First assessments on marine geomorphology and bio-colonization through the study and survey of the wreck of the Hospital Ship “Po” along the Albanian coasts (Vlora Bay)

Vlora bay (Albania) hosts the 134.98 m long wreck of the Hospital Ship “Po” (sunk in 1941), located within the Karaburun-Sazan Marine Protected Area. This study investigates the interactions between this anthropogenic structure and the surrounding marine environment through a multidisciplinary observational approach. Resting at 33–37 m depth on a gently inclined shelf, the wreck provides a hard substrate that supports benthic colonization. High-resolution geophysical surveys (MBES, SSS) and Structure-from-Motion (SfM) photogrammetry, integrated with biological and sedimentological censuses, were used to develop two preliminary quantitative metrics: the Geometric-Geomorphological Wreck Effect Index (IGEOM) and the Ecological Hotspot Value Index (IECO). Geomorphological observations reveal a 2.1 m deep upstream scour hollow and a 45 m downstream depositional tail. Sedimentological data indicate a local granulometric gradient, with fine silt-clay fractions (< 63 μm) found near the wreck, contrasting with a medium sand baseline (350 μm). These features are consistent with deposition under regional sediment-supply conditions influenced by the Vjosa River (from NW), though direct sediment-source attribution requires further investigation. Ecologically, the site exhibits high benthic diversity (H' = 2.91) and structural complexity. Findings include the presence of the non-indigenous Pterois miles, evidence of Ommastrephidae spawning and a sighting of the critically endangered Monachus monachus, generating the hypothesis that the wreck may occasionally function as a local feeding area. The proposed indices (IGEOM and IECO) offer preliminary, site-specific descriptors summarizing observational data, with potential applicability to future monitoring and Marine Spatial Planning (MSP) frameworks, pending multi-site validation.

Vlora bay, situated along the Albanian coast at the transition zone where the southern Adriatic Sea meets the northern Ionian Sea, represents a site of paramount ecological and strategic importance. The bay is characterized by a high diversity of marine habitats, significantly influenced by its proximity to the Vjosa River, one of Europe's last remaining undammed river systems, which drives complex sedimentological and hydrological dynamics within the basin (Modugno et al., 2019, 2021; Pano et al., 2006). This unique environmental setting led to the establishment of the Karaburun-Sazan Marine Protected Area (MPA) in 2010, Albania's first and largest MPA, recognized for its exceptional biodiversity and conservation value (Maiorano et al., 2011; Kashta et al., 2011). Within this MPA and its contiguous waters lies a significant concentration of submerged cultural heritage, primarily consisting of shipwrecks from World War II (Meçollari, 2022). Among these, the Italian Hospital Ship “Po” (Figure 1), originally a passenger vessel than during war operations converted into an Italian hospital military navy, holds a prominent position. The “Po” was sunk on March 14, 1941, during the Greek-Italian War after being struck by an aerial torpedo from a British Swordfish bomber, while engaged in the evacuation of wounded soldiers (Meçollari, 2022).

The withdrawal of Italian forces began in December 1940 after being defeated on the Greek front. It included the withdrawal of its wounded, mainly through the field hospital in Radhima, Vlore. To evacuate the wounded soldiers the Italian Royal Navy Hospital Ship “Po” arrived in Vlora bay. They were attempting to transport them from Vlora to Italy in March 1941. Hospital Ship “Po” dropped the anchor 5 cables (1000 m) off the coast of Radhima (Viroi Cape) and the wounded soldiers were planning to embark the next day. On board the Hospital Ship “Po” was a medical team of the Italian Red Cross. Among them was Edda Ciano, the wife of the Italian Minister of Foreign Affairs, Count Galeazzo Ciano who was also the daughter of Italian Prime Minister Benito Musolini. British forces successfully attacked the Taranto Naval Base in November 1940 which allowed them to deploy the 815-th “Swordfish” aircraft Squadron in Paramithi and “Blenheims” of the 211th aircraft Squadron in Corfu. The attack on the Italian Royal Navy Hospital Ship “Po” was the first combat action in Vlora bay by the British Air Force. On the evening of 14 March 1941 in the airspace of Vlora bay there appeared five “Swordfish” aircraft of the 815-th Squadron. Prior to this action the aircraft “Blenheims” of the 211th Squadron took off from Corfu and conducted a bombardment in Vlora Port to distract the Anti-Air Italian Artillery. The night of 14 April was illuminated by a full moon. That same night, by order of the Navy's Headquarter, Hospital Ship “Po” switched all her lights on board the ship off so as not to be recognized by British aircrafts. By 23:00 most of the crew was in their cabins, while the Red Cross Members were resting. A few minutes after 23:00 several aircraft were identified approaching from the direction of Gjuheza Cape in formation. Later an explosion occurred. It was 23:13 when a torpedo hit “Po” in astern on the starboard side. Three Red Cross female members in panic threw themselves into the sea. The ship started to tilt and the hatch in the astern began to take water. Many sailors were trapped. The ship sank in 10 min. The British “Swordfish” squadron was under the command of LT Michael Torrens-Spence. Hospital Ship “Po” sank at a depth of about 37 m. On board that night were 240 people, crew and medical personnel. From the sinking of Hospital Ship “Po” 23 people died, among them three members of the Red Cross. Edda Ciano survived. The same night the British “Swordfish attacked with torpedoes and sank the Italian Support Ship “Santa Maria”, which was later recovered and re-entered service. The switching off lights of Hospital Ship “Po” on the night of March 14, 1941 and her sinking does not represent a crime under the International Humanitarian Law. A ships protected status, while hiding herself or being escorted by combat ships, loses the status of protection under the Hague Conventions of 1907.

Beyond their historical and cultural significance as “sites of memory,” such wrecks function as artificial structures that can influence the local marine environment (Macreadie et al., 2011; Perkol-Finkel and Benayahu, 2005; Ambrose and Anderson, 1990; Riggio et al., 2000). It is widely established that artificial reefs, including shipwrecks and decommissioned offshore platforms, enhance the structural complexity of sandy and flat marine habitats, providing important refuge and supporting secondary marine production (Baine, 2001; Feary et al., 2011; Ambrose and Anderson, 1990; Riggio et al., 2000) thanks to the presence of some new vertical structures and possibility of refuges and caves provided by the metal structures of the wreck. Research in the Mediterranean basin indicates that artificial submerged structures are rapidly colonized by benthic assemblages, with community composition and succession rates dictated by environmental forcing such as depth, light availability, orientation relative to currents and local hydrodynamic regimes (Perkol-Finkel et al., 2006). Furthermore, shipwrecks may act as “stepping stones” or primary nodes for the expansion of Non-Indigenous Species (NIS), particularly those migrating from the Red Sea via the Suez Canal (Rilov and Galil, 2009; Galil et al., 2018). Despite their undeniable environmental and historical value, shipwrecks along the Albanian coast remain poorly investigated till nowadays. Quantitative geomorphological and biological data are currently scarce (Savini et al., 2011; Coll et al., 2010), which limits the effective integration of underwater cultural heritage into broader Marine Spatial Planning (MSP) frameworks. This knowledge gap is particularly critical in regions experiencing rapid coastal development and climate-driven environmental stressors (Dailianis et al., 2018; Modugno et al., 2019, 2021). This important World War II wreck, as many other shipwrecks in Vlora bay, provides a unique opportunity to bridge these gaps through a multidisciplinary observational approach that links underwater cultural resources with geomorphology and marine ecology. The wreck of the Hospital Ship “Po” represents a relevant site within MSP considerations in Vlora bay due to its accessibility and potential for sustainable underwater tourism.

While previous fundamental works (e.g., Savini et al., 2011; Maiorano et al., 2011) provided a robust regional baseline characterizing the soft-bottom benthic fauna and the geomorphology of the wider Vlora Bay, no historical biological or quantitative data existed for the shipwreck itself. Therefore, rather than attempting to validate pre-existing biological data of the wreck, this study utilized the regional literature as a conceptual starting point to design a dedicated 2-year monitoring program (June 2022–September 2024). This effort aimed to establish the very first ecological and geomorphological baseline of the “Po” wreck acting as an artificial reef. The systematic survey effort included: a) 12 scientific dives dedicated to benthic taxonomic censuses (revealing >80% biogenic coverage used to compute the IECO index); b) 20 dives deploying divers and ROVs for sub-centimetric 3D photogrammetry; c) 12 targeted sediment samples (6 proximal, S1–S6; and 6 far-field reference sites, C1–C6), where 12 samples were also collected for sorting analysis for meio- and macrobenthos (Van Veen Grab).; and d) 12 visual census surveys, by both divers both ROV, that cataloged 28 fish species living near the wreck. e) Furthermore, this monitoring led to the discovery of exceptional occurrences, including the non-indigenous Pterois miles, evidence of Ommastrephidae spawning, and a sighting of the critically endangered Monachus monachus.

This study, which builds upon preliminary assessment work (UNEP-MAP, 2019, 2020; Modugno et al., 2019, 2021), represents a collaborative effort between Italian and Albanian institutions and CMAS diving associations (Italian FIPSAS-CMAS CICF “Il Blu del Mare” and Albanian CMAS Underwater Diving Federation).

The research aims to: (1) provide an observational description of the interaction between the “Po” wreck and the seafloor, particularly regarding scour and depositional patterns; (2) document the biodiversity and benthic colonization patterns populating the wreck; (3) evaluate the role of the wreck as a refuge/feeding area for endangered species or a gateway for invasive species; and (4) propose a methodological approach for the survey and index-based quantification of shipwrecks in the Adriatic-Ionian region. By integrating high-resolution geophysical surveys (MBES and SSS) from previous scientific research on site (Savini et al., 2011), sedimentological observations and the calculation of quantitative biological indices, this project seeks to provide an initial conceptual framework for describing the “wreck-effect.” This approach considers shipwrecks not merely as historical artifacts but as non-living structures that gradually become living, as they become encapsulated over time by marine organisms and sediments that may influence local geomorphological processes and act as ecological indicators within the Mediterranean coastal zone.

Study site

Vlora bay is the largest bay in Albania (Figure 2), situated along the middle of the Albanian coastline and in the very southern part of the Adriatic Sea. The high altitude of the surrounding terrain allows the maritime surveillance of SLOCs from Mid Adriatic Sea through Central Mediterranean, while the accommodation capacities of the Vlora bay allow immediate response in Otranto Chanel (the narrow sea channel that separates the Sazan Island from Cape Gjuhëzës, the tip of the Karaburun Peninsula), Adriatic Sea and Central Mediterranean. The Vlora bay is a key strategic position supporting maritime activity in both the East and West Mediterranean Sea. Vlora bay has a surface of 315 km 2, extending from the North-West toward the South-East. The southern part of Vlora bay is called bay of Dukat. The bay is connected to the Adriatic Sea through two entrances, divided by Sazan Island, called Northern and Southern Channels. The bay of Vlora is one of the densest maritime areas containing shipwrecks in Albania, followed by Shengjin and the Durres area. Vlora bay, due to its history, geography, depth, etc., is also one of the most attractive tourist destinations in Albania. Over the course of two World Wars, including in later years, around 45 ships were sunk in the bay of Vlora and surrounding area, 27 of them inside the bay. Most of the shipwrecks are under 50 m depth.

The Hospital Ship “Po” wreck is situated at 40°2′53.0 ′′ N, 19°7′54.2 ′′ E (Meçollari, 2011), roughly 1.2 km from the Radhimë coast in Vlora bay. This event, which resulted in other many fatalities, marks the site as a significant element of underwater cultural heritage and a primary submerged resource within the bay (Meçollari, 2022). The wreck lies at a depth of 33–37 m on a gently inclined shelf floor (slope < 1°) characterized by a relatively homogeneous depositional environment (Savini et al., 2011). This area is influenced by terrigenous inputs associated with the Vjosa River delta, located approximately 15 km to the north-west. Hydro-dynamically, the site is located within an area influenced by cyclonic circulation patterns. Currents at the site, moving in a NW-SE direction, typically range from 0.1 to 0.25 m/s (0.2–0.5 knots) (Savini et al., 2011). These currents are influenced by the bay's micro-tidal regime, which exhibits an average tidal range of 20–30 cm (Pano et al., 2006). Significant modifications to the hydrodynamic regime occur during flood events of the Vjosa River, which has an average discharge of about 195 m 3 /s (Pano et al., 2006). Such events increase turbidity and sediment transport, potentially influencing the interaction between the wreck and the surrounding seafloor (Celik et al., 2010). Water temperature shows seasonal variability between winter and summer conditions, as reported in regional studies. Visibility generally varies depending on seasonal conditions and sediment input, with higher values observed during periods of reduced turbidity (Meçollari, 2022).

Geophysical surveys and geomorphological investigation

Detailed seafloor maps were produced (thanks to previous project, Savini et al., 2011) using a Teledyne RESON SeaBat 7125 multibeam echosounder (MBES; 400 kHz) and an Edgetech 4200 side-scan sonar (SSS; 300/600 kHz dual frequency), both mounted on a small vessel (Savini et al., 2011; Supplementary material Annexes 1, 2). MBES data were processed with QPS Qimera and Fledermaus to generate a 0.5 m-resolution bathymetric grid. To enhance the visualization of morphological features around the wreck, digital terrain models (DTMs) were further analyzed through shaded-relief and slope visualizations (Savini et al., 2011; Supplementary material Annexes 1, 2). The SSS mosaics were georeferenced and processed to emphasize backscatter anomalies, which were used to identify the wreck's orientation and to support the interpretation of possible scour-related and depositional features around the structure (Savini et al., 2011; Supplementary material Annexes 1, 2). To distinguish morphological features potentially associated with the wreck from background seabed variability, the investigation was framed within the known geomorphological setting of Vlora bay. The site's location on a gently inclined shelf floor (slope < 1°) was used as a geomorphological reference context for assessing whether the observed features differed from the expected background seabed conditions (Savini et al., 2011).

A total of 12 sediment samples were collected using a Van Veen grab (0.1 m 2) and analyzed for grain size using a Malvern Mastersizer 3000. As reported in Figure 3, the targeted sampling design included 6 proximal stations distributed along the entire length of the wreck (134.98 m), with three stations on the port side (S1–S3) and three on the starboard side (S4–S6). This proximal array was integrated with 6 external far-field reference sites (collectively indicated as C1–C6 in Figure 3) located 200 m upstream to evaluate the consistency of the local granulometric distribution against the regional sedimentary background, in order to evaluate the consistency of the local granulometric distribution with the regional sedimentary influence of the Vjosa River both for grin size, both for benthic organisms. Given the relative limited number of samples, the sedimentological dataset was used to characterize a targeted localized pattern rather than to provide an exhaustive reconstruction of the wreck-induced sedimentary system.

Biological and photogrammetry diving surveys

A total of 12 sediment samples, took at the same time of the 12 sediment withdrawals, were collected for the characterization of macro-benthic fauna (Table 1; Supplementary material Annex 7), all samples were sieved through multiple mesh sizes and the sampled fauna was sorted and identified to the species level under stereo microscope (sorting), when possible, following the European Register of Marine Species and updated Mediterranean checklists A sampling plan was defined based on the preliminary analysis, involving 12 sampling stations (S1–S6 and C1–C6, Figure 3). Samples were collected using a Van Veen grab with a sampling surface of 0.1 m 2 and a volume of 0.02 m 3. Three replicates were performed at each station at depths ranging from 33 to maximum 54 meters. All samples were strained through multiple sieves with mesh sizes of 10 mm, 5 mm, 1 mm, and 0.45 mm. Benthic fauna was sorted and identified to the species level when possible, covering groups such as Foraminifera, Cnidaria, Nemertea, Sipuncula, Mollusca, Annelida, Crustacea Decapoda, Bryozoa, Echinodermata, Hemichordata, and Tunicata (Table 1).

From June 2022 to September 2024, a total of 32 scientific SCUBA dives were conducted, organized into seasonal summer campaigns (e.g., August 2022, July 2023, and June–September 2024). These included 12 dives dedicated to biological visual censuses (made both by underwater scientific researchers and by ROVs) and 20 dives focused on photogrammetric acquisition by divers and ROV (Chasing Gladius Mini S) deployment (Figures 4, 5c). To maximize bottom time and safety for divers, operations adhered strictly to CMAS and HSE scientific diving standards, utilizing Nitrox EAN32 (Enriched air nitrox with the following binary composition: 32% O 2 68% nitrogen; with a maximum operating depth (MOD) of 33.7 meters, rounded to 33 meters to be conservative, calculated on a maximum partial pressure of oxygen pO 2 of 1.4 bar) breathing gas and planned decompression procedures (Figures 4, 5e, 6b).

Dives for 3d shooting lasted up to 4 h each, targeting the wreck's upper structures and main decks at a maximum operational depth of 28 m (while the hull bottom rests at 33–37 m). During scientific dives (total diving runtime of about 2 h each), benthic cover was quantified using a random bionomic square design (Figure 5d). The structure (Figures 5a, b) was divided a priori into three main sections (Bow, Midship, Stern) (Figure 4), with n = 20 random bionomic squares (0.25 m 2) sampled per section (Figures 4, 5d). To provide a baseline, n = 20 bionomic squares were also sampled on the adjacent surrounding seabed (total N = 80). Furthermore, surfaces were classified as “exposed” or “sheltered” based on their orientation relative to the dominant NW-SE hydrodynamic flow. For 3D modeling, over 2.000 high-resolution images per dive-survey were processed (Figures 6c–e) using both Metashape (Agisoft) and Reality Capture (Unreal Engine) SfM softwares. Sub-centimetric accuracy was achieved by placing calibrated metric scale bars (Figure 5d) on the wreck prior to image acquisition.

Organisms within benthic squares were identified to the lowest possible taxonomic level (Table 2 and Supplementary material Annex 7) using WoRMS and regional Mediterranean guides (Ballesteros et al., 2007; Coll et al., 2010). Pelagic and demersal species were recorded via scientific divers observations, through the application of the Underwater Visual Census (UVC) protocol (Brock, 1954) and Chasing Gladius Mini S video transects (Figures 4, 5c) (50 m linear distance from the wreck) and diver-operated stereo-cameras (Figure 5e). The presence of Pterois miles was confirmed through direct observations during dives and photo identification. The identification and the mapping of Pterois miles were done using recent data from citizen science and observations made during events such as the 2024 CMAS World Underwater Photography and Video Championships.

Data integration, index development, and statistical analysis

To integrate information on the interaction between the wreckage and the marine environment, a multi-parameter data integration approach was developed using ArcGIS Pro (v3.0) to synthesize geophysical, sedimentological, and biological observations. This approach complemented qualitative descriptions by formulating two preliminary quantitative descriptors: the Geometric-Geomorphological Wreck Effect Index (IGEOM) and the Ecological Hotspot Value Index (IECO).

The Geometric-Geomorphological Wreck Effect Index (IGEOM) was calculated by integrating Digital Terrain Model (DTM) derivatives (slope and shaded-relief) with local sedimentological parameters (Savini et al., 2011). The reasoning, mathematical formulations, and procedural steps are provided in the Supplementary material (Annexes 1, 2). The index incorporates: (i) the maximum scour depth relative to the Reference Seabed Level (RSL), (ii) the planimetric area of the interpreted downstream depositional tail, and (iii) the granulometric contrast observed between proximal fine fractions (< 63 μm) and distal medium sands (350 μm). These variables were normalized and combined to provide a site-specific summary of the morphological footprint interpreted around the wreck as a seabed obstacle. CEDAS was used as a supporting tool to examine sediment-movement scenarios and hydrodynamic conditions relevant to the interpretation of the wreck-seafloor setting.

The Ecological Hotspot Value Index (IECO) was developed to summarize the biological relevance of the site by integrating (i) benthic percentage cover, (ii) taxonomic richness (S), and (iii) the occurrence of conservation-relevant species (e.g., Monachus monachus) or non-indigenous species (e.g., Pterois miles). The IECO serves as a unitless interpretive metric that summarizes site-specific biodiversity observations without assigning formal hotspot classes or validated ecological thresholds. Biodiversity was further characterized using the Shannon-Wiener index (H ‘) and Pielou's evenness (J ') in PRIMER v7 (Shannon, 1948; Pielou, 1966). The reasoning, mathematical formulations, and procedural steps are provided in the Supplementary material (Annexes 1, 2).

Differences in assemblage composition among the four sampled habitat levels (bow, midship, stern, and adjacent seabed) were tested using a one-way PERMANOVA in PRIMER v7 with the PERMANOVA+ add-on. Before the test, species abundance data were square-root transformed to down-weight highly abundant species and a Bray-Curtis resemblance matrix was constructed. The assumption of homogeneity of multivariate dispersions was priorly verified using the PERMDISP routine. Statistical significance was assessed using 9,999 permutations of residuals under a reduced model, an approach suitable for ecological datasets. Non-metric Multi-Dimensional Scaling (nMDS) was applied to visualize spatial gradients in community composition (Figure 7).

Finally, given the small number of sediment samples (n = 12: S1–S6 near the wreck and C1–C6 far from it), exploratory correlations between median sediment grain size (D 50) and total benthic cover were assessed using Spearman's rank correlation coefficient (ρ) to examine possible eco-geomorphological associations (Supplementary material Annex 1 and Figure 3).

Geometric-geomorphological wreck effect index

To synthesize the morphological footprint interpreted around the shipwreck on the seafloor, a preliminary Geometric-Geomorphological Wreck Effect Index (IGEOM) was calculated. At this stage, the metric is intended as a site-specific descriptive tool and does not provide calibrated thresholds, cross-site comparability, or predictive capability. In this specific case, the index was derived by integrating geomorphological information extracted from geophysical data with granulometric characteristics gathered during the field surveys. The index integrates three primary components describing relative depositional extent, scour intensity and sedimentological contrast, selected because they are consistent with established concepts in fluid dynamics and sediment transport (Sumer and Fredsøe, 2002; Sumer et al., 2001; Whitehouse, 1998; Best, 2005; Dalrymple et al., 2012). Specifically, the index incorporates: (i) the Scour Intensity Factor, derived from the maximum scour depth (2.1 m) measured relative to the Reference Seabed Level (RSL); (ii) the Depositional Footprint Ratio, calculated from the planimetric extension (45 m) of the interpreted downstream depositional tail; and (iii) the Textural Gradient Variance, which expresses the contrast between proximal fine fractions (< 63 μm) and distal medium sands (350 μm). Each component was normalized to a 0–1 scale and combined to obtain a dimensionless site-specific summary value. This approach summarizes discrete bathymetric and sedimentological observations into a single interpretive value describing the local morphological expression of the wreck-effect at the study site.

The Geometric-Geomorphological Wreck Effect Index (IGEOM) is calculated as the geometric mean of three dimensionless ratios representing horizontal deposition, vertical erosion and sedimentological contrast. The formal equation is: IGEOM =((Lt / Lw)·(ds / H)·(d 50, far / d 50, lee))1/3 where Lw is the wreck length, Lt is the deposition tail length, ds is the scour depth, H is the local water depth, d 50, far is the reference median grain size, and d 50, lee is the median grain size in the wake. The geometric mean ensures equal weight (1/3 in logarithmic space) for each component and prevents compensation among unrelated geomorphological processes. The complete reasoning, mathematical formulations, and procedural steps are provided in the Supplementary material (Annexes 1, 2).

Ecological hotspot value index (IECO)

The Ecological Hotspot Value Index (IECO) was developed to summarize the biological relevance of the shipwreck through a site-specific metric for describing its role as an anthropogenic structure within the local ecological context (Macreadie et al., 2011; Ferrario et al., 2016; Myers et al., 2000; Halpern et al., 2008). The IECO integrates multiple ecological proxies into a single interpretive value, providing a structured summary of site-specific biodiversity observations related to the wreck. In this specific case, the index is designed as a site-specific descriptor integrating heterogeneous ecological observations, and its values should be interpreted within the local environmental context of Vlora bay. The IECO formulation incorporates three primary ecological components: (i) Benthic Assemblage Complexity, derived from percent cover and taxonomic richness (S) recorded across the bow, midship, and stern sections; (ii) Colonization Characteristics, describing the observed development of the sessile community on the 134.98 m hard substrate; and (iii) Conservation Relevance, accounting for the occurrence of conservation-relevant taxa, such as the critically endangered Mediterranean monk seal (Monachus monachus), alongside the presence of non-indigenous species (NIS) such as Pterois miles.

The index formulation is defined as the geometric mean of three normalized components: IECO = (D · C · K)1/3 All biological parameters were mathematically normalized to allow comparison without introducing external ecological benchmarks. Specifically, the benthic diversity component (D) is normalized against its theoretical upper bound (D = H ′/ln S); the coverage component (C) is scaled to 1 based on percentage (C = 100 × observed coverage); and the key-species component (K = k /3) is a presence-only indicator based on the count (k ε{0,1,2,3}) of the three selected taxonomic categories. The geometric mean ensures that the components contribute in a co-necessary way: if one component approaches zero, the aggregated value declines proportionally, thus preventing compensation among unrelated ecological indicators. Details on definitions, normalizations, and numerical computations can be found in the Supplementary material (Annex 2).

This study represents a a very first and initial step toward a broader multi-wreck assessment strategy within the Albanian Marine Protected Areas (MPAs), and presents a quantitative description of the Po wreck that may inform future comparative research.

Wreck geomorphology and sediment dynamics

The Hospital Ship “Po” measures 134.98 m in length and 16.24 m in width (Meçollari, 2022). It rests in an upright navigation trim on the seabed, oriented at 285° (WNW-ESE), placing its longitudinal axis nearly perpendicular to the primary NW-SE water currents. High-resolution MBES and SSS data (from also previous research investigation, Savini et al., 2011) indicate a distinct scour hollow on the upstream (NW) side, reaching a maximum depth of 2.1 m below the Reference Seabed Level (RSL). The RSL was established based on a 200 m NW baseline to provide a reference for the comparison of local seabed morphology. Conversely, on the downstream (SE) side, a sediment accumulation feature interpreted as a depositional trail extends approximately 45 m from the stern, characterized by a positive relief relative to the surrounding seabed. Side-scan sonar backscatter analysis reveals small sedimentary ridges (ripples with wavelengths λ = 0.6–1.2 m) aligned with the dominant water current. These bedforms disappear within a 5 m radius of the wreckage, suggesting that the ship may act as a local hydrodynamic obstacle influencing near-bed flow conditions.

The sedimentological analysis highlights a granulometric gradient in proximity to the wreck (Table 3). Sediment samples transitioned from medium sand (D50 = 350 μm) at a distance of about 200 m (reference sites: C1–C6) to fine silt-clay fractions (< 63 μm) immedCiately adjacent to the wreckage (S1–S6: 3 near the left side, 3 near the right side). This localized presence of fine-grained material is consistent with deposition under regional sediment-supply conditions influenced by the Vjosa River, situated approximately 15 km to the north. While the integration of bathymetric DTMs, sonar backscatter intensity, and grain size data strongly supports a regional hydrodynamic-sedimentary interpretation of spatially structured erosion and deposition, direct sediment-source attribution at the wreck scale would require further validation through time-series monitoring or sediment-trap data (Pano et al., 2006).

Benthic assemblages and seabed life patterns

Quantitative analysis of bionomic square samples and of the 3D photogrammetric models indicates that over 80% of the shipwreck's hard surface is covered by dense biological assemblages (Underwood and Chapman, 2006; Ferrario et al., 2016). The benthic community is dominated by a multi-layered biogenic crust, whose main representative taxa are listed in Table 1. Sponges such as Crambe crambe and Spirastrella cunctatrix, sea squirts like Botryllus schlosseri and calcified bryozoans like Myriapora truncata and Reteporella septentrionalis represent the dominant functional groups.

The taxonomic checklist of the benthic and sessile organisms colonizing the artificial hard substrates of the Hospital Ship “Po” wreck comprises a total of 54 identified taxa. Preliminary census data from the total scientific dives and 2 of the marine sessile and motile macro-organisms recorded on the artificial hard substrates of the “Po” shipwreck. The table displays the total number of identified species (n = 54) and the estimated mean percentage of coverage (% covering) partitioned by Kingdom and Phylum, highlighting a cumulative substrate coverage of 80.00%. This specific biocoenosis forms the dense, multi-layered biogenic crust responsible for the >80% epibenthic coverage recorded during the bionomic square surveys, representing the core biological dataset utilized for the calculation of the Ecological Hotspot Value Index (IECO). The wreck-associated community is taxonomically distributed across two Kingdoms and eleven Phyla. The Kingdom Plantae/Algae is represented by 5 taxa (Rhodophyta: 3; Chlorophyta: 2), primarily consisting of photophilic algae and turf-forming species that predominantly colonize the horizontal decks and upper structures exposed to light. The Kingdom Animalia heavily dominates the structural complexity with 49 taxa, highlighting the wreck's role as a localized biodiversity hotspot. The faunal assemblage is primarily characterized by Porifera (9 species, including encrusting sponges such as Crambe crambe and Spirastrella cunctatrix), Cnidaria (7 species), and Mollusca (7 species, including structure-forming bivalves such as Mytilus galloprovincialis and Ostrea edulis, mainly concentrated on vertical surfaces). Other significant contributors to the three-dimensional complexity of the artificial reef include Tunicata (6 species, e.g., Botryllus schlosseri), calcified Bryozoa (5 species, e.g., Myriapora truncata and Reteporella septentrionalis), Annelida-Polychaeta (5 species), Echinodermata (5 species), Crustacea (4 species), and Foraminifera (1 species). This specialized hard-substrate community is ecologically distinct from the surrounding soft-bottom baseline and demonstrates the advanced stage of ecological succession achieved by the shipwreck over eight decades of submersion. The distribution of these taxa shows a spatial zonation associated with the wreck's geometry: Mediterranean mussels (Mytilus galloprovincialis) and European flat oyster (Ostrea edulis) are primarily concentrated on vertical surfaces and upper hull sections, while photophilic algae and turf-forming species are more frequently observed on horizontal decks (Figures 5a, b, d, 6a, b).

Multivariate statistical tests confirmed that the four investigated habitat levels hosted significantly different species assemblages (one-way PERMANOVA, pseudo- F 3, 76 = 4.87, p < 0.001). The assumption of homogeneity of multivariate dispersions was met (PERMDISP, p > 0.05). Pair-wise PERMANOVA comparisons indicated that the bow section exhibited significantly different and more structurally complex assemblages compared to both the midship and stern sections (p < 0.01 for all pairwise tests against the bow). The nMDS (Figure 7) ordination plot visually confirmed this separation, showing distinct clustering of samples by wreck section (2D stress = 0.21, a value considered acceptable given the high number of samples and the ecological complexity of the dataset).

Sections exposed to the NW-oriented flow show higher species diversity (Shannon-Wiener Index H ′ = 2.91) compared to more sheltered areas (H ′ = 2.14) (Shannon, 1948; Pielou, 1966; Glasby and Connell, 2001; Moschella et al., 2005) (Table 2). Vertical and current-exposed areas are associated with a higher occurrence of filter-feeding organisms and with more diverse communities compared to horizontal surfaces. This ecological heterogeneity appears to be associated with the geomorphological features described above.

Exploratory statistics indicated a strong rank correlation between sediment texture and biological cover, generating the hypothesis of an eco-geomorphological association (Spearman's ρ = −0.83, p = 0.041, n = 6), indicating that higher biological cover was generally associated with smaller D 50 values (finer sediments) observed in the wake of the wreck.

(Note: The complete taxonomic checklist of the 151 surveyed benthic taxa on seabed is provided in Supplementary material Annex 7 and the complete list of the 54 surveyed taxa on wreck's structures is provided in Supplementary material Annex 7 too).

Open-water and bottom-benthic species

Systematic underwater surveys (Underwater Scientific Dives performed by Scientific Divers) and ROV video transects (Figures 4, 5c) identified a total of 28 fish species belonging to 14 families (Table 4). Juvenile individuals of Diplodus vulgaris, Serranus scriba, and Mullus surmuletus were observed in association with the wreck structure. The distribution of these species appears to be associated with local sediment variability around the wreck; in particular, Mullus surmuletus was predominantly observed foraging within fine-grained sediments (< 63 μm) in the area interpreted as the depositional tail, compared to observations on medium sands in the surrounding seabed.

Photographic sightings by a Citizen science project by CMAS Albania during the 2024 World Championship of underwater photo and video (Figure 8) and a preliminary census in the fishing port area (conducted by CMAS Albanian Diving Federation and INCA Albania) confirmed the presence of Pterois miles, with up to seven individuals observed simultaneously, representing one of the northernmost reported occurrences of the species in the Adriatic Sea (Katsanevakis et al., 2020a,b; Rilov and Galil, 2009).

In addition to mobile fauna, reproductive-related observations were recorded in association with the wreck. Large gelatinous egg clusters of Ommastrephidae likely Illex coindetii (Figure 9) were documented adhering to the wreck's vertical surfaces at depths of 22–25 m (Figure 9). The occurrence of these egg clusters on vertical surfaces may be associated with local hydrodynamic conditions and surface orientation, although the underlying processes were not directly measured (Boyle and Rodhouse, 2005). A summary of the key taxonomic indicators documented at the site is provided in Table 5.

Notably, a transient individual of the critically endangered Monachus monachus was captured on camera on August 12, 2023, resting within the sheltered cavities of the wreck's upper structure. This represents the first confirmed sighting in Vlora bay since 1996 (Figure 10a). The first real photographic sighting dates back to 2020–2021 thanks to the photographs of Dr. S. Modugno (Figure 10a) and the subsequent (2022–2024) reports by the Karaburun-Sazan MPA monitoring manager Nexhip Hysolakoj, NAPA, UNDP (Figures 10b, c).

These observational records suggest that the Po wreck is utilized by a variety of species, contributing to local habitat heterogeneity. The data generate the hypothesis that the structure may provide an occasional nursery area for coastal fish, support cephalopod reproduction, and function as a temporary resting site for severely endangered marine mammals, pending further long-term behavioral studies.

Index-based quantification of wreck effects

To summarize geomorphological characteristics observed around the vessel, the Geometric-Geomorphological Wreck Effect Index (IGEOM, Table 6) was derived using field parameters, including wreck dimensions, local bathymetry, scour depth, depositional-tail geometry, and granulometric variability. Using the empirical values reported in the primary dataset (wreck length 134.98 m, water depth range of 33–37 m, scour depth 2.1 m, deposition tail length 45 m, median grain size D50 = 350 μm at the reference site, and D50 < 63 μm in the interpreted depositional area), the three dimensionless components of the index were computed and combined using a geometric mean (Sumer and Fredsøe, 2002; Sumer et al., 2001; Whitehouse, 1998; Best, 2005; Dalrymple et al., 2012). The resulting index value of 0.52 represents a site-specific summary of the observed geomorphological characteristics and should not be interpreted as a calibrated measure of seabed modification or as a basis for cross-site comparability. We specify that the value of 0.52 was obtained using a reference observation depth (H) of 28 m (2.1 / 28 = 0.075), corresponding to the upper deck within the total depth range of 33–37 m. As discussed in Supplementary material Annexes 2, 3, this value provides a unit less synthesis of the measured scour depth, depositional extent, and textural differences, and should be interpreted as a descriptive, site-specific metric.

The Ecological Hotspot Value Index (IECO, Table 6) was employed to summarize ecological observations associated with the wreck function integrating three primary biological components: a normalized Shannon diversity score, benthic percent cover, and the occurrence of selected indicator species (specifically Pterois miles, Ommastrephidae egg masses, and Monachus monachus) (Shannon, 1948; Pielou, 1966; Underwood and Chapman, 2006; Ferrario et al., 2016). Given the observed seafloor and organism data (taxonomic richness with H ′ = 2.91 in exposed locations, epibenthic coverage exceeding 80% on structural surfaces, and the detection of three high-value biological indicators), the IECO value is calculated at 0.83 (Myers et al., 2000; Halpern et al., 2008). As detailed in Supplementary material Annex 2, this value represents a numerical summary of benthic diversity, observed colonization patterns, and the co-occurrence of selected biological indicators. This information contributes to the interpretation of the site's ecological characteristics and is integrated within the descriptive framework adopted in this study. Both indices are derived from the geomorphological, sedimentological, and ecological data collected in this study and should be interpreted as site-specific descriptors that may inform future comparative analyses. Their complete mathematical definitions and computational steps are systematically archived in Supplementary material Annexes 1, 2.

Artificial reef behavior and sediment dynamics

Although still absolutely preliminary and to be confirmed with further research and monitoring, it can be stated that the Hospital Ship “Po” wreck could be considered an anthropogenic structure that likely influences local near-bed flow conditions, exhibiting characteristics frequently described for artificial reefs in shelf environments. The interaction between the hull's geometry and the NW-oriented flow could be consistent with the presence of the upstream scour hollow and the downstream depositional feature observed in the geophysical data. A key observational finding of this study is the presence of fine-grained silt-clay fractions (< 63 μm) within the area interpreted as the depositional tail, contrasting with the 350 μm medium sands of the open shelf. The fine sediments observed near the wreck are possibly consistent with the deposition under regional sediment-supply conditions influenced by the Vjosa River. This last aspect is certainly the most interesting and still to be confirmed, and we hope for future work and projects to further explore this statement: so the data support a regional hydrodynamic-sedimentary interpretation, direct sediment-source attribution at the local scale remains a newsworthy hypothesis to be tested.

The wreck appears to act as a local zone of sediment accumulation, potentially intercepting part of the suspended sediment load during high-discharge events and favoring its deposition within the sheltered wake zone. Regarding the long-term evolution of the site, the fact that the “Po” has remained in situ since 1941 suggests that these geomorphological features may have persisted over time. Although hydrodynamic thresholds were not directly measured, the observed scour morphology reflects long-term exposure to environmental forcing, generating the hypothesis that the wreck represents a persistent geomorphological feature within the local coastal system.

Benthic colonization and assemblage structure

The colonization patterns observed on the wreck are consistent with other artificial reefs in the Mediterranean Sea. Filter-feeding organisms are commonly associated with artificial reef structures and may benefit from prevailing current conditions (Badalamenti et al., 2011). The higher biodiversity and density recorded on vertical surfaces are likely due to thigmotactic preferences and a reduced predation risk (Perkol-Finkel and Benayahu, 2005). Based on our observations, the variance in assemblage composition and percentage cover between exposed and sheltered areas of the wreckage is consistent with this hypothesis. Surfaces exposed to the main flow direction and vertical orientations appear to provide favorable conditions for filter feeders and structural biogenic assemblages (Underwood and Chapman, 2006; Ballesteros et al., 2007; Moschella et al., 2005; Ferrario et al., 2016).

Invasive and endangered species

The presence of Pterois miles is consistent with the dramatic actual ongoing expansion of Lessepsian species into the Adriatic Sea. This expansion may be facilitated by rising sea temperatures and by the availability of artificial structures, including wrecks, as potential stepping-stone habitats (Bianchi and Morri, 2000; Rilov and Galil, 2009; Katsanevakis et al., 2020a,b).

The sighting of Monachus monachus is noteworthy, as artificial structures may provide occasional resting opportunities for this species in areas where natural habitats are fragmented (González, 2015; Karamanlidis and Dendrinos, 2015). In particular, the wreck appears to function as a feeding and hunting area for several species, such as monk seal, given the proven thigmotactic attraction effect on fish and marine organisms.

The “Po” wreck may therefore contribute to local habitat heterogeneity used by alien thermophilic species and occasionally by endangered marine mammals, highlighting potential conservation issues related to artificial structures in the warming Mediterranean.

Wrecks as long-term observatories of change and cultural-ecological assets

This research suggests that World War II wrecks sunk under the sea, can represent valuable sites for innovative and multidisciplinary marine scientific study as long-term examples of ecological change and geomorphological processes. Combining 3D modeling by photogrammetry with geophysical data (from MBES+SSS acoustic sonar survey) allows for non-invasive monitoring of the site. This is important for preserving underwater cultural heritage under the UNESCO 2001 Convention (Maarleveld et al., 2013). The “Po” wreck illustrates how underwater cultural heritage can act as an artificial structure associated with marine life. It can also provide a fixed reference point for studying geomorphological and ecological responses to river inputs, climate, and human influence over time. Integrating wreck observations and monitoring into Marine Spatial Planning may connect underwater cultural heritage conservation, biodiversity protection and tourism in Vlora bay, and similar Mediterranean areas.

Despite the multidisciplinary approach, this study has several limitations, due to the fact that it is a preliminary work and also carried out with the preventive contribution of previously carried out works and not calibrated for the specific purpose of this study. Geophysical and biological surveys occurred during a short period and represent snapshots of the wreck-environment relationship. The lack of time series data, such as long-term current data or seasonal biological monitoring, limits the ability to distinguish short-term variability from longer-term ecological and geomorphological changes. Also, the estimation of benthic coverage is sensitive to large spatial variation, even with careful sampling. The authors strongly hope that future studies can be organized to continue and validate the results and thesis emerged during this preliminary work and can, therefore, include long-term monitoring systems to better characterize the seasonal and inter-annual dynamics (also combining careful current measurement and geomorphological studies of the area), thus providing answers to what has emerged so far.

Role and limitations of the two indices

The Geometric-Geomorphological Wreck Effect Index (IGEOM) and the Ecological Hotspot Value Index (IECO) were created to summarize qualitative and quantitative observations—geomorphological, sedimentological, and ecological aspects—using concise, unitless numerical measures. Their main purpose is to support the structured recording of site-specific data by providing a consistent way to summarize different indicators that would otherwise remain heterogeneous (Myers et al., 2000; Halpern et al., 2008). The indices described in Supplementary material Annex 1, 2 should therefore be regarded as preliminary tools. Because they have not been calibrated across multiple wrecks or reference distributions, they cannot yet be used to define thresholds, make statistical claims, or compare different wreck sites. Therefore, the values calculated for the “Po” wreck provide only a descriptive summary of the data collected at that location. They should not be interpreted as validated measures of geomorphological impact or ecological condition. This structured numerical synthesis can still support internal consistency in reporting and may help integrate different observations into wider monitoring or Marine Spatial Planning (MSP) workflows in the future (Halpern et al., 2008). So, these indices remain site-specific. Their application to other wrecks would require validation, assessment of component behavior under different environmental conditions, and, if necessary, recalibration of the normalization schemes. Only through multi-site testing and time-series analysis will it be possible to evaluate their robustness, domain of applicability, and potential use for comparative purposes.

The “Po” shipwreck represents an artificial structure associated with benthic colonization and sediment redistribution, which appears to have developed persistent geomorphological features since its sinking. The site is characterized by increased biological diversity and structural complexity compared to the surrounding soft-bottom seabed, including observational records of protected and non-indigenous species. The interaction between the anthropogenic structure and the local environment, summarized by the preliminary Geometric-Geomorphological Wreck Effect Index (IGEOM = 0.52) and the Ecological Hotspot Value Index (IECO = 0.83), is possibly consistent with the influence of regional currents and sediment inputs from the Vjosa River. Although the outcomes of this 2-year monitoring program are highly consistent with the regional sedimentary and biological dynamics described by previous authors for the natural soft-bottoms (Savini et al., 2011; Maiorano et al., 2011), the current dataset remains intentionally observational. As a preliminary baseline, it is not yet sufficient to draw definitive causal conclusions regarding the direct hydro-geomorphological relationship between the wreck's biogenic cover, the specific suspended load from the Vjosa River, and the overall complex hydrodynamics of Vlora Bay. These quantitative findings offer a hypothesis-generating framework and a preliminary observational baseline for future long-term monitoring of the Karaburun-Sazan Marine Protected Area (MPA) in Vlora bay (south Albania). Furthermore, the application of the total digital twin (from 3D SfM underwater model reconstruction, in addition to MBES+SSS acoustic sonar surveys) of wreck and environment, plus the application of such indices (IGEOM, IECO) highlights the potential of historical shipwrecks to serve as localized ecological and geomorphological observatories, providing descriptors that align with Marine Spatial Planning (MSP) objectives and the goal of achieving Good Environmental Status (GES) under the European Marine Strategy Framework Directive (MSFD).

Future studies on shipwreck sites may consider the following research directions:

(1) broaden surveys to include additional wrecks (e.g., Teti and Probitas);

(2) deploy current meters and sediment traps;

(3) continue and complete the 3D modeling of the wreck “Po”;

(4) develop virtual reality models for public communication and integration into Marine Spatial Planning; and

(5) propose site-specific protection zones that balance cultural heritage conservation with sustainable tourism.

Ultimately, the methodological workflow adopted in this study—starting from the critical integration of pre-existing regional sonar and soft-bottom ecological baselines, followed by targeted multi-scale monitoring (sub-centimetric 3D photogrammetry, benthic square sampling, and visual censuses of key species), and culminating in the calculation of the IGEOM and IECO indices—lays the foundation for a standardized assessment protocol. By coupling this preliminary framework with future high-resolution hydrodynamic modeling and direct current measurements, this approach can be systematically replicated as a robust scientific protocol to evaluate the “wreck-effect” on other submerged vessels. This perspective elevates historical shipwrecks from isolated artifacts to integrated networks of long-term environmental observatories, providing actionable tools for Marine Spatial Planning and conservation strategies.