http://orcid.org/0000-0003-3788-8528
Figures
Abstract
The Biscayne Bay Coastal Wetlands (BBCW) project of the Comprehensive Everglades Restoration Plan (CERP) aims to reduce point-source freshwater discharges and spread freshwater flow along the mainland shoreline of southern Biscayne Bay. These actions will be taken to approximate conditions in the coastal wetlands and bay that existed prior to construction of canals and water control structures. An increase in pink shrimp (Farfantepenaeus duorarum) density to ≥ 2 individuals m-2 during the wet season (i.e., August-October) along the mainland shoreline was previously proposed as an indication of BBCW success. This study examined pre-BBCW baseline densities and compared them with the proposed target. Densities were monitored by seasonal (wet, dry) throw-trapping (1 m2 replicated in triplicate) at 47 sites along ~22 km of the southwestern Biscayne Bay coastline over 10 years (2007–2016). Densities varied across years and were most often higher in dry seasons. Quantile regression revealed density limitation by four habitat attributes: water temperature (°C), depth (m), salinity (ppt), and submerged aquatic vegetation (SAV: % cover). Procrustean analyses that tested for concordance between the spatial and temporal distributions of shrimp densities and habitat metrics found that water temperature, water depth, and salinity explained ~ 28%, 28%, and 22% of density variability, respectively. No significant relationship with SAV was observed. Hierarchical clustering was used to identify spatially and temporally similar groupings of pink shrimp densities by sites or season-years. Significant groupings were then investigated with respect to potentially limiting habitat attributes. Six site and four year-season clusters were identified. Although habitat attributes significantly differed among spatial clusters, within-cluster median pink shrimp densities did not correlate with within-cluster minima, maxima, medians, or standard deviations of habitat attributes. Overall, pink shrimp density ( = 0.86, SD = 1.32 shrimp m-2) was significantly lower (t(α = 0.10,2),939 = -26.53, P <0.0001) than the 2 shrimp m-2 CERP Interim Goal target. Pink shrimp density corresponded significantly with salinity and appeared limited to density < 2 shrimp m-2 by salinity < ~18 ppt. Salinity is an environmental attribute that will be directly influenced by CERP implementation.
Citation: Zink IC, Browder JA, Lirman D, Serafy JE (2018) Pink shrimp Farfantepenaeus duorarum spatiotemporal abundance trends along an urban, subtropical shoreline slated for restoration. PLoS ONE 13(11): e0198539. https://doi.org/10.1371/journal.pone.0198539
Editor: Suzannah Rutherford, Fred Hutchinson Cancer Research Center, UNITED STATES
Received: May 18, 2018; Accepted: September 28, 2018; Published: November 7, 2018
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All excel format files are available from the Dryad database (accession number https://doi.org/10.5061/dryad.f861k8s).
Funding: Various work orders to the National Marine Fisheries Service Southeast Fisheries Science Center from the US Army Corps of Engineers, Jacksonville, Florida, to Joan Browder and Joseph Serafy helped support this work. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Biscayne Bay is a coastal lagoon located adjacent to the city of Miami, Florida, USA. Its watershed was heavily modified during the 20th century and is currently highly managed to prevent urban, suburban, and agricultural flooding while also meeting agricultural, commercial, and residential freshwater demands. The Comprehensive Everglades Restoration Plan (CERP) seeks to restore the quality, quantity, timing, and distribution of freshwater deliveries to southern Florida nearshore areas [1], including Biscayne Bay. The Biscayne Bay Coastal Wetlands (BBCW) project, a CERP component, is designed to restore a more natural hydrology and salinity regime along the bay’s southwestern shoreline [2,3]. Three actions are needed to make this improvement: (1) increasing the total volume of freshwater deliveries; (2) diverting part of point-source freshwater discharge (i.e., canal discharge) into coastal wetlands to enter the bay as overland sheet flow; and (3) altering the present timing of deliveries by extending discharges through the wet season (May-October) and into the dry season (November-April) [3,4]. The REstoration COordination and VERification (RECOVER) team established Interim Goals (IGs) to link ecological indicator metrics to CERP activities and thus evaluate restoration performance and realization of post-implementation ecological benefits at 5-yr intervals [1].
The pink shrimp Farfantepenaeus duorarum is one of many ecological indicators selected to assess ecological impacts of CERP implementation [1,5]. Pink shrimp was selected to assess effects of salinity change because previous work suggested linkage to salinity [1,5]. As reviewed by Zink et al. [6], higher pink shrimp productivity can be expected at salinities within polyhaline (18–30 ppt: [7]) and euhaline (30–40 ppt) ranges. Expansion of southwestern Biscayne Bay estuarine habitat was expected to benefit pink shrimp residing there [8]. Increased pink shrimp abundance might result indirectly from increased seagrass cover and spatial extent [5,8] that might occur from reduction in stress from extreme salinity fluctuations. Higher abundance of pink shrimp has been reported in areas exhibiting higher and more stable salinity [8–10], which also exhibit more continuous seagrass cover [11]. Areal expansion of shoal grass (Halodule wrightii) cover could further amplify pink shrimp abundance due to an apparent affinity for this seagrass species [12]. The stated pink shrimp IG for southwestern Biscayne Bay is “2 shrimp m-2 in nearshore optimal habitat (i.e., seagrasses)” during August-October peak abundance periods [1]. This IG was based upon a peak density of ~1.8 shrimp m-2 observed in September during a 2 yr (2003–2004) pilot study [8].
Historically, most freshwater delivery to Biscayne Bay was through transverse glades, broad natural channels through the Miami Coastal Ridge that allowed Everglades Basin surface water drainage [13,14] and groundwater seepage [15–20]. These natural drainage features fed fresh water from the Everglades through the coastal ridge via transverse glades into creek networks that spread surface water flows along the bay’s shoreline. Canalization converted the freshwater delivery system to one dominated by pulsed point-source (i.e., canal mouth) discharges that altered benthic submerged aquatic vegetation (SAV), infaunal, epifaunal, and nekton communities [11,21–26] and lowered the water table, reducing groundwater seepage [19,20].
BBCW salinity goals for southwestern Biscayne Bay (Shoal Point to Turkey Point: Fig 1) prescribe oligohaline (0.5–5 ppt) and mesohaline (5–18 ppt) regimes at the shoreline, trending towards 20 ppt (polyhaline, 18–30 ppt) 500 m from the coast [4]. These conditions are anticipated to enrich estuarine faunal assemblages as well as increase estuarine species distributions and abundances [4,27,28]. Expansion of continuous submerged aquatic vegetation (SAV) habitats dominated by Halodule wrightii, a species commonly associated with low and variable salinity, is foreseen [11,24–26,29,30]. BBCW implementation goals for benthic habitat include increased spatial extent of nearshore seagrass beds, especially seaward expansion of H. wrightii [30]. Increased overlap of optimal salinity conditions with preferred benthic SAV habitats would provide synergistic benefits to estuarine fauna such as pink shrimp [4,5,27,31].
The second panel depicts the same sites color-coded to match significant site clusters (Figs 2B and 4). Cluster 6 (red) exhibited similar spatial density patterns that were, on average, significantly higher than all other cluster groups. Conversely, Cluster 1 (light blue) exhibited significant, consistently lower densities than all other groups except for Cluster 4 (dark green). Clusters 2–5 exhibited intermediate average densities (Fig 4).
The purpose of this study was to investigate spatiotemporal trends in pink shrimp density along the southwestern Biscayne Bay shoreline. We investigated the plausibility of the post-CERP establishment of ≥ 2 shrimp m-2 IG. Further, we addressed presumptions that (1) pink shrimp peak abundance occurs during the wet season and (2) nearshore mesohaline salinity goals would yield increased pink shrimp abundance in the nearshore zone. Pink shrimp density relationships to species-specific and total SAV % cover and canopy height were also investigated. Our focus was on evaluating temporal (i.e., seasonal and inter-annual) and spatial (i.e., among sites) pink shrimp density trends relative to habitat attributes. This was achieved by (1) using quantile regression to identify habitat attributes that potentially limit pink shrimp density, (2) organizing pink shrimp density and habitat observations via heatmaps to visually assess spatiotemporal variability and trends, (3) using Procrustean analysis to measure concordance between shrimp density and habitat attribute matrices, (4) employing hierarchical clustering analysis to identify spatiotemporal density clusters, and (5) investigating distributional aspects (median, minimum, maximum, and standard deviation) of habitat attribute values (temperature, salinity, water depth, and SAV % cover) within density clusters to link density patterns to the environment. These analyses employ data from wet and dry seasons of 10 years of epifaunal community sampling at 47 sites within 50 m of shore spanning ~22 km of shoreline.
Materials and methods
Study area
Biscayne Bay is a large (1,110 km2), shallow (depths generally < 3 m), subtropical lagoon system extending approximately 56 km north to south along the southeast coast of Florida, USA (Fig 1). Where coastal urban development is low, the bay’s mainland shoreline consists of mangrove-seagrass ecotone punctuated by natural tidal creeks, freshwater canals, and other channels [32]. Overland freshwater discharges and groundwater seepage create a salinity gradient across the bay perpendicular to the shoreline with three salinity zones: (1) western nearshore areas usually affording the lowest salinities; (2) the bay central axis marked by near oceanic salinities; and (3) oceanic salinities near the eastern passes through barrier islands to the open ocean [24,29,33]. Tidal ranges are generally 0.5 to 1 m [34,35].
Field surveys
Epifaunal communities and SAV habitats were surveyed seasonally at fixed sampling sites (n = 47) along the southwestern Biscayne Bay nearshore zone (0–50 m) from Shoal Point to Turkey Point (Fig 1). Surveys were conducted in public waters under authority of Biscayne National Park (Study #: BISC 06016, Permit #: BISC-2017-SCI0022). Surveys were conducted in dry (January-March sampling) and wet (July-September sampling) seasons that characterize south Florida’s climate. The primary sampling unit was the 20 m buffer around GPS coordinates that identified permanent sampling sites. Sites were located in the shallow, open water along the western shoreline mangrove-seagrass ecotone, an area likely to be directly affected by CERP implementation. During each survey, the 47 fixed sampling sites were visited within 3 hr of high tide over 5 to 7 days within a 2 week period. Water quality and habitat parameters, including water temperature (°C), salinity (ppt), pH, dissolved oxygen saturation (%), dissolved oxygen concentration (mg L-1), water depth (m), and sediment depth (m), were recorded at each site. This study analyzed farfantepenaid and associated chemical and physical sampling data collected from 2007 through 2016.
Benthic habitats were assessed for species-specific SAV % cover by visual assessment of 10 replicate 0.5 m2 quadrats per site [24,25]. In addition, canopy height (maximum seagrass blade length) was measured to provide a topography metric. Species-specific and total SAV % cover data following the methods of Lirman et al. [25] were obtained for the period 2008 to 2016.
Epifaunal communities were sub-sampled at each site (n = 3) using an open-ended, rigid-sided aluminum box (i.e., throw-trap) measuring 45 cm tall by 1 m2 [36,37]. Two 3-mm stretch-mesh cover nets affixed to opposite sides of the throw-trap upper surface prevented epifauna escape during deployment in water deeper than 45 cm. Once deployed, the throw-trap was cleared of trapped epifauna by sweeping (n = 4) its interior from alternating directions with a metal-framed seine fitted with 3 mm stretch-mesh, while gently tapping the substrate with the seine frame. Organisms collected from each sub-sample throw-trap deployment were bagged and numbered separately for storing and processing. Samples were frozen during storage until processing. No protected species were sampled.
Epifauna identification and measurement
Taxonomic identifications and size measurements were conducted in the laboratory. Organisms collected from each replicate throw-trap deployment at a site were maintained and processed independently. Where possible, carapace length (CL, mm) and total length (TL, mm) were recorded for each farfantepenaeid shrimp. Shrimps >8.0 mm CL were identified to species primarily using petasma and thelycum (i.e., sexual) morphology, although other characteristics were also used [38–41]. Shrimps <8.0 mm CL were identified to genus due to low degree of sexual morphological development [39].
Statistical analysis
All statistical analyses were performed using the R statistical package (The R Foundation, https://www.r-project.org/). Statistical analyses were performed with a Type 1 error criterion of α = 0.10 to reduce potential Type 2 errors. Shrimp density (#/ m2) at each site was calculated as the sum of observed shrimps from the triplicate throw-trap sub-samples divided by 3 m2. Density data were natural logarithm (x + 1) transformed before analysis to reduce influence of outliers.
Potential habitat limitations on pink shrimp density.
As a statistical interpretation of the ecological concept of Leibig’s Law of the Minimum [42,43], quantile regression (QR) has been applied to identify species distribution or abundance limitation by specific habitat attributes by focusing specifically on the upper bound of the abundance vs. habitat attribute relationship [44–47]. Pink shrimp density was first plotted against individual habitat factors to graphically assess potential limiting factors. QRs (function ‘rq’ of package ‘quantreg’) fit to the 0.5 and 0.9 density percentiles were used to statistically identify a subset of habitat attributes that suggested limitation at the median and upper edge of the density distribution. Analyses initially considered water temperature (°C), salinity (ppt), pH, dissolved oxygen saturation (%), dissolved oxygen concentration (mg/L), water depth (m), sediment depth (m), and the following SAV metrics: Thalassia testudinum % cover, H. wrightii % cover, total seagrass % cover, total SAV % cover, and total SAV canopy height. As in previous studies in the same region [24–26], Syringodium filiforme was rarely encountered (n = 10, 1.6% of total samples) and thus was not further considered. Ultimately, potential limiting variables were confined to the following four: temperature, salinity, water depth, and SAV % cover. Due to lack of SAV data for 2007, specific analyses that included SAV covered only the period from 2008 forward.
Multiple QR functional response shapes were investigated including linear, quadratic, cubic, log-linear, natural cubic splines (function ‘ns’ of package ‘splines’) [48], and additive quantile smoothing spline (AQSS) response curves (functions ‘rqss’ and ‘qss’ of package ‘quantreg’ [49,50]). Natural cubic splines were constructed with 3 (0.25, 0.50, and 0.75 quantiles of the predictor), 2 (0.33 and 0.66 quantiles of the predictor) and 1 (0.5 quantile of the predictor) internal knots [48]. QR coefficient confidence intervals were constructed and tested for significance by xy-pair bootstrapping (function ‘summary.rq’ of package ‘quantreg’). Natural cubic spline QRs were modeled without intercepts; these QRs were considered significant if each individual spline describing a sub-range of the data was significant.
Spatiotemporal relationships.
Heatmaps were generated to visualize spatiotemporal trends in pink shrimp density and the habitat attributes found by QR to potentially limit pink shrimp density. Observation data were converted to 47 row by 20 column matrices to display their spatial (47 sampling sites) and temporal (10 year x 2 seasons = 20 year-seasons) patterns, and color gradients were used to represent the magnitude of density.
Procrustean analyses allowed direct testing of statistical concordance between matrices of shrimp density and habitat attributes [51–53]. Procrustean analysis minimizes the residual sum of squares between a target matrix (X: here, shrimp density) and a second, fitted matrix (Y: here, habitat attributes), superimposed on it by scaling, rotating, and dilating [51,52]. The Procrustean Sum of Squares (PSS, also known as Gower’s Statistic: m2X,Y) represents the minimized residual sum of squares from the fitting procedure and is used to assess Procrustean fit ranging from 0 to 1, with higher values presenting poorer fit [51–53]. The PSS metric is equivalent to 1 –r2, where r is a Pearson correlation coefficient [52]. Because the method hinges on one-to-one relationships between the matrices being compared, Procrustean analysis [54,55] cannot accommodate missing values. Following Adams et al. [54] and Arbour and Brown [55], missing habitat attribute values were imputed with linear regressions that included site, season, and year as potential factors. PROTEST (function ‘protest,’ package ‘vegan’, permutation n = 9999) provided statistical significance of Procrustean fits between density and habitat attribute matrices [51].
Pink shrimp density and habitat attributes among density clusters.
Hierarchical clustering procedures were used to identify groups with similar density patterns among sites or year-seasons [56,57]. Bray-Curtis dissimilarity matrices were constructed (‘vegdist’ function, ‘vegan’ package) with respect to site (i.e., spatial) and year-season (i.e., temporal) density observations. Hierarchical agglomerative clustering (function ‘hclust’) using the “Ward.D2” agglomeration method identified spatially and temporally similar density groupings. A prioiri statistical significance of clusters was tested via similarity profiling (function ‘simprof’ of package ‘clustsig’) (permutations = 999, number of expected groups = 1000) of identified density cluster memberships [57]. Permutational multivariate ANOVA (PERMANOVA: function ‘adonis2’ of package ‘vegan’) testing provided a posteriori cluster significance [58,59]. PERMANOVA was also used to investigate inter-annual and seasonal density differences using year-season cluster membership as a categorical nesting factor. To investigate potential dispersion influences on PERMANOVA significance, multivariate homogeneity of dispersions analysis (function ‘betadisper’ of package ‘vegan’) was used to test for inter-cluster differences in dispersion (i.e., distance to centroid) [60]. The density heat map was rearranged to display site and year-season cluster membership.
Pink shrimp density and habitat attributes previously detected via QR as potentially limiting to pink shrimp density were investigated among site and year-season clusters. First, medians and confidence intervals (CI) of density and habitat attributes were computed for each site cluster and year-season cluster. Confidence intervals (CIs) about median values were computed as:
(1)
where IQR = interquantile ranges and n = sample size, as described in McGill et al. [61] and Chambers et al. [62]. Plots of density and habitat attributes’ median, CIs, minimum, and maximum values were used to visualize their distributions within site and year-season clusters. Density and habitat attributes were analyzed with respect to site or year-season clusters. Nonparametric tests were used because parametric normality and equality of variance assumptions were usually violated. Kruskal-Wallis tests were used to investigate differences in distribution shape and range (i.e., location: [63]) of density and habitat attributes among site or year-season clusters. Post-hoc Tukey-type nonparametric Conover multiple comparison tests (function ‘posthoc.tukey.conover.test’ of package ‘PMCMR’) were used to test for significant pairwise differences. These tests were implemented as χ2 distributions to correct for data ties, and p-values were Bonferroni-corrected [63,64]. A series of correlation analyses was used to identify habitat attribute distribution characteristics that associated with site or year-season cluster median densities. Pearson correlation analyses were applied to median, minimum, maximum, and standard deviation of habitat attributes within site or year-season clusters.
Results
A total of 3,178 farfantepenaeid shrimp specimens were collected. The distribution of shrimp sizes suggested a gear capture inefficiency for individuals <5mm CL; therefore, data only for shrimps ≥5 mm CL (2,417 shrimps) were retained for further analysis (S1 Fig). Of the retained shrimp, 1,573 (65.1%) were identified as pink shrimp, F. duorarum, and the remaining 844 (34.9%) were identified as farfantepenaeids due to difficulties with species identification of individuals <8 mm CL. Of the 1,937 individuals with measured CL, 1,931 individuals (79.9%) were considered juveniles (≤17.5 mm CL) and the remaining 36 individuals as subadults.
Pink shrimp density observations ranged from 0 to 13.0 shrimp m-2; 105 instances (11.2%, N = 940 samples) of densities ≥2 shrimp m-2 were observed, while no farfantepenaeid shrimps were observed in 377 samples (40.1%). Overall, shrimp density averaged 0.86 (SD = 1.32) shrimp m-2 and was significantly lower (t(α = 0.10,2),939 = -26.53, P <0.0001) than the 2 shrimp m-2 CERP Interim Goal threshold. Average density in any year-season was always < 2.0 shrimp m-2 (Table 1), although the highest year-season density (2014 Dry: 1.62 ± 2.02 shrimp m-2; Table 1) did not significantly differ from 2 shrimp m-2 (t(α = 0.10,2),46 = -1.30, p > 0.10). Averaged over all sites, mean dry season shrimp densities were higher than those of the subsequent wet season 50% of the time. Averaged over year-seasons, the highest mean site density was 2.15 (±1.95) shrimp m-2 at site 33 (Table 2). Seven sites (7, 10, 12, 33, 34, 43, and 44: 14.9%) exhibited temporally averaged densities that did not significantly differ from 2.0 shrimp m-2 (t(α = 0.10,2),19 = -0.080, -1.69, -0.30, 0.34, -1.63, -1.57, -0.32, respectively; p > 0.10). Most sites (n = 33, 70.2%) exhibited average densities below 1 shrimp m-2 (Table 2).
Temperatures ranged from 12.49 to 36.06°C. Average temperatures demonstrated a clear pattern of cooler (22.73 ± 2.65°C) and warmer (30.65 ± 1.53°C) values for dry and wet seasons, respectively (Table 1). The dry season record was punctuated by an extreme cold front event that occurred during the 2010 field sampling. No pattern of variation in average temperatures among sites was readily discernable (Table 2). Salinities ranged from 2.48 to 39.71 ppt; overall average salinity was 23.64 (± 7.05) ppt (Table 2). Spatially averaged wet season salinities were generally lower than those of dry seasons, although 2011, 2014, and 2015 wet seasons were notable exceptions with higher average salinity than both the preceding and following dry seasons (Table 1). Sampling sites’ mean salinity and standard deviation of salinity were negatively correlated (Pearson r = -0.63, t = -5.49, d.f. = 45, p < 0.0001; S2 Fig). Prior to epifaunal sampling during the 2011 and 2015 wet seasons, hypersaline (>40 ppt) conditions of considerable duration were recorded by continuous salinity loggers deployed within this sampling domain [65]. Only four temporally averaged site salinities were mesohaline, most (n = 42) were polyhaline, and one was euhaline (Table 2). Water depths ranged from 0.19 to 1.5 m and averaged 0.74 (± 0.20) overall (Table 2) with no appreciable trends among year-seasons or among sites. Total SAV % cover ranged from 4.57 to 100% and averaged 66.57% (± 21.97) with no clear year-season variation patterns (Tablea 1, 2). A planktonic microalgal bloom event was observed in parts of the Biscayne Bay coastal area during the 2013 wet season [65,66].
Habitat limitations on pink shrimp density
Of the multiple habitat attributes investigated, significant QR analysis results revealed that temperature (°C), salinity (ppt), water depth (m), and SAV (% cover) potentially limited pink shrimp density (Table 3, Fig 2). QR of density vs. temperature yielded a single-knot natural cubic spline relationship which was roughly dome-shaped and maximized at 26.6°C, with tails that tapered off at higher and lower temperatures (Table 3, Fig 2A). Temperatures between 21.08 and 31.33°C did not appear to limit pink shrimp densities to <2 shrimp m-2 (Table 3, Fig 2A). Although a series of functional shapes was considered for the QR density vs. salinity response curve, only the linear and log-linear responses were both significant and ecologically plausible [67]. The log-linear response, which suggested more severe density limitation below10 ppt and appeared more asymptotic at salinities above 10 ppt (Table 3, Fig 2B), seemed more plausible than the linear response. Salinities < ~18 ppt limited shrimp density to <2 shrimp m-2 (Table 3, Fig 2B). QR of pink shrimp density against water depth (m) yielded a 3 knot (0.25, 0.5, 0.75 quantile) splined bimodal relationship with steep increases in limitation below ~0.6 m and above ~ 1.0 m (Table 3, Fig 2C). Apparent constraint of density to <2 shrimp m-2 occurred at water depths less than 0.43 m and greater than 1.05 m (Table 3, Fig 2C). Shrimp density had a logarithmic linear relationship with SAV % cover (Table 3, Fig 2D). SAV cover less than 45% limited density to <2 shrimp m-2 (Table 3, Fig 2D). For the four habitat attributes, significant QRs were observed at the 0.9 quantile, but not at the 0.5 quantile (Table 3).
Pink shrimp density (shrimp m-2) and back-transformed 0.50 and 0.90 quantile regressions lines of predicted density (LN x+1) plotted against A) temperature (°C), B) salinity (ppt), C) water depth (m), and D) submerged aquatic vegetation (SAV: % cover). Predicted regression lines depict relationships reported in Table 2.
Spatiotemporal relationships
Heatmap visualization of pink shrimp spatiotemporal density trends revealed a general absence of pink shrimp from sites 13 to 28 (approximately Black Point to Fender Point, Fig 1) and sites 45 to 47 (near Turkey Point, Fig 1) across all year-seasons (Fig 3A). Within these groups of sites, only 16 (4.4%, N = 360) and 4 (6.7%, N = 60) instances of pink shrimp densities >2 shrimp m-2 were observed, respectively. Generally, higher densities were observed at sites 31 through 44 and sites 1 to 12, where 42 (15%, N = 280) and 40 (16.7%, N = 240) instances, respectively, of densities >2 shrimp m-2 were observed across all year-seasons. Densities were particularly low during the 2007, 2009, 2013, and 2014 wet seasons (< 0.5 shrimp m-2: Table 1), when shrimp were absent from a high proportion of samples (55.3, 51.1, 70.2, and 76.6%, respectively: Fig 3A). Other year-seasons (2008 dry, 2009 dry, 2012 dry, 2012 wet, 2014 dry, and 2015 wet: Fig 3A) exhibited relatively high average density (> 1 shrimp m-2), because the low and zero-catch observations from Black Point to Fender Point (Fig 1) were offset by higher density observations at most other sites (Fig 3A). Average densities in these six year-seasons yielded the highest year-season average densities, all >1 shrimp m-2 (Table 1).
Heatmaps depicting spatial (i.e., site) and temporal (i.e., year-season) patterns in A and B) pink shrimp density (shrimp m-2: LN[x+1]), C) temperature (°C), D) salinity (ppt), E) depth (m), and F) SAV (% cover). Shrimp density patterns in panel B are organized by site and year-season cluster membership, as denoted by color bars along the left and top margins and presented in the legend. Black cells in A) and B) highlight 0 shrimp m-2 observations while in C) through F) black bars represent missing values. Year-season label colors depict ecological perturbations: red = hypersalinity event, blue = cold snap, green = algal bloom. Labels on the left margin of (A) refer to canal outlets (blue) and coastline features (black) depicted in Fig 1.
Cluster analyses were performed to organize year-season elements and site-specific elements into groups according to shrimp density. Then, the heatmap of shrimp density (Fig 3A) was rearranged to reflect both the site and year-season cluster groupings (Fig 3B). The cluster groupings will be discussed in the following section. Heatmaps were developed to visualize spatiotemporal trends in temperature, salinity, water depth, and SAV (Fig 3C, 3D, 3E and 3F). Procrustean analyses revealed significant concordance of the shrimp density matrix (Fig 3A) with water depth, temperature, and salinity habitat attribute matrices (Fig 3C, 3D and 3E; Table 4) but not with the SAV matrix (Fig 3F, Table 4). Water depth and temperature exhibited the highest correlations, followed by salinity (Table 4). Each comparison yielded a high residual sum of squares (high m2X,Y values), indicating relatively weak explanatory power of individual habitat attributes (Table 4). Procrustean fitting procedures explained 28.3, 27.1, and 22.1% of the variability in density for water depth, temperature, and salinity, respectively.
Pink shrimp density and habitat attributes among density clusters
SIMPROF testing identified six significant site clusters and four significant year-season clusters (Fig 3B). PERMANOVA testing of cluster membership confirmed SIMPROF site (F5,41 = 4.765, p = 0.001, R2 = 0.368) and year-season (F7316 = 3.727, p = 0.001, R2 = 0.411) clustering. Two site clusters (i.e., 2 and 6: Fig 3B) together included most (66%, n = 31) of the sampling sites. One large year-season cluster included most year-seasons (75%, n = 15: Fig 3B). Smaller year-season clusters (80%, n = 4) were mostly comprised of wet seasons having pink shrimp densities <0.5 shrimp m-2. Substantial differences in shrimp densities of members of different clusters likely drove the significantly differing multivariate dispersion among site (F5,41 = 14.886, p < 0.0001) and year-season (F3,16 = 22.987, p = <0.0001) clusters. PERMANOVA testing detected significant season (F1,9 = 1.912, p = 0.0063), but not year (F9,9 = 1.020, p = 0.4279), categorical temporal effects. Multivariate dispersions differed greatly among years (F9,10 = 5.56*1029, p < 0.0001). Significant seasonal multivariate dispersion (F1,18 = 7.047, p = 0.0161) was also observed with greater variability in the wet season than in the dry season.
Significant differences in shrimp density distributions were detected among both site and year-season clusters (Table 5, Fig 4A, Fig 5A). Site clusters represented three relative median density levels: high (~0.7 shrimp m-2: site cluster 6); intermediate (~0.3 shrimp m-2: site clusters 2, 3, and 5); and low (0.0 and ~0.14 shrimp m-2: site clusters 1 and 4, respectively: Table 5, Fig 4A). Year-season clusters also exhibited three relative density levels: high (0.51 shrimp m-2: year-season cluster 2), intermediate (0.29 shrimp m-2: year-season cluster 4) and low density (0.0 shrimp m-2: year-season clusters 1 and 3: Table 6, Fig 5A). Significant differences among site clusters defined by differences in shrimp density were also observed for salinity, water depth, and SAV distributions (Table 5; Fig 4B, 4C and 4D), but temperature distributions did not differ among site clusters. All four habitat attributes exhibited significant differences among year-season clusters (Table 6, Fig 5B, 5C, 5D and 5E).
Median (± CI) and maximum, and minimum values of A) density (shrimp m-2: LN([x+1]), B) salinity (ppt), C) water depth (m), and D) submerged aquatic vegetation (SAV: % cover) in shrimp density site clusters. Point colors coincide with Figs 1 and 3B. Letters denote statistically similar groups. Horizontal line of A) depicts the 2-shrimp m-2 CERP Interim Goal.
Median (± CI) and maximum, and minimum values of A) density (shrimp m-2: LN([x+1]), B) temperature (°C), C) salinity (ppt), D) water depth (m), and E) submerged aquatic vegetation (SAV: % cover) relative to shrimp density year-season clusters. Point colors coincide with Fig 3B. Letters denote statistically similar groups.
Discussion
Analysis of 10 years of monitoring data revealed few instances (11.2%) of pink shrimp densities > 2 shrimp m-2, the IG for focal Biscayne Bay pink shrimp populations [1]. All but one spatially averaged year-season density and all but a few temporally averaged site densities in the 2007–2016 database were significantly below 2 shrimp m-2. CERP implementation is expected to result in more favorable salinity conditions for pink shrimp, leading to higher shrimp densities [1,5,27]. Reductions in extreme salinity variability in the southern half of the study area (i.e., Black Point to Convoy Point: Fig 1) could lead to average densities > 2 shrimp m-2 across the entire study spatial domain. However, our results suggest that ~10ppt salinity (i.e., low mesohaline to oligohaline: Fig 2B) was a threshold below which pink shrimp densities were severely limited. Above this threshold, the positive response to salinity continued more gradually. The shape of the response may have been affected by the lack of pink shrimp density observations at salinities >39.71 ppt in this study. Limitation of pink shrimp densities at salinities <10 ppt does not support CERP post-restoration IGs of >2 shrimp m-2 with reduction of Biscayne Bay nearshore salinity regimes to oligohaline and low mesohaline conditions. Pink shrimp densities reported here represent an underestimate of true density as a substantial number of shrimps (~25%) were removed from analysis due to concern about gear sampling inefficiency of shrimps <5 mm CL.
Spatial pink shrimp density patterns from Black Point to Convoy Point (Fig 1) were represented by membership in one low density site cluster (cluster 1) and one intermediate density site cluster (cluster 2). This zone is strongly influenced by canal discharges [11,23–26,33,68]. Both rapid (<60 min to 2 d) and extreme (~25 ppt) salinity reductions can occur along this stretch of coastline [22,24,25,29,69,70]. Such salinity fluctuations can alter fish community assemblages [23] and may affect foraging behavior and survival [22,23]. Pink shrimp may avoid these conditions as they have been reported to migrate to avoid large-volume riverine inflows [71]. Rapid salinity reductions of greater than 20 ppt cause near complete pink shrimp mortality in laboratory settings [72–74]. Low and intermediate density site clusters 1 and 2 included sites 45, 46, and 47 (Figs 3 and 4), which are located near Turkey Point (Fig 1), well south of the canal zone (Black Point to Convoy Point) and so were not impacted by low and variable salinity conditions. These sites had moderate minimum salinities (≥11.06 ppt), the highest temporally-averaged salinity across all sites (≥29.57 ppt: Table 2), and generally high SAV cover (Table 2, Fig 3F). The cause of their low shrimp density is unknown, but does not seem to be related to limitation due to salinity or SAV cover.
Site cluster 6, comprised of sites located away from canal mouths (Fig 1), had the highest median density and the second highest minimum salinity (9.62 ppt). Previous field [9,75] and modeling [35,76] studies describe the area represented by the more northern sites of this cluster as an area of relatively high shrimp abundance. These northern sites were situated immediately across the bay from a wide ocean inlet known as the Safety Valve. This inlet has been considered a primary postlarval immigration pathway [35,76] and may have contributed to high densities at northern sites [77,78]. Shrimp cumulative size frequency distributions differed between northern (sites 1–17) and southern (sites 18–47) sampling sites (S1 Fig). Comparison of these size distributions suggested that juvenile shrimps were more abundant in the north. This difference does not necessarily mean that greater recruitment was occurring in the north because differential growth and/or mortality also could have caused the size distribution difference. The inclusion of southerly sites (33, 34, 37, 40, 41, 43, and 44) within site cluster 6 (Figs 1 and 3B) contradicts the notion of southern recruitment limitation.
Some southern sites in high density cluster 6 are located near mangrove creeks that drain more natural watersheds. Other cluster 6 sites, both northern and southern, are located near, but not immediately adjacent to, canals with relatively small freshwater discharges (Military Canal: 1994–2003 annual mean canal output = 21.9 cfs; Cutler Drain C-100: 1994–2003 annual mean output = 46.1 cfs; [33]). However, sites immediately adjacent to the low-volume-discharge canals (i.e., 6 and 32) do not have high shrimp density and clustered with intermediate and low-density sites (clusters 1 and 2, respectively: Figs 1 and 3).
Most year-seasons (75%) were aggregated within one large cluster, indicating a general lack of inter-annual and inter-season variability in Biscayne Bay juvenile pink shrimp densities. However, year-season cluster 2, which had the highest median cluster shrimp density, consisted mainly (60%) of dry season sampling events. Smaller shrimp were associated with the higher dry season densities (S1 Fig), presumably because of higher dry season postlarval recruitment. Observation of higher densities in the dry rather than wet season is contrary to the pink shrimp IG, which focused on improving ‘peak’ fall (wet) season abundance [1]. The pink shrimp IG was based upon previous findings of a summer/fall (i.e., wet season) peak in abundance [8,9]. Others reported peak juvenile abundances in late fall/early winter (i.e., November/December) [12] or estimated maximal Biscayne Bay juvenile pink shrimp populations occurring in November (i.e., late fall) [75,79]. Differences in sampling gear and spatial domain and the short durations (≤2 yr) of the four pervious studies [8,9,75,79] complicate comparisons with the present study. Although of greater duration, the present study’s bi-annual sampling effort may have insufficient resolution to precisely identify the period of peak pink shrimp density or its year to year variation.
A lack of understanding of Biscayne Bay pink shrimp recruitment complicates study of their abundance patterns. The only study on recruitment reported a late fall through early winter (i.e., October through March: [80]) peak in recruitment, but was too short in duration (1 yr) to provide any information on inter-annual trends or consistency. This peak agreed with juvenile abundance studies reporting a late fall/early winter peak [75,79]. A model of pink shrimp postlarval recruitment from Tortugas and Marquesas spawning grounds found that oceanographic processes favored potential recruitment pathways up the Atlantic coast of the Florida Keys during late wet season and early dry season months [81]. Modeling of larval permit Trachinotus falcatus originating from spawning grounds near those of pink shrimp also found similar recruitment patterns for the Florida Keys and Biscayne Bay [82]. Oceanographic, coastal, and climatic conditions affect pink shrimp adult reproductive activity [83,84], and larval abundances [85] and interact with behavior to influence early life stage recruitment to nearshore areas [81,86–93]. Use of pink shrimp and other offshore spawning species as indicators of ecological conditions in their nearshore nursery grounds is complicated by life cycles affected by prior external conditions [5,32].
According to QR analysis, pink shrimp densities were significantly limited by four habitat attributes (water temperature, salinity, water depth, and SAV % cover). Two of these (i.e., salinity regime and SAV % cover) can be influenced by freshwater management within this study domain. Procrustean analysis confirmed the influence of each habitat attribute except SAV % cover. These habitat attributes vary at different time scales; for example, water depth can differ by as much as 1.3 m within 12 hr during extreme tidal cycles, whereas SAV % cover may integrate salinity, nutrients, water clarity, and other influential factors from 6 mo. to 1 yr or more. Responses on different time scales may affect the ability to discern relationships.
The correlation between spatiotemporal pink shrimp density and water depth was the strongest in this study. This was unexpected given the narrow spatial sampling domain along the mangrove-seagrass ecotone and the small range of variation in water depth in the data. Associations between nearshore pink shrimp abundance and depth have been previously reported [75,80,94,95]. Other studies that focused on very nearshore areas (<100 m from shore) also found higher abundances there [8–10,12,96]. Recruiting postlarval pink shrimp often concentrate in SAV near the low-tide mark along shorelines [9,12,97–102]. Other pink shrimp habitat investigations [80,94] found multiple habitat attributes (e.g., salinity, salinity standard deviation, standard deviation of turbidity, temperature, median sediment size, dissolved oxygen concentration, water depth, and benthic habitat characteristics) can influence pink shrimp abundance. As re-iterated by Zink et al. [6], Costello et al. [12] stated that “…factors other than salinity per se control abundance of the euryhaline juveniles…” The present results clearly indicate water depth is one of these factors.
Fluctuating tidal depths and/or reduced detection probability at greater depths likely contributed to the domed shape of the QR relationship [103]. Short-term variations in water depth are caused by astronomical tides and meterological/climatological events [104], which may override any effects of variation in freshwater inflow, especially in a relatively open bay area such as south-central Biscayne Bay. Hydrological alterations can both increase inundation [105] or decrease water depth via sediment accretion [106]. The multiple broad openings connecting Biscayne Bay to the open ocean promote dominance of water depth by astronomical tides [107,108]. Given the exposed nature of that part of Biscayne Bay’s coastline, changes in freshwater inflow within the plausible range that may occur along this coastline are unlikely to appreciably alter water depth within the domain. On longer time scales vertical movements of land surface and coastal geomorphology also can influence nearshore water depth and hydroperiod [104]. Sea level rise may also interact with hydrological changes to further alter coastal water levels and marsh inundation [109]. Although these processes are slow, sea level rise effects on water depth could influence nearshore pink shrimp density within the CERP time frame.
In this study, temperature emerged as the second most influential habitat attribute. The 90th QR dome shaped response curve was shifted towards cooler water temperatures, which reflected the previously discussed higher densities observed during the dry (winter) season. As has already been discussed, seasonality, and thus temperature, relationships to density are dependent upon recruitment patterns. However, it is important to note here that pink shrimp are known to ‘overwinter’ in estuaries as far north as North Carolina [98] and can occur at higher abundance during winter periods at the southern limit of their range [110].
Along with temperature, salinity was indicated by Kinne [111,112] as a major factor influencing organismal abundance in estuaries. Re-analyses by Zink et al. [6] of data presented by Brusher and Ogren [113] and Minello [114] found increasing pink shrimp abundance with increasing salinity and no statistical difference between polyhaline and mesohaline conditions. It was unexpected to find no limitation of pink shrimp density at high salinities (>35 ppt), but a shortcoming of this study is lack of hypersaline observations (>40ppt). Perhaps the upper limit of salinity observations in our data (i.e., none >39.71 ppt) was not sufficient to reflect a reduction of densities that may occur under hypersaline conditions [5].
The Biscayne Bay pink shrimp IG proposed a pink shrimp preference for seagrasses and presumed that increased % cover of seagrasses would increase the seaward spatial extent of H. wrightii [30] and pink shrimp abundance [1,5,30]. Presently, total SAV QRs yielded the most plausible relationship between pink shrimp density and the benthic habitat metrics investigated. Procrustean analysis test results did not support this. The seemingly weak statistical relationships with either total or species-specific SAV metrics was unexpected. Pink shrimp associations with H. wrightii have been previously reported [12,96,97,115], while other studies have reported high pink shrimp densities associated with total SAV biomass or % cover [10,116,117]. Although one study reports negative impacts of drift and attached algal biomass [118], the positive relationships reported by most studies suggest a stronger relationship between pink shrimp density and either species-specific or total SAV than found in this study.
Several environmental perturbations occurred during the period covered by this study, but only one may have affected pink shrimp density. Variability in climatic conditions led to both wetter and drier than normal wet seasons (Fig 3D). However, departure from typical salinity regimes did not seem to influence temporal density patterns appreciably. For example, the second highest wet-season pink shrimp density (1.29 ± 1.65 shrimp m-2: Table 1) coincided with 2012 record rainfall that reduced salinities (3.34 to 22.08 ppt) across the spatial domain (Fig 3D). Conversely, the highest wet season pink shrimp average density (1.45 ± 2.25 shrimp m-2: Table 1) occurred during the 2015 wet season, previously denoted as a ‘hypersaline’ period [65]. Record dry season rainfall during 2016 yielded the lowest average dry season salinity (Table 1, Fig 3D) while pink shrimp densities that dry season were moderate (0.84 ± 1.18 shrimp m-2: Table 1). Despite their differing salinity conditions, these year-seasons were assigned to the same shrimp density cluster (Fig 3B). The 2013 wet season clustered separately from the others, suggesting a negative impact of microalgal bloom conditions [65,66] on pink shrimp density. No discernable impact on pink shrimp densities was observed after passage of an extreme cold front in the 2010 dry season.
Due to the field sampling design, the present study results may be applicable only to shallow areas < 100 m from the shoreline. Application of study results to areas further offshore should proceed with caution, if at all, due to potential interaction with other habitat attributes that influence trends in pink shrimp density. The present study was also limited by the apparent low catchability of very recently settled pink shrimp by the throw trap gear suggested by reduced numbers of pink shrimp from 3 to 5 mm CL (S1 Fig). Pink shrimp postlarvae are generally considered settled in their nursery habitat by 3 mm CL [9,12,86–88]. Pink shrimp postlarvae settle in shallow (≤1 m), calm water areas along shorelines [12,115], which would suggest they should be readily available to the present field sampling program that samples nearshore waters generally < 1 m deep.
The RECOVER Biscayne Bay IG has set >2 shrimp m-2 as a target wet season pink shrimp density to be achieved with CERP BBCW implementation. But achievement of BBCW and CERP salinity IGs, which include low mesohaline (<10 ppt) and even oligohaline conditions (<5 ppt), may not support increased pink shrimp density. The Biscayne Bay pink shrimp IG may need modification to clarify whether the ≥ 2 shrimp m-2 target refers to all monitoring observations or a seasonal or annual average density across the entire shoreline and to reconsider spatial and seasonal abundance patterns. The present 10 yr analysis of pink shrimp density patterns could be used to refine the RECOVER Biscayne Bay IG.
Supporting information
S1 Fig.
Histograms depicting size frequencies (mm CL) of A) all farfantepenaeid shrimps collected, B) those collected north of Black Point, and C), those collected south of Black Point, D) those collected in the dry season, and E) those collected in the wet season. Vertical dashed line in A) separates smaller sizes (to left of line) that were removed from analysis due to suspected catchability concerns. Shrimp size frequency differences were detected between the two regions (D2-tailed = 0.230, p < 0.0001) and between the wet and dry seasons (D2-tailed = 0.092, p < 0.0001)
https://doi.org/10.1371/journal.pone.0198539.s001
(TIF)
S2 Fig.
Salinity distributional trends of A) mean salinity (ppt) and B) standard deviation of salinity (ppt) across sampling sites across all year-seasons sampled while C) depicts the scatter of salinity mean and standard deviation values as well as the significant correlation trend line between them
https://doi.org/10.1371/journal.pone.0198539.s002
(TIF)
Acknowledgments
This study used monitoring data collected over a substantial period by many individuals; without their efforts, this study would not have been possible. We also thank the National Park Service for its support of this research. This study was conducted during pursuit of the PhD degree of I. Zink; he would like to thank committee members D. Die and J. Luo for their comments and suggestions during review of earlier versions of this manuscript. The authors would also like to thank two anonymous reviewers for their suggestions prior to submission of the manuscript for publication. Funding from the U. S. Army Corps of Engineers under the Comprehensive Everglades Restoration Plan supported this study.
References
- 1.
RECOVER. Appendix to the Recover Team’s Recommendation for Interim Goals and Interim Targets for the Comprehensive Everglades Restoration Plan. West Palm Beach, FL, USA: South Florida Water Management District; Jacksonville, FL, USA: U.S. Army Corps of Engineers; 2005. Available from: http://141.232.10.32/pm/recover/recover_docs/igit/igit_mar_2005_report/09_ig_it_intro_to_app.pdf
- 2.
USACE, SFWMD. Central and Southern Florida Project Comprehensive Everglades Restoration Plan Biscayne Bay Coastal Wetlands Phase 1: Final Integrated Project Implementation Report and Environmental Impact Statement. Volume 1 –Main Report. West Palm Beach, FL, USA: South Florida Water Management District; Jacksonville, FL, USA: U.S. Army Corps of Engineers; 2011. Available from: https://www.nrc.gov/docs/ML1227/ML12270A058.pdf
- 3.
USACE. Biscayne Bay Coastal Wetlands project: Facts and Information. Jacksonville, FL, USA: USACE, 2014. Available from: http://141.232.10.32/docs/BBCW_FS_Sept2014.pdf.
- 4.
RECOVER. Southern estuaries performance measure: salinity. CERP System-wide Performance Measure, Southern Estuaries–Salinity Documentation Sheet. West Palm Beach, FL, USA: South Florida Water Management District; Jacksonville, FL, USA: U.S. Army Corps of Engineers; 2008. Available from: http://141.232.10.32/pm/recover/recover_docs/perf_measures/010709_se_salinity.pdf
- 5. Browder JA, Robblee MB. Pink shrimp as an indicator for restoration of Everglades ecosystems. Ecol. Ind. 2009;9, S17–S28.
- 6. Zink IC, Browder JA, Lirman D, Serafy JE. Review of salinity effects on abundance, growth, and survival of nearshore life stages of pink shrimp (Farfantepenaeus duorarum). Ecological Indicators 2017;81: 1–17.
- 7. Anonymous. The Venice system for the classification of marine waters according to salinity. Limnology and Oceanography 1958;3(3): 346–347.
- 8.
Browder JA, Robblee M, Hall J, Uzgiden L, Eyo P, Moore D, Daniels A. Biscayne Bay coastal and nearshore community baseline study to develop biological performance measures, Part 1: Faunal density and community composition of the nearshore zone; 2005. Annual Report C13401 to South Florida Water Management District. NOAA, NMFS, SEFSC, Miami, FL, USA. 91 pp. Available from: https://www.sefsc.noaa.gov/P_QryLDS/download/PRD134_PRD-03_04-12.pdf?id=LDS
- 9. Diaz GA. Population dynamics and assessment of pink shrimp (Farfantepenaeus duorarum) in subtropical nursery grounds. PhD Dissertation, University of Miami, 2001. Available from: http://access.library.miami.edu/login?url=http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqdiss&rft_dat=xri:pqdiss:3032374
- 10.
Santos RO. Spatial dynamics of vegetated seascapes and the influence on fish and crustaceans. PhD Dissertation, University of Miami, 2014. Available from: https://scholarlyrepository.miami.edu/cgi/viewcontent.cgi?article=2353&context=oa_dissertations
- 11. Santos RO, Lirman D, Serafy JE. Quantifying freshwater-induced fragmentation of submerged aquatic vegetation communities using a multi-scale landscape ecology approach. Mar. Ecol. Prog. Ser. 2011;427: 233–246.
- 12.
Costello TJ, Allen DM, Hudson JH. Distribution, seasonal abundance, and ecology of juvenile northern pink shrimp, Penaeus duorarum, in the Florida Bay area. NOAA Tech. Memo; 1986. NMFS-SEFSC-161. U.S.DOC, NOAA, NMFS, SEFSC, Miami, FL. 84 pp. Available from: https://repository.library.noaa.gov/view/noaa/5744/noaa_5744_DS1.pdf
- 13. Davis JH Jr. The natural features of southern Florida, especially the vegetation and the Everglades. Fla. Geol. Surv. Bull. 1943; 25.
- 14. Halley RB, Shinn EA, Hudson JH, Lidz BH. Pleistocene barrier sand seaward of ooid shoal complex near Miami, Florida. The American Association of Petroleum Geologists Bulletin 1977;64(4): 519–526.
- 15. Shaler NS. The Topography of Florida, with a not by Alexander Agassiz. Bulletin of the Museum of Comparative Zoology 1895;16(Geological Series II): 139–159.
- 16.
Smith HM. Notes on Biscayne Bay, Florida, with reference to its adaptability as the site of a marine hatching and experiment station. In: US Commission of Fish and Fisheries Report of the Commissioner for the Year Ending June 30, 1895. Washington, D.C., USA: Government Printing Office; 1896. Washington, D.C. Pp. 169–191.
- 17.
Kohout FA. Relation of seaward and landward flow of ground water to the salinity of Biscayne Bay. M.Sc. Thesis, University of Miami. 1967.
- 18.
Kohout FA, Kolipinski MC. Biological zonation related to groundwater discharge along the shore of Biscayne Bay, Miami, Florida. In: Laugg G.H., Ed. Estuaries. Washington, D.C., USA: American Association for the Advancement of Science Publication No. 83; 1967. Pp. 488–499.
- 19. Langevin CD. Simulation of submarine ground water discharge to a marine estuary: Biscayne Bay, Florida. Groundwater 2003;41(6): 758–771.
- 20. Stalker JC, Price RM, Swart PK. Determining spatial and temporal inputs of freshwater, including submarine groundwater discharge, to a subtropical estuary using geochemical tracers, Biscayne Bay, South Florida. Estuaries and Coasts 2009;32: 694–708.
- 21.
Brook IM. The effect of freshwater canal discharge on the stability of two seagrass benthic communities in Biscayne National Park, Florida. Oceanologica Acta Special Volume Dec. 1982: Proceedings of the international symposium on coastal lagoons. Bordeaux, France, Sept. 8–14, 1981. Pp. 63–72.
- 22. Irlandi E, Maciá S, Serafy J. Salinity reduction from freshwater canal discharge: effects on mortality and feeding of an urchin (Lytechinus variegatus) and a gastropod (Lithopoma tectum). Bull. Mar. Sci. 1997;61(3): 869–879.
- 23. Serafy JE, Lindeman KC, Hopkins TE, Ault JS. Effects of freshwater canal discharge on fish assemblages in a subtropical bay: field and laboratory observations. Mar. Ecol. Prog. Ser. 1997;160: 161–172.
- 24. Lirman D, Deangelo G, Serafy J, Hazra A, Smith-Hazra D, Herlan J, Luo J, Bellmund S, Wang J, Clausing R. Seasonal changes in the abundance and distribution of submerged aquatic vegetation in a highly managed coastal lagoon. Hydrobiologia 2008;596: 105–120.
- 25. Lirman D, Deangelo G, Serafy JE, Hazra A, Smith-Hazra D, Brown A. Geospatial video monitoring of nearshore benthic habitats of western Biscayne Bay (Florida) using the Shallow-Water Positioning System (SWaPS). J. Coast. Res. 2008;24(1A): 135–145.
- 26. Lirman D, Thyberg T, Santos R, Schopmeyer S, Drury C, Collado-Vides L, Bellmund S, Serafy J. SAV communities of western Biscayne Bay, Miami, Florida, USA: Human and natural drivers of seagrass and macroalgae abundance and distribution along a continuous shoreline. Estuaries and Coasts 2014;37: 1243–1255.
- 27. Browder JA, Alleman R, Markley S, Ortner P, Pitts PA. Biscayne Bay conceptual ecological model. Wetlands 2005;25(4): 854–869.
- 28. McManus LC, Yurek S, Teare PB, Dolan TE, Serafy JE. Killifish habitat suitability as a measure of coastal restoration performance: integrating field data, behavioral trials and simulation. Ecological Indicators 2014;44: 173–181.
- 29. Lirman D, Crooper WP. The influence of salinity on seagrass growth, survivorship, and distribution within Biscayne Bay, Florida: field, experimental, and modeling studies. Estuaries 2003;26(1): 131–141.
- 30.
RECOVER. Performance measures: southern coastal systems submerged aquatic vegetation (Florida Bay and Biscayne Bay). West Palm Beach, FL, USA: South Florida Water Management District; Jacksonville, FL, USA: U.S. Army Corps of Engineers; 2004. Available from: http://141.232.10.32/pm/recover/recover_docs/ret/pm_se_sub_aqua_veg.pdf
- 31. Sklar FH, Browder JA. Coastal environmental impacts brought about by alterations to freshwater flow in the Gulf of Mexico. Environmental Management 1998;22(4): 547–562. pmid:9582391
- 32. Serafy JE, Valle M, Faunce CH, Luo J. Species-specific patterns of fish abundance and size along a subtropical mangrove shoreline: an application of the delta approach. Bull. Mar. Sci. 2007;80(3): 609–624.
- 33. Caccia VG, Boyer JN. Spatial patterning of water quality in Biscayne Bay, Florida as a function of land use and water management. Marine Pollution Bulletin 2005;50: 1416–1429. pmid:16194552
- 34.
Schneider JJ. Tidal relations in the south Biscayne Bay area, Dade County, Florida. Tallahassee, FL, USA: USGS; 1969. Available from: http://dpanther.fiu.edu/sobek/FI07010201/00001
- 35. Wang JD, Luo J, Ault JS. Flows, salinity, and some implications for larval transport in south Biscayne Bay, Florida. Bull. Mar. Sci. 2003;72(3): 695–723.
- 36. Kushlan JA. Sampling characteristics of enclosure fish traps. Trans. Am. Fish. Sco. 1981;110(4): 557–562.
- 37.
Robblee RB, Jewell SD, Schmidt TW. Temporal and spatial variation in the pink shrimp, Penaeus duorarum, in Florida Bay and adjacent waters of Everglades National Park. Homestead, FL, USA: South Florida Research Center, Everglades National Park; 1991.
- 38. Perez-Farfante I. Western Atlantic shrimps of the Genus Penaeus. Fish. Bull. 1969;67(3): 461–591.
- 39.
Perez-Farfante I. Special Scientific Report–Fisheries No. 599: Diagnostic characters of juveniles of the shrimp Penaeus aztecus aztecus, P. duorarum, and P. brasiliensis (Crustacea, Decapoda, Penaeidae). Washington, D.C., USA: U.S. Fish and Wildlife Service; 1970.
- 40.
Perez-Farfante I. Illustrated key to Penaeoid shrimps of commerce in the Americas. NOAA Tech. Report NMFS 64; 1988. Available from: https://spo.nmfs.noaa.gov/sites/default/files/legacy-pdfs/tr64.pdf
- 41.
Perez-Farfante I, Kensly B. Penaeoid and Sergestroid shrimps and prawns of the world: Keys and diagnoses for the families and genera. Paris, France: Memoirs du Museum Nationale d’Histoire Naturelle, Tome 175; 1997.
- 42. Taylor WP. Significance of extreme or intermittent conditions in distribution of species and management of natural resources, with a restatement of Liegbig’s Law of Minimum. Ecology 1934;15(4): 374–379.
- 43. Thompson JD, Weiblen G, Thompson BA, Algaro S, Legendre P. Untangling multiple factors in spatial distributions: lilies, gophers, and rocks. Ecology 1996;77(6): 1698–1715.
- 44. Cade BS, Terrell JW, Schroeder RL. Estimating effects of limiting factors with regression quantiles. Ecology 1999;80(1): 311–323.
- 45. Cade BS, Noon BR. A gentle introduction to quantile regression for ecologists. Front. Ecol. Environ. 2003;1(8): 412–420.
- 46. Lancaster J, Belyea LR. Defining the limits to local density: alternative views of abundance-environment relationships. Freshwater Biology 2006;51: 783–796.
- 47. Scharf FS, Juanes F, Sutherland M. Inferring ecological relationships from the edges of scatter diagrams: Comparison of regression techniques. Ecology 1998;79(2): 448–460.
- 48.
Wood SN. Generalized Additive Models: an introduction with R. Boca Raton, FL, USA: CRC Press; 2006.
- 49. Koenker R, Ng P, Portnoy S. Quantile smoothing splines. Biometrika 1994;81(4): 673–680.
- 50.
Koenker R. Chapter 2—Additive models for quantile regression: an analysis of risk factors for malnutrition in India. In: Vinod HD, editor. Advances in Social Science Research Using R. Berlin, Germany: Springer; 2010. Pp. 23–33.
- 51. Jackson DA. PROTEST: A PROcrustean randomization TEST of community environment concordance. Ecoscience 1995;2(3): 297–303.
- 52. Peres-Neto PR, Jackson DA. How well do multivariate data sets match? The advantages of a Procrustean superimposition approach over the Mantel test. Oecologia 2001;129: 169–178. pmid:28547594
- 53. Lisboa FJG, Peres-Neto PR, Chaer GM, Jesus EC, Mitchell RJ, Chapman SJ, Berbara RLL. Much beyond Mantel: Bringing Procrustes association metric to the plant and soil ecologist’s toolbox. PLoS ONE 2014;9(6): e101238. pmid:24971901
- 54. Adams DC, Rohlf FJ, Slice DE. Geometric morphometrics: Ten years of progress following the ‘revolution.’ Italian Journal of Zoology 2004;71(1): 5–16.
- 55. Arbour JH, Brown CM. Incomplete specimens in geometric morphometric analyses. Methods in Ecology and Evolution 2014;5: 16–26.
- 56.
Legendre P, Legendre L. Numerical Ecology, 2nd Ed. Amsterdam, Netherlands: Elsevier; 1998.
- 57.
Clarke KR, Gorley RN, Somerfield PJ, Warwick RM. Change in marine communities: An approach to statistical analysis and interpretation. 3rd ed. Plymouth, UK: PRIMER-E Ltd.; 2014.
- 58. Anderson MJ. A new method for non-parametric multivariate analysis of variance. Austral Ecology 2001;26: 32–46.
- 59. McArdle BH, Anderson MJ. Fitting multivariate models to community data: A comment on distance-based redundancy analysis. Ecology 2001;82(1): 290–297.
- 60. Anderson MJ. Distance-based test for homogeneity of multivariate dispersions. Biometrics 2006;62(1): 245–253. pmid:16542252
- 61. McGill R, Tukey JW, Larsen WA. Variations of box plots. The American Statistician 1978;32(1): 12–16.
- 62.
Chambers JM, Cleveland WS, Kleiner B, Tukey PA. Graphical methods for data Analysis. Belmont, CA, USA: Wadsworth International Group; 1983.
- 63.
Sokal RR, Rohlf FJ. Biometry, 3rd ed. New York City, NY, USA W.H. Freeman and Co.; 1995.
- 64.
Conover WJ. Practical nonparametric statistics, 3rd ed. New York, NY, USA: John Wiley & Sons, Inc.; 1999.
- 65.
Lirman, D., Bellmund, S., Browder, J., Serafy, J., 2016. 4th Annual Report: IBBEAM–Integrated Biscayne Bay Ecological Assessment and Monitoring. Miami, FL, USA: NOAA NMFS SEFSC; 2016. Available from: https://www.sefsc.noaa.gov/P_QryLDS/download/PRB91_PRBD-2016-15.pdf?id=LDS
- 66.
Wachnicka, A., Browder, J., 2016. Changes in spatial and temporal distribution of epiphytic and planktonic algal communities along environmental gradients in nearshore areas of Biscayne Bay (Florida, U.S.A.). Final report to the National Research Council of the National Academy of Sciences. Available from: https://www.academia.edu/23835222/CHANGES_IN_SPATIAL_AND_TEMPORAL_DISTRIBUTION_OF_EPIPHYTIC_AND_PLANKTONIC_ALGAL_COMMUNITIES_ALONG_ENVIRONMENTAL_GRADIENTS_IN_NEARSHORE_AREAS_OF_BISCAYNE_BAY_FLORIDA_U.S.A._?auto=download
- 67.
Zink IC. Nearshore salinity and juvenile pink shrimp (Farfantepenaeus duorarum): Integrating field observations, laboratory trials, and habitat suitability simulations. PhD Dissertation, University of Miami. 2017. Available from: https://scholarlyrepository.miami.edu/cgi/viewcontent.cgi?article=3030&context=oa_dissertations
- 68. Stabenau E, Renshaw A, Luo J, Kearns E, Wang JD. Improved coastal hydrodynamic model offers insight into surface and groundwater flow and restoration objectives in Biscayne Bay, Florida, USA. Bull. Mar. Sci. 2015;91(4): 433–454.
- 69.
Wang JD, Cofer-Shabica SV. The effects of freshwater canal discharges on salinities in Biscayne National Park. Report to the National Park Service; 1988.
- 70. Lirman D, Orlando B, Maciá S, Manzello D, Kaufman L, Biber P, Jones T. Coral communities of Biscayne Bay, Florida and adjacent offshore areas: diversity, abundance, distribution and environmental correlates. Aquatic Conserv.: Mar. Freshw. Ecosyst. 2003;13: 121–135.
- 71. Ingle RM. Intermittent shrimp sampling in Apalachicola Bay with biological notes and regulatory applications. Proc. Gulf Carib. Fish. Inst. 9th Annu. Sess., 1957. pp. 6–17.
- 72. Williams AB. The influence of temperature on osmotic regulation in two species of estuarine shrimp (Penaeus). The Biological Bulletin 1960;119(3), 560–571.
- 73. Zein-Eldin ZP, Aldrich DV. Laboratory studies of shrimp tolerance to salinity and temperature. Proc. Gulf Carib. Fish. Inst. 1964;16: 121.
- 74. Criales MM, Zink IC, Browder JA, Jackson TL. The effect of acclimation salinity and age on the salinity tolerance of pink shrimp postlarvae. J. Exp. Mar. Biol. Ecol. 2011;409:283–289.
- 75. Ault JS, Diaz GA, Smith SG, Luo J, Serafy JE. An efficient sampling survey design to estimate pink shrimp population abundance in Biscayne Bay, Florida. N. Am. J. Fish. Manage. 1999;19: 696–712.
- 76. Ault JS, Luo J, Smith SG, Serafy JE, Wang JD, Humston R, Diaz GA. A spatial dynamic multistock production model. Can. J. Fish. Aquat. Sci. 1999;56(Suppl. 1): 4–25.
- 77. Hovel KA, Fonseca MS, Myer DL, Kenworthy WJ, Whitfield PE. Effects of seagrass landscape structure, structural complexity and hydrodynamic regime on macrofaunal densities in North Carolina seagrass beds. Mar. Ecol. Prog. Ser. 2002;243: 11–24.
- 78. Blanco-Martínez Z, Pérez-Castañeda R. Does the relative value of submerged aquatic vegetation for penaeid shrimps vary with proximity to a tidal inlet? Preliminary evidence from a subtropical coastal lagoon. Marine and Freshwater Research 2017; 68: 581–591. http://dx.doi.org/10.1071/MF15207
- 79.
Berkeley, S.A., Campos,W.L., Fisheries assessment of Biscayne Bay; 1984. Final Report to Department of Environmental Resources and Management, Miami-Dade County, Miami, FL, USA. 1984. 290 pp. Available from: http://www.aoml.noaa.gov/general/lib/cedardoc.html
- 80. Criales MM, Bello MJ, Yeung C. Diversity and recruitment of Penaeoid shrimps (Crustacea: Decapoda) at Bear Cut, Biscayne Bay, Florida, USA. Bull. Mar. Sci. 2000;67(2): 773–788.
- 81. Criales MM, Cherubin LM, Browder JA. Modeling larval transport and settlement of pink shrimp in south Florida: dynamics of behavior and tides. Marine and Coastal Fisheries: Dynamics, Management, and Ecosystem Science 2015;7(1): 148–176.
- 82. Bryan DR, Luo J, Ault JS, McClellan DB, Smith SG, Snodgrass D, Larkin MF. Transport and connectivity modeling of larval permit from an observed spawning aggregation in the Dry Tortugas, Florida. Environmental Biology of Fishes 2015;98(11): 2263–2276.
- 83. Cummings WC. Maturation and spawning of the pink shrimp, Penaeus duorarum Burkenroad. Trans. Am. Fish. Soc. 90(4): 1961;462–468.
- 84. Kennedy FS, Barber G. Spawning and recruitment of pink shrimp, Penaeus duorarum, off eastern Florida. J. Crust. Biol. 1981;1(4): 474–485.
- 85. Jones AC, Dimitriou DE, Ewald JJ, Tweedy JH. Distribution of early developmental stages of pink shrimp, Penaeus duorarum, in Florida waters. Bull. Mar. Sci. 1970;20(3): 634–661.
- 86. Tabb DC, Dubrow DL, Jones AE. Studies on the biology of the pink shrimp, Penaeus duorarum Burkenroad, in Everglades National Park, Florida. State of Florida Board of Conservation 1962; Tech. Ser. No. 37.
- 87. Allen DM., Hudson JH, Costello TJ. Postlarval Shrimp (Penaeus) in the Florida Keys: Species, Size, and Seasonal Abundance. Bull. Mar. Sci. 1980;30(1), 21–33.
- 88. Criales MM, Yeung C, Jones DL, Jackson TL, Richards WJ. Variation of oceanographic processes affecting the size of pink shrimp (Farfantepenaeus duorarum) postlarvae and their supply to Florida Bay. Estuar. Coast. Shelf Sci. 2003;57, 457–468.
- 89. Criales MM, Wang J, Browder JA, Robblee MB. Tidal and seasonal effects on transport of pink shrimp postlarvae. Mar. Ecol. Prog. Ser. 2005;286: 231–238.
- 90. Criales MM, Wang JD, Browder JA, Robblee MB, Jackson TL, Hittle C. Variability in supply and cross-shelf transport of pink shrimp (Farfantepenaeus duorarum) postlarvae into western Florida Bay. Fishery Bulletin 2006;104, 60–74.
- 91. Criales MM, Browder JA, Mooers CNK, Robblee MB, Cardenas H, Jackson TL. Cross-shelf transport of pink shrimp larvae: interactions of tidal currents, larval vertical migrations and internal tides. Mar. Ecol. Prog. Ser. 2007;345: 167–184.
- 92. Criales MM, Robblee MB, Browder JA, Cardenas H, Jackson TL. Nearshore concentration of pink shrimp (Farfantepenaeus duorarum) postlarvae in northern Florida Bay in relation to nocturnal flood tide. Bull. Mar. Sci. 2010; 86(1): 53–74.
- 93. Criales MM, Robblee MB, Browder JA, Cardenas H, Jackson TL. Field observations on selective tidal-stream transport for postlarval and juvenile pink shrimp in Florida Bay. J. Crust. Biol. 2011;31(1): 26–33.
- 94. Rubec PJ, Kiltie R, Leone E, Flamm RO, McEachron L, Santi C. Using delta-generalized additive models to predict spatial distributions and population abundance of juvenile pink shrimp in Tampa Bay, Florida. Marine and Coastal Fisheries 2016;8(1): 232–243.
- 95. Sheridan P, Minello TJ. Nekton use of different habitat types in seagrass beds of lower Laguna Madre, Texas. Bull. Mar. Sci. 2003;72(1): 37–61.
- 96.
Johnson DR, Browder JA, Harper D, Wong S. A meta-analysis and synthesis of existing information on higher trophic levels in Florida Bay (model validation and prediction): Final report on year 2 of a two-year project. Miami, FL, USA: NOAA NMFS SEFSC; 2002. Available from: https://www.nps.gov/ever/learn/nature/upload/ASS99-1FinalReportSecure.pdf
- 97.
Eldred B. Biological shrimp studies (Penaeidae) conducted by the Florida State Board of Conservation Marine Laboratory. In: Gorsline DS, editors. Proceedings of the First National Coastal and Shallow Water Research Conference, October 1961; 1962. Pg. 411–414.
- 98. Williams AB. A contribution to the life histories of Commercial Shrimps (Penaeidae) in North Carolina. Bull. Mar. Sci. 1955;5(2), 116–146.
- 99.
Costello TJ, Allen DM, Florida Bay ecology project. In: Annual Report of the Bureau of Commercial Fisheries Biological Laboratory, Galveston, Texas. U.S. Fish and Wildlife Circular 246; 1966. p. 15–17. Available from: https://www.biodiversitylibrary.org/item/34421#page/21/mode/1up
- 100.
Allen, D.M., Hudson, J.H., 1970. A sled-mounted suction sampler for benthic organisms. Special Scientific Report–Fisheries No. 614, Bureau of Commercial Fisheries, USFWS, USDOI. Washington, D.C., USA. 16 pp.
- 101. Murphey PL, Fonseca MS. Role of high and low energy seagrass beds as nursery areas for Penaeus duorarum in North Carolina. Mar. Ecol. Prog. Ser. 1995;121: 91–98.
- 102.
Pitre ST. Ontogeny of habitat selection by pink shrimp, Farfantepenaeus duorarum. B.Sc. Thesis, Loyola University, 2001. Available from: http://louisianadigitallibrary.org/islandora/object/loyno-etd%3A82
- 103.
Robblee MB, Dorazio RM, Browder JA, Roy V, Daniels A. Comparison of juvenile pink shrimp abundances in south Florida mangrove and seagrass dominated estuaries. In: Greater Everglades Ecosystem Restoration Planning, Policy and Science Meeting Program and Abstracts; 2008. Pp. 364–365. Available from: http://conference.ifas.ufl.edu/GEER2008/pdf/Abstract_BOOK.pdf
- 104. Rozas LP. Hydroperiod and its influences on nekton use of the salt marsh: A pulsing ecosystem. Estuaries 1995;18(4), 579–590.
- 105. Roth BM, Rose KA, Rozas LP, Minello TJ. Relative influence of habitat fragmentation and inundation on brown shrimp Farfantepenaeus aztecus production in northern Gulf of Mexico salt marshes. Marine Ecology Progress Series 2008;359, 185–202.
- 106. DeLaune RD, Jugsujinda A, Peterson GW, Partrick WH Jr.. Impact of Mississippi River freshwater reintroduction on enhancing marsh accretionary processes in a Louisiana estuary. Estuarine, Coastal, and Shelf Science 2003;58(3), 653–662.
- 107.
Wang JD. Analysis of tide and current meter data for model verification. In: Sündermann J, Holz K-P, editors. Mathematical Modelling of Estuarine Physics. Berlin: Springer-Verlang. 1980. pp. 157–171.
- 108. Wang JD, Luo J, Ault JS. Flows, salinity, and some implications for larval transport in south Biscayne Bay, Florida. Bull. Mar. Sci. 2003;72(3): 695–723.
- 109. Stabenau E, Engel V, Sadle J, Pearlstine L. Sea-level rise: Observations, impacts, and proactive measures in Everglades National Park. Park Science 2011;28(2), 26–30.
- 110. Pérez-Castañeda R, Defeo O. Growth and mortality of transient shrimp populations (Farfantepenaeus spp.) in a coastal lagoon of Mexico: role of the environment and density-dependence. ICES J. of Mar. Sci. 2005;62: 14–24.
- 111. Kinne O. The effects of temperature and salinity on marine and brackish water animals II: Salinity and temperature-salinity combinations. Oceanogr. Mar. Biol. Amm. Rev. 1964;2: 281–339.
- 112.
Kinne O. Physiology of estuarine organisms with special reference to salinity and temperature: general aspects. In: Lauff GH, editor. Estuaries. Washington, D.C.:American Association for the Advancement of Science Publication No. 83; 1967. p. 525–540.
- 113. Brusher HA, Ogren LH. Distribution, abundance, and size of penaeid shrimps in the St. Andrew Bay system, Florida. Fish. Bull. 1976;74(1): 158–166.
- 114.
Minello TJ. Nekton densities in shallow estuarine habitats of Texas and Louisiana and the identification of essential fish habitat. In: Benaka, L.R., Ed. Fish Habitat: Essential Fish Habitat and Rehabilitation. American Fisheries Society Symposium 22. Bethesda, MD, USA: American Fisheries Society; 1999. Pp. 43–75.
- 115.
Eldred B, Ingle RM, Woodburn KD, Hutton RF, Jones H. Biological observations on the commercial shrimp, Penaeus duorarum Burkenroad, in Florida waters. St. Petersburg, FL, USA: Florida State Board of Conservation Professional Papers Series No. 3.; 1961.
- 116. Gore RH, Gallaher EE, Scotto LE, Wilson KA. Studies on decapod crustacea from the Indian River region of Florida XI: Community composition, structure, biomass and species-areal relationships of seagrass and drift algae-associated macrocrustaceans. Estuarine, Coastal and Shelf Science 1981;12: 485–508.
- 117.
Yokel BJ. Animal abundance and distribution in Rookery Bay Sanctuary, Collier County, Florida. PhD Dissertation, University of Miami, 1983.
- 118. Pérez-Castañeda R, Blanco-Martínez Z, Sánchez-Martínez JG, Rábago-Castro JL, Aguirre-Guzmán G, Vázquez-Sauceda MDLL. Distribution of Farfantepenaeus aztecus and F. duorarum on submerged vegetation habitats along a subtropical coastal lagoon (Laguna Madre, Mexico). Journal of the Marine Biological Association of the United Kingdom 2010;90(3): 445–452.