Search

Large-area land cover maps are produced to satisfy different information needs. Land cover maps having partial or complete spatial and/or temporal overlap, different legends, and varying accuracies for similar classes, are increasingly common. To address these concerns and combine two 30-m resolution land cover products, we implemented a harmonization procedure using a Latent Dirichlet Allocation (LDA) model. The LDA model used regionalized class co-occurrences from multiple maps to generate a harmonized class label for each pixel by statistically characterizing land attributes from the class co-occurrences. We evaluated multiple harmonization approaches: using the LDA model alone and in combination with more commonly used information sources for harmonization (i.e. error matrices and semantic affinity scores). The results were compared with the benchmark maps generated using simple legend crosswalks and showed that using LDA outputs with error matrices performed better and increased harmonized map overall accuracy by 6–19% for areas of disagreement between the source maps. Our results revealed the importance of error matrices to harmonization, since excluding error matrices reduced overall accuracy by 4–20%. The LDA-based harmonization approach demonstrated in this paper is quantitative, transparent, portable, and efficient at leveraging the strengths of multiple land cover maps over large areas.

Operational use of the ecosystem service (ES) concept in conservation and planning requires quantitative assessments based on accurate mapping of ESs. Our goal is to review spatial assessments of ESs, with an emphasis on the socioecological drivers of ESs, the spatial datasets commonly used to represent those drivers, and the methodological approaches used to spatially model ESs. We conclude that diverse strategies, integrating both spatial and aspatial data, have been used to map ES supply and human demand. Model parameters representing abiotic ecosystem properties can be supported by use of well-developed and widely available spatial datasets. Land-cover data, often manipulated or subject to modeling in a GIS, is the most common input for ES modeling; however, assessments are increasingly informed by a mechanistic understanding of the relationships between drivers and services. We suggest that ES assessments are potentially weakened by the simplifying assumptions often needed to translate between conceptual models and widely used spatial data. Adoption of quantitative spatial data that more directly represent ecosystem properties may improve parameterization of mechanistic ES models and increase confidence in ES assessments.

Understanding species–habitat relationships is critical for wildlife management, providing information on habitat requirements, distribution, and potential land use impacts. In this paper, we examined occurrence and abundance–habitat relationships for moose (Alces alces), a species of economic and ecological importance, across its range in southern and central Ontario. We (1) evaluated and tested competing hypotheses for predicting moose distribution and abundance, and (2) examined the ability of remotely derived environmental indicators to characterize and extrapolate moose habitat throughout the Ontario moose range. To do so, remotely sensed environmental indicators – including vegetation productivity calculated using the Dynamic Habitat Index (DHI), land cover, topography, snow cover, and natural and anthropogenic disturbances – were used to estimate moose occurrence and abundance derived from moose aerial survey data. A 2-step Hurdle model was used to accommodate for zero-inflated data, separately modeling moose occurrence and abundance in a common model framework. Our results indicate that remotely sensed indicators are able to estimate moose occurrence with a moderate degree of certainty; however, these environmental indicators did not successfully estimate moose abundance. The approach outlined in this paper provides a useful framework for hypothesis testing of remote sensing environmental drivers at broad scales, as well as for estimating moose occurrence at the regional level.

Boreal forests maintain regionally important biodiversity and globally important ecosystem services, such as carbon storage and freshwater resources. Many boreal systems have limited anthropogenic disturbances and are preserved, in effect, to date largely by their harsh climates and remoteness. As of 2011, almost 10% of Canada is subject to some manner of formal protection, with 4.5% of this protected area found within the boreal zone. The management of existing parks and protected areas (PPAs) is shared amongst many federal, provincial, and territorial jurisdictions. Although there are currently low levels of anthropogenic development in some portions of the boreal zone (especially the north), if expansion of protected areas is of interest, there are challenges to traditional PPA networks that may be more prominent in the boreal zone than elsewhere: (1) the boreal zone is home to charismatic mammal species with area requirements much larger than typical PPAs; (2) the boreal zone is characterized by natural disturbance regimes that impact large areas; and (3) projected changes to climate for the boreal zone are among the greatest in the world, creating temporal considerations for conservation planning exercises. There is currently no PPA assessment specific to boreal Canada. To address this lack of an assessment, we developed a conservation gap analysis of the current PPA system with respect to a variety of environmental surrogates (ecozones, land cover, vegetation productivity, and landscape structure). The amount of formally protected land varied within each surrogate, with few commonly reported features meeting national or international conservation targets. Furthermore, few reserves met the areal requirements that have been previously recommended to protect large mammals or accommodate the disturbance regimes present. We also discuss considerations and implications of area-based versus value-based protection objectives. While recognizing that there are still sci

Ecological and conservation research has provided a strong scientific underpinning to the modeling of ecosystem services (ESs) over space and time, by identifying the ecological processes and components of biodiversity (ecosystem service providers, functional traits) that drive ES supply. Despite this knowledge, efforts to map the distribution of ESs often rely on simple spatial surrogates that provide incomplete and non-mechanistic representations of the biophysical variables they are intended to proxy. However, alternative data sets are available that allow for more direct, spatially nuanced inputs to ES mapping efforts. Many spatially explicit, quantitative estimates of biophysical parameters are currently supported by remote sensing, with great relevance to ES mapping. Additional parameters that are not amenable to direct detection by remote sensing may be indirectly modeled with spatial environmental data layers. We review the capabilities of modern remote sensing for describing biodiversity, plant traits, vegetation condition, ecological processes, soil properties, and hydrological variables and highlight how these products may contribute to ES assessments. Because these products often provide more direct estimates of the ecological properties controlling ESs than the spatial proxies currently in use, they can support greater mechanistic realism in models of ESs. By drawing on the increasing range of remote sensing instruments and measurements, data sets appropriate to the estimation of a given ES can be selected or developed. In so doing, we anticipate rapid progress to the spatial characterization of ecosystem services, in turn supporting ecological conservation, management, and integrated land use planning.

Ecosystem classifications map an area into relatively homogenous units for environmental research, monitoring, and management. However, their effectiveness is rarely tested. Here, three classifications are (1) defined and characterized for Canada along summertime productivity (moderate-resolution imaging spectrometer fraction of absorbed photosynthetically active radiation) and wintertime snow conditions (special sensor microwave/imager snow water equivalent), independently and in combination, and (2) comparatively evaluated to determine the ability of each classification to represent the spatial and environmental patterns of alternative schemes, including the Canadian ecozone framework. All classifications depicted similar patterns across Canada, but detailed class distributions differed. Class spatial characteristics varied with environmental conditions within classifications, but were comparable between classifications. There was moderate correspondence between classifications. The strongest association was between productivity classes and ecozones. The classification along both productivity and snow balanced these two sets of variables, yielding intermediate levels of association in all pairwise comparisons. Despite relatively low spatial agreement between classifications, they successfully captured patterns of the environmental conditions underlying alternate schemes (e.g.; snow classes explained variation in productivity and vice versa). The performance of ecosystem classifications and the relevance of their input variables depend on the environmental patterns and processes used for applications and evaluation. Productivity or snow regimes, as constructed here, may be desirable when summarizing patterns controlled by summer- or wintertime conditions, respectively, or of climate change responses. General purpose ecosystem classifications should include both sets of drivers. Classifications should be carefully, quantitatively, and comparatively evaluated relativ

Canada is the second largest country in the world by land area, at nearly 10 million km2 in size. Monitoring biodiversity and associated ecosystems for a nation the size of Canada requires approaches that enable broad scale national assessments. Over the past five years the Universities of British Columbia (UBC) and Victoria (UVic) with the Canadian Forest Service (CFS) of Natural Resources Canada (NRCan), have investigated the role remote sensing can play in the assessment of biodiversity across Canada. This research includes the national level application of indices which capture different aspects of species habitats, and the production of regionalizations or environmental domains which allows for the assessment of, for example, the representation of park networks which can be used to inform national biodiversity planning.

Aim: Several lines of evidence suggest that beta diversity, or dissimilarity in species composition, should increase with productivity: (1) the latitudinal species richness gradient is most closely related to productivity and associated latitudinal beta-diversity relationships have been described, and (2) the scale dependence of the productivity-diversity relationship implies that there should be a positive productivity-beta-diversity relationship. However, such a pattern has not yet been demonstrated at broad scales. We test if there is a gradient of increasing beta diversity with productivity. Location: Canada. Methods: Canada was clustered into regions of similar productivity regimes along three remotely sensed productivity axes (minimum and integrated annual productivity, seasonality of productivity) and elevation. The overall (β j), turnover (β sim) and nestedness (β nes) components of beta diversity within each productivity regime were estimated with pairwise dissimilarity metrics and related to cluster productivity with partial linear regression and with spatial autoregression. Tests were performed for all species, productivity breadth-based subsets (e.g. species occurring in many and a moderate number of productivity regimes), and pre- and post-1970 butterfly records. Beta diversity between adjacent clusters along the productivity gradients was also evaluated. Results: Within-cluster β j and β sim increased with productivity and decreased with seasonality. The converse was true for β nes. All species subsets responded similarly; however, productivity-beta-diversity relationships were weaker for the post-1970 temporal subset and strongest for species of moderate breadth. Between-cluster beta diversity (β j) and nestedness (β nes) declined with productivity. Main conclusions: As predicted, beta diversity of communities within productivity regimes was observed to increase with productivity. This pattern was driven largely by a gradient of species turnover. Ther

Understanding changes in landscape productivity patterns is critical for management of ecosystems, characterising biodiversity, and monitoring climate change. Vegetation productivity, a key functional component of terrestrial ecosystems, can be readily monitored using remote sensing and can also be combined with other spatial information, such as land cover and topography, to provide a more comprehensive understanding, at the landscape scale, of ecosystem dynamics. Ontario is the second largest and most populated province in Canada covering approximately 1 million km 2 with a population of 13 million. Befitting such a large and complex area, Ontario is characterised by high environmental and land cover diversity. In order to assess temporal trends in vegetation productivity across the province we utilised a series of 10-day composites of Medium Resolution Imaging Spectroradiometer (MERIS) fraction of Photosynthetically Active Radiation (fPAR) over a 6-year period (2003-2008). Indicators of annual productivity, seasonality and the minimum amount of vegetated cover were all calculated from the fPAR time-series to evaluate vegetation condition. To define the level of observed natural variability and areas which fall within, and beyond, the variability thresholds, we utilised the Theil-Sen's non-parametric statistical trend test. Results indicate that over the 6-year period decreasing trends in productivity dominated the Ontario landscape. Using spatial data layers of potential drivers of vegetation variability, we found that different land cover types exhibited significantly different trend amplitudes and proportions with elevation a significant correlate for most of the observed trends. Distance from nearest road or human settlement, indicative of anthropogenic drivers, was not significantly related to the observed trends. Recent and older burn areas were linked to decreasing and increasing trends in satellite derived productivity respectively. Our findings suggest th

Canada is dominated by remote wilderness areas that make important conservation contributions, but are currently only protected de facto by their inaccessibility. Mechanisms for the identification and formal protection of such areas can help ensure that they continue to function naturally and provide essential ecosystem services. However, a lack of spatially explicit, publicly available sources of data on anthropogenic disturbances and natural resource extraction challenges the development of detailed wilderness inventories. We suggest that landscape structure can be used to classify areas of natural landscapes, as trained by the landscape structure of protected areas, and demonstrate this approach by mapping de facto protected areas in Canada's boreal forest. Overall, between 50%, based on landscape structure, and 80%, based on anthropogenic infrastructure alone, of Canada's boreal zone exists in large, intact blocks. The true extent of boreal wilderness likely falls within this range, as existing infrastructure datasets may omit disturbance and the protected area network in far northern areas proved inadequate to train effective wilderness classifications. We anticipate that such efforts may be improved by refining the identification of training areas or by classifying along additional landscape metrics. Nevertheless, the areas identified are valuable candidates for protected area expansion, and can contribute to a reserve network that meets national and regional conservation targets and is representative of the range of vegetation productivities, which was used as a biodiversity surrogate. Our general approach need not be limited to the boreal forest, as it has the potential to successfully identify relatively undisturbed (or less disturbed) areas over a range of systems and across levels of human influence.

Earth observation data and approaches are increasingly being utilized to improve our insights into the ecological processes that influence biological diversity. Physically based indices such as the fraction of absorbed photosynthetically active radiation (fAPAR) intercepted by vegetation are particularly useful in describing variations in productivity and seasonality that can, in turn, be related to species abundances and distributions. The increasing availability of time series of fAPAR data from 2000 from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensors, as well from other satellites, such as ENVISAT, has resulted in a motivation to extend techniques, originally developed for MODIS biophysical data, to other sensors. In this letter we investigate and demonstrate the application of the dynamic habitat index (DHI) methodology to the MEdium Resolution Imaging Spectrometer (MERIS) Global Vegetation Index (MGVI) across the over 1 million km2 province of Ontario, Canada. Results indicate the three DHI components varied significantly in their magnitude, principally because of MODIS fAPAR estimates being larger than those observed by MERIS fAPAR. However the relationship was, in general, temporally stable across the years and the residuals tended to be spatially consistent. We conclude that following intercalibration the production of consistent indicators of habitat and biodiversity from different data sources is possible, thus supporting global terrestrial ecological research, policy support and management.

Salmon are critical to the ecology and livelihood of the Pacific Northwest, and are declining throughout much of their range. While much of their life cycle occurs in open ocean, freshwater conditions also contribute to population trends. Because stream habitats are connected to uplands by water flow, salmon can be influenced by the characteristics of terrestrial systems. We analyzed the relationships between the population trends of Pacific salmon (1953-2006) and land cover, fragmentation, and forest age derived from remotely-sensed, landscape level datasets. Analyses included 425 populations of all native salmon species in 156 watersheds on Vancouver Island, British Columbia, Canada. Vancouver Island salmon escapements exhibited general patterns of decline, which may be largely controlled by broad-scale marine conditions. The spatial variation in these population trends was related to landscape variables at watershed and riparian scales with regression trees. Results were found to be species specific, but characteristics indicating a legacy of historic and current forest management (such as fragmented forests and non-forested or early-successional forest cover) generally had negative effects, driven by a small subset of highly fragmented watersheds. Chum and coho had strong negative relationships with fragmentation, pink had a strong positive relationship with wetland abundance, and Chinook and sockeye were most closely related to geomorphology. There was no 'single best' scale of analysis. Salmon trends were generally more closely related to variables estimated over the entire watershed, however, the relative importances of watershed and riparian level predictors varied by both variable and species. Efforts to restore salmon habitat will be complicated by marine and freshwater processes, terrestrial conditions throughout watersheds, and the idiosyncratic requirements of each species.

Ecological regionalizations, such as ecoregions or environmental clusters, are often used as coarse filters for conservation. To be effective biodiversity surrogates, regionalizations should contain distinct species assemblages. This condition is not frequently evaluated and regionalizations are rarely assessed comparatively. We used a national dataset of Canadian butterfly collections to evaluate four regionalizations (ecoregions, land cover and productivity regime classifications, and a spatial grid) at two thematic resolutions using analysis of similarity (ANOSIM) and species indicator values. Overall, the spatially constrained schemes (ecoregions and grids) best captured patterns of butterfly community composition and species affinities, indicating that butterfly communities are strongly structured by space at the continent scale. In contrast, when comparing regions only within spatial or environmental neighbourhoods (i.e., comparing between regions that are adjacent along geographic or environmental gradients), all regionalizations performed similarly. Adjacency in environmental space is thus as important as physical adjacency at determining community dissimilarity. Productivity regimes and land cover will be useful biodiversity surrogates when considered in conjunction with space or within a spatially constrained area. This finding was confirmed with two ecoregional case studies (of the Algonquin-Lake Nipissing and Thompson-Okanagan Plateau ecoregions), which also revealed that the relative performance of regionalizations depends upon the context of the study area. We conclude that including species data can improve the efficiency of environmental surrogates for systematic conservation planning.

Productivity is an important driver of broad-scale diversity gradients and community composition. Surprisingly, it is rarely used as a biodiversity proxy in protected area network assessments. In this research, we evaluate if biases exist in the locations of Canada's protected areas with respect to productivity and assess the distribution of anthropogenic threats to protection along productivity and topographic axes. Productivity was expressed as the annual minimum and integrated fraction of absorbed photosynthetically active radiation (fPAR) and related seasonality (defined as variation in fPAR over a year), as acquired by remote sensing. Results indicate that, overall, protected areas are slightly biased to lower productivities (higher seasonalities), but are relatively unbiased along elevation. Sites at high elevations, high seasonalities, or low productivities are extremely well represented, but they account for a very small proportion of Canada's area overall and under protection. However, the productivity characteristics of protected areas are heterogeneous across reserve characteristics and space; biases are greater and more variable when considering individual International Union for Conservation of Nature (IUCN) categories, protected area size classes, or ecozones. The spatial and environmental correlates of anthropogenic pressures (proximity of roads and settlements in the greater park ecosystem) to protected areas in the forested ecozones were evaluated with partial linear regression models. Threats were strongly spatially structured. Reserve characteristics (productivity, elevation, size, IUCN category) shared about half of this spatial structure, explaining ∼25% of the variation in threat distributions. Small and productive protected areas tended to occur closer to human populations and roads, respectively, and are thus expected to face greater threats to biodiversity. High-elevation protected areas generally occurred farther from both roads and settlem

Regionally intensive human activities related to resource extraction (i.e., harvesting, oil and gas extraction) are increasing the occurrence of edges found in some forested landscapes. Edges between different land cover types represent important transition zones for abiotic and biotic processes. However, boundary detection methods often identify edges solely in areas of high contrast, such as transitions between forest and non-forest areas, and are insensitive to the relative contrast and orientation of different transitions. Edge contrast and orientation can determine the magnitude and even the occurrence of ecological edge effects and should be measured to provide information on landscape condition and habitat potential. Wombling was applied to the wetness component of a tasselled cap transformation (TCT) of a Landsat scene acquired over a portion of the eastern slopes of the Rocky Mountains in Alberta, Canada. By incorporating wombled edge contrast and orientation, and edge class transition type obtained from a land cover dataset, the nature of all transitions between land cover classes within the image was characterized and quantified. The consistency between edges identified by wombling and other common methods of edge delineation (such as spatial clustering) and methods of edge quantification (such as landscape pattern indices, or LPIs) was also assessed. Land cover transitions showed a broad range of edge contrast. Comparisons of edge contrast and the LPI edge density showed a positive correlation (r2 = 0.33); however, the strength of this relationship varied with the dominant land cover type (e.g., r2 = 0.016 for broadleaf open forest to r2 = 0.48 for dense coniferous forest). Stratifying edge contrast to higher values (i.e., >1 standard deviation) increased agreement with edge density, indicating that the LPI is preferentially relating high contrast edges. This study demonstrates how unique edge characteristics may be generated from a remotely sensed conti

The combination of forest inventory and satellite-derived landscape composition and structure provides otherwise unavailable information on regional forest conditions and enables investigation of the cumulative effects of forest management over time. Forest pattern results from a range of both natural and anthropogenic factors. This study characterizes the forest pattern of Vancouver Island, British Columbia (>32,000 km 2) at the watershed scale, using new national datasets of Canadian forest composition and fragmentation, and relating these to forest inventory-derived age structure. Vancouver Island is extensively forested, possessing highly valuable and productive forests, and is managed to meet a range of stakeholder interests. Forest fragmentation metrics were derived from a satellite-derived, 25-m spatial resolution land cover map (i.e., grain) to represent patterns within 1 km analysis units (i.e., extent) for the entire forested area of Canada. We summarized forest fragmentation island-wide and compared trends between forest-dominated (>85% forest) and less-forested (<85% forest) watersheds. We also explored these patterns with partial canonical correspondence analyses to determine the independent and shared relationships of landscape composition, forest fragmentation, and spatial variables with forest age structure. Less-forested watersheds are more fragmented than forest-dominated watersheds, as indicated by more (5.6 versus 2.7 per km 2) and smaller (36 versus 63 ha) forest patches, and a greater edge density (82 versus 55 m/ha). Of the 1283 watersheds examined, 91% were forest-dominated; island-wide trends are thus similar to those of the forest-dominated watersheds. Forest age is related to landscape composition and forest fragmentation, which collectively explain 27% and 53% of the age structure of all and less-forested watersheds, respectively. In both sets of watersheds, young forest stands (1-120 years) are associated with broadleaf forests, agreeing