Biotope modelling technical summary – Nordland VI

This page provides a brief technical summary of biotope classification and modelling in the Nordland VI mapping area.


See map.

Video data analysis and biotope classification

All CAMPOD video data from Nordland VI were reviewed by biologists at the Institute of Marine Research who recorded all the visible fauna down to the lowest taxonomic level possible.  The species observations were then analysed using multivariate statistics (Detrended Correspondence Analysis – DCA) to reveal natural groupings which provide the information for the next stage, biotope classification.  Identification of the most appropriate number and best grouping of biotopes was done with the help of the statistical clustering technique k-means analysis. Classified point observations (representing approximately 200 m of pooled video observations) can be seen in the figure below while the physical characteristics of each biotope are listed in the table, together with typical fauna. Note that not all observation points are visible at this map scale.

Biotope Typical fauna* Physical environment Biology in brief
Continental shelf
 

 

   

Biotope 2

Thenea sp. (Demospongia), Quasilina sp. (Demospongia), Bonelliidae (Echiura), Bolocera tuediae (Actiniaria), Phycis blennoides (Teleostei)

Mixed (sand dominated, with gravel) sediments on continental shelf plain/marine valley above 600 m. Small sponges and echiurans.

Biotope 6
Axinella infundibuliformis (Demospongia), Primnoa resedaeformis (Octocorallia), Craniella zetlandica (Demospongia), Paragorgia arborea (Octocorallia), Tethya citrine (Demospongia). Mixed (gravelly/sandy) sediments sometimes with mud and including compacted sediments. Outer continental shelf plain and upper continental slope above 500 m. Sponges and corals.

Biotope 8
Parastichopus tremulus (Holothuroidea), Kophobelemnon stelliferum (Octocorallia), Flabellum macandrewi (Scleractinia), Ditrupa arietina (Polychaeta), Trisopterus esmarkii (Teleostei). Sandy/muddy sediments on continental shelf plain and (shallow) marine valley above 500 m, often in lower lying areas. Seapens and cup corals.

Biotope 9
Geodia atlantica (Demospongia), Geodia barrette (Demospongia), Antho dichotoma (Demospongia), Phakellia sp. (Demospongia), Stryphnus ponderosus (Demospongia). Gravelly/sandy sediments, sometimes with cobbles and boulders, above 500m, mainly more elevated areas of continental shelf plain or broad slopes. Soft bottom sponge aggregations.
Continental slope and deep sea plain      

Biotope 4
Crossaster papposus (Asterpoidea), Gorgonocephalus sp (Ophiuroidea), Drifa glomerata (Octocorallia), Reinhardtius hippoglossoides (Teleostei), Lycodes esmarkii (Teleostei). Mixed (muddy) sediments, continental slope and marine valleys, including areas of steep slopes. Cauliflower corals and basket stars.

Biotope 3
Cerianthidae purple (Actiniaria), Ophiopleura borealis (Ophiuroidea), Asconema setubalense (Hexactinellida), Lucernaria bathyphila (Stauromedusae), Candelabrum sp. (Hydroidea). Continental slope 600-1700 m with mixed sediments. Cerianthids and Ophiopleura.

Biotope 5
Bathybiaster sp. (Asteroidea), Pontaster tenuispinus (Asteroidea), Chondrocladia gigantea (Demospongia), Gersemia sp (Octocorallia), Lycodonus sp. (Teleostei). Mixed (muddy/sand) sediments) continental slope 500 – 2000 m. Bathybiaster and Pontaster.

Biotope 7
Bythocaris leucopis (Decapoda), Caulophacus arcticus (Hexactinellida), Elpidia glacialis (Holothuroidea), Pourtalesia jeffreysi (Echinoidea), Neohela monstrosa (Amphipoda). Muddy sediments below 1500 m. Deep sea shrimp and glass sponges.

Biotope 1
Hymenaster pellucidus (Asteroidea), Bathycrinus carpenter (Crinoidea), Lycodes frigidus (Teleostei), Hexactinellida indet (Hexactinellida), Paraliparis bathybius (Teleostei). Muddy sediments, continental slope, generally below 1000 m. Hymenaster and Sea lilies.

* The typical species are defined on the basis of the combination of the frequency of occurrence (how common the species is) and mean abundance. The top five taxa of species being most common with highest abundances in a group were selected.

Physical environment and biotope modelling

The physical characteristics listed in Table 1 above are those which were found to be most relevant to each biotope through statistical analysis and modelling. These characteristics are determined from multibeam data (bathymetry and backscatter), derived terrain variables such as slope, and geological maps produced by the Geological Survey of Norway for MAREANO. Through further analysis we identified which of these environmental variables would serve as the best predictor variables to model the overall distribution of biotopes – these were bathymetry, sediment grain size, landscape, sedimentary environment, backscatter and broad-scale slope. Biotope modelling employed Maximum Entropy Modelling (MAXENT) a machine learning modelling technique which produces separate maps for each biotope which are then combined into a composite map showing the overall distribution of biotopes. In each model 25 % of the data were retained for model testing. A summary of the performance statistics for the individual and combined models are given below for information. The overall accuracy is simply the percentage of correctly classified points in the final composite model. The producers and users accuracy measures provide an indication of how the accuracy varies between classes and are used to produce the Kappa Statistic which is another standard measure of classification performance. With a Kappa value of 0.56 this model has fair-good performance. These performance statistics should improve when the limited Nordland VI dataset is combined with other MAREANO data to produce a harmonised biotope for adjacent areas, however initial indications already suggest that the map shows some similarities to that for neighbouring Nordland VII which is based on a larger dataset extending all the way to Troms II.

Class 1 2 3 4 5 6 7 8 9 Total Producers accuracy
1
27
0 0 0 0 0 12 0 0 39 0.69
2 0
118
0 0 0 28 0 15 20 181 0.65
3 0 0
45
10 21 0 1 0 0 77 0.58
4 1 2 1
36
0 0 2 1 0 43 0.84
5 6 0 3 11
28
97 0 9 20 165 0.59
6 0 31 0 8 0
97
0 9 20 165 0.59
7 12 0 1 0 2 0
61
0 0 76 0.80
8 0 16 0 1 0 4 0
53
4 78 0.68
9 0 54 0 0 0 25 0 7
72
158 0.46
Total 46 221 50 66 51 154 76 85 116 865  
Producers accuracy 0.59 0.53 0.90 0.55 0.55 0.63 0.80 0.62 0.62    

Overall accuracy 62 %: Kappa statistics 0.56 (fair-good).