Introduction

Marine ecology and hydrodynamics are intrinsically linked; with hydrodynamic features strongly shaping the environment for marine taxa (Lowe and Falter 2015). This relationship spans the entirety of the marine environment, with examples ranging from the interactions between tide and wave exposure on bivalve settlement (Esqueda-González et al. 2022), to eddy formation and the feeding of fin whales (Cotté et al. 2011). Hydrodynamic drivers similarly influence biological processes in coral reef environments. For example, currents transport off-reef dissolved organic matter, detritus, and plankton into reef waters, where these external resources are assimilated by primary producers, suspension feeders, and planktivores, driving the productivity of numerous marine organisms (Hamner and Hauri 1981; Sebens et al. 1998; Brocke et al. 2015; Morais et al. 2021; Gahan et al. 2023). Furthermore, currents are key drivers of the completion of the bi-partite lifecycles of many reef organisms influencing the potential for retention or dispersal of larvae, and the return of larvae from open water to reefs at settlement (Brandl et al. 2019; Gouezo et al. 2021). Hydrodynamics are also one of the main drivers of sediment dynamics on coral reefs (Schlaefer et al. 2021) which can influence numerous biological processes (Tebbett and Bellwood 2019) ranging from the settlement of coral larvae (Leong et al. 2018) to turf algae interactions with herbivorous reef fishes (Tebbett et al. 2017). Hydrodynamics and flushing times are also seen to have an influence on water temperature (Green et al. 2019b) and hence coral bleaching, when high air temperatures and long residence times allow stagnant water to heat without the influence of cooler offshore water (Reid et al. 2020). Lizard Island is one such coral reef on the Great Barrier Reef (GBR) where an understanding of the physical environment is essential to explain observations within ecological studies. However, despite Lizard Island being one of the most ecologically well studied reefs in the world (Vail et al. 2023), there is limited hydrodynamic knowledge of the area at ecologically relevant scales (i.e. within 2–12 m of the reef (Deoraj et al. 2023)).

Larger-scale circulation patterns, and hydrodynamic drivers, have been examined on multiple coral reefs around the world. For example, Andréfouët et al. (2023) described the circulation of Apataki Atoll in French Polynesia to understand the physical environment of black pearl farms using an array of hydrodynamic instruments including temperature/pressure loggers and ADCPs (acoustic Doppler current profilers). Again, in French Polynesia, Bruyère et al. (2023) characterised the hydrodynamics around the Gambier Islands using similar instrumentation but including Marotte HS current meters. The results were used to validate a hydrodynamic model to interpret larval dispersal. Similarly, in Western Australia Green et al. (2019a, b) examined Scott Reef to understand the dynamics of cold-water tidal bores into the deep lagoon in relation to nitrate transport and physical processes driving the spatial variability in temperature during the 2016 global coral bleaching event. Similar approaches were used by Grimaldi et al. (2022, 2023) to measure physical processes across Mermaid Reef Atoll in the Rowley Shoals, Western Australia, to investigate the hydrodynamic processes driving the circulation patterns underpinning temperature variation across reef zones. However, Bruyère et al. (2022) examined the processes controlling the hydrodynamics of lagoons in New Caledonia. They sought to understand the ocean-lagoon exchanges, circulation, sea level dynamics, temperature, and salinity variability. Finally, Lalau et al. (2022) investigated Poé-Gouaro-Déva lagoon in New Caledonia to understand the water budget and to calculate a flushing time for the lagoon. Using an array of hydrodynamic instruments, including Marotte HS current meters, Lalau et al. (2022) identified the main drivers for the variation in flushing time and how this impacts eutrophication within the lagoon.

Despite the contribution of these studies to our general understanding of hydrodynamic processes on reefs, our understanding of the influence of water movement at small, reefal scales (within 2–12 m of the reef) is limited, especially in intensively studied locations such as Lizard Island. The overarching goal of this study, therefore, is to better understand coral reef ecology in the context of on-reef, ecological-scale hydrodynamics.

The Lizard Island region is home to a long-term monitoring array supported by Australia’s Integrated Marine Observing System (IMOS). This mooring is located on the continental slope to the north of Lizard Island (14.69138889° S, 145.46638889° E) and has been deployed from October 2007 to May 2014, and November 2019 to present. This long-term monitoring array was designed to monitor the variability and position of the Western Boundary currents of the Coral Sea (Ridgway et al. 2018). It does not provide information at ecologically relevant spatial scales, of a few metres, regarding the circulation of the local Lizard Island region. Using the observations from the IMOS monitoring arrays, eReefs (a collaborative group between CSIRO, the Great Barrier Reef Foundation, Australian Institute for Marine Science, Bureau of Meteorology, and the Queensland Government) created multiple hydrodynamic models at various resolutions to resolve the circulation of the GBR. Of these, eReefs GBR1 model has the highest resolution at 1 km scale. This far exceeds ecologically relevant reef scales (the reef scale being characterised by Deoraj et al. (2023) as a resolution of approximately 2–12 m; although even finer scales of 0.1 to 2 m are likely to be even more important). Resolving reef-scale processes requires sufficient resolution to examine features such as lagoonal passages less than 70 m wide, as seen within the Lizard Island lagoon, to accurately represent their influence on the circulation. Clearly, the eReefs GBR1 hydrodynamic model is unable to resolve the circulation within the Lizard Island lagoon at reefal scales.

Furthermore, only a few studies, conducted by Leis (1986), Frith et al. (1986), and Johansen (2014), have quantified the currents around the Lizard Island group. All were limited in time and space. Current speed and direction to windward and downwind of Lizard Island were studied by Leis (1986), to understand plankton abundance and distribution, using cruciform drogues at two sites external to the lagoon (at the water surface and at a depth of 7.5 m). Surface currents were primarily wind driven, with a predominant northerly current flow during the winter south-easterly (SE) trade wind season. At the same time, Frith et al. (1986) quantified supply currents to Lizard Island from Carter Reef using Aanderaa current meters at three stations external to the Lizard Island lagoon; again, flow was predominantly north to north-westerly during the winter with the influence of the SE trade winds. However, the summer months saw an increase in the variability, but currents were still highly wind driven. More recently, Johansen (2014) deployed load cell sensors around the perimeter of Lizard Island to measure current speed; on exposed, oblique, and sheltered reef slopes at 3 m, 6 m, and 9 m depths (27 sites total). Johansen (2014) also found that the currents surrounding Lizard Island were predominantly wind driven, with the greatest influence in shallow, exposed locations; tidal currents in contrast were minimal. Despite gaining essential hydrodynamic information, these studies were spatially or temporally constrained, being restricted to study sites external to the lagoon and/or over relatively short periods of time (just 6 to 22 days). Finally, using Marotte HS current meters, Tebbett et al. (2022b) provided preliminary data on mean current flows within the lagoon in relation to cyanobacterial outbreaks, observing a higher percentage of cyanobacterial mat cover in areas with lower current speeds. The maximum current speed for microbial mat persistence was 0.24 m s−1.

Our study differs markedly from these earlier works in that the primary goal was to characterise water movement across the Lizard Island lagoon at a scale dictated by ecology, i.e. in the locations where fish and corals live. The emphasis was on describing near-reef water movement (within 1–3 m of the reef) and then to determine the extent to which physical processes, namely wind and tide, come together to influence this water flow. At present, the circulation of the Lizard Island lagoon is largely unknown. The overarching aim of this study, therefore, is to deepen our understanding of how physical processes influence the circulation at the reef scale, especially within the Lizard Island lagoon. This will be achieved by using observations from 23 to 27 Marotte HS current meter records spanning a 2-year period and analysing patterns in accordance with wind and tide frequencies. In addition to describing water movement, flushing times within the lagoon will be calculated, under a range of wind and tidal conditions, allowing flushing time to be hindcast for the 2016 coral bleaching event at Lizard Island, to explore the potential role of hydrodynamics during coral bleaching. These observations will provide a key step in bridging the gap between hydrodynamic and ecological knowledge at Lizard Island.

Methods

Study site

The Lizard Island group is a complex and dynamic coral reef system, approximately 27 km offshore from the Australian mainland on the Great Barrier Reef (GBR). It is composed of a main granite island (Lizard Island) of approximately 10 km2 (at 14.66888889°S, 145.45944444°E), and three smaller islands (Palfrey Island, South Island, and Bird Islet). This island group is surrounded by a fringing reef that semi-encloses a 2–15-m-deep lagoon (Fig. 1). Herein, we define the Lizard Island lagoon as the area extending from inside the reef crest to the shoreline of each island. The lagoon is broadly exposed to the prevailing SE trade winds during the Austral winter. This dominant wind direction in the winter causes a prominent north to north-westerly flow between Lizard Island and the Ribbon Reefs within the Great Barrier Reef lagoon, with very little cross-shelf flow (Frith et al. 1986). The variability of weather conditions, and currents, increases in summer during the monsoon season, with increase in the cross-shelf motion (Frith et al. 1986). Lizard Island has a maximum elevation of 359 m which protects the lagoon from north to north-westerly winds during the Austral summer. Recorded water movement outside the lagoon at Lizard Island has a maximum speed of 0.38 m s−1, median 0.1–0.15 m s−1 (Leis 1986). The tidal signal is dominated by semidiurnal and diurnal frequencies (Frith et al. 1986) with a tidal range of 0.01–2.71 m during our study. Around the Lizard Island group, tidal currents dominate the circulation, in a predominant cross-shelf direction (Frith et al.1986). However, the non-tidal circulation is predominantly wind driven, with a significant relationship between wind stress and current velocity (Frith et al. 1986).

Fig. 1
figure 1

Deployment of Marotte HS current meters across the Lizard Island lagoon. The inset shows the location of Lizard Island in relation to the Great Barrier Reef, Australia. Colours indicate the three deployments: orange = Deployment 1, purple = Deployment 2, and turquoise = Deployment 3. Current meters have been grouped spatially (in black circles) and given a label based upon their spatial distribution. Current meters without groups are described in terms of site description. The light grey shading indicates areas of coral reef, while the dark grey shading indicates islands within the Lizard Island group. Yellow dot is the IMOS weather station. Individual deployment maps in (ESM Figs. S1, S2, S3)

Full size image

Current meter deployment

To measure the on-reef circulation of the Lizard Island lagoon, within the broader context of the greater Great Barrier Reef lagoon, an array of Marotte HS current meters (Marine Geophysics Laboratory, James Cook University) was deployed throughout the reef and lagoon area. The Marotte HS current meter is a drag-tilt current meter using an accelerometer and magnetometer to measure both current speed and direction. It has a speed range of 0.05–1.20 m s−1, with the minimum systematic error being ± 0.03 m s−1 between speeds of 0.10–0.70 m s−1. With these specifications, the Marotte HS current meter was an optimal sensor for current speeds around the Lizard Island. The near-reef water movement within the Lizard Island lagoon was quantified during three extended current meter deployments, using bottom-mounted Marotte HS current meters around the reef and across the lagoon over a period of approximately 2 years, from the beginning of 2021 to the beginning of 2023.

The first deployment, from the 1st March 2021 to 31st July 2021 (5 months), collected data across the end of the monsoon season and the first half of the trade wind season. A total of 29 Marotte HS current meters were deployed around the Lizard Island lagoon (Fig. 1; Electronic Supplemental Material (ESM) Fig. S1). This is henceforth designated Deployment 1. During the second deployment (Deployment 2), 25 current meters were placed in similar positions to Deployment 1 from 16th February 2022 to 27th June 2022 (over 4 months; Fig. 1). Again, at the end of the monsoonal season and during the predominant SE trades. Current meters in Deployments 1 and 2 were placed in locations within the lagoon chosen to describe the water movement in areas of biological significance (i.e. where most lagoonal coral reefs and fish populations are located). This was supplemented by additional current meters within the eastern channel entrance to the lagoon, the Bird-South and Palfrey-South reef slopes, and North Reef (Fig. 1; ESM Fig. S2). However, for this study we will solely be analysing current meters within the lagoonal area.

The third deployment (Deployment 3) consisted of 23 Marotte HS current meters from 2nd July 2022 to January 2023 (5 months). Their locations varied from the previous two deployments as some of the current meters were now positioned to capture the circulation within the lagoon, as well as the inflow and outflow during different wind regimes (Fig. 1; ESM Fig. S3). This deployment spanned the end of the SE trade wind season and the beginning of the summer monsoon season. The reduction in the number of Marotte HS current meters throughout each deployment was due to occasional flooding of the sensor housing, and rejection of instruments during data checking (instruments yielding anomalous data were not redeployed). Because our study was designed to quantify currents in the immediate proximity of reefs, i.e. to understand their influence on fish and coral communities, most were relatively shallow and within metres of a coral margin.

Across the span of these 2 years, La Niña conditions dominated, becoming stronger throughout the Austral summer and weaker or inactive during the Austral winter (as per Bureau of Meteorology ENSO forecast history). No other major weather events were observed during this time.

All Marotte HS current meters were calibrated before deployment following the Marotte HS-I User Guide. Current meters were deployed on snorkel or SCUBA by members of the Reef Function Hub at predetermined GPS locations (ESM Table S1). Each current meter was attached using a shackle to a 10-kg concrete block placed on the lagoon floor or on the sand adjacent to the reef (ESM Fig. S4). This allowed for the acquisition of near-bottom current measurements, approximately 30–50 cm above the substratum. The Marotte HS current meters produced a high-frequency time series, recording at 1 Hz. These data were processed using MarotteHSConfig programming software and averaged over 10-min time intervals, reducing the interference from high-frequency orbital wave velocities.

Wind data

Wind data were obtained from the Integrated Marine Observing System (IMOS) weather station at Lizard Island (https://thredds.aodn.org.au/thredds/catalog/IMOS/FAIMMS/Lizard_Island/Relay_Pole_2/catalog.html, date accessed: 17/02/2023). Wind speed (in km hr−1; 10-min scalar average) and direction (in degrees; 10-min vector average) were extracted for the period of 1st March 2021 until January 31st 2023. The weather station is in the lagoon (14.69138889° S, 145.46638889° E) on the north-western side of Bird Islet. Although optimally placed to capture the prevailing SE winds, the northerly winds associated with monsoon troughs may be partially blocked by Lizard Island.

Furthermore, this IMOS wind time series was discontinuous through the current meter deployment period. Therefore, wind data output within the eReefs GBR1 hydrodynamic model was explored to fill in measurement gaps (https://ereefs.aims.gov.au/ereefs-aims#ereefs-hydro-model, date accessed: 17/02/23). The U (east–west) and V (north–south) components of the model wind were extracted for the duration of the deployments at the location of the weather station. Overall, the magnitude and variability in direction and speed were relatively consistent between datasets (ESM Fig. S5). Specifically, the IMOS wind speed had a mean of 27.45 km hr−1 ± 11.03 standard deviation (STD) which was comparable to the mean speed of 27.87 km hr−1 ± 11.07 for the GBR1 model. Furthermore, there was a strong positive correlation between the observed and modelled U wind components (Pearson correlation, r = 0.85) and also a strong positive correlation between the observed and modelled V wind component (Pearson correlation, r = 0.79). Given the agreement between the observed and modelled wind data, it was determined that the GBR1 data could be used to fill the gaps in the observations (ESM Fig. S5e). Wind conditions were divided according to strength and direction. Strong SE winds were characterised by speeds of greater than 40 km hr−1 and occurred 12% of the time during our study. Weak SE winds were described as speeds less than 11 km hr−1, occurring for 20% of our study time. Typical SE wind conditions were defined as speeds between 11 and 40 km hr−1, occurring 46% of the time. North-westerly (NW) wind and no wind conditions occurred 18% and 4% of the time, respectively.

Tide/sea surface elevation

Sea surface elevation data were obtained from water pressure readings acquired from the IMOS weather station Relay Pole 2 and from the eReefs GBR1 hydrodynamic model (ESM Fig. S6b). However, the available IMOS sea surface elevation data only corresponded to the duration of Deployment 2. Therefore, the sea surface elevation data from the eReefs GBR1 model were again extracted for the location of the IMOS relay pole 2 and compared to the measurements to assess the validity of using the model data in place of measurements. For the comparison, all data were converted from the recorded water depth to Mean Sea Level (MSL), based on the pressure sensor deployed at 10 m depth. There was a strong positive correlation between the depth measurements and the modelled depth (Pearson correlation, r = 0.99; ESM Fig. S6.). Consequently, the tidal series from GBR1 sea surface elevation could be confidently used to provide tidal information across all deployments.

Quality control

All current meters went through a robust quality control procedure. Firstly, U and V time series from all current meters were visualised to look for data dropouts, spikes, or flatlines, and where these issues occurred, they were flagged as bad data. Next, current vectors were visualised in space, through time, and across the three deployments to look for congruence in measurements taken at similar locations across deployments. This resulted in one current meter being excluded from the analyses. Specifically, current meter B1686 in deployment 3, which was positioned at 5.7 m depth at the Lagoon Entrance, did not have a clear tidal signal, while current meters in approximately the same position in the previous deployments, and proximal current meters in the same deployment all had a clear tidal signal.

Data analysis

Wind and tide interactions

The combined influence of the wind and tide on the currents at Lizard Island was explored by averaging over selected times when the flood and ebb current speeds were likely to be peaking, during prevailing wind conditions. It was calculated that peak flood and ebb current speeds occurred in the hour at the midpoint between low and high tide, and vice versa (see ESM Fig. S7 for example). These peak flood and peak ebb current speeds were averaged across all flood and ebb tides in Deployment 2. This deployment was chosen to visualise because of the consistent prevailing SE winds throughout the deployment.

Tidal analysis

A classical harmonic tidal analysis with nodal corrections was conducted for both the IMOS observed sea surface elevation time series and each current meter U and V velocity components to evaluate the tidally driven component of the circulation at Lizard Island. We used the MATLAB package: T_Tide Harmonic Analysis Toolbox, which models ‘the tidal signal as the sum of a finite set of sinusoids at specific frequencies related to astronomical parameters’ (Pawlowicz et al. 2002).

Notably, tidal flows within a complex reef structure would resonate, diffract, and refract around the highly complex reef structures and bathymetry of the Lizard Island lagoon, confounding the tidal signal. This means that the tidal analysis could not extract the entire tidal signal from the measurements of current meters located within the lagoon and adjacent to the reef matrix, limiting the scope of the tidal analysis, i.e. a fraction of the actual tidal signal is still likely to be present in the residual value. The residual from the IMOS observed sea surface elevation time series, and each current meter U and V velocity components likely contain low-frequency wind influences as well as high- and low-frequency physical processes. However, the nonlinear interactions between wind and tide are outside the scope of this study.

The tidal signal predicted for each current meter was then compared against the original current meter observational data to estimate (conservatively) a percentage current flow that may be attributed to the tide. These calculations were undertaken using the tidal signal predicted percentage for the U and V components from the tidal analysis. These two percentages were added together (similar to that of a complex number: \(U+iV\)), then divided by 2, as seen in the following Equation:

$$\text{Predicted} \%= \frac{{U}_{\text{tide}}+{V}_{\text{tide}}}{2}$$
(1)

where Utide is the predicted tidal signal percentage for the U component out of a total of 100%, and Vtide is the predicted tidal signal percentage for the V component out of a total of 100%.

Wind and low-frequency components

A low pass filter was applied to the current meter data to extract the low-frequency current component. This analysis was conducted using the ‘cmglowpass’ tool from the MATLAB package: CMGTooL toolbox (Xu et al. 2002). A filter of 33 h was applied using the pl33 process (Limeburner et al. 1985) to remove high-frequency components, producing a time series of low-frequency signals within the current meter data. The low pass filter of 33 h was also applied to the wind data time series, using the combination of IMOS and eReefs GBR1 wind datasets, to understand the low-frequency component of the wind and its subsequent potential influence on the currents within the Lizard Island lagoon.

Transport and flushing time

The transport (or volumetric flow) and flushing time (also called renewal time in the literature; Andréfouët et al. 2001; Lucas & Deleersnijder 2020) were calculated following Lalau et al. (2022). Flushing time is defined as ‘the time for all water in the domain of interest to be replaced’ (Lucas & Deleersnijder 2020). It is represented by the time taken for a volume of water equal to the contents of the area of interest to be replaced, and is the lower bound of the turnover time for a system (Andréfouët et al. 2001). The flushing time (Tf) was derived from the ratio between volume (in m3) of the lagoon at time t (Vt) and the volumetric flow (measured in m3 s−1) entering and leaving the lagoon at time t (Qt) as per the equation (Gallagher et al. 1971; Monsen et al. 2002; Lucas & Deleersnijder 2020; Lalau et al. 2022):

$${T}_{f}=\frac{{V}_{t}}{{Q}_{t}}$$
(2)

The volume of the lagoon (Vt) was calculated using Eq. 3 (Lalau et al. 2022) from the area (A) of the lagoon defined as the reef crest to the shoreline, with the boundary at the Back Reef being the reef exposed during Lowest Astronomical Tide (LAT). A high-resolution bathymetry dataset (at MSL), from Dr. Robin Beaman (Hamylton et al. 2015), provided the depth (HMSL), while eReefs GBR1 provided the sea surface elevation at time t (ΔHt).

$${V}_{t}=A \times ({H}_{\text{MSL}}+\Delta {H}_{t})$$
(3)

To calculate transport (or volumetric flow), Qt, the width (W) and depth (h(x)) of the channels of inflow and outflow of the lagoon was calculated using the high-resolution bathymetry and the sea surface elevation at time t (Δht). The current velocity (Ct) measured by the Marotte HS current meters was then averaged per hour, and integrated over the cross-sectional area of each point of exit (Eq. 4; Lalau et al. 2022). For Qt, we used the volume of water leaving the lagoon to calculate the flushing time because current meters were better placed to capture the flow out of the lagoon.

$${Q}_{t}= {\int }_{0}^{W}\left(h\left(x\right)+\Delta {h}_{t}\right)\text{d}x \times {C}_{t}$$
(4)

Different wind and tide conditions were isolated and analysed for patterns in the circulation, along with different combinations of channel inflow and outflow to understand their influence on transport and flushing time. Current meters placed in the Lagoon, Loomis Reef Channel, Lagoon Entrance, and Palfrey-South areas were used as key locations to understand the potential influence of waves passing over the Bird-South crest. The main inflow over the Bird-South crest was calculated indirectly as the difference between the transport in and out of the lagoon, as per the laws of conservation (the flow that enters must equal the flow that leaves). The directionality of currents entering or leaving the lagoon were calculated using principal axes of flow.

Using the linear regression between wind velocity and current velocity for areas of inflow and outflow of the lagoon, transport and flushing time was hindcast for the period of the 2016 coral bleaching event, including the peak water temperature on 3rd March 2016 and April and May 2016 which saw ∼8 Degree Heating Weeks (DHW) (Bainbridge 2017).

Results

Wind and tide interactions

There were complex interactions between the tide and wind conditions at Lizard Island. Across both flood and ebb tide during Deployment 2, SE trade winds prevailed at approximately 27 km hr−1 (or ~ 15 kts). During the flood tide, water tended to flow through the Lagoon Entrance area and into the lagoon (Fig. 2a). This water flow occurred at both deeper (> 4 m) and shallower depths (2–4 m), suggesting an incoming flow through the entire water column at this location, despite directionality being topographically influenced (Fig. 2a). There was also a flow over the Bird-South crest, most likely in the form of wind-induced waves, as well as an along-slope flow in this area (Fig. 2a). Furthermore, there was an approximate NW flow across the Loomis Reef Channel, Lagoon, and Back Reef areas at shallow depths between 0 and 4 m. The pattern of flow at shallow depths is most likely wind-induced shallow currents due to the prevailing SE wind conditions (Fig. 2a), occurring approximately 78% of the time over the 2 years of deployment.

Fig. 2
figure 2

Averaged flood and averaged ebb tide current meter vectors for Deployment 2 indicating speed and direction of water flow: a during flood tide; b during ebb tide. Vector colours represent depth of the individual current meter (red = 0–2 m; orange = 2–4 m; pink =  > 4 m). The yellow vector represents the wind speed and direction taken from the IMOS Relay Pole 2 weather station, located on Bird Islet (yellow dot). The standard red vector (bottom right) represents 0.1 m s−1 of current velocity to the east. The brown colouration indicates shallow sand/reef areas. Bathymetry < 17 m is based on a 2-m resolution plot from Dr. R. Beaman (Hamylton et al. 2015), depths validated and refined following Philipps (2023); bathymetry > 17 m is based on a 30-m resolution (Beaman 2023). Legend represents depth (m MSL) from these two bathymetry datasets

Full size image

In contrast, when the tide was in ebb, water exited the lagoon through the Lagoon Entrance area in an easterly direction (Fig. 2b). During the ebb tide, there may have been some water exiting the lagoon through a narrow shallow channel in the Palfrey-South area which shows a stronger flow and hence larger vector magnitude in the ebb tide (Fig. 2b) than in the flood tide (Fig. 2a). Once again, a NW flow was present across the Loomis Reef Channel, Lagoon and Back Reef at shallow depths of 0–4 m (Fig. 2b). Similar to the flood tide (Fig. 2a), this circulation across the western part of the lagoon was likely due to wind-induced shallow currents.

There was a strong positive correlation (r > 0.7) between wind speed and current speed across all deployments at locations including Loomis Reef Channel, Lagoon, and a few sites at Bird-South Slope and Bird-South Back Reef. The currents at shallow locations (< 4 m) around the lagoon that were exposed to the prevailing SE trade winds (i.e. not sheltered behind reef structures) were generally predominantly wind driven at wind speeds above approximately 15 km hr−1 (Fig. 3a). Currents at sites that were sheltered from the SE winds (i.e. behind reef structures) and in deeper locations (> 4 m) also increased in velocity at wind speeds above approximately 20 km hr−1 (Fig. 3b). The difference between shallow and deep sites is most marked in changes to the maximum and minimum currents (Fig. 3).

Fig. 3
figure 3

Scatter plot between wind speed (km hr−1) and current speed during Deployment 2 for: a all shallow and exposed (to wind) current meters and b all deeper and sheltered (from wind) current meters. Red vertical lines indicate the wind speed at which the current speed appears to become increasingly wind driven. Colours represent tidal range and indicate a limited effect of tide level on current speeds

Full size image

Tidal influences

The tides around Lizard Island were well characterised by the IMOS sea surface elevation measurements near Bird Islet (ESM Table S3). The tidal analysis accounted for 96.6% of the variability in sea surface elevation data. The main tidal constituents evident for the Lizard Island group were the M2 (principal lunar) and S2 (principal solar) semidiurnal components with an amplitude of 0.5034 m and 0.3116 m, and phase of 159.07° and 15.06°, respectively (ESM Table S3). The main diurnal tidal constituents were K1 (lunar-solar declinational) and O1 (principal lunar) with an amplitude of 0.3351 m and 0.1753 m, and phase of 135.81° and 83.64°, respectively (ESM Table S3). However, the tidal analyses of the data from the current meters within the lagoon were not always congruent with the analysis from the IMOS data due to the confounding effects of the complex reef structure. Therefore, it was likely that the tidal flow within the lagoon was also represented by these primary tidal constituents extracted from the IMOS sea surface elevation data.

In general, there were similar trends in the tidal signal in the Lizard Island lagoon across the three deployments (Fig. 4; ESM Fig. S8). For Deployment 1, the tidal current at Bird-South Slope ran along the slope with a maximum of ~ 8% of the 0.12—0.2 m s−1 current speed attributed to the tide in all deployments (Fig. 4). Approximately 40% of total current speed for Bird-South Slope was attributed to the tide in Deployment 1 (Fig. 5). In the lagoon, there was a strong tidal flow in and out of the lagoon at Lagoon Entrance sites with relatively fast speeds of 0.12—0.2 m s−1 for the inflow (Fig. 4). This was also shown in a higher percentage tidal contribution of ~ 40–60% across all deployments for the mid-channel Lagoon Entrance sites (Fig. 5). Lagoon Entrance current meters with a lower tidal contribution were shallower and/or in close proximity to the reef and, therefore, more likely to be influenced by the wind and topography.

Fig. 4
figure 4

Current roses of the predicted tidal signal from select current meters from Deployment 1. The direction of spokes represents the direction of tidal water flow and spoke length represents the frequency in that direction. Spokes are coloured according to the speed of predicted tidal currents. Dark grey shading indicates islands within the island group (Lizard, Palfrey, South, and Bird Island; see Fig. 1), while the lighter grey shading represents areas of coral reef. The visualised current meters were chosen based on their greater percentage variability in Utide and Vtide components explained by the tidal analysis from each location around the lagoon (cut off = 20%). Deployment 2 and 3 analyses in ESM (Fig. S8)

Full size image
Fig. 5
figure 5

Percentage of current speed attributed to predicted tide speed from tidal analysis for all deployments. Each point represents a current meter and colour indicates the estimated percentage tidal contribution, 0–70% shown out of a possible 100%. Note the preponderance of high tidal values in the Lagoon Entrance, Deep Channel, and windward Bird-South Slope. Dark grey shading indicates islands within the island group (Lizard, Palfrey, South, and Bird Island; see Fig. 1), while the lighter grey shading represents areas of coral reef. Table of values in Supplementary (ESM Table S4)

Full size image

Further into the lagoon, the tidal flow visibly follows the topography of the lagoon, heading into/out of the Deep Channel area. This was best quantified by the deeper current meters of Deployment 3 (ESM Fig. S8b), where there was approximately 30–40% predicted tidal contribution to the total current speed (Fig. 5). The channel tidal current continues to follow the topographical features, heading through the Loomis Reef Channel (Fig. 4). Tidal currents in this area were best quantified during Deployment 3 (Fig. 5) where 30% of the current speed was attributed to the predicted tide (Fig. 5), in comparison with Deployments 1 and 2 which were predicted to contribute 0–20% (Fig. 5).

Another lagoon inflow and outflow may be present through a shallow channel over the Palfrey-South crest near to Palfrey Island. This was evident through the north-east/north-west to south-west directionality of the current roses at Palfrey-South (Fig. 4). However, there was a great range in percentage contribution of predicted tide to the total current speed across the current meters deployed in the Palfrey-South area (range: ~ 10–30%; Fig. 5). Outside of the lagoon on the western side in the Back Reef, there was an approximate north–south tidal flow, especially in the northern area of this zone across all three deployments (Fig. 4).

Low-frequency wind influences

Currents in shallow water (2–4 m) at the reef interface were largely wind driven, especially during a strong SE wind regime (wind strength greater than 40 km hr−1) (Fig. 6a). These currents had a higher magnitude than deeper currents at 4–6 m and > 6 m (approximate 2 – 3-fold difference in magnitude) (Fig. 6a). This was indicative of reduced influence from wind-shear stress with depth. In shallow waters, the direction of the currents largely followed the direction of the wind; however, there was some topographically driven variability (Fig. 6a). At depth in the Deep Channel, there was evidence of a consistent deeper outflow from the lagoon towards the Lagoon Entrance (Fig. 6a). Furthermore, it is likely that there is another outflow through the shallow, narrow passage in Palfrey-South (Fig. 6a). During these strong SE wind conditions, wind-driven currents at 2–4 m depth are extremely strong (e.g. 0.13 m s−1 to 0.31 m s−1) through Loomis Reef Channel, a key outflow point (Fig. 6a). At Bird-South Slope, there is a significant alongshore flow at 2–4 m depth.

Fig. 6
figure 6

Visualisation of low-frequency lagoonal currents in Deployment 3 during: a the strong prevailing SE wind period (greater than 40 km hr−1) and; b a NW wind period (commonly seen during the monsoon season; greatest wind strength of approximately 28 km hr−1). Vectors represent current speed and direction for individual current meters. Colours represent current meter depth: orange 2–4 m, pink 4–6 m and purple > 6 m. Yellow vector represents the wind speed and direction at the IMOS Relay Pole 2 on Bird Islet (yellow dot). The standard red vector (bottom right) represents 0.1 m s.−1 current velocity. The brown colouration indicates shallow sand/reef areas. Bathymetry < 17 m based on plot from Dr. R. Beaman (Hamylton et al. 2015); bathymetry > 17 m follow (Beaman 2023)

Full size image

The circulation pattern under NW wind conditions varied markedly from the circulation under SE winds. The NW wind conditions occur intermittently throughout the Austral monsoon season (e.g. 28 km hr−1, a strong NW wind strength, Fig. 6b). Similar to those in the SE wind regime, shallow currents at the reef interface (2–4 m depth) are also wind driven. As such, during NW winds, the direction of the currents in the Back Reef and Loomis Reef Channel areas flowed eastwards (Fig. 6b), as opposed to westward or north westwards under SE winds (Fig. 6a). This pattern continues in areas including Bird-South Slope and Palfrey-South (Fig. 6b). There was an evident reduction in current vector magnitude across the Lizard Island lagoon under NW wind conditions, in comparison with that of the strong SE wind conditions, likely due to the reduction in wind strength (Fig. 6). Weak SE wind conditions are a combination of the two, with a SE direction but weak current strength (ESM Fig. S9).

Flushing time and the role of waves

During the typical SE wind conditions (~ 25 km hr−1) which occurred about 46% of the time during our deployments, there was similar transport of water into the lagoon on spring and neap tides (Table 1). However, there was a large difference between the recorded inflow and outflow transport (Table 1). Specifically, the outflow was five–six-fold greater than the inflow during the 25 km hr−1 SE wind conditions. The discrepancy presumably being driven by wind-forced waves over the SE facing reef crests between Bird-South and South-Palfrey. The flushing time for these typical conditions was approximately 1–2 days (Fig. 7; Table 1).

Table 1 Inflow and outflow transport, and flushing time calculated across SE and NW wind conditions and different tidal cycles. Inflow and outflow of no wind period using only the Lagoon Entrance and Loomis Reef Channel as areas of measurement
Full size table
Fig. 7
figure 7

Daily averaged flushing time (in days) for all deployments as a variable of: a wind direction (degrees from North) and; b wind speed (km hr−1). Flushing times represented are only those under 10 days. Full figures with outliers included in ESM (Fig. S10). Red symbols represent flushing times during the 2016 coral bleaching event, including the hottest day (red circle) (3rd March 2016; Bainbridge 2017) and the average of the ~ 8 DHW (red triangle) period

Full size image

During stronger SE wind conditions (> 40 km hr−1), there was again similar transport between the spring and neap tides (Table 1). However, there was even greater disparity between the recorded transport in and out of the lagoon, with outflows approximately 50–100-fold higher than inflows (Table 1). This supports the notion that the strong SE conditions, which occurred approximately 12% of the time, represent an extreme of the typical SE conditions. The flushing time (at approximately 12 h) was also reduced during the strong SE conditions, indicating that water renewal times significantly decrease with increase in wind speed (Fig. 7b).

Transport in and out of the lagoon was nearly identical during a neap tidal cycle during NW wind conditions (Table 1), which occurred 18% of the time during our deployments. This indicates that the flow over the Bird-South crest may not be as significant during NW wind conditions compared to during the SE wind regime. The flushing time of the lagoon dramatically increased with the change in wind direction (NW) to approximately 5 days over both tidal cycles (Fig. 7; Table 1). Periods of no wind were rare during our deployments, only occurring approximately 4% of the time. The inflow and outflow transport were equal during periods of no wind when investigating flow solely through the Lagoon Entrance and Loomis Reef (Table 1), suggesting that tidal flow (i.e. no wind influence) predominantly occurs through these two channels. There is also limited wave input during this wind regime.

Coral bleaching conditions

The 3rd of March 2016 was the hottest day of the 2016 coral bleaching period (Bainbridge 2017). However, it had an estimated flushing time of approximately 2.9 days which is within the normal range during SE to NW periods (Table 1; ESM Table S6), suggesting that physical drivers, wind and currents, were within normal limits. This is consistent throughout the remainder of March 2016, with the transport flow and flushing time being similar to other values calculated for the duration of our study (Table 1; Fig. 7; ESM Fig. S11). For the April and May period, that saw ~ 8 DHW, the flushing time was 0.93 days, which was less than the period leading to the time of the DHW (Table 1; ESM Fig. S11). This suggests that the lagoon was being flushed at a rate similar to that during typical SE wind conditions (Fig. 7), indicating that it is unlikely that reduced circulation of Lizard Island lagoon was a major contributing factor to the extensive coral bleaching seen during this time.

Discussion

We characterised water movement around the reef matrix of the Lizard Island lagoon. Within the lagoon, the movement of water in the shallows, especially at the reef interface, was largely wind driven and was highly correlated with the wind speed. In deeper water, while currents are still primarily wind driven, tidal influences are more dominant.

Wind and wave influences

Coral reef organisms are strongly influenced by hydrodynamics at the reef interface. In the top 4 m, wind-influenced currents are dominant across the Lizard Island lagoon. This was previously noted by Johansen (2014) on exposed reefs at Lizard Island, who suggest that there was an exponential relationship between wind strength and current speed. SE winds occurred for 78% of the time across deployments; these are the prevailing conditions at Lizard Island (Tebbett et al. 2023). The depth of wind influences on the currents at Lizard Island is also similar to the wind driven currents of Cleveland Bay, Townsville, Australia, which were largely restricted to the top 5 m (Larcombe et al. 2001). At our study sites, shallow (< 4 m) locations that are exposed to SE winds (i.e. seaward facing reef structures) are dominated by wind driven currents at wind speeds above approximately 15 km hr−1. At these wind speeds, it is likely that wind-driven circulation dominates over tidal and other physical drivers influencing the circulation within the Lizard Island lagoon. At deeper depths and in sites sheltered from the prevailing SE trade winds, wind-driven circulation is less important and tidal influences may predominate.

During periods of SE wind, increased speed is likely to be a significant driver of circulation and flushing times in the Lizard Island lagoon. Inflow volumes into the lagoon were up to 100 times less than outflow volumes. This difference occurred in both spring and neap tidal cycles, and was greater under stronger SE wind speeds. This difference is likely due to the input of waves over the Bird-South crest, and to a lesser extent the South-Palfrey crest. There was a pattern of decreased flushing time with increased wave energy. This is again likely driven by the increased wind speed from the SE, leading to an increase in significant wave height breaking and flowing over the crest between Bird and South Islands and, to a lesser extent the shorter South-Palfrey crest.

This importance of wind-driven wave transport and its influence on flushing time has been observed across multiple coral reef lagoons and atolls. Lalau et al. (2022) found that the flushing time of the Poé-Gouaro-Déva lagoon, New Caledonia, ranged from 4 to 29.5 h, with periods of high offshore waves caused faster flushing times. At Ningaloo Reef, Western Australia, flushing time also depended upon waves; from 3 h with 2 m waves to more than a day with smaller waves (Taebi et al. 2011). Likewise, the flushing time across atolls in the Tuamotu Archipelago in French Polynesia varied widely with atoll morphology, ranging from 0.6 to 600 days for a significant wave height of 1.5 m (Andréfouët et al. 2001). Despite the great variation in flushing times, flushing times consistently decreased with increased significant wave height (Andréfouët et al. 2022).

Flushing time is biologically important. Flushing links directly to residence time and is a critical component of temperature and nutrient exchange, which can have significant biological implications, for example, during heatwaves leading to coral bleaching (Shredrawi et al. 2017), the assimilation of pelagic subsidies (Valenzuela et al. 2021), and cyanobacterial mat formation (Tebbett et al. 2022b). As such, our findings from the hindcast 2016 coral bleaching event found flushing time was unlikely a major contributor to the mass bleaching in the Lizard Island as times were within the normal range (0.93 days for ~ 8 DHW). Given the biological importance of lagoonal flushing, this is a key area of future research for the Lizard Island lagoon.

Tidal influences

Tidal influences were observed mostly at depths > 4 m within the lagoon as well as in areas further from reef structures. There are, however, limitations to the analyses herein, related to the highly complex reef structure and bathymetry of the lagoon. As the tide moved through the lagoon channels and over reefs, it would have been affected by resonance (Tee 1976; Tarya et al. 2010), refraction (Larcombe et al. 2001), and diffraction. All create hydrodynamic noise within the reef system that is recorded by the current meters. It is therefore difficult to accurately predict the total tidal influence on the system. The filtering in the tidal analysis should result in an isolated tidal signal, and a residual around the tidal signal that represents the high-frequency wind effect, as well as low-frequency trends. However, due to the distortion of the tidal signal by the reef matrix, a tidal signature still probably remained within the residual, so wind effects could not be extracted cleanly from the residual. Therefore, it is likely that we were unable to produce a completely reliable quantitative measure of the tidal influence on the circulation and the values may underpredict the true tidal circulation. Johansen (2014) suggested that the tidal component of circulation external to the Lizard Island lagoon contributed 3.3 ± 0.4 cm s−1. Using the mean of predicted tidal contribution to current speed (of 20.4%) from our data, compared to the average current velocity in the Lagoon Entrance and Loomis Reef Channel of approximately 7 cm s−1, a tidal speed contribution can be calculated to be approximately 1.4 cm s−1. This result is lower than that of Johansen (2014), who measured tidal influences in the presence of no wind (i.e. assuming the sole influence was tidal). This again suggests that the tidal analyses herein are likely to underestimate the actual tidal contribution.

Maximum current speeds found at North Scott Atoll, Western Australia, reached 2.0 m s−1. These currents recorded in the channel outflow were tidally influenced (Green et al. 2018) and far exceed those recorded at Lizard Island. However, at the Rowley Shoals, in Western Australia, Mermaid Reef had current velocity of a comparable order of magnitude to our observations at Lizard Island, with a range of 0—0.3 m s−1, again, most likely occurring due to the combination of wave and tidal processes (Grimaldi et al. 2022).

The tidal analysis herein did identify a clear tidal signal through the Lagoon Entrance channel indicating an area of inflow/outflow for the lagoon, and other, less clear tidal inflow/outflow areas at Loomis Reef Channel and Palfrey-South. During periods of no wind at Lizard Island, tidal flushing through these channels is likely to be the sole driver of water in and out of the lagoon. This was also found in New Caledonia, in that wind and tide are important drivers during no wave periods (Lalau et al. 2022).

Barotropic current

A barotropic current at Palfrey-South could also be a key contributor to water movement. During periods of strong SE wind (wind speed greater than 40 km hr−1), the current at Palfrey-South displayed a direction ‘out’ of the lagoon, in a SW direction. A barotropic pressure gradient may thus be forcing water out of the lagoon during strong wind conditions. Strong wind conditions such as those seen during the prevailing SE trade wind season may cause a build-up of water within the Lizard Island lagoon from waves and wind-driven currents, creating a pressure gradient between the inside and outside of the lagoon (cf. Hench et al. 2008), causing the directionality of water flow near perpendicular to the wind direction.

Ecological implications

The results from this study may help explain ecological observations in previous research conducted within the Lizard Island lagoon. We provide a broader context for cross-reef observations, such as those by Tebbett et al. (2022b) who discovered that the locations with the highest cyanobacterial mat cover occurred at the Back Reef which had some of the lowest current speeds across all locations and environmental conditions in our study. This further supports the findings of Tebbett et al. (2022b) of low current speeds being associated with high microbial growth. Our data also cast light on the work of Collins et al. (2022) who found that apogonid reef fish productivity was high in sites around the Deep Channel, Lagoon, and Loomis Reef Channel. It is likely that these areas are important locations for water movement bringing in a fresh influx of plankton and nutrients into the Lizard Island lagoon, while simultaneously protecting the fishes from the direct influence of strong SE wind driven currents. In a similar vein, Valenzuela et al. (2021) found that larger planktivorous fish inhabited the Bird-South Slope and Back Reef which are areas likely to be more exposed to supply currents. The Bird-South Slope, in particular, is directly exposed to the prevailing SE winds likely driving the surface currents at the reef interface, bringing plankton in from off-reef locations. The Back Reef currents typically flow in a northerly direction and hence it is likely that the currents travelling NW along the reef slope at Palfrey-South form an ‘island wake’ or eddy (Wolanski et al. 1984) to the lee of Palfrey Island during SE wind conditions, again bringing supply currents to the area.

The circulation of the Lizard Island lagoon may also influence coral dynamics. For example, in the erosion of dead coral colonies after bleaching. Morais et al. (2022) found that on the Bird-South Slope and in Lagoon/Loomis Reef Channel locations, 50% of dead colonies were lost after 31–33 months. These higher current speed areas contrast with the sheltered Back Reef, where 50% loss of dead colonies took 45.7 months. The differing current speeds may help explain the difference in time taken to erode the dead coral colonies. Tebbett et al. (2022a) found similar location-specific patterns in living corals with higher recruitment of Acropora corals in semi-exposed locations around the Lizard Island group. In this case, the Palfrey-South area was particularly interesting, with a mean density of 17.7 individual recruits per m2 far exceeding the typical 0.6–5.2 recruits per m2 elsewhere. At this location, there is a SW outflow through the shallow channel that is likely to meet a dominant NW flow along the reef slope. The difference in current direction of these two flows may create eddy-like behaviour that ‘traps’ Acropora larvae in the area, providing a potential explanation for the increased settlement and recruitment of corals in this area. At this same location, Morais et al. (2023a) also found that coral growth was some of the highest at Lizard Island lagoon, at approximately 150 cm yr−1 growth. This was in comparison with less than 100 cm yr−1 growth in all other locations around the Lizard Island lagoon. The circulation in this area may not only provide a suitable coral recruitment environment, but also a habitat that supports exceptional coral growth.

Our study may also be able to suggest possible areas of refuge for large reef fishes as Kerry and Bellwood (2017) identified wind speed as a potential factor driving fishes to seek refuge under tabular structures; reducing their energy demands. If wind and wind-driven currents are driving fishes to seek refuge, our results can indicate areas around the Lizard Island lagoon where this behaviour is most likely to take place (cf. Johansen et al. 2007).

Despite the transport (or volumetric) flow and flushing time being a limited driver in the 2016 coral bleaching event at Lizard Island (ESM Fig. S10), the general circulation patterns could have been influential during this event. Morais et al. (2023b) quantified the percentage of coral bleaching after the 2016 event and found that Lagoon Entrance areas, Bird-South Back Reef, Loomis Reef Channel and particular Lagoon areas experienced 100% bleaching. This was in comparison with Pigeon Point, Lizard Head, Bird-South Slope, and the Back Reef that experienced bleaching rates of 53–71% (Morais et al. 2023b). Pigeon Point, Lizard Head, and Bird-South Slope generally receive supply currents due to the SE trade winds that may aid in reducing the water temperature. Similarly, in the Back Reef, there is the potential for supply currents bringing cooler water as currents travel in a northerly direction likely due to the island wake caused by Palfrey Island (cf. Wolanski et al. 1984). By contrast, current velocities are typically slower in the ‘front lagoon’ (e.g. Bird-South Back Reef and Lagoon Entrance), essentially allowing corals to sit in the warm water; likely contributing to the 100% coral bleaching in these areas (Morais et al. 2023b). Furthermore, Loomis Reef Channel is a key area of outflow from the Lizard Island lagoon during the SE wind regime and also experienced 100% coral bleaching during the 2016 bleaching event (Morais et al. 2023b). It is likely that this area was continually subjected to the warm water outflow from the lagoon during the critical ~ 8 DHW warming period.

Overall, like most reef systems, Lizard Island has a highly complex hydrodynamic regime, dominantly influenced by tide and wind. The tidal influence is most strongly expressed in deeper locations around the reef and in areas further away from reef structures. Wind is the predominant influence on current speeds, particularly in locations less than 4 m deep, where the majority of marine organisms such as reef fishes and corals reside. Current direction, however, is highly dependent upon the local topography. It must be also noted that these influences are not independent of each other as wind and tide invariably come together to drive the hydrodynamics of the Lizard Island lagoon. Observing and describing the influence of these two processes in shaping the circulation of Lizard Island may provide a valuable first step in bridging the gap between hydrodynamic and ecological knowledge at Lizard Island.