EMODnet Seabed Habitats Glossary
A guide to terms and phrases used within EMODnet Seabed Habitats - you may use the links below to jump to a specific bookmark in the glossary.
The degree to which a measured value (either spatial resolution/precision or attribute) conforms to a true or accepted value. Accuracy is a measure of correctness. It is distinguished from precision, which measures exactness
In a habitat mapping context, accuracy describes how closely a map predicts the actual habitat observed on the seabed at a given location.
Non-spatial information describing the characteristics of a map object (vector geographic feature in a GIS), usually stored in a database table and linked to the map object by a unique identifier. For example, a polygon feature drawn on a habitat map will have the type of habitat present as one of its attributes, where values are any text entry taken from the EUNIS classification scheme (e.g. A1.23). In raster datasets, an attribute is the information associated with each unique value of a raster cell that describes it value.
A database or tabular file containing information about a set of geographic features, usually arranged so that each row represents a feature and each column represents one feature attribute. In raster datasets, each row of an attribute table corresponds to a certain zone of cells having the same value. In a GIS, attribute tables are often joined or related to spatial data layers, and the attribute values they contain can be used to find, query, and symbolize features or raster cells.
Bathymetry is the study of underwater depth, leading to topographic maps of the ocean floor. A bathymetric map or chart usually shows seabed relief or terrain as contour lines or false colour composite images, using a colour ramp (shades of different colours) to indicate different depths.
Any deviation from a flat bed, generated by the flow of a transporting agent (water, ice or air). Bed-forms range in size from ripples in the sand, a few centimetres apart and a few millimetres high, to 'dunes' with wavelengths of hundreds of meters, kilometres in length and a few to tens of meters high.
The sound you make.
Generally used to describe a map that shows the distribution of broadly-defined habitat types over a relatively large geographic area. It is often used instead of the more technically correct mapping expression small-scale (e.g. 1:250,000). Broad-scale also implies that the primary purpose of a map is to present an overview of a large area. As opposed to fine-scale.
Cartographic modelling is a general, but well-defined methodology that is used to address diverse applications of GIS in a clear manner. It is a technique used for both vector and raster based GIS, and as the term suggests, cartographic modelling involves models (i.e. of geospatial information) represented in cartographic form (i.e. as maps). Cartographic modelling is used to simultaneously analyse both the spatial and thematic characteristics of geospatial information.
A class is a set of entities grouped together on the basis of shared attribute values.
Continuous data can be sub-divided into 'chunks' or classes. For example, a measure of tidal stream strength might range from 0-15 but could be divided into three classes for easier comparison - Low (0-5.0), Medium (5.1-10) and High (10.1-15). Classes can be created from complex multivariate data (such as the list of difference species, their abundance, type of seabed and wave exposure data at a specific location) by grouping together locations with similar values for each variable. These classes are the data format is required for habitat maps.
Systematic arrangement of habitats into more broadly defined classes on the basis of an analysis of their attributes; the arrangement is often repeated to create a hierarchy with broadly defined habitats at the top, sub-dividing into more and more detailed, narrowly defined classes further down the hierarchy.
A statement about how reliable a map user thinks the map is given its purpose. This is not a mathematical definition like accuracy or uncertainty, but is a judgement made by the map-user and may therefore vary for any map. However, this judgment can be supported by evidence from:
- Accuracy measures
- Supporting maps show underlying evidence used to interpret map
- Evaluation of all contributing data
- Independent validation
- Expert opinion
- User support: Generally found to be acceptable by stakeholders and the map has stood the test of time
A single thematic map, usually in the context of one of many layers in a GIS. A coverage can be a data layer or overlay, storing geographic features. In a coverage, features are stored as both primary features (points, arcs, polygons) and secondary features (tics, links, annotation). In the MESH context coverage is also taken to mean that the coverage is full (100% cover) over the survey area.
In a data collection context, coverage is also taken to mean that the actual area of a survey for which remote sensing data are available. For example, an aerial photograph may cover all the area (100% coverage) but a sonar survey may record along a series of non-overlapping corridors with gaps between resulting in only 50% coverage.
A spheroid used to approximate the shape of the Earth in order that a coordinate system may be used to locate objects. As the earth is an imperfect sphere, different datums may be more accurate in certain geographic locations.
- WGS84, commonly used for worldwide coordinate systems;
- ED50, which focuses on accuracy in Europe;
- NAD27/83 for North American cartography.
A Data Exchange Format (DEF) defines the characteristics of data to be exchanged between parties. These characteristics will typically cover the data format (e.g. ESRI shapefile), the geographic co-ordinate system used (e.g. Ordnance Survey of Great Britain). the attributes required (e.g. depth, seabed type, habitat class) and their format (e.g. habitat type - text data no longer than 20 characters; depth - numerical data with no more than 2 decimal places).
A digital elevation model (DEM) is a digital representation of the ground surface topography or terrain. It is also widely known as a digital terrain model (DTM). A DEM can be represented as a raster (a grid of squares) or as a triangular irregular network. A DEM - sometimes also called a digital surface model (DSM) - generally refers to a representation of the earth's surface (or subset of this), including features such as vegetation, buildings, bridges, etc. The DEM often comprises much of the raw dataset, which may have been acquired through techniques such as photogrammetry, LiDAR, IfSAR, land surveying, and remote sensing techniques. DEMs are used often in geographic information systems, and the most common basis for digitally-produced relief maps. (From Wikipedia).
A model in which the parameters and variables are not subject to random fluctuations, so that the system is at any time entirely defined by the initial conditions chosen. In a habitat mapping context, the data layers are selected and fixed so that the user can vary the model parameters to explore the relationship between the habitat type and the physical data layers in the knowledge that the physical data itself will not be changed by the modelling process.
When a user views a remotely-sensed image of the seabed, they will be able to see patterns that reflect changes in the surface structure. For example, when a user views an aerial photograph, the surface structures are clearly visible. A user can draw polygon shapes around these surface features to encapsulate an area that 'looks' the same and can be truthed / interpreted as a habitat type. This process of drawing boundaries around habitat types is called direct mapping; it is also known as segmentation.
Basic information that should provide a sufficient description of a data set to enable the user to establish whether the data meet their requirements. Typically discovery metadata will answer the 'who? what? where? and when? questions' for a dataset. Technically the term discovery metadata refers to a high level set of metadata elements.
The ability to distinguish two or more habitat classes. There are two main stages in the habitat mapping process where discrimination must take place to map habitats as separate entities. Firstly, the field sampling system must detect features which can be used to separate classes. Secondly, the classes must be separable by their response to remote sensors or by differentiated on the basis of physical variables.
The broad-scale European predictive habitat mapping products produced by EMODnet Seabed Habitats. Created using a top-down approach and provided in the EUNIS classification system. For more information, see "About EUSeaMap".
In phase I of EMODnet Seabed Habitats, the project itself was also known as "EUSeaMap".
Based on experiment and observation but not necessarily on proven scientific data, or based entirely on practical experience.
Empirical models are developed using field observations and practical experience to produce maps that can be tested (by field samples or observations) to 'prove' their reliability.
A ESRI™ vector data storage format for storing the location, shape, and attributes of geographic features. A shapefile is stored in a set of related files and contains one feature class. Shapefiles spatially describe points, polygons and polylines. The term shapefile is generally used to mean to a collection of files with '.shp', '.shx', '.dbf', and other extensions on a common prefix name (i.e. 'habitat.*'). The actual shapefile relates specifically to files with the '.shp' extension; this file alone is incomplete for dissemination, as it depends on the other supporting files.
The EUNIS habitat types classification is a comprehensive pan-European system to facilitate the harmonised description and collection of data across Europe through the use of criteria for habitat identification; it covers all types of habitats from natural to artificial, from terrestrial to freshwater and marine.
Level A describes all MARINE HABITATS and is broken down into 8 sub-sections: A1 = Littoral rock and other hard substrata, A2 = Littoral sediment, A3 = Infralittoral rock and other hard substrata, A4 = Circalittoral rock and other hard substrata, A5 = Sublittoral sediment, A6 = Deep-sea bed, A7 = Pelagic water column, A8 = Ice-associated marine habitats.
Habitat is defined as: Plant and animal communities as the characterising elements of the biotic environment, together with abiotic factors operating together at a particular scale.
EMODnet is a DG MARE funded network of organisations supported by the EU's integrated maritime policy. These organisations work together to observe the sea, process the data according to international standards and make that information freely available as interoperable data layers and data products. It is split by themes into the following "lots":
- Human Activities
- Seabed Habitats
For more information, please visit EMODnet Central Portal.
Generally used to describe a map that shows the distribution of detailed habitat types over a relatively small geographic area. It is often used instead of the more technically correct mapping expression large scale (e.g. 1:10,000). Fine-scale also implies that the primary purpose of a map is to explore the distribution of small features with a measurable degree of accuracy. As opposed to broad-scale.
An integrated collection of computer software and data used to view and manage information about geographic places, analyse spatial relationships, and model spatial processes. A GIS provides a framework for gathering and organizing spatial data and related information so that it can be displayed and analysed. GIS are generally much more powerful than computer assisted drafting (CAD), although the distinction between them is not well defined.
Aligning geographic data (map features) to a known coordinate system so they can be viewed, queried, and analysed with other geographic data. Georeferencing may involve shifting, rotating, scaling, skewing, and in some cases warping, rubber sheeting, or orthorectifying the features so that they fit the co-ordinate system.
Direct observations and samples of the seabed provide information that can be used to interpret remotely sensed images; the observations are the 'truth' with regard to the habitats actually present on the seabed. Observations used in this way provide ground truth data. The process of using ground truth data for interpretation is often termed ground truthing. During this process the relationship between properties of the remote images at the observation/sample sites (in the form of points, irregular digitised areas or buffer areas around points) is determined. These relationships are then applied to the whole image to predict the distribution of habitat types. Ground truthing is distinct from ground validation.
Observations can be used to test the predictive power of a habitat map. The validation dataset is displayed on top of the habitat map and the predicted habitat class is compared with the actual class from the validation observation. This is the basis for many measures of accuracy and uncertainty. The ground validation dataset should not be the one also used for ground truthing, although this rule is often broken when sample data are sparse. These issues are discussed more fully in "How good is my map?" in the archived MESH guide to mapping.
The organisation of different habitats into specific class types. These classes can vary in their level of detail from broad descriptions (e.g. Shallow subtidal sand) to highly detailed descriptions that include specific organisms (Zostera marina seagrass beds on clean coarse sand in sheltered water less than 5 m). In general, detailed habitat types will be more geographically limited to small areas of the seabed.
A habitat requires specific physical (and chemical) environmental conditions to support its biological community. Searching a map of physical variables to find the specific environmental conditions to suit a habitat is called suitability analysis. In other words, the appropriateness of an area to support a particular habitat is determined.
A similar approach can be applied to find the suitable conditions to support an individual species.
Habitat distribution is normally defined by multiple environmental variables (depth, light penetration, incident energy etc). Habitat suitability models bring together these multiple data layers to simultaneously select the most appropriate area and therefore predict habitat (or species) occurrence.
The amount of diversity (or variability) of a variable (e.g. sediment type) or class (e.g. habitat type) within a geographic area. In offshore areas, extensive sediment plains tend to dominate the seabed with little variability in sediment type - low heterogeneity. Shallow inshore areas are often a mixture of rocky outcrops, gravel beds and many sediment types depending upon the tidal streams and/or wave action - there is much variability over small areas leading to high heterogeneity. Heterogeneity has practical implications for mapping at two levels. Firstly, there is heterogeneity within the minimum mapping unit of a map in which case what is displayed on the map must either try to represent this diversity using a mixed class or display some measure of diversity or the map class must simplify the situation on the ground by showing the predominant class. Secondly, the map may show many small polygons of different classes clustered together in which case there is likely to be reduced confidence in the exact position of each class. The map shows what might be found in the general region.
The ability of two or more systems, or components to exchange information, and to use the information that has been exchanged. It is commonly used to describe the sharing of data on the internet where a mapping website will draw data from another website to give context to its data; for example, a website may draw a coastal outline from a central repository rather than try to maintain its own copy.
The estimation of surface values at unsampled points based on known surface values of surrounding points. Interpolation can be used to estimate elevation, rainfall, temperature, chemical dispersion, or other spatially-based phenomena. Interpolation generates a continuous surface from a series of discrete individual points.
The visual representation of a geographic dataset in any digital map environment. Conceptually, a layer is a slice or stratum of the geographic reality in a particular area, and is more or less equivalent to a legend item on a paper map. On a road map, for example, roads, national parks, political boundaries, and rivers might be considered different layers.
In ESRI™ ArcGIS, a layer is a reference to a data source, such as a shapefile, coverage, geodatabase feature class, or raster, that defines how the data should be symbolized on a map.
In the particular case of mapping, LiDAR is an airborne surveying technique which utilises the travel time of laser light to measure altitude of the aircraft above the ground. When the aircraft has a level flight path, variations in the measured altitude are actually changes in the height of the ground. LiDAR is the airborne equivalent of sonar. Infrared light is adapted for ground detection, whilst green light, due to its penetration ability through water, offers a way to carry out clear water bathymetry surveys.
MESH was an international marine habitat mapping programme that ran from 2004-2007 and was carried out by a consortium of 12 partners across the UK, Ireland, the Netherlands, Belgium and France. MESH aimed to produce seabed habitat maps for north-west Europe and develop international standards and protocols for seabed mapping studies. Many of the products from MESH have since been absorbed into EMODnet Seabed Habitats.
A set of metadata elements that are needed to describe a particular type of data. Metadata standards are usually defined by official standards organisations, but they can also be defined by an organisation or project for a specific purpose: the MESH metadata standard (www.searchMESH.net/metadata) are the set of metadata elements necessary to fully describe seabed habitat map data.
This is the smallest size of a habitat, which although not precisely defined in the EUNIS classification system, is taken to be about 5 x 5 m for marine habitats. Any feature smaller than this is regarded as an attribute of the habitat. Note that, whilst the MHU is well within the resolution of many remote sensing systems, it is unlikely that the MHU will correspond with the minimum mapping unit unless the map scale is very small.
This is the smallest object size that is represented on a map (smaller objects being either 'lost' or subsumed into a larger unit). In a single thematic map, such as a habitat map, this will be the minimum size of any habitat class that can be represented on a map at any given scale. Note that this means the size of the MMU will change with scale. This is distinct from the minimum habitat unit (MHU).
The dictionary definition of the term 'model' that is most closely related to habitat mapping is: a simplified representation used to explain a real world system. Therefore any map representation of the seabed based on a systematic investigation of remotely sensed data correlated to ground truth information can be regarded as a model. Modelling is the process by which data are simplified to produce the map. All maps should be regarded as predictive and require testing against the real world. Cartographic modelling and mathematical modelling provide a more rigorous basis for maps than conceptual models that are based on expert opinion. Cartographic and mathematical modelling are more reproducible since there is no element of subjectivity in the analysis. Models can be re-run with changed parameters based on new evidence.
The 1992 OSPAR Convention is the current instrument guiding international cooperation on the protection of the marine environment of the North-East Atlantic. It combined and up-dated the 1972 Oslo Convention on dumping waste at sea and the 1974 Paris Convention on land-based sources of marine pollution. The work under the convention is managed by the OSPAR Commission, made up of representatives of the Governments of 15 Contracting Parties and the European Commission, representing the European Community.
The smallest unit of information in an image or raster map, usually square or rectangular, the four sides of a pixel enclose a small, homogeneous area. Pixel is often used synonymously with cell. In remote sensing, a pixel is the fundamental unit of data collection. A pixel is represented in a remotely sensed image as a cell in an array of data values. The term mixel is used where a pixel is known to consist of more than one habitat class on the ground.
On a map, a closed shape defined by a connected sequence of x,y coordinate pairs, where the first and last coordinate pair are the same and all other pairs are unique. In the context of a habitat map a polygon is taken to be homogeneous within that enclosed area (i.e. every point within the shape shares the same attributes).
In habitat mapping, there is confusion between this term and accuracy. Precision can be defined as the variability between repeated measurements but this has limited application to habitat mapping. However, in habitat mapping its more general usage is to define the likely error of a boundary (e.g. ±100 m, ±10 m). It could also be applied to the level in a hierarchy that a record has been assigned to (i.e. a EUNIS level 4 class is less precise than a level 5 class).
All habitat maps derived from classification and modelling are predictive because the actual habitat present on the seabed has not been observed/sampled at all points of the map. The mapped habitat distributions are based on statistical links, assumptions and hypotheses between the source data (from remote sensing) and the classes to be mapped (determined from ground truth samples). This underlying predictive nature of maps is often conveniently ignored, but all maps must be judged by their predictive power.
The ability of a map to correctly predict what will be found on the seabed in each minimum mapping unit of the map. Note predictive power need not only apply to the mapped class – the predictive power of a map could be widened to include (for example) rarer biotopes associated with the dominant mapped biotope.
A measure of the likelihood that a particular outcome, such as a spatial pattern or event, will occur given a set of possible outcomes. Probability values range from 0 for impossible outcomes to 1 for completely certain outcomes. The probability that a tossed coin will land heads-up, for example, is 0.5, since landing heads-up is one of two possible outcomes.
In a seabed habitat mapping context, it is possible to determine the probability of a single habitat occurring at every location in a map: those areas where the prevailing physical conditions are suitable for the habitat will lead to a higher probability of occurrence than areas that are clearly unsuitable. When individual habitat probability maps are overlaid, for each location it is possible to see 'most likely' habitat (highest probably) to be found, followed by the next most likely and so on to the least likely habitat.
In ecological research, a proxy (variable) is something that can be easily measured and is known to be a substitute for a variable that cannot be easily measured. Clearly, the relationship between the proxy (variable) and the real variable must be understood, and be consistent. For example, ocean surface colour recorded by satellite sensors is used as a proxy for water clarity from which the depth of light penetration may be calculated. Often, the proxy is probably not in itself of any great interest.
The RAMSAR Convention on Wetlands. Ramsar, Iran, 1971
The Convention on Wetlands, signed in Ramsar, Iran, in 1971, is an intergovernmental treaty which provides the framework for national action and international cooperation for the conservation and wise use of wetlands and their resources. There are presently 153 Contracting Parties to the Convention, with 1634 wetland sites, totalling 145.6 million hectares, designated for inclusion in the RAMSAR List of Wetlands of International Importance.
A spatial data model that defines space as an array of equally sized cells (pixels)arranged in rows and columns, and comprised of single or multiple bands. Each cell contains an attribute value and location coordinates. Unlike a vector structure, which stores coordinates explicitly, raster coordinates are contained in the ordering of the matrix. Groups of cells that share the same value represent the same type of geographic feature. Many single map and multi-map operations in a GIS involve raster maths that are not possible with vector format layers.
Raster format is a graphics image or data file representing a generally rectangular grid of pixels, or points of colour, on a computer monitor, paper, or other display medium. The colour of each pixel is individually defined; for example colour images often consist of coloured pixels defined by three values — one each for red, green and blue. Raster graphics are distinguished from vector graphics in that vector graphics represent an image through the use of geometric objects such as curves and polygons.
The quality of a raster image is determined by the total number of pixels (resolution), and the amount of information in each pixel (often called colour depth). Raster graphics cannot be scaled to a higher resolution without loss of apparent quality. This is in contrast to vector graphics being dimensionless), which easily scale to the quality of the device on which they are rendered.
Is the ratio or relationship between a distance between the same two points on the earth's surface. A map scale of 1/100,000 or 1:100,000 means that one unit of measure on the map equals 100,000 of the same unit on the earth. For example, if two points are 10,000 cm (100 m) apart on land and 1 cm apart on a map, the map's scale is 1:10,000. A small ratio (1:250,000) is a small scale and shows objects to be small. A large ratio (1:10,000) is a large scale and shows objects to be large. There is often confusion over these terms and, where possible, MESH uses the terms broad-scale to indicate that the map scale is small and fine-scale to indicate that it is large.
Sediment is solid fragmental material which has been eroded, transported and deposited by wind, water or ice; chemically precipitated; or secreted by organisms. It forms in loose, unconsolidated layers. Sediment type varies according to various parameters, including lithology, particle size distribution, compaction/ bulk density, cohesiveness, shell content, and moisture content. It reflects a history of erosion, transportation and accumulation in different sedimentary environments.
'Condition' is the defined quality or status of the biological features at a site. Assessing the current status of the features against the defined standard is termed site condition monitoring. Repeating these assessments over time will establish whether any changes in status may be linked to any trends in the physical environment or human activities.
A method for classifying images whereby thematic classes are defined by the characteristics for pixels within an image that correspond to training areas in the field chosen to represent known features (e.g. habitats). Each pixel within the image is then assigned to a thematic class using one of several decision rules.
It is a major tool for integrating ground-truth samples and remotely-sensed images. The user supervises feature classification by setting up prototypes (collections of field sample points) for each feature, class, or habitat type to be mapped. During this process the relationship between properties of the remote images spatially associated with the sample sites (in the form of points, irregular digitised areas or buffer areas around points) are then applied to the whole image.
It is the actual area of the seabed imaged by a remote sensor during one 'pass'. Side scan and multibeam sonars project a beam of sound out to either side of the vessel's towpath to ensonify a wide region of the seafloor. Both right and left sonar channels added together make up the swath. Swath width is dependent on the water depth and can also be changed by altering the range setting of the sonar device. Swath width is an important factor in determining the lane spacing of vessel tracks in achieving the desired coverage of the survey area.
A similar principle applies to airborne remote sensing devices (cameras, LiDAR etc) where the sensor images an area of the seabed. The area imaged will depend on the altitude of the aircraft and the range settings of the sensor. Coverage will depend on these settings and the flight path of the aircraft.
A defined area of an image where the habitat class is known (normally from ground truth sampling) that is used for gathering image data to derive the relationship (correlation) between the habitat class and image. This relationship can then be used to seek other areas of the image that match and may be assumed to be the same habitat type. These original areas are 'training' the classification process.
In mathematics, truncation is the term used for reducing the number of digits right of the decimal point, by discarding the least significant ones. In taxonomy, there is a similar concept, reducing the precision to which an organism is identified/classified. For example, truncating the barnacle Balanus balanoides to progressively higher taxonomic levels such as Genus (Balanus), Family (Balanidae) or Order (Sessilia).
The degree to which the measured value of some quantity is estimated to vary from the true value. Uncertainty can arise from a variety of sources, including limitations on the precision or accuracy of a measuring instrument or system; measurement error; the integration of data that uses different scales or that describe phenomena differently; conflicting representations of the same phenomena; the variable, unquantifiable, or indefinite nature of the phenomena being measured; or the limits of human knowledge. Uncertainty is the opposite of confidence.
A position fixing method for establishing the real position of an instrument (towed camera, towed sonar fish, a diver) below a survey vessel. It uses a sonar transponder - responder fitted to the deployed instrument and a transceiver mounted on a vessel at a known and surveyed position below the water line. The direction of origin of received signal from the instrument mounted transponder-responder indicates the towed instrument position relative to the support vessel, while the time delay between transceiver emitted signal and transponder/ responder return signal provides its distance. Positioning of the vessel mounted transceiver is related by survey to the position of GPS antenna, enabling accurate positioning relative the GPS defined position of the vessel.
A GIS data format where objects are represented by a coordinate-based data model that represents geographic features as points, lines, and polygons. Each point feature is represented as a single coordinate pair, while line and polygon features are represented as ordered lists of vertices. Many attributes are associated with each vector feature, as opposed to a raster data model, which associates only one attribute with a grid cell feature.
Vector graphics is the use of non-dimensional geometrical primitives such as points, lines, curves, and polygons, which are all based upon mathematical equations to represent images in computer graphics. It is used by contrast to the term raster graphics, which is the representation of images as a collection of pixels (dots) which represent a given surface on the ground.
Extensible Markup Language (XML) files are often used for exchanging information, both on and off the Web. Like HTML (Hypertext Markup Language) files, XML files use start and end tags to format their content. However, XML tags define the structure of elements in a document, whereas HTML tags define how elements should look. XML is extensible because you can extend it by adding custom tags.