Informing the in-water humpback whale (Megaptera novaeangliae) interaction management program in the Ningaloo Marine Park, Western Australia

Whale and dolphin watching tourism continues to expand globally, generating substantial economic benefits for coastal communities. However, ensuring that tourism activities remain safe, sustainable, and do not adversely affect cetacean populations remains a significant management challenge. This study evaluated humpback whale (Megaptera novaeangliae) distribution in relation to the intensity of commercial in-water interactions at the commencement of a tourism program off Ningaloo Marine Park (NMP), Western Australia. Aerial surveys were conducted in 2016, 2017, and annually from 2022 to 2024 to assess humpback whale distribution and abundance, while vessel-based Electronic Monitoring System data were used to quantify commercial in-water interactions between 2016 and 2024. Availability-corrected estimates of whale abundance were calculated, and spatial point-pattern analyses were used to assess interannual differences in whale distribution. Generalised additive models were used to examine the influence of calf presence on the probability of whales occurring within the mother-calf protection zone. Availability-corrected estimates indicated that humpback whale abundance varied among years, with abundance increasing between 2022 and 2024. The number of whale groups increased from 440 in 2016 to 866 in 2024, while the proportion of groups containing calves differed significantly among years (χ 2 = 68.88, df = 4, p < 0.001). Spatial analyses revealed no significant interannual differences in fine-scale whale distribution. Commercial in-water interactions increased from 596 in 2016 to 1,387 in 2019, declined to approximately 600 between 2020 and 2022, and subsequently increased to approximately 1,000 in both 2023 and 2024. Following the implementation of a mother-calf protection zone in 2016, humpback whale groups with calves were found to be more than eight times as likely to occur within the protection zone than groups without calves (β = 2.11 ± 0.11 SE, z = 20.08, p < 0.001; odds ratio = 8.27). The proportion of in-water interactions occurring within the protection zone increased from 68% in 2016 to 93% in 2024. Areas of concentrated in-water interactions represent potential hotspots of vessel and swimmer congestion that may increase pressure on humpback whales and elevate risks to visitor safety. These findings provide an evidence base to support adaptive management of commercial in-water humpback whale tourism, particularly for vulnerable mother-calf pairs, and contribute to strategies aimed at minimizing industry-related impacts while maintaining safe and sustainable tourism practices.

Commercial whale and dolphin watching tourism has grown considerably around the world over the past decades (O’Connor et al., 2009; Cisneros-Montemayor et al., 2010). While tours generally involve observing whales and dolphins from vessels, some tours include in-water interactions with cetaceans (Hoarau et al., 2020; Kessler and Harcourt, 2010). An in-water interaction occurs when a human enters the ocean and encounters a wild marine mammal in its natural habitat. In-water interactions can be accidental or intentional including through dedicated commercial tours. The central challenge for tour operators and managers as these additional activities become more popular is to ensure that whale watching remains sustainable, economically viable, and safe for the public, while avoiding negative impacts on the whales themselves.

Breeding Stock D (BSD) humpback whales undertake a consistent annual migration from high latitude Antarctic feeding grounds to low latitude breeding and calving areas along the Western Australian (WA) coast (Chittleborough, 1965). Humpback whales are present along the Ningaloo Coast from approximately May through October; however, the exact timing of the migration varies from year to year (Chittleborough, 1965). The migration corridor extends up the Western Australian coast and generally occurs within approximately the 200 to 300 m bathymetry from the coastline (Salgado Kent et al., 2012). The southern migration commences around August, with mothers and calves using known resting and nursing grounds along the way, including Exmouth Gulf adjacent to the Ningaloo Coast (Jenner et al., 2001; Irvine et al., 2018; Chittleborough, 1953; Bejder et al., 2019; Videsen et al., 2017; Christiansen et al., 2016).

The most recent population estimate for BSD is ~30,000 individuals, based on data collected in 2008 (Salgado Kent et al., 2012), suggests strong recovery from a near complete collapse due to intensive commercial whaling in the 20 th century (Bejder et al., 2016). Given an estimated population growth rate of ~10% per year at the time of the population estimate (Bannister and Hedley, 2001; Salgado Kent et al., 2012) the current population size is likely now considerably larger.

During migration along the Western Australian coast, humpback whales are exposed to multiple anthropogenic pressures, including vessel traffic, seismic exploration, fisheries interactions (Ferreira et al., 2023) and tourism-related activities, with a number of these pressures occurring repeatedly throughout the whales’ migration. Consequently, there is a need to develop an understanding of humpback whales and their use of the Western Australian coastline in relation to these pressures, to inform and improve management measures. There is growing recognition of this need along the Pilbara coast of Western Australia as it represents a key section of the migration pathway, includes several areas known as resting and staging sites for humpback whale mothers and is experiencing increasing pressures through industrial and recreational activity. Ningaloo Marine Park (NMP) is a site where a commercial tourism industry for boat based in-water interaction with humpback whales has begun, adding to the pressures in NMP. This activity is licenced and managed by the Western Australian Department of Biodiversity, Conservation and Attractions (DBCA). To ensure best practice management relies on assessing the effectiveness of management measures.

The approach and proximity of the vessels and swimmers can potentially affect whale behaviour and may result in negative impacts, such as the disruption of important activities such as resting, breeding and nursing. Elsewhere, such as in Tonga, research has shown that in-water tourism results in short-term effects on humpback whale behaviour (Fiori et al., 2020; Kessler et al., 2013). Studies conducted in 2016 and 2017, found similar short term behavioural responses by humpback whales in NMP during in-water interactions (Sprogis et al., 2020; Wall and Harcourt, 2017). Humpback whales displayed avoidance strategies, such as increased swim speeds and changes in travel direction, as well as a higher frequency of agonistic behaviours when a commercial tour vessel was <100 m away compared to when a vessel was >100 m away (Sprogis et al., 2020). Following the introduction of the in-water interaction program, several injuries and hazardous encounters between swimmers and humpback whales have been reported in NMP (Sprogis et al., 2020). Injuries to swimmers have also been reported during in-water interactions with humpback whales in the Kingdom of Tonga (Fiori et al., 2020) and the Reunion Islands (Barra et al., 2020; Hoarau et al., 2020). Elsewhere in Australia humpback whale in-water interactions are offered in Hervey Bay, Queensland and Jervis Bay in New South Wales. The humpback whales in Hervey Bay have shown significant short-term behavioural changes during in-water interactions, including direction changes and a decrease in resting during interactions when compared to whale watching tours (Stack et al., 2021).

To minimise disturbance to neonate calves and young in the shallow coastal waters of NMP (Irvine et al., 2018), a mother-calf protection zone (MCPZ) was introduced at the commencement of the commercial in-water interactions in 2016. While the MCPZ is intended to reduce disturbance, its effectiveness has not been evaluated.

Using five years of aerial survey data collected between 2016 and 2024, together with records of commercial in-water interactions, we evaluated whether humpback whale abundance, distribution, group composition and interaction intensity have changed following the introduction of commercial in-water interaction tourism in NMP. Specifically we assessed (1) temporal changes in humpback whale abundance and distribution, (2) use of the MCPZ by humpback whale groups with and without calves, (3) trends in interaction intensity and (4) the implications of these findings for adaptive management of commercial in-water humpback whale interactions in NMP.

Study area

NMP, gazetted as a marine park in 1987, and whose management is guided by the Ningaloo Marine Park Management Plan (Department of Conservation and Land Management, 2005), is located in the Pilbara region of Western Australia (Figure 1) and encompasses approximately 2,435 km² of coral reef, lagoon and shelf habitats. The park supports an important migration corridor and resting habitat for BSD humpback whales during their annual migration.

Commercial in-water interaction tourism commenced in 2016 under a regulated management program administered by the Western Australian Department of Biodiversity, Conservation and Attractions. The program includes a restricted number of licensed operators, specific conditions for vessel and swimmer approaches, and a designated MCPZ.

The MCPZ was implemented in the North West Cape area from the seaward edge of the reef out to 1 km west of Ningaloo Reef (Irvine and Salgado Kent, 2017; Department of Biodiversity Conservation and Attractions, 2020) as this was the area with the highest density of mother-calf pairs identified through aerial surveys (Irvine et al., 2018; Salgado Kent and Irvine, 2018). In 2017, the MCPZ was extended to include Coral Bay in the south and Bundegi in the north.

Research and monitoring forms part of the Management Program to inform ongoing management. This includes providing baseline information on humpback whale use of NMP and industry interactions as well as assessing specific research questions to ensure the Program continues to meet its objectives (Department of Biodiversity Conservation and Attractions, 2020).

Aerial surveys

Data collection

Each study year included ten systematic aerial surveys along 17 line transects (Figure 1). Survey periods varied: July 25 th - August 26 th 2016, August 7 th - October 20 th 2017, July 25 th - September 27 th 2022, July 15 th - September 26 th 2023, and July 12 th -September 26 th 2024. The survey design followed that of 2016 and 2017 described in (Irvine and Salgado Kent, 2017; Salgado Kent and Irvine, 2018) and covered the study area within state waters (3 nautical miles/5.6 km) with transects spaced 5 km apart. The transects ran angled perpendicular to the coast and bathymetric contours and covered the area between the seaward (westerly) reef edge and the western boundary of NMP ~3 nm from the reef or shore.

Aerial surveys were conducted using a single engine, high wing aircraft (Cessna 172), flying at an altitude of 245 m (800 ft) and a speed of 176 km hr-1 (95 knots). Survey personnel consisted of a pilot and two observers, one on each side of the aircraft, i.e. covering the port and starboard sides, linked via an intercom system that could be isolated from the pilot when necessary. Observations were recorded on a time-coded audio recorder synchronised with a hand-held GPS. Owing to GPS memory limitations, aircraft latitude and longitude were recorded every two seconds rather than every second. Aircraft positions corresponding to the intervening one-second observations were estimated post-flight by linear interpolation between successive GPS fixes, assuming constant aircraft speed and heading over each two-second interval. All devices (GPS, audio recorder and digital camera) were calibrated to ±1 sec accuracy prior to each flight.

At the beginning of each transect, the following environmental conditions were recorded: sea state (Beaufort scale), wind speed (knots), wind direction, cloud cover (oktas), visibility, turbidity, glare intensity. These conditions were updated during the transect whenever they changed. Angle of drift of the aircraft from the flight path (Lerczak and Hobbs, 1998) was recorded by the pilot. Two modes were used during the surveys: i) passing mode where observations were recorded without deviating from the transect line; and ii) closing mode where the aircraft broke from transect to collect more detailed information about group size and composition. Once the information was collected in closing mode, passing mode was resumed by returning to the location where the aircraft broke from transect. In these surveys, passing mode was used on all transects. The aircraft transitioned to closing mode to confirm humpback whale group composition only when a group was sighted within 1 km either side of the transect.

Observers determined whale group locations, while in passing mode and prior to transitioning into closing mode, by measuring the vertical angle from the horizon with Suunto PM-5/360PB clinometers and the horizontal angle from the aircraft’s travel direction using an angle board to the whale group. The position of each whale group was calculated post survey following Salgado Kent et al. (2012). In addition to location, group size, group composition, direction of travel, and calf colour (light grey or adult colour) were also recorded. Group composition was described in terms of the number of adults and calves present. Here, a calf was defined as a whale within proximity to another whale and visually estimated to be less than 2/3 of the length of the accompanying adult (Clapham and Mayo, 1987). Neonate and post-neonate calves were determined based on their colour and size compared with their mother, with light grey calves approximately ≤1/3 length of the mother classed as neonates and calves the same colour as the mother and > 1/3 length of the mother classed as post-neonates (Irvine et al., 2018). Surveys were restricted to Beaufort Sea states ≤ 4, as higher wind speeds and whitecaps reduce sighting probability.

The position (latitude and longitude) of each group of whales was calculated post survey using the R function ‘destPoint’ in the package geosphere (Hijmans, 2024) in R v4.5.2 (R Development Core Team, 2025). ‘destPoint’ requires radial distances from the position on the trackline in which the observation was made. Distances were calculated by: RD = h * tan (θ), where RD is the radial distance, h is the height of the observer from the surface of the ocean, and θ is the vertical angle up to the group from directly below the aircraft (calculated by subtracting the declination angle from the horizon from 90°) (Lerczak and Hobbs, 1998). The measured angles from the aircraft to the group were corrected for the course and drift of the aircraft by: AW = AC + MHA ± DA, where AW is the true angle to the whale, AC is the aircraft’s course, MHA the measured horizontal angle and DA the angle of the aircraft’s drift that was subtracted or added depending upon the side of the aircraft the sighting was on (as defined and described in Lerczak and Hobbs, 1998 and applied in Salgado Kent et al., 2012). Perpendicular distance to the whale groups was then calculated using the formula.

Where x = perpendicular distance for Distance analysis, r = radial distance and ∅ = angle between track line and whale group position.

Abundance estimation

Data collected systematically from humpback whale aerial surveys conducted in 2016, 2017, 2022, 2023 and 2024 were used to estimate the abundance of humpback whales that migrated through NMP in each of those seasons. Sightings recorded during both passing and closing modes were included only when the distance used in the detection function corresponded to the group’s position at first detection while the aircraft was on transect. Given that during the full migration season whales could be traveling north (early in the season) or south (later in the season) we considered scenarios to avoid double counting of individuals. Three scenarios were evaluated to account for whales of unknown travel direction: (1) allocation to the southward migration, (2) random allocation according to the observed directional proportions within each survey, and (3) exclusion from directional analyses. Scenario 2 was adopted for subsequent analyses because it retained ambiguous observations while avoiding the extreme assumptions of Scenarios 1 and 3. A sensitivity analysis comparing the three approaches is provided in Supplementary Material.

Abundance was estimated using conventional line-transect distance sampling implemented in Distance (Buckland et al., 2004, Buckland et al., 2001; Thomas et al., 2010). This is preferable to raw encounter rates or uncorrected counts, which assume constant detectability and would be biased by changes in sight-ability. Distance sampling also provides a formal framework for estimating density and abundance while accounting for imperfect detection (Buckland et al., 2001).

These abundance estimates are only the proportion of the population that travelled through NMP during the study years at the time of the surveys. These are not estimates for the entire BSD population. Abundance estimates were corrected for whales being unavailable for detection when they were underwater and not visible to observers (Salgado Kent et al., 2012; Hodgson et al., 2017). Estimates of N ^ g s abundance was calculated using a Horvitz-Thompson like estimator:

Where n is the number of detections during any one flight, A is the size of the study region, p(x) is the average probability of detection for each survey, L is the total survey effort and w is the total perpendicular right-truncation distance. The right-truncation distance is the distance out towards a selected boundary of either side of the aircraft beyond which observations are excluded from the analysis. This data truncation helps to prevent a poor fit of the detection function at large distances from the transect line. Left truncation was evaluated using sensitivity analyses. Inspection of near-track detections indicated a blind strip within 50 m of the transect line, and detections within this distance were excluded. Detection functions were evaluated using right truncation distances of 1.5, 2.0 and 2.5 km. Final model selection was based on goodness-of-fit, AIC and stability of abundance estimates. Candidate detection functions were compared using AIC, goodness-of-fit statistics and visual inspection of Q–Q plots. The most parsimonious model within ΔAIC ≤2 and with acceptable goodness-of-fit was selected. More details are provided in the Supplementary Material.

To account for humpback whales that were missed due to being unavailable for detection, we used availability bias values from previous studies; 0.64 (Salgado Kent et al., 2012) and 0.63 (Hodgson et al., 2017). The availability bias correction factor from Salgado Kent et al. (2012) was estimated from vessel surveys calculated at a similar location to our aerial surveys. The availability bias correction factor from Hodgson et al. (2017) was calculated from aerial surveys at a different location (Stradbroke Island, Queensland). Calculation of perception bias was not possible as the surveys were single platform surveys consisting of a single observer on each side of the aircraft, rather than paired observers on each side. Abundance estimates were corrected for availability bias but not perception bias. If perception bias occurred, abundance estimates are likely negatively biased. Although survey protocols, aircraft platform, observer configuration, and environmental constraints were standardised among years to minimise variability in detectability, residual differences in observer detection probability may contribute to uncertainty in abundance estimates and interannual comparisons. Consequently, abundance estimates should be interpreted as minimum corrected estimates.

Binomial Generalised Additive Models (GAM) with a logit link functions were used to investigate the intra- and inter-annual variation and migration direction variation in the number of humpback whale groups, and the proportion of groups with calves. Variation was modelled using a GAM smooth term applied to humpback whale groups and groups with calves which was used to identify the peak timing of number of humpback whale groups and the proportion of groups with calves observed from each migration direction. Furthermore, these variations were used to identify the approximate transition point in time when the north and south migration crossed. This point enables the investigation of inter-annual variation in migration timing (Chittleborough, 1953).

We also tested whether the presence of a calf in a humpback whale group was associated with the occurrence of the group being observed in the MCPZ. A binomial GAM was used to examine the probability of humpback whale groups occurring within the MCPZ, with calf presence, year and day-of-year (DOY) included as predictors. Seasonal variation was modelled using a smooth term applied specifically to calf groups. Annual abundance estimate growth rate was estimated from 2022–2024 only, as the surveys in those years were more systematic in temporal coverage of the humpback whale migration period.

Comparison of distribution between years

To compare the spatial distribution of humpback whale groups migrating through NMP between years, we used spatial point-pattern methods in R 4.5.2 (R Development Core Team, 2025), specifically Ripley’s K-function K (t) (Dixon, 2013) in the spatstat package (Baddeley et al., 2015). Treating each group as a point within the survey effort window, we tested whether whale groups were more clustered or dispersed than expected under a random distribution and whether that clustering or dispersion had changed between years. This can help to detect if there are any shifts in the migration corridor as the humpback whales pass through NMP.

We compared the spatial distribution of humpback whale groups between survey years. For each pair of years, we combined the two-point patterns within an observation window and randomly permuted the year labels of all points (1000 simulations). A global test statistic was calculated as the integrated squared difference between the two K-functions across distances up to 2 km. Pairwise tests across all five years yielded a 5 × 5 matrix of p-values, which we additionally adjusted for multiple comparisons using the Benjamini–Hochberg false discovery rate procedure (Haynes, 2013).

Data collection for in-water interactions

Commercial in-water humpback whale interaction data were collected using an Electronic Management System (EMS) installed as a licence condition on each licensed commercial tour vessel. This system was used to record details of all interactions including date, interaction start, geographical position, whale group size, number of swimmers, start and end time of interactions, whether there were repeated swims with the same whale group, and behavioural state of the whales before and after in-water interactions. These data were then imported into the main EMS database and quality controlled for missing information during the import process. The time spent with any whale or whale group must not exceed 60 minutes (Department of Biodiversity Conservation and Attractions, 2020). Multiple operators may interact with the same whale or group, but not at the same time, as long as the total cumulative time does not exceed this limit (Department of Biodiversity Conservation and Attractions, 2020). It is possible for an operator to engage in multiple in-water interactions with the same whale or whale group within the allotted time limit, however, each of the in-water interactions is treated separately. Behavioural states were defined as travelling (exhibiting directional movement where the direction can change, with low levels of surface-active behaviours), resting (laying stationary submerged or at the surface, not swimming or exhibiting surface behaviours) or surface active (energetic breaches as well as the repetitive slapping of a pectoral fin or flukes on the water surface (Kavanagh et al., 2017).

Distribution of in-water interactions and identification of high use areas

EMS data collected each year between 2016 and 2024 were used to investigate the spatial density distribution of the humpback whale in-water interactions. The locations recorded at the beginning of the humpback whale interactions (latitude and longitude) were overlayed on a map of NMP with 1km 2 grid cells. Kernel densities of the number of encounters per km 2 were calculated in QGIS 3.40.11 using Python 3.7 using cell sizes of 500m x 500m and a search radius of 1000m (from the centre of the grid cell). Kernel density plots were produced for all years combined. A threshold of the mean density of humpback whale in-water interactions plus two standard deviations were calculated for the study area. Using the R package 4.5.2 (R Development Core Team, 2025) and sfhotspot (Ashby, 2023) the Gi* statistic was calculated (Getis and Ord, 1992) to identify high and low cluster values spatially. The Gi* statistic was used via the hotspot gistar function in the sfhotspot R package, in combination with the density threshold, to identify hotspots of recorded commercial in-water humpback whale interactions across all years combined for the study area. The number of commercial tour operator vessels in each grid cell were overlayed on the plots.

Aerial survey summary

Survey timing varied among years, with surveys conducted between (Figure 2). The number of humpback whale groups recorded increased from 460 in 2016 to 866 in 2024.

The peak number of humpback whale groups for each migration direction was derived from the GAM predictive curves. The timing of the northern migration peak of humpback whale groups always occurred at the beginning of the study period each year (Figure 3). The timing of the southern migration peak humpback whale groups varied between years from mid-August to mid-September. The timing of the peak number of humpback whale groups from an unknown migration direction occurred around the transition time from northern to southern migration (Figure 3).

The proportion of humpback whale groups containing calves differed significantly among years (χ² = 68.88, df = 4, p< 0.001). After accounting for seasonal variation, the effect of year was reduced (Figure 4).

The peak proportion of humpback whale groups with calves from the predictive GAM curves on the northern migration were near the end of August for 2016 and 2017, but varied between end of July (2022) to the end of September (2023 and 2024) (Figure 5), whereas the peak proportion of humpback whale groups with calves on the southern migration varied between the end of August (2016) to the end of October (2017) and were approximately at the end of September in 2022, 2023 and 2024 (Figure 5). The probability that a group contained a calf differed significantly among years (χ² = 42.96, df = 4, p< 0.001), with higher proportions observed in 2016, 2017 and 2023 compared with 2022 and 2024.

Humpback whale groups with calves were predominantly observed in the MCPZ (Figure 6A). Calf presence in a humpback whale group was a highly significant predictor of MCPZ use (β = 2.11 ± 0.11 SE, z = 20.08, p< 0.001). Humpback whale groups with calves had a significantly higher odds of occurring within the MCPZ than humpback whale groups without calves (odds ratio = 8.27). Humpback whale groups without calves exhibited a low baseline probability of MCPZ use (β = −1.51 ± 0.12 SE, p< 0.001), corresponding to an estimated probability of approximately 18%.

A significant non-linear seasonal effect was detected for calf groups (edf = 3.34, χ² = 440.5, p< 0.001). No significant seasonal effect was detected for non-calf groups (Figure 6B).

Among year differences in MCPZ use were not statistically significant (all Year effects p > 0.05). The model explained 13.9% of deviance in MCPZ occupancy (adjusted R² = 0.178), with humpback whale groups with calves and seasonal timing accounting for the majority of explained variation.

Comparison of distribution between years

The Ripleys K(t) function statistic (Dixon, 2013) pairwise random-labelling tests of K-functions showed that while all years exhibited spatial clustering, there were no statistically significant differences in pattern between any survey years (all p > 0.05) (Figure 7; Table 1). While some pairs, such as 2016–2022 and 2016–2024 (p = 0.118), showed modest differences, these did not exceed random expectation. The largest p-value was observed between 2022 and 2024 (p = 0.848).

Abundance estimates

Abundance estimates for northern migration were relatively robust to direction allocation assumptions, with variation across scenarios ranging from 1–22% among years. In contrast, estimates for the southern migration were substantially more sensitive, varying by 30–47% depending on allocation scenario. Weighted log-linear models confirmed a significant Scenario × Direction interaction, indicating that allocation disproportionately influenced southern migration estimates. Despite this directional sensitivity, total seasonal migration (north + south combined) was comparatively stable across scenarios, with inter-scenario variation substantially smaller than directional differences. Abundance estimates are presented based on Scenario (2).

The Cramer-von Mises goodness of fit tests for the most parsimonious models using the hazard rate and half normal key functions for each year were not significant and therefore indicate a good fit (Table 2). Group size was retained in the selected detection functions for all years and migration directions (Table 2). Turbidity was retained in the 2022 southern migration, 2023 north and south, and 2024 northern migration (Table 2).

Line-transect distance sampling with a right and left truncation distance of 2 km and 50 m respectively, were used to produce individual abundance estimates for each survey year (Table 3). Mean group size ranged from 1.86 – 2.03 North and 1.70 – 2.20 South. Applying availability corrections using surfacing probabilities of 0.64 (Salgado Kent et al., 2012) and 0.63 (Hodgson et al., 2017) increased the abundance estimates proportionally, resulting in corrected individual abundance estimates (Table 3).

In 2016 the predicted northern migration peak occurred before the first survey, whereas the predicted southern migration peak occurred near the end of the survey period (Figure 8A). In 2017, the surveys captured the majority of the southern migration and the end of the northern migration. The highest predicted abundance estimate for the northern migration was at the beginning of the survey period, whereas the predicted peak abundance estimate for the southern migration was at the beginning of September (Figure 8B). The more systematic surveys conducted in 2022, 2023 and 2024 captured a more even distribution of both the northern and southern migration (Figures 8C–E). The predicted peak abundance estimates for the northern migration are at the beginning of the survey period, while those for the southern migration varied between the beginning and mid-September (Figures 8B–E). No transition from northern migration to southern migration was predicted in 2016, however in 2017, 2022, 2023 and 2024, the transition from northern migration to southern migration was around mid-August.

Comparisons of abundance estimates among years are presented for 2022-2024. From 2022–2024, northern migration abundance increased between 2022 and 2023 and decreased slightly in 2024. Southern migration abundance also increased between 2022 and 2023, with slight decline in 2024.

In-water humpback whale interactions

In 2016 13 licences were allocated to existing commercial whale shark operators in Exmouth and Coral Bay to conduct in-water humpback whale interactions during the migration season, which generally runs from July to October. In the first two years of the trial, 2016 and 2017, the number of in-water humpback whale interactions were similar at 596 and 594 respectively before doubling to 1173 in 2018 and continuing to increase to 1387 in 2019 (Figure 9). In 2020 in-water humpback whale interactions dropped to less than half of 2019 (643), before decreasing further in 2021 and 2022 to similar levels to those in 2016 and 2017 (Figure 9). In 2023 in-water humpback whale interactions increased to 965 and then slightly decreased to 892 in 2024 (Figure 9).

The proportion of in-water humpback whale interactions inside the MPCZ increased from 68% in 2016 to 93% in 2022 and has remained relatively stable at around 93% (Figure 10).

The distribution of high-density areas for both in-water interactions off Tantabiddi (Figure 11A) and vessels engaged in the interactions (Figure 11B) and Coral Bay (Figures 12A, B) followed the seaward edge in the MCPZ, with the highest concentrations close to the Tantabiddi and Coral Bay boat ramps.

This study provides the first multi-year assessment of humpback whale distribution, abundance, group composition and commercial in-water interaction intensity within NMP following the introduction of commercial in-water interaction tourism. We found that the spatial distribution of humpback whales remained stable between 2016 and 2024, humpback whale groups containing calves were substantially more likely to occur within the MCPZ than groups without calves, hotspots of commercial in-water interactions coincided with areas of high calf-group occurrence, and abundance increased between the standardised surveys conducted from 2022 to 2024 consistent with the broader recovery of BSD reported elsewhere. Understanding the pressures on humpback whales from in-water interactions in NMP is important to inform and adapt the management program along the Ningaloo coast to continue to meet its key objectives. This study demonstrates the application of EMS data combined with data collected from aerial surveys, to investigate the pressure of the in-water interactions on the migrating humpback whales in NMP and to identify areas of high-use.

Annual variation in the proportion of groups containing calves was observed although differences in survey timing may have contributed to some of this variability. Interannual fluctuations in reproductive output has been documented in other humpback whale populations (Clapham and Mayo, 1987; Baker et al., 1987; Pallin et al., 2018), and may reflect broader environmental variability (Li et al., 2023; Van Weelden et al., 2021) affecting female body condition and reproductive success. For example, in Hawaii, a study between 2008–2018 showed that the humpback whale encounter rates fluctuated due to environmental influences (Cartwright et al., 2019). Between 2013 and 2018 mother and calf encounter rates dropped by 76.5%, which suggested a rapid reduction in the reproductive rate of humpback whales in the region (Cartwright et al., 2019).

The proportion of humpback whale groups with calves observed in the MCPZ varied between years. However, we illustrated that humpback whale groups with calves have a significantly higher odds of being observed in the MCPZ, than humpback whale groups without calves. Interestingly humpback whale groups with calves were more than eight times more likely to be observed in the MCPZ than those groups without calves. Consistent with other studies around the world, mothers with calves are more likely to be observed in shallow sheltered waters closer to the shore, than those without calves as has been seen in Jervis Bay, Australia (Bruce et al., 2014), Ecuador (Félix and Botero-Acosta, 2011) and the Dominican Republic (Mattila et al., 1994).

The spatial point-pattern analyses indicated that the fine-scale distribution of humpback whales within NMP migration corridor has remained relatively stable across all five survey years. This supports the view that humpback whales consistently use the same nearshore migration pathway along the reef slope and adjacent coastal waters, despite changes in abundance, demographic composition, or fluctuations in vessel traffic and in-water activity. A shift in the spatial distribution of the humpback whales and the in-water humpback whale interactions in NMP might indicate increased pressure on the humpback whales from in-water interactions. An example of such a shift may have occurred along the Hawaiian coast, where wintering humpback whales experience considerable levels of recreational and tour boat traffic. There was some concern that the movements of mother-calf pairs off the coast of Maui were affected by levels of boat traffic (Glockner-Ferrari & Ferrari 1985). As the nearshore calf encounter rate decreased, there was an increase in sightings three to four km offshore, which suggested a shift to the less congested offshore waters (Salden, 1988). A similar shift has not been detected in NMP.

In-water humpback whale interactions

Between the start of the in-water humpback whale interaction trial in 2016 and 2019 there was an increasing trend for the number of in-water interactions each season. COVID-19 pandemic, quarantines and the restrictions imposed on travel to WA, both nationally and internationally in 2020 and 2021, contributed to a drop in tourism interactions to less than half that of 2019. Even though the WA border reopened for international travel on 3 rd March 2022, the number of in-water interactions did not increase that year, and were lower than the initial year of the trial in 2016. Most licensed tour operators also hold licences to conduct in-water interactions with whale sharks in NMP. Whale shark interactions start earlier in the year in April, as the whale sharks are present in NMP earlier than the humpback whales, and extend to July. The whale shark interaction season has expanded into August as whale sharks have been found to continue to be present. Further investigation on how the whale shark and humpback whale interactions overlap, spatially and temporally could be beneficial to management especially if high use areas for whale sharks and humpback whale in-water interactions overlap (Tyne et al., 2025).

From 2016 to 2024 the highest proportion of in-water humpback whale interactions occurred in the MCPZ and increased each year. Humpback whale interactions are prohibited with groups containing calves less than half the length of an accompanying adult (Department of Biodiversity Conservation and Attractions, 2020). This licence condition is to protect young, neonate calves from disturbance and to minimise the risk of tourists being injured by mothers protecting their young, vulnerable calves.

Mother–calf pairs are likely to be particularly vulnerable to repeated disturbance because calves have limited energy reserves and depend on frequent nursing, resting and close maternal association (Bejder et al., 2019). Repeated vessel approaches, swimmer entries or interruptions may increase energetic costs for both mothers and calves through avoidance responses, longer dives, changes in direction (Sprogis et al., 2020). Evidence from other swim-with-whale locations supports this concern. In Hervey Bay, in-water interactions altered humpback whale behaviour and were associated with more direction changes and reduced resting compared with whale-watching encounters, raising concerns about energetic costs on a resting ground (Stack et al., 2021). In Tonga, research on Vava’u has shown behavioural responses of humpbacks to vessel and swimmer approaches in a tourism setting focused heavily on mother–calf pairs (Fiori et al., 2020). In Réunion Island, mother–calf groups were frequently exposed to whale-watching and swim-with activities, and reports indicate that vessel and swimmer presence can reduce resting and nursing opportunities (Hoarau et al., 2020).

The implications are broader than whale welfare alone. High swimmer and vessel density also increases risk to people, where animals may change direction suddenly or where multiple vessels compete for position. If crowding reduces encounter quality, increases safety risks, or leads to visible disturbance, it may also undermine the long-term sustainability of the tourism industry.

It is noticeable that the highest number of in-water humpback whale interactions was concentrated in the same 1km 2 grid cell each year, the centre of which is approximately 4.2km south of North Passage, with a maximum density of 55 interactions in 2024, 51 in 2019, 40 in 2018, and 38 in 2023. This is likely due to tour operators being able to offer successful in-water humpback whale interactions due to proximity to the boat-launch area, reducing vessel effort, thus minimising travel time and fuel costs. It is also noticeable that the grid cell where the highest density of interactions occurs, also includes the highest number of vessels involved in those interactions. Concerns of congestion on the water with vessels and swimmers, and in the air with spotter planes has been raised, and could potentially increase pressures on targeted humpback whales that are being interacted with and increase potential risks to swimmers. While such effects have been proposed in studies from other humpback whale tourism destinations (Stack et al., 2021; Fiori et al., 2020; Hoarau et al., 2020), the extent to which they affect whales at NMP remains uncertain and warrants further investigation. Consequently, the potential for cumulative effects on calf development, maternal condition, recruitment, or population dynamics should be viewed as a precautionary consideration.

The concentration of interactions close to reef structures raises the possibility that vessel congestion could restrict whale movement in some situations. In Panama, crowding of humpback whales from vessels was suggested to be restricting their breaching behaviour (Amrein et al., 2020). Furthermore, there is the potential to reduce the quality of visitor experience based on the presence of more than one vessel with swimmers (Department of Biodiversity Conservation and Attractions, 2020). Consequently, the effects of congestion could have implications for Ningaloo’s international reputation as a world leader in the management of marine fauna interactions if the target animals are impacted or the visitor experience is compromised. However, the present study did not quantify boxing in, vessel crowding, avoidance responses, entrapment risk, or swimmer safety incidents. As such, these issues should be regarded as potential management concerns and given due consideration in future evaluations.

The management implications of this study arise directly from the observed spatial overlap between humpback whale groups with calves and in-water interactions in the MCPZ. Humpback whale groups with calves were significantly more likely to occur within the MCPZ than groups without calves, indicating that the MCPZ encompasses habitat that is disproportionately important for vulnerable mother–calf pairs. This finding provides empirical support for spatial protection measures within the MCPZ and suggests that management actions should prioritise habitats used most frequently by humpback whale groups with calves.

Although the present study did not quantify behavioural responses or demographic consequences, the spatial overlap identified here also supports consideration of other management measures aimed at reducing repeated exposure of mother–calf pairs to in-water interaction activity. Such measures could include limits on the number of vessels and swimmers, restrictions on repeat approaches, or adaptive management responses triggered by vessel congestion. More broadly, the results support the use of spatially explicit management approaches in which tourism pressure is aligned with habitat use patterns. Because humpback whale groups with calves were not distributed uniformly throughout the study area, management interventions can be targeted toward locations where conservation benefits are likely to be greatest. This approach allows protection efforts to focus on biologically important habitat while minimising unnecessary restrictions in areas less frequently used by humpback whale groups with calves. Together, these findings suggest that the effectiveness of the MCPZ should not be evaluated solely based on its spatial extent, but also on the degree to which in-water interaction activity within and around the zone is managed to reduce pressure on the whale groups it was designed to protect. Continued monitoring of humpback whale distribution, interaction intensity and compliance with management regulations would help provide evidence for assessing whether current protection measures remain effective as tourism activity increases.

As humpback whales migrate along the WA coast, they are exposed to a range of pressures in addition to commercial whale watching and in-water interactions (Ferreira et al., 2023). These include underwater noise pollution, shipping traffic, fisheries interactions, recreational vessel activity (Groom and Coughran, 2012; How et al., 2021; Peel et al., 2018; Vanderlaan and Taggart, 2007; Schilling et al., 2023). A number of these pressures may occur repeatedly throughout the whales’ migration path and there is a paucity of knowledge on the contemporary humpback whale distribution in relation to these pressures. Consequently, there is a need to develop an understanding of humpback whales and their use of the Western Australian coast in relation to these pressures, to inform and to improve the management of these pressures.

This study provides the first multi-year assessment of humpback whale distribution, abundance, calf-group occurrence and commercial in-water interaction intensity within NMP. An increase in the abundance of humpback whales migrating through NMP during the study period is apparent, consistent with previously reported growth in the Western Australian humpback whale population. Humpback whale groups containing calves were disproportionately associated with the MCPZ, and hotspots of commercial interactions occurred within these areas. This supports a precautionary approach, where adaptive management action should aim not only to prevent disturbance but reduce cumulative interaction pressure within the MCPZ, that is used disproportionately by humpback whale groups with calves, while also ensuring safe and sustainable interactions for people and whales. Although no evidence of changes in whale distribution was detected over the study period, the observed spatial overlap between vulnerable groups and tourism activity highlights the value of continued monitoring to assess whether current management measures remain effective as tourism activity evolves.