Species distribution models (SDMs) are tools that combine species observations of occurrence, abundance, or biomass with environmental variables to predict the distribution of a species in unsampled locations. To produce accurate predictions of occurrence, abundance or biomass distribution, a wide range of physical and/or biological variables is desirable. Such data is often collected over limited or irregular spatial scales, and require the application of geospatial techniques to produce continuous environmental surfaces that can be used for modelling at all spatial scales. Here we provide a review of 102 environmental data layers that were compiled for the entire spatial extent of Fisheries and Oceans Canada’s (DFO) Maritimes Region. Variables were obtained from a broad range of physical and biological data sources and spatially interpolated using geostatistical methods. For each variable we document the underlying data distribution, provide relevant diagnostics of the interpolation models and an assessment of model performance, and present the final standard error and interpolation surfaces. These layers have been archived in a common (raster) format at the Bedford Institute of Oceanography to facilitate future use. Based on the diagnostic summaries in this report, a subset of these variables has subsequently been used in species distribution models to predict the distribution of deep-water corals, sponges, and other significant benthic taxa in the Maritimes Region.Cite this data as: Beazley, Lindsay; Guijarro, Javier, Lirette; Camille; Wang, Zeliang; Kenchington, Ellen (2020). Characteristics of Environmental Data Layers for Use in Species Distribution Modelling in the Maritimes Region. Published July 2023. Ocean Ecosystems Science Division, Fisheries and Oceans Canada, Dartmouth, N.S. https://open.canada.ca/data/en/dataset/34a917cb-a0e3-403c-91c7-af3dc20628b1

Description:This dataset contains layers of predicted occurrence for 65 groundfish species as well as overall species richness (i.e., the total number of species present) in Canadian Pacific waters, and the median standard error per grid cell across all species. They cover all seafloor habitat depths between 10 and 1400 m that have a mean summer salinity above 28 PSU. Two layers are provided for each species: 1) predicted species occurrence (prob_occur) and 2) the probability that a grid cell is an occurrence hotspot for that species (hotspot_prob; defined as being in the lower of: 1) 0.8, or 2) the 80th percentile of the predicted probability of occurrence values across all grid cells that had a probability of occurrence greater than 0.05.). The first measure provides an overall prediction of the distribution of the species while the second metric identifies areas where that species is most likely to be found, accounting for uncertainty within our model. All layers are provided at a 1 km resolution.Methods:These layers were developed using a species distribution model described in Thompson et al. 2023. This model integrates data from three fisheries-independent surveys: the Fisheries and Oceans Canada (DFO) Groundfish Synoptic Bottom Trawl Surveys (Sinclair et al. 2003; Anderson et al. 2019), the DFO Groundfish Hard Bottom Longline Surveys (Lochead and Yamanaka 2006, 2007; Doherty et al. 2019), and the International Pacific Halibut Commission Fisheries Independent Setline Survey (IPHC 2021). Further details on the methods are found in the metadata PDF available with the dataset.Abstract from Thompson et al. 2023:Predictions of the distribution of groundfish species are needed to support ongoing marine spatial planning initiatives in Canadian Pacific waters. Data to inform species distribution models are available from several fisheries-independent surveys. However, no single survey covers the entire region and different gear types are required to survey the range of habitats that are occupied by groundfish. Bottom trawl gear is used to sample soft bottom habitat, predominantly on the continental shelf and slope, whereas longline gear often focuses on nearshore and hardbottom habitats where trawling is not possible. Because data from these two gear types are not directly comparable, previous species distribution models in this region have been limited to using data from one survey at a time, restricting their spatial extent and usefulness at a regional scale. Here we demonstrate a method for integrating presence-absence data across surveys and gear types that allows us to predict the coastwide distributions of 66 groundfish species in British Columbia. Our model leverages the use of available data from multiple surveys to estimate how species respond to environmental gradients while accounting for differences in catchability by the different surveys. Overall, we find that this integrated method has two main benefits: 1) it increases the accuracy of predictions in data-limited surveys and regions while having negligible impacts on the accuracy when data are already sufficient to make predictions, 2) it reduces uncertainty, resulting in tighter confidence intervals on predicted species occurrences. These benefits are particularly relevant in areas of our coast where our understanding of habitat suitability is limited due to a lack of spatially comprehensive long-term groundfish research surveys.Data Sources:Research data was provided by Pacific Science’s Groundfish Data Unit for research surveys from the GFBio database between 2003 and 2020 for all species which had at least 150 observations, across all gear type and survey datasets available.Uncertainties:These are modeled results based on species observations at sea and their related environmental covariate predictions that may not always accurately reflect real-world groundfish distributions though methods that integrate different data types/sources have been demonstrated to improve model inference by increasing the accuracy of the predictions and reducing uncertainty.

Monitoring data from DFO invasive species monitoring programs, along with occurrence information from online databases and the scientific literature, have been paired with high resolution environmental data and oceanographic models in species distribution models that predict present-day and project future distributions of 24 non-indigenous species (NIS) on North America`s east coast, and 31 NIS on its west coast. Future distributions were predicted for 2100, under Representative Concentration Pathway 8.5 from the Intergovernmental Panel on Climate Change’s fifth Assessment Report. Present-day and future richness of these species (i.e., hotspots) have been estimated by summing the occurrence probabilities of NIS. This data set includes the present-day and year 2100 species distribution modeling results for each species, and the estimated species richness.Cite this data as: Lyons DA., Lowen JB, Therriault TW., Brickman D., Guo L., Moore AM., Peña MA., Wang Z., DiBacco C. Data of: Updated species distribution models for marine invasive species hotspot identification. Published: November 2023. Coastal Ecosystems Science Division, Fisheries and Oceans Canada, Dartmouth, N.S. https://open.canada.ca/data/en/dataset/1439dcb3-82a6-40fd-a9a4-8f045b20ff5b

A spatial representation of the general distribution of wild mountain sheep (bighorn and thinhorn sheep) in British Columbia. Populations that extend into neighbouring provinces and states are also included. The distribution polygons are divided by species into bighorn and thinhorn sheep.

Since 2005, Fisheries and Oceans Canada has been collecting monitoring data for aquatic invasive species (e.g. https://open.canada.ca/data/en/dataset/8d87f574-0661-40a0-822f-e9eabc35780d, https://open.canada.ca/data/en/dataset/503a957e-7d6b-11e9-aef3-f48c505b2a29, https://open.canada.ca/data/en/dataset/8661edcf-f525-4758-a051-cb3fc8c74423). This monitoring data, as well additional occurrence information from online databases and the scientific literature, have been paired with high resolution environmental data and oceanographic models in species distribution models that predict the present-day and future potential distributions of 12 moderate to high risk invasive species on Canada’s east and west coasts. Future distributions were predicted for 2075, under Representative Concentration Pathway 8.5 from the Intergovernmental Panel on Climate Change’s fifth Assessment Report. Present-day and future richness of these species (i.e., hotspots) has also been estimated by summing their occurrence probabilities. This data set includes the occurrence locations of each species, the present-day and future species distribution modeling results for each species, and the estimated species richness. This research has been published in the scientific literature(Lyons et al. 2020).Lyons DA, Lowen JB, Therriault TW, Brickman D, Guo L, Moore AM, Peña MA, Wang Z, DiBacco C. (In Press) Identifying Marine Invasion Hotspots Using Stacked Species Distribution Models. Biological InvasionsCite this data as: Lyons DA., Lowen JB, Therriault TW., Brickman D., Guo L., Moore AM., Peña MA., Wang Z., DiBacco C. Data of: Species distribution models and occurrence data for marine invasive species hotspot identification. Published: November 2020. Coastal Ecosystems Science Division, Fisheries and Oceans Canada, Dartmouth, N.S. https://open.canada.ca/data/en/dataset/1bbd5131-8b34-4245-b999-3b4c4259d74f

DescriptionConservation of marine biodiversity requires understanding the joint influence of ongoing environmental change and fishing pressure. Addressing this challenge requires robust biodiversity monitoring and analyses that jointly account for potential drivers of change. Here, we ask how demersal fish biodiversity in Canadian Pacific waters has changed since 2003 and assess the degree to which these changes can be explained by environmental change and commercial fishing. Using a spatiotemporal multispecies model based on fisheries independent data, we find that species density (number of species per area) and community biomass have increased during this period. Environmental changes during this period were associated with temporal fluctuations in the biomass of species and the community as a whole. However, environmental changes were less associated with changes in species’ occurrence. Thus, the estimated increases in species density are not likely to be due to environmental change. Instead, our results are consistent with an ongoing recovery of the demersal fish community from a reduction in commercial fishing intensity from historical levels. These findings provide key insight into the drivers of biodiversity change that can inform ecosystem-based management.The layers provided represent three community metrics: 1) species density (i.e., species richness), 2) Hill-Shannon diversity, and 3) community biomass. All layers are provided at a 3 km resolution across the study domain for the period of 2003 to 2019. For each metric, we provide layers for three summary statistics: 1) the mean value in each grid cell over the temporal range, 2) the probability that the grid cell is a hotspot for that metric, and 3) the temporal coefficient of variation (i.e., standard deviation/mean) across all years.Methods:The analysis that produced these layers is presented in Thompson et al. (2022). The analysis uses data from the Groundfish Synoptic Bottom Trawl Research surveys in Queen Charlotte Sound (QCS), Hecate Strait (HS), West Coast Vancouver Island (WCVI), and West Coast Haida Gwaii (WCHG) from 2003 to 2019. Cartilaginous and bony fish species caught in DFO groundfish surveys that were present in at least 15% of all trawls over the depth range in which they were caught were included. This depth range was defined as that which included 95% of all trawls in which that species was present. The final dataset used in our analysis consisted of 57 species (Table S1 in Thompson et al. 2022).The spatiotemporal dynamics of the demersal fish community were modeled using the Hierarchical Modeling of Species Communities (HMSC) framework and package (Tikhonov et al. 2021) in R. This framework uses Bayesian inference to fit a multivariate hierarchical generalized mixed model. We modeled community dynamics using a hurdle model, which consists of two sub models: a presence-absence model and a biomass model that is conditional on presence. Our list of environmental covariates included bottom depth, bathymetric position index (BPI), mean summer tidal speed, substrate muddiness, substrate rockiness, whether the trawl was inside or outside of the ecosystem-based trawling footprint, and survey region (QCS & HS vs. WCVI & WCHG)), mean summer near-bottom temperature deviation, mean summer near-bottom dissolved oxygen deviation, mean summer cross-shore and along-shore current velocities near the seafloor, mean summer depth-integrated primary production, and local-scale commercial fishing effort.Layers are provided for three community metrics. All metrics should be interpreted as the value that would be expected in the catch from an average tow in the Groundfish Synoptic Bottom Trawl Research Surveys taken in a given 3 km grid cell. Species density (sometimes called species richness) should be interpreted as the number of the 57 species that would be caught in a trawl. Hill-Shannon diversity is a measure of diversity that gives greater weight to communities where biomass is spread equally across species. Community biomass is the total biomass across all 57 species that would be expected to be caught per square km in an average tow. Data Sources:Research data was provided by Pacific Science's Groundfish Data Unit for research surveys from the GFBio database between 2003 and 2019 that occurred in four regions: Queen Charlotte Sound, Hecate Strait, West Coast Haida Gwaii, and West Coast Vancouver Island. Our analysis excludes species that are rarely caught in the research trawls and so our estimates would not include the occurrence or biomass of these rare species.Commercial fishing data was accessed through a DFO R script detailed here: https://github.com/pbsassess/gfdata. Local scale commercial fishing effort was calculated from this data. The substrate layers were obtained from a substrate model (Gregr et al. 2021). The oceanographic layers (bottom temperature, dissolved oxygen, tidal and circulation speeds, primary production) were obtained from a hindcast simulation of the British Columbia continental margin (BCCM) model (Peña et al. 2019).Uncertainties:Species that are not well sampled by the trawl surveys may not be accurately estimated by our model. The model did not include spatiotemporal random effects, which likely underestimates spatiotemporal variability in the region. It is also important to underline covariate uncertainty and model uncertainty. The hotspot estimates provide one measure of model uncertainty/certainty.

Kernel density estimation (KDE) utilizes spatially explicit data to model the distribution of a variable of interest. It is a simple non-parametric neighbour-based smoothing function that relies on few assumptions about the structure of the observed data. It has been used in ecology to identify hotspots, that is, areas of relatively high biomass/abundance, and in 2010 was used by Fisheries and Oceans Canada to delineate significant concentrations of corals and sponges. The same approach has been used successfully in the Northwest Atlantic Fisheries Organization (NAFO) Regulatory Area. Here, we update the previous analyses with the catch records from up to 5 additional years of trawl survey data from Eastern Canada, including the Gulf of Saint Lawrence. We applied kernel density estimation to create a modelled biomass surface for each of sponges, small and large gorgonian corals, and sea pens, and applied an aerial expansion method to identify significant concentrations of these taxa. We compared our results to those obtained previously and provided maps of significant concentrations as well as point data co-ordinates for catches above the threshold values used to construct the significant area polygons. The borders of the polygons can be refined using knowledge of null catches and species distribution models of species presence/absence and/or biomass.

In this dataset, we share maps of annual dominant tree species (also known as leading tree species) from 1984-2022 covering the entirety of Canada's 650 Mha forested ecosystems using Landsat time-series imagery at a 30-m spatial resolution. It is developed within the framework of Canada’s National Terrestrial Ecosystem Monitoring System (NTEMS). Classifications are based on regionally representative Random Forests model using local training samples from Canada's National Forest Inventory (Hermosilla et al., 2024). Descriptive metrics provide information on spectral, geographic, climatic, and topographic characteristics. Initial annual tree species classifications were subjected to a time series post-classification process using the forward-backward Hidden Markov Model to improve the temporal consistency of tree species transitions within the time series. Assessment of the annual species maps using independent validation data resulted in an overall accuracy of 86.1% ± 0.14% (95%-confidence interval). These data allow consistent comparison of trends and rates of change in tree species composition nationally and across regions using a common time frame, spatial resolution, and analytical approach.Hermosilla, T., Wulder, M.A., White, J.C., Coops, N.C., Bater, C.W., Hobart, G.W., 2024. Characterizing long-term tree species dynamics in Canada's forested ecosystems using annual time series remote sensing data. Forest Ecology and Management, 122313. https://doi.org/10.1016/j.foreco.2024.122313 (Hermosilla et al. 2024)

High-resolution map of leading tree species distribution for Canada’s forested ecosystems (2019). Leading tree species map produced from a 2019 Landsat image composite, geographic and climate data, elevation derivatives, and remote sensing derived phenology following the framework described in Hermosilla et al. (2022). Regional classification models were generated based on Canada’s National Forest Inventory using a 150x150 km tiling system. The leading tree species are defined by representing the most voted tree species from the Random Forests classification models (i.e. the class with the highest class membership probability).The data represents leading tree species of Canada's forested ecosystems in 2019. An image compositing window of August 1 ± 30 days was used to generate the best-available-pixel (BAP) image composites utilized as source data for the classification.The science and methods developed to generate the information outcomes shown here, that track and characterize the history of Canada’s forests, were led by Canadian Forest Service of Natural Resources Canada, developed within the framework of Canada’s National Terrestrial Ecosystem Monitoring System (NTEMS), partnered with the University of British Columbia, augmented by processing capacity from Digital Research Alliance of Canada.For an overview on the data, image processing, and methods applied, as well as information on independent accuracy assessment of the data, see Hermosilla et al. (2022) https://doi.org/10.1016/j.rse.2022.113276When using this data, please cite as: Hermosilla, T., Bastyr, A., Coops, N.C., White, J.C., Wulder, M.A., 2022. Mapping the presence and distribution of tree species in Canada’s forested ecosystems. Remote Sensing of Environment 282, 113276.

Effective fisheries and habitat management processes require knowledge of the distribution of areas of high ecological or biological significance. On the Scotian Shelf and Slope, a number of benthic ecologically or biologically significant areas consisting of habitat-forming species such as sponges and deep-water corals have been identified. However, knowledge of their spatial distribution is largely based on targeted surveys that are limited in their spatial extent. We used a species distribution modelling approach called random forest (RF) to predict the probability of occurrence and biomass of sponges, sea pens, and large and small gorgonian corals across the entire spatial extent of Fisheries and Oceans Canada’s (DFO) Maritimes Region. We also modelled the rare sponge Vazella pourtalesi, which forms the largest known aggregation of its kind on the Scotian Shelf. We utilized a number of data sources including DFO multispecies trawl catch data and in situ benthic imagery observations. Most models had excellent predictive capacity with cross-validated Area Under the Receiver Operating Characteristic Curve (AUC) values ranging from 0.760 to 0.977. Areas of suitable habitat were identified for each taxon and were contrasted against their known distribution and when applicable, the location of closure areas designated for their protection. Generalized additive models (GAMs) were developed to predict the biomass distribution of each taxonomic group and serve as a comparison to the RF models. The RF and GAM models provided comparable results, although GAMs provided superior predictions of biomass along the continental slope for some taxonomic groups. In the absence of data observations, the results of this study could be used to identify the potential distribution of sensitive benthic taxa for use in fisheries and habitat management applications. These results could also be used to refine significant concentrations of these taxa as identified through the kernel density analyses.Cite this data as: Beazley, Lindsay; Kenchington, Ellen; Murillo-Perez, Javier; Lirette, Camille; Guijarro-Sabaniel, Javier; McMillan, Andrew; Knudby, Anders (2019). Species Distribution Modelling of Corals and Sponges in the Maritimes Region for Use in the Identification of Significant Benthic Areas. Published July 2023. Ocean Ecosystems Science Division, Fisheries and Oceans Canada, Dartmouth, N.S. https://open.canada.ca/data/en/dataset/356e92f3-5bf3-4810-98b1-3e10cd7742aa