The objective of the study was to describe the spatial distribution of krill in eastern Canadian waters using a statistical modelling approach in support of the identification of important habitat for the western North Atlantic (WNA) blue whale (Balaenoptera musculus). Generalized Additive Models (GAMs) were used to predict ‘Significant Aggregations of Krill’ (SAK), i.e., areas where dense krill aggregations would have a greater probability of occurring. SAK cover less than 2% of the entire spatial domain and their location varied among krill categories and seasons. These SAK are interpreted as areas where environmental conditions promote krill aggregation on a regular basis and therefore are potentially important for WNA blue whale foraging in eastern Canadian waters. Plourde, S., Lehoux, C., McQuinn, I.H., and Lesage, V. 2016. Describing krill distribution in the western North Atlantic using statistical habitat models. DFO Can. Sci. Advis. Sec. Res. Doc. 2016/111. v + 34 p.

Fish Habitat Assessment Output: 1 of 16 High Water Level (75.4m ASL) - Spawning Habitat - High Vegetation Association Species (All Temperature Windows) Habitat suitability was assessed for the Bay of Quinte Area of Concern, at a 3 m grid resolution, using the Habitat Ecosystem Assessment Tool (HEAT), temperature algorithms, vegetation models, and water level input. Habitat classifications were based on three variables: depth (elevation), vegetation, and substrate; and modified by temperature suitabilities. The final suitability maps were based on documented habitat and temperature associations for the fish in the area. Different life stages (spawning requirements, nursery habitat, adult habitat) were modeled for the years of 1972-2011. Suitability values were scaled from 0 (not suitable) to 1 (highly suitable) and converted to suitability classes of very low, low, medium, and high. The final maps for each guild – life stage combination are maximum suitability values from the 39-year period modelled.

Canada's National Forest Inventory (NFI) sampling program is designed to support reporting on forests at the national scale. On the other hand, continuous maps of forest attributes are required to support strategic analyses of regional policy and management issues. We have therefore produced maps covering 4.03 × 106 km2 of inventoried forest area for the 2001 base year using standardised observations from the NFI photo plots (PP) as reference data. We used the k nearest neighbours (kNN) method with 26 geospatial data layers including MODIS spectral data and climatic and topographic variables to produce maps of 127 forest attributes at a 250 × 250 m resolution. The stand-level attributes include land cover, structure, and tree species relative abundance. In this article, we report only on total live aboveground tree biomass, with all other attributes covered in the supplementary data (http://nrcresearchpress.com/doi/suppl/10.1139/cjfr-2013-0401). In general, deviations in predicted pixel-level values from those in a PP validation set are greater in mountainous regions and in areas with either low biomass or sparse PP sampling. Predicted pixel-level values are overestimated at small observed values and underestimated at large ones. Accuracy measures are improved through the spatial aggregation of pixels to 1 km2 and beyond. Overall, these new products provide unique baseline information for strategic-level analyses of forests (https://nfi.nfis.org) Collection: - **[Canada's National Forest Inventory (NFI) 2006](https://open.canada.ca/data/en/dataset/e2fadaeb-3106-4111-9d1c-f9791d83fbf4)**

This dataset provides wall-to-wall maps of forest structure across Canada's 650 million hectare forested ecosystems for the year 2022, generated at a spatial resolution of 30 m. It is developed within the framework of Canada’s National Terrestrial Ecosystem Monitoring System (NTEMS). Structure estimates include key attributes such as canopy height, canopy cover, and aboveground biomass, derived using a combination of airborne lidar and Landsat-based spectral composites. Structure models were trained using the - lidar-plot framework - (Wulder et al. 2012), which integrates co-located airborne lidar data and ground plot measurements with Landsat time-series composites (Hermosilla et al. 2016). A Nearest Neighbour imputation approach was applied to estimate structural attributes across the full extent of Canada's forested area. These nationally consistent products are intended to support strategic-level forest monitoring and assessment and are not designed for operational forest management. For further details on the methods, accuracy assessment, and source data, see Matasci et al. (2018). Matasci, G., Hermosilla, T., Wulder, M.A., White, J.C., Coops, N.C., Hobart, G.W., Bolton, D.K., Tompalski, P., Bater, C.W., 2018. Three decades of forest structural dynamics over Canada's forested ecosystems using Landsat time-series and lidar plots. Remote Sensing of Environment, 216, 697-714. https://doi.org/10.1016/j.rse.2018.07.024 (Matasci et al. 2018)

Dominant Species Map 2015 The data represent dominant tree species for British Columbia forests in 2015, are based upon Landsat data and modeling, with results mapped at 30 m spatial resolution. It is developed within the framework of Canada’s National Terrestrial Ecosystem Monitoring System (NTEMS). The map was generated with the Random Forests classifier that used predictor variables derived from Landsat time series including surface reflectance, land cover, forest disturbance, and forest structure, and ancillary variables describing the topography and position. Training and validation samples were derived from the Vegetation Resources Inventory (VRI), from a pool of polygons with homogeneous internal conditions and with low discrepancies with the remotely sensed predictions. Local models were applied over 100x100 km tiles that considered training samples from the 5x5 neighbouring tiles to avoid edge effects. An overall accuracy of 72% was found for the species which occupy 80% of the forested areas. Satellite data and modeling have demonstrated the capacity for up-to-date, wall-to-wall, forest attribute maps at sub-stand level for British Columbia, Canada. BC Species Likelihood 2015 The tree species class membership likelihood distribution data included in this product focused on the province of British Columbia, based upon Landsat data and modeling, with results mapped at 30 m spatial resolution. The data represent tree species class membership likelihood in 2015. The map was generated with the Random Forests classifier that used predictor variables derived from Landsat time series including surface reflectance, land cover, forest disturbance, and forest structure, and ancillary variables describing the topography and position. Training and validation samples were derived from the Vegetation Resources Inventory (VRI) selecting from a stratified pool of polygons with homogeneous internal conditions and with low discrepancies when related to remotely sensed information. Local models were applied over 100x100 km tiles that, to avoid edge effects, considered training samples from the 5x5 neighbouring tiles. An overall accuracy of 72% was found for the species which occupy 80% of the forested areas. As an element of the mapping process, we also obtain the votes received for each class by the Random Forest models. The votes can be understood as analogous to class membership likelihoods, providing enriched information on land cover class uncertainty for use in modeling. Tree species class membership likelihoods lower than 5% have been masked and converted to zero. When using this data, please cite as: Shang, C., Coops, N.C., Wulder, M.A., White, J.C., Hermosilla, T., 2020. Update and spatial extension of strategic forest inventories using time series remote sensing and modeling. International Journal of Applied Earth Observation and Geoinformation 84, 101956. DOI: 10.1016/j.jag.2019.101956 ( Shang et al. 2020).

Forest Elevation(Ht) Covariance 2015 Coefficient of variation of first returns height (%). Represents the variability in canopy heights relative to the mean canopy height. It is developed within the framework of Canada’s National Terrestrial Ecosystem Monitoring System (NTEMS). Products relating the structure of Canada's forested ecosystems have been generated and made openly accessible. The shared products are based upon peer-reviewed science and relate aspects of forest structure including: (i) metrics calculated directly from the lidar point cloud with heights normalized to heights above the ground surface (e.g., canopy cover, height), and (ii) modelled inventory attributes, derived using an area-based approach generated by using co-located ground plot and ALS data (e.g., volume, biomass). Forest structure estimates were generated by combining information from lidar plots (Wulder et al. 2012) with Landsat pixel-based composites (White et al. 2014; Hermosilla et al. 2016) using a nearest neighbour imputation approach with a Random Forests-based distance metric. These products were generated for strategic-level forest monitoring information needs and are not intended to support operational-level forest management. All products have a spatial resolution of 30 m. For a detailed description of the data, methods applied, and accuracy assessment results see Matasci et al. (2018). When using this data, please cite as follows: Matasci, G., Hermosilla, T., Wulder, M.A., White, J.C., Coops, N.C., Hobart, G.W., Bolton, D.K., Tompalski, P., Bater, C.W., 2018b. Three decades of forest structural dynamics over Canada's forested ecosystems using Landsat time-series and lidar plots. Remote Sensing of Environment 216, 697-714. Matasci et al. 2018) Geographic extent: Canada's forested ecosystems (~ 650 Mha) Time period: 1985–2011

High-resolution annual forest land cover maps for Canada's forested ecosystems (1984-2022). It is developed within the framework of Canada’s National Terrestrial Ecosystem Monitoring System (NTEMS). The annual time series of forest land cover maps are national in scope (entire 650 million hectare forested ecosystem) and represent a wall-to-wall land cover characterization yearly from 1984 to 2022. These time-series land cover maps were produced from annual time-series of Landsat image composites, forest change information, and ancillary topographic and hydrologic data following the framework described in Hermosilla et al. (2022), which builds upon the approach introduced in Hermosilla et al. (2018). The methodological innovations included (i) a refined training pool derived from existing land cover products using airborne and spaceborne measures of forest structure; (ii) selection of training samples proportionally to the land cover distribution using a distance-weighted approach; and (iii) generation of regional classification models using a 150x150 km tiling system. Maps are post-processed using disturbance information to ensure logical class transitions over time using a Hidden Markov Model. Hidden Markov Models assess individual year class likelihoods to reduce variability and possible noise in year-on-year class assignments (for instances when class likelihoods are similar). Hermosilla, T., Wulder, M.A., White, J.C., Coops, N.C., 2022. Land cover classification in an era of big and open data: Optimizing localized implementation and training data selection to improve mapping outcomes. Remote Sensing of Environment. Vol. 268, No. 112780. https://doi.org/10.1016/j.rse.2021.112780. (Hermosilla et al. 2022) Hermosilla, T., M.A. Wulder, J.C. White, N.C. Coops, G. W. Hobart, (2018). Disturbance-Informed Annual Land Cover Classification Maps of Canada's Forested Ecosystems for a 29-Year Landsat Time Series. Canadian Journal of Remote Sensing. 44(1) 67-87.DOI: 10.1080/07038992.2018.1437719 (Hermosilla et al. 2018).

This product is a 1km resolution composite over the North American domain, which, for areas with radar coverage, can distinguish the occurrence, type and intensity of precipitation. This product uses two 1km radar composites as input: a North American composite cleaned using dual polarization technology, another particle classification radar composite (precipitation) and surface temperature from the High Resolution Deterministic Prediction System (HRDPS). The SPTP product is produced every 6 minutes.

100 Class - Canadian Ecological Domain Classification from Satellite Data. Satellite derived data including 1) topography, 2) landscape productivity based on photosynthetic activity, and 3) land cover were used as inputs to create an environmental regionalization of the over 10 million km2 of Canada’s terrestrial land base. The outcomes of this clustering consists of three main outputs. An initial clustering of 100 classes was generated using a two-stage multivariate classification process. Next, an agglomerative hierarchy using a log-likelihood distance measure was applied to create a 40 and then a 14 class regionalization, aimed to meaningfully group ecologically similar components of Canada's terrestrial landscape. For more information (including a graphical illustration of the cluster hierarchy) and to cite this data please use: Coops, N.C., Wulder, M.A., Iwanicka, D. 2009. An environmental domain classification of Canada using earth observation data for biodiversity assessment. Ecological Informatics, Vol. 4, No. 1, Pp. 8-22, DOI: https://doi.org/10.1016/j.ecoinf.2008.09.005. ( Coops et al. 2009).