Beyond migrants and residents: complex partial migrations in differentially migrating shortnose sturgeon

Abstract

Background

Partial migration refers to populations that contain both resident and migrant individuals. Partially migratory populations exhibit broad behavioral plasticity, with individual variation stemming from a combination of environmental influences, inheritable genetics, and sexual dimorphism. The United States federally endangered shortnose sturgeon (Acipenser brevirostrum) have been observed to form partially migratory populations in the Gulf of Maine, with sex-biased migration behavior and complex differential timing. However, migrations in southern populations are poorly understood, as are the influences behind migratory life history determination for this species. Therefore, the aim of this study was to investigate the spatiotemporal characteristics and influences of partial migration behaviors in several shortnose sturgeon populations in the southern extent of its range.

Methods

Long-term (2011–2024) acoustic telemetry data from shortnose sturgeon across the southern United States (i.e., Georgia, South Carolina, North Carolina) were used to identify behaviorally distinct migratory contingents using an analysis of similarities with several migration metrics and residency indices. Population- and sex-level influences on migratory life history determination were then assessed using generalized linear and multinomial models.

Results

Multiple behavioral contingents were present within both the migratory and resident groups of southern shortnose sturgeon spawning populations. Individuals classified as resident did not move to other rivers, and differential behaviors between resident contingents were observed where some fish traveled into the lower river and connected estuarine/coastal environments of their spawning river system while others remained upriver. Migrants were detected in at least one river system beyond their putative spawning river, and differential behaviors between migrant contingents related to residency patterns, connectivity, and travel distances during coastal inter-river migration. Furthermore, females had a greater probability of being migratory than males, and larger populations displayed more migratory life histories than smaller populations.

Conclusions

Shortnose sturgeon populations in the southern United States exhibited extensive behavioral plasticity, with sex- and population-level influences on migratory life history determination. Behavioral diversity is critical for stability and persistence of partially migrating populations, and our study demonstrates such behavioral complexity for shortnose sturgeon. Species management must account for the rull range of migratory capabilities observed in our study.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40462-026-00653-x.

Background

Partial migration is a commonly observed migratory phenomenon, in which a population is composed of two behavioral groups – one that migrates, and another that remains resident [1]. Populations that exhibit these behaviors are widespread across animal taxa, including birds [2, 3], ungulates [4], insects [5], and fishes [6]. In fishes, partial anadromy refers to a population whose residents remain in the natal river while migrants travel into saltwater environments for superior growth opportunities before subsequently returning to freshwater to spawn [7, 8]. Further, differential migration describes a particular form of partial migration in fishes in which differential behaviors are related to size, age, or sex [9]. Understanding the determination of migratory life history is an important aspect of fish population ecology because migratory behaviors present interesting tradeoffs. For example, migratory individuals often grow large with high reproductive potentials, but can experience relatively high stress and mortality whereas residents can be smaller and less fecund but have a higher probability of reaching maturity [1, 6, 10]. Despite these differences, each group plays a critical role in the overall persistence and stability of the partially migrating population [11].

Determination of migratory life history within fishes can be influenced by several factors, including heritable genetic behaviors, body condition, juvenile growth rates, environmental conditions, habitat quality, population density, and sex [7, 8, 10]. For example, a partially anadromous salmonid population within a river with little available habitat and few resources displayed a suite of migratory life histories, whereas a conspecific population in a larger river with a diverse array of habitats and more resources was predominantly resident to that river [12]. Furthermore, several fish species have demonstrated that growth and environmental experiences during early life history influence whether individuals become residents or migrants [13–16]. With this, such early life experiences can be dependent on population density because high densities can limit available habitat and growth, often promoting migration as a release from such conditions [1, 7]. Early life history determination can subsequently have carryover effects through ontogeny, influencing life-long migratory behavior and life history parameters such as annual growth, maximum body size, longevity, condition, and morphometrics [17–19]. Overall, migratory life history determination can be dynamic with several nuanced and interacting causes and consequences, emphasizing the importance of empirical study of these behaviors.

Among the factors observed to affect the migratory behavior of partially anadromous fishes, sex has proven highly influential as females are often more migratory than males [7, 10]. This has been attributed to the differential increases in reproductive fitness associated with larger body size for females relative to that of males. For females, the larger body sizes afforded through anadromous migrations greatly enhance reproductive fitness through increases in fecundity and egg size, effectively offsetting costs (i.e., physiological stress, predation, mortality, etc.) incurred from migrating [7, 10]. However, male reproductive fitness is not as strongly linked to body size, so costs associated with anadromous migration make the freshwater resident strategy more selectively advantageous [7, 10]. In fact, empirical evidence suggests sex may be more influential in the expression of migratory life history than other factors for some species. For example, a section of the genome of Oncorhynchus mykiss has been closely associated with becoming an anadromous migrant (i.e., steelhead trout) as opposed to being resident (i.e., rainbow trout) [20]. However, further studies indicated that sex was more strongly correlated to migratory life history than genetics in this species, with female biased migration and male dominated residence [21]. Although migration is behaviorally plastic and may depend on the interaction of several factors, sex appears particularly influential for fishes determining whether to migrate or remain resident in partially anadromous populations.

Shortnose sturgeon (Acipenser brevirostrum) are a federally endangered, long-lived, iteroparous fish species that occur in riverine, estuarine, and marine environments throughout its native range along the Atlantic coast of North America [22, 23]. The species exibits diadromy, with adults often residing downriver in or near estuarine environments before migrating upriver to freshwater reaches of their natal river for spawning [24–27]. Shortnose sturgeon populations can possess multiple distinct movement contingents, with divergent behaviors relating to the extent of migration within and beyond the river of tagging. For example, shortnose sturgeon tagged in the Hudson River, New York were divided into river-residents and those that seasonally explored the lower harbor [28], and rivers within the Gulf of Maine displayed partial migration [29, 30]. In the partially migratory systems, resident individuals were observed to remain in one river, whereas migrants seasonally moved among rivers, with inter-river migrations seemingly centered around returning to their spawning river [29, 30]. Gulf of Maine migrants also exhibited several movement behaviors, varying by time of non-spawning river emigration, time of arrival to the spawning river, duration of spawning river residency, and time spent migrating and exploring other systems during inter-river migration [29, 31]. Furthermore, shortnose sturgeon movement patterns appeared sex-linked because females with developing eggs at tagging were more likely to migrate in the Gulf of Maine [29], and movements into the harbor in the Hudson River appeared similarly female-biased [28]. These studies suggest the importance of sex in shortnose sturgeon migratory life history determination and highlight the variety of behaviors that migrants can exhibit.

Although inter-river migration has been observed in southern populations, the spatial patterns and influences on these behaviors are yet to be formally described in the region. Genetic analyses have indicated that shortnose sturgeon captured in freshwater and estuarine locations in Georgia (GA), South Carolina (SC), and North Carolina (NC) may form a genetically related “southeastern metapopulation” [32]. Mark-recapture and preliminary acoustic telemetry data have confirmed that shortnose sturgeon move among rivers in the region [33, 34], and that the coastal movements of southern shortnose sturgeon (~ 425 km; [35]) are the longest documented for the species. However, there is a paucity of data on the inter-river migratory behaviors of southern populations, including: factors that influence migratory life history determination, prevalence of inter-river migration among southern populations, and identification of behavioral patterns among migrants. Our study addresses these knowledge gaps by using long-term (10 + years), multi-agency collaborative acoustic telemetry detection and sex determination data to determine (1) the behavioral variability within migratory group(s) of shortnose sturgeon in southern populations, and (2) the influence of sex and spawning population in migratory life history determination.

Methods

Fish capture and telemetry

From 2011 to 2024, shortnose sturgeon were captured using bottom anchored gill and trammel nets fished in the estuarine and freshwater habitats of river systems throughout GA (Altamaha and Savannah rivers), SC (ACE Basin [Ashepoo, Combahee, and Edisto rivers], Cooper, Santee, and Winyah Bay [Black, Sampit, Great Pee Dee, and Waccamaw] rivers), and NC (Cape Fear River). A variety of netting gears were used across years and river systems, and descriptions of net design and sampling locations can be found in Ingram & Peterson [27] and Post et al. [34]. Briefly, nets were commonly constructed of monofilament, gill net mesh sizes ranged from 2- to 14-inch stretch measure but were often between 3- and 7-inch stretch measure, and trammel nets were comprised of an inner section of 3-inch stretch and two outer sections of 12-inch stretch measure. These netting methods have been demonstrated to capture shortnose sturgeon of all size and age classes (Fig. S1 [36, 37]). Tagging effort was distributed across the study area from 2011 to 2013 during a multi-state investigation among institutions of GA, SC, and NC, and tagging continued in SC through 2024. Captured individuals were measured for fork and total length and were scanned for a passive integrated transponder (PIT) tag for identification of previously captured fish. If none was detected, a PIT tag was implanted into the dorsal musculature of the fish and a small (~ 1 cm) section of fin tissue was extracted for genetic analysis. A random subset of captured adult shortnose sturgeon (fork length > 500 mm; [25, 36]) were selected for acoustic telemetry transmitter implantations surgery (V9-2x [69-kHz; 410 day estimated battery life] or V16-6x [69-kHz; 1633–3650 day estimated battery life]; Innovasea Systems, Inc.). Detections of these telemetered individuals were tracked via an extensive acoustic telemetry array that spanned across major rivers, estuaries, and select coastal locations from southern GA through NC (Fig. 1). Extensive acoustic receiver coverage was maintained throughout the study area from 2011 to 2017; however, receiver stations were removed from GA rivers and select locations in SC throughout 2017 and 2018. Although the array was contracted in latter years, broad-scale receiver coverage in SC and NC was maintained for the duration of our study (Fig. 1).

Fig. 1.

Acoustic telemetry detections of adult shortnose sturgeon (n = 206) tagged in rivers of Georgia, South Carolina, and North Carolina during Timeframe 1 (A; detections from 2011–2017) and 2 (B; detections from 2017–2024). Small, opaque circles represent individual acoustic receivers that detected individuals during each timeframe. Larger, translucent circles indicate assigned node locations (i.e., receiver groupings by river system/coastal position where receivers were active throughout the timeframe) used in network analyses. Acoustic receiver and node colors correspond to separate river systems and coastal locations. Note although Altamaha River detections occurred early in Timeframe 2 (B), those receivers were removed in 2017 and thus the river was not treated as a node location thereafter

Data analysis

Detection database preparation

Raw detection data were imported into program R 4.4.3 [38] and were filtered to exclude “false” detections, including detections prior to a transmitter’s activation date or before the minimum interval between consecutive detections [35, 39]. Unrealistic detections based on the timing and locations of successive detections and any telemetered shortnose sturgeon that remained stationary at its tagging site were also filtered. Individuals that were detected for less than 365 days after tagging were also excluded from further analysis, as shortnose sturgeon movements are seasonal [27, 29] and thus detection chronologies that spanned across multiple seasons are necessary to obtain an accurate depiction of broad-scale migratory behavior. Detections were grouped into broad locations based on the river system (i.e., freshwater river and adjacent lower estuary) or location in marine waters along the coast. The detection dataset was truncated to include a maximum of one detection per individual per location per day. Futher, sex was assigned to each individual based on either observation during acoustic surgery or, when gonads were not identifiable, by molecular techniques [40, 41]. Briefly, molecular sex was determined by amplification (female) or non-amplification (male) of the AllWSex2 marker with appropriate positive controls to ensure that any non-amplification of the marker was not due to PCR failure. Ambiguous molecular results, such as atypically low amplification, were re-run once. If the ambiguity persisted, the sex of this sturgeon remained unknown.

To avoid biases due to the uneven telemetry effort from 2011 to 2024, analyses were conducted on individuals within the following timeframes: Timeframe 1 = shortnose sturgeon acoustically tagged in the collaborative effort in 2011–2013 and passively tracked from 2011 to 2017 (when acoustic receiver arrays were maintained throughout the study area); Timeframe 2 = individuals tagged in SC from 2014 to 2024 and whose tracks included the 2017–2024 contracted array; and an Overall Timeframe including all tagged fish and detections. Evaluating these timeframes independently and then together allowed us to identify whether behavioral observations differed due to differential effort to tag and detect fish in systems throughout the study region, while also looking for broad-scale and persistent behavioral trends.

To further assign fish to ecologically meaningful groups, the detection history of each telemetered shortnose sturgeon was visually inspected to identify upriver putative spawning movements. Putative spawning migrations were conservatively identified based on previously described spatiotemporal characteristics of upriver movements believed to be indicative of spawning runs for each system within our study. Briefly, published putative spawning migrations for southern populations consisted of substantial upriver movement (> 80 river kilometers [rkm]) during the spring spawning window (December–April) in most rivers (Table S1; [27, 34, 42]). However, in the Cooper River, spawning behavior has been observed in the tailrace of the Pinopolis Dam at rkm 77 [34, 43]. Therefore, shortnose sturgeon that migrated to to the Pinopolis Dam in the spring were grouped into the Cooper River spawning population. Given that shortnose sturgeon are assumed to demonstrate natal philopatry [44], individuals were assigned to the natal population within the river where each fish made putative upriver spawning migration(s). These filtering criteria and groupings allowed us to make unbiased and accurate observations of the spatial extent of coastal inter-river migrations, while drawing comparisons among populations and sex.

Network metrics

To visualize spatial characteristics and habitat use throughout inter-river migration, we created spatially projected network diagrams (referred to hereafter as “networks”) for every telemetered individual detected in more than one river system (referred to hereafter as “migrant”) using the igraph package in R [45]. In our networks, the previously defined broad detection locations (Fig. 1) were treated as individual “nodes,” and the size of each node was determined by the number of days that the individual was detected in that location. Connectivity among nodes is indicated by “edges,” which identify instances when a migrant’s previous and subsequent detections occur between two nodes. Edge thickness illustrates how many times a migrant made the movement, as edge weight becomes greater and therefore thicker with a higher frequency of movement directly between nodes. These networks were spatially projected using the R package ggmap [46].

To understand the overall migratory behavior, regional connectivity, and coastal movement distance capabilities of migrants as displayed by these networks, we selected specific metrics for further analysis based on previous studies of migratory fishes [47, 48]. These metrics included node density, number of edges, and distance to the furthest node from the node corresponding to a migrant’s river of spawning (referred to hereafter as a migrant’s “spawning node”). Node density was the proportion of nodes a migrant was detected on when considering all possible nodes within the study system at the time of its migration, representing the regional connectivity displayed by a migrant throughout its transmitter’s lifetime. The number of edges within a network reflected the connectivity among nodes during a given migratory event. Distance from the spawning node to the furthest node of detection is indicative of the overall migratory distance capacity of each migrant as it traveled along the coast. These distances were calculated in ArcGIS Pro 3.1.0 (Esri Inc., Redlands, California) through the Distance Accumulation function, using an Atlantic coastline barrier to restrict measurements to in-water movements among nodes to ensure biological relevance. Inter-river migration distance estimates among nodes were conservative because node centrums were arbitrarily assigned within the downriver estuarine and tidally influenced segment of each river system, but observed migrations often took migrants further upriver than these node centrums. Moreover, marine migration measurements account for the minimum in-water distances between each node, which likely underestimates true migration distances traveled.

Residency indices

Residency patterns were further defined through the calculation of a residency index (RI) for each migrant shortnose sturgeon. To do this, we used the last-observation-carried-forward (LOCF; [49]) method where an individual was assumed to be within its node of last detection on days when it went undetected. However, if an individual moved from one node to another after a period of going undetected, those undetected days were considered days of active migration within the estuarine and marine habitats and were therefore labelled as “estuarine/marine residence.” We then calculated RI values by dividing the number of days detected at each node/location by the total number of days that the transmitter was at liberty.

Analysis of similarities

To identify potential behavioral contingents among migrant shortnose sturgeon, we used an analysis of similarities (ANOSIM) with spatial characteristics of migration behavior, habitat use, and sex. An ANOSIM is a robust non-parametric multivariate analysis that analyzes both within-group similarities and between-group dissimilarities within a data matrix [50]. For our data matrix, rows corresponded to each migrant shortnose sturgeon, and columns corresponded to network metrics (node density, number of edges, distance to furthest node), RI values (spawning river, estuarine/marine), and sex. A dissimilarity matrix was created using the daisy function within the cluster package in R [51], using the Gower distance metric [52]. Gower distance is a flexible method for calculating dissimilarity distances among migrants as it allows for the mixture of categorical and continuous variables within the data matrix, and for the inclusion of individuals that may have undefined metrics such as unknown sex [52]. The differentiation of groups was then tested using the anosim function within the vegan package [53]. The ANOSIM model results provided an R statistic, which relates to both the intra- and inter-group dissimilarities. An R statistic ranges from − 1 to 1, where values approaching 1 indicate strong separation among groups (high inter-group dissimilarity and high intra-group similarity), suggesting an effective grouping structure. Values near 0 denote little to no separation among groups, implying that the samples may belong to a single homogeneous group. Negative values indicate greater similarity between groups than within them, suggesting that the grouping may not reflect true underlying patterns. We used the ANOSIM significance value threshold of < 0.05 in determining statistical significance.

In each ANOSIM, we tested a different grouping method to identify the migrant groups with the highest migratory behavioral separation. Groupings were based on the migratory distance capabilities observed for each migrant in relation to its spawning river, with varying migration distances differentiating each group. The coastal migration distances used to create bins differentiating groups were selected based on distances that were ecologically relevant given distances among southern river systems while also offering large enough sample sizes to draw statistical inferences among groups. Additionally, we grouped fish by spawning river to assess behavior by population. Thus, ANOSIM groups were as follows: coastal migration distances of less than (short-distance) or greater than (long-distance) 100 km, 150 km, and 200 km; a multifaceted strategy including short-distance (up to 100 km), moderate distance (between 100 and 200 km), and long-distance (over 200 km) groups; and the spawning river grouping. We then selected the best grouping strategy based on which ANOSIM provided the Statistic R closest to 1, and the lowest significance value.

To account for the differences in networks inherent with a changing acoustic receiver array, we first performed separate ANOSIMs within each timeframe. An overall ANOSIM was then constructed to investigate behavioral groupings over time. Results of each ANOSIM were visualized in ordination space with non-metric multidimensional scaling (NMDS) plots, with 95% confidence ellipses calculated for the overall model through the ConfidenceEllipse R package [54].

Resident contingent identification

In addition to migratory contingents, we sought to identify distinct resident contingents based on differing use of salinity profiles throughout the putative spawning river, similar to shortnose sturgeon in the Hudson River [28]. To do this, we compiled salinity data attained from environmental monitoring stations throughout the study area including: hydrographic monitoring stations from the Georgia Coastal Ecosystems Long Term Ecological Research (GCE-LTER) in the Altamaha River, a National Estuarine Research Reserve water quality station in Winyah Bay, and several United States Geological Survey stream gauges in the Waccamaw, Cooper, and Savannah rivers [Table S2]. Salinity datasets were filtered to daily low tide measurements during winter/spring months (December–April), which has been previously identified as the timeframe that southern resident shortnose sturgeon are able to move into higher salinity environments of the lower river and estuary [27, 34, 44, 55]. We then calculated the percentage of days during this timeframe that the daily low tide salinity at each station was observed mesohaline (> 5 parts per thousand [ppt]) and oligohaline (> 3 ppt). We used these thresholds because shortnose sturgeon have historically demonstrated a freshwater-centric life cycle [24, 56, 57], with populations seasonally concentrated between 0 and 3 ppt [44, 56]. Furthermore, although explicit studies of salinity effects on adults are lacking, juvenile shortnose sturgeon begin to experience negative effects on weight gain and preliminary osmoregulatory changes at > 5 ppt [58, 59], and other juvenile diadromous sturgeons have been observed to undergo preliminary physiological and osmoregulatory changes at ≥ 3 ppt [60, 61]. Therefore, rkm thresholds for freshwater vs. euryhaline occurrences were identified based on river-specific salinity intrusion patterns to determine differential space use among resident shortnose sturgeon across salinity gradients.

Sex and population influence on migratory behavior

To understand the basis of partial migration in southern shortnose sturgeon, we first tested for a potential sex-bias in migration as is common in partially migrating fishes and was observed with shortnose sturgeon in the Gulf of Maine. To do this, we first filtered out all shortnose sturgeon for which we could not determine sex or putative spawning population to ensure ecological relevance of model results. We then assigned telemetered individuals either “resident” or “migrant.” In this study, any shortnose sturgeon that was only ever detected in one river system was assigned “resident,” and individuals that were detected in more than one system were assigned “migrant.” These migrant-resident classifications based observation(s) of inter-river movement mirrors that of previous shortnose sturgeon partial migration studies in the Gulf of Maine [29, 30]. Individuals were assigned a binary response based on this classification, with “0” indicating a resident and “1” indicating a migrant. These data were then analyzed with a generalized linear model (GLM) using the glm function within the stats package in program R, with sex as the predictor variable. The level of statistical significance was set to p < 0.05, and pairwise comparisons and significant differences were identified and visualized by calculating the estimated marginal means using the emmeans package [62]. All comparisons of statistical significance and model result visualizations hereafter followed this methodology.

To understand the basis of differential migration within southern shortnose sturgeon populations, we assigned individuals to behavioral contingents as determined by ANOSIM results for migrants and within-river salinity partitioning for residents. We then tested whether contingents were biased towards a particular sex, spawning population, or both, with a set of multinomial models using the nnet package in program R [63]. We first considered the interactive effect of sex and spawning population, and if the interaction was insignificant, we interpreted the model that included the additive effect of these predictor variables.

Results

From 2011 to 2024, 299 shortnose sturgeon were captured and acoustically tagged within the Altamaha (n = 40), Savannah (n = 102), Edisto (n = 13), Cooper (n = 73), Santee (n = 2), Winyah Bay system (n = 67), and Cape Fear (n = 2) rivers. These telemetered individuals produced a total of 24,708,527 raw detections. After detection filtering, the data set was truncated to include 23,449,693 detections of 206 individuals. The 93 individuals filtered from the dataset were detected for less than 1 year, potentially due to a combination of tag failure, tag loss, or mortality. Of the remaining 206 shortnose sturgeon, 158 (76.7%) remained resident to one system throughout their transmitter’s lifetime, and 48 (23.3%) were observed to perform inter-river migration(s). The resident group was comprised of 55 females, 95 males, and 8 individuals of unknown sex, and the migratory group included 39 females, 7 males, and 2 unknowns. Acoustic detections occurred as far south as the Altamaha River, and as far north as the Cape Fear River (Fig. 1). Although acoustic receivers were maintained south and north of these systems, no detections occurred outside of this range.

After data filtering, Timeframe 1 consisted of 81 shortnose sturgeon that made putative spawning runs and were therefore assigned to a river population: 29 to the Altamaha River, 20 to the Savannah River, 30 to the Cooper River, and 2 to the Great Pee Dee River. In Timeframe 1, migratory individuals that were detected in river(s) beyond their spawning river system included 4 (13.4%) of shortnose sturgeon from the Altamaha River, 9 (45.0%) from the Savannah River, 0 from the Cooper River, and 1 (50.0%) from the Great Pee Dee River (Table S3). Timeframe 2 consisted of 1 shortnose sturgeon assigned to the Altamaha River using putative spawning movement, 44 assigned to the Savannah River, 27 assigned to the Cooper River, and 26 assigned to the Great Pee Dee River. In Timeframe 2, migratory individuals that were detected beyond their spawning river system included 1 (100.0%) shortnose sturgeon from the Altamaha River, 10 (22.7%) from the Savannah River, 0 from the Cooper River, and 13 (50.0%) from the Great Pee Dee River (Table S3). Throughout both timeframes, 27 individuals displayed no putative spawning movements. This included 10 migrants and 17 residents, tagged across several river systems, including the Altamaha (n = 2), Savannah (n = 11), Edisto (n = 1), Santee (n = 1), and Winyah Bay (n = 12) river systems. Given that previous literature has hypothesized that shortnose sturgeon inter-river migration centers around returning to the putative spawning river [29, 30], individuals without any observation(s) of a putative spawning migration were excluded from ANOSIM analyses. This was done because the true migratory extent of these individuals could not be determined without observing a coastal migration to return to the putative spawning river. Therefore, 38 migrants overall were included in ANOSIM analysis (i.e., 14 from Timeframe 1 and 24 from Timeframe 2).

Classification of distinct movement contingents

Migratory individuals displayed a wide variety of behaviors, with behavioral variation relating to migration distance capabilities, connectivity throughout the study area, and residency patterns (Figs. 2 & S2-S5). According to the ANOSIM, the most effective grouping strategy in each timeframe was to divide shortnose sturgeon migrants according to the following migratory capacities: short-distance migrants (< 150 km; Fig. 3; Table 1) and long-distance migrants (> 150 km; Fig. 3; Table 1). In addition to migratory distance capabilities, short-distance migrants were also characterized by low regional connectivity as observed with low node densities and network edge numbers, relatively low residence in the estuarine/marine environment, and relatively high spawning river residence (Fig. 2; Table 2). This behavioral contingent included 22 (57.9%) of the 38 migrant shortnose sturgeon analyzed via ANOSIM (Table 3): 8 from Timeframe 1, and 14 from Timeframe 2. Conversely, long-distance migrants were characterized by high levels of connectivity with large node density and network edge values, relatively high residence in the estuarine/marine environment, and relatively low spawning river residence (Fig. 2; Table 2). This behavioral contingent included 16 (42.1%) of the 38 migrants analyzed via ANOSIM (Table 3): 6 from Timeframe 1 and 10 from Timeframe 2 (Fig. 4).

Fig. 2.

Examples of paired detection plots and network diagrams for short- and long-distance migrants. Plots represent examples of individual shortnose sturgeon, with the migration contingent, putative spawning river, and transmitter ID identified in the title for each panel in the left column. Bolded letters in the top left of each panel indicate the individual represented in these figures; panels A and B represent individuals from Timeframe 1 (detections from 2011–2017) and panels C and D were from Timeframe 2 (detections from 2017–2024). Y-axes on detection plots represent river systems and coastal locations where receivers were maintained during respective tracking timeframes. Locations are in south to north order starting at the bottom of the axes. Grey bars indicate previously identified spawning windows of southern populations (December–April). Network diagrams display directed movements with black lines (“edges”) and habitat use among locations as node size increases with the more detections in that location

Fig. 3.

Non-metric multidimensional scaling (NMDS) ordination plots displaying the behavioral divergence of short- and long-distance migrant shortnose sturgeon. Shape reflects each migrant’s spawning river, and color-coding reflects the migration behavioral group. Timeframe 1 represents migrants detected from 2011–2017, Timeframe 2 represents migrants detected from 2017–2024, and the Overall timeframe represents migrants detected from 2011–2024. Ellipses in the Overall timeframe plot containing all migrant shortnose sturgeon (n = 38) represent 95% confidence ellipses

Table 1.

Shortnose sturgeon migratory grouping strategies tested with analysis of similarities (ANOSIM) by timeframe. Groups were differentiated by coastal distance capabilities of migration for each individual from its putative spawning river, and the last group in each timeframe represented putative spawning river populations. The ANOSIM Statistic R and Significance value reflect how well the groupings explain inter-group dissimilarity, intra-group similarity, and the statistical significance of the groups. Groups in bold represent the optimal grouping strategy for that timeframe, and asterisks denote a group that contained the same individuals as the optimal group for that timeframe

Study timeframe	Grouping strategy	Statistic R	Significance value
Timeframe 1	<100 km, >100 km	0.256	0.056
	<150 km, >150 km	0.936	0.001
	<200 km, >200 km	*	*
	<100, 100–200, >200 km	0.759	0.001
	Spawning river	0.32	0.02
Timeframe 2	<100 km, >100 km	0.822	0.001
	<150 km, >150 km	0.912	0.001
	<200 km, >200 km	0.831	0.001
	<100, 100–200, >200 km	0.854	0.001
	Spawning river	0.837	0.001
Overall timeframe	<100 km, >100 km	0.587	0.001
	<150 km, >150 km	0.786	0.001
	<200 km, >200 km	0.781	0.001
	<100, 100–200, >200 km	0.726	0.001
	Spawning river	0.485	0.001

Table 2.

Summary of migration metrics between short- and long-distance shortnose sturgeon migrants as determined by the analysis of similarities (ANOSIM)

Migration contingent	ANOSIM metric	Mean	Standard deviation
Short-distance migrants	Node density	0.27	0.53
	Number of edges	2.363	1.002
	Distance to furthest node	58.677	35.843
	Spawning river residency	0.714	0.361
	Estuarine/marine residency	0.005	0.006
Long-distance migrants	Node density	0.612	0.157
	Number of edges	9.375	4.08
	Distance to furthest node	255.136	82.355
	Spawning river residency	0.094	0.11
	Estuarine/marine residency	0.042	0.069

Table 3.

Behavioral description and prevalence of each migratory contingent observed in southern shortnose sturgeon, by putative spawning population and sex

Migratory contingent	Contingent behavior description	Observations by population	Observations by sex
River resident	Migration restricted to the putative spawning river system, does not enter the lower river reaches of euryhaline salinities	Altamaha River (n=12)	Females (n=28)
		Savannah River (n=22)	Males (n=64)
		Cooper River (n=54)	Unknown (n=4)
		Winyah Bay (n=0)
		Unknown (n=8)
Euryhaline resident	Migrates through entirety of the putative spawning system including lower river and adjacent coastal habitat, never enters other estuaries or river systems	Altamaha River (n=13)	Females (n=27)
		Savannah River (n=23)	Males (n=31)
		Cooper River (n=3)	Unknown (n=4)
		Winyah Bay (n=14)
		Unknown (n=9)
Short-distance migrant	Migrates throughout entirety of the putative spawning river, and geographically proximate (<150 km) river system(s)	Altamaha River (n=2)	Females (n=22)
		Savannah River (n=6)	Males (n=4)
		Cooper River (n=0)	Unknown (n=2)
		Winyah Bay (n=14)
		Unknown (n=6)
Long-distance migrant	Migrates throughout entirety of the putative spawning river, and river systems relatively far (>150 km) from the putative spawning river	Altamaha River (n=3)	Females (n=17)
		Savannah River (n=13)	Males (n=3)
		Cooper River (n=0)	Unknown (n=0)
		Winyah Bay (n=0)
		Unknown (n=4)

Fig. 4.

Select examples of the migratory ranges of shortnose sturgeon, with each map depicting the movement of an individual that displayed one of the differential migration patterns observed in this study. Panel titles indicate the migration contingent displayed, spawning population of example individual, and the approximate tracking duration for that individual. Inset boxes in the lower panels indicate the map extent of the upper panels

Behavioral variation was also observed within the resident group pertaining to downriver movement into areas of varying salinity throughout the study (Table 3; Figs. 4 & S6). Euryhaline residents were categorized as those that were detected for > 2 days in the lower mesohaline and oligohaline reaches of their putative spawning river (Fig. S7-S10), at times moving downriver and beyond the array, seemingly into the coastal/marine environment (Figs. 4 & S6). However, these individuals were never detected on the receiver array of any other river system, suggesting that each individual’s movements were restricted to coastal waters in or near their spawning river system. This behavioral contingent included 53 (37.6%) of the resident shortnose sturgeon with putative spawning population assignments (Table 3). In contrast, river resident shortnose sturgeon were not observed to use these euryhaline lower river reaches (Figs. 4 & S6), despite substantial tracking durations for these individuals (mean days at liberty = 1389 days, standard deviation ± 612 days). This behavioral contingent included 88 (62.4%) of the resident shortnose sturgeon with putative spawning population assignments.

Sex and population influence on migratory behavior

After filtering out individuals without sex or putative spawning population assignments, we were left with 171 adult shortnose sturgeon for migratory life history determination analyses. Among these individuals, GLM results indicated that females were significantly more likely to be observed as migrants (0.39 probability, 95% Confidence Interval [CI]: 0.29–0.50) than were males (0.06, 95% CI: 0.03–0.13; p-value < 0.001) (Fig. 5). However, no migration was observed in individuals that exhibited putative spawning movements in the Cooper River, suggesting this population may be particularly resident. To address the confounding effects of the complete lack of behavioral variability within the Cooper River population, we performed another GLM excluding Cooper River shortnose sturgeon (Fig. 5). This second GLM indicated that females (0.47, 95% CI: 0.35–0.59) were again significantly more likely to be observed as migrants than males (0.12, 95% CI: 0.05–0.24; p-value < 0.001).

Fig. 5.

Predicted influence of sex on migration vs. residency as observed with generalized linear models (GLM). Analyses were conducted using all telemetered shortnose sturgeon (A) and all except those assigned to the Cooper River (B) to account for the lack of behavioral variability in this entirely resident population. Each point represents the probability point estimate for that sex, and error bars represent 95% confidence intervals. Statistically significant differences according to post hoc Tukey’s multiple comparisons tests are indicated by letters above each treatment

Multinomial model results indicated no interactive effect of sex and spawning population; however, we observed significant additive effects with these predictors when assessing shortnose sturgeon migratory life history determination (Fig. 6). Female and male shortnose sturgeon had comparable probabilities of exhibiting river resident behavior (p-value = 0.160; Fig. 6). However, males were significantly more likely than females to be euryhaline residents (p-value = 0.044; Fig. 6). Females were significantly more likely to be both short-distance (p = 0.032; Fig. 6) and long-distance migrants (0.018; Fig. 6) than males.

Fig. 6.

Predicted influence of sex and spawning population on the migratory life history determination exhibited by shortnose sturgeon as observed with a multinomial model considering the additive effect of each. Each point represents the probability point estimate for that sex to take on the migratory contingent of that panel, and error bars represent 95% confidence intervals. Statistically significant differences according to post hoc Tukey’s multiple comparisons tests are indicated by letters above each treatment

Multinomial model results further indicated several intricate inter-population differences in migratory life history determination. Shortnose sturgeon putatively spawning in the Cooper River were significantly more likely than all other populations to be river residents (p-values < 0.001; Fig. 6), with 54 out of 57 individuals exhibiting this life history. Further, Cooper River individuals displayed the lowest probability of exhibiting all other life histories, as only three of these shortnose sturgeon were euryhaline resident and none were migratory (Table 3; Fig. 6). In contrast, none of the shortnose sturgeon putatively spawning in the Great Pee Dee River (n = 28) were observed to be river residents (Table 3) as each of these individuals were detected below the river’s mouth in the Winyah Bay estuary or in other river systems that also drained to Winyah Bay. Each of the other putative spawning populations displayed similar probabilities of euryhaline residency, and Great Pee Dee River putative spawning individuals were most likely to be short-distance migrants (Fig. 6). Shortnose sturgeon spawning in the Altamaha and Savannah river exhibited the highest probabilities of long-distance migration, as neither of the other populations contained these individuals (Fig. 6).

Discussion

Partial migration and migratory contingent population structure, where contingents diverge in movement behavior and space use, have been observed in several diadromous fishes such as rainbow/steelhead trout and striped bass, among many others [9, 12, 64, 65]. Although shortnose sturgeon have been observed to possess differential behavior in relation to migration or residency as well as migratory timing [29], this study further demonstrates the nuanced variability in the species’ migratory behavior. In this study, we observed female-biased partial migration in southern shortnose sturgeon populations. Furthermore, we used an innovative statistical framework to identify four migratory contingents across four putative spawning populations that differed in their overall space use, migratory distance capabilities, and regional connectivity. Contingent determination was also sex-biased, demonstrating the differential migration dynamics of shortnose sturgeon. The presence of several distinct migratory contingents that spread population members across a broad spatial range (i.e., the “portfolio effect” [9, 66, 67]) preserves population stability over time by protecting against disturbances in any one refuge location (Fig. 7). Although such behavior is evolutionarily advantageous, it is vital to understand and document differential behavior among contingents to effectively describe the ecology of a species and manage contemporary populations.

Fig. 7.

A map of the study region that displays the portfolio effect for migrant populations of shortnose sturgeon. Pie charts indicate the proportion of migrants that make up each migratory contingent among those that made putative spawning migrations in that river. Arrows indicate the furthest location of detection for each migrant from that spawning river, are color-coded by migration contingent, and the thickness of each line reflects how many migrants are represented by that line as the spatial extent of their migratory behavior. Sample sizes of putative spawning adults per river are listed next to each spawning river

Migratory life history determination

Partial migration of diadromous fishes has repeatedly been observed to be female-biased, affording females sufficient resources for increased body size and therefore higher reproductive potential [7, 10]. This sexual dimorphism was observed with shortnose sturgeon in the Gulf of Maine, and this study demonstrated the same sex-biased partial migration at the southern extent of its range. In fact, the GLM controlling for complete residency in the Cooper River population estimated that females had an approximately 47% probability of inter-river migration, compared with 12% for males (Fig. 5). Furthermore, the differential migration patterns we observed further demonstrated the influence of sex on migratory life history determination of shortnose sturgeon among migration contingents. Although females and males demonstrated similar probabilities of exhibiting river residency, males were more likely to be euryhaline residents, seasonally moving into lower river reaches and adjacent estuarine/coastal environments of euryhaline salinities without entering another river system (Figs. 4, 6, & S6). However, females that entered the same coastal environments were more likely to become both short- and long-distance migrants (Fig. 6). These observations suggest that southern shortnose sturgeon males and females may be similarly capable of tolerating elevated salinity when entering estuarine/marine habitats from the freshwater riverine environment; an interesting contrast to the behaviors observed in Hudson River shortnose sturgeon [28] and other partially anadromous species where male-biased resident groups often remain in freshwater [7, 8]. Future studies should seek to analyze stable isotope ratios of southern shortnose sturgeon to determine how migratory contingents use each biome, and whether separate contingents occupy distinct ecological niches. For example, stable isotope analysis could further elucidate whether euryhaline residents, short-distance, and long-distance migrants feed in marine habitats. Understanding how migrant shortnose sturgeon use the estuarine/marine environment (i.e., resource acquisition, migratory pathway, both) – and how this compares to euryhaline residents – could have substantial implications for the ecological understanding of each sex and migration contingent. Overall, shortnose sturgeon within our study displayed a broad spectrum of life history and behavioral tactics, and sex played a pivotal role in both partial and differential migratory life history determination.

Beyond sex, putative spawning population was also highly influential in differential migratory life history determination. For example, three of the four putative spawning populations that we observed contained a contingent of migrants, and the two most abundant river populations in this study (i.e., the Altamaha and Savannah rivers; [36, 37, 68]) displayed similar life history determination patterns with all four migratory life histories (Figs. 6 and 7). However, none of the telemetered shortnose sturgeon putatively spawning in the Cooper River (n = 57) were detected in any other system. In fact, this population was almost entirely river resident (n = 54; Table 3; Fig. 4), further emphasizing the sedentary nature of this population. Furthermore, there has only been one confirmed instance through mark-recapture of an adult shortnose sturgeon from the Cooper River being encountered in another river in four decades of capture data (B. Post unpublished data). Previous studies have suggested that density-dependence may be responsible for increased migratory behavior in shortnose sturgeon [30]. Interestingly, The Cooper River population is among the smallest in this study region, with little to no juvenile recruitment observed for decades and an adult population seemingly persisting on the species’ long lifespan [42, 69, B. Post unpublished data]. This is likely a result of the Pinopolis Dam at rkm 77 impeding upriver spawning migrations and altering spawning substrate and environmental characteristics downstream, potentially hindering spawning success and early life stage survival [43]. These results suggest that the low densities of shortnose sturgeon in the Cooper River could be at least partially responsible for these highly resident adult behaviors, as opposed to the nuanced migratory contingent structure of the more abundant populations. Moreover, partial migration has also been linked to habitat complexity and food availability [9, 12]. Despite the seemingly limited available habitat to Cooper River shortnose sturgeon, wide-spread residency and low migratory complexity suggests that this anthropogenically altered river may still afford enough resources to suppress migration in the remaining population. Future research is needed to determine whether population and environmental characteristics such as density-dependence, habitat characteristics, or food availability are responsible for these behaviors.

Many factors may influence migrant individuals to become short- vs. long-distance migrants, and the fine-scale behavior during migrations. Across taxa, heritable genetic traits have been observed to influence not only whether an individual will migrate, but also fine-scale behaviors such as migratory timing, distances traveled, orientation, and destination [70]. For example, several diadromous salmonids have demonstrated genetic influence on the spatiotemporal characteristics of marine migration [71–73]. This raises interesting questions regarding whether heritable genetic traits similarly influence the behavioral variations observed in diadromous shortnose sturgeon. If genetics do influence fine-scale migration behavior, then conservation of each migrant contingent is necessary to safeguard genetic diversity of these partially migrating populations. Future studies could pair detailed migratory tracking with assays of functional genomic markers linked to migratory traits to identify labile developmental periods and quantify lifetime carryover effects on contingent behavior, as has been demonstrated in other partially migratory fishes [21, 70, 74, 75].

Ecological and conservation implications of partial migration and differential migration

The inter-river migrations of reproductively active adult shortnose sturgeon during spawning windows has been listed as a research priority for the species [22, 23]. Observing “effective straying,” in which migrants spawn in one or more non-natal rivers, would have a profound ecological influence for this federally endangered species by providing gene flow among systems and potential recolonizion in rivers where populations have been locally extirpated. Although several behavioral distinctions exist among the migratory contingents that we observed in this study, every telemetered shortnose sturgeon displayed putative spawning movement in only one river. In doing this, several migrants departed from and passed numerous river systems that actively support a spawning population to reach their spawning river. These results reinforce our method of assigning fish to a spawning population and suggest that shortnose sturgeon may indeed have high natal philopatry. This pronounced spawning river fidelity suggests that there may be a limited or lack of effective straying, despite numerous opportunities throughout several inter-river migrations. Although it is difficult to ascertain whether effective straying could happen beyond the feasible timeframe of this study, the lack of observable straying in this study suggests the behavior may be rare. These results emphasize the importance of effective management and conservation for each contemporary southern river population of shortnose sturgeon because recolonization through rescue effects for extirpated populations may not be realistic given currently observed migratory dynamics. Future research should continue to leverage long-term telemetry as well as complementary inter-population genetics to further elucidate whether straying occurs among any southern shortnose sturgeon populations.

Behavioral divergence among migration contingents could have profound implications for the life history characteristics of shortnose sturgeon. Migration and habitat use can influence adult body size, longevity, fecundity, and lifetime reproductive output [1, 76] and would therefore be expected to differ among shortnose sturgeon of each migratory contingent. For example, in partially anadromous salmonids, residents are commonly smaller than migrants [7, 10]. This has been similarly observed with shortnose sturgeon as river-residents in the Hudson River were shorter in length than those that moved downriver into the harbor [28], and migration appeared to offer superior opportunities for growth and body condition in the Gulf of Maine [30]. Further, separate contingents of steelhead and rainbow trout that foraged in freshwater, coastal, and marine environments displayed differential growth and lipid storage [64]. Potentially contingent-specific life history characteristics of shortnose sturgeon such as body size, and the hypothesized size-associated fecundity, suggest members of each contingent may contribute to their spawning population at varying rates. More research is needed to ascertain whether life history parameters are unique among resident (i.e., river resident vs. euryhaline resident) and migrant (i.e., short-distance vs. long-distance) contingents of shortnose sturgeon.

Similarities in coastal migratory patterns between southern and northern populations

Several similarities were apparent among the migratory patterns that we observed in southern shortnose sturgeon populations and what has been previously observed in populations at the northern extent of its range. For example, coastal migrations from Altamaha and Savannah river individuals typically took place during the putative spawning window (Figs. 2 & S2-S5), centered around returning to the putative spawning river, and often resulted in an upstream putative spawning run. This is analogous to descriptions of inter-river migrations in the Gulf of Maine, as these movements appeared to be a part of spawning migrations in this region as well [29]. Also similar to observations in northern populations, inter-annual variation existed in whether or not southern migrants migrated in a given year, relating to differential spawning periodicity. Some individuals made annual migrations across the coast to putatively spawn while others did so infrequently, often remaining resident to a non-natal system during non-spawning years (Figs. 2 & S2-S5). Return migrations to an individuals putative spawning river without subsequent upriver spawning movements were rare. A “skipped-spawning” strategy is hypothesized to occur when migrants are not of suitable body condition to invest properly in gonadic development for reproduction [77]. In these instances, migrants forgo migration, and therefore spawning, to instead increase body size and condition to maximize reproductive fitness at their next spawning opportunity [6, 77, 78]. The behavioral plasticity of migrant shortnose sturgeon suggests that individuals make the “decision” to migrate or remain resident annually based on the behavior that will maximize fitness given an individual’s condition at that time (i.e., the conditional strategy [1]). These commonalities in migratory timing and appartent link between coastal movements and putative spawning migrations emphasize the parellel behaviors of migrant shortnose sturgeon at the southern and northern extents of its range.

Another behavioral similarity among southern and northern populations was the consistency in migratory behavior throughout the study. Specifically, short- and long-distance migrants were repeatedly observed displaying the behaviors of their contingent, and residents were never observed migrating to another river despite years of tracking (Figs. 2 & S2-S5). The only exceptions to this were three long-distance migrants that displayed inter-annual deviations in the extent of their migration, at times migrating short distances (< 150 km) within a migration season before seeking freshwater summer refuge habitat (Fig. S4). However, each of these individuals consistently travelled long distances (> 150 km) from their spawning river in the migration season(s) prior to and after these observed seasons of short-distance migration (Fig. S4). No adult shortnose sturgeon in our study was observed to switch between short- and long-distance migratory tactics long-term (Figs. 2 & S2-S5), suggesting life-long migratory life history determination may occur early in life prior to sexual maturation. These observations align with the “threshold model,” which posits that all individuals within a partially migratory population have the underlying genetic predisposition to take on a particular migratory life history, but that biological and environmental conditions at an early life history “switch-point” must surpass a certain threshold to trigger life-long migratory behavior [9, 79]. This theory affords high behavioral complexity and plasticity within such populations and suggests population-wide behavior may be sensitive to environmental change. More research is needed to confirm at what developmental stage shortnose sturgeon undertake their first inter-river migration and the biological and environmental characteristics that influence this behavior.

Notable exceptions to these patterns were observed with several migrants from the Great Pee Dee River. Rather than migrating to and from the Great Pee Dee River for putative spawning purposes, some of the migrants in this river migrated both to and from the Santee River within the same migratory season (Fig. S5). Furthermore, eight of the 14 migrants from the Great Pee Dee River only migrated once, with these singular migratory events taking place in late fall of 2018, just after Hurricane Florence made landfall in this region. These observations suggest that some Great Pee Dee River short-distance migrants take advantage of the Santee River’s proximity (~ 20 km in-water distance), potentially for enhanced winter foraging or as refuge during times of environmental stress without spending additional time beyond their putative spawning river. The behavioral variation of Great Pee Dee River shortnose sturgeon compared to those from the Altamaha, Savannah, and Gulf of Maine rivers emphasize the behavioral plasticity of the species. Future research is needed to identify whether additional differences exist among migrants of separate populations.

Future directions

Southern shortnose sturgeon migratory contingents were largely defined by observed disparities in space use in and beyond the spawning river. However, traditional diadromous categorizations are defined strictly by how an individual uses freshwater and saltwater biomes. Although we accounted for the variable salinity intrusion within each of our study systems when identifying between euryhaline and river resident behavior (Table S2; Figs. S7-S10), finer-scale observations of salinity mosaics in relation to downriver movements could further elucidate resident contingent structure and diadromous behaviors. Further, sample sizes for each population of study (n < 100) were relatively low and may not be in proportion to the river-specific population size. Therefore, multinomial results represent the importance of spawning population for determining migration contingent, and broad-scale population-level trends in the relative presence of each contingent within each population. Future long-term tracking studies are needed to attain a greater understanding of these diadromous migratory life histories.

Although our field efforts and migratory classifications were exhaustive and informed by other studies of shortnose sturgeon and migratory fishes, our study could be an underestimation of the true migratory potential of southern shortnose sturgeon populations. For example, migrant designations in the partial migration analysis were conservative because we could only account for the migrations that we observed via acoustic telemetry. Therefore, some individuals may have migrated without being detected, migrated post-transmitter expiration, or succumbed to mortality during migration before being detected in another river system. Additionally, expanding telemetry efforts to be more comprehensive and consistent could similarly benefit future studies. In particular, maintaining the receiver stations that were removed in Timeframe 2 as well as tagging in other coastal rivers that may support smaller populations of shortnose sturgeon (e.g., Ogeechee and Satilla rivers, GA) could lead to additional detection of migrants. Overall, this study provides detailed descriptions of the sex- and population-level patterns in southern shortnose sturgeon migration ecology, and future efforts informed by this work could provide even more nuanced detection of inter-river migration in this region.

Conclusions

Although partial migration has been described for shortnose sturgeon populations at the northern extent of its range, our work details the prevalence of partially migratory behavior at the species’ southern extent. The presence of these behaviors throughout its range demonstrates that partial migration is a consistent and important aspect of shortnose sturgeon life history and ecology. Furthermore, the identification of several behaviorally distinct contingents indicates novel levels of migratory complexity for shortnose sturgeon. The migratory behaviors and coastal movement capabilities described herein represent substantial updates to shortnose sturgeon ecology, and these updates must be reflected in management practices if the species is to be properly conserved.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

JDN conceived the study, curated and analyzed the data, and led the writing of the manuscript. MJH and AGF provided funding and logistical support, supervision, data acquisition, and wrote the manuscript. EW and WCP secured and distributed funding, managed data collection, and edited the manuscript. DJF and TLD performed genetic analyses for sex assignments and edited the manuscript. ECI managed data collection, curated early telemetry data, and edited the manuscript.

Funding

Funding for this project was provided by the National Marine Fisheries Service, Grant No. NA10NMF4720036, NA14NMF4720012, NA22NMF4720099, and NA23NMF4720166, and by Duke Energy as part of the Yadkin-Pee Dee River Hydroelectric Project (FERC Project No. 2206-030), and by the United States of America Army Corps of Engineers, Savannah District as part of the Savannah Harbor Expansion Project, and by United States of America Army Corps of Engineers, Charleston District.

Data availability

The datasets generated and analyzed during the current study are not publicly available due to the sensitivity surrounding endangered species locations and because these data are owned by several institutions. Data underlying this manuscript can be made available upon reasonable request.

Declarations

Ethics approval and consent to participate

All research methods were conducted pursuant to National Marine Fisheries Service Permits No. 16482-01, 15677, 16442, 16375-01, and 20528-05, as well as under all scientific collection permits from the Georgia Department of Natural Resources. Research was conducted under animal use protocols approved by the University of Georgia’s Institutional Animal Care and Use Committee.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.