Resident Harbor Porpoises (Phocoena phocoena vomerina) in the Salish Sea: Photo-Identification Shows Long-Term Site Fidelity, Natal Philopatry, and Provides Insights into Longevity and Behavior

Abstract

Harbor porpoises have been well studied in locations worldwide, but relatively little research has focused on site fidelity of individuals, which influences population structure and determines appropriate management and conservation measures. This study investigated the site fidelity and behavior of harbor porpoises through long-term, land-based photo-identification around Fidalgo Island, Washington, USA. Individuals were consistently re-sighted for up to 11 years, including natal philopatry. There was individual variation in seasonal site fidelity, with some individuals seen year-round, though general porpoise presence in the study area and individual re-sightings were greatly reduced during summer. Females had calves yearly for up to 3 consecutive years, with females as old as at least 14 successfully producing calves. There was some evidence for mother–calf associations post-weaning, but none long-term. Behavior was affected by tide, where porpoises traveled more and foraged and socialized less during ebb tide compared to other tidal states. Mating attempts with aerial behaviors were observed 84 times and seen year-round in every month and every season. This study provides documentation of a local, resident population of harbor porpoise, and emphasizes the need for more fine-scale studies like this to ensure the appropriate management and conservation of harbor porpoise populations and/or stocks worldwide.

1. Introduction

Patterns of site fidelity of individual animals, or their tendency to return to, or remain in, a particular habitat, influence the social, genetic, and community structure of a population [1,2]. The movement of animals is driven by processes across multiple spatial and temporal scales, and influences the fate of individuals as well as the structure and dynamics of populations, communities, and ecosystems [2]. The harbor porpoise (Phocoena phocoena) is one of the most abundant coastal species of cetacean worldwide, with a wide geographic distribution spanning much of the northern hemisphere [3,4]; however, information on the site fidelity and behavior of individuals is lacking for this species. As a small, highly mobile, coastal marine predator, harbor porpoises live a large portion of their lives in close proximity to human activities, and knowledge of site fidelity and habitat use patterns is critical in assessing impacts from anthropogenic threats [5].

To date, research on harbor porpoises along the Western coast of the United States (USA) has been focused on large-scale abundance, stranded animals, and stock structure [6,7,8,9,10,11,12,13,14,15], or on mating behavior [16,17,18]. Similarly, the multitude of research in European waters has primarily focused on large-scale abundance, stranded animals, distribution, and the effects of pollutants and anthropogenic impacts (e.g., [19,20,21,22,23]), as well as mating behavior [18]. However, there is increasing evidence for fine-scale population structure in a variety of harbor porpoise populations worldwide. Research has shown that harbor porpoise behavior and distribution can be structured over varying spatial scales and across different temporal scales, including diel and seasonal differences [24,25,26,27,28,29,30,31]. Some harbor porpoise populations may also live in relatively small and restricted geographic areas [6,32,33], and significant genetic differences have been observed within geographic areas that were previously thought to be one genetic stock [14,15,34]. When population structure is unknown, it is difficult to define what effect the loss of individuals will have; if there are small or restricted porpoise populations, a localized conservation concern may exist [35]. Understanding fine-scale variations within and between populations is critical to ensure the appropriate management and conservation of harbor porpoise stocks.

Globally, the conservation status of the species differs between populations. In Atlantic and European waters there are harbor porpoise populations that are classified as either vulnerable, endangered, or critically endangered, all of which indicate that they are threatened with extinction [36,37]. Along the West Coast of the USA and the Salish Sea (a multinational body of water spanning between Washington (WA), USA, British Columbia, Canada, and many Indigenous Tribes and First Nations), harbor porpoises are in a somewhat better status [33], ranging from being a species of special concern in Canada under the Species At Risk Act [38], to the USA where it is a Priority Species under Washington Fish and Wildlife’s Priority Habitat and Species Program [39] and a candidate for listing as a Washington State endangered, threatened, or sensitive species. In the Salish Sea, harbor porpoises are year-round residents with a high abundance and consistent presence [40,41,42,43,44], and in WA comprise the Washington inland waters stock (designated by the National Oceanographic and Atmospheric Association (NOAA)). This stock is of particular interest given their drastic decline in numbers by the 1970s [43,45,46] followed by a subsequent recovery [9,46,47]. Their status as a species of concern at various governmental levels in this region is based on the large amount of data gaps about many aspects of their biology, ecology, behavior, sociality, and regionally specific threats [33], combined with the fact that, due to their coastal habitat and prey preferences, harbor porpoises are vulnerable to anthropogenic impacts such as fishery interactions [35], noise disturbance [22,48], disease [49], and pollution [50,51,52]. They are also important as a prey species to mammal-eating Bigg’s killer whales which frequent this region [53,54].

Dedicated long-term research is critical to help fill in the data gaps for harbor porpoises in this region and better understand fine-scale population structure, movements, site fidelity, seasonal variation, and life history that will inform effective management and conservation to ensure their continued success in the Salish Sea (see review [33]). The long-term tracking of individuals can provide valuable data on fine-scale site fidelity, as well as life history characteristics that are critical to understanding these population dynamics. Photographic identification (photo-ID) allows for the tracking of individual animals over long periods of time and can provide vital data on population dynamics and individual variation, offering insights into site fidelity, heredity, longevity, and ranging patterns, while also providing a unique opportunity to explore and observe behavioral patterns and anomalies as seen in other well-studied small cetacean populations (e.g., bottlenose dolphins, Tursiops truncatus) [55,56,57,58,59] and spotted dolphins, Stenella frontalis [60,61,62].

Harbor porpoises are known for being evasive and shy, making them a more difficult species to study. This behavior, along with their small size, cryptic coloration, and few readily observable distinguishing marks [63,64,65,66,67], has meant that they have historically been overlooked as candidates for photo-ID studies at the local population level. In spite of these qualities, pigmentation patterns on the side/flank of harbor porpoises have been known to be unique to individual harbor porpoises since the early 1970s [64,68]. Research utilizing these unique markings for identification has only recently been implemented in a small number of locations, demonstrating that these marks can be successfully used for photo-ID over shorter (months) and longer (years) periods of time in both USA and European waters [32,67,69,70,71,72]. Using photo-ID, our previous research successfully demonstrated site fidelity of Pacific harbor porpoises (Phocoena phocoena vomerina) over 3 years and year-round use of the waters off Fidalgo Island, Washington, USA, part of the Salish Sea [32].

This paper provides further data from the on-going, long-term, land-based, photo-ID study of harbor porpoises in Burrows Pass off Fidalgo Island, between January 2014 and October 2024, expanding on Elliser et al.’s [32] initial study. The objectives of this study included using photo-ID to determine the degree of individual site fidelity, along with related life history and behavioral observations. To our knowledge, this study is one of the longest consecutive datasets available on individual free-swimming harbor porpoises in the world and can provide invaluable insights into their behavior, site fidelity, and ecology.

2. Materials and Methods

2.1. Study Area

The land-based study site was located on Fidalgo Island, WA. Burrows Pass is a stretch of water flowing between Burrows and Fidalgo Islands, connecting Burrows Bay to Rosario Strait (Figure 1). The center of Burrows Pass is located at 48.49 N, 122.69 W. The depth varies from 12 to 40 m and the pass is approximately 1300 m long, with a maximum width of approximately 915 m at the western end opening to Rosario Strait and a minimum width of approximately 460 m at the eastern end heading into Burrows Bay. The observation point was situated about 6 m above sea level and provided a good vantage point of the entire Pass.

2.2. Data Collection

Photographic and observational data were collected between January 2014 and October 2024 using Canon EOS T3i, T5i, and 5D cameras fitted with either a 100–400 mm or 100–600 mm zoom lens. Data were collected year-round; however, effort was not evenly distributed throughout seasons due to poor weather (e.g., increased precipitation, wind, etc.), particularly during the winter months. For the first 6 months of the study (January 2014–June 2014), data were collected approximately once per month as part of a feasibility study [32], increasing to an average of 2–3 times per week.

Observation periods were defined as the total amount of time spent in the field on a given day, and the vast majority lasted 2 h; however, due to additional animal activity, poor weather, time constraints, etc., occasionally, sessions were shorter or longer (range = 1–3 h). However, over the entire study, the average time spent in an observation was approximately 2 h (average = 1.99 h). Observations were conducted on separate weekdays during daylight hours, though occasionally two observations were conducted on a single day when necessary. Observations were typically not conducted in conditions over a Beaufort 3, or in winds > 20 mph. During observations, environmental data were collected including temperature, wind direction and speed, Beaufort state, tidal rip/current, sun/cloud cover, lunar phase, glare, and tidal state, as well as the number of boats passing through.

Little is known about harbor porpoise sociality and group structure [32]; therefore, for the purpose of this study a group was defined based on behavior and proximity: all animals in sight and/or involved in the same activity [73]. Because of the small size and topography of Burrows Pass, all individual porpoises were likely within range of acoustic, if not visual, communication. Thus, all porpoises within the Pass were considered a group unless individuals remained spatially separated while in the Pass (i.e., individuals stayed at opposite ends of the Pass) and showed no observable interactions with one another, in which case they were considered as two separate groups [32]. Generally, the number of individuals in the study site was three or less and their behavior and relative cohesiveness made it clear they were interacting with one another. A sighting was defined as the duration of time in which a porpoise or group of porpoises were observed within the study area during an observation period. A sighting was considered over after 10 min of not seeing any individual in the group. The end time of the sighting was then set at 5 min after the last sighting of an individual or group. The vast majority of the time there was one sighting of harbor porpoises at a time, but occasionally two sightings (defined using the criteria above) would overlap in time.

A minimum, maximum, and best estimate for group size and number of calves (defined as the observers’ best approximation based on the range of animals observed during a sighting) was recorded and updated after photographic analysis. Even if suitable identification photos were not obtained, photos could be used to confirm or update group size due to dorsal fin shape variability among individuals and small group sizes [32]. The final group size estimates presented are based on these best estimates. Calves were identified by visual estimation of body length (up to two-thirds the length of the adult [74]), along with typical calf behavior of swimming in echelon position. Calves included neonates and young (under 1 year old), while all other animals (including juveniles/subadults) that could not be visually distinguished from adult animals were classified as adults.

Sighting data (number of adults and calves, behavioral state (overall behavioral state that porpoises were engaged in during a sighting: probable foraging/foraging, traveling, socializing, unknown/other), behaviors observed (i.e., specific behaviors such as surface chases, fish chases, wake surfing, mating, aerials, porpoising, spy hops), and location within Burrows Pass) were recorded throughout the sighting, and environmental variables (see above) were recorded at the start of each sighting. Behavioral state was determined by observation of the following specific behaviors: probable foraging/foraging (surface chases, long dive intervals, fish chases/catches, and/or presence of bird mills/birds diving/catching fish in close proximity to porpoises), socializing (aerials, wake surfing, play, mating), travel (directional movement/porpoising of animals into or out of study area), and unknown (no clear behavioral state observed) [32]. All data were entered into FinBase, a customized Microsoft Access database used to store, manage, and visualize data from photo-ID surveys [75].

Seasons were defined as follows: winter (December–February), spring (March–May), summer (June–August), and fall (September–November). Effort was not evenly distributed across seasons; therefore, sightings were corrected for effort in order to compare sighting rates for harbor porpoise seasonal habitat use. Sightings per unit effort (SPUE) were calculated as the number of harbor porpoise sightings divided by total number of observations.

2.3. Data Analysis

Photos were processed and analyzed using the methodology described in Elliser et al. [32], with only photos of suitable quality (good clarity and focus, lighting, proximity of animal) and animals that were ‘highly distinctive’ (D1) and ‘distinctive’ (D2) being used for identification [67]. A matrix of 8 different identification categories was used to accurately identify individual harbor porpoises. Pigmentation and scars/lesions were the primary identification features used, alongside any secondary features (e.g., nicks, notches, or scars on the peduncle or dorsal fin trailing edge) that were present. Confirmation features (overall coloration, fin shape, fin size, fin base width) were used to confirm identifications if needed, but were not used as the primary method of identification [32].

Harbor porpoise markings are not bilaterally symmetrical; thus, there was a possibility that a single harbor porpoise could be identified as two different animals when only one side of the animal was photographed. To reduce this type of error, dorsal fin shape, secondary features, and/or other confirmation marks were used to help match the left and right sides of individual animals and reduce this potential bias. When both sides of an animal were photographed within one sighting, these markings could be used to confirm the left and right sides were of the same individual (this was aided by the small group sizes and variety in dorsal fin shapes, allowing researchers to match the left and right sides of a particular individual).

Once identified, individual porpoises were assigned a unique identifier (name) and added to a catalog of known individuals. Along with their unique identifier, each individual entry in the catalog included the following: (a) the best photo of that animal’s left and/or right side, (b) when the individual was first observed and/or its birth year, (c) its distinctiveness category, (d) sex, if known, (e) calves and their birth years, if applicable, and (f) relationship to other porpoises in the catalog, if known (e.g., mothers and calves). The number of sightings was calculated for each individual. Sometimes an individual would be in more than one sighting during an observation period (in one day). Due to the close proximity of time and being in the same location, these subsequent sightings were not included in an individual’s total sighting count, i.e., an individual was not counted as sighted more than once per day.

Porpoises seen with calves in echelon position were assumed to be females. Individuals involved in mating attempts were sometimes able to be sexed as well. Harbor porpoise mating behavior is highly lateralized and conserved around the world, with males exclusively coming up on the left-hand side of the female [17,18]. Thus, even if physical evidence (i.e., visible penis) was not visible to confirm the sex of an animal during a mating attempt, the animal on the left was presumed to be male and the animal on the right assumed to be female.

Data analyses were computed in Python using the latest stable release (0.14.4) of the statsmodels package [76]. Additionally, we used the latest stable release (0.11.2) of the scikit-posthocs package [77] for performing post hoc pairwise comparisons.

We tested for the multicollinearity of environmental variables by calculating the variance inflation factors (VIFs). VIFs > 10.0 are typically used to indicate multicollinearity [78]. The VIFs for our environmental variables were below this value (all VIFs < 2.5).

Individual binomial logistic regression analyses using a generalized linear model (GLM) were run to examine the predictive impact of the available environmental data for each of the recorded outcomes: presence/absence of porpoises as well as the observation of social, foraging, and/or travel behavior of observed porpoise groups. A full model including all three of the following environmental factors as predictors was set initially: season, tide state (slack low, flood, slack high, ebb), and rip tide strength (low, medium, high). To determine the relative significance of the predictor variables within the full model, the p-values from a Wald test were examined for each variable. Using Akaike’s Information Criterion (AIC), and considering Cox and Snell’s pseudo R² as a secondary measure of model fit, the original full model was compared to the reduced models containing only the predictors deemed significant (p < 0.05) for that outcome based on the Wald tests. In all cases, the reduced model showed a lower AIC value indicating a better fit, and was therefore selected for further analyses of that particular outcome.

For the analysis of seasonal differences in size of groups, a nonparametric Kruskal–Wallis test was used due to the skewed distribution of porpoise group sizes. Pairwise post hoc comparisons were performed using Dunn’s test, with a Holm stepwise adjustment applied.

3. Results

3.1. Effort

Over 2290 h were spent on observations from Jan 2014 to October 2024. Effort was lowest in the winter due to poor weather conditions (e.g., wind, precipitation, etc.) that restricted time for fieldwork. Despite the fact that effort was highest in summer and that there were a relatively large number of sightings, that season had the lowest sightings per unit effort (SPUE = number of sightings/total number of observations), percentage of observations with sightings, and the lowest amount of time observing porpoises (

Table 1. Total number of observations, sightings, percentages of amount of observations with sightings, and amount of time observing porpoises (during sightings). Sightings per unit effort (SPUE) is number of sightings divided by number of observations (average time spent per observation = 1.99 h). Italicized indicates lowest percentages and SPUE in summer.

Season	Number of Observations	Number of Sightings	Percentage (%) of Observations with Sightings	Percentage (%) of Time Observing Porpoises	SPUE
Winter	192	334	86.5%	42.1%	1.74
Spring	322	610	89.8%	53.9%	1.89
Summer	338	488	75.7%	31.3%	1.44
Fall	282	487	90.4%	53.2%	1.73
Total	1134	1919	85.19%	45.0%	1.69

).

3.2. Group Size

Group size ranged from 1 to 20 (x = 2.58 ± 1.72), with a median of 2, and was significantly different between seasons (Kruskal–Wallis, H = 81.436, p < 0.001). Group size was significantly smaller in summer (2.08 ± 1.40) compared to spring (2.82 ± 1.99), fall (2.8 ± 1.74), and winter (2.55 ± 1.42) (Dunn’s test p < 0.001). The largest groups of 16 and 20 were seen in fall and spring, respectively, which also had the highest group size averages. The vast majority (78.9%) of sightings in all seasons had three or fewer individuals (fall 74.9%, spring 74.4%, summer 88.7%, winter 78.7%).

3.3. Site Fidelity and Re-Sighting History

Between January 2014 and October of 2024, 195 individual porpoises, including calves, were identified (Figure 2). Because the re-sightings for calves during the first year of life will be highly correlated with the mother, the first year of sightings and re-sightings for 31 calves were removed, leaving 164 individuals (28 calves were removed completely as they were not re-sighted/able to be matched after the first year; three calves remained, and their re-sightings post-weaning were counted). The majority (68.9%, n = 113) of these 164 individuals were re-sighted two or more times.

Of these 113 re-sighted individuals, 57.5% were sighted in at least two consecutive or non-consecutive years (n = 65,

Table 2. Sighting and re-sighting histories of non-calf harbor porpoises in Burrows Pass, WA, showing range of total sightings and mean re-sighting rates for all individuals, those seen at least two times, and those seen in at least 2 years.

	Number of Total Sightings	Mean Re-Sighting Rate
All non-calf individuals (n = 164)	1–125	5.4
Re-sighted at least twice (n = 113)	2–125	9.1
Re-sighted in at least 2 different years (n = 65)	2–125	12.8

). Sightings per calendar year for an individual ranged from 0 to 31.

The number of years an individual was re-sighted (not including missing years) ranged from 2 to 9 years, with an average of 2.9 years. If missing years are included (it being more likely we missed taking a photo of them vs. them not being present in the area), which may better indicate truer levels of site fidelity, then the range is 2 to 11 years, with an average of 4.0 years. This indicates at least 9 or 11 years of site fidelity for some individuals. Examples of long-term matches of individuals are provided in Figure 3.

3.4. Seasonality

Logistic regression analysis revealed that season (Wald χ² = 32.102, p < 0.001) as well as rip tide strength (Wald χ² = 9.218, p = 0.010) both significantly affected harbor porpoise presence. Porpoises were more likely to be seen in fall, winter, or spring compared to summer (p < 0.001, p = 0.022, and p < 0.001, respectively). Additionally, harbor porpoises were more likely to be seen in low (p = 0.009) and medium (p = 0.005) rip tides when compared to high rip tides. Tide level (Wald χ² = 0.805, p = 0.848) was not found to have a statistically significant impact on porpoise presence. The VIFs of all three environmental variables fell below the cutoff value used to indicate multicollinearity (all VIF values < 2.5).

Over half (55.4%, n = 36) of the 65 individuals seen in multiple years (not including a calf’s first year) were also seen in more than one season within a calendar year (66.2%, n = 43, if calves are included). Fourteen of these individuals (38.9%) were seen in every season, some with all four season sightings coming within a calendar year and others across multiple years. Some individuals were seen in multiple seasons across years (e.g., spring in one year, fall in another), while others were seen in only one season across years (e.g., only fall of multiple years).

To look at seasonal trends within animals showing a stronger degree of site fidelity, re-sightings were calculated for individuals with 10 or more sightings across multiple years (Figure 4). There was a strong seasonal trend, with the highest re-sightings in spring (n = 253) and fall (n = 170). Summer (n = 86) and winter (n = 62) had about half or less compared to spring and fall. Six of these individuals were seen in every season within a calendar year (Figure 4).

3.5. Mothers and Calves

There were 19 presumed females, 18 documented by presence of calves consistently in echelon position and 1 inferred female from an observed mating attempt (

Table 3. Known females and their calving history, including the number or re-sightings of the mother, the number of calves she was seen with, the years the female was seen (gray boxes), and the years she had calves (letters in boxes). UN means that the calf was not named (due to lack of good photos or identifying marks), other letters are the three letter abbreviation of their name. Individuals with a * were of the three-generation family. Note that the lack of a calf in a given year is not necessarily indicative of no calf that year; the mother may have only been sighted before having a calf, or the calf did not survive, or the female was not seen enough times to determine. ** indicates that the previous year’s calf was still with the female through November of the current year (indicating the calf was over a year old and still associated with the mother).

Mom	Re-Sights	# Calves	2014	2015	2016	2017	2018	2019	2020	2021	2022	2023	2024
ARC	2	1											UN
AUR	4	2							UN			AUS
DOL	1	1										X
DRA	2	2							UN				DAM
FAL	19	1						UN
FLI	7	2							UN			UN
MIL	39	2						MIN	MEL
PEA	2	2		UN					PAP
PIA *	45	4	UN			PIC *	UN		PIZ	**		PER	PIO
POI	124	4						PHA		PHS	UN		UN
PIC *	10	1				born	NA					PIP *
POL	2	1											UN
RAI	71	6				UN		RAN	UN	**	RAB	UN	UN
SLE	11	1							UN
SIC	4	1						UN
SLI	41	2							UN				UN
SMO	14	1										UN
SLO	2	1										UN

). Each of the eighteen females had one to six calves over the study period. Four females had calves in a set of consecutive years, with one female having five calves in six years. Some females were not seen every year; thus, it is possible that they had calves that could not be documented.

Three of the most re-sighted females, PIA, POI, and RAI (Figure 3), were likely adults when they were first observed (2014, 2018, and 2018, respectively). PIA was sighted with a calf in 2014, indicating she was at least 3 years old. She had five confirmed calves, including the most recent in 2024 (Table 3). Conservatively, this means she is at least 14 and still having calves. Estimating that POI (4 calves) and RAI (6 calves) were at least 3 when first sighted in 2018 (each had a calf in 2019), this would make them at least 10 years old and still having calves (each had one in 2024, Table 3).

At least three calves have shown site fidelity long past weaning. PHA is a known male and son of POI, with 47 sightings over 5 years. He has been seen in multiple sightings with POI since weaning (once in 2021, once in 2023, and four in 2024), including a mating attempt with his mother on 5/17/24 (Figure 5). PIC is a daughter of PIA with 10 sightings over 8 years, including 5 years not seen. In addition, PIC was sighted in 2023 with calf PIP, making this a three-generation family (PIA, PIC, PIP), with the grandmother PIA also having a calf in 2023. PIZ is of unknown sex, the offspring of PIA, with 31 sightings over 5 years, including 1 year not seen. In addition, two calves showed close association (similar to echelon position) with their mothers during sightings over a year after birth, PIZ (two sightings in November) and an unnamed (UN) 2020 calf of RAI (three sightings in October and November), indicating that they were still associated with their mother at over a year old. Both mothers, PIA and RAI, were not observed with new calves that year (2021).

3.6. Behavior

Harbor porpoises are often cryptic in their behavior at the surface, such that it is often unclear in what overall behavioral state they are in (listed as “Unknown” in

Table 4. The percentage of porpoise time in behavioral states by season and tide (number of sightings with the given behavior or tide divided by the total number of sightings). There can be more than one behavior observed during a sighting (i.e., foraging and social); thus, the overall totals may be over 100% as those sightings were counted for each behavioral state in the calculations. Bold/italicized indicates significantly higher levels, as described in text.

	Travel	Foraging	Social	Unknown
Winter	35.9% (n = 120)	32.9% (n = 93)	5.10% (n = 17)	34.1% (n = 114)
Spring	23.6% (n = 144)	36.9% (n = 225)	6.10% (n = 37)	42.8% (n = 261)
Summer	26.2% (n = 128)	23.4% (n = 114)	7.20% (n = 35)	48.2% (n = 235)
Fall	24.4% (n = 119)	38.4% (n = 187)	9.90% (n = 48)	40.7% (n = 198)
Slack low	31.28% (n = 56)	35.19% (n = 63)	6.15% (n = 11)	34.08% (n = 61)
Flood	18.31% (n = 156)	40.02% (n = 341)	11.62% (n = 99)	41.31% (n = 352)
Slack High	19.67% (n = 24)	46.72% (n = 57)	9.02% (n = 11)	40.98% (n = 50)
Ebb	36.57% (n = 325)	20.88% (n = 157)	2.13% (n = 16)	44.15% (n = 332)

). Behavior, categorized as travel, foraging, social, or other, was examined in relation to season, tide, and rip tide (Table 4).

Travel behavior was influenced by season (Wald χ² = 12.849, p = 0.005), tide (Wald χ² = 37.238, p < 0.001), and rip tide (Wald χ² = 7.333, p = 0.026). Travel was more likely to be observed in winter than any other season (vs. spring p < 0.001, vs. summer p = 0.026, and vs. fall p = 0.009). Travel was also more likely to be seen during ebb tide vs. flood or slack high (p < 0.001), as well as being more likely to be seen during slack low vs. flood (p = 0.008) and slack low vs. slack high (p = 0.019). Travel was observed less frequently during medium rip tides as compared to low rip tides (p = 0.01) (Table 4).

Foraging behavior was influenced by season (Wald χ² = 29.137, p = 0.005) and tide (Wald χ² = 72.386, p < 0.001). Rip tide was found not to be statistically significant (Wald χ² = 0.779, p = 0.677). Foraging behavior was more likely to be observed in fall and spring than it was in either winter (p = 0.019 and p = 0.001, respectively) or summer (p < 0.001 and p < 0.001, respectively). Additionally, foraging behavior was less likely in ebb tide as compared to all other tide states (flood, slack high, and slack low (p < 0.001). Observed foraging differences across non-ebb tide states were not statistically significant (Table 4).

Social behavior was influenced by tide (Wald χ² = 32.870, p < 0.001) and rip tide (Wald χ² = 10.536, p < 0.005). Season was found not to be statistically significant (Wald χ² = 3.510, p = 0.320). Observed social behavior was significantly less likely during ebb tide as compared to all other tide states (ebb vs. flood (p < 0.001), vs. slack high (p < 0.001), and vs. slack low (p = 0.005)). Social behavior was more likely to be observed during medium rip tides vs. high (p = 0.048) and vs. low (p = 0.002) rip tides (Table 4).

Mating attempts involving aerial behavior were observed during 84 sightings between 2015 and 2024. Typically, there was one attempt per sighting, but occasionally there would be multiple (2–4) attempts observed. Mating attempts were documented year-round. Attempts were seen in every month across all years, although almost, but not all, in one concurrent year (

Table 5. The number of mating attempts observed between 2015 and 2024, and what months and seasons they were documented in. Confirmed visually and with photo documentation when possible.

Year	Number of Mating Attempts	Number of Months	Seasons
2015	1	1	Sp
2017	4	3	Sp, F, W
2018	4	4	Sp, F, W
2019	9	5	Sp, Su, F, W
2020	17	8	Sp, Su, F, W
2021	17	9	Sp, Su, F, W
2022	3	3	Su, F
2023	11	7	Sp, Su, F, W
2024	18	8	Sp, Su, F, W
Total	84	Seen in every month	In all seasons

). Mating attempts were observed in all seasons across years, and in all seasons during a calendar year in five years (Table 5). All mating attempts were confirmed visually, and when possible with photographic documentation.

4. Discussion

This study has successfully demonstrated that harbor porpoises in the Salish Sea can be identified over long periods of time (over a decade) using photo-ID, and that these data can provide unique insights into site fidelity, life history characteristics, behavior, habitat use, and population structure. We have demonstrated consistent long-term site fidelity of up to 11 years and consistent year-round, multi-year use of a habitat, including natal philopatry, by individually identified free-ranging harbor porpoises, indicating a resident, local population. Understanding this type of fine-scale population structure is vitally important in order to consider the appropriate biological context for the animals [33,79] and make biologically meaningful conservation decisions.

4.1. Group Size

Trends in group size were consistent with previous work, including summer having the smallest average group size, though the overall average of x = 2.58 was slightly higher than previously observed (x = 2.32) [32]. Harbor porpoises are most commonly observed in group sizes of 1–3 animals throughout their range [7,9,16,32,80,81]. However, large aggregations of harbor porpoises have been seen along the west coast of Canada in the Port of Prince Rupert (ranging from 30 to 1000 individuals) [30] and in the US waters of the Salish Sea (ranging from 20 to >300 individuals), and are more common than previously thought [82]. We did not observe any large aggregations in our local study area, likely due to the topography of the habitat that is not conducive to large groupings. In previous work north of our study site in the San Juan Islands area of the Salish Sea, group sizes were lower, with an average 1.87 individuals [43]. With a mean group size of 2.58, our findings are on the higher end of the local and global ranges, demonstrating the potential variability in harbor porpoise group sizes in different parts of the world, and even different locations within a population range.

4.2. Site Fidelity and Re-Sighting History

The regular and long-term site fidelity documented here indicates that there is a local, resident population of harbor porpoises that regularly utilize Burrows Pass. Almost 70% of non-calf harbor porpoises were re-sighted in Burrows Pass at least twice, with 57.5% of these animals showing site fidelity over multiple years. Long-term site fidelity within these re-sighted individuals ranged from 2 to 11 years, with as many as 31 sightings in a calendar year. It is important to note that these re-sightings are likely an under-representation of the true re-sighting rates. Usable photos of the animals are not always obtainable, due to behavior and/or distance to the photographer; thus, we are likely to have missed re-sightings of individuals (i.e., the porpoises were present, but we could not identify the individuals during a sighting). The local population of harbor porpoises that are resident and utilize Burrows Pass regularly is therefore likely larger than that documented here.

Until recently, there has been very little published information on site fidelity in individual harbor porpoises. Recent research has demonstrated varying degrees of site fidelity in harbor porpoises, ranging from weeks to years [32,33,69,70,71,72,83]; however, to our knowledge, only one other study has long-term photo-ID data similar to that shown here. Stichting Rugvin has been conducting photo-ID in the Eastern Scheldt, Netherlands, since 2009 [71]. They have documented harbor porpoise site fidelity of up to 14 years [F. Zanderink, pers. comm.].

Looking beyond photo-ID, other recent research supports the suggestion that long-term site fidelity over multiple years may be more widespread within this species than has previously been believed. At the population level, environmental DNA (eDNA) research has indicated that harbor porpoises in coastal and inland waters of Alaska show natal philopatry at multiple geographic scales, and have limited gene flow/dispersal between neighboring areas [14,15]. Two regions in the Southeast Alaska stock are suggested to have demographically independent populations (DIPs) based on genetic data, trends in abundance, and distribution, and it is thought that multiple other DIPs exist within the currently recognized harbor porpoise stock [84,85]. Along the California, (CA), WA, and Oregon (OR) coast, recent genetic studies have proposed creating a Central OR stock from the Northern CA/Southern OR stock, indicating more genetic differentiation within the stock than previously thought [13].

While some harbor porpoise populations may be more fluid and migrate seasonally [24,25,86,87], it is evident that some harbor porpoises do remain resident to specific areas, showing strong site fidelity and genetic differentiation, and that this is not isolated to one region or even one part of the world. Clearly, harbor porpoise populations and movement patterns are not universal across their range, and should be investigated for each population. Further research with photo-ID and genetics in our study area and the surrounding waters is necessary to determine whether there are other smaller, genetically or socially distinct populations or communities within the larger Salish Sea population. Understanding these dynamics is critical for the conservation of this species and redefining stock assessments, supporting the importance of fine-scale data in effectively managing and conserving this small coastal species [15,26,31,33].

4.3. Seasonality

Season and rip tide strength affected the presence/absence of harbor porpoises in Burrows Pass, with summer lower than every other season and porpoises more often seen in low and medium rip compared to high, consistent with our previous work [32]. The lower presence during high rip may indicate that the rip strength in Burrows Pass is above the upper limit for adult harbor porpoises in tidal stream habitats [32]. While harbor porpoises were observed year round, over all 11 years of the study we consistently observed that, despite increased effort, harbor porpoise presence was lower during summer in the months of June–August, with a significant drop in presence typically occurring in mid-late July. Seasonal variations in harbor porpoise presence and relative densities are well documented around the world, with different locations seeing peaks in harbor porpoises across different seasons [24,25,88,89,90], indicating that seasonal variations in abundance and density are likely population-specific and may be due to location-specific drivers, such as environmental conditions or prey availability.

Burrows Pass has the characteristics of a tidal stream habitat and is an important foraging location for harbor porpoises [32], with up to 34% of sightings in the current study including foraging behavior. This is likely an under-representation, as we could only document foraging behavior when there were observable surface behaviors (such as extended dive times, surface chases, fish catches, bird presence/diving). Harbor porpoises in Burrows Pass have been observed feeding on both small and large fish species (Elliser unpublished data, [91]). Worldwide, tidal stream habitats have been shown to have conditions that likely improve foraging opportunities for porpoises [26]. Feeding in distinctive areas, and thus reducing their need for movement between foraging patches, has been hypothesized to be a beneficial strategy for harbor porpoises [31,92], and is likely driven by prey availability in different locations. Prey movements and seasonal peaks in specific prey types have been documented as influences on harbor porpoise presence and seasonal movements [24,25,31,93]; thus, it is possible that a similar effect is occurring here and driving the reduced sightings in summer. Additionally, females are commonly observed returning to the Pass with new calves in August to early September; thus, it is possible that at least some of the animals showing strong site fidelity spend more time elsewhere to have their calves before returning to Burrows Pass.

Our study site was located adjacent to a busy marina, with boats traveling through the Pass as the main thoroughfare to Rosario Strait and the San Juan Islands. The number of boats in the area was higher during the summer than in other seasons; however, throughout the duration of the study, porpoises seem to have little or no direct behavioral response to vessel traffic in the Pass [32], a trend which is also seen in the southern Salish Sea (Puget Sound, D. Anderson pers. comm.). To the north, in the Port of Prince Rupert, Canada, harbor porpoises are consistently present despite the presence of large vessels in this very busy port, suggesting some level of acclimatization to localized disturbance [30], which has also been documented in European waters [94]. While harbor porpoises in the Salish Sea may appear to be more acclimated to boats than in other areas of their range, it is unknown whether boat presence may affect harbor porpoise in less overt and/or indirect ways in Burrows Pass which are not immediately visible through direct behavioral responses and might influence their seasonal presence [32]. For example, harbor porpoises with DTAGS have shown cessation of echolocation that was correlated with high boat-noise levels, leading to significantly fewer prey captures [95], and in another study, although little habitat displacement occurred, the distribution of harbor porpoises was significantly affected by shipping-noise levels [94]. It is important to note that some regularly re-sighted individuals have also been re-sighted during summer months when porpoise sightings are low, so while they are not completely absent from the area they are, in general, greatly reduced in time, presence, and number of re-sightings.

The lower re-sighting rate of individuals observed in winter is likely more of an artifact from difficulties in data collection, rather than the absence of animals. While effort was reduced due to poor weather conditions, the SPUE and percentage of observations with sightings were similar to spring and fall. However, the time spent observing porpoises was about 10% lower in winter, corresponding with travel being seen significantly more in winter than in any other season. While traveling, porpoises are typically directional and moving quickly through the area. While we can document this behavior, it is difficult to obtain photos that are usable for ID because the porpoises spend less time in the pass, providing few opportunities for good ID photos. The combination of these factors, along with less-ideal environmental conditions (such as higher Beaufort, glare, and low-light levels), make obtaining good ID photos more difficult in winter. It is likely that individuals with high re-sighting rates in spring and fall are also around more in the winter than we were able to document.

4.4. Mothers and Calves

Our study shows that females that utilize Burrows Pass are capable of successfully reproducing yearly for at least 3 consecutive years. It is likely that this is similar for many parts of the Salish Sea, as previous work on stranded individuals in the region suggested harbor porpoises may be able to reproduce yearly [10]. It is understood that harbor porpoises vary between populations on yearly or bi-yearly birthing patterns. Atlantic harbor porpoises seem to reproduce yearly, whereas it was thought that in the Pacific they reproduced bi-yearly [96]. It is unclear if the trend for yearly birthing seen in the Salish Sea extends to the coast and Pacific Ocean.

In Burrows Pass, new calves are most often first sighted in August and September, though one was observed as early as April [32]. This corresponds with stranding data for the area, which shows the beginning of calving season being May–late July and gestation being around 10.8 months [10]. Calves in Burrows Pass were consistently seen in close proximity (often echelon position) with their mothers through spring of the following year. This supports the suggestion that calves should stay in close proximity during the 8–9 months of lactation [97,98]; however, not much is known about their associations beyond that time frame.

Our study is the first to show site fidelity in harbor porpoise calves of at least 5 years, indicating natal philopatry in this area. This supports recent genetic evidence from southeast Alaska, which suggests levels of natal philopatry and limited dispersal in specific areas [14,15]. Additionally, in our study there were a few instances where calves were sighted with their mothers more than a year after their birth, and thus after weaning occurred. Those five sightings occurred in fall of the following year, and the mother and calf were observed in close proximity, similar to echelon position, not just sighted within the same group. Interestingly, in these years the mother had not been seen with a new calf, which may influence how long a calf will remain associated with the mother. In one unique case, the male porpoise PHA was re-sighted in groups with his mother six times after he was older than one year. He was observed attempting to mate with his mother in one sighting; thus, these associations may be more related to reproduction than a true social association. No other associations between mothers and grown calves were observed, though the sample size was small. The potential lack of these associations may be linked to the unknown social structure of harbor porpoises, as longer-term mother–calf associations have been documented in highly social small cetacean species such as bottlenose [57] and spotted dolphins [62]. In this population, it seems that calves may remain in their natal range; however, this does not equate to continued associations with their mothers. This may derive from a relatively weak mother–calf bond compared to other odontocete species.

Longitudinal drone work in Denmark that followed an individual female and her calves over nine years [72] has shown there is considerable separation between the mother and calf starting as early as even a month after birth, and the mother–calf bond seems to be maintained by the calves more so than the mother [72]. If the bond between mother and calf is not as strong in the early months of life, it seems unlikely that it would extend beyond the weaning process. While current evidence suggests that there may be little association between mothers and calves post-weaning in harbor porpoises, our data shows that it cannot be completely dismissed, and may depend on the individuals, inter-birth intervals, and other variables.

Even less is known about harbor porpoise calf associations with individuals other than their mother. In the Salish Sea, harbor porpoise calves are seen during large aggregation events that last from hours to months [82], and group size with calves is larger in Burrows Pass [32], indicating that calves do have some interaction with other porpoises, either directly or indirectly. Hamel et al. [72] observed no social associations between calves and other porpoises other than their mother, suggesting that the mother may be the only individual that interacts with the calf prior to weaning. An important factor to consider, however, is that the Hamel et al. [72] study focused on only one individual female and her offspring over nine years, and, thus, may not be representative of typical harbor porpoise mother–calf relationships. In addition, this study was conducted between April and November, leaving a four-month period over the winter where it is possible more interactions could have occurred. Differing maternal styles have been shown in bottlenose dolphins, where different behavioral patterns including proximity maintenance, discipline, pectoral fin contact, and initiation of separations and reunions with calves are exhibited [99,100,101]. It is possible such individual variations exist in harbor porpoise females as well, which could account for some of the differences in mother–calf relationships or associations observed.

Together, these data indicate that calves in Burrows Pass consistently associate with their mother and likely do associate to some unknown degree with other porpoises during the lactation period, whereas long-term associations with their mother post-weaning are uncommon and may vary by individual. While these data help to provide some evidence for association patterns (and possible differences between populations), the social structure of harbor porpoises is still poorly understood. More research is needed to fully understand this aspect of harbor porpoise society, which may vary depending on the population, habitat, and other spatiotemporal factors.

4.5. Behavior

Tide was not a significant factor for overall presence/absence, similar to our previous work [32]; however, when looking at what the animals were doing while present in the Pass, tide greatly influenced behavior. There was a strong correlation with behavior and ebb tide. The porpoises were significantly more likely to be traveling during ebb and slack low, and significantly less likely to be foraging or socializing, correlating with the much lower number of sightings that included behaviors like mating, porpoising, surface chases, and wake surfing. In addition, the presence of calves was found to be significant in relation to tide in Burrows Pass, with calves more likely to be present during slack high tides [32]. This could affect behavioral trends, as mothers may restrict their movements when they have young calves that may find it more difficult to swim in various tidal flows.

Trends for harbor porpoise behavior in relation to tide seem to be site-specific throughout their worldwide range. In San Francisco Bay, harbor porpoises were seen typically traveling out of the bay in response to ebb currents [81]. Buzzing detections (and potential foraging activity) in the Wadden Sea showed site-specific diel and tidal patterns, even at small spatial scales [102]. In Oregon, USA, porpoises were seen foraging at a rocky reef site during ebb phase [103]. In southwest Britain, foraging occurred at high tide, with travel occurring throughout the tidal states in one location, whereas behavioral activity did not vary by tidal state in another close study site [104]. General conclusions about marine top predator distribution need to take into account local differences in topography and tides that can have various effects at different sites [102,105] and can affect behavior and habitat use. It has also been shown that significant changes in porpoise behavior can be found in short and long-term observations [102]. Our data supports the increasing evidence that trends in harbor porpoise behavior are highly variable at small spatiotemporal scales, and that long-term datasets are necessary to tease apart the strong seasonal and spatial variation in diel and tidal effects that are likely site-specific [102].

Social behavior was observed the least in Burrows Pass compared to foraging and travel, but that is more likely because harbor porpoises are relatively less active at the surface, making it difficult to identify social behaviors. Recent work with drones in our study area has revealed that social interactions (like mating attempts) were occurring underwater, when from the surface they looked like regular surfacing behaviors. Thus, the amount of social activity presented here is likely a large under-representation. For this study, most of the social behavior documented was mating attempts, as those are often the most conspicuous. Male harbor porpoises often, but not always, leap out of the water when mating, and it is a highly lateralized behavior worldwide, where males approach the left side of the female 100% of the time [16,18]. Other less-frequently observed social behaviors included wake surfing, spy hops, surface splashing (not related to foraging), and play. Social behavior was observed year-round, including mating attempts occurring in every month and every season, which has been observed in San Francisco, CA, USA (except in June due to fog restricting visibility) [16]. Due to the fast, high-intensity maneuvers required by males during mating, year-round practice of female approaches with correct timing may be important for immature as well as mature males [16]. Practice may start before weaning. In Denmark, mother–calf sexual interactions have been documented between a 9–10 month old unweaned calf and its mother [18]. We provide evidence of an adult male son (PHA) attempting to mate with his mother (POI). It is unknown if the male knew she was his mother, or if this is a common occurrence in harbor porpoises. Mating attempts in Burrows Pass appeared to be observed the most in fall, although this was not statistically significant. However, the estimated conception dates for Salish Sea harbor porpoise range between mid-August and the end of December [10], supporting the idea that mating attempts may be more prevalent during fall.

4.6. Longevity

Long-term photo-ID data can also provide important insights into the longevity of harbor porpoises, which is currently estimated to be anywhere from 8 to 25 years of age, with most averages given as 8–12 years [80,106]. By tracking the calving history of known females and correlating this to the average age of sexual maturity in female harbor porpoises of 3–4 years [80,106,107], it is estimated that several of the animals in our study site are conservatively 10–14 years old and still successfully reproducing. Similarly, there has been a female harbor porpoise in Denmark that was at least 12 years old [40], as well as an individual that is at least 13 years old in the Netherlands [F. Zanderdink, pers. comm.] that were still successfully reproducing. We have documented an individual that is at least 14 years old in the wild, which has also been documented in the Netherlands [F. Zanderink, pers. comm.]. These data have profound implications for the life history of free-swimming harbor porpoises, and suggest that a minimum life expectancy of 8–12 years may be an underestimate for otherwise healthy animals in at least some populations.

5. Conclusions

Through long-term photo-ID, we have identified a local population of harbor porpoises in the Salish Sea that show up to over a decade of consistent site fidelity and at least five years of natal philopatry, and provide unique insights into their behavior, associations, reproduction, and longevity. This long-term dataset provides the opportunity to explore community, social, and population structure through analyses of associations and mark–recapture data that will be a focus of future research. Photo-ID on harbor porpoises has long been overlooked as a viable strategy for tracking harbor porpoises; however, this study shows that in at least some locations it can provide a wealth of data on this cryptic species. Other populations around the world have shown the potential for this type of long-term work, including in San Francisco Bay, USA [69,81], Denmark [18,72], the Netherlands [71], and British Columbia, Canada [33], and there are likely many more.

These results are especially important because of the pronounced variability and flexibility of harbor porpoises between and within population boundaries in their distribution, habitat use, and behavior, even at small spatiotemporal scales. Not all harbor porpoise populations are the same, and attempting to utilize uniform conservation strategies, without appropriate biological context for the animals [33,79], may not effectively protect a population, particularly where a localized conservation concern (small community/population) may exist for this species [35]. From a variety of research techniques, it is increasingly evident that many harbor porpoise populations along the west coast of the USA are structured at a finer scale than previously thought [6,14,15,32,33,34].

The long-term site fidelity, natal philopatry, and fine-scale variability in habitat use observed in this study support the idea that there are smaller communities or populations of harbor porpoises that may be socially or genetically distinct within the larger Salish Sea population. Cetaceans in USA waters are federally managed and protected as stocks delineated by NOAA. This research highlights the vital need for more research and the possible reassessment of stock structure in the Salish Sea [33]. The site fidelity, behaviors observed, and habitat use documented here also continue to show that Burrows Pass is likely a biologically important area [32,108], and we suggest that areas like these be identified elsewhere in the Salish Sea and provided additional protections for the continued conservation of this species. Long-term research such as this study provides the data needed to make these decisions and is vital for the survival of populations where large data gaps exist, in order to prevent potential future population declines [33]. Finally, as this is a multinational body of water, collaboration between researchers across borders is necessary to provide accurate context for the biologically meaningful conservation of Salish Sea harbor porpoises [33], similar to what has been accomplished between countries in Europe [69]. Future research on harbor porpoises should be undertaken with these conservation goals in mind, incorporating photo-ID where possible as a powerful tool to better understand site-specific site fidelity, population, and community structures in order to provide biologically meaningful measures for their protection.