High-resolution species distribution modelling of two coastal biogenic habitat-forming species in an Ecologically and Biologically Significant Area of the Bay of Fundy, Canada
Data files
Oct 03, 2025 version files 3.74 GB
-
BRT_models.R
9.38 KB
-
model_data_boltenia.csv
881.65 KB
-
model_data_modiolus.csv
881.94 KB
-
Raster_layers.zip
3.74 GB
-
README.md
18.33 KB
Abstract
High-resolution species distribution models (SDMs) were developed for two benthic invertebrate species of marine conservation significance across a 113 km2 Ecologically and Biologically Significant Area (EBSA) of the Bay of Fundy, Canada. The stalked tunicate, Boltenia ovifera, and horse mussel, Modiolus modiolus, can form coastal biogenic habitat and are vulnerable to disturbance. A near-seabed imaging survey (depths ranging from 8 to 79 m) provided presence, absence, and abundance data for both species. Boosted Regression Tree SDMs combined these data with 11 environmental variables. Presence-probability distributions were generated; however, abundance patterns could not be adequately modelled. Oblique geographic coordinates, which incorporate location of samples as information, proved useful for predicting species presence, along with seabed rugosity, maximum current speed and bathymetry for B. ovifera, and maximum and minimum current speed along with seabed rugosity for M. modiolus. High-resolution SDMs (in this case, 5-m grid) provide enhanced spatial context for ocean managers towards marine spatial planning in high-use coastal marine environments where bottom contact fisheries access and other coastal development must be balanced against marine conservation objectives.
Dataset DOI: 10.5061/dryad.sf7m0cgjc
Description of the data and file structure
There are two data files, one for each of two benthic invertebrates that were censused using a near-seafloor optical imaging system. The species are Boltenia ovifera (file: model_data_boltenia.csv) and Modiolus modiolus (file: model_data_modiolus.csv). Within each data file, the numerical count and presence/absence of the species is listed for each photographic image analyzed from the survey data set. For each station, values are provided for eleven different environmental variables extracted from a set of raster data layers that cover the survey area domain. Twenty-two raster layers were used in initial analyses narrowed down to 11 variables from multicollinearity tests. In addition to the 11 environmental variables, a set of 15 oblique geographic coordinate values are provided for each image location. The BRTmodels.R file provides the R code used to run boosted regression tree models with oblique geographic coordinates based on the .csv files. The Raster_layers.zip file provides the raster for all 22 environmental predictors at the original 1-m resolution and resampled 5-m resolution, along with the final model outputs indicating the probability of presence and prevalence of both Boltenia ovifera and Modiolus modiolus. In the EnvVariables_5m folder the .tif files are accompanied by metadata files (.tfw,.aux,.ovr,.xml). These 5m layers were used in final species distribution model iterations, so these extra files are the metadata (i.e., spatial statistics). With these additional files the 5m layers will load in a GIS environment much quicker than a standalone tif.
Files and variables
The variables below are defined based on file name.
File: model_data_modiolus.csv
Description: The count and presence of Modiolus modiolus along with derived environmental variable values according to an analyzed image and its associated location.
Variables
- Count: Count of the number of Modiolus modiolus observed in an image.
- Presence: Presence (1) or absence (0) of Modiolus modiolus in an image.
- station: Drift transect station number (49 transects total).
- OGC1: The first oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC2: The second oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC3: The third oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC4: The fourth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC5: The fifth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC6: The sixth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC7: The seventh oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC8: The eighth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC9: The ninth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC10: The tenth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC11: The eleventh oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC12: The twelfth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC13: The thirteenth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC14: The fourteenth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC15: The fifteenth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- Bathymetry: The depth value (chart datum - m) derived from the bathymetry layer at an image location.
- Backscatter: The backscatter (db) derived from the backscatter layer at an image location.
- Northness: The northness (unitless) value from the northness layer (derived from the bathymetry layer) at an image location.
- CurrentMax: Maximum current speed (m/s) value at an image location.
- CurrentMin: Minimum current speed (m/s) value at an image location.
- ShoreDist: The distance from shore (m) value derived from coastal shapefiles at an image location.
- Habitat: Binary substrate classification (hard [1]/soft [0]) value derived from classification shapefiles (Substrate layer) at an image location.
- RDMV: The Relative Deviation from the Mean Value (RDMV - unitless) value from the RDMV layer (derived from the bathymetry layer) at an image location.
- Rugosity: The rugosity (unitless) value from the rugosity layer (derived from the bathymetry layer) at an image location.
- Eastness: The eastness (unitless) value from the eastness layer (derived from the bathymetry layer) at an image location.
- Slope: The slope (degrees) value from the slope layer at an image location.
File: model_data_boltenia.csv
Description: The count and presence of Boltenia ovifera along with derived environmental variable values according to an analyzed image and its associated location.
Variables
- Count: Count of the number of Boltenia ovifera observed in an image.
- Presence: Presence (1) or absence (0) of Boltenia ovifera in an image.
- station: Drift transect station number (49 transects total).
- OGC1: The first oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC2: The second oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC3: The third oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC4: The fourth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC5: The fifth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC6: The sixth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC7: The seventh oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC8: The eighth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC9: The ninth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC10: The tenth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC11: The eleventh oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC12: The twelfth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC13: The thirteenth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC14: The fourteenth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- OGC15: The fifteenth oblique coordinate angle computed at an image location (see Møller et al. 2020 for details).
- Bathymetry: The depth value (chart datum - m) derived from the bathymetry layer at an image location.
- Backscatter: The backscatter (db) derived from the backscatter layer at an image location.
- Northness: The northness (unitless) value from the northness layer (derived from the bathymetry layer) at an image location.
- CurrentMax: Maximum current speed (m/s) value at an image location.
- CurrentMin: Minimum current speed (m/s) value at an image location.
- ShoreDist: The distance from shore (m) value derived from coastal shapefiles at an image location.
- Habitat: Binary substrate classification (hard [1]/soft [0]) value derived from classification shapefiles (Substrate layer) at an image location.
- RDMV: The Relative Deviation from the Mean Value (RDMV - unitless) value from the RDMV layer (derived from the bathymetry layer) at an image location.
- Rugosity: The rugosity (unitless) value from the rugosity layer (derived from the bathymetry layer) at an image location.
- Eastness: The eastness (unitless) value from the eastness layer (derived from the bathymetry layer) at an image location.
- Slope: The slope (degrees) value from the slope layer at an image location.
File: BRT_models.R
Description: R file containing the R script to configure and run the presence probability boosted regression tree models using oblique geographic coordinates (OGCs, see Møller et al. 2020 for details).
File: Raster_layers.zip
Description: Raster files (.tif) of all 22 environmental variables used for initial species distribution modelling as well as the final model outputs.
Sub-file: EnvVariables_1m
- aspect_1m.tif: The aspect (identifying the direction the downhill slope faces - in radians) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Backscatter_1m.tif: The backscatter (db) raster layer at a 1-m spatial resolution from the Ocean Mapping Group (open source), clipped to the study area.
- Bathymetry_1m.tif: The bathymetry (m) raster layer at a 1-m spatial resolution from the Ocean Mapping Group (open source), clipped to the study area.
- Bathymetry_mean_1m.tif: The mean bathymetry (neighbourhood size of 3x3 - in m) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Bathymetry_sd_1m.tif: The standard deviation of the bathymetry (neighbourhood size of 3x3 - in m) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Bathymetry_slope_1m.tif: The slope (degrees) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Cos_aspect_1m.tif: The cosine of the aspect (radians) raster layer at a 1-m spatial resolution derived from the aspect layer, clipped to the study area.
- Current_max_1m.tif: The maximum current speed (m/s) raster layer at a 1-m spatial resolution derived from Page et al. 2015, clipped to the study area.
- Current_mean_1m.tif: The mean current speed (m/s) raster layer at a 1-m spatial resolution derived from Page et al. 2015, clipped to the study area.
- Current_min_1m.tif: The minimum current speed (m/s) raster layer at a 1-m spatial resolution derived from Page et al. 2015, clipped to the study area.
- Current_sd_1m.tif: The standard deviation of the current speed (m/s) raster layer at a 1-m spatial resolution derived from Page et al. 2015, clipped to the study area.
- Curvature_1m.tif: The curvature (pixel value) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Eastness_1m.tif: The eastness (unitless) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Northness_1m.tif: The northness (unitless) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- RDMV_1m.tif: The Relative Deviation from the Mean Value (RDMV - unitless) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Roughness_1m.tif: The roughness (pixel value) raster layer at a 1-m spatial resolution derived from the backscatter layer, clipped to the study area.
- Ruggedness_1m.tif: The ruggedness (unitless) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Rugosity_1m.tif: The rugosity (unitless) raster layer at a 1-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Shore_distance_1m.tif: The distance from shore (m) raster layer at a 1-m spatial resolution from open-source shapefiles, clipped to the study area.
- Sin_aspect_1m.tif: The sine of the aspect (radians) raster layer at a 1-m spatial resolution derived from the aspect layer, clipped to the study area.
- Slope_1m.tif: The slope (degrees) raster layer at a 1-m spatial resolution from the Ocean Mapping Group (open source), clipped to the study area.
- Substrate_1m.tif: The binary substrate classification (hard/soft) raster layer at a 1-m spatial resolution from open-source shapefiles, clipped to the study area.
Sub-file: EnvVariables_5m (accompanied by metadata files: .tfw, .aux, .ovr, .xml)
- aspect_5m.tif: The aspect (identifying the direction the downhill slope faces - in radians) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Backscatter_5m.tif: The backscatter (db) raster layer at a 5-m spatial resolution from the Ocean Mapping Group (open source), clipped to the study area.
- Bathymetry_5m.tif: The bathymetry (m) raster layer at a 5-m spatial resolution from the Ocean Mapping Group (open source), clipped to the study area.
- Bathymetry_mean_5m.tif: The mean bathymetry (neighbourhood size of 3x3 - in m) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Bathymetry_sd_5m.tif: The standard deviation of the bathymetry (neighbourhood size of 3x3 - in m) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Bathymetry_slope_5m.tif: The slope (degrees) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Cos_aspect_5m.tif: The cosine of the aspect (radians) raster layer at a 5-m spatial resolution derived from the aspect layer, clipped to the study area.
- Current_max_5m.tif: The maximum current speed (m/s) raster layer at a 5-m spatial resolution derived from Page et al. 2015, clipped to the study area.
- Current_mean_5m.tif: The mean current speed (m/s) raster layer at a 5-m spatial resolution derived from Page et al. 2015, clipped to the study area.
- Current_min_5m.tif: The minimum current speed (m/s) raster layer at a 5-m spatial resolution derived from Page et al. 2015, clipped to the study area.
- Current_sd_5m.tif: The standard deviation of the current speed (m/s) raster layer at a 5-m spatial resolution derived from Page et al. 2015, clipped to the study area.
- Curvature_5m.tif: The curvature (pixel value) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Eastness_5m.tif: The eastness (unitless) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Northness_5m.tif: The northness (unitless) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- RDMV_5m.tif: The Relative Deviation from the Mean Value (RDMV - unitless) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Roughness_5m.tif: The roughness (pixel value) raster layer at a 5-m spatial resolution derived from the backscatter layer, clipped to the study area.
- Ruggedness_5m.tif: The ruggedness (unitless) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Rugosity_5m.tif: The rugosity (unitless) raster layer at a 5-m spatial resolution derived from the bathymetry layer, clipped to the study area.
- Shore_distance_5m.tif: The distance from shore (m) raster layer at a 5-m spatial resolution from open-source shapefiles, clipped to the study area.
- Sin_aspect_5m.tif: The sine of the aspect (radians) raster layer at a 5-m spatial resolution derived from the aspect layer, clipped to the study area.
- Slope_5m.tif: The slope (degrees) raster layer at a 5-m spatial resolution from the Ocean Mapping Group (open source), clipped to the study area.
- Substrate_5m.tif: The binary substrate classification (hard/soft) raster layer at a 5-m spatial resolution from open-source shapefiles, clipped to the study area.
Sub-file: Model_outputs
- boltenia_brt_ogc16_prob.tif: The probability of presence (ranging from 0-1) species distribution model output raster layer of Boltenia ovifera at a 5-m spatial resolution.
- boltenia_brt_prevalence.tif: The prevalence (present/absent) species distribution model output raster layer of Boltenia ovifera at a 5-m spatial resolution.
- modiolus_brt_ogc16_prob.tif: The probability of presence (ranging from 0-1) species distribution model output raster layer of Modiolus modiolus at a 5-m spatial resolution.
- modiolus_brt_prevalence.tif: The prevalence (present/absent) species distribution model output raster layer of Modiolus modiolus at a 5-m spatial resolution.
Code/software
The file BRT_models.R contains the code for both species' distribution models.
Access information
Other publicly accessible locations of the species data:
Data was derived from the following sources:
- http://omg.unb.ca/projects/South_West_NB/html/
- https://search.open.canada.ca/openmap/f2c493e4-ceaa-11eb-be59-1860247f53e3
- http://www.snb.ca/geonb1/e/DC/catalogue-E.asp
- Page, F. H., Losier, R., Haigh, S., Bakker, J., Chang, B. D., McCurdy, P., Beattie, M., et al. 2015. Transport and dispersal of sea lice bath therapeutants from salmon farm net-pens and well-boats. DFO Can. Sci. Advis. Sec. Res. Doc. 2015/064. xviii +148 p.
Oblique geographic coordinates source:
- Møller, A. B., Beucher, A. M., Pouladi, N., and Greve, M. H. 2020. Oblique geographic coordinates as covariates for digital soil mapping. SOIL 6, 269–289. https://doi.org/10.5194/soil-6-269-2020.