Date Published: March 14, 2019
Publisher: Public Library of Science
Author(s): Michael A. Dance, Jay R. Rooker, Heather M. Patterson.
Habitat shifts that occur during the life cycles of marine fishes influence population connectivity and structure. A generalized additive modeling approach was used to characterize relationships between environmental variables and the relative abundance of red snapper Lutjanus campechanus over unconsolidated substrate on the continental shelf (<150 m) of the U.S. Gulf of Mexico (GoM) at three different life stages: juvenile (age-0, <125 mm FL), sub-adult (age-1-2, 125–300 mm FL), and adult (age-2+, >300 mm FL). Fisheries independent data (2008–2014) were used to develop separate models for both the eastern and western GoM, and final models were used to predict the relative availability of suitable habitat for each life stage across the two regions. Predictor variables included in final models varied by age class and region, with depth, dissolved oxygen, longitude, and distance to artificial structure common to most models. Depth was among the most influential variables in all models, and preferred depth increased with increasing size/age. Regional differences in fish-habitat relationships were also observed, as relative abundance of larger red snapper over unconsolidated substrates was more closely linked to artificial structure in the eastern GoM. The location of predicted high quality habitat for juvenile red snapper was greatest on the inner Texas shelf and a smaller area east of the Mississippi River Delta, suggesting these two areas may represent important nursery grounds for the respective regions. Clear ontogenetic shifts in the spatial distribution of predicted high quality habitat were evident in both the eastern (expansion from west to east with age) and western (shift from inshore to offshore) GoM. Given the unique population dynamics between the eastern and western GoM, improving our understanding of spatial and temporal variability in habitat quality may be important to maintaining connectivity between juvenile and adult habitats, and may enhance recovery and management of red snapper stocks in the GoM.
Habitat use by aquatic organisms often reflects a series or suite of behavioral decisions to maximize fitness and ultimately survival to the next age class [1–3]. In the most basic sense, these decisions are based on maximizing growth while minimizing mortality [2, 4, 5]. Thus, habitat selection is inherently affected by ontogeny, as increases in body size influence both the resources an animal is able to exploit and the associated predation risk involved [1, 6, 7]. In response, multiple habitats or regions are often needed for an animal to complete its life cycle [2, 3, 8, 9], creating age-structured or size-structured populations that are spatially segregated . This spatial segregation complicates our ability to identify and conserve essential habitats, as the importance of habitat patches, landscapes (seascapes), and/or regions to each life stage may vary [4, 11, 12].
Fishery independent catch data for red snapper were obtained from trawl surveys conducted over unconsolidated substrates in shelf waters (< 150 m) of the GoM during both summer (June-July) and fall (September-October) as part of the Southeast Area Monitoring and Assessment Program (SEAMAP) from 2008 to 2014. This survey uses a stratified random sampling design to select locations based on depth and shrimp statistical zones . The SEAMAP sampling protocol for trawl surveys systematically avoids untrawlable habitat (hard bottom habitats and artificial reef sites) to avoid hang ups, and selected sampling stations that occur in such habitat are relocated to the nearest trawlable location (up to 1.8 km from such structure) . As a result, SEAMAP trawl surveys primarily targeted unconsolidated substrates (e.g. non-hard bottom sediments including gravel, sand, mud, coralgal, marl, and shell). A 12.2 m otter trawl was towed at each station for approximately 30 minutes , and the linear distance of each trawl tow was then used to calculate the area swept at each station as a measure of sampling effort. Prior to 2008, SEAMAP surveys were weighted more heavily from the Texas-Mexico border to the Florida-Alabama border, resulting in a far greater number of samples in the western versus the eastern GoM; thus, the time period of 2008–2014 was chosen for this study to better account for spatial variability in the abundance of red snapper in each region. Separate models for the eastern and western GoM were developed for each age class, and deviance explained ranged from 34.7–54.3% with the highest for juvenile stage models in both regions (Table 1). Deviance explained decreased with age from juvenile (49.1%) to adult (34.7%) for eastern GoM models. In contrast, the deviance explained was 54.3% for the juvenile model in the western GoM, and the adult model (40.5%) explained a greater percent of the deviance than the sub-adult model (34.9%). Predictor variables included in final models varied widely with only depth retained in all models. However, several variables were common to at least 5 of the 6 models including dissolved oxygen, longitude, and distance to artificial structure. The relative importance of each predictor variable varied by both life stage and region, and models from each region are presented by life stage below. Depth is an important driver of reef fish community structure [41, 42] and was among the most influential variables of red snapper relative abundance across all age classes and regions. Many tropical reef fish species (e.g. snappers, groupers, grunts) undergo ontogenetic migrations from nearshore nurseries (inner shelf) to offshore spawning areas (outer shelf) . Similarly, juvenile red snapper inhabit shallow water habitats, albeit these nurseries occur over sand/mud/shell substrates on the inner shelf [23, 43] as opposed to seagrass and mangrove habitats commonly used by snappers in tropical nurseries. These low-relief habitats are hypothesized to provide structure without the high densities of predators common at larger offshore reef habitats , suggesting shallow habitats may reduce predation and increase survival during vulnerable early life stages until individuals attain sufficient size to shift to larger structure in deeper water on the mid to outer continental shelf . The observed progression to deeper benthic habitat with age is similar to dispersal patterns described for other large reef fish in subtropical systems (e.g. gag grouper; Mycteroperca microlepis) , suggesting that cross-shelf shifts during ontogeny are not unique to red snapper. Although it is also possible that observed shifts in spatial distribution and abundance may be due to differential survival due to higher fishing mortality in shallower nearshore waters , movement across the shelf during ontogeny is well documented in red snapper and observed depth patterns in both regions are consistent with known shifts with age to habitats of increasing complexity. Interestingly, we found that adult red snapper in the eastern GoM (where fishing pressure is greater) were associated with shallower depths (20–60 m) than those in the western GoM, where adults were most abundant at the shelf edge (100–150 m), suggesting the inshore-offshore shift with age may be less pronounced in the eastern GoM. This notion is consistent with recent findings by Powers et al.  off the Alabama coast and may reflect the differences in the spatial distribution of reef habitat (both artificial and natural) between the eastern and western GoM, as the majority of artificial structures in the eastern GoM are located in relatively high densities over the mid shelf in 20–60 m of water off Mississippi, Alabama, and north Florida, and age-2+ red snapper are often abundant on or near these structures [26, 49, 50]. In contrast, artificial structures in the western GoM are more widely dispersed across the shelf, and natural reefs are primarily located in deeper water (100–150 m) near the shelf edge where numerous natural banks hold high biomass of large reef fish, including red snapper [50, 51]. Source: http://doi.org/10.1371/journal.pone.0213506