	85.1	1.52	41	0.34	0.12
	Lighthouse Reef – Halfmoon Caye	Closed	52.19	1.34	43	0.02	1.09
	Lighthouse Reef – Sandbore Caye	Open	45.65	1.77	31	0.13	0.61
	South Water Caye	Open	119.56	1.36	106	0.09	0.23
British West Indies	Montserrat – West	Open	20.04	1.5	48	0.15	0.17
Colombia	Coastal – Isla Mangle	Open	133.2	1.62	50	0.16	0
	Coastal – Isla Tintipan	Open	126.6	1.74	50	0.14	0
	Coastal – Isla Grande	Closed	287.7	2.65	49	0.51	0.04
	Coastal – Isla Tesoro	Closed	267	2.88	50	0.58	0.16
	∗Seaflower – Albuquerque	Open	12	1.14	50	0.02	1.28
	Seaflower – Serranilla	Open	1	1.14	63	0.03	1.37
	Seaflower – Providencia	Open	264.78	1.06	43	0.14	0.79
Cuba	Guanahacabibes – Maria la Gorda	Closed	70.56	3.13	40	0.15	0.2
	Guanahacabibes – Reef	Closed	66.9	3	40	0.08	1.8
	Punta Frances North	Closed	39	2.79	40	0.3	0
	Punta Frances South	Closed	100	3.38	40	0.25	0
	Jardines de la Reina	Closed	33.47	2.7	62	0.34	0.71
Dominican Republic	Bayahibe	Open	249.23	0.49	42	0.07	0
	Boca Chica	Open	2661.95	0.56	44	0.11	0
	Buen Hombre	Open	312.93	1.89	40	0.13	0
	Santo Domingo – La Caleta	Closed	2962	2.34	40	0.38	0
French West Indies	Guadeloupe – Grande cul de Sac	Open	159.07	3.07	41	0.39	0.02
	Guadeloupe – Petit Terre	Closed	95.18	2.21	48	0.42	0.08
	Martinique – Reef 1	Open	226.9	2.96	42	0.76	0
	Martinique – Reef 2	Open	137.64	3.15	47	1.02	0
Jamaica	Coastal – Dragon Point	Open	330.54	1.54	25	0.44	0
	∗Coastal – East Portland	Open	329.92	1.68	37	0.24	0
	Coastal – Negril	Closed	513.64	1.44	24	0.38	0
	Coastal – Ocho Rios	Open	377.82	1.24	34	0.15	0
	Pedro Bank	Open	15.58	3.12	54	0.28	0.67
Turks and Caicos	South Caicos – Back	Open	67	0.19	25	0	0.4
	South Caicos – Reserve	Closed	61.55	0.52	20	0	0.8
U.S.A.	Florida – Miami	Open	13,785.08	0.4	38	0.03	0.18
	Florida Keys – Lower Keys	Open	92.55	0.76	43	0.16	0.33
	Florida Keys – Tip	Open	125.67	1.5	40	0.23	0.45
	∗Florida Keys – Middle Keys 1	Open	171.88	0.79	43	0.23	0.49
	Florida Keys – Middle Keys 2	Open	135.58	0.85	42	0.12	0.29
	Florida Keys – Middle Keys 3	Open	107.14	0.73	41	0.15	0.39
	Florida Keys – Upper Keys 1	Open	250.88	0.59	42	0.07	0.38
	Florida Keys – Upper Keys 2	Open	226.29	0.96	44	0.18	0.66
	Florida Keys – Upper Keys 3	Open	199.79	1.38	48	0.17	0.58

Reefs with an asterisk denote where eDNA samples were also collected. Protection status refers to open or closed to fishing. Market gravity score is derived from Cinner et al. (2018) ‘Global Gravity of Coral Reefs Spatial Layer.’ Reef complexity (0–5) is scored for each BRUVS (via BenthoBox online annotation tool [https://benthobox.com/]) and averaged per reef. Moray MaxN and Shark MaxN are the mean per reef.

Table 2.

Tweedie GLM to predict moray abundance on BRUVS including market gravity

Factors	Degrees of freedom (DF)	Deviance	Residual DF	Residual deviance	F-value	p value	Deviance explained
Null			66	25.25
log10(Gravity)	1	5.01	65	20.24	33.89	<0.0001	0.20
Protection status	1	0.48	64	19.77	3.22	0.08	0.02
Reef complexity	1	8.56	63	11.20	57.91	<0.0001	0.34
log10(Gravity) ∗ Protection status	1	0.02	62	11.18	0.13	0.72	0.001
Sum of deviance explained							0.56

Analysis of deviance table for the Tweedie GLM to determine the effects of key anthropogenic (log-transformed market gravity score, protection status [open or closed to fishing]) and habitat (mean reef complexity score [0–5]) factors on the reef-level occurrence (mean MaxN) of morays on reefs throughout the greater Caribbean (n = 67 reefs). An asterisk denotes the interaction between factors.

Figure 1.

Relationships between market gravity, morays, sharks, and reef complexity in the Caribbean

(A) Relationship between market gravity and moray occurrence (n = 67 reefs in 12 nations).

(B) Relationship between market gravity and shark occurrence on the same reefs.

(C) Relationship between market gravity and mean reef complexity.

(D) Relationship between moray and shark MaxN. Red lines depict the linear trendlines.

Correlation coefficients and p values calculated via Pearson (A, B, C) and Spearman's rank (D; non-linear relationship).

Table 3.

Tweedie GLM to predict shark abundance on BRUVS

Factors	Degrees of freedom (DF)	Deviance	Residual DF	Residual deviance	F-value	p value	Deviance explained
Null			66	47.61
log10(Gravity)	1	12.47	65	35.14	4.49	<0.0001	0.26
Protection status	1	0.01	64	35.12	2.88	0.87	0.0003
Reef complexity	1	3.68	63	31.45	28.26	0.01	0.08
Protection status ∗ log10(Gravity)	1	0.29	62	31.16	0.64	0.43	0.01
Sum of deviance explained							0.35

Analysis of deviance table of Tweedie GLM to determine the effects of key anthropogenic (protection status [open or closed to fishing], log-transformed market gravity score) and habitat (mean reef complexity score [0–5]) factors on the reef-level occurrence (mean MaxN) of sharks on reefs throughout the greater Caribbean (n = 67 reefs). An asterisk denotes the interaction between factors.

Figure 2.

Relationship between mean country-level occurrence of sharks and morays on BRUVS

Jurisdictions where all shark fishing and trade is prohibited at the time of sampling are labeled as ‘Shark sanctuaries’. Red circles denote non-shark sanctuaries, blue triangles denote shark sanctuaries, and vertical and horizontal error bars denote standard deviation.

Table 4.

Tweedie GLM to predict moray abundance on BRUVS including shark abundance

Factors	Degrees of freedom (DF)	Deviance	Residual DF	Residual deviance	F-value	p-value	Deviance explained
Null			66	28.00
Shark MaxN	1	7.24	65	20.76	30.81	<0.0001	0.26
Protection status	1	0.53	64	20.23	2.24	0.14	0.02
Reef complexity	1	4.83	63	15.40	20.57	<0.0001	0.17
Sum of deviance explained							0.45

Analysis of deviance table for the Tweedie GLM to determine the effects of key anthropogenic (protection status [open or closed to fishing]), habitat (mean reef complexity score [0–5]), and ecological (mean shark MaxN) factors on the reef-level occurrence (mean MaxN) of morays on reefs throughout the greater Caribbean (n = 67 reefs).

Figure 3.

Co-occurrence probability matrix between morays and different shark species on BRUVS

Each cell represents the spatial co-occurrence probability of a pair of species. Significant probabilities are denoted by color (red: negative, blue: positive). Non-significant co-occurrence is denoted by the color gray.

Recording animals on BRUVS is partially dependent on a positive response to bait, which can be inhibited in predator-rich habitats by predator avoidance behavior (Sherman et al., 2020) and interspecific competition (Sabando et al., 2020). It is therefore important to compare detections from multiple survey methods, including techniques that don't require a behavioral response from the subject (Sherman et al., 2020). Six reefs subject to a range of market gravity (4.81–329.92) were surveyed for moray extra-organismal environmental DNA (eDNA) by vacuum filtering 14–21 replicate 2L water samples collected at random locations per reef across hydrophilic polyethersulfone filters and using universal 12S teleost fish primers (Miya et al., 2015) to generate sequence reads. We reasoned that eDNA surveys would be less sensitive than BRUVS to any difference in moray behavior between reefs because eDNA detection does not depend on individual morays approaching a bait source. A total of 19,168 moray sequence reads, belonging to at least six species, were obtained (Table S1). Each water sample was scored from 0 to 3 representing the minimum number of individual morays detected. We conservatively assumed that all detections of the same species on one filter were from one individual, thus our metric is equivalent to moray species richness per sample. A GLM with a negative binomial error distribution explored the relationship between this metric and ‘high gravity’ (log-transformed market gravity >2; East Portland, Jamaica and Middle Florida Keys, U.S.A.) and ‘low gravity’ categories (log-transformed ∼ 1; Middle Exumas, San Salvador, Cape Eleuthera [Bahamas] and Albuquerque Island in the Seaflower Biosphere Reserve [Colombia]). Market gravity explained 16.9% of the deviance in the minimum number of moray individuals detected per sample (p = 0.01; Table 5). Significantly more individuals were detected in samples collected on reefs subject to high market gravity (at least six moray species were detected overall, with 1–3 individuals in 17.6% of water samples) than reefs subject to lower market gravity (at least two moray species were detected overall, with 1–2 individuals in 3.5% of samples; Table S1).

Table 5.

Negative binomial GLM to predict moray eDNA detection

Factors	Degrees of freedom (DF)	Deviance	Residual DF	Residual deviance	p-value	Deviance explained
Null			116	38.72
log10(Gravity)	1	6.54	115	32.19	0.01	0.17

Analysis of deviance table for the negative binomial GLM to determine the effect of log-transformed market gravity (low [~1] or high [> 2]) on the minimum number of moray individuals detected in water samples (n = 118) collected across 6 reefs in the greater Caribbean (U.S.A., Jamaica, Bahamas, Colombia [Seaflower Biosphere Reserve]).

This study is the largest effort to date to assess moray occurrence on coral reefs. The positive effect of increasing reef complexity was expected and likely reflects moray preference for reef structures that provide both refuge and prey ambush opportunities (Gilbert et al., 2005). While a preference for complex structure is typical of Caribbean coral reef predators (Valdivia et al., 2017), the positive effect of market gravity on moray sightings is unusual. This factor or its components (local human population density, proximity to humans) nearly always have a strong negative effect on reef predators (Cinner et al., 2018; Graham et al., 2017; MacNeil et al., 2020; Valdivia et al., 2017). We hypothesize that the positive effect of market gravity on morays could be an indirect result of a local reduction in moray predators, competitors, or both on reefs subject to higher human pressure. The negative correlation between shark and moray MaxN we observed at the reef and national level is consistent with this ‘release’ hypothesis, with uniformly low moray MaxN observed in jurisdictions where sharks are protected at the national level and common on all sampled reefs (‘Shark Sanctuaries’, n = 21 reefs in The Bahamas and Sea Flower Biosphere Reserve, Colombia; Figure 2). Local shark MaxN had a significant negative effect in the moray GLM, which was further supported by the inverse relationship between sharks and morays as demonstrated by their negative co-occurrence probabilities at the reef-level (Figure 3). Morays were less likely to occur on reefs where several shark species (Caribbean reef, nurse, blacknose, tiger, blacktip, unknown) were present. While more research on the ecological relationships between morays, sharks, and other predators is needed to further test this hypothesis, it is mechanistically viable because a wide variety of shark species include morays in their diet (Delorenzo et al., 2015; Estupiñán-Montaño et al., 2017; Lowe et al., 1996; McElroy et al., 2006; Sears and Sikkel, 2016; Torres-Rojas et al., 2010; Wetherbee et al, 1996, 1997) (e.g., Families: Carcharhinidae, Sphyrnidae, Ginglymostomatidae) and probably also compete with them for prey. Anecdotal observations at St. Paul's Rocks (Brazil) linked the decline of sharks around the island with a subsequent increase in sightings of the whitespot moray (Muraena pavonina) (Luiz and Edwards, 2011). Nonetheless, the better fit of the market gravity model than the substituted shark MaxN model suggests that there may be other factors that are contributing to the positive effect of market gravity on morays. It is possible that removals of moray predators and competitors other than sharks (e.g., large teleosts) may also be important. Other viable hypotheses outside of release that could potentially explain this pattern include that prey availability is greater near humans or that human activities benefit morays (e.g., if fishing provides opportunities for morays to scavenge discarded catch).

It is possible that morays are less inclined to approach BRUVS on reefs where other predator taxa are common due to elevated predation risk or to avoid aggressive interactions with competitors (Sherman et al., 2020). Indeed, we found that mean moray MaxN was significantly lower on BRUVS sets where at least one shark was also recorded than on sets where sharks were absent (negative binomial GLM, p < 0.001; Figure 4 and Table S2). However, shark presence on BRUVS sets did not always deter moray detection: in 98 analyzed BRUVS when sharks and morays co-occurred, they did not overlap in frame on 76 occasions (76.6%). Moreover, on 23 BRUVS where morays and sharks occurred in frame together, the moray left the field of view and did not return after a shark arrived in only five instances (22%; Video S1). Morays generally ignored sharks when they were able to remain partially hidden in reef structure, when the shark(s) passed in the background, or when the individual shark was small, perhaps because the shark was too small to be a direct threat (Video S2). Detection of moray eDNA also exhibited the same pattern observed in the BRUVS; morays were detected less often in water samples from reefs subject to low market gravity. This indicates that instead of just hiding more often (in which case their eDNA would still be detected), and consequently not being visually detected on BRUVS, morays were also less common at these sites. The combined eDNA and BRUVS results argue for reduced moray populations and changes in their behavior on reefs with lower market gravity and more predators.

Figure 4.

Set-level abundance of morays in the presence or absence of sharks

Mean moray MaxN ±SE on BRUVS sets (n = 2,052 BRUVS across 51 reefs in 11 nations) where sharks were absent or at least one shark was present, on reefs where at least one shark was detected. p value calculated via negative binomial GLM (Table S2).

Morays are functionally unique in that they are able to move undetected within complex reef structure and ambush relatively large prey (Gilbert et al., 2005; Mehta and Wainwright, 2007; Mouillot et al., 2013). Changes in moray abundance and the risk of moray ambush from reef structure could affect prey populations and corresponding risk-sensitive behavior. It may also reduce cooperative hunting with other fishes (e.g., grouper [Epinephelinae]) that capture prey flushed out of structure by morays (Bshary et al., 2006). Yet the ecological roles and importance of morays are not as extensively studied as most other reef predator taxa making it difficult to predict these effects (Gilbert et al., 2005; Mehta and Wainwright, 2007; Mouillot et al., 2013). We show that eDNA and BRUVS are potentially informative survey approaches for these cryptic predators, which are frequently undercounted using underwater visual censuses (Gilbert et al., 2005). Our study indicates that morays are more common on Caribbean coral reefs subject to high market gravity, making them unusual among the major predator taxa (Cinner et al., 2018; Graham et al., 2017; MacNeil et al., 2020; Valdivia et al., 2017). A better understanding of moray ecology is needed to assess the causes and consequences of these profound changes in Caribbean coral reef predator assemblages, while also acknowledging that emerging export markets for morays (Mañez and Paragay, 2013) and human impacts that reduce reef complexity and prey biomass may ultimately deplete morays close to people.

Limitations of the study

This study used a large sample of BRUVS set randomly on reefs distributed throughout the greater Caribbean and provided corroboration of the BRUVS results with more limited eDNA sampling. Our eDNA sampling sites (n = 6) were opportunistically selected and half of the sites were within The Bahamas. Expanding the number and geographic coverage of reefs sampled for moray eDNA could overcome this limitation in the future. Although BRUVS were deployed during daylight hours in this study, it is suggested that some species of reef-associated morays and sharks actively forage during twilight and night hours (Bardach et al., 1959; Chapman et al., 2007; Gruber et al., 1988). Further studies could therefore incorporate nighttime surveys to explore the relationship between these species under low-light conditions. This study focused on testing for associations between morays and three factors that generally have strong effects on reef fish: market gravity, protection status, and reef complexity. While these factors explain much of the deviance in moray sightings on BRUVS there are many other important factors that could be investigated. For example, our model does not include any metric for prey availability. Morays can have broad diets that include a diversity of reef fish, crustaceans, and octopus, some of which are not well sampled using BRUVS and may require targeted sampling efforts using other methods. There remain several potential mechanisms that could be driving the positive association between moray occurrence and market gravity. We discussed a ‘release’ hypothesis in relation to the loss of sharks, but this could also apply to the loss of other potential predators and competitors. Our study only compared morays and sharks due to limitations on the time to annotate BRUVS for other species, and it would be insightful to also model the co-occurrence of morays, sharks, and other large predators on BRUVS in order to resolve the strongest interactions. It is also a viable hypothesis that human activities increase the prey available to morays. Although increased market gravity typically reduces overall reef fish biomass (Cinner et al., 2018), this does not necessarily translate to fewer prey for morays. It is possible that low biomass reefs are composed of more suitable prey species for morays than high-biomass, predator-rich reefs. It is also possible that human activities directly enhance moray presence (e.g., if morays frequently scavenge fishery discards or if SCUBA tour operators feed morays). Overall, future studies should model the relationships of a wider variety of potential moray MaxN covariates to better resolve the key drivers behind the association detected in our analyses. Additional information on moray ecology and their location-specific interactions with humans are necessary to select and quantify the most appropriate potential covariates.

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Demian Chapman (dchapman@fiu.edu).

Materials availability

No materials were newly generated for this paper.

Data and code availability

The BRUVS sampling locations and corresponding ecological, anthropogenic, and habitat factors are listed in Table 1. The number of moray eel sequence reads obtained from water samples are listed in Table S1.

Methods

All methods can be found in the accompanying Transparent methods supplemental file.

Published: March 19, 2021