The CoDeMap scheme is organized into three main components: 1) Morphology, 2) Substrate, and 3) Biology (Figure 1).

Internally, the main components are systemized hierarchically, with a series of subcomponents, levels, and classes (Figure 2). Seafloor morphology, type of substrate, and distribution of individual species or communities can be mapped separately and then merged into a single map of benthic habitats by using GIS software.

More specifically, the CoDeMap morphology components consist of subcomponents organized in descending order of size and level of detail 1) Physiography, 2) Geoforms, and 3) Bedforms. Within each subcomponent, levels are used to classify features from a broader scale (Level 1, L1) to small-scale features (L2, L3). In particular, the Physiography sub-component contains one level (PL1), while the Geoforms sub-component contains three levels (GL1, GL2, and GL3), and the Bedforms sub-component contains one level (BFL1). The Substrate and Biology components include three levels each, SL1, SL2, SL3 and BL1, BL2, BL3 respectively (Figure 2).

The sum of the three components (Morphology, Substrate and Biology) returns a benthic habitat map where each class is uniquely and unequivocally represented. The maximum number of levels describing a habitat class is 11, but not all levels must perforce contain information.

Each class of the scheme is identifiable by a univocal alphanumeric code and a label (Figure 2). Both the complexity of the code and the detail expressed with the label of the features increase from L1 to L3. Within L3 of substrate and biology, classes can be more specific (e.g. B020302 – Codium adherens) or more generic (e.g. B020301- Green algae) in bold in the scheme. To limit the proliferation and redundancy of classes in the scheme, codes can be combined to describe situations characterized by multiple classes, ordered by prevalence. The combination is permitted if more classes coexist and there is a representativeness of at least 25%. For example, an area characterized by coralligenous (spatial coverage=75%) and Posidonia oceanica (spatial coverage=25%) can be described as B0907+B040403. Conversely, a Posidonia oceanica meadow (coverage=75%) with interspersed coralligenous outcrops (coverage=25%) can be codified as B040403+B0907.

The legend can be customized to include only codes, only labels, or a combination of both. Additionally, users can select which levels to display according to the complexity and purpose of the representation (see paragraph 3).

The South Adriatic Sea has been investigated by the CNR-ISMAR throughout the last 20 years by the acquisition of a large amount of geophysical data (multibeam and seismic), seabed samples (grab samples, box cores), and video from ROV. The interpretation of these data provided the basis to produce a geomorphological map of the South Adriatic continental margin (Campiani et al., 2024), and a benthic habitat map published in Prampolini et al. (2021). In this application, we have classified these two products using the CoDeMap BHCS producing several maps representing the morphology, the substrate, the biology, and the benthic habitat map of the basin.

Figure 3A represents the Morphology component classified according to the Physiography sub-component. In Figure 3B, the morphologic classification includes the level of Physiography (PL1) and the three levels of Geoforms (GL1, GL2, GL3), increasing the detail and the complexity of the seascape. The complete South Adriatic morphology is charted in Figure 4, where all the sub-components (Physiography, Geoforms and Bedforms) were used to build the map. These three representations of the South Adriatic morphology enhance the increase in scale, detail and complexity of the depicted seascape, by applying different levels of the CoDeMap BHCS and consequently, changing the information represented on the map.

The South Adriatic continental margin has been classified according to information on substrate and biology at different levels of detail. The map in Figure 5 derives from the interpretation of seabed reflectivity and samples; the latter permitted to specify substrate texture and biological communities living on the seafloor. The substrate is described by all levels of the CoDeMap Substrate component and represented by using distinct colors range and tones according to the texture. The Biology component is depicted through a halftone screen superimposed on the substrate.

The Tricase Canyon is a submarine feature that cuts through the Apulian continental slope on the western side of the Ionian Sea (Mediterranean Sea). Morphological and substrate components were mapped by interpreting the Digital Elevation Model (DEM) and classifying seabed acoustic reflectivity using Remote Sensor Object-Based Image Analysis (RSOBIA). A ground-truthing activity involving seabed samples and ROV images validated the results of the automatic classification and helped analyze the biological component. In fact, the deeper areas of the canyon host white corals such as Madrepora oculata, Desmophyllum dianthus, and Desmophyllum pertusum (=Lophelia pertusa). Corals have been mainly observed on the top of blocks interpreted as the result of several mass-transport deposits (Prampolini et al., 2020). Figure 6 shows the benthic habitat map of the Tricase Canyon described using the CoDeMap BHCS.

For the Tricase Canyon case, we decided to use all 11 levels in the legend displaying just the code, integrating all three components into a single seafloor representation. Table 2 contains the labels for the various classes.

The Dohrn Canyon is in the center of the Gulf of Naples, a submarine canyon of ecological, functional and oceanographic interest since featured by important upwelling currents affecting more coastal waters. It hosts deep bioconstructions, specifically cold-water corals and oysters. Available information documents the presence of living specimens of the scleractinians M. oculata, D. pertusum, and D. dianthus. These communities are also associated with large bivalves such as Neopycnodonte zibrowii and Acesta excavata. This coexistence of deep corals and large bivalves represents a unique biotope for the Mediterranean Sea, threatened by severe anthropogenic threats (Taviani et al., 2019).

In this application of the CoDeMap BHCS, we described the ROV transect coded “ANOMCITY_ROV01”, conducted during the CNR oceanographic cruise ANOMCITY 2016 aimed to characterize and map the bioconstructions populating the Dohrn Canyon (Oliveri et al., 2016). The transect develops along the flank of the northern branch of the canyon following a South-North direction, revealing the coexistence of cold-water corals and deep oysters (Figure 7).

From 2000 to 2024 several research projects and scientific papers focused on the area in the South Adriatic Sea that runs along the Apulian coast and continental shelf (Italy). In this case study, we considered the seabed stretching from Mola di Bari to Fasano municipalities hosting Posidonia oceanica meadows, coralligenous bioconstructions and coral reefs in shallow waters, as well as deep oyster reefs in the continental shelf at approximately at 100 meters water depth. The seabed substrate ranges from hardgrounds to bioclastic coarse and fine sediments (from gravelly sands to sandy muds), while the continental shelf is characterized by flat surfaces, megaripple fields, comet marks, and erosional remnants. We integrated the maps produced through the years from the coast to the deep-sea into a comprehensive map intended to support conservation initiatives, such as the establishment of new Natura 2000 sites (Grande et al., 2024). The main challenge lies in homogenizing maps derived from multiple sources, produced at different scales and using different devices (see Supplementary Figure S1 in the Supplementary Materials for the original data).

Figure 8 compares the CoDeMap and EUNIS classification schemes in the production of benthic habitat maps for Apulian coastal waters and continental shelf. Both frameworks ensure a consistent representation of the study area, even in contexts where data availability is limited. Within EUNIS, these areas can be classified at Level 1 as “Marine benthic habitats,” whereas CoDeMap adds further detail by incorporating physiographic features such as the “Continental shelf.” The key distinction between the two approaches lies in the reliance on biozones in EUNIS versus geomorphological classes in CoDeMap. In Figure 8A, geomorphological features are clearly delineated, with biological data embedded within their abiotic setting. For instance, the oyster reef offshore Monopoli is shown to coincide with an erosive remnant area, offering valuable insights into the reef’s formation. Conversely, EUNIS highlights the spatial distribution of biozones: Figure 8B clearly illustrates the extent of habitats across the infralittoral and circalittoral zones.

In terms of completeness, CoDeMap generally provides a more detailed account of the original data. This is exemplified by the mesophotic coral reef described by Corriero et al. (2019), classified under CoDeMap as a coral reef dominated by Phyllangia americana mouchezii and Polycyathus muellerae, whereas EUNIS categorizes it as “MC2 – Mediterranean circalittoral biogenic habitat.” In this case, the absence of specific EUNIS classes necessitates classification at Level 3, resulting in a coarser description. By contrast, CoDeMap captures the reef’s character more precisely, including the identification of its dominant species.

When harmonizing maps from different sources, however, some information may be lost, potentially leading to misclassification or the use of categories that do not fully match the context. This is illustrated in Figure 8 with the mapping of Posidonia oceanica from the “Inventory and Cartography of Posidonia Meadows” (POR 2000–2006). The dataset shows a mosaic of Posidonia oceanica and matte within the circalittoral zone. Under the EUNIS scheme, the area can only be mapped as MB252 “Biocenosis of Posidonia oceanica” (Figure 8B), which omits details on the presence of matte, since the available data are insufficient to classify the habitat at Level 5 (e.g., MB2523 “Facies of dead mattes of Posidonia oceanica without much epiflora”). CoDeMap, by contrast, allows classes to be combined, thus retaining the full complexity of the original dataset.

Finally, CoDeMap consistently provides detailed information on substrate composition (e.g., gravelly and muddy sands or sandy muds across the continental shelf), whereas EUNIS categories do not always capture substrate variability exhaustively—for example, “MC45 Mediterranean circalittoral mixed sediment”.