Abstract
The Baltic Sea observing system combines ship-based, autonomous, and remote-sensing components, yet major observational gaps persist in the shallow and dynamic coastal zone. Here, we present the first evaluation of a micro-autonomous underwater vehicle (micro-AUV) for high-resolution mapping of the Baltic Sea water column and assess its potential integration into regional coastal observing systems. Eleven missions were conducted in 2025 across three contrasting environments—the shallow estuarine Bay of Matsalu, the Suur Strait, and the coastal slope of the Baltic Proper—covering depths from 1 to over 100 m. The micro-AUV operated reliably under diverse hydrographic conditions, capturing fine-scale frontal, submesoscale, and benthic-layer structures, as well as a dual upwelling cell. Comparisons with reference profiles confirmed the high accuracy of temperature and salinity measurements, and strong correlations for chlorophyll-a fluorescence and turbidity. The vehicle’s endurance (up to 40 km or ~7 h at 1.5 m s−1) enables efficient mesoscale surveys, while its maneuverability allows operation in very shallow or complex coastal areas where gliders and floats cannot be used. The results demonstrate that micro-AUVs can fill a critical observational gap between ship-based surveys, gliders, FerryBoxes, and coastal stations by extending coastal observations offshore and resolving vertical structures inaccessible to remote sensing. Despite limitations in endurance and communication range, micro-AUVs represent a promising, cost-efficient addition to integrated coastal observing systems, supporting targeted process studies and rapid-response missions in dynamic environments such as the Baltic Sea.
1 Introduction
Ocean observing systems have developed considerably in recent decades. The conventional research vessel surveys have been increasingly complemented by autonomous observations. As a result, the majority of the Global Ocean Observing System now relies on observations from unmanned platforms (Tanhua et al., 2019). The coastal ocean plays a crucial role in marine ecosystems and human society. It is the zone where land–sea interactions are most intense, and where strong physical, chemical, and biological gradients shape the marine environment. Coastal waters are strongly influenced by riverine inputs, atmospheric forcing, and human activities. Measuring this complex ecosystem is particularly challenging because of its high spatial and temporal variability, which requires dedicated and adaptive observational approaches. Integrated initiatives such as JERICO () have been established to address these challenges by developing sustained, harmonized coastal observing systems that combine fixed, drifting, and mobile autonomous platforms to improve the understanding and monitoring of coastal dynamics. In the present study, we test the integration potential of a new micro-AUV platform within such coastal observing systems, using the Baltic Sea as an example.
The Baltic Sea exemplifies these challenges: as an inland, shallow sea with a large catchment area, it is strongly affected by freshwater inflows, nutrient loads, and limited water exchange with the North Sea, all of which amplify its natural variability and sensitivity to human pressures (; Reckermann et al., 2022; ). Baltic countries have conducted coordinated ship-based monitoring for decades under HELCOM (Helsinki Commission), producing long-term time series at selected stations. In contrast, observations from autonomous platforms in the Baltic Sea have not been coordinated and have largely evolved according to the scientific interests and needs of individual research groups.
The variability of water column properties in the Baltic Sea occurs at various spatial and temporal scales. For example, conditions in the deep layer of the Central Baltic are relatively stable in synoptic and seasonal timescales (; Rolff et al., 2022; Stockmayer and Lehmann, 2023), whereas higher variability is found in the smaller sub-basins (Schmidt et al., 2021; ; ), straits (; Muchowski et al., 2023; Purkiani et al., 2024), river mouths (Ylöstalo et al., 2016; ; Pein and Staneva, 2024), and along coastal slopes where mesoscale processes are more active (; ; ).
Capturing this range of variability requires high-resolution and flexible observation methods that can resolve processes often missed by traditional monitoring. Flow-through systems, commonly known as FerryBoxes, have become an important component of the Baltic Sea observing system (Petersen, 2014; ; ). Permanent and temporary moorings, equipped with fixed sensors or profiling devices, have significantly advanced the understanding of temporal variability in the region (Prien and Schulz-Bull, 2016; Mohrholz, 2018; ; ). Argo floats have been used to complement HELCOM monitoring in offshore areas (; Walczowski et al., 2020; ). In addition, underwater gliders have been deployed in the Baltic Sea for more than a decade, primarily for process studies and targeted case investigations (; ; ; Salm et al., 2023). These in situ observations are complemented by remote sensing, which provides a synoptic spatial view of surface-layer properties (She et al., 2007; Ostrowska et al., 2022; Soomets et al., 2022).
The use of the above-mentioned observing platforms is limited in coastal zones, river mouths, and straits. Argo floats would quickly strand in shallow areas, and underwater gliders cannot be operated at depths shallower than ~30 m — approximately one-third of the Baltic Sea’s area (). Measurements from research vessels and small boats are restricted in both spatial and temporal resolution. FerryBoxes provide high-resolution data along ship tracks, but these tracks are fixed to commercial routes and only cover the upper layer. Similarly, remote sensing is limited to the surface layer, and its data availability is further constrained by cloud cover. Moorings offer high-resolution temporal records but lack spatial coverage, particularly in dynamic coastal zones. Consequently, a gap remains in the current observing system for the coastal zone from the shoreline to ~30 m depth. In this study, we present results from missions with a micro-AUV, a platform capable of collecting high-resolution observations from shallow near-coastal areas (a few meters deep) down to 300 m.
The main aim of the present study is to test the applicability of a micro-AUV in a brackish environment with complex topography. Specifically, our goals are to:
- Assess the user experience of operating a micro-AUV in areas with strong gradients in the eastern Baltic Sea;
- Present the characteristics of the micro-AUV and evaluate its potential role in the Baltic Sea observing system;
- Identify use cases where the micro-AUV platform offers the greatest efficiency and added value.
To achieve these goals, we carried out 11 AUV missions in three contrasting areas: the shallow estuarine Bay of Matsalu, the Suur Strait connecting the Gulf of Riga and the Moonsund (Väinameri), and the coastal slope of the northeastern Baltic Proper. While the missions revealed interesting hydrographic features, the primary focus of this study is not on oceanographic interpretation. Instead, these observations serve to demonstrate the applicability and potential of micro-AUVs as platforms for high-resolution measurements in the Baltic Sea and other estuarine and coastal environments. To our knowledge, this is the first study to apply a micro-AUV to investigate the vertical structure of the Baltic Sea water column and to assess its suitability for integration within the regional marine observing system.
2 Data and methods
2.1 Description of the used equipment
The YUCO-PHYSICO micro-AUV (Figure 1) is an autonomous underwater vehicle built by the French company Seaber, designed to measure parameters of the marine environment. The device is lightweight (9.5 kg in air) and compact (123 cm long, 12 cm in diameter), making it easy to deploy from a boat or directly from the shore. It can operate autonomously throughout the water column at speeds of 1.5–3 m s-1 and dive to a depth of 300 m.
Figure 1
The device runs on rechargeable lithium-ion batteries and uses an electric motor with a propeller for propulsion. Steering is achieved via movable fins located in the tail section. It is equipped with a Doppler velocity logger (Waterlinked A50 DVL), which enables it to maintain a fixed height above the seabed (up to 50 m) and to compensate for the effects of currents on its movement, as well as with pressure sensor that allows it to maintain a fixed depth.
The AUV carries a multiparameter probe (AML−RT3, manufactured by AML Oceanographic) mounted in the nose section, to which up to three separate sensors (i.e. temperature-conductivity plus two additional sensors) can be connected. The sensors used in the present study include: conductivity (XCH2-CT STD with accuracy 0.01 mS/cm, range 0–90 mS/cm and response time 25 ms), temperature (XCH2-CT STD with accuracy 0.005 °C, range -5 to 45°С and response time 100 ms), turbidity (XCH2-TRB-TUR-06 with accuracy 2% reading or 0.2 NTU, range 0–1500 NTU and response time 200 ms), chlorophyll-a with blue excitation sensor (XCH2-CHL-ABTUR-06 with linear accuracy R2 = 0.99, range 0-500 μg/L and response time 200 ms) (AML ).
The AUV’s missions are configured using Seaplan software. During operation, communication with the device is maintained via the SEACOMM remote control, which operates over Ultra High Frequency (UHF) radio. SEACOMM enables the operator to request the AUV’s location coordinates, read system messages, interrupt the current mission, or command the vehicle to return to the SEACOMM position. Under typical conditions, reliable communication can be maintained within a range of approximately 1–2 km, depending on sea state, antenna height, and environmental conditions.
The AUV supports multiple operational modes, including constant-altitude and constant-depth navigation, saw-tooth and undulating trajectories between predefined depth and altitude limits, helical ascent or descent, and free drifting within the water column or at the surface. In addition to mission trajectory definitions, the user specifies additional safety parameters, namely a maximum allowable depth and a minimum allowable altitude above the seabed, which the AUV continuously attempts to maintain during the mission.
If either the maximum depth or the minimum altitude threshold is exceeded, the AUV automatically aborts the mission and surfaces as a safety measure. The operational risk increases when missions are designed to run close to these predefined limits, particularly in areas with complex or rapidly changing bathymetry. In such environments, risks can be minimized by applying larger safety margins, for example by operating at higher altitudes above the seabed.
In the present study, the AUV was intentionally operated with very small altitude thresholds (sometimes only 0.5–1 m above the seabed) in order to assess the performance limits and operational capabilities of the platform in shallow-water conditions. Under these conservative settings, a single bottom contact occurred when the vehicle descended in an area where the seabed depth changed abruptly.
Other operational risks, such as collision with surface vessels or entanglement in fishing nets, were considered during mission planning. In Matsalu Bay, where a significant number of fishing nets are present, the operational area was visually inspected from a motorboat prior to deployment. Areas with intense shipping activity were generally avoided; however, the fairway between the mainland and Muhu Island, used by 1–2 ferries per hour, was crossed during certain missions.
To minimize collision risk in this area, the AUV was programmed to cross the fairway at sufficient depth using a fixed-altitude operating mode, thereby remaining well below vessel drafts and associated disturbance. These operational measures were effective in reducing the risk associated with anthropogenic obstacles and vessel traffic during the reported missions.
Raw AUV data were first converted to time-series format and quality-controlled to remove invalid values. Derived variables and positional information were then used to construct both vertical sections along predefined transects and horizontal maps of surface properties.
For the vertical sections, observations were projected onto the transect line and interpolated using linear interpolation within a predefined lateral corridor (typically 100–200 m wide) around the section track. The data were subsequently gridded at regular horizontal and vertical intervals (horizontal spacing on the order of tens of meters and vertical resolution of approximately 0.5 m) to produce continuous fields of temperature, salinity, chlorophyll-a, and turbidity.
For comparison, vertical CTD profiles were recorded using two CTD sonde SBE 19plus V2 SeaCAT Profiler (Seabird Scientific, 2025b). One SBE sonde was brand-new and factory-calibrated, while the other unit was older, but regularly factory calibrated. The device has a sampling rate of 4 Hz and uses pump-controlled flow, which minimizes salinity spiking, improves sensor response time, and enhances data quality during profiling. The probe has integrated pressure sensor (range: 0–350 dbar; accuracy: ± 0.1% of full scale), conductivity sensor (range: 0–9 S/m; accuracy: ± 0.0005 S/m), and temperature sensor (range: –5 to +35 °C; accuracy: ± 0.005°) (Seabird Scientific, 2025b).
The SBE19plus V2 SeaCats were equipped with two auxiliary sensors: an ECO FLNTURT optical sensor, which measures chlorophyll−a fluorescence (sensitivity: 0.025 µg/L; range: 0–50 µg/L) and turbidity (sensitivity: 0.013 NTU; range: 0–100 NTU) (Seabird Scientific, 2025a), and SBE43 dissolved oxygen sensor (initial accuracy: ± 2% of saturation; range: up to 120% of surface saturation; typical stability: 0.5% per 1000 h of deployment; response time: 2–5 s with a 0.5−mil membrane) (Seabird Scientific, 2025c).
All water temperature and salinity values are expressed as Conservative Temperature and Absolute Salinity, respectively. Wind speed, wind direction, air temperature, and sea-surface temperature data from coastal stations were provided by the Estonian Environmental Agency. Water temperature was measured at Virtsu, Haapsalu, and Dirhami; air temperature at Haapsalu and Dirhami; and wind speed and direction at Heltermaa and Osmussaar (Figure 2). All bathymetry data used in this study originate from the EMODnet Bathymetry Digital Terrain Model (DTM) ().
Figure 2
2.2 Description of surveys conducted
The surveys were conducted in three areas: Matsalu Bay within the Moonsund system, the Suur Strait, and the coastal slope of the northeastern Baltic Proper (Figure 2, Table 1). The first five missions were conducted in April, missions 6–10 in July, and the final mission on 1 October 2025.
Table 1
| Mission | Date | Area | Distance covered (km) | Max depth (m) | Pattern |
|---|---|---|---|---|---|
| 1 | 14.04.2025 | Matsalu | 23 | 1.8 | Fixed depth |
| 2 | 15.04.2025 | Matsalu | 27 | 4.5 | Fixed depth, sawtooth |
| 3 | 16.04.2025 | Matsalu | 31 | 2.2 | Fixed depth |
| 4 | 17.04.2025 | Suur Strait | 12 | 21.7 | Fixed depth, sawtooth |
| 5 | 18.04.2025 | Matsalu | 31 | 2.9 | Fixed depth |
| 6 | 18.07.2025 | Suur Strait | 33 | 21.5 | Sawtooth, fixed depth, fixed altitude |
| 7 | 19.07.2025 | Suur Strait | 23 | 21.6 | Sawtooth, fixed depth, fixed altitude |
| 8 | 20.07.2025 | Suur Strait | 26 | 21.4 | Sawtooth, fixed depth, fixed altitude |
| 9 | 21.07.2025 | Suur Strait | 32 | 21.5 | Sawtooth, fixed depth, fixed altitude |
| 10 | 22.07.2025 | Suur Strait | 26 | 21.3 | Sawtooth, fixed depth, fixed altitude |
| 11 | 1.10.2025 | Baltic Proper | 33 | 100.5 | Sawtooth, fixed depth, helix, neutral drift from bottom layer to surface |
Summary of AUV missions conducted in 2025 in the northeastern Baltic Sea.
Matsalu Bay, part of the Moonsund system, is a shallow, brackish, and nutrient-rich basin stretching 18 km from west to east and 6 km from north to south. The average depth of this estuarine water body is about 1.5 m, with a maximum of 3.5 m (e.g., ). The main freshwater input to Matsalu Bay—and to the Moonsund more broadly—comes from the Kasari River, which enters at the easternmost part of the bay. Salinity in the inner bay ranges from 0 to 2 g kg−1, increasing to 4–6 g kg−1 in the western part (Vetemaa et al., 2006). Similar gradients are also expected in other physical and biogeochemical parameters. Matsalu Bay is part of a nature-protection area, designated mainly for bird conservation, and access is restricted during parts of the year to minimize disturbance.
The Moonsund is a shallow marine area connected to both the Baltic Proper and the Gulf of Riga through five straits. Water exchange through these straits, together with the alternation of different water masses and frequent sediment-resuspension events, makes the physical and biogeochemical properties of Moonsund waters highly variable in time (; Suursaar et al., 2009).
The Suur Strait connects the Moonsund with the Gulf of Riga. Hydrographic conditions in the strait are strongly controlled by the position of the frontal zone. When the current is directed northward, water from the Gulf of Riga fills the strait, whereas during prevailing southward currents, water masses from the Moonsund dominate (Otsmann et al., 1997; ).
2.2.1 Matsalu Bay
The first three missions, as well as the fifth, were conducted in Matsalu Bay. Mission 1 began in the western part of the bay, from which the AUV proceeded toward the inner bay and finished in the open waters of the Moonsund (Figure 3). Water depth ranged from 2.3–2.4 m in the easternmost part to 6–7 m in the west. Wave activity was minimal, and currents inside the bay were likely weak, allowing the AUV to follow its pre-programmed route with high accuracy. The vehicle was configured to operate at a depth of 0.8 m and a speed of 2.0 m s−1, both of which it successfully maintained. The mission was terminated after 90% of the battery capacity had been used.
Figure 3
Mission 2 started in the inner part of the bay, continued into the open Moonsund, and concluded at the bay mouth (Figure 3). The AUV was configured to operate at a speed of 1.8 m s−1 and a depth of 0.8 m for most of the mission. In the south–north-oriented section of the Moonsund, the depth control alternated between a fixed depth of 0.8 m and an altitude of 2.0 m above the bottom (sawtooth pattern), allowing vertical profiles to be collected along the route. A vertical speed of 0.1 m s−1 was applied during descents and ascents. The AUV maintained its position above the bottom within ±0.1 m of the altimeter reading; however, upon reaching the target altitude, it briefly overshot downward by 0.3–0.4 m due to inertia. Similarly, when ascending toward the 0.8 m depth setting, the vehicle surfaced slightly shallower than intended, by about 0.2–0.3 m.
Mission 3 followed the same initial route as Mission 2, but the transect in the open Moonsund was shortened and an additional zig-zag pattern was executed near the bay mouth. The AUV was configured to operate at a speed of 1.8 m s−1, maintaining a depth of 0.8 m inside the bay and 1.5 m in the open Moonsund.
Mission 5 was conducted in the mouth of Matsalu Bay to capture details of the frontal zone. The AUV was configured to operate at a speed of 1.8 m s−1 and a depth of 0.8 m. Water depth along the route was generally 2–4 m, with occasional shallows of 1.7–1.8 m. The mission was interrupted when a risk of collision with fishing nets was identified near the southern coast. After adjusting the plan, a new mission file was uploaded from the boat, and the survey continued until completion.
Overall, the Matsalu Bay missions demonstrated the reliability and precise navigation of the micro-AUV in a very shallow marine environment. The successful completion of multiple missions confirmed the vehicle’s suitability for detailed mapping of coastal gradients and frontal structures in estuarine settings.
2.2.2 Suur Strait
Mission 4, along with missions 6–10, was conducted in the Suur Strait area. All missions in this area were initiated near Kesselaid Island (Figure 2). Mission 4—the only April survey in the Suur Strait—aimed to map the vertical structure of the water column across and along the narrow, deeper ridge between Kesselaid and Muhu Islands (Figure 4). The mission began with a profiling transect across the ridge, followed by a southward leg and concluding with a northward profiling section. The AUV operated at a speed of 1.5 m s−1, except for the leg from the westernmost point of the cross-ridge section southward, where the speed was increased to 3.0 m s−1. Profiling was configured between 2 m above the seabed and 1 m below the surface. Vertical speeds were set to 0.25 m s−1 across the ridge and 0.20 m s−1 along the ridge. Due to inertia, the AUV occasionally reached altitudes of 1.0–1.3 m and 1.2–1.3 m, respectively. In one profile where bottom depth shoaled during descent, the vehicle approached as close as 0.5 m to the seabed. Similarly, during ascents the AUV briefly reached depths of 0.4–0.5 m.
Figure 4
Missions 6–10 investigated the Suur Strait south of Kesselaid (Figures 2, 4), covering an along-strait transect of 13.5–17 km and one to three cross-strait sections. Vertical profiling—alternating between fixed depth and fixed altitude—was applied, except during transits between sections, when fixed-depth mode was used. The along-strait transect crossed the mainland–Muhu Island fairway, which has regular ship traffic; to avoid collision risks, the AUV was operated in fixed-altitude mode over a 1.3–1.4 km stretch. The horizontal spacing between descents and ascents was 150–200 m. Vertical speed was set to 0.25 m s−1, with dives targeting 2 m above the seabed. Due to inertia, the vehicle often reached an altitude of 1 m and, when descent coincided with shoaling bottom topography, occasionally approached as close as 0.5 m. On one occasion, the AUV aborted the mission after contacting the bottom, but the mission was updated and successfully continued.
The AUV’s speed was generally 1.5 m s−1, while higher speeds were used for transits between sections. Another mission was aborted when the AUV failed to submerge within the programmed time window. Since the mission was almost over, mission was not continued. All missions were carried out during daylight hours, except for Mission 9, which concluded around midnight. Thanks to the AUV’s lights, the recovery procedure at night proved as straightforward as during the day.
Overall, the Suur Strait missions demonstrated the capability of the micro-AUV to perform repeated vertical profiling and long transects in areas with variable topography. The vehicle maintained stable navigation, precise depth and altitude control, and reliable data acquisition throughout the missions.
2.2.3 Baltic Proper
Mission 11 was conducted in the northeastern Baltic Proper, near the mouth of the Gulf of Finland (Figure 5). The transect extended from the offshore area, with depths reaching 103 m, to the shallow coastal zone of about 3 m. The primary objective was to test the performance of the AUV in deeper offshore conditions, including dives from the sea surface to 100 m depth across two pycnoclines: the seasonal thermocline and the permanent halocline (e.g. ). A previous glider study in the central Baltic Proper showed that, due to the strong halocline and the presence of turbidity maxima, the glider altimeter may erroneously register the halocline as the seafloor and stop descending further ().
Figure 5
The horizontal speed of the AUV was set to 1.5 m s−1 and the vertical speed to 0.25 m s−1. The mission began with a helical dive to 20 m depth, followed by a period of neutral floating until the vehicle resurfaced by its own buoyancy. In the next phase, the AUV was programmed to descend to an altitude of 4 m above the seafloor, but due to inertia it reached 3 m altitude (99 m depth) before resurfacing by its own buoyancy to the sea surface. Subsequently, the vehicle was directed SSE and instructed to follow a sawtooth trajectory between 1 m depth and 3 m altitude above the bottom. As in earlier missions, the AUV approached within ~2 m of the seafloor due to inertia. Importantly, the vehicle successfully traversed both the seasonal thermocline and the permanent halocline, demonstrating its ability to operate through strong pycnoclines in the offshore Baltic Proper. Overall, the offshore mission confirmed that the micro-AUV can operate reliably in deeper, strongly stratified environments.
2.3 Endurance and power consumption
The AUV records battery fullness as a percentage of total capacity. For each mission, battery consumption was normalized by the distance travelled, providing a measure of energy cost per kilometer:
Consumption = ΔBattery%/Distance [km]
To investigate the factors influencing energy expenditure, we examined the dependence of consumption on commanded speed, the mean conservative temperature of seawater, and the variability of vehicle attitude expressed through roll, pitch, and yaw (see Equations 1–5). Maneuvering effort was quantified from angular rates of roll, pitch and yaw, computed as time derivatives of the attitude angles:
The variances of these rates were calculated as:
and combined into a single maneuvering effort metric:
The relationship between battery consumption and the predictor variables was assessed using a multiple linear regression model.
where Y denotes the battery consumption per kilometer travelled in a mission, mean commanded AUV speed (m s−1), Temperature is the mean water temperature (°C), AttVar is the maneuvering effort defined above, and ϵ is the residual error. β0-β3 are regression coefficients.
The explanatory power of each predictor was evaluated using a variance decomposition approach. For each variable, the reduction in the coefficient of determination (R²) after its removal from the model was calculated:
where R²_full is the coefficient of determination for the full regression model and R²_drop is the value obtained when the respective predictor was excluded. This method allowed us to quantify the relative importance of commanded speed, temperature, and maneuvering effort in determining battery consumption, as well as the fraction of variability that remained unexplained.
The analysis demonstrated that battery consumption per kilometer correlated most strongly with the mean commanded speed (Figure 6), while water temperature and maneuvering effort provided additional but weaker contributions (Table 2). Based on the regression between commanded speed and battery cost (Figure 6), the AUV endurance range spans from ~40 km (~7.8 h) at the minimum commanded speed of 1.5 m s−1 to ~14 km (~1.3 h) at the maximum commanded speed of 3.0 m s−1, assuming 90% of the battery capacity is available for propulsion. Temporal endurance can be extended if the vehicle is allowed to float part of the mission.
Figure 6
Table 2
| Component | Explained (%) |
|---|---|
| Speed | 76.2 |
| Temperature | 2.1 |
| AttitudeRateVar | 0.6 |
| Unexplained (residual) | 21.1 |
Variance decomposition of battery consumption per kilometer, showing the relative contributions of speed, temperature, and attitude variability, as well as the residual unexplained fraction.
2.4 Sensor performances
The sensors installed on the AUV have fast response times and high accuracy (see Section 2.1). To verify measurement reliability, AUV sensor data were compared with reference measurements obtained from SBE probes. Simultaneous vertical profiles were collected in the Baltic Proper during Mission 11. In addition, chlorophyll-a and turbidity data were validated in the Matsalu Bay in April during Missions 1–3 and 5.
Comparison with a brand-new, factory-calibrated SBE sonde (Figure 7) showed a strong agreement between the AUV and SBE measurements. The water column structure and vertical gradients were very well captured by the AUV sensors. In the upper layer (2–15 m depth), the differences in temperature and salinity were 0.0235 °C and 0.0078 g kg−1 (SBE – AUV), respectively, while in the deeper layer (82–88 m depth) the corresponding differences were −0.0125 °C and 0.0036 g kg−1. Although simultaneous measurements at exactly the same time and position were intended, perfect alignment was not achievable due to boat drift and other operational factors. Considering these factors and the presence of strong gradients in the area, these small deviations demonstrate that the AML sensors installed on the AUV provide highly accurate data suitable for resolving fine-scale oceanographic features in the brackish Baltic Sea and similar environments.
Figure 7
To compare the AUV’s chlorophyll-a and turbidity measurements with those from the SBE19 sonde, we conducted reference measurements with the SBE19 in the upper layer of Matsalu Bay. The measurements were performed as close as possible in time and location to the AUV transects. Each SBE19 measurement lasted a few minutes at a fixed site, while the corresponding AUV data were extracted within a 100 m radius around the SBE location. For the comparison, the median values from both datasets were used. The resulting median turbidity and chlorophyll-a values for the AUV and SBE19 are shown in Figure 8. The comparison shows a strong correlation between the two instruments, although the absolute values differ, likely due to differences in sensor design and measurement techniques between the two systems (; Richardson et al., 2025). In order to convert turbidity (NTU) values to suspended matter concentrations or to determine local chlorophyll-a values, regression relationships should be established between the optical AUV measurements and values obtained from water samples, which were not collected in the present study. However, the strong relationship between the classical SBE sonde and the AUV demonstrates the good performance of the AUV’s optical sensors.
Figure 8
3 Survey results
3.1 Meteorological conditions before and during AUV missions
Surveys were conducted in three time windows in April, July, and October 2025 (see Table 1). The water properties in the coastal areas of the Baltic Sea are strongly dependent on the atmospheric forcing (; ; ). Wind speed and direction, as well as sea-surface and air temperature, during and within five days preceding each mission period, are presented in Figure 9.
Figure 9
Both the Matsalu and Suur Strait missions were conducted during an atmospheric heatwave. Daily maximum air temperatures at the nearby coastal station in Haapsalu were between 3 and 8 °C in the preceding days, but increased to 15–16 °C on the day before and during Mission 1 (Figure 9b). Exceptionally high daily maximum air temperatures, reaching up to 20 °C, were recorded during Missions 3 and 5 (Figure 9b). The heatwave led to pronounced warming of the water in the nearby enclosed Bay of Haapsalu, where temperatures rose from 4–6 °C before the missions to 9–10 °C on the day of Mission 5 (Figure 9b). Similar rapid warming was observed by the AUV in Matsalu Bay, whereas in the more open Moonsund area the warming was weaker (see Section 3.2). Light winds, predominantly from easterly directions, prevailed during Missions 1–5 (Figure 9a).
The July heatwave began two days before Mission 6 and persisted during Missions 6 and 7, when daily maximum air temperatures reached 28–29 °C (Figure 9d). Although the heatwave slightly weakened, it continued throughout Missions 8–10. This period of elevated air temperature led to a warming of the daily maximum sea-surface temperature in Suur Strait at Virtsu, increasing from 19–20 °C before the heatwave to 21–22 °C during Missions 6–10. Winds remained light throughout the missions (Figure 9c).
In contrast, Mission 11 was conducted during the cooling season. Nighttime air temperatures dropped to 2–7 °C (Figure 9f), resulting in a decrease of the sea-surface temperature at the Dirhami coastal station from about 14.5 °C to 12.5 °C within one week (Figure 9f). Before and during the mission, weak easterly winds prevailed (Figure 9e).
3.2 Matsalu Bay
Missions 1–3 and 5 were conducted in Matsalu Bay in April (Figure 3). The inner part of the bay was characterized by warmer, fresher, and more turbid water, whereas the western part was colder, saltier, and less turbid. During the five-day observation period, a marked warming occurred due to prevailing warm weather. At the beginning, water temperature was about 5 °C in the west and up to 7 °C in the inner bay, while by the final day it had increased to 11.7 °C inside the bay and exceeded 7 °C in the west during Mission 5. It should be noted that Mission 5 was focused on the mouth of the bay; thus, temperatures in the inner bay were likely even higher, while those in the western part were somewhat lower than those recorded at the mouth. Salinity, turbidity, and chlorophyll-a showed strong contrasts between the inner bay and the western, open Moonsund, ranging from 2.0 to 6.6 g kg−1, 1 to 30 NTU, and 1 to 14.5 mg m−3, respectively.
A frontal zone was present between the two water masses and was specifically targeted during Mission 5 (Figure 10). The front was oriented northwest–southeast, with fresher and more turbid water extending westward along the northern part of the bay, while more saline waters from the open Moonsund intruded into the bay from the southern side of the mouth. Within the 4 × 4 km mouth area, temperature ranged from 7.7 to 11.7 °C, salinity from 3.2 to 6.5 g kg−1, chlorophyll-a from 0.3 to 10.4 mg m−3, and turbidity from 0 to 12.5 NTU. The strongest frontal gradients reached 0.005 °C m−1 (0.5 °C per 100 m), 0.005 g kg−1 m−1 (0.5 g kg−1 per 100 m), 0.03 mg m−3 m−1 (3 mg m−3 per 100 m), and 0.02 NTU m−1 (2 NTU per 100 m), respectively.
Figure 10
Salinity provides a reliable descriptor of the water masses in the study area, with the freshest water representing the inner Matsalu Bay and the most saline water representing the open Moonsund. Based on this assumption, the correlations calculated for Mission 5 data suggest that the relative contribution of these two water masses largely determines the observed variability in temperature (R² = 0.83, n = 345 015), chlorophyll-a (R² = 0.93, n = 345 015), and turbidity (R² = 0.97, n = 345 015). This finding supports the interpretation that the distribution of water properties was primarily determined by the location of the front and frontal mixing between the two water masses.
Overall, the Matsalu Bay observations highlight the strong horizontal gradients between the fresher bay water and the more saline Moonsund water. These results demonstrate the AUV’s capability to capture estuarine front dynamics and related biogeochemical gradients at spatial resolutions unattainable by traditional ship-based surveys.
3.3 Suur Strait
The Suur Strait area was visited with the AUV once in spring (Mission 4) and on five consecutive days in summer (Missions 6–10). Mission 4 surveyed the area north of Kesselaid, while the subsequent missions focused on the Suur Strait south of Kesselaid (Figure 4). The collected data revealed both the interactions between Baltic Proper, Moonsund and Gulf of Riga waters and a submesoscale patchiness.
During the only spring survey, an interleaving, slightly warmer and more turbid water mass was observed at the coastal slope of Muhu Island. In the cross-strait section this feature appeared as a 6–8 m thick wedge extending about 1 km offshore (Figure 11), while in the along-strait section it manifested as a 4 m thick patch with a meridional extent of roughly 2 km (Figure 12). The mechanisms responsible for the formation of this patch are beyond the scope of the present study. Nevertheless, the observation demonstrates that the AUV is capable of capturing submesoscale structures.
Figure 11
Figure 12
Unlike in spring, temperature strongly enhanced stratification in July. Surface temperatures exceeded 23 °C, while bottom waters remained below 17 °C in the deep part on the Gulf of Riga side (Figure 13). Salinity varied between 5.9 and 6.7 g kg−1. The Moonsund functions as a shallow junction where the fresher, nutrient-richer Gulf of Riga water meets the saltier, nutrient-poorer Baltic Proper water (), and it also receives riverine input from the Kasari River via Matsalu Bay (see previous section) and smaller rivers. The Suur Strait is a frontal zone between Moonsund and Gulf of Riga waters. Despite the net freshwater flux from the Kasari, one might expect saltier water on the northern side of the meridional section and fresher water on the southern side (); however, the high-resolution AUV observations reveal a more complex hydrography.
Figure 13
At least five distinct water types (see Figure 13b) can be distinguished: (i) a cold (<17 °C), dense sub-thermocline Gulf of Riga water (WG cold) with salinity of 6.4-6.5 g kg−1; (ii) a saltier (~6.7 g kg−1), moderately warm (~18 °C) Baltic Proper water that had transited the shallow (~5 m) central Moonsund and experienced mixing en route (WBP); (iii) a fresher (~6.4–6.5 g kg−1) and marginally warmer (~18.5 °C) Moonsund water formed prior to the heatwave by mixing of Gulf of Riga, Baltic Proper, and riverine waters (WM cold); and, above these, two warm surface waters (~22–23 °C) developed under strong atmospheric heat flux: a fresher Gulf of Riga warm water (WG warm) and a relatively saltier Moonsund warm water (WM warm). Together, these waters and their mixtures produce lateral and vertical gradients that differ from the simple north–south salinity contrast expected from steady estuarine exchange and underscore the value of synoptic, high-resolution AUV sampling.
Considerable displacement and modification of water masses occurred during the five-day survey. On 21–22 July the surface layer was mostly occupied by fresher WG warm water, which already appeared in the southern part of the section on 20 July (Figure 13a), while in the rest of the surface layer WM warm was present. The core of the Baltic Proper water (WBP) exhibited the highest salinity on 18 July and became progressively fresher thereafter, decreasing from about 6.71 g kg−1 on 18 July to 6.66 g kg−1 on 22 July. The WBP was beneath the fresher and slightly warmer WM cold. Both the location and thickness of the WBP core varied over time. On 20 July (Figure 13b) the WBP intruded between the cold Gulf of Riga water (WG cold) and the warm Gulf of Riga water (WG warm) in the southern part of the section. Warm and saline patches of buoyant subsurface waters have previously been observed in the Gulf of Riga, with their formation linked to water exchange through the Irbe Strait (Stipa et al., 1999; ). The present AUV observations suggest that similar lenses also form in the Suur Strait. identified such submesoscale lenses using a fixed vertical profiler. However, their spatial extent, temporal evolution, and influence on the water properties of the Gulf of Riga remain poorly understood due to the absence of targeted observational campaigns.
Whereas in the Matsalu Bay frontal area the spatial distributions of chlorophyll-a and turbidity could largely be explained by the horizontal position of the front between Moonsund and Matsalu Bay waters, in the Suur Strait the vertical structure of the water column played a more prominent role. The chlorophyll-a concentrations were linked both to the thickness of the upper mixed layer and to the prevailing water mass, which varied between missions and spatially. For example, the fresher water mass that entered the southern part of the study area (Figure 13b) contained lower chlorophyll-a and turbidity levels (Figure 13c).
A striking feature was the persistent near-bottom turbidity maximum in the Suur Strait, which was observed during all five summer missions and across most of the section. The thickness of the near-bottom turbidity maximum layer was typically 2–4 m, with magnitudes reaching up to 12–13 NTU (Figure 13d). Wave-induced sediment resuspension occurs frequently in the Moonsund region () and related increases in turbidity have been observed by both remote sensing and fixed mooring (; Raag et al., 2014). In such cases, the water column is usually well mixed. However, in the Suur Strait a persistent near-bottom turbidity maximum was also observed under calm wind and stratified conditions, indicating that the elevated turbidity was not caused by wave-induced stress. It was more likely associated with a fluffy layer consisting of recently deposited, organic-rich material—planktonic detritus, bacteria, and other fine organic and inorganic particles (; ). This layer is easily resuspended compared with the underlying consolidated sediments (Orvain et al., 2003). Consequently, turbidity increases observed by remote sensing or in situ measurements in the upper layer may not necessarily indicate sediment resuspension, but rather the redistribution of this fluffy layer. From an observing-system perspective, such near-bottom processes cannot be detected by surface-based or remotely sensed observations, underscoring the unique contribution of AUV measurements to resolving benthic boundary-layer dynamics.
Overall, the Suur Strait observations reveal a complex and dynamic hydrographic system shaped by the interplay of Gulf of Riga, Moonsund, and Baltic Proper waters. The micro-AUV successfully resolved the fine-scale lateral and vertical structure of multiple interacting water masses, as well as transient submesoscale lenses and persistent near-bottom turbidity layers. These results demonstrate the capability of AUV-based surveys to capture short-term variability and benthic boundary-layer processes.
3.4 Offshore area
Mission 11 extended from shallow coastal waters to the offshore region (Figure 5) and encompassed all three principal layers of the eastern Baltic Sea: the warm and relatively fresh surface layer, the cold and more saline intermediate layer, and the warmer and saltiest deep layer. Temperature and salinity ranged from 5.1 to 16.0 °C and from 6.2 to 10.5 g kg−1, respectively, corresponding to a potential density difference of up to 4.3 kg m−3 between the surface and deep layers (Figures 14a–d). The layers were separated by two distinct pycnoclines — the thermocline and the halocline — located approximately at 5–25 m and 50–80 m depth, respectively. Both pycnoclines shoaled toward the coast, likely as a result of easterly winds (Figure 9e) preceding the survey, which are known to induce upwelling and raise the halocline along the southern coast (). Two zones were evident where colder and saltier thermocline water reached the surface: one within a ~3 km coastal zone at the nearshore end of the section, and another near the steep coastal slope, where depth decreased from 70 to 15 m over a distance of about 4 km (Figures 15a, b). These observations reveal a dual upwelling structure, in which the thermocline surfaces not only directly adjacent to the coastline—as classically expected—but also along the coastal slope. Moreover, the AUV measurements showed that although the upper mixed layer extended to more than 20 m depth in the offshore area, the water column remained stratified in the coastal zone, where the total depth was less than 15 m and mixing did not reach the seafloor.
Figure 14
Figure 15
The lower temperatures observed on the southern side of the section were not solely related to the vertical uplift of the pycnoclines caused by upwelling. Extensive shallow areas in the southern part of the region, including the Moonsund, cool more rapidly than the deeper offshore waters during autumn. It is likely that this colder, denser water cascades downslope until it reaches a depth corresponding to its density equilibrium (), thereby contributing to vertical exchange between the surface and deeper layers as well as to lateral exchange between coastal and offshore regions. The influence of seasonal cooling is also evident in the temperature–salinity relationship (not shown): for instance, water with a salinity of 6.6 g kg−1 had a temperature of approximately 12–13.5 °C in the southern part of the section, whereas in the offshore area it was 15–16 °C (Figures 15a, b). Although autumnal cooling had begun offshore as well, it progressed more slowly than in the shallow coastal zone. The maximum temperature in the entire section was found within a 3 m thick submesoscale patch centered around 18–21 m depth in the offshore area, where the negative buoyancy was compensated by a local vertical salinity gradient.
Elevated chlorophyll-a concentrations were confined to the upper layer and closely followed the depth of the upper mixed layer. Because the thermocline was shallower and the mixed layer thinner near the coast, the chlorophyll-a maximum was correspondingly thinner in this region and did not extend to the seafloor. Turbidity showed a distinct maximum within the halocline, at depths of 65–80 m. The elevated turbidity values were closely associated with the salinity structure and occurred at salinities between 9.4 and 9.9 g kg−1. This turbidity maximum was likely not caused by suspended matter but rather associated with the redoxcline within the oxic–anoxic transition zone, where enhanced turbidities represent the optical signature of colloidal sulfur particles generated during the mixing of oxic and sulfidic waters (e.g., ).
Overall, the offshore mission demonstrated the micro-AUV’s capability to perform deep dives through multiple pycnoclines and to resolve the vertical and lateral structure of the stratified water column in the Baltic Proper. The measurements revealed dual upwelling structure, differential cooling of coastal and offshore waters, and the general fine-scale horizontal and vertical structures of temperature, salinity, chlorophyll-a, and turbidity.
4 Discussion
4.1 Operational characteristics of a micro-AUV
The tested AUV is capable of covering a maximum distance of approximately 40 km during a 7-hour mission without recharging. The battery recharging time is in a similar order. Consequently, continuous 24-hour coverage of a specific area could, in principle, be maintained with two AUVs operating alternately. However, if recharging cannot begin immediately after recovery—due to transport or logistical delays — three AUVs would likely be required to ensure uninterrupted temporal coverage.
In comparison, an underwater glider operating in the Baltic Sea can typically travel 10–15 times farther without recharging than the AUV, although its velocity is about ten times slower (approximately 0.3 m s−1 versus 3.0 m s−1; Salm et al., 2023). Unlike gliders, however, AUVs can operate effectively in very shallow waters: our missions occasionally reached depths shallower than 2 m, and in principle, operation at ~1 m depth is possible, though this increases the risk of bottom collision and may compromise data quality.
To put these performance characteristics into a practical context, we next estimate the time required for the AUV to traverse a typical mesoscale eddy with a diameter of 5–10 km (; ). At 1.5 m s−1 the crossing requires ~1–2 h, whereas at 3.0 m s−1 it is reduced to 0.5–1 h. With a representative background current of 0.1 m s−1 (Suhhova et al., 2018; ), water would be displaced by only 0.2–0.7 km during the AUV transect, which is small compared to the eddy diameter. This confirms that the vehicle can provide a high-resolution, synoptic view of mesoscale eddies in the Baltic Sea.
By contrast, autonomous underwater gliders typically operate at horizontal velocities of only 0.3 m s−1 (; ; Salm et al., 2023). At these speeds, an eddy crossing would require 5–10 h, during which water could be advected 2–4 km by background currents. Such displacements are a substantial fraction of the eddy radius, meaning that glider sections cannot be regarded as fully synoptic. The AUV therefore offers a distinct advantage over gliders for resolving the spatial structure of mesoscale features in the Baltic Sea.
While the deepest parts of the Baltic Sea were not explored in this study, our results show that the AUV can easily pass through pycnoclines to depths of 100 m, suggesting that deeper regions such as the Gotland Deep could also be surveyed. The Landsort Deep remains the only location in the Baltic Sea that exceeds the vehicle’s 300 m depth rating.
By default, the AUV samples at a rate of 2 Hz, though this can be increased up to 20 Hz when higher temporal resolution is required. Given an operational speed of 1.5–3 m s−1, the default rate corresponds to a horizontal resolution of several meters along its track. The vertical speed was primarily set to 0.25 m s−1, resulting in a horizontal spacing of approximately 300 m between profiles for 50 m-deep dives, and 600 m for 100 m-deep dives. Pairaud et al. (2025) reported that slightly higher vertical speeds (up to 0.3 m s−1) can also be used successfully, although further increases tend to cause difficulties maintaining the desired depth limits. The AUV performed helix-shaped ascents with a radius of approximately 15 m, followed by neutral floating until surfacing, during which the vertical velocity was about 0.13 m s−1. Assuming a typical current velocity of 0.1 m s−1 in the Baltic Proper (Suhhova et al., 2018; ), and a barotropic current structure, the resulting horizontal drift during neutral floating would be on the order of the depth. However, due to the vertically layered current structure observed in deeper areas (), the actual drift there is likely smaller than the depth. In conclusion, the AUV enables both vertical and horizontal high-resolution measurements, making it well suited for investigating fine-scale processes in estuarine environments.
The temporal and horizontal resolution of measurements collected along the AUV’s track—typically on the order of a few meters at the default 2 Hz sampling rate, and improving to tens of centimeters at the maximum rate of 20 Hz—is comparable to that achieved by underwater gliders, whereas FerryBoxes and research vessel (RV)–mounted flow-through systems typically provide resolutions of 150–250 m (). The vertical resolution of the AUV is similar to that of underwater gliders and moored profiling systems. A major advantage of the AUV over these other observing platforms lies in its high maneuverability that enables three-dimensional mapping with spacing between adjacent measurement lines of only a few tens of meters, while maintaining meter-scale resolution along each track. This maneuverability also allows the AUV to navigate effectively in complex coastal and estuarine regions (Pairaud et al., 2025).
Unlike underwater gliders, the AUV does not transmit measurement data or its position over satellite or mobile networks. Instead, communication is limited to a UHF handheld command unit, which provides short-range location information when the vehicle is at the surface and within a range of 1–2 km, which could be approximately doubled with a longer antenna.
The absence of long-range measurement data and location transmission introduces a risk of vehicle loss and necessitates the presence of personnel at sea, which may not always be feasible during adverse weather conditions. Implementing real-time communication and positioning capabilities via satellite or mobile networks would significantly enhance operational flexibility, enabling safer, more efficient missions with reduced personnel requirements and greater adaptability to varying environmental conditions. To provide a backup source of location information in case the AUV drifts beyond UHF communication range, we plan to attach a small Argos satellite tag, similar to those used for tracking marine mammals (e.g. Prawirasasra et al., 2022), in future studies.
In principle, hydrodynamic conditions in the Baltic Sea impose only minor operational limitations on the micro-AUV, as typical current velocities (Raudsepp et al., 2011; ) are substantially lower than the vehicle’s cruising speed. This contrasts with underwater gliders, whose much lower speeds make them more sensitive to advection by ambient currents and may limit their ability to maintain planned survey trajectories under stronger flow conditions. Practical operational limits are therefore more closely related to deployment and recovery conditions, as well as surface communication constraints. Strong winds, waves, or poor visibility can complicate launch and retrieval operations and may restrict mission execution despite otherwise favorable underwater conditions.
Data can be downloaded and processed within approximately 10–30 minutes, provided that robust data-processing routines are established. For example, when operating at the maximum AUV speed, completing a cross-section of a mesoscale eddy with a diameter of 5–10 km and producing an initial visualization of the results would typically require 1–1.5 hours from deployment of the AUV.
Various sensors can be installed on the AUV; however, due to the small size of the micro-AUV, the payload capacity—that is, the number and weight of sensors that can be carried simultaneously—is limited. In this study, we used AML-3 sonde, which has three sensor ports, fitted with conductivity-temperature, chlorophyll-a fluorescence, and turbidity sensors. Because of differing sensor weights, re-ballasting is required whenever the sensor configuration is changed. For instance, the AML X2change oxygen sensor is heavier than the chlorophyll-a fluorescence and turbidity sensors, necessitating the addition of extra buoyancy to maintain proper trim. Furthermore, because the sensors are mounted on the nose section of the AUV, replacing them also requires rebalancing. For the Yuco AUV, this can be achieved to certain extent by adjusting the position of the internal battery, without opening the vehicle’s dry compartment.
The potential generation of underwater noise by the micro-AUV was not evaluated in this study. The vehicle is fully electrically powered and expected to produce relatively low acoustic emissions compared to conventional vessel-based surveys; however, quantitative assessment of underwater noise and its possible environmental implications was beyond the scope of the present work.
In summary, the tested micro-AUV demonstrated reliable operation across a broad range of coastal and offshore conditions, from 1–2 m shallow waters to depths exceeding 100 m, while maintaining sub-meter vertical and meter-scale horizontal resolutions. With an endurance of about 7 hours and a 40 km range, it enables efficient, high-resolution surveys of mesoscale to submesoscale features when deployed in repeated missions. The platform’s maneuverability, including the ability to perform helix-type dives, allows flexible control of horizontal sampling density and precise navigation across strong thermohaline gradients. While communication is currently limited to a short-range (1–2 km) UHF link, extending this range would substantially enhance operational flexibility and reduce the risk of instrument loss.
4.2 The potential role of micro-AUVs in observing systems
AUVs have the potential to play a significant role in integrated marine observing systems, and to complement existing observing platforms (Figure 16). Their high maneuverability, fine-scale spatial resolution, and ability to operate in shallow and complex coastal environments make them particularly valuable for targeted process studies, event-driven surveys, and adaptive sampling. AUVs can bridge the observational gap between fixed-point platforms and basin-scale observing systems, providing high-resolution three-dimensional data essential for understanding short-term variability and small-scale physical and biogeochemical processes. Together with Argo floats, research vessels, underwater gliders, and moored profilers, micro-AUVs contribute to characterizing the vertical structure of the water column, a key aspect of the marine environment that cannot be resolved by remote sensing, FerryBoxes, surface drifters, or coastal stations (Figure 16).
Figure 16
Due to their limited autonomy, small AUVs cannot make a substantial contribution to the collection of long-term time series, which are more effectively maintained by conventional ship-based monitoring, FerryBox systems, Argo floats, moored observatories, and coastal stations. Similarly, their limited operational range prevents them from supporting basin-scale studies independently. However, AUVs can provide valuable contextual and complementary information by capturing high-resolution spatial variability in the vicinity of fixed mooring sites, thereby helping to explain the temporal fluctuations observed in these long-term datasets.
Moreover, AUVs can be effectively deployed in key locations such as straits, river mouths, and narrow estuaries, where their maneuverability and fine-scale sampling capability enable detailed investigations that are challenging for other observing platforms. They can also be effectively used to map anthropogenic pressures, such as construction, dredging, wastewater discharge, or thermal plumes from power plant cooling water, as well as to conduct surveys around structures like wind turbines, jetties, and breakwaters (; Raag et al., 2014; Schultze et al., 2020; ), where high-resolution spatial mapping is required. AUVs can further help bridge the gap in high-resolution observations between underwater gliders, which typically operate in waters deeper than 25–30 m, and coastal stations located in very shallow environments. Unlike slow-moving gliders, micro-AUVs can operate effectively in areas with strong currents, maintaining their planned trajectories and sampling patterns even under dynamic flow conditions. Thus, micro-AUVs could be effectively used to study (sub)mesoscale coastal processes and coastal–offshore exchange studies (; ; Salm et al., 2023). They can also make a significant contribution to benthic observations due to their ability to follow bathymetry accurately, maintaining a constant altitude above the seafloor and enabling detailed mapping of near-bottom processes, such as groundwater discharges into the sea (; Szymczycha et al., 2023).
4.3 Potential use cases
Based on the missions conducted in the diverse locations of the Baltic Sea—ranging from shallow estuarine areas to deep offshore waters—several promising applications for micro-AUVs can be identified. These use cases demonstrate how such vehicles can complement existing observing platforms and enhance understanding of regional hydrography and ecosystem dynamics.
4.3.1 Coastal zones
AUVs are well suited for observing coastal gradients where strong spatial and temporal variability occurs over short distances. Their maneuverability and high-resolution sampling capability enable mapping of hydrographic structures that are difficult to resolve with other platforms. One example in the Baltic Sea is the complex coastal archipelago region (Miettunen et al., 2024).
4.3.2 River plumes and estuarine regions
Micro-AUVs can track freshwater dispersion, chlorophyll-a, oxygen, and suspended matter dynamics at river mouths and estuarine plumes—environments that are often too shallow for glider operation. In the Baltic Sea context, they can support studies of riverine influence and mixing in areas such as Matsalu Bay, as illustrated in the present study.
4.3.3 Harbor activities
Because of their compact size and precise navigation, micro-AUVs can operate safely in the vicinity of harbors to monitor and assess dredging or dumping activities, sediment resuspension, and related changes in turbidity. Their fine-scale spatial mapping capability can provide valuable information on the spreading of dredging-induced material. Dredging is regularly conducted in the Baltic Sea (Raag et al., 2014; ; Virtasalo et al., 2018) and is considered a potentially significant pressure on the marine environment.
4.3.4 Pollutant, wastewater, and thermal discharges
AUVs can map and quantify anthropogenic inputs such as wastewater outflows, thermal plumes from power-plant cooling systems, and other discharges (; ). Their maneuverability allows repeated surveys close to inputs, providing high-resolution data for tracking anthropogenic signals and evaluating environmental impacts.
4.3.5 Fronts, filaments, and (sub)mesoscale features
AUVs enable targeted, adaptive sampling of dynamic ocean features such as fronts, eddies, upwellings, and their filaments. In the present study, frontal structures in Matsalu Bay and submesoscale variability in the Suur Strait and the Baltic Proper were successfully resolved, demonstrating their potential to quantify short-term variability and small-scale exchange processes.
4.3.6 Benthic boundary layer and seafloor mapping
By maintaining a constant altitude of a few meters above the seafloor, AUVs can map near-bottom water properties—including groundwater discharge, high-turbidity layers, and sediment resuspension events. In the present study, a 2–4 m thick turbid layer above the seafloor was successfully detected.
4.3.7 Offshore renewable energy and infrastructure monitoring
AUVs could monitor offshore wind farms, bridges and other marine infrastructure that influence hydrographic conditions around installations (; ). Their ability to repeatedly follow programmed tracks enables consistent baseline monitoring and change detection over time.
4.3.8 Under-ice surveys
With careful navigation and mission planning, AUVs could somewhat operate beneath sea ice to collect temperature, salinity, and biogeochemical data from otherwise inaccessible sub-ice environments—an application particularly relevant for northern Baltic conditions during winter.
4.3.9 Event-based or adaptive missions
AUVs are well suited for rapid-response operations following exceptional events such as inflows, heatwaves (as illustrated in the present study), algal blooms, or pollution incidents (; von Nordheim et al., 2020; ). Their flexibility makes them suitable for adaptive sampling strategies.
4.3.10 Water-properties mapping in nature reserves
AUVs can conduct surveys in sensitive areas where the presence of boats or people disturbs local fauna, such as seals and birds. AUVs operate quietly and have minimal surface presence, avoiding the bird escape effect (e.g. ) and noise disturbance for seals (e.g. Prawirasasra et al., 2022). For example, the shallow Matsalu Bay, visited in this study, has seasonal restrictions on human presence, yet an AUV can collect data there with minimal disturbance.
4.3.11 Complementation of other observing platforms
AUVs can complement gliders, long-term moorings, and FerryBox systems. They can extend glider transects into shallower coastal areas, collect vertical profiles to complement FerryBox surface data, or map the surroundings of mooring sites to enhance interpretation of time-series variability. In addition, micro-AUVs can effectively extend coastal station observations toward offshore regions, providing a dynamic link between nearshore fixed-point measurements and wider-scale autonomous observations.
4.3.12 Acting as other platforms
For short periods, AUVs can emulate the functionality of other observing platforms, providing comparable data type and resolutions. They can be programmed to drift passively and act as a surface drifter (), collect data continuously along a route like a FerryBox system (), follow a sawtooth trajectory like an underwater glider, or maintain position while performing repeated vertical profiles, functioning as a profiling mooring.
5 Conclusions
This study demonstrated that the YUCO-PHYSICO micro-AUV can operate reliably across diverse Baltic Sea environments—from 1–2 m deep estuaries to >100 m deep, strongly stratified offshore waters—while providing high-resolution physical and biogeochemical measurements. The missions in Matsalu Bay, the Suur Strait, and the Baltic Proper showed that the AUV successfully captured estuarine fronts, submesoscale features, a dual upwelling structure, and near-bottom high turbidity layer. The estimated endurance ranges from ~40 km (~7 h) at 1.5 m s−1 to ~14 km (~1.3 h) at 3.0 m s−1. Sensor comparisons with reference measurements confirmed high data quality suitable for brackish-water applications.
Micro-AUVs can fill a critical observational gap in the Baltic Sea between conventional monitoring, fixed moorings, FerryBoxes, and underwater gliders, particularly in shallow and dynamic coastal zones. Integration into existing observing systems would be beneficial, as the micro-AUV represents an efficient platform for high-resolution surveys and rapid-response missions. The main limitations are short endurance, limited payload capacity, and short-range communication. Extending communication to mobile or satellite networks would make micro-AUVs more attractive for operational programs and end users.
Statements
Data availability statement
The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request. Access will be granted under a data-use agreement in accordance with institutional and national data-sharing policies. Requests to access the datasets should be directed to taavi.liblik@taltech.ee.
Author contributions
TL: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Validation, Visualization, Writing – original draft, Writing – review & editing. FB: Data curation, Formal analysis, Visualization, Writing – original draft, Writing – review & editing.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This study was co-funded by the Environmental Investment Centre (project no. RES.4.07.24-0018). The research was conducted using the research infrastructure “Marine Technology and Hydrodynamics Research Infrastructure MarTech” funded by the Estonian Research Council (TARISTU24-TK6).
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
1
AleniusP.NekrasovA.MyrbergK. (2003). Variability of the baroclinic Rossby radius in the Gulf of Finland. Cont. Shelf Res.23, 563–573. doi: 10.1016/S0278-4343(03)00004-9
2
AML Oceanographic (2025). X2Change Swappable Sensors Datasheet. Available online at: https://amloceanographic.com/documents (Accessed November 1, 2025).
3
AnderssonS.EkP.BergM.GrundestamJ.LindblomE. (2016). Extension of two large wastewater treatment plants in Stockholm using membrane technology. Water Pract. Technol.11. doi: 10.2166/wpt.2016.034
4
AstokV.OtsmannM.SuursaarÜ. (1999). Water exchange as the main physical process in semi-enclosed marine systems: the Gulf of Riga case. Hydrobiologia393:11–18 doi: 10.1023/A:1003517110726
5
CahillB.ChrysagiE.Vortmeyer-KleyR.GräweU. (2024). Deconstructing co-occurring marine heatwave and phytoplankton bloom events in the Arkona Sea in 2018. Front. Mar. Sci.11. doi: 10.3389/fmars.2024.1323271
6
CarpenterJ. R.RodriguesA.SchultzeL. K. P.MerckelbachL. M.SuzukiN.BaschekB.et al. (2020). Shear instability and turbulence within a submesoscale front following a storm. Geophys. Res. Lett.47. doi: 10.1029/2020GL090365
7
ChristiansenN.CarpenterJ. R.DaewelU.SuzukiN.SchrumC. (2023). The large-scale impact of anthropogenic mixing by offshore wind turbine foundations in the shallow North Sea. Front. Mar. Sci.10. doi: 10.3389/fmars.2023.1178330
8
ChristiansenC.EdelvangK.EmeisK.GrafG.JähmlichS.KozuchJ.et al. (2002). Material transport from the nearshore to the basinal environment in the southern Baltic Sea I. Processes and mass estimates. J. Mar. Syst.35, 133–150 doi: 10.1016/S0924-7963(02)00126-4
9
ChrysagiE.BasdurakN. B.UmlaufL.GräweU.BurchardH. (2022). Thermocline salinity minima due to wind-driven differential advection. J. Geophys. Res. Ocean.127. doi: 10.1029/2022JC018904
10
ChubarenkoI.DemchenkoN. (2010). On contribution of horizontal and intra-layer convection to the formation of the Baltic Sea cold intermediate layer. Ocean Sci.6, 285–299. doi: 10.5194/os-6-285-2010
11
CieślikiewiczW.DudkowskaA.Gic-GruszaG.JędrasikJ. (2018). Assessment of the potential for dredged material dispersal from dumping sites in the Gulf of Gdańsk. J. Soils Sediments18, 3437–3447 doi: 10.1007/s11368-018-2066-4
12
DanielssonÅ.JönssonA.RahmL. (2007). Resuspension patterns in the baltic proper. J. Sea Res.57:257–269 doi: 10.1016/j.seares.2006.07.005
13
DvornikovA. Y.MartyanovS. D.RyabchenkoV. A.EreminaT. R.IsaevA. V.SeinD. V. (2017). Assessment of extreme hydrological conditions in the Bothnian Bay, Baltic Sea, and the impact of the nuclear power plant “hanhikivi-1” on the local thermal regime. Earth Syst. Dyn.8. doi: 10.5194/esd-8-265-2017
14
EMODnet Bathymetry Consortium (2022). EMODnet Digital Bathymetry (DTM 2022). doi: 10.12770/ff3aff8a-cff1-44a3-a2c8-1910bf109f85
15
FarcyP.DurandD.CharriaG.PaintingS. J.TamminemT.CollingridgeK.et al. (2019). Toward a European coastal observing network to provide better answers to science and to societal challenges; the JERICO research infrastructure. Front. Mar. Sci.6. doi: 10.3389/fmars.2019.00529
16
FischerH.MatthäusW. (1996). The importance of the Drogden Sill in the Sound for major Baltic inflows. J. Mar. Syst.9, 137–157. doi: 10.1016/S0924-7963(96)00046-2
17
FliessbachK. L.BorkenhagenK.GuseN.MarkonesN.SchwemmerP.GartheS. (2019). A ship traffic disturbance vulnerability index for Northwest European Seabirds as a tool for marine spatial planning. Front. Mar. Sci.6. doi: 10.3389/fmars.2019.00192
18
GolubkovM.NikulinaV.GolubkovS. (2019). Effects of environmental variables on midsummer dinoflagellate community in the Neva Estuary (Baltic Sea). Oceanologia61, 197–207 doi: 10.1016/j.oceano.2018.09.001
19
HaavistoN.TuomiL.RoihaP.SiiriäS.-M.AleniusP.PurokoskiT. (2018). Argo floats as a novel part of the monitoring the hydrography of the bothnian sea. Front. Mar. Sci.5. doi: 10.3389/fmars.2018.00324
20
HELCOM (2023). Eutrophication. Thematic assessment 2016-2021. Third HELCOM holistic assessment of the Baltic Sea. Balt. Sea Environ. Proc.192.
21
HoltermannP.PrienR.NaumannM.UmlaufL. (2020). Interleaving of oxygenized intrusions into the Baltic Sea redoxcline. Limnol. Oceanogr.65, 482–503 doi: 10.1002/lno.11317
22
IvaskM.MeristeM.KuuA.KuttiS.SizovE. (2012). Effect of flooding by fresh and brackish water on earthworm communities along Matsalu Bay and the Kasari River. Eur. J. Soil Biol.53:11–15 doi: 10.1016/j.ejsobi.2012.08.001
23
JaskulakM.SotomskiM.MichalskaM.MarksR.ZorenaK. (2022). The effects of wastewater treatment plant failure on the gulf of gdansk (Southern baltic sea). Int. J. Environ. Res. Public Health19. doi: 10.3390/ijerph19042048
24
KarimovaS. (2017). Observations of asymmetric turbulent stirring in inner and marginal seas using satellite imagery. Int. J. Remote Sens.38, 1642–1664 doi: 10.1080/01431161.2017.1285078
25
KarlsonB.AnderssonL. S.KaitalaS.KronsellJ.MohlinM.SeppäläJ.et al. (2016). A comparison of FerryBox data vs. monitoring data from research vessels for near surface waters of the Baltic Sea and the Kattegat. J. Mar. Syst.162, 98–111. doi: 10.1016/J.JMARSYS.2016.05.002
26
KarstensenJ.LiblikT.FischerJ.BumkeK.KrahmannG. (2014). Summer upwelling at the Boknis Eck time-series station, (1982 to 2012) – A combined glider and wind data analysis. Biogeosciences11, 3603–3617. doi: 10.5194/bg-11-3603-2014
27
KikasV.LipsU. (2016). Upwelling characteristics in the Gulf of Finland (Baltic Sea) as revealed by Ferrybox measurements in 2007–2013. Ocean Sci.12, 843–859. doi: 10.5194/os-12-843-2016
28
KrayushkinE. V.LavrovaO. Y.NazirovaK. R.ElizarovD. A. (2023). Three-dimensional structure and dynamics of coastal eddy dipoles in the southeastern baltic sea: results of remote sensing and oceanographic experiments in summer 2021. Cosm. Res.61, 130–140 doi: 10.1134/S0010952523700582
29
KrechikV.MyslenkovS.KapustinaM. (2019). New possibilities in the study of coastal upwellings in the southeastern Baltic sea with using thermistor chain. Geogr. Environ. Sustain.12, 44–61 doi: 10.24057/2071-9388-2018-67
30
KreuzburgM.ScholtenJ.HsuF. H.LiebetrauV.SültenfußJ.RapagliaJ.et al. (2023). Submarine groundwater discharge-derived nutrient fluxes in eckernförde bay (Western baltic sea). Estuaries Coasts46, 1190–1207 doi: 10.1007/s12237-023-01202-0
31
LassH. U.MohrholzV.KnollM.PrandkeH. (2008). Enhanced mixing downstream of a pile in an estuarine flow. J. Mar. Syst.74, 505–527. doi: 10.1016/j.jmarsys.2008.04.003
32
LassH. U.MohrholzV.SeifertT. (2001). On the dynamics of the pomeranian bight. Cont. Shelf Res.21, 1237–1261 doi: 10.1016/S0278-4343(01)00003-6
33
LeeD. K.CenturioniL. (2018). Water following characteristics of Global Drifter Program drifters with and without subsurface float. Deep. Res. Part I Oceanogr. Res. Pap.137, 20–29 doi: 10.1016/j.dsr.2018.05.003
34
LeipeT.LoefflerA.EmeisK. C.JaehmlichS.BahloR.ZiervogelK. (2000). Vertical patterns of suspended matter characteristics along a coastal-basin transect in the western Baltic Sea. Estuar. Coast. Shelf Sci.51, 789–804 doi: 10.1006/ecss.2000.0715
35
LeppärantaM.MyrbergK. (2009). “ Topography and hydrography of the baltic sea,” in Physical Oceanography of the Baltic Sea. Eds. LeppärantaM.MyrbergK. ( Springer Berlin Heidelberg, Berlin, Heidelberg), 41–88. doi: 10.1007/978-3-540-79703-6_3
36
LiblikT.LipsU. (2011). Spreading of suspended matter in a shallow sea area influenced by dredging activities and variable atmospheric forcing: results of in-situ measurements. J. Coast. Res., 561–566.
37
LiblikT.LipsU. (2017). Variability of pycnoclines in a three-layer, large estuary: the Gulf of Finland. Boreal Environ. Res.22, 27–47.
38
LiblikT.NaumannM.AleniusP.HanssonM.LipsU.NauschG.et al. (2018). Propagation of impact of the recent major baltic inflows from the eastern gotland basin to the gulf of Finland. Front. Mar. Sci.5. doi: 10.3389/fmars.2018.00222
39
LiblikT.SkudraM.LipsU. (2017). On the buoyant sub-surface salinity maxima in the Gulf of Riga. Oceanologia59, 113–128. doi: 10.1016/J.OCEANO.2016.10.001
40
LiblikT.StoicescuS.-T.BuschmannF.LiloverM.-J.LipsU. (2023). High-resolution characterization of the development and decay of seasonal hypoxia in the Gulf of Riga, Baltic Sea. Front. Mar. Sci.10. doi: 10.3389/FMARS.2023.1119515
41
LiblikT.VäliG.SalmK.LaanemetsJ.LiloverM. J.LipsU. (2022). Quasi-steady circulation regimes in the Baltic Sea. Ocean Sci.18, 857–879. doi: 10.5194/OS-18-857-2022
42
LiloverM.-J.ElkenJ.SuhhovaI.LiblikT. (2017). Observed flow variability along the thalweg, and on the coastal slopes of the Gulf of Finland, Baltic Sea. Estuar. Coast. Shelf Sci. 23–33 doi: 10.1016/j.ecss.2016.11.002
43
LiloverM.-J.LipsU.LiblikT.SuhhovaI.VäliG. (2025). Variability of stratification and currents at the submesoscale detected from CTD and current profiles in the central Gulf of Finland. Boreal Environ. Res.30, 77–92. doi: 10.60910/ber2025.jtg5-j683
44
LinJ.ZouX.HuangF.YaoY. (2021). Quantitative estimation of sea surface temperature increases resulting from the thermal discharge of coastal power plants in China. Mar. pollut. Bull.164. doi: 10.1016/j.marpolbul.2021.112020
45
LipsI.LipsU.LiblikT. (2009). Consequences of coastal upwelling events on physical and chemical patterns in the central Gulf of Finland (Baltic Sea). Cont. Shelf Res.29, 1836–1847. doi: 10.1016/j.csr.2009.06.010
46
LyngsgaardM. M.MarkagerS.RichardsonK.MøllerE. F.JakobsenH. H. (2017). How well does chlorophyll explain the seasonal variation in phytoplankton activity? Estuaries Coasts40, 1263–1275. doi: 10.1007/s12237-017-0215-4
47
MatosT.MartinsM. S.HenriquesR.GoncalvesL. M. (2024). A review of methods and instruments to monitor turbidity and suspended sediment concentration. J. Water Process Eng.64. doi: 10.1016/j.jwpe.2024.105624
48
MerchelM.WalczowskiW.RakD.WieczorekP. (2024). The use of Argo floats as virtual moorings for monitoring the South Baltic Sea. Oceanologia66:99–110 doi: 10.1016/j.oceano.2024.01.002
49
MeyerD.LipsU.PrienR. D.NaumannM.LiblikT.SchuffenhauerI.et al. (2018). Quantification of dissolved oxygen dynamics in a semi-enclosed sea – A comparison of observational platforms. Cont. Shelf Res.169, 34–45. doi: 10.1016/J.CSR.2018.09.011
50
MiettunenE.TuomiL.WesterlundA.KanarikH.MyrbergK. (2024). Transport dynamics in a complex coastal archipelago. Ocean Sci.20:69–83 doi: 10.5194/os-20-69-2024
51
MohrholzV. (2018). Major baltic inflow statistics – revised. Front. Mar. Sci.5. doi: 10.3389/fmars.2018.00384
52
MuchowskiJ.JakobssonM.UmlaufL.ArneborgL.GustafssonB.HoltermannP.et al. (2023). Observations of strong turbulence and mixing impacting water exchange between two basins in the Baltic Sea. Ocean Sci.19:1809–1825 doi: 10.5194/os-19-1809-2023
53
OrvainF.HirP.SauriauP.-G. (2003). A model of fluffy layer erosion and subsequent bed erosion in the presence of the bioturbator. J. Mar. Res.61, 821–849. doi: 10.1357/002224003322981165
54
OstrowskaM.FicekD.StoltmannD.Stoń-EgiertJ.ZdunA.KowalewskiM.et al. (2022). Ten years of remote sensing and analyses of the Baltic Sea primary production, (2010–2019). Remote Sens. Appl. Soc Environ.26. doi: 10.1016/j.rsase.2022.100715
55
OtsmannM.AstokV.SuursaarÜ. (1997). A model for water exchange between the Baltic Sea and the Gulf of Riga. Nord. Hydrol.28, 351–364. doi: 10.2166/nh.1998.31
56
PairaudI.PettonS.SelletH.HelleringerC.Le BihanC.CharriaG. (2025). Using a micro-AUV for monitoring coastal river plume areas. CEANS 2025 Brest (Brest), 1–10. doi: 10.1109/OCEANS58557.2025.11104618
57
PeinJ.StanevaJ. (2024). Eutrophication hotspots, nitrogen fluxes and climate impacts in estuarine ecosystems: A model study of the Odra estuary system. Ocean Dyn.74, 335–354 doi: 10.1007/s10236-024-01607-w
58
PetersenW. (2014). FerryBox systems: State-of-the-art in Europe and future development. J. Mar. Syst.140, 4–12 doi: 10.1016/j.jmarsys.2014.07.003
59
PrawirasasraM. S.JüssiM.MustonenM.KlausonA. (2022). Underwater noise impact of a ferry route on dive patterns of transiting Baltic ringed seals. Est. J. Earth Sci.71, 201–213 doi: 10.3176/earth.2022.14
60
PrienR. D.Schulz-BullD. E. (2016). Technical note: GODESS – a profiling mooring in the Gotland Basin. Ocean Sci.12, 899–907. doi: 10.5194/os-12-899-2016
61
PurkianiK.JochumsenK.FischerJ.-G. (2024). Observation of a moderate major Baltic Sea inflow in December 2023. Sci. Rep.14. doi: 10.1038/s41598-024-67328-8
62
RaagL.SipelgasL.UiboupinR. (2014). Analysis of natural background and dredging-induced changes in TSM concentration from MERIS images near commercial harbours in the Estonian coastal sea. Int. J. Remote Sens.35, 6764–6780. doi: 10.1080/01431161.2014.963898
63
RaudseppU.LaanemetsJ.HaranG.AlariV.PavelsonJ.KõutsT. (2011). Flow, waves and water exchange in the Suur Strait, Gulf of Riga, in 2008. Oceanologia53, 35–56. doi: 10.5697/oc.53-1.035
64
ReckermannM.OmstedtA.SoomereT.AigarsJ.AkhtarN.BełdowskaM.et al. (2022). Human impacts and their interactions in the Baltic Sea region. Earth Syst. Dyn.13:1–80 doi: 10.5194/esd-13-1-2022
65
RichardsonE. T.KrausT. E. C.SturgeonC. L.O’DonnellK.BergamaschiB. A. (2025). Identifying conditions associated with outliers produced by three different chlorophyll fluorometers: A comparison of instrumentation and development of correction formulae. Limnol. Ocean. Methods23, 673–687. doi: 10.1002/lom3.10705
66
RolffC.WalveJ.LarssonU.ElmgrenR. (2022). How oxygen deficiency in the Baltic Sea proper has spread and worsened: The role of ammonium and hydrogen sulphide. Ambio51:308–2324 doi: 10.1007/s13280-022-01738-8
67
SalmK.LiblikT.LipsU. (2023). Submesoscale variability in a mesoscale front captured by a glider mission in the Gulf of Finland, Baltic Sea. Front. Mar. Sci.10. doi: 10.3389/FMARS.2023.984246
68
SchmidtB.WodzinowskiT.BulczakA. I. (2021). Long-term variability of near-bottom oxygen, temperature, and salinity in the Southern Baltic. J. Mar. Syst.213, 103462. doi: 10.1016/J.JMARSYS.2020.103462
69
SchultzeL. K. P.MerckelbachL. M.HorstmannJ.RaaschS.CarpenterJ. R. (2020). Increased mixing and turbulence in the wake of offshore wind farm foundations. J. Geophys. Res. Ocean.125, e2019JC015858. doi: 10.1029/2019JC015858
70
Seabird Scientific. (2025a). ECO FLNTU Fluorometer and Scattering Meter datasheet. Available online at: https://www.seabird.com/asset-get.download.jsa?id=69833850019 (Accessed November 1, 2025).
71
Seabird Scientific. (2025b). SBE 19plus V2 SeaCAT Profiler datasheet. Available online at: https://www.seabird.com/asset-get.download.jsa?id=69833853385 (Accessed November 1, 2025).
72
Seabird Scientific. (2025c). SBE 43 Dissolved Oxygen Sensor datasheet. Available online at: https://www.seabird.com/asset-get.download.jsa?id=69833850207 (Accessed November 1, 2025).
73
SheJ.HøyerJ. L.LarsenJ. (2007). Assessment of sea surface temperature observational networks in the Baltic Sea and North Sea. J. Mar. Syst.65, 314–335 doi: 10.1016/j.jmarsys.2005.01.004
74
SoometsT.TomingK.PaavelB.KutserT. (2022). Evaluation of remote sensing and modeled chlorophyll-a products of the Baltic Sea. J. Appl. Remote Sens.16. doi: 10.1117/1.jrs.16.046516
75
StipaT.TamminenT.SeppäläJ. (1999). On the creation and maintenance of stratification in the Gulf of Riga. J. Mar. Syst.23, 27–49. doi: 10.1016/S0924-7963(99)00049-4
76
StockmayerV.LehmannA. (2023). Variations of temperature, salinity and oxygen of the Baltic Sea for the period 1950 to 2020. Oceanologia65, 466–483 doi: 10.1016/j.oceano.2023.02.002
77
SuhhovaI.LiblikT.LiloverM.-J.LipsU. (2018). A descriptive analysis of the linkage between the vertical stratification and current oscillations in the Gulf of Finland. Boreal Environ. Res.23, 83–103.
78
SuursaarÜ.KutserT.ApsR.KullasT.VahtmäeE.MetsamaaL.et al. (2009). Hydrodynamically induced or modified patterns derived from satellite images in the coastal waters of Estonia. J. Coast. Res. 1602–1606
79
SzymczychaB.BöttcherM. E.DiakM.Koziorowska-MakuchK.KulińskiK.MakuchP.et al. (2023). The benthic-pelagic coupling affects the surface water carbonate system above groundwater-charged coastal sediments. Front. Mar. Sci.10. doi: 10.3389/fmars.2023.1218245
80
TanhuaT.McCurdyA.FischerA.AppeltansW.BaxN.CurrieK.et al. (2019). What we have learned from the framework for ocean observing: Evolution of the global ocean observing system. Front. Mar. Sci.6. doi: 10.3389/fmars.2019.00471
81
VetemaaM.EschbaumR.VerliinA.AlbertA.EeroM.LillemägiR.et al. (2006). Annual and seasonal dynamics of fish in the brackish-water Matsalu Bay, Estonia. Ecol. Freshw. Fish15, 211–220. doi: 10.1111/J.1600-0633.2006.00134.X
82
VirtasaloJ. J.KorpinenS.KotilainenA. T. (2018). Assessment of the influence of dredge spoil dumping on the seafloor geological integrity. Front. Mar. Sci.5. doi: 10.3389/fmars.2018.00131
83
von NordheimL.KotterbaP.MollD.PolteP. (2020). Lethal effect of filamentous algal blooms on Atlantic herring (Clupea harengus) eggs in the Baltic Sea. Aquat. Conserv. Mar. Freshw. Ecosyst.30. doi: 10.1002/aqc.3329
84
WalczowskiW.MerchelM.RakD.WieczorekP.GoszczkoI. (2020). Argo floats in the southern Baltic Sea. Oceanologia62:478–488 doi: 10.1016/j.oceano.2020.07.001
85
YlöstaloP.SeppäläJ.KaitalaS.MaunulaP.SimisS. (2016). Loadings of dissolved organic matter and nutrients from the Neva River into the Gulf of Finland – Biogeochemical composition and spatial distribution within the salinity gradient. Mar. Chem.186, 58–71. doi: 10.1016/J.MARCHEM.2016.07.004
Summary
Keywords
AUV, Baltic Sea, estuarine front, estuary, submesoscale processes, upwelling
Citation
Liblik T and Buschmann F (2026) Application of a small AUV for mapping water column properties in the Baltic Sea. Front. Mar. Sci. 13:1743589. doi: 10.3389/fmars.2026.1743589
Received
10 November 2025
Revised
09 February 2026
Accepted
23 February 2026
Published
17 March 2026
Volume
13 - 2026
Edited by
Musa Bashir, University of Liverpool, United Kingdom
Reviewed by
Rui Caldeira, Agência Regional para o Desenvolvimento da Investigação Tecnologia e Inovação (ARDITI), Portugal
Aracelis Narayan Rajnauth, Agência Regional para o Desenvolvimento da Investigação Tecnologia e Inovação (ARDITI), Portugal
Updates
Copyright
© 2026 Liblik and Buschmann.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Taavi Liblik, taavi.liblik@taltech.ee
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.