The Impact of Anchoring on Seafloor Integrity: An Integrated Assessment within a Major Bunkering Area of the Maltese Islands

Abstract

International shipping is a fundamental component of the global economy. As the industry expands, the demand for high-capacity vessels increases, raising concerns about their impact on the marine environment. While awaiting access to port facilities, vessels often anchor to save fuel and prevent drifting, but this practice is a significant cause of mechanical disturbance to the seafloor and benthic habitats. Identifying and quantifying anchoring pressure is essential for better managing and mitigating the damage to the seafloor. The Automatic Identification System (AIS) can be utilized to assess anchoring pressure by transmitting vessel information (e.g., position, type and size) to other vessels and coastal stations. This research evaluates anchoring pressure in a strategically located bunkering area around the Maltese Islands using AIS data collected from an antenna at the University of Malta. An arbitrary index was developed to determine anchoring pressure, and the AIS data was used to create GIS maps showing the location and size of vessels within the bunkering area, as well as plots depicting anchoring pressure by vessel type, seasonality, and density. This study serves as a blueprint for future assessments of anchoring pressures from various maritime activities in other areas around the Maltese Islands and provides a decision support tool for national policy-making related to Descriptor 6 (Seafloor Integrity) of the Marine Strategy Framework Directive (MSFD), the Maritime Spatial Planning Directive (MSPD), and the management plan for Sites of Community Interest (SCI) and Marine Protected Areas (MPAs).

1. Introduction

Global trade prosperity is driven by international shipping, and the shipping industry has grown into a multi-billion-dollar industry [1,2]. It is estimated that 80% of global cargo is transported via shipping [2], with 90% of the EU’s external cargo originating from seaborne sources [3]. Shipping in the Mediterranean region has increased over the past decades [4]. It is estimated that a total of 200,000 vessels cross the Mediterranean Sea annually with 30% of these vessels unloading commercial goods in one of the many ports dotted around the coast of the Mediterranean Region [4].

It is standard practice for vessels to anchor while waiting for instructions from port operators to access port facilities and services [1]. Additionally, anchoring reduces fuel consumption and emissions as well as prevents a vessel from drifting [1]. Consequently, anchoring is responsible for mechanical damage caused to the seafloor and benthic habitats and ecosystems (e.g., coral reefs and seagrass meadows) [1,5]. Damage occurs when an anchor is deployed and retrieved and when the mooring chain/rope swings and sways, resulting in abrasion (anchor scour) [6,7].

Additionally, an anchor can shift on the seabed, and if the hold of the anchor is lost, it can drag across the seafloor to create furrows [1,7,8]. Therefore, it is vital to locate and quantify anchoring pressures to the seabed, pursuant to mitigating the damage to the seafloor resulting from maritime activities by devising measures aimed at reducing conflicts between maritime activities, uses and the marine environment [5].

The Marine Strategy Framework Directive (MSFD) (2008/56/EC), together with the Marine Spatial Planning Directive (MSPD) (2014/89/EC), are the two pillars of the Integrated Maritime Policy that strive to ensure that the marine environment is protected and conserved while allowing for maritime industries to thrive [9,10,11].

The MSFD follows an ecosystem-based management (EBM) approach which recognizes that the interactions between biophysical and human components of an ecosystem are complex to unpack and resolve [10]. The MSPD was established to support EU countries in implementing maritime spatial plans within their marine waters, and to organize and manage present and potential future maritime activities so that various ecological, economic, and social objectives can be achieved (Article 4 and Article 5) [11].

Seafloor Integrity is defined in Descriptor 6 of the MSFD as “a level that ensures that the structure and functions of the ecosystems are safeguarded and benthic ecosystems, in particular, are not adversely affected” [10]. According to Rice et al. (2010) the descriptor term “seafloor” comprises both the physical and chemical parameters of the seabed as well as the benthic biological communities, whilst “integrity” refers to the spatial connectedness of an ecosystem, implying that these ecosystems are not unnaturally fragmented and that ecosystems are functioning naturally [12]. Four seafloor substrates are defined by MFSD based on their physical properties, namely soft, gravel, hard, and biogenic substrates [13,14]. All these substrates support an abundant and diverse habitat of marine organisms, with their habitat building often dependent upon the physical properties of the seabed (e.g., geology and topography) and site-specific hydrodynamic processes (e.g., sediment transport, deposition, infilling and admixing).

Biogenic substrates, such as seagrass meadows, are recognized as the most vulnerable to physical damage [15]. Numerous factors such as climate change (e.g., marine heatwaves and extreme storm events), the introduction of invasive species, and human activities (e.g., coastal development, urban/industrial discharges, bottom-trawling and anchoring) have contributed to the rapid decline of seagrasses [16,17,18,19,20]. In the Mediterranean, studies on habitat health of seagrasses record a range of results that swing from recovery rates in some locations for fast growing species (e.g., Zostera spp.) [21], to an overall decline for slow growing endemic Posidonia oceanica (L.) Delile of between 13% and 50% [22,23]. P. oceanica meadows play a crucial role in the Mediterranean Region by supporting over 25 ecosystem services to marine communities, marine food webs, and coastal ecosystems [20,24,25,26]. Its economic value for the Mediterranean is estimated to range from €150/km²/year for carbon sequestration (EC, 2019a), €2 million/year for human wellbeing to €190 million/year for the fisheries industry [25,27,28].

Under MSFD, the objective of Seafloor Integrity (D6.2) is to prevent human pressures on the seabed from hindering the ecosystem components, functions, and resilience [15]. However, MFSD also allows for the occurrence of human activities, including those that may impact the seafloor, as long as these activities remain sustainable and in compliance with the MSFD [12,15]. The impact assessment of these activities presents unique challenges due to the patchy spatial distribution of human activities, benthic ecosystems, and seafloor monitoring efforts. This variability creates a mosaic of differing degrees and types of seafloor disturbance [14,15]. Consequently, interpolating data to map a cohesive regional assessment of the impacts remains complex.

To achieve Seafloor Integrity, understanding the structure and dynamics of benthic communities is crucial for assessing ecosystem functionality. This is a mandatory obligation of the Maltese Islands as signatory of both the MSFD (transposed as Marine Policy Framework Regulations S.L. 549.52 in 2011) and the MSPD (transposed as Maritime Spatial Planning Regulations L.N. 341 of 2016 under the Development Planning Act Cap. 552).

2. Study Area

The Maltese Islands are situated in central Mediterranean Sea, at approximately 90 km south of the Sicilian coast and 260 km north of the North African coast [28,29]. The Maltese Archipelago consists of three main islands, Malta, Gozo, and Comino, including several small uninhabited islands (Figure 1a). The seabed topography between Malta and Sicily is considered to be shallow (approximately 150 m), with a gentle slope running in a North-South direction [30]. There are 18 Marine Protected Areas (MPAs) designated within the 25-nautical-mile Maltese Fisheries Management Zone (FMZ) protecting various habitats and species, and eight of which are designated as Special Protected Areas (SPAs).

Bunkering is an important maritime commercial activity within the Maltese Islands. The archipelago sits very close to one of the main shipping routes found in the Mediterranean (Sicily Channel), and given its location, maritime transport is of great economic importance [4,31,32]. The Maltese Islands are a well-established bunkering hub, providing vessels with essential services, such as bunkers, without deviating the vessel too much from its original route [33]. To facilitate bunkering operations, the Maltese Islands have five designated offshore bunkering areas (Figure 1b), strategically located in its territorial waters, and with a combined area of 58.27 km² [4,34]. Upon arriving in Maltese territorial waters, shipping vessels are assigned to one of these five bunkering areas, depending on the weather conditions and vessel type [33]. Currently. the Maltese Islands have a fleet of 17 bunker barges owned by several private companies and permitted to provide bunkering services [33,34,35].

This research investigates the anchoring pressure in Malta’s sole northern bunkering and anchorage zone, known as Bunkering Area 1. Within this area lies a significant reef, locally called Is-Sikka l-Bajda, which is among the largest reefs in Maltese territorial waters (Figure 1b). Dense P. oceanica meadows, settled on both matte and bedrock with small isolated patches, are found on Is-Sikka l-Bajda and its surrounding area [36]. This bunkering area also falls within the confines of one of Malta’s largest marine Sites of Community Interest (SCI) and Marine Protected Areas (MPAs) i.e., Żona fil-Baħar bejn Il-Ponta ta’ San Dimitri (Għawdex) u Il-Qaliet (MT 105). This MPA protects the P. oceanica meadows, maerl and Is-Sikka l-Bajda reefs. Despite such a high level of environmental protection, coexisting with intense pressure from bunkering activities, no comprehensive assessment of the scale and extent of this impact was ever conducted for this bunkering area.

3. Materials and Methods

3.1. ROV Surveys

An ROV (remotely operated vehicle) footage, recorded over Is-Sikka l-Bajda within the AMAre project, enabled this study to use the same footage to assess some of the damage caused to P. oceanica meadows as a result of vessels anchoring within Bunkering Area 1. The ROV footage was collected using the Saab Seaeye Falcon ROV, which is a lightweight, easy to control ROV. It utilizes four horizontal thrusters and one vertical thruster to maneuver through the water column.

The AMAre project funds allowed for two surveys to take place (Figure 2). The first survey was conducted on 11 May 2019, running from North to South, and the second survey was conducted on 29 May 2019, running from East to West. Each survey lasted approximately 6 h, and the footage collected from each survey was approximately 4.5 h long. During each survey, the mobile application Locus Map—installed on an Android device—took GPS readings at 10-s interval. At the end of each survey, the GPS coordinates recorded by Locus Maps were exported as a comma-separated value (.csv) file.

The ROV footage was recorded in five-minute increments and stored on the internal hard-drive of the video-recording system. After the survey, the ROV footage was downloaded from the internal hard drive, retaining the five-minute increments. The footage was subsequently analyzed by AquaBiotech Ltd. to produce GIS maps and a report on the damage caused to the seafloor through anchoring, as required by the AMAre Project. The impacts from anchoring were classified as either a hit (anchor halos) or a scour (linear impacts resulting from anchor chains). The information gathered from the ROV footage included the number of hits and scours as well as the distribution and health of the P. oceanica meadows.

3.2. AIS Data

The AIS data used in this study were downloaded from the MySQL AIS database that is housed and maintained by the Oceanography Research Group within the Department of Geosciences at the University of Malta. This study has only utilized AIS data since it is highly unlikely that small-sized vessels, not recorded through AIS means, would drop anchor within Bunkering Area 1. The coordinates of Bunkering Area 1 were taken from the ‘Designated Bunker Supply Anchorage Area’ as issued by TM [37]. The outer coordinates of Bunkering Area 1 (top-left coordinate and bottom-right coordinate) were used to establish a polygon-shaped ‘search’ area (Figure 3).

A search query was conducted within MySQL AIS database to find all vessels within the established search area that registered a speed of 0 knots between July 2016 and June 2019. A total of 16,367 AIS entrieswere exported into a spreadsheet with details on the MMSI number, date and time, vessel position, speed, direction, heading, estimated time of arrival, destination, and navigation status. A second spreadsheet was exported, including the MMSI number, vessel name, call sign, IMO number, ship type, deadweight, length, breadth, draught, country flag of registration, year of build, and record index. Both files were saved as a .csv to facilitate the generation of GIS maps. The first table consisted of vessels registered as authorized bunker barges and supply vessels, retaining only AIS entries for stationary vessels.

A third table was created for vessels with the navigation status “At Anchor”, and excluded entries with different statuses and multiple AIS entries for the same MMSI number within the same timeframe (Figure 3). For authorized bunker barges and supply vessels, AIS entries were retained if their position varied significantly from their previous position within the same timeframe. This third table was split into six-month intervals, thus producing an additional six tables. The gross tonnage and summer deadweight tonnage were also added to these tables, and this entailed searching for each vessel on the Marine Traffic webpage to acquire such information. These tables would enable the visual representation of the number and size of vessels anchoring in Bunkering Area 1 between July 2016 and June 2019. A flow chart summarizing the data extraction process is presented in Figure 4.

Before each survey, both AIS and VMS data for 2018–2019 were overlaid over a bathymetric map of Is-Sikka l-Bajda (Figure 5). This exercise highlighted the areas with the highest concentration of vessels that were known to be stationary (0 knots). This information would determine the path that the ROV would take while out in the field. The hypothesis proposed that the predetermined paths would cover areas with varying levels of anchoring pressure on P. oceanica meadows and the seafloor, thereby providing a comprehensive spectrum for comparative analysis. Each path would include a different number of stationary vessels and span depths ranging from 15 to 30 m.

3.3. Data Mapping

The GIS maps in this study were generated using QGIS (v3.8 “Zanzibar”), using the AIS tables imported as delimited text layers, along with the GPS coordinates from ROV surveys. The Maltese Islands base map, updated in August 2019, was downloaded from GEOFABRIK and imported as a vector file. Two P. oceanica habitat maps from the Malta Inspire Geoportal were also imported as vector files.

The coordinates for Bunkering Area 1 and the aquaculture zone were imported as delimited text layers and converted to shapefiles. AIS entry layers were symbolized as “Graduated” to display point sizes according to vessel gross tonnage, categorized using the “Pretty Breaks” function. Legends displayed the total number of vessels per six-month interval and the vessel size category.

GIS maps were also created to illustrate the extent and intensity of anchor damage observed in the ROV footage. These maps show the number of anchor hits, scours, impact intensity (low or high), areas affected by anchor hits and scours, and the distribution of P. oceanica along the ROV tracks. When there was an overlap between the anchor damage maps and AIS data, the coordinates of the recorded damage and the time registered on the ROV footage were noted to allow for the extraction of direct examples of the anchor damage caused to P. oceanica meadows.

3.4. Calculations of the Shipping Index

An arbitrary index was created to determine the anchoring pressure within Bunkering Area 1. Anchoring pressure was computed by multiplying the total residence time, the total number of vessels and the total tonnage of the vessels found in a grid cell so as to generate a Shipping Index. To compute this Shipping Index, all the AIS entries of the vessels that had their navigation status set to “At Anchor” were required. The first Microsoft Excel file downloaded from the MySQL AIS database was sorted according to the navigation status of the vessels and any vessels that had their navigation status set to anything other than “At Anchor” were removed. Also, the IMO number and gross tonnage corresponding to the MMSI number of these vessels were added to this file. Each MMSI number was checked against the second Microsoft Excel file downloaded from the MySQL AIS database to retrieve the IMO number of the vessels. The MMSI/IMO number was used to retrieve the gross tonnage of the vessel from the Marine Traffic webpage. The IMO number was added to keep track of the vessel just in case a vessel had to be referenced. The IMO number never changes, unlike an MMSI number when, for example, the ownership of the vessel changes. To facilitate data processing, this table was split into two separate tables. The first table contained information on the recorded vessels, including their MMSI number, the IMO number, and the gross tonnage. The second table contained the AIS entries of the vessels, including their date and time, MMSI number, and the position of the vessel in terms of latitude and longitude.

3.5. Spatial Plots for Density of Anchored Vessels and Shipping Index

A density plot was generated to determine the areas that were most frequented by the anchored vessels within Bunkering Area 1 between July 2016 and June 2019. The table that contained the data for the recorded anchored vessels was amended so as to retain only the latitude and longitude of these vessels and was saved as a text (.txt) file. The two tables containing the GPS coordinates from both surveys were also amended to retain only the latitudes and longitudes data. A script code, executable through both MATLAB and Octave, was written to read and extract the latitude and longitude of these generated text files. A total of 792 points were initially extracted from the anchored vessels text file. However, after applying the script to establish a domain matrix, points outside this matrix were removed, resulting in 786 points available for plotting.

The script code computed the total number of vessels found in each grid cell and displayed the results as colored grid cells through a pseudocolor plot. The anchored vessels (red) and survey paths (blue and magenta) were plotted on top of the pseudocolor plot.

An arbitrary index was then created to determine the anchoring pressure within Bunkering Area 1. Anchoring pressure was computed by multiplying the total residence time, the total number of vessels and the total tonnage of the vessels found in a grid cell so as to generate a Shipping Index. The MMSI/IMO number was used to retrieve the gross tonnage of the vessel from the Marine Traffic webpage. A script code, executable by MATLAB, was written to read and extract the information from the tables and a spatial domain matrix (similar to that for density plots) was established. The components of the index were calculated for a 25 by 31 grid having a 0.0025 grid resolution.

The sizes of the vessels (in terms of Gross Tonnage), were grouped into 12 categories (

Table 1. The 12 broad categories groupings vessel sizes in terms of Gross Tonnage (G.T.).

Group	Group Vessel Size	Group	Group Vessel Size
01	0 ≤ GT < 500	02	500 ≤ GT < 10,000
03	10,000 ≤ GT < 20,000	04	20,000 ≤ GT < 30,000
05	30,000 ≤ GT < 40,000	06	40,000 ≤ GT < 50,000
07	50,000 ≤ GT < 60,000	08	60,000 ≤ GT < 70,000
09	70,000 ≤ GT < 80,000	10	80,000 ≤ GT < 90,000
11	90,000 ≤ GT < 100,000	12	100,000 ≤ GT < 200,000

), based on the same categories established when producing the GIS maps at six-month intervals. The maximum residence time for the same MMSI was set to 2.5 days to prevent false readings. Pseudocolor plots were generated for each vessel size grouping depicting the total residence time of the vessels, the total number of vessels, and the total tonnage of the vessels for each grid cell. Additionally, text outputs of the results were generated from which the shipping index could be calculated.

A separate Microsoft Excel file was created for the 12 vessel size categories, with each file having a spreadsheet for the total number of vessels, total residence time, and total tonnage. Data normalization scaled the values in each spreadsheet within a zero to one range, and the index was computed by multiplying the normalized total number of vessels, averaged residence time and averaged tonnage values. This method was then replicated for the remaining ship gross tonnage categories. A final script code was written to establish a domain matrix showing the spatial distribution of the anchoring pressure over Bunkering Area 1, in terms of the total number of vessels, average time of residence and average tonnage.

4. Results

4.1. Anchoring Pressure Based on the AIS Data

According to the AIS data analysis, a total of 791 vessels were anchored within Bunkering Area 1 between July 2016 and June 2019. General cargo (174), oil products tankers (172), bulk carrier (98) and oil/chemical products tankers (84) were the vessels with the most anchorage activity within Bunkering Area 1. The large presence of these vessel types is consistent with the area used as a designated bunkering area for ship-to-ship transfers between bunker barges and supply vessels, and those vessels receiving the goods.

The period of the year that had the highest number of anchored vessels was between October and March (the winter months), with a combined total of 686 vessels. November was the month that had the highest number of anchored vessels (166), and this was followed by February (145) and January (131). GIS mapping of vessel positions and sizes, based on Gross Tonnage (GT) over six-month intervals from July 2016 and June 2019, revealed that most vessels anchoring in the southern region of the bunkering area were small, with GTs ranging between 0 and 10,000 (Figure 6 and Figure 7). Consequently, these small vessels were frequently anchored in areas where P. oceanica is present. In contrast, vessels with higher GTs were predominantly found anchoring towards the northern, southern and seaward outer margins of the bunkering area, where P. oceanica was not present.

Inter-annual differences are also evident when comparing the GIS data over six-month periods. In July–December 2016 (Figure 6a), most of the larger vessels occupied the seaward outer margin of the bunkering area. Over subsequent years (Figure 7a and Figure 8a), vessels moved further to the outer margins. It is worth noting that in 2016 anchoring took place in what is now a designated aquaculture zone. From January to June (Figure 6b, Figure 7b and Figure 8b) vessels progressively shifted from the seaward outer margins of the bunkering area to the northern and southern areas of the bunkering area. Between January and June 2017 (Figure 6b), anchoring again occurred near the designated aquaculture zone.

4.2. Shipping Index

All the GT Groups displayed the highest concentrations of total residence time, total tonnage and the total number of vessels to the south of the bunkering area. GT Group 2 registered the highest total residence time in term of hours (14,256), the total tonnage in terms of GT (11,903,029) and the total number of vessels (2308). With regards to anchoring pressure, pseudocolor plots show that anchoring pressure is mainly focused along the southern margin of the bunkering area and is mostly derived from those vessels in GT Groups 2–5 (Figure 9a–d). Only GT Groups 3 and 4, particularly displayed high concentrations to the north of the bunkering area in the total tonnage plot (Figure 9b,c).

4.3. Comparison of ROV Damage Survey with AIS Data

Spatial comparisons with GIS mapping between the ROV damage survey and AIS data produced two observations. The first ROV survey, covering the period between July and December 2017, showed chain scour to P. oceanica meadows (Figure 10a). A second data match between the ROV footage and AIS was for the period January and June 2018, in which evidence of bedrock breakage from anchoring damage was observed (Figure 10b).

5. Discussions

These results confirm how Bunkering Area 1 is primarily used during the winter months, when winds and currents are stronger and hence more impactful upon bunkering operations. This is especially significant given that Malta’s prevailing wind is from the northwest [38].

The high concentration of vessels found in the south of the bunkering area could be for the same reason, i.e., deep waters, with the seabed characterized by medium sand [30]. These physical conditions provide a relatively flat and unobstructed seafloor, allowing for an anchor to be deployed without any risk of becoming entangled or ensnared [39]. Soft bottoms also recover rapidly when anchor damage occurs. Additionally, muddy or sandy substrates provide a better anchor hold because an anchor can be easily embedded in these substrates [39].

In contrast, rocky substrates are regarded as weak holding grounds, and in strong weather conditions, anchors often lose their hold and drag across the seafloor, causing damage [39]. Vessels may also be anchoring in these locations because the island of Gozo and the northern tip of mainland Malta may provide additional shelter from strong weather conditions (winds and currents) as a result of land shadowing. In the case of storm events, land shadowing creates relatively calmer seas than the rest of the bunkering area. This is evident in Figure 7b and Figure 8a, which display high concentrations of vessels to the north and the south of the bunkering area.

Plotting results confirm that even though anchoring has occurred throughout the bunkering area, the highest concentration of anchored vessels lies within the southern extent of the bunkering area, whilst anchoring within the middle sections of the bunkering area is sparse. These sections coincide with the location of Is-Sikka l-Bajda, where water depths are shallower than the rest of the bunkering area and, thus, do not represent an ideal anchorage area.

All the GT Groups displayed the highest concentrations of total residence time, total tonnage and the total number of vessels to the south of the bunkering area. GT Group 2 registered the highest total residence time in terms of hours (14,256), the total tonnage in terms of GT (11,903,029) and the total number of vessels (2308). With regards to anchoring pressure, pseudocolor plots show that anchoring pressure is mainly focused along the southern margin of the bunkering area and is mostly derived from those vessels in GT Groups 2–5 (Figure 9a–d). Only GT Groups 3 and 4, particularly displayed high concentrations to the north of the bunkering area in the total tonnage plot (Figure 9b,c).

A comparison of these plots reveals that the anchoring pressure from GT Group 2 was not as intense as shown in the other plots. The anchoring pressure from this GT Group extends across the whole of the bunkering area, except for where Is-Sikka l-Bajda is located and some other small patches along the northern margin of the bunkering area. Only GT Groups 6, 7 and 8 display some anchoring pressure to the north of the bunkering area. These findings confirm that the southern part of the bunkering area is heavily used as an anchorage zone, experiencing significant anchoring pressure from shipping activities. Pseudocolor plots reveal that this pressure primarily comes from smaller vessels, ranging between 500 and 30,000 GTs.

The southern part of the bunkering area features by deep, coastal waters and medium sand substrates, making it a popular anchoring area. The unobstructed seafloor allows anchors to embed easily, providing good holding grounds. Similarly, the northern part of the bunkering area also experiences anchoring pressure from vessels across most GT Groups, likely due to comparable water depths and benthic substrates to those in the southern bunkering area.

Pseudocolor plots indicate that the middle of the bunkering area, where Is-Sikka l-Bajda is located, experiences less anchoring pressure than initially speculated. However, ROV surveys in this study have revealed damage to the existing reef in the area. This is consonant with past news articles that had reported anchor and chain damage to the seafloor and P. oceanica meadows in the Is-Sikka l-Bajda region [40,41]. Demicoli (2017) does however also state that the Maltese competent authority responsible for protecting, conserving, and preserving the marine environment is working on management plans to mitigate the damage caused by these shipping activities [40].

During this study period, two vessels were located near the ROV track, causing anchor and chain scour damage. In the first data match for this period, no damage was captured, possibly due to the ROV’s camera direction. In the second match, anchor chain scour damage was visible as a straight line on the seafloor (Figure 10c). No overlaps of data were encountered for the months ranging between July 2018 and June 2019.

5.1. Limitations and Resolutions

A few operational limitations were encountered. Since the conducted surveys were linear in nature, coverage of the area was limited to the path taken by the vessel; initially, the surveys were meant to follow pre-determined transects. This did not happen while out in the field for various reasons, such as time and funding constraints and technical issues such as power shortages and fluctuating weather conditions.

This study was not granted permission to directly access the MySQL AIS database so as to ensure that the integrity of the database is maintained. Instead, the information needed was downloaded from the database in ‘static’ form as Microsoft Excel files. When it came to treating the AIS data so as to remove multiple AIS entries for the same MMSI number and with the same timestamp, this proved to be a difficult and time-consuming operation, particularly when it came to treating those AIS points for bunker barges and supply vessels. This study processed the data downloaded from the MySQL AIS database, and although it is highly unlikely, it could be that the master of a vessel may have switched off the AIS before switching the status of the AIS to “At Anchor”. Therefore, this was an error that could not be mitigated within the current study.

Another issue was that there was a lot of missing data when it came to the table containing the vessel descriptors/metrics. The acquisition of any missing or additional data (e.g., ship type and gross tonnage) proved to be laborious and time-consuming; each vessel had to be searched for manually on the MarineTraffic webpage. Consequently, the ship type was not factored within the computation of this arbitrary index. Additionally, some vessels had changed their MMSI number from the time the vessel was picked up in the Maltese territorial waters and searched for on the MarineTraffic webpage. The IMO number, which was available for most vessels in the table downloaded from the MySQL AIS database, had to be referenced to ensure that the new MMSI number listed in the webpage referred to the vessel with the old MMSI number. Once again, this exercise proved to be laborious and time-consuming. Also, some vessels had no information tied to them, and unfortunately, these vessels had to be eliminated since they could not be used in this study.

5.2. Recommendations for Future Research on Anchor Pressure

This research succeeded in showing how AIS data can be used to semi-quantify anchoring pressure within an area of interest. Even though the protocols and approaches followed need to be refined and optimized, this study can serve as a basis for other future studies. The multi-modal nature of its methods provided the ideal opportunity to apply and synergize various techniques and datasets in different formats. Future research about anchor pressure may benefit from the following recommendations for improvement as follows:

• Equip the ROV with an underwater acoustic positioning system (USBL).This would directly provide more accurate location tracking and would also overcome the limitation of relying on third-party digital tools to detect ROV’s tentative position in the water. The ROV model deployed in the current study lacked a USBL;
• Maintain a constant ROV depth and ensure the camera is consistently oriented downward. This would facilitate the effective application of machine learning techniques for classifying P. oceanica meadows and for quantifying anchoring impacts on the seabed;
• Direct access to the MySQL AIS database. This would streamline processing, eliminating the need for static Excel files and reducing manual corrections;
• Integrate more comprehensive data sources and collaborate with maritime databases. This would address missing vessel information and improve data completeness;
• Refine the methodology for computing the Shipping Index. This would include using precise measurements for anchored vessel times, incorporating ship type considerations and improving script accuracy in order to produce more accurate and meaningful results.

6. Conclusions

The ecological impacts arising from anchoring and mooring activities on marine benthic ecosystems are not exclusively attributed to the shipping industry, even though the current study focuses exclusively on this economic sector, with additional related ‘contributions’ also being ascribed to recreational vessels and to anchored aquaculture industry facilities.

Through the analysis of GIS maps and AIS data, this study determined that vessels mainly anchor along the southern margins of Bunkering Area 1, with smaller vessels being the main contributors to anchoring pressure within this area as indicated by the Shipping Index values. The area along the southern margins of the bunkering area is characterized by deep, coastal waters having a medium sand substrate. This implies that vessels can anchor in this area without fear of losing their anchor. Also, this study determined that small clusters of a high concentration of vessels are found along the northern margins of the bunkering area.

Moreover, the results of this study indicate that the central part of the bunkering area is not under as much anchoring pressure as this study had initially hypothesized. However, this does not mean that anchoring pressure is completely absent in this area, especially since the GIS maps, AIS data and Shipping Index show that some localized anchoring pressure has occurred in the past. From the AIS data, it resulted that anchoring in Bunkering Area 1 mostly occurs during the winter months, when the winds and currents are much stronger. The main vessel types that visited the bunkering area during the period this study focused on were general cargo, oil products tankers, bulk carriers and oil/chemical products tankers.

Producing a detailed map of Is-Sikka l-Bajda would aid port operators and vessel masters in better determining which are the best areas for vessels to anchor. If possible, future anchoring would avoid completely Is-Sikka l-Bajda and areas where P. oceanica meadows are present. Ideally, vessels should anchor in deeper waters, and if this is not possible, a designated area should be established that is solely used as an anchorage area to minimize further damage to the area. Understandably, this may be difficult to implement since anchored vessels must retain a safe swinging distance between them.

A better understanding of the seafloor beneath those areas that displayed a high concentration of vessels, notably those areas along the northern and southern margins of the bunkering area, would enable managers and policy-makers to mitigate the anchoring pressure to these areas, especially if anchoring is occurring over sensitive benthic environments.

A number of studies e.g., [42,43] consider that MSP provides an opportunity to develop anchoring management alternatives that support local authorities in identifying, comparing, ranking, and selecting the most suitable anchoring locations. According to these studies, such a selection should consider the “optimal use of the maritime space”, taking into account the socio-economic incentives, as well as the protection of the marine environment [44].

Byrnes and Dunn (2020) list a comprehensive list of management measures that could be adopted in order to mitigate the ecological pressures posed by boating and shipping operations within the marine environment [45]. The measures they proposed, specifically to counter the impacts arising from anchoring and mooring (i.e., designated anchorages, adoption of environmentally-friendly moorings, zoning plans, regulations, community education, and restoration) are mainly MSP-oriented.

This study hopes to serve as a case study that can be used and integrated into decision-making techniques by managers and policy-makers. Moreover, the Maltese authorities must work hand in hand to ensure that the marine environment is safeguarded while allowing maritime activities in the Maltese territorial waters to flourish.