Description:Spatial information on ecologically important species is needed to support marine spatial planning initiatives in British Columbia’s (BC) marine environment. For data deficient taxa, such as shark species, species distribution models that integrate presence-absence data from different sources can be used to predict their coastwide distributions. Here we provide spatial estimates of the distribution of Blue Shark (Prionace glauca), Salmon Shark (Lamna ditropis), Pacific Sleeper Shark (Somniosus pacificus) and Bluntnose Sixgill Shark (Hexanchus griseus). These estimates were generated using spatial generalized linear mixed effects models and are based on data from two scientific surveys and the commercial hook and line, midwater trawl and bottom trawl fisheries. For each species, we provide predicted probability of occurrence and prediction uncertainty at a 3 km resolution for the British Columbia coast, and parameter estimates for model covariates (depth, slope, year, data source). Results show variable predicted distributions across species, with Blue Shark and Pacific Sleeper Shark showing higher probability of presence along the continental slope, while Salmon Shark show low probability of occurrence coastwide and Bluntnose Sixgill Shark show the highest probability of occurrence in the Strait of Georgia. The results from this study can support ongoing marine spatial planning initiatives in the BC and support the conservation and management of these important species.Methods:Data Sources The species distribution models (SDMs) are based on data from two fishery independent scientific surveys and from the commercial hook and line fishery, which are all conducted within Canadian Pacific waters. The scientific surveys include the Fisheries and Oceans Canada (DFO) hard bottom longline surveys and the International Pacific Halibut Commission (IPHC) fishery-independent setline survey. The study area is bound by the outer convex hull of these three data sources. Other DFO research surveys, such as the groundfish synoptic bottom trawl surveys, midwater trawl surveys and sablefish trap surveys were investigated as potential data sources, but were found to have insufficient presence observations for the species of interest to warrant their inclusion in the analysis. For more information on the details of the source data please refer to Proudfoot et al. 2024.Modelling Approach and Comparison For each species, we fit a suite of generalized linear mixed effects models (GLMMs) using the sdmTMB package (Anderson et al. 2022). For each species, we fit four models, each with a different set of fixed effects/environmental predictors. Additionally, we compared the predictive power of four models for each species, with each model having a different combination of environmental predictors (i.e., slope, depth, slope + depth, none). A summary of the candidate models is provided in Table 2 of Proudfoot et al. 2024. For each species, we selected the model with the highest predictive accuracy (assessed using the predicted log likelihood based on the cross-validation) as the best fit.Spatial Species Distribution Predictions We made predictions of species occurrence using the selected model and a 3 km resolution spatial prediction grid. Our predictions were made for the entire BC coast, and species distribution predictions were made using models fit to the full dataset, as opposed to models fit using cross-validation. We made predictions with year set to 2014 (the approximate midpoint of the dataset) and type set to IPHC (the dataset with the most even spatial distribution of data points).Uncertainties:Because limited survey and commercial catch data exists for deep areas off the continental shelf, predictions in these areas are likely more uncertain than predictions on the shelf. To illustrate this, uncertainty (standard deviation derived from the 500 simulated values from the joint precision matrix of selected models) was mapped across the full study area for each species. Additionally, because these models are based on data that likely do not span the full spatiotemporal extent of the species’ habitat (i.e., mid depths, surface waters, and data across all seasons may not be captured), these results illustrate a snapshot of occurrence but do not account for more complex migration and movement patterns undertaken by these species.

This dataset contains two geotiff layers. The first layer (1) represents the coastwide distribution of Dungeness crab as predicted from a geostatistical model. The model predicts the mean coastwide probability of Dungeness crab detection using trap sampling gear. The second layer (2) represent the uncertainty in those predictions. Detailed descriptions of these data products can be found in Nephin et al. (2023) and the code used to produce them can be found at https://gitlab.com/dfo-msea/dungeness-sdm/.The objectives of this work was to model the habitat of Dungeness crab (_Metacarcinus magister_), a data-limited coastal marine species, to evaluate the efficacy of data integration when making predictions to geographic areas larger than the area covered by any one data source. In British Columbia, Dungeness crab are sampled regionally and sporadically with a variety of sampling gears and survey protocols, making them an ideal case study to investigate whether the integration of disparate surveys can improve habitat predictions. To that aim, we assemble data from dive, trawl, and baited-trap surveys to generate six candidate generalized linear mixed-effect models with spatial random fields. This dataset contains the mean (1) and difference (2) between the Survey-effect and Gear-effect model predictions.

A Highway Profile defines the number of through travel lanes, including passing and truck lanes, and whether the road is divided or not. It is a Linear feature

Multi-model ensembles of snow depth based on projections from twenty-eight Coupled Model Intercomparison Project Phase 5 (CMIP5) global climate models are available for 1900-2100. Specifically, the 5th, 25th, 50th, 75th and 95th percentiles of the monthly, seasonal and annual ensembles of snow depth (m) are available for the historical time period, 1900-2005, and for emission scenarios, RCP2.6, RCP4.5 and RCP8.5, for 2006-2100. Note: Projections among climate models can vary because of differences in their underlying representation of earth system processes. Thus, the use of a multi-model ensemble approach has been demonstrated in recent scientific literature to likely provide better projected climate change information.

Multi-model ensembles of mean precipitation based on projections from twenty-nine Coupled Model Intercomparison Project Phase 5 (CMIP5) global climate models are available for 1901-2100. Specifically, the 5th, 25th, 50th, 75th and 95th percentiles of the monthly, seasonal and annual ensembles of mean precipitation (mm/day) are available for the historical time period, 1901-2005, and for emission scenarios, RCP2.6, RCP4.5 and RCP8.5, for 2006-2100. Note: Projections among climate models can vary because of differences in their underlying representation of earth system processes. Thus, the use of a multi-model ensemble approach has been demonstrated in recent scientific literature to likely provide better projected climate change information.

Seasonal and annual multi-model ensembles of projected change (also known as anomalies) in sea ice thickness, based on an ensemble of twenty-six Coupled Model Intercomparison Project Phase 5 (CMIP5) global climate models are available for 1900-2100. Projected change in sea ice thickness is with respect to the reference period of 1986-2005 and expressed as a percentage (%). The 5th, 25th, 50th, 75th and 95th percentiles of the ensemble of sea ice thickness change are available for the historical time period, 1900-2005, and for emission scenarios, RCP2.6, RCP4.5 and RCP8.5, for 2006-2100. Twenty-year average changes in sea ice thickness (%) for four time periods (2021-2040; 2041-2060; 2061-2080; 2081-2100), with respect to the reference period of 1986-2005, for RCP2.6, RCP4.5 and RCP8.5 are also available in a range of formats. The median projected change across the ensemble of CMIP5 climate models is provided. Note: Projections among climate models can vary because of differences in their underlying representation of earth system processes. Thus, the use of a multi-model ensemble approach has been demonstrated in recent scientific literature to likely provide better projected climate change information.

Seasonal and annual multi-model ensembles of projected change (also known as anomalies) in surface wind speed based on an ensemble of twenty-nine Coupled Model Intercomparison Project Phase 5 (CMIP5) global climate models are available for 1900-2100. Projected change in wind speed is with respect to the reference period of 1986-2005 and expressed as a percentage (%). The 5th, 25th, 50th, 75th and 95th percentiles of the ensemble of wind speed change are available for the historical time period, 1900-2005, and for emission scenarios, RCP2.6, RCP4.5 and RCP8.5, for 2006-2100. Twenty-year average changes in wind speed (%) for four time periods (2021-2040; 2041-2060; 2061-2080; 2081-2100), with respect to the reference period of 1986-2005, for RCP2.6, RCP4.5 and RCP8.5 are also available in a range of formats. The median projected change across the ensemble of CMIP5 climate models is provided. Note: Projections among climate models can vary because of differences in their underlying representation of earth system processes. Thus, the use of a multi-model ensemble approach has been demonstrated in recent scientific literature to likely provide better projected climate change information.

Multi-model ensembles of sea ice concentration based on projections from twenty-eight Coupled Model Intercomparison Project Phase 5 (CMIP5) global climate models are available for 1900-2100. Specifically, the 5th, 25th, 50th, 75th and 95th percentiles of the monthly, seasonal and annual ensembles of sea ice concentration as represented as the percentage (%) of grid cell area, are available for the historical time period, 1900-2005, and for emission scenarios, RCP2.6, RCP4.5 and RCP8.5, for 2006-2100. Note: Projections among climate models can vary because of differences in their underlying representation of earth system processes. Thus, the use of a multi-model ensemble approach has been demonstrated in recent scientific literature to likely provide better projected climate change information.

Seasonal and annual multi-model ensembles of projected change (also known as anomalies) in snow depth based on an ensemble of twenty-eight Coupled Model Intercomparison Project Phase 5 (CMIP5) global climate models are available for 1900-2100. Projected change in snow depth is with respect to the reference period of 1986-2005 and expressed as a percentage (%). The 5th, 25th, 50th, 75th and 95th percentiles of the ensemble of snow depth change are available for the historical time period, 1900-2005, and for emission scenarios, RCP2.6, RCP4.5 and RCP8.5, for 2006-2100. Twenty-year average changes in snow depth (%) for four time periods (2021-2040; 2041-2060; 2061-2080; 2081-2100), with respect to the reference period of 1986-2005, for RCP2.6, RCP4.5 and RCP8.5 are also available in a range of formats. The median projected change across the ensemble of CMIP5 climate models is provided. Note: Projections among climate models can vary because of differences in their underlying representation of earth system processes. Thus, the use of a multi-model ensemble approach has been demonstrated in recent scientific literature to likely provide better projected climate change information.

Multi-model ensembles of sea ice thickness based on projections from twenty-six Coupled Model Intercomparison Project Phase 5 (CMIP5) global climate models are available for 1900-2100. Specifically, the 5th, 25th, 50th, 75th and 95th percentiles of the monthly, seasonal and annual ensembles of sea ice thickness (m) are available for the historical time period, 1900-2005, and for emission scenarios, RCP2.6, RCP4.5 and RCP8.5, for 2006-2100. Note: Projections among climate models can vary because of differences in their underlying representation of earth system processes. Thus, the use of a multi-model ensemble approach has been demonstrated in recent scientific literature to likely provide better projected climate change information.