Feeding Ecology of Fishes Associated With Artificial Reefs in the Northwest Gulf of Mexico

Habitat selection is important for reef-associated fishes and is considered optimal if it maximizes the energy gained from prey while minimizing the risks of predation (Mittelbach 2002). How fish select a habitat to occupy is a complex process affected by intraspecific and interspecific competition, the presence of predators, substrate complexity and type, environmental conditions (e.g., temperature), and the quantity and quality of available food (Wootton 1998). Interactions between these factors will in turn impact habitat quality; for example, quality may decline with increased competition or improve with greater prey resources. Prey abundances and species vary among habitats; thus, residence on habitats with low-quality prey may hinder fish growth, ultimately leading to reduced condition and potentially reduced reproductive output (Adams and Breck 1990; Busacker et al. 1990; Wootton 1998).

The Red Snapper Lutjanus campechanus is a reef-associated species that is important in fisheries and fish communities of the Gulf of Mexico (hereafter, Gulf). Red Snapper in the Gulf have been overfished for at least 30 years and remain overfished according to the most recent benchmark stock assessment (Cass-Calay et al. 2015), although it appears that overfishing is no longer occurring in the eastern and western subunits of the Gulf stock (Cass-Calay et al. 2015). Red Snapper are found throughout the Gulf and occupy several habitat types, including natural hard-bottom reefs, soft-bottom substrates like mud and sand, and artificial reefs (Gallaway et al. 2009). Determining which habitats provide high-quality prey resources for Red Snapper in the Gulf may assist managers in maximizing Red Snapper reproductive output by implementing policies and regulations to protect those habitats.

Many of the natural reefs found on the Louisiana shelf edge are associated with active salt-diapirs buried beneath the slope, whereas other reefs are associated with bare bedrock or are dominated by encrusting coralgal organisms (Rezak et al. 1985; Holcombe et al. 2002). In addition to natural reefs in the Gulf, artificial reefs in the form of toppled, standing, and partially removed oil platforms are among the world's largest artificial reef systems, and the majority of these platforms are found in waters off the Louisiana coast (Kasprzak and Perret 1996). It is estimated that these platforms provide an additional 12.1 km² of reef fish habitat in the Gulf, a small contribution compared to the approximately 2,700 km² of natural reef habitat in the northern Gulf (Parker et al. 1983; Gallaway et al. 2009). Although oil and gas platforms are present in the eastern Gulf, the bulk of artificial reefs in this area are structurally different and consist of sunken ships, planes, cars, dry docks, tanks, and other small man-made items, such as concrete balls and pyramids (Minton and Heath 1998).

With two radically different types of reefs capable of supporting Red Snapper in the Gulf, it is important both to evaluate the functional role of the two habitat types in the life history of this valuable species and to incorporate any differences into management decisions. Given the recent progression of fisheries management from single‐species-based to ecosystem‐based methods, an understanding of the biological and environmental interactions of target species within a system is of utmost importance (Pikitch et al. 2004). A key aspect of these interactions is the feeding ecology of the target species, which provides insights into the energy flow through a system and the niche filled by a particular species (Layman et al. 2007).

Recent studies of Red Snapper feeding ecology conducted at more complex natural substrate habitats are lacking. The majority of studies focusing on adult Red Snapper feeding ecology only at natural reefs are over 20 years old (Camber 1955; Moseley 1966; Davis 1975; Nelson 1988), whereas more recent studies have been focused at artificial reefs (Ouzts and Szedlmayer 2003; Szedlmayer and Lee 2004; McCawley et al. 2007; Wells et al. 2008). A comparison of Red Snapper diets between artificial reefs (oil platforms) and three natural reefs on the eastern Louisiana shelf edge indicated that the diets were more similar between habitats than anticipated and that overlap in the prey items consumed was high, although prey diversity was higher at the natural reefs (Simonsen et al. 2015). Studies comparing Red Snapper diets between natural and artificial reefs in the northeastern Gulf have produced conflicting results, suggesting that differences do (Davis et al. 2015) and do not (Tarnecki and Patterson 2015) exist. In the western Gulf off the Texas coast, Red Snapper diets were found to be similar among natural reefs and standing and toppled oil platforms (Downey 2016). The natural reefs sampled by Simonsen et al. (2015), Tarnecki and Patterson (2015), Davis et al. (2015), and Downey (2016) are of lower habitat complexity than the natural reefs on the western part of the Louisiana shelf edge (Rezak et al. 1985; Gardner and Beaudoin 2005; Streich et al. 2017; T. A. Langland, personal observation), so additional information on Red Snapper diets from more complex reefs is needed to further assess differences in feeding ecology.

Despite the availability of data, no other study has compared Red Snapper feeding ecology between individual natural reefs. The present study addresses this by comparing Red Snapper feeding ecology among seasons, fish sizes, and natural reefs, and between natural and artificial reefs located on the Louisiana shelf edge. We used stomach content analysis to assess short-term diet trends, and we used analysis of nitrogen (δ¹⁵N) and carbon (δ¹³C) stable isotopes in muscle tissue to examine trophic variability. Foraging patterns were also determined by examining the habitat utilization of prey items found in the diets. We hypothesized that Red Snapper feeding ecology would differ among the natural reefs as well as between the natural and artificial reefs due to differences in substrate and habitat complexity.

METHODS

Study area

Red Snapper populations were sampled at three natural reefs of varying habitat complexity and at oil platforms within one artificial reef planning area in the northwestern Gulf (Figure 1). The three natural reefs sampled (Jakkula, McGrail, and Bright) are large, complex habitats and are part of an extensive network of reefs located on the northwestern Gulf continental shelf edge. The artificial reefs sampled are located in Lease Block 272 within the East Cameron Artificial Reef Planning Area (hereafter, East Cameron) and are placed on top of lithified or partially lithified delta mud (J. H. Cowan Jr., unpublished data). Differences in bathymetry, relief, surface area, underlying structure, surrounding and overlying sediment, the presence of reef-building corals, and habitat complexity are apparent among all four reefs (Supplementary Table S.1 available separately online with this article).

FIGURE 1. Map showing the three natural reefs (Jakkula, McGrail, and Bright) and the East Cameron Artificial Reef Planning Area (East Cameron) that were selected for sampling of Red Snapper in the northwestern Gulf of Mexico.

Sampling

Detailed sampling procedures were described by Schwartzkopf and Cowan (2017). Briefly, sampling was conducted twice quarterly between September 2011 and October 2013 as the weather and schedule permitted. Each Red Snapper was placed into one of four size-classes: size-class 3 (300–399 mm FL), size-class 4 (400–499 mm FL), size-class 5 (500–599 mm FL), and size-class 6 (>600 mm FL). Seasons were defined as winter (December–February), spring (March–May), summer (June–August), and fall (September–November). A water temperature profile was obtained at each site with a Sea-Bird Electronics 25 Sealogger conductivity–temperature–depth profiler. The temperature associated with a sample was the bottom temperature during each collection period.

Laboratory procedures

Stomachs were removed from Red Snappers, placed in individual jars, and immediately frozen to hinder further digestion. In the laboratory, stomachs were thawed, placed in a 10% solution of formalin for a minimum of 48 h, and transferred to a 70% ethanol solution. During examination, stomach contents were removed, identified, and sorted to the lowest taxonomic level, dried at 60°C, and weighed to the nearest 0.0001 g.

Epaxial muscle tissue from the left flank above the pectoral fin was taken from each Red Snapper and stored at –80°C until processed in the laboratory. Tissue samples were dried at 60°C for 24 h in a drying oven and then were homogenized with a ball-mill grinder. Samples were prepared according to the sample preparation guidelines supplied by the Stable Isotope Facility (SIF) at the University of California–Davis (UCD). Samples were analyzed for δ¹³C and δ¹⁵N with a PDZ Europa ANCA-GSL elemental analyzer interfaced to a PDZ Europa 20–20 isotope ratio mass spectrometer (Sercon Ltd., Cheshire, UK) at the UCD-SIF. Final isotopic ratios were reported relative to international standards (Vienna PeeDee belemnite for δ¹³C values; standard atmospheric N₂ for δ¹⁵N values) and were calculated as

where R represents the ratio of heavy to light isotope.

Statistical analysis

Percent composition by dry weight (%DW) was the primary method chosen for diet analysis because it can be used to assess the nutritional contribution of prey items (Bowen 1996; McCawley and Cowan 2007). Stomach contents were sorted into seven major prey categories for analysis: fishes, crabs, shrimp, zooplankton, other invertebrates, incidental items (items thought to be eaten incidentally during foraging, such as rocks), and tunicates. The diets contained large amounts of unclassifiable material, defined as material that did not include any bones, hard parts, or recognizable features for classification into one of the prey categories. This material was subsequently excluded from further analyses.

Habitat associations were assigned to each prey item found in the diets of Red Snapper at Jakkula, Bright, and East Cameron based upon the published literature and personal communications (Table S.2). Habitat associations could not be determined for unidentified fishes, crabs, shrimp, zooplankton, and crustaceans. After each identifiable prey item was assigned a habitat type, the %DW was summed to examine the contribution of each habitat type to the diets at each site.

Individual stomachs were treated as replicates, and data were log _e (x + 1) transformed to account for zeros in prey categories and to place more importance on rare prey species. A similarity resemblance matrix using Bray–Curtis similarity coefficients was then constructed on the transformed data to assess the similarity between weight proportions of prey items for each individual. The Bray–Curtis similarity coefficient was chosen because it (1) is the most common similarity coefficient for biological community data and (2) obeys natural biological guidelines better than other coefficients (Clarke and Gorley 2015). The Bray–Curtis similarity coefficient ranges from 1 to 100, with 100 representing perfect similarity. A permutational ANOVA (PERMANOVA) was used to test for significant differences in prey weight proportions among sites, among seasons, and among size-classes. Site, season, and size-class were used as fixed-effect factors, and the options chosen were type III sums of squares, permutation of residuals under a reduced model, and 9,999 permutations. If a significant (P < 0.05) main effect or interaction was found in the PERMANOVA, a subsequent pairwise test was carried out to determine which factors differed from one another. All diet statistical analyses were conducted with the PRIMER (Plymouth Routine in Multivariate Ecological Research) version 7 statistical package (Anderson et al. 2008; Clarke and Gorley 2015).

Stable isotope ratios were analyzed with both linear mixed models and the Stable Isotope Bayesian Ellipses in R (SIBER) program developed by Jackson et al. (2011). Values of δ¹³C and δ¹⁵N were compared between sites and seasons with a multivariate analysis of covariance (MANCOVA) that included FL as a covariate. Pairwise differences between least-squares (LS) means were examined for significance after a Tukey–Kramer adjustment.

The SIBER method was used to compare intraspecific isotope niche structure among sites, seasons, and size-classes. The corrected standard ellipse area (SEA _c ) was calculated for each site, season, or size-class group ("siar" package in R; Parnell and Jackson 2003) as defined relative to the SEA as

To limit underestimation bias and maintain a sample size greater than 10 individuals per group (Jackson et al. 2011), the site, season, and size-class effects were analyzed separately (i.e., three analyses, with four groups per analysis) rather than conducting a cross-classified analysis.

A distribution of Bayesian estimates for each site, season, and size-class group was generated using the "siber.ellipses" function in the "siar" package (Parnell and Jackson 2003). Simulated values of group means (10⁶) and covariance matrices were generated, from which the Bayesian SEA (SEA _B ) was calculated. The relative positioning and extent of group SEA _c values were quantified by calculating the overlap between groups in a pairwise fashion. The area of overlap (in ‰²) was defined simply as the shared area encompassed by the two standard ellipses being compared. To make this area more interpretable relative to the total area encompassed by the ellipses, the area of overlap was also presented as a percentage of each standard ellipse that was considered in the comparison.

RESULTS

Sampling Summary

In total, 651 Red Snapper were collected from Jakkula (n = 81), McGrail (n = 27), Bright (n = 215), and East Cameron (n = 328) during this study. From the 651 individuals, 636 tissue samples were analyzed for δ¹³C and δ¹⁵N values: 322 from East Cameron, 209 from Bright, 79 from Jakkula, and 26 from McGrail (). Among those 651 individuals, 362 contained stomach contents, 273 stomachs were empty due to regurgitation, and 16 stomachs were truly empty. Of the 362 fish with stomach contents, 261 contained identifiable prey items, and 101 contained only unclassifiable material or bait. Of the 261 usable stomachs, 23 were collected from Jakkula, 8 were collected from McGrail, 104 were collected from Bright, and 126 were collected from East Cameron (). The stomach contents of Red Snapper from McGrail were not included in the diet analysis due to the small sample size and uneven distribution, with only three stomachs collected in September 2011 and five collected in July 2012. The diet data from McGrail are included in Table S.3. The tissue samples from McGrail, however, were included in the stable isotope analysis. Sizes of Red Snapper used in analyses ranged from 432 to 802 mm FL at Jakkula, from 446 to 754 mm FL at McGrail, from 406 to 681 mm FL at Bright, and from 304 to 724 mm FL at East Cameron. More detailed information about the size-class distributions of Red Snapper caught at natural and artificial reefs in this study is provided by Schwartzkopf and Cowan (2017).

TABLE 1. Number of samples (stomach contents/muscle tissue) per season and size-class (size-class 3 = 300–399 mm FL; size-class 4 = 400–499 mm FL; size-class 5 = 500–599 mm FL; size-class 6 = >600 mm FL) used in the diet and stable isotope analyses of Red Snapper collected at three natural reefs and one artificial reef area (East Cameron) within the northwestern Gulf of Mexico.

General Site Differences

Based on %DW, fish were the most abundant prey type in the diets of Red Snapper at Jakkula (60%), Bright (68%), and East Cameron (44%). At Jakkula, tunicates (most likely the pelagic tunicate Pyrosoma atlanticum) were the second most abundant prey type (19%), whereas incidental items were the second most abundant type at Bright (16%), and zooplankton was the second most abundant type at East Cameron (23%). Because %DW is the most commonly used index to quantify Red Snapper diets, we do not present results for percent by number, percent frequency of occurrence, or percent index of relative importance, which generally revealed the same patterns as %DW.

Overall, 57 prey types were identified in the stomachs of Red Snapper collected at Jakkula, Bright, and East Cameron (). There was little overlap between prey types consumed at East Cameron and those consumed at Bright. There were only eight prey types—not including any unidentified prey categories—found in the diets at all three sites. Red Snapper at Bright had the most varied diet, with a total of 36 identified prey types and 19 unique prey species (i.e., only found at that particular site; ). The diet at East Cameron contained 30 identifiable prey types and 15 unique prey species, whereas the diet at Jakkula was the least varied, with only 10 identifiable prey items and 1 unique prey species ().

TABLE 2. Percent contribution by dry weight (%DW) of prey types in the diets of Red Snapper at three reef sites (Jakkula: n = 23; Bright: n = 104; East Cameron: n = 126) within the northwestern Gulf of Mexico. Unclassifiable material was not included in the calculation of %DW for each prey type. Asterisks represent prey types that were unique to the diet at a specific location.

Red Snapper exhibited significantly different isotope niche areas between sites (SIBER: P < 0.05). When combined across seasons and size-classes, fish from Bright and East Cameron showed significantly smaller SEA _B values than Jakkula and McGrail (SIBER: P < 0.05). Site groups showed a relatively low degree of overlap, an indication that site groups were distinct in C–N space. Jakkula and McGrail showed the highest degree of overlap between site ellipses (0.61‰²; Figure 2). This overlap encompassed 60.6% of the total SEA _c for the Jakkula group and 75.9% of the total SEA _c for the McGrail group. Minor overlap was also observed between Bright and McGrail (0.09‰²; 22.2% of Bright, 10.6% of McGrail) and between East Cameron and Jakkula (0.05‰²; 12.7% of East Cameron, 6.2% of Jakkula; Figure 2). No overlap in ellipses was observed between Bright and East Cameron or between Bright and Jakkula (Figure 2).

FIGURE 2. Scatter plot of white muscle δ¹⁵N (‰) values at a given δ¹³C (‰) value from individual Red Snapper collected at three natural reef sites (Jakkula, McGrail, and Bright) and one artifical reef area (East Cameron), overlaid with the corresponding corrected standard ellipse area (SEA _c ) for each site. The SEA _c values from which overlap was calculated were not scaled and thus contained 40% of the data for each site group.

The contributions of hard-substrate-, soft-sediment-, water-column-, variety-habitat-, and benthic-associated prey varied among sites (). Hard-substrate- and water-column-associated prey contributed the most to the diets at Jakkula and Bright, whereas soft-sediment- and water-column-associated prey were the largest contributors to the diet at East Cameron (). After benthic-associated prey, the smallest contributor to the diets at Jakkula and Bright was soft-sediment-associated prey, whereas hard-substrate-associated prey contributed the least to the diet at East Cameron ().

TABLE 3. Summary of percent dry weight contributions of hard-substrate-, soft-substrate-, water-column-, variety-habitat-, and benthic-associated prey items in the diets of Red Snapper at three reef sites (Jakkula: n = 23; Bright: n = 104; East Cameron; n = 126) within the northwestern Gulf of Mexico.

Given the habitat classifications of the natural and artificial reefs (Table S.1) and the habitat associations of prey items present in the diets (Table S.2), Red Snapper at both natural reefs (Bright and Jakkula) were feeding on prey found on and above the reef, while Red Snapper at the artificial reef area (East Cameron) were feeding on prey found on the surrounding seafloor and up in the water column.

Seasonal Site Differences

Season and the site × season interaction were significant contributors to the variation in Red Snapper diets (PERMANOVA: P < 0.05; ). Fish were consumed in greater proportions at both natural reefs compared to the artificial reef area in all seasons except winter (Figure 3). The diets varied between Bright and East Cameron and between Jakkula and East Cameron during spring (PERMANOVA: P < 0.05). These differences were driven by a larger biomass of zooplankton consumed at East Cameron in the spring (Figure 3). During summer, the diets differed between Bright and East Cameron (PERMANOVA: P < 0.05), and the difference was driven by the greater proportion of crabs and smaller proportions of fishes and tunicates consumed at East Cameron (Figure 3). The diets also varied between Bright and East Cameron and between Jakkula and East Cameron during fall (PERMANOVA: P < 0.05); these differences were attributable to the smaller proportion of fish and the greater proportion of shrimp consumed at East Cameron (Figure 3). No statistical differences were found between sites during winter, but the largest biomass of tunicates was consumed at Jakkula. Raw %DW values for Figure 3 are included in Table S.4.

TABLE 4. Permutational ANOVA (PERMANOVA) model testing the effects of reef site (Jakkula, Bright, and East Cameron), Red Snapper size-class (size-class 3 = 300–399 mm FL; size-class 4 = 400–499 mm FL; size-class 5 = 500–599 mm FL; size-class 6 = >600 mm FL), and season (spring, summer, fall, and winter) on the Red Snapper diet.

FIGURE 3. Percent contributions (bubbles) by dry weight of the seven prey categories in the diets of Red Snapper collected at two natural reef sites (Jakkula and Bright) and one artificial reef area (East Cameron) during spring, summer, fall, and winter (inverts = invertebrates).

Site, season, and the site × season interaction were significant contributors to the variation in Red Snapper δ¹³C and δ¹⁵N values (MANCOVA: P < 0.05; ). Fork length was not a significant covariate for δ¹³C and δ¹⁵N values (MANCOVA: P < 0.05; ). Fish from East Cameron exhibited consistently lower mean δ¹³C values across seasons, which were significantly different from the δ¹³C of fish at both Bright and Jakkula in spring, McGrail in summer and fall, and Bright in winter (MANCOVA: P < 0.05; Figure 4). Red Snapper from Bright exhibited significantly lower mean δ¹⁵N values relative to the other sites in every season except spring, when fish from Bright and McGrail did not differ (MANCOVA: P < 0.05; Figure 4). The differences observed between least-squares mean δ¹³C values, however, were very small: the maximum range of δ¹³C values observed in a given season was just 0.8‰ in spring (Figure 4).

TABLE 5. Results of the multivariate analysis of covariance (MANCOVA) model testing for the effects of reef site (natural: Jakkula, McGrail, and Bright; artificial: East Cameron) and season (spring, summer, fall, and winter) on Red Snapper δ¹³C and δ¹⁵N values, with FL as a covariate. The reported F-value and P-value correspond to Wilks' lambda statistic.

FIGURE 4. Scatter plots of white muscle δ¹⁵N (‰) values at a given δ¹³C (‰) value for individual Red Snapper collected at three natural reef sites (Jakkula, McGrail, and Bright) and one artifical reef area (East Cameron) during each season.

The only site with fish that exhibited significant intrasite δ¹³C or δ¹⁵N value differences among seasons was East Cameron. Pairwise comparisons indicated that fish at East Cameron exhibited significantly lower δ¹⁵C values during the spring relative to fall and summer (MANCOVA: P < 0.05). All pairwise comparisons for the MANCOVA model are included in Table S.5.

Seasonal Size Differences

The season × size-class interaction was also a significant contributor to the variation in Red Snapper diets (PERMANOVA: P < 0.05; ). During spring, the diets differed between size-class 3 and size-classes 4, 5, and 6 (PERMANOVA: P < 0.05). Fish belonging to size-class 3 consumed the largest proportion of zooplankton during spring, while fish in size-class 6 consumed very little zooplankton and the largest proportion of fish (Figure 5). During summer, the diets differed between size-class 3 and size-class 6 (PERMANOVA: P < 0.05) due to the large consumption of crabs by size-class 3 and the large consumption of incidental items by size-class 6 (Figure 5). During fall, the diets differed between size-class 5 and size-class 6 (PERMANOVA: P < 0.05) due to the greater proportion of shrimp and the smaller proportions of zooplankton and fish consumed by size-class 6 (Figure 5). There were no differences during winter, as high proportions of tunicates were present in the diets of all size-classes. Raw %DW values from Figure 5 are included in Table S.6.

FIGURE 5. Percent contributions (bubbles) by dry weight of the seven prey categories in the diets of Red Snapper in size-class 3 (300–399 mm FL), size-class 4 (400–499 mm FL), size-class 5 (500–599 mm FL), and size-class 6 (>600 mm FL) collected from northwestern Gulf of Mexico reefs during each season (inverts = invertebrates).

Although the extent of isotopic variation by size-class was examined within seasons, size-classes were not evenly distributed across seasons, and the limited number of replicates precluded SEA _c calculations or statistical comparisons. From a strictly qualitative perspective, larger size-classes showed higher isotopic variation in every season (Figure 6). Variability was more pronounced with regard to δ¹⁵N, whereas δ¹³C tended to be similar between size-classes during a given season (Figure 6).

FIGURE 6. Scatter plots of white muscle δ¹⁵N (‰) values at a given δ¹³C (‰) value for individual Red Snapper in size-class 3 (300–399 mm FL), size-class 4 (400–499 mm FL), size-class 5 (500–599 mm FL), and size-class 6 (>600 mm FL) collected from northwestern Gulf of Mexico reefs during each season.

DISCUSSION

Red Snapper diets, isotopic niches, and feeding patterns varied between natural and artificial reefs in the northwestern Gulf. Distinct isotopic niches for Red Snapper were found among natural reefs, whereas diet did not vary between natural reefs. Differences in Red Snapper isotopic niches between natural reefs indicated that individuals may remain in residence on a specific reef for an extended period. The discrepancy between diets and isotopic niches also suggests that if a larger sample size of stomachs with identifiable prey had been available for Jakkula and McGrail, differences might have been detected among diets at the natural reefs. Red Snapper diets were found to differ between the East and West Flower Garden Banks, which are complex natural reefs located along the same reef tract as the natural reefs sampled in the present study (Nelson 1988). The present findings, coupled with the observations from Nelson (1988), indicate that further investigation is warranted to compare the feeding ecology of Red Snapper among large, complex natural reefs, particularly those located in the northwestern Gulf.

The consumption of different prey species between natural and artificial reefs likely contributed to the variability in Red Snapper isotope niches between these two habitats. Because phytoplankton have lower δ¹³C values than benthic algae (Moncreiff and Sullivand 2001; Fry 2006), lower δ¹³C values imply a greater planktonic contribution. The lower δ¹³C values observed for Red Snapper at artificial reefs correspond to a greater biomass of zooplankton consumed at this habitat, especially during spring. Red Snapper have also been observed to feed on lower-trophic-level prey at artificial reefs compared to natural reefs (Davis et al. 2015). This predominance of lower-trophic-level prey at artificial reefs, however, also leads to the expectation that Red Snapper will exhibit lower δ¹⁵N values, yet fish from the artificial reefs tended to have higher δ¹⁵N values at a given δ¹³C value than fish from the natural reefs. This result contrasts with other studies that have reported little difference in Red Snapper δ¹⁵N and δ¹³C values between natural and artificial reefs (Simonsen et al. 2015; Tarnecki and Patterson 2015). The lack of difference in Red Snapper δ¹⁵N and δ¹³C values observed by Simonsen et al. (2015) and Tarnecki and Patterson (2015) may be due to the similar diets they found for Red Snapper between natural and artificial reefs. Basal resources were also similar between the natural and artificial reefs sampled by Simonsen et al. (2015), so it is possible that the easternmost banks on the Louisiana shelf edge reef track, which were in closer proximity to the artificial reef sites sampled in that study, have similar isotopic baseline differences.

Given the discrepancies, isotopic baseline differences between study sites must also be considered as important determinants of the relative positioning of Red Snapper isotopic niche spaces. Their location in waters closer to shore makes the artificial reef sites in the present study more susceptible to influences from the Mississippi River plume. Previous studies have shown that high δ¹⁵N values are associated with the river water itself, particulate organic matter, entrained plankton, and shrimp associated with the plume (Wissel and Fry 2005; Dorado et al. 2012). Additionally, increases in deepwater coral δ¹⁵N values have been attributed to increased nutrient loading of river waters, demonstrating both the extent of river influence and its effect on N isotopic compositions (Williams et al. 2007; Prouty et al. 2014). If this enriched N source was incorporated into the local food web at East Cameron, then consumers such as Red Snapper would be expected to show higher δ¹⁵N values relative to individuals from areas with a lesser riverine influence, which was observed in the present study.

Our results support previous findings that (1) Red Snapper at natural reefs feed on prey items occurring on and above the reef (Camber 1955); and (2) Red Snapper are not gaining large amounts of nutrition derived from artificial reefs (Gallaway et al. 1981; McCawley et al. 2007; Simonsen et al. 2015). Hard-substrate-associated prey items consumed by Red Snapper were reported to have the highest average caloric density among all prey types, followed by water-column- and soft-sediment-associated prey types (McCawley et al. 2007). Assimilation efficiency of fish prey was reported to be greater than that of invertebrate prey for Walleyes Sander vitreus (Kelso 1972). If the hard-substrate prey items in the present study also have the highest caloric densities, and if the higher amount of fish and lower number of invertebrates consumed at the natural reefs lead to greater assimilation efficiencies, then Red Snapper diets at the natural reefs are more calorically rich and provide prey with a higher assimilation efficiency than diets and prey at the artificial reefs.

Seasonal and size-specific diets were present for Red Snapper in this study. The greatest biomass of zooplankton was consumed in spring at artificial reefs and during spring for all Red Snapper size-classes. Red Snapper have previously been observed to consume large quantities of zooplankton at artificial reefs during spring (McCawley and Cowan 2007). The highest values of primary production on the Louisiana continental shelf have been observed in spring and at the beginning of summer (April–July; Lohrenz et al. 1997). Red Snapper feed opportunistically (Moseley 1966; McCawley et al. 2007); therefore, the large amounts of zooplankton consumed in spring could have been due to higher abundances from increased primary production.

Further investigation was carried out to examine trends in the diets of Red Snapper size-classes at the artificial reef area because the size-class 3 fish used in our analyses were only collected at the artificial reefs. This is not surprising, as artificial reefs in the Gulf are heavily used by young Red Snapper (ages 2 and 3) relative to natural reefs (Gallaway et al. 2009; Karnauskas et al. 2017). Consumption of zooplankton in spring was high for size-classes 3, 4, and 5 but not for size-class 6 (Table S.7). Red Snapper on platforms in the western Gulf are most abundant in depths over 60 m (Reynolds 2015), and smaller Red Snapper are located higher in the water column and have a greater affinity for the platform structure than larger individuals (Render 1995; E. M. Reynolds, National Oceanic and Atmospheric Administration, unpublished data). Red Snapper were located throughout the entire water column during late spring on a small, man-made artificial reef in the eastern Gulf (Williams-Grove and Szedlmayer 2017). Staying close to the structure and utilizing the water column might put smaller Red Snapper in greater contact with abundant zooplankton during spring.

Larger Red Snapper have been observed to show less affinity toward platforms (Render 1995); therefore, they have the ability to forage throughout the water column and along the bottom and may be less likely to forage on zooplankton. A small contribution of zooplankton (%DW) to the diets of Red Snapper larger than 600 mm was found at artificial reefs in the eastern Gulf (McCawley et al. 2007). Large predators were observed to be highly abundant in depths of 30–60 m on platforms in the western Gulf (Reynolds 2015), which could influence Red Snapper size structure on platforms. The observations of size distributions on platforms are only qualitative; thus, more research on how different sizes of Red Snapper vertically distribute themselves on platforms, and whether distributions are also impacted by the presence of large predators, would help to clarify foraging behavior. Conversely, zooplankton were almost absent from Red Snapper diets at both natural reefs (Bright and Jakkula), as %DW was less than 3.5% for all size-classes in all seasons (Tables S.8, S.9). Because natural reefs on the Louisiana shelf edge have a greater horizontal area than artificial reefs (Rooker et al. 1997), Red Snapper are able to distribute horizontally at the natural reefs instead of vertically (Langland 2015), which could explain why they foraged on prey types found on the reef, with less contribution by zooplankton.

Habitat-specific availability of prey species is likely the main factor driving the differences between Red Snapper diets at the natural and artificial reefs in this study. Invertebrate and fish communities have been found to vary with habitat and substrate type (Jackson and Jones 1999; Valesini et al. 2004; Wildsmith et al. 2005) and depth (Wenner et al. 1983; Williams and Hatcher 1983; Hyndes et al. 1999; Zimmer et al. 2000; Harris et al. 2001). The substrate on the natural reefs in this study was dominated by fine to coarse sands and coral reef patches (Gardner and Beaudoin 2005), whereas the artificial reef area was located on top of lithified or partially lithified delta mud (J. H. Cowan Jr., unpublished data). The high percentage of hard-substrate-associated prey items in Red Snapper diets at the natural reefs and the high percentage of soft-sediment-associated prey items in the diets at the artificial reef site reflect the substrate type of the habitat occupied by individual Red Snapper. Distinct fish assemblages at each habitat were also observed with concurrent video surveys conducted during this study, corroborating the idea that these two habitats supported different prey species (Langland 2015). Although the surrounding depths varied between the sampled natural and artificial reefs, Red Snapper were collected at the crests of the natural reefs, which are of similar depths as the artificial reefs; thus, depth is unlikely to explain the observed diet differences.

The disagreement between studies on Red Snapper diets can be attributed to differences in underlying substrate type, distance, and depth, and possibly to artificial reef structure. Red Snapper diets have not been found to differ between natural and artificial reefs when underlying substrates are alike (Tarnecki and Patterson 2015; Downey 2016; substrates estimated by B. D. Schwartzkopf using data from usSEABED [Buczkowski et al. 2006] and from Dufrene 2005 and M. K. Streich, Harte Research Institute for Gulf of Mexico Studies, personal communication), when sampling depths are similar (Simonsen et al. 2015; Downey 2016), and when distances between reefs are short (Downey 2016). In contrast to Tarnecki and Patterson (2015), differences were detected between Red Snapper diets at natural and artificial reefs in the northeastern Gulf when substrates were likely similar (Davis et al. 2015). There appear to be greater distances between sites, with more sites possibly located in deeper waters in the Davis et al. (2015) study compared to the Tarnecki and Patterson (2015) study, which may explain the differing results. It is also possible that the substrate at the easternmost sites in deeper waters studied by Davis et al. (2015) is mud (usSEABED; Buczkowski et al. 2006). The bulk of artificial reefs in the northeastern Gulf consist of sunken ships, planes, cars, dry docks, tanks, and small, man-made items, such as concrete balls and pyramids (Minton and Heath 1998); thus, it is also possible that Tarnecki and Patterson (2015) and Davis et al. (2015) sampled diverse types of artificial reefs, which may source different prey items. Future studies should describe the underlying substrates at sampling sites and address what types of artificial reefs are sampled so that better comparisons can be made between studies.

Because only one artificial reef planning area was sampled during this study, we acknowledge a lack of replication. To evaluate differences between more complex natural reefs on the Louisiana shelf edge, platforms closer to the banks should be sampled. For example, the High Island A-389A (HI-A389A) platform is located in the Flower Garden Banks National Marine Sanctuary and is situated on top of similar substrates as the natural reefs in this area (Scanlon et al. 2005). Reef fish assemblages were observed to be similar between the East and West Flower Garden Banks and HI-A389A (Rooker et al. 1997), whereas a low diversity of crab and shrimp species was found on HI-A389A compared to both the East and West Flower Garden Banks (Wicksten 2005). It is possible that Red Snapper consume similar fish prey between artificial and natural reefs in this area while differing in their consumption of invertebrate species.

Future studies should sample additional artificial reef sites of both sand and mud substrates and various distances away from the natural reefs on the Louisiana shelf edge (such as HI-A389A) to clarify whether substrate differences or distance between natural and artificial reefs could be driving diet variability. Future isotopic studies of Red Snapper would benefit from baseline source sampling to elucidate the relative importance of various sources and determine whether the trophic level is similar between habitat types or Gulf regions. Larger sample sizes for Red Snapper stomachs from individual natural reefs would also help to determine whether the prey species consumed affect isotopic niche variations. Unidentifiable material did represent a significant portion of the diets at both habitats. Techniques for DNA barcoding have proven successful at identifying previously unidentifiable fish in Red Snapper diets (Foss 2016) and would be beneficial for use in future studies to further discern differences in consumed fish species.

The quality of habitats underlying reefs is important, as Red Snapper diets reflect the substrate on which they live. Increasing habitat complexity and percent hard substrate have been shown to increase fish abundance and diversity (Gratwicke and Speight 2005). Given the high prey diversity observed in diets at the natural reefs, the large areas and complex habitats of natural reefs may provide enough space to support higher abundances and diversity of prey items relative to less-complex artificial reefs on mud substrates. Large, complex natural reefs in the northwestern Gulf therefore offer a higher habitat quality in the form of prey resources for Red Snapper in comparison with artificial reefs located on mud substrate. Concurrent studies found that Red Snapper energy reserves and reproductive potential were significantly greater at the natural reefs compared to the artificial reefs in the northwestern Gulf (Glenn et al. 2017; Schwartzkopf and Cowan 2017), and this could have resulted from more favorable feeding conditions. The high-quality habitat at the natural reefs on the Louisiana shelf edge, coupled with observed high energy reserves and reproductive potential of Red Snapper, should be considered during decisions on habitat protection policies and regulations. With many variables impacting Red Snapper feeding ecology in the Gulf, decisions about future artificial reef placement should involve careful consideration of factors such as the underlying substrate.

wentznuthe1971.blogspot.com

Source: https://www.tandfonline.com/doi/full/10.1080/19425120.2017.1347117

Feeding Ecology of Fishes Associated With Artificial Reefs in the Northwest Gulf of Mexico

METHODS

Study area

Published online:

Sampling

Laboratory procedures

Statistical analysis

RESULTS

Sampling Summary

Published online:

General Site Differences

Published online:

Published online:

Published online:

Seasonal Site Differences

Published online:

Published online:

Published online:

Published online:

Seasonal Size Differences

Published online:

Published online:

DISCUSSION

0 Response to "Feeding Ecology of Fishes Associated With Artificial Reefs in the Northwest Gulf of Mexico"

ارسال یک نظر

Iklan Atas Artikel

Iklan Tengah Artikel 1

Iklan Tengah Artikel 2

Iklan Bawah Artikel