Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 742 2016 SOUTHEASTERN NATURALIST 15(4):742–759 Behavioral Patterns of Common Bottlenose Dolphins (Tursiops truncatus truncatus) Within the Galveston–Port Bolivar Ferry Lane Alexandria E. Rivard1,*, Frances P. Gelwick2, and Wyndylyn von Zharen3 Abstract - The objective of this study is to assess Tursiops truncatus truncatus (Common Bottlenose Dolphin) group behavior as a function of spatial, temporal, and vessel proximity variables within the Galveston–Port Bolivar ferry lane, in lower Galveston Bay, TX. This area is subjected to vessel traffic entering the Houston, Texas City, and Galveston ship channels and at risk for environmental accidents. We used the Galveston–Port Bolivar ferries as a platform of opportunity to observe group behavior within the ferry lane. We conducted 1412 hours of observation between 1 June and 30 November 2013 and then utilized canonical correspondence analysis to evaluate behavioral state as a function of spatial, temporal, and vessel-proximity variables. Principal response curve (PRC) analysis showed significant variation in behavioral states over time in the Bolivar and Galveston zones as compared to the passage zone. Chi-square goodness-of-fit tests showed significant deviations from expected behavior observation across zone and time block. These findings demonstrate finescale behavioral variability in an area of high anthropogenic activity. Introduction The number of vessels at sea and in major ports has risen steadily since the 1950s (Ross 2005). Significant maritime shipping occurs in the Gulf of Mexico, both offshore and within the Intracoastal Waterway, and accounts for $129 billion of cargo movement annually (Adams et al. 2004). Several major shipping lanes exist within the Gulf of Mexico, primarily those through the Florida Straits and the ports of Houston and New Orleans (Azzara 2012). In addition to commercial traffic, more than 5 million households in Alabama, Florida, Louisiana, Mississippi, and Texas participated in recreational boating in 2012 (USCG 2012). Despite the consistent commercial and recreational traffic, the northern Gulf of Mexico is home to 31 stocks of Tursiops truncatus truncatus (Montagu) (Common Bottlenose Dolphin) within its bays and estuaries, which are regularly exposed to encounters with vessels (Waring et al. 2015). Exposure to vessel traffic can alter habitat use of Common Bottlenose Dolphins. For instance, a significant decline in dolphin abundance was observed for vessel tourism sites in Shark Bay, Australia, when compared with control sites where traffic was lower (Bejder et al. 2006). In Milford Sound, New Zealand, boat traffic was determined to be the dominant factor responsible for a decline in dolphin visits 1Department of Marine Biology, Texas A&M University at Galveston, Galveston, TX 77553. 2Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, TX 77843. 3Department of Marine Science, Texas A&M University at Galveston, Galveston, TX 77553. *Corresponding author - arivard@email.tamu.edu. Manuscript Editor: Jeremy Pritt Southeastern Naturalist 743 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 (Lusseau 2005). In the short-term, animals may dive longer to avoid vessel encounters or leave the area as boats arrive (Lusseau 2003). In Sarasota Bay, FL, resident Common Bottlenose Dolphins exhibited longer inter-breath intervals when approached by boats compared to when there were no boats within 100 m (Nowacek et al. 2001). When approached by a vessel, animals changed direction and swimming speed more frequently than when there were no boats in the vicinity (Nowacek et al. 2001). However, not all behavioral responses are avoidant. In some instances, the animals may be attracted to vessels to ride the bow wave of the ship (Fish and Hui 1991) or forage behind commercial trawlers (Leatherwood 1975). Nevertheless, sound exposure, stemming from watercraft and other sources of marine noise, may cause a rise in stress hormones and potentially lead to long-term health problems (Romano et al. 2004). The type of vessel and its operation are also important variables impacting Common Bottlenose Dolphin behavior (Nowacek et al. 2001, Mattson et al. 2005, Miller et al. 2008, Weilgart 2007). In Hilton Head Island, SC, group behavior changed following 55% of encounters with watercraft, 67% of encounters with jet skis, 100% of encounters with shrimp boats, and 11% of encounters with large ships (Mattson et al. 2005). In the Pelagie Archipelago in Sicily, Italy, animals were disturbed by fast-moving vessels but not by sailboats (Papale et al. 2012). Approaches by powerboats can frequently disrupt foraging behavior (Lemon et al. 2008). Such findings indicate that behavioral disruption is not simply a function of the size or noise level of a ship, but also the manner in which it is operated. The Marine Mammal Protection Act of 1972 (MMPA) (16 U.S.C. § 1361 - 1421h) prohibits approaching marine mammals with a vessel, but enforcement has proven challenging and is often inconsistent due to the spatial scale over which violations may occur (Roman et al. 2013). While intentional take of marine mammals is prohibited in the United States, unnecessary accidental take of individuals should be avoided to the extent possible (Roman et al. 2013). Beneficial foraging habitat can be associated with areas of high vessel traffic. In some areas of the Gulf of Mexico, dolphins have shown a preference for foraging in dredged channels as compared to natural seagrass beds because their prey are unable to hide in dredged areas where vegetation was removed (Allen et al. 2001). Habitat preference depends on a myriad of other factors, but availability of desirable prey in the vicinity of some Gulf of Mexico habitats does appear to outweigh the potential negative impacts of heavy vessel traffic (Allen et al. 2001). However, the level of traffic activity that may cause animals to leave an otherwise desirable habitat is unknown. The particularly heavy and steady stream of ships, including regular ferry traffic, in Galveston, TX (Merrick and Harrald 2007) presents a venue for regular interactions between vessels and animals. Evidence suggests that habituation to anthropogenic activity can have adverse effects on cetacean hearing, as well as limit the effectiveness of their evasive responses to vessels (Richardson and Würsig 1997). Moreover, because shipping corporations operating in this area often transport hazardous cargo and traffic is heavy, this location represents an area generally deemed Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 744 as unsafe for vessel operation (Merrick and Harrald 2007). Documenting behavioral patterns under current conditions will provide a behavioral standard against which behavior can be compared if vessel traffic changes markedly or an accident occurs. Within the Gulf of Mexico, fine-scale population structure and patterns of long-term site fidelity have been documented (Miller and Baltz 2010, Sellas et al. 2005). The Common Bottlenose Dolphin population in Galveston Bay is relatively small. In one survey, approximately 240 individuals were recorded, though many were not sighted a second time and may have been transient (Fertl 1994). Seasonal fluctuations in abundance exist: spring and fall surveys showed higher encounter rates than summer and winter surveys (Fertl 1994). Weller (1998) supports the low encounter rate for the winter months, noting that Common Bottlenose Dolphins move toward inshore waters in the summer in the northern Gulf of Mexico, though it is unclear why the encounter rate from Fertl (1994) was low in the summer or if abundance fluctuations can be tied to vessel traffic. There were no significant patterns of preferred associations with other individuals in Galveston Bay, suggesting high group fluidity within the region (Bräger et al. 1994). This study aims to describe patterns for group behaviors of Common Bottlenose Dolphins as a function of spatial location, group characteristics, month and time of day, and characteristics of vessel traffic. Using ferry vessels as our observation platform, we collected data pertaining to dolphin groups within the Galveston–Port Bolivar ferry lane, which crosses the entrance to the Houston ship channel. These data describe group behavior in an area of high vessel traffic, and can serve as a standard for group behaviors in the event of an environmental perturbation. Materials and Methods Field site description Galveston Bay is the largest estuary in the Gulf of Mexico, and home to 3 major ports: Houston, Texas City, and Galveston. These ports are connected by the Houston Ship Channel, which enters Galveston Bay between the eastern tip of Galveston Island and the western end of the Bolivar Peninsula (Steichen et al. 2012). This entry point is crossed continually throughout the year by the Galveston– Port Bolivar ferry. Approximately 7000 ships visit the Port of Houston via the Houston Ship Channel annually (Merrick and Harrald 2007). The level of traffic through the port is expected to rise following the completion of the Panama Canal expansion in 2015; container ship traffic alone is expected to increase by 15 percent (Harrison and Trevino 2013). The Galveston–Port Bolivar ferry bridges the gap in State Road 87 by traveling 2.7 miles between the Galveston ferry dock and Port Bolivar ferry dock (Fig. 1; Weisbrod and Lawson 2003). Ferries cross the same route year round, 24 hours a day. A one-way trip from dock to dock takes approximately 20–30 minutes. Figure 1 (following page). The 3 zones of observation: Galveston ferry dock (upper circle), the passage channel (lines connecting circles), and Bolivar ferry dock (lower circle). (Ferry channel aerial map from Texas Natural Resources Information System.) Southeastern Naturalist 745 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 746 Between 1 and 5 vessels operate concurrently, with more vessels operating during times of peak travel. The current ferry fleet consists of 6 vessels, and in 2000, accommodated over 2 million vehicles and more than 6.5 million passengers (Weisbrod and Lawson 2003). Field observations Data collection took place continually from 7:00 (or sunrise if later) to 19:00 (or sunset if earlier) between 1 June 2013 and 30 November 2013, weather permitting. We selected this time period to include periods of high and low sighting rates and to exclude the springtime calving and breeding season (Fertl 1994, Weller 1998). We did not conduct observations when visibility was less than 100 m. All observations were made from the front of the outdoor viewing deck on the second level of the Galveston–Port Bolivar ferry. This platform was ~20 feet above water level and provided a 270-degree view. In cases of rain or high wind when visibility was still greater than 100 m, we conducted observations from the forward windows of the second-level cabin (visibility 180 degrees), alternating between port and starboard sides in 2-minute intervals. We collected data in 3 spatial zones: the passage channel (P), Galveston dock (G), and Bolivar dock (B) (Fig. 1). We defined the docks as the area within a 100-m radius of the ferry landing, and the passage channel as the area of ferry operation more than 100 m from the docks (Fig. 1). The spatial zone recorded for an observation was the location of the majority of the individuals in the group being observed. In all zones, we documented only dolphin groups within 100 m of the ferry vessel. When the group was split evenly at the boundary between 2 observation zones, we recorded observations only for the group members within the zone nearest to the ferry. Groups were defined using a 10-m–chain rule as per Smolker et al. (1992), wherein we considered all animals within 10 m of another animal part of the same group. We defined calves as those individuals estimated to be less than 1.5 m in length (Leatherwood and Reeves 1989). For all sighted groups, we visually estimated the distance from the dolphin in the group nearest to (a) the shore; (b) all vessels within 100 m of the group, including other ferries on the route; and (c) the observer’s ferry. Distance estimation, particularly over the water, is rife with human error (Baird and Burkhart 2000). To minimize such error, all observers were trained in distance estimation, and distances were estimated as ranges (Baird and Burkhart 2000). Lemon et al. (2006) estimated the distances at which dolphins change behavior as a vessel approaches, and ranges were established in accordance with these findings. Therefore, groups more than 100 m away and vessels more than 100 m from the group were not documented. All distances were categorized as: very close (V; ≤ 10 m), close (C; 10–30 m), intermediate (I; 30–50 m), far (F; 50–100 m), or open water (O; ≥100 m). The open-water category was used only to document the distance of the group from shore. We recorded the predominant group activity (PGA) (by ≥50% of the individuals in the group) when first sighting a group and for every 2-minute interval thereafter Southeastern Naturalist 747 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 if the group remained in sight and within 100 m (Mann 1999). When assessing a small population of a highly mobile species in a dynamic environment with heavy vessel traffic, it is often difficult to determine when observations become independent. The 2-minute interval selected provided sufficient time for PGA and other variable conditions to change and individuals to join or leave the group, as group fluidity in Galveston Bay is high (Bräger et al. 1994). However, some problems with independence may persist. PGA was characterized as: resting (R), foraging (F), socializing (S), or travelling (T) as described in Ballance (1992). We defined resting as low activity with no coordinated direction of movement by the group; foraging as the active pursuit of prey; socializing as observable body contact, either positive or aggressive; and travelling as the coordinated movement of the group in 1 direction (Ballance 1992). All observers were trained extensively in behavioral identification first in the laboratory and then in the field, until they consistently identified the correct behavior for all groups sighted. For each observation, we recorded the following data: time of day sighted; zone of sighting (P, B, or G); total number of individuals in the group; number of calves; number of vessels operating an engine in each distance category (V, C, I, or F); ferry proximity to the group (V, C, I, or F); distance of the group from shore (V, C, I, F, or O). Observations were distributed equitably across 3 time blocks: morning (M; 7:00 [or sunrise]–11:00), afternoon (A, 11:00–15:00), and evening (E, 15:00–19:00 [or sunset]). In the latter portion of the field season when sunrise occurred after 7:00 and sunset occurred before 19:00, we initiated or terminated observations based on the sunrise and sunset times. Analysis The aim of this study is to describe Common Bottlenose Dolphin behavior as a function of spatial, temporal, and vessel-proximity variables. To that end, we used multivariate analysis as a means to evaluate PGA as a function of all candidate explanatory variables simultaneously. We conducted multivariate analysis using CANOCO v5.0 and then evaluated variation in PGA based on time block and zone using Pearson’s chi-square tests in R version 3.2.3 (R Core Team 2015). In order to create a quantifiable measure of total vessel traffic, we converted the count of vessels in the vicinity of a group to an ordinal “vessel score”. The count of vessels in each distance range V, C, I, or F is given by NV, NC, NI, or NF, respectively. Vessels very close (V) were weighted by 4, close (C) by 3, intermediate (I) by 2, and far (F) by 1. The ferry score S was assigned the value of 4 if very close (V) to the group, 3 if close (C), 2 if intermediate (I), and 1 if far (F). The vessel score was then calculated using the equation: Vessel score = 4 NV + 3 NC + 2 NI + 1 NF + S We performed canonical correspondence analysis (CCA) using CANOCO v5.0 (ter Braak and Šmilauer 2012) to evaluate the association of values for candidate explanatory independent variables with respect to PGA at each sighting (dependent variable). Such constrained ordination methods result in a best-fit model of relationships among variables in a dataset; CCA axes correspond to the directional Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 748 gradients of greatest variation in the dependent variable as explained by combinations of explanatory variables (Van den Brink and ter Braak 1999). We used the “Forward Selection” procedure in CANOCO to rank the environmental variables based on the strength of their relationship to gradients in behaviors (ter Braak and Šmilauer 2012). The initial analysis identified individual significant explanatory variables in order to quantify their total combined explanatory value. We then classified explanatory variables into temporal, spatial, and vessel-proximity categories to partition the variation in PGA that can be explained uniquely by each category (Table 1). We assessed the seasonal trends in behaviors using a partial redundancy analysis (RDA) to construct principal response curves (PRC). The principal response curves analysis is a multivariate method for the analysis of repeated measurement designs, and is designed to test and display treatment effects that change across time. It is based on reduced-rank regression (redundancy analysis) that is adjusted for changes across time in the base/reference treatment. This protocol allows the method to focus on the time-dependent treatment effects. The temporal factor (time) is set as a covariate in the analysis so that the main effect of treatment (location of observations, with P, passageway, as the base/reference and G and B, Galveston and Bolivar, as the terminal port locations to be compared to P) and the interaction with time is included in the ordination model. Each repeatedly recorded location forms a separate whole-plot and utilizes the observations made in time for its splitplots. This permutation test then permutes whole-plots, but not the split plots. The principal component of this analysis (here axes 1 and 2) is plotted against time in the diagram. (Van den Brink and Ter Braak 1998, 1999). All multivariate tests were run using Monte Carlo simulations in CANOCO 5 software (ter Braak and Šmilauer 2012) and calculating the F-ratio at the 5% significance level (comparing the original data to output from 499 permutations for tests). We used pseudo-F values for partial RDA when controlling for the effects of covariates. In consideration of the family-wise Type I error rates for multiple testing, we implemented the false discovery rate in the software program using the approach of Benjamini and Hochberg (1995). Table 1. The variable levels included in variance partitioning tests of conditional effects of 3 types of variable groups: temporal, spatial, and vessel proximity to the group. DS = distance to shore, DF = distance to ferry. Temporal Spatial Vessel proximity Date Bolivar ferry dock DF very close Morning Passage DF close Afternoon Galveston ferry dock DF intermediate Evening Ds very close DF Far Sight time Ds close Total vessel score Ds intermediate # vessels very close Ds far # vessels close Ds open water # vessels intermediate # vessels far Southeastern Naturalist 749 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 We used Pearson’s chi-square tests to determine how occurrence of PGA deviated from the expected value on the basis of zone and time block. We calculated expected values as the product of the row (zone and time block, respectively) and column (PGA) totals in the contingency table divided by the sample size. The chisquare test determines whether PGA deviated significantly from these expected values. This analysis allows for more precise identification of where deviations in behavior occur within spatial and temporal metrics. Results Between 1 June 2013 and 30 November 2013, we conducted 1412 hours of observations. We documented dolphin groups on 24,780 occasions, in which calves comprised ~9% of the individuals sighted. Group size ranged from 1 to 74 individuals, and the number of calves in a group ranged from 0 to 15 (Fig. 2). The greatest number of groups was sighted in the passage zone during the afternoon. The fewest number of groups was documented at the Galveston ferry dock during the evening (Table 2). Figure 2. Frequency histogram of group size (top) and number of calves per group (bottom). Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 750 CCA with forward selection of variables explained 10.2% (adjusted for degrees of freedom) of the total variation in behavior states across all observations (Table 3). Variation is due to differences associated with the likelihood that a given behavior occurs in conjunction with the explanatory variable. All explanatory variables were significant (Table 3). The largest percentage of the total variation in behavioral state due to a single variable was from the Bolivar ferry dock, which accounted for 2.2% of the total variation (21% of the total explained variation). The open-water distance to shore variable contributed an additional 1.7% of the total variation (16.4% of the explained variation). With variance partitioned into our 3 sub-groups of explanatory variables (spatial, temporal, and vessel proximity), they combined to significantly explain 8.5% of the variation in behavior states (Table 4) as follows: spatial variables uniquely accounted for 3.4% (40.0% of total explained variation), temporal variables uniquely accounted for 2.6% (30.6% of total explained variation), vessel-proximity variables uniquely accounted for 1.8% (21.2% of total explained variation), and variation shared among these groups accounted for 0.7% (8.2% of total explained variation). Table 2. The number of dolphin groups documented in each zone by time block. Time block Bolivar ferry dock Galveston ferry dock Passage Morning 2286 2090 3706 Afternoon 3692 1305 3961 Evening 3318 658 3764 Table 3. The results of the forward-selection global permutation test CCA showing the total percentage of variation in behavior and percentage of explained variation that is attributable to each variable. All selected variables explain 10.3% (adjusted for degrees of freedom to 10.2%) of the total variation in behavior. DS = distance to shore, DF = distance to ferry. % of total % of explained Name variation variation pseudo-F P P (adj) Bolivar ferry dock 2.2 21.0 548.0 0.002 0.01480 DS Open water 1.7 16.4 436.0 0.002 0.01480 Date 1.5 14.2 382.0 0.002 0.01233 Group size 1.3 13.0 356.0 0.002 0.01057 Total vessel score 1.1 11.1 307.0 0.002 0.00925 Morning 0.8 7.4 206.0 0.002 0.00822 #Calves 0.5 4.5 125.0 0.002 0.00740 Df very close 0.3 3.3 92.6 0.002 0.00673 Ds very close 0.2 2.1 58.5 0.002 0.00617 #Vessels very close 0.1 1.3 38.0 0.002 0.00569 Sight time 0.1 1.1 29.8 0.002 0.00529 Ds far 0.1 1.1 29.8 0.002 0.00493 #Vessels close 0.1 1.1 31.0 0.002 0.00462 Ferry distance less than 0.1 0.7 20.6 0.002 0.00435 Passage less than 0.1 0.7 18.9 0.002 0.00411 Evening less than 0.1 0.6 17.5 0.002 0.00389 Ds intermediate less than 0.1 0.2 7.1 0.002 0.00370 Ds close less than 0.1 0.2 7.1 0.002 0.00370 Df far less than 0.1 0.2 6.8 0.002 0.00322 Southeastern Naturalist 751 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 The CCA ordination diagrams simultaneously plot the significant associations of explanatory variables and behavior states (Fig. 3). CCA axis 1 accounted for 5.7% of total variation in behavioral state (Fig. 3); axis 2 for 3.6% (Fig. 3); and axis 3 for 1% (not shown). Centroids indicate the location of optima for behavioral states as related to explanatory variables. Longer vector arrows indicate a quantitative explanatory variable that is more strongly associated with the behavior state centroid closest to it (positively in the direction of the arrow head, and negatively in the opposite direction) along the axis. Centroids for categorical explanatory variables located farthest from the center of the ordination are most strongly associated with the behavior centroid closest to it along the axis. For example, along CCA axis 1, foraging and traveling are the 2 behavioral states best explained and most strongly contrasted with one another, as they sit on opposite sides of the axis. Resting was best explained along CCA axis 2 and negatively associated with time spent in traveling or foraging (Fig. 3). Foraging, plotted towards the right on CCA axis 1, therefore was most strongly associated with morning observations, larger group sizes, greater numbers of calves, and vessels that were close to the group (Fig 3). Conversely traveling, plotted towards the left on CCA axis 1, therefore was associated with open water and groups very close to the ferry in the passage zone in open water (Fig. 3). In the evening, dolphin groups were resting or socializing farther from the ferry and farther from shore. Along CCA axis 2, resting was strongly positively associated with observations in the Bolivar zone and groups farther from the ferry in the evening, and negatively associated with total vessel score (Fig. 3). PRC analysis showed significant trends over time (Fig. 4) for variation in behavior states, comparing spatial distribution of observations in the 2 ports to observations in the passage zone. When combined, the temporal and spatial explanatory variables accounted for 8.2% (adjusted 7.0%) of the total variation in behavioral state. PRC axis 1 accounted for 60% of the total explained variation. PRC axis 1 shows the dominant behavioral trends over time; behaviors in Galveston and Bolivar zones were more similar to one another in June, July, and August, primarily consisting of resting and foraging; socializing was more commonly observed in the passage zone (Fig. 4A). In July, September, and November, notable increases in traveling occurred in both ports, but in October foraging and resting increased in the Galveston zone; smaller increases in traveling occurred in the Bolivar zone during September, October, and November (Fig. 4A). PRC axis 2 accounted for 35.5% of explained variation in behavior and indicated the strongest contrasts Table 4. Percentage of the variation in observed dolphin behaviors that can be explained by each subgroup of variables (spatial, temporal, vessel proximity), equally by any of the groups (shared), and by all 3 groups combined (total explained). Group % of variation % of explained variation Pseudo-F value P value Spatial 3.4 40.0% 146.0 0.002 Temporal 2.6 30.6% 166.0 0.002 Vessel proximity 1.8 21.2% 66.1 0.002 Shared 0.7 8.2% 136.0 0.002 Total explained 8.5 100.0% Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 752 in behaviors between the two ferry docks, especially in June and July when foraging and travel were more common in Galveston, and socializing and resting were more common in Bolivar (Fig. 4B). Behavioral trends were more homogeneous Figure 3. Results of canonical correspondence analysis of behavioral s t a t e (l a rge circles) as the dependent variable. Explanatory variables closer to the centroid for a dependent variable are more likely to be associated with the observation of that behavior. Explanatory variables significant on CCA axes 1 and 2 were distance to ferry (Ferry- Dis; vector head plotted in upper left quadrant and open triangles for categorical variables V-, C-, I-, and F-DistFerry plotted from upper left to lower right quadrants), zone (Bolivar, Galveston, and Passage, dark filled triangles), vessel proximity (VessV4, VessC3, VessI2, and VessF1; vector heads plotted in upper right quadrant), time of day (Morning, Aftnoon, Evening;gray filled triangles from upper right to lower left quadrants), month (gray filled triangles plotted from upper left to lower and upper right quadrants), total vessel score (TotVessS; vector head plotted in upper right quadrant), group size (Grp- Size; vector head plotted in lower right quadrant), number of calves (Num-Calf; vector head plotted in upper right quadrant), and categorical distance to shore (V-, C-, I-, and ODistShore; open triangles plotted from upper right to lower right quadrants). Total variation for axes 1, 2, and 3 were, respectively, 5.7%, 3.6%, and 1.0% of total variation; permutation tests for Axis 1 and all 3 canonical axes together were, respectively, pseudo-F= 1505, P = 0.002 and pseudo-F = 150, P = 0.002. Calf; vector head plotted in upper right quadrant), and categorical distance to shore (V-, C-, I-, and O-DistShore;open triangles plotted from upper right to lower right quadrants). Total variation for axes 1, 2, and 3 were, respectively, 5.7%, 3.6%, and 1.0% of total variation; permutation tests for Axis 1 and all 3 canonical axes together were, respectively, pseudo-F= 1505, P = 0.002 and pseudo-F = 150, P = 0.002. Southeastern Naturalist 753 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 over time for observations in Bolivar than Galveston, being more similar to those in the passage, and consisting primarily of socializing and traveling (Fig. 4B). PRC axis 3 accounted for only 4.5% of explained variation and was not significant. PRC results show that, across all behaviors, traveling was most strongly associated with observations in the passage zone, resting was most strongly associated with Figure 4. Principal Response Curve(PRC) analysis results using the passage zone (P) as the baseline (horizontal line plotted at 0.0) for comparison of temporal trends in behaviors associated with each zone. The vertical axis at the right side of the diagram shows the gradient for behaviors on the ordination axis aligned with the respective PRC axis; Forage (For), Travel (Tvl), Socialize (Soc), and Rest (Rst) across time periods (months indicated by arrows along time axis) for Bolivar ferry dock (B) and Galveston ferry dock (G) zones on the first (PRC1) and second (PRC2) principal responsecurve axes. The first and second axes, respectively, depict 60% and 36% (permutation tests respectively pseudo-F = 1263, P = 0.002 and pseudo-F = 770, P = 0.002) of the total explained variation in behavior. Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 754 observations at the Bolivar ferry dock, and foraging was most strongly associated with observations at the Galveston ferry dock. Chi-square goodness-of-fit tests detected significant deviation from the expected occurrence of behavioral state across zone (χ2 [6, n = 24,780] = 2175.75, P < 0.01) and time block (χ2 [6, n = 24,780] = 965.39, P < 0.01), and was consistent with the PRC trends. We observed resting behavior more frequently than expected and traveling less frequently than expected in Bolivar (Table 5). Traveling was observed more frequently than expected in the passage (Table 5). In Galveston, socializing was less frequent than expected, and foraging more frequent (Table 5). Across time blocks, foraging was more frequent than expected in the morning and less frequent in the evening (Table 6). Resting was observed less frequently than expected in the morning and more frequently in the evening (Table 6). Discussion This study represents a relatively fine-scale analysis of Common Bottlenose Dolphin behavior in an area with high levels of vessel activity. While several studies suggest that heavy vessel traffic can cause animals to vacate or avoid the area (Bejder et al. 2006; Lusseau 2003, 2005), Common Bottlenose Dolphins are sighted regularly in Galveston. However, long-term exposure to vessel traffic can cause hearing loss and leave individuals vulnerable to long-term health problems (Richardson and Würsig 1997, Romano et al. 2004). Ultimately, the variables assessed accounted for 8.5% of the variance in behavior. Each variable was statistically significant, though the percentage of total variation explained was low. These findings suggest that while shore position, time of day, and vessel-proximity variables are important factors in group behavior, they represent only a small portion of the contributing variables in a complex ecosystem. The fluid social structure documented in Galveston Bay (Bräger et al. 1994) may also influence group behavioral changes in a complex way that is difficult to Table 6. Observed (O) and expected (E) occurrence of each behavioral state based on time block. Significant deviations in behavior were detected (X2 [6, n = 24,780] = 965.39, P < 0.01). Foraging Resting Socializing Traveling Morning O:3804, E:2791.5 O:1141, E:1526.1 O:457, E:640.9 O:2680, E:3123.5 Afternoon O:2771, E:3094.1 O:1673, E:1691.5 O:769, E:710.3 O:3745, E:3462.1 Evening O:1984, E:2673.4 O:1865, E:1461.5 O:739, E:613.8 O:3152, E:2991.4 Table 5. Observed (O) and expected (E) occurrence of each behavioral state based on the zone. Significant deviations in behavior were detected based on the zone of observation (X2 [6, n = 24,780] = 2175.75, P < 0.01). Foraging Resting Socializing Traveling Bolivar O:3077, E:3210.8 O:2833, E:1755.3 O:860, E:737.2 O:2526, E:3592.7 Galveston O:2004, E:1400.0 O:454, E:765.3 O:104, E:321.4 O:1491, E:1566.4 Passage O:3478, E:3948.3 O:1392, E:2158.4 O:1001, E:906.5 O:5560, E:295.3 Southeastern Naturalist 755 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 capture in brief observations. Variance partitioning showed variation in behaviors was related primarily to spatial (40.0%) and temporal (30.6%) variables, and to a lesser extent, to vessel proximity to groups (21.2%). These findings may suggest that the local population has been to some extent desensitized to vessel traffic. PRC analysis indicated trends in behaviors across months. The 3 observation zones deviate most from each other during June, July, and August. Dolphins in Galveston and Bolivar tended to be resting and foraging, whereas socializing was more consistently associated with observations in the passage zone. Particularly early in the summer, dolphins in Galveston were more commonly engaged in foraging and those in Bolivar were more commonly engaged in socializing. Behavior in Galveston and Bolivar diverged most from the passage during the summer months, possibly due to recreational boater disruption of these activities in the passage. Vessels are known to disturb dolphin groups, increasing dive time and inter-breath intervals (Lusseau 2003, Mattson et al. 2005, Nowacek et al. 2001). The chi-square test showed that groups were more likely to rest in Bolivar and forage in Galveston. Groups in Galveston may be seeking areas of less traffic such as the Bolivar ferry dock to rest in order to avoid such disruptions, though this does not necessarily indicate the groups are stressed by vessel traffic. The overall trend suggests that the Galveston ferry dock and Bolivar ferry dock serve as areas for dolphins to rest, forage, and socialize, particularly during the summer months when overall vessel activity is highest. The impact of vessel traffic on group behavior is further illuminated by vessel activity within various distance ranges. Groups were frequently recorded foraging very close (V) and close (C) to vessels, likely indicating that vessel-associated foraging is common within the survey area and contributing to the high total vessel score associated with foraging behavior. However, when the ferry was very close to the group, traveling behavior predominated. The fact that groups foraged in the presence of trawlers but traveled in the presence of the ferry suggests that the type of vessel and its operation influence group behavior (Mattson et al. 2005, Miller et al. 2008, Nowacek et al. 2001, Weilgart 2007). When vessels were far or intermediate from the group and when the ferry was far from the group, socializing and resting behaviors predominated, likely due to the minimal vessel disruption. Further research is necessary to determine how the specific type and activity of the vessel impacts group behavior in this area. Distance to shore was significantly related to group behavior. At the open-water distance (>100 m), traveling behavior dominated. The open-water distance to shore was only possible in the passage as docks were defined as 100-m circles around the dock center. Because traveling was often observed in the passage zone, it is unsurprising that the open-water distance is tightly correlated with traveling. Groups very close to shore were most often observed foraging. This finding may indicate a predilection for foraging in shallower waters towards the edges of the dredged channels, as observed in Clearwater, FL (Allen et al. 2001). In shallow water, prey’s vertical mobility is limited and may make hunting more efficient. This finding may also indicate a preference for shrimp-boat associated foraging, which tends to occur Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 756 in shallower areas. Finally, groups intermediate (30–50 m) and far (50–100 m) from shore were most often socializing or resting. These ranges likely offers the best habitat to engage in these behaviors with minimal disruptions from passing vessels, while also avoiding the hazards of shallower water such as rocks, debris, or water too shallow to allow for swimming. A notable shift away from foraging in the evening and with later sighting times was documented in both the CCA and chi-square results. This preference for foraging in the morning has been observed in several Common Bottlenose Dolphin populations globally (Alford 2005, Allen et al. 2001, Bräger 1993). Social behavior occurred with the expected frequency throughout the day, and may be indicative of the high group fluidity previously documented in Galveston (Bräg er et al. 1994). A shift in behavior, away from traveling and towards foraging, resting, and socializing, occurred moving from summer into fall. Recreational vessel traffic tends to be high in Galveston during June, July, and August. Watercraft and jet skis tend to alter behavior more frequently than large ships (Mattson et al. 2005), and powerboat approaches may cause groups to cease foraging to avoid the oncoming vessel (Lemon et al. 2008). As the presence of these recreational vessels decreases in the fall, it is logical that avoidance behavior such as traveling observed in the presence of these types of vessels would decline and foraging, resting, and socializing would consequently increase. Further study on how the type and activity of the vessel impacts group behavior will help to determine whether recreational traffic is more likely to alter group behavior than commercial boats. Group size was largest when groups were engaged in foraging, and the number of calves was higher in larger groups. While larger groups may simply be more likely to have calves in them, this finding also may be a function of the high level of ship traffic in the survey area. Mothers with calves are considered the most vulnerable individuals (Buckstaff 2006). Because the ferry lane is constantly subjected to a high level of vessel activity as a function of its location at the mouth of Galveston Bay, adults with calves may be seeking larger social groups for added protection in response to the greater number of vessels. However, the higher number of calves in large groups may be an artifact due to mothers with calves bringing their young into foraging groups as a matter of necessity, because calves remain with their mothers for 3 to 6 years (Buckstaff 2006). A third potential explanation is that calves are learning to forage through observation in these large groups. Observational learning was noted in wild and captive Common Bottlenose Dolphins (Sargeant and Mann 2009, Yeater and Kuczaj 2010). The dolphin population in Galveston warrants further research; behavioral patterns beyond those detected in this study are likely to emerge with further study. Evaluating the impacts of anthropogenic activity on this population and in similar high-activity areas will improve understanding of how vessel traffic influences group behavior, and how groups in areas of heavy traffic are affected. Particularly because Galveston is a very active commercial port, establishing a standard for behavioral patterns there will prove useful in the event of an accident that impacts the marine environment. Southeastern Naturalist 757 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 Acknowledgments We thank the Texas Department of Transportation and the Galveston–Port Bolivar ferry crewmembers and security officials for permitting the use of their vessels, and research assistants J. Anklam, M. Bache, K. Clark, G. Giannotti, K. Gillis, E. Jones, A. Love, and L. Miller for their assistance with data collection. Literature Cited Adams, C.M., E. Hernandez, and J.C. Cato. 2004. The economic significance of the Gulf of Mexico related to population, income, employment, minerals, fisheries, and shipping. Ocean and Coastal Management 11–12:565–580. Alford, L. 2005. The occurrence and foraging activity of Bottlenose Dolphins and Harbour Porpoises in Cardigan Bay SAC, Wales. M.Sc. Thesis. The University of Wales, Bangor, Wales. 68 pp. Allen, M.C., A.J. Read, J. Gaudet, and L.S. Sayigh. 2001. Fine-scale habitat selection of foraging Bottlenose Dolphins, Tursiops truncatus, near Clearwater, Florida. Marine Ecology Progress Series 222:253–264. Azzara, A.J. 2012. Impacts of vessel-noise peturbations on the resident Sperm Whale population in the Gulf of Mexico. Ph.D. Dissertation. Texas A&M University, College Station, TX. 104 pp. Baird, R.W., and S.M. Burkhart. 2000. Bias and variability in distance estimation on the water: Implications for the management of whale watching. IWC Meeting, Adelaide, Australia. Ballance, L.T. 1992. Habitat-use patterns and ranges of the Bottlenose Dolphin the Gulf of California, Mexico. Marine Mammal Science 8:262–274. Bejder, L., A. Samuels, H. Whitehead, N. Gales, J. Mann, R. Connor, M. Heithaus, J. Watson- Capps, C. Flaherty, and M. Krützen. 2006. Decline in relative abundance of Bottlenose Dolphins exposed to long-term disturbance. Conservation Biology 20:1791–1798. Benjamini, Y., and Y. Hochberg. 1995. Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of Royal Statistical Society Series B 57:289–300. Bräger, S. 1993. Diurnal and seasonal behavior patterns of Bottlenose Dolphin (Tursiops truncatus). Marine Mammal Science 9:434–438. Bräger, S., B. Würsig, A. Acevedo, and T. Henningsen. 1994. Association patterns of Bottlenose Dolphins (Tursiops truncatus) in Galveston Bay, Texas. Journal of Mammalogy 75:431–437. Buckstaff, K.C. 2006. Effects of watercraft noise on acoustic behavior of Bottlenose Dolphins, Tursiops truncatus, in Sarasota Bay, Florida. Marine Mammal Science 20:709–725. Fertl, D. 1994. Occurence patterns and behavior of Bottle-nosed Dolphins (Tursiops truncatus) in the Galveston ship channel, Texas. The Texas Journal of Science 46:299–317. Fish, F.E., and C.A. Hui. 1991. Dolphin swimming: A review. Mammal Review 21:181–195. Harrison, R., and M. Trevino. 2013. Evaluating the impacts of the Panama Canal expansion on Texas gulf ports. Research Report SWUTC/13/476660-00062-1. Center for Transportation Research, Austin, TX. 56 pp. Leatherwood, S. 1975. Some observations of feeding behavior of Bottle-nosed Dolphins (Tursiops truncatus) in the northern Gulf of Mexico and (Tursiops cf. T. gilli) off southern California, Baja California, and Nayarit, Mexico. Marine Fisheries Review 37:10–16. Southeastern Naturalist A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 758 Leatherwood, S., and R.R. Reeves. 1989. The Bottlenose Dolphin. Academic Press, San Diego, CA. Lemon, M., T P. Lynch, D.H. Cato, and R.G. Harcourt. 2006. Response of travelling Bottlenose Dolphins (Tursiops aduncus) to experimental approaches by a powerboat in Jervis Bay, New South Wales, Australia. Biological Conservation 127:363–372. Lemon, M., D. Cato, T. Lynch, and R. Harcourt. 2008. Short-term behavioural response of Bottlenose Dolphins, Tursiops aduncus, to recreational powerboats. Bioacoustics 17:171–173. Lusseau, D. 2003. Male and female bottlenose dolphins, Tursiops spp., have different starategies to avoid interactions with tour boats in Doubtful Sound, New Zealand. Marine Ecology Progress Series 257:267–274. Lusseau, D. 2005. Residency pattern of bottlenose dolphins, Tursiops spp., in Milford Sound, New Zealand is related to boat traffic. Marine Ecology Progress Series 295:265–272. Mann, J. 1999. Behavioral sampling methods for cetaceans: A review and critique. Marine Mammal Science 15:102–122. Mattson, M.C., J.A. Thomas, and D.S. Aubin. 2005. Effects of boat activity on the behavior of Bottlenose Dolphins (Tursiops truncatus) in waters surrounding Hilton Head Island, South Carolina. Aquatic Mammals 31:133–140. Merrick, J.R.W., and J.R. Harrald. 2007. Making decisions about safety in US ports and waterways. Interfaces 37:240–252. Miller, C.E., and D.M. Baltz. 2010. Environmental characterization of seasonal trends and foraging habitat of Bottlenose Dolphins (Tursiops truncatus) in northern Gulf of Mexico bays. Fishery Bulletin 108:79–86. Miller, L.J., M. Solangi, and S.A. Kuczaj. 2008. Immediate response of Atlantic Bottlenose Dolphins to high-speed personal watercraft in the Mississippi Sound. Journal of the Marine Biological Association of the United Kingdom 88:1139–1143. Nowacek, S.M., R.S. Wells, and A.R. Solow. 2001. Short-term effects of boat traffic on Bottlenose Dolphins, Tursiops truncatus, in Sarasota Bay, Florida. Marine Mammal Science 17:673–688. Papale, E., M. Azzolin, and C. Giacoma. 2012. Vessel traffic affects Bottlenose Dolphin (Tursiops truncatus) behaviour in waters surrounding Lampedusa Island, south Italy. Journal of the Marine Biological Association of the United Kingdom Special Issue 8:1877–1885. R Core Team. 2015. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online at https://www.R-project. org. Accessed 11 November 2016. Richardson, W.J., and B. Würsig. 1997. Influences of man-made noise and other human actions on cetacean behaviour. Marine and Freshwater Behaviour and Physiology 29:183–209. Roman, J., I. Roman, M. Altman, C.D. Daly, M. Campbell, and A. Jasny. 2013. The Marine Mammal Protection Act at 40: Status, recovery, and future of US marine mammals. Annals of the New York Academy of Sciences 1286:29–49. Romano, T.A., M.J. Keogh, C. Kelly, et al. 2004. Anthropogenic sound and marine mammal health: Measures of the nervous and immune systems before and after intense sound exposure. Canadian Journal of Fisheries and Aquatic Sciences 61:1124–1134. Ross, D. 2005. Ship sources of ambient noise. IEEE Journal of Oceanic Engineering 30:257–261. Southeastern Naturalist 759 A.E. Rivard, F.P. Gelwick, and W. von Zharen 2016 Vol. 15, No. 4 Sargeant, B.L., and J. Mann. 2009. Developmental evidence for foraging traditions in wild Bottlenose Dolphins. Animal Behavior 78:715–721. Sellas, A.B., R.S. Wells, and P.E. Rosel. 2005. Mitochondrial and nuclear DNA analyses reveal fine-scale geographic structure in Bottlenose Dolphins (Tursiops truncatus) in the Gulf of Mexico. Conservation Genetics 6:715–728. Smolker, R.A., A.F. Richards, R.C. Connor, and J.W. Pepper. 1992. Sex differences in patterns of association among Indian Ocean Bottlenose Dolphins. Be haviour 123:38–69. Steichen, J.L., R. Windham, R. Brinkmeyer, and A. Quigg. 2012. Ecosystem under pressure: Ballast-water discharge into Galveston Bay, Texas (USA) from 2005 to 2010. Marine Pollution Bulletin 64:779–789. ter Braak, C.J.F. and P. Šmilauer. 2012 Canoco Reference Manual and User’s Guide: Software for Ordination (Version 5.0). Microcomputer Power, Ithaca, NY, USA. 496 pp. United States Coast Guard (USCG). 2012. National recreational boating survey 2012. Available online at http://www.uscgboating.org/statistics/national-recreational-boatingsafety- survey.php.Accessed 11 November 2016. Van den Brink, P.J., and C.J.F. ter Braak. 1998. Multivariate analysis of stress in experimental ecosystems by Principle Response Curves and similarity analysis. Aquatic Ecology 32:163–178. Van den Brink, P.J., and C.J.F. ter Braak. 1999. Principal response curves: Analysis of timedependent multivariate responses of biological community to stress. Environmental Toxicology and Chemistry 18:138–148. Waring, G.T., E. Josephson, K. Maze-Foley, and P.E. Rosel. 2015. US Atlantic and Gulf of Mexico Marine Mammal Stock Assessments - 2014. NOAA Technical Memorandum NMFS-NE-231:361. Weilgart, L.S. 2007. A Brief Review of Known Effects of Noise on Marine Mammals. International Journal of Comparative Psychology 20:159–168. Weisbrod, R.E., and C.T. Lawson. 2003. Ferry systems: Planning for the revitalization of US cities. Journal of Urban Technology 10:47–68. Weller, D.W. 1998. Global and regional variation in the biology and behavior of Bottlenose Dolphins. Ph.D. Dissertation. Texas A&M University, College Station, TX. 142 pp. Yeater, D., and S. Kuczaj. 2010. Observational learning in wild and captive dolphins. International Journal of Comparative Psychology 23:379–385.