British Columbia Mine Reclamation Symposium

Biodiversity management : establishing pre-mining baseline conditions on historical mining disturbances… Knopff, Kyle. H.; Franklin, C. W.; Luini, G. L.; Vasiga, D. J. 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


59367-Knopff_K_et_al_Biodiversity_management_pre-mining.pdf [ 264.95kB ]
JSON: 59367-1.0374926.json
JSON-LD: 59367-1.0374926-ld.json
RDF/XML (Pretty): 59367-1.0374926-rdf.xml
RDF/JSON: 59367-1.0374926-rdf.json
Turtle: 59367-1.0374926-turtle.txt
N-Triples: 59367-1.0374926-rdf-ntriples.txt
Original Record: 59367-1.0374926-source.json
Full Text

Full Text

BIODIVERSITY MANAGEMENT: ESTABLISHING PRE-MINING BASELINE CONDITIONS ON HISTORICAL MINING DISTURBANCES TO DERIVE BACK-CASTED WILDLIFE HABITAT SUITABILITY MODEL METRICS   Kyle. H. Knopff, PhD. 1 C. W. Franklin, MSc. 2 G.L. Luini, BSc. 2 D.J. Vasiga, BSc. 3   1Golder Associates, Calgary, Alberta 2Teck Coal Limited, Sparwood, British Columbia 3Teck Coal Limited, Vancouver, British Columbia   ABSTRACT Teck Coal Limited (Teck) is aiming towards a Net Positive Impact (NPI) on biodiversity in the areas where they operate. Teck’s NPI target applies to a condition prior to commencement of major mining activities at each operation, and biodiversity losses are measured from this pre-mining condition. Because mining at most of Teck’s five Elk Valley operations in southeastern British Columbia has been ongoing for decades, detailed pre-mining ecosystem mapping and inventory and wildlife habitat suitability prior to mining are unavailable. We recreated pre-mining ecosystem conditions using a modified Terrestrial Ecosystem Map (TEM) based on historical aerial imagery combined with available pre-mining data. The TEM permitted further investigation of pre-mining habitat conditions for ten wildlife species using species-specific Habitat Suitability Index (HSI) models. This novel approach led to a comprehensive understanding of ecosystem conditions and wildlife habitat in the pre-mining landscape, which enables Teck to quantify biodiversity losses. Mitigation, such as reclamation, will be applied to rehabilitate terrestrial ecosystems that provide habitat function for a wide range of wildlife species, similar to the pre-mining condition. This work provides the foundation for Teck’s pursuit of NPI in the Elk Valley and integrates with other novel approaches Teck implements to measure and mitigate adverse impacts to biodiversity.   KEYWORDS Net positive impact, terrestrial ecosystem mapping, habitat suitability index. INTRODUCTION Biodiversity encompasses the abundance and variety of life, including genes, species, and ecosystems. Biodiversity is threatened globally because of human activities. Goals of no net loss (NNL) or net positive impact (NPI) for biodiversity have increasingly been established by governments and financial institutions. Examples include the United States wetland policy (Environmental Law Institute 2002), France’s NNL policy for biodiversity (Quetier et al. 2014), Canada’s fisheries protection policy statement (Fisheries and Oceans Canada 2013), and the International Finance Corporation’s policy of demonstrating NNL in natural habitats and NPI in critical habitats prior to obtaining funding for development projects (IFC 2012).  Some resource extraction and manufacturing companies have also adopted corporate biodiversity policies, and by 2013 at least 32 companies around the world had made public commitments to NNL or NPI. Teck Resources Limited (Teck) is among these companies and maintains an aspirational vision of achieving NPI on biodiversity in the areas where Teck operates (Teck Resources 2015). Achieving NPI means that biodiversity will be better off as a consequence of Teck’s operations. The temporal target Teck has applied for NPI is at the end of the active-closure stage of mining, before a site goes into the long-term post-closure stage.  In the Elk Valley of south-eastern British Columbia, Canada, Teck’s commitment applies to five open pit steelmaking coal mines: Coal Mountain Operation (CMO), Elkview Operation (EVO), Line Creek Operation (LCO), Greenhills Operation (GHO), and Fording River Operation (FRO). Biodiversity management plans developed for each of these Elk Valley operations are blueprints that will lead to NPI by maintaining or re-establishing self-sustaining landscapes and ecosystems that will achieve an agreed set of viable, appropriate and diverse long-term land uses.  Although challenging to implement in practice, achieving NPI for biodiversity is conceptually straight-forward through application of biodiversity mitigation hierarchies (BBOP 2017, BC ENV 2014). Applying a mitigation hierarchy involves understanding the potential impact of a mine with no mitigation and then applying avoidance, minimization and rehabilitation to reduce the magnitude, duration, and frequency of potential adverse impacts. If a residual impact persists after the first three steps of the mitigation hierarchy have been applied, gains through offsets and other conservation actions must then be achieved to demonstrate NPI (Franklin et al. 2018). Teck’s biodiversity management plans focus on quantifying impacts to ecosystem and biodiversity elements (EBE) and identifying appropriate mitigation to achieve NPI (Franklin et al. 2018). Ecosystems are a major focus of the biodiversity management plans because assessing and managing biodiversity at the ecosystem level means that large numbers of biodiversity elements are addressed together. For example, impacts on wildlife guilds and plant communities that depend on different ecosystems can be captured simultaneously using an ecosystem level approach to biodiversity management. However, impacts to biodiversity at the species level sometimes can be distinct from ecosystems, and targeted actions for individual species are occasionally required. Consequently, some species level EBE also are included in Teck’s biodiversity management plans. In pursuit of NPI, a method of calculating losses and gains for both ecosystems and the species-level EBEs is required. Credible loss-gain calculations are an important part of demonstrating NPI (BBOP 2012), and, for Teck, these must be capable of quantifying habitat conditions in a comparable way for both the pre-mining landscape and the post-mining landscape at the end of the active closure stage. Teck’s NPI goal accounts for adverse impacts to biodiversity from each operation since its inception, whether or not Teck owned the mine during its initial development. Because Teck’s operations in the Elk Valley have been active for decades under multiple owners and because baseline information describing and mapping pre-mining conditions for active operations was either incomplete or unavailable when Teck’s biodiversity vision was adopted in 2013, finding a way to measure biodiversity loss from a pre-mining condition was an important initial challenge faced by Teck when working towards NPI in the Elk Valley. In this paper, we presents methods for calculating biodiversity losses associated with existing mining operations for ecosystems and individual wildlife species. Further refinements of ecosystem level loss-gain calculations using a Vegetation Quality Assessment are presented in Boyle et al., 2018 at this symposium and form an equally important discussion regarding the quantification of NPI. ECOSYSTEMS Some of Teck’s coal mining operations in the Elk Valley began as underground operations in the early 1900s, though most of the impacts to biodiversity occurred when open pit mining began in the late 1950s and early 1960s. We compiled information about pre-mining ecosystem conditions for each operation using internet searches, local industry and government libraries, university databases and data warehouses, and in-house Teck databases to identify imagery, plot data, and other information that could be used to support ecosystem mapping. The following information sources were identified and applied: • 1949 to 1952 panchromatic air photos (1:31,000 scale) University of Calgary Air Photo Archive; • 1969-1981 panchromatic and color air photos (1:10,000, 1:20,000 and 1:32,000 scale)  from GeoBC Base Map Online Store ( • 3-D PurVIEW models aligned using Terrain Recourses Information Management II (TRIM II) data ( developed by Colbert Tam Consulting (Vancouver). • Current Terrestrial Ecosystem Mapping (TEM) completed by Teck for natural ecosystems adjacent to the existing mining disturbance footprints where pre-mining ecosystems needed to be mapped. • Current Predictive Ecosystem Mapping (PEM) available for the Elk Valley in the form of pre-development (1950’s) and recent (2014) digital map products (Ketcheson 2014 a,b). • Slope and aspect derived from a 1950’s digital elevation model created from pre-mining aerial imagery (Golder 2014). • Regional Biogeoclimatic Subzone/Variant (BGC) Mapping version 9 and current classification from the regional ecologist (MacKillop 2015). • Pre-mining plot data (Province of British Columbia 2015a,b; Biogeoclimatic Ecosystem Classification Program of British Columbia, 2014). • Historic fire perimeter data (MFLNRO 2015a). • Historic forest cover data (BC Department of Lands and Forests 1973-78) • Waterbodies and Watercourse mapping (IHS 2004). • Freshwater Atlas of British Columbia’s hydrological features (MFLNRO 2015b). • Government and consultant data collected between 1930 and 2015 within and adjacent to areas disturbed by mining and adjacent locations. Plot classifications were converted to the newest classification systems applicable to the Biogeoclimatic Ecosystem Classification (BEC) subzones and variants within the study area (BC MOFR 2008, MacKenzie and Moran 2004, MacKenzie 2012, Mackillop 2015). Available aerial photographs were scanned to 14 microns (1,800 pixels per inch) by the BC Ministry of Forest, Lands & Natural Resource Operations (MFLNRO) - Digital Image Services. Photo dates ranged from 1969 – 1981 to cover all of Teck’s Elk Valley operations. We used the most recent and best resolution pre-mining photos for TEM interpretation at each operation.  We also used imagery from 1949 - 1952 from the University of Calgary Air Photo Archive in cases where disturbance associated with coal mines commenced prior to 1969. We mapped pre-mining ecosystem conditions using British Columbia’s Biogeoclimatic Ecosystem Classification (BEC) system, which considers climate, soil, and vegetation to group ecosystems at broad regional levels (Meidinger and Pojar 1991). We first undertook bioterrain mapping to delineate surficial materials, landforms, and hydrological processes. Bioterrain line work for all operations was based on historic digital elevation models, ortho-corrected air photos, and 3-D Purview imagery derived from ortho-corrected air photos and was completed at a 1:20,000 scale (Howes and Kenk 1997, RIC 1998). Soil moisture was not mapped on terrain features, although soil drainage was recorded to help inform the delineation of ecosystem units. Bioterrain mapping produces information about terrain materials and their texture, origin, depth, slope, relevant geological processes and drainage, which are important for ultimate ecosystem classification.  Next, we delineated ecosystems according to standard BEC rules and codes used to describe ecosystems for the region (MacKillop 2015). Non-forested ecosystems were included and follow the Biogeoclimatic Ecosystem Classification of Non-forested Ecosystems in British Columbia (MacKenzie 2012). Mapping methods and terminology follow RIC (1998) and the associated quality assurance and quality checking guidance (RISC 2000). Ecosystem attributes were assigned to polygons through photo interpretation. Three-dimensional stereo imagery allowed for delineation of structural stage (e.g., shrub from forest, herb meadows from krumholtz, old forest from young forest) and non-vegetated areas such as rock outcrops and talus slopes. Where plot data were available, these were used to direct or confirm visual air photo interpretation. We carefully evaluated all plot data and other data sources when applying them spatially because data collected before the use of GPS devices can be difficult to accurately georeference. To the extent possible, mapping delineated small but important non-forested ecosystem units such as grasslands and wetlands, which may occur as small areas relative to comparably large and continuous forest polygons. Some of this detailed mapping work for historic grasslands and wetlands had already been completed for the region (e.g., Golder 2014; EBA 2005), and was used to guide historical TEM mapping at each operation. Historical PEM (Ketcheson 2014a) included a soil moisture model, stream-buffered riparian areas, manually digitized wetlands, rock outcrops and grasslands that were also used to guide mapping decisions. Complex polygons were sometimes used (RIC 1998). No more than three ecosystem units were described within each complex polygon. Complex polygons are not spatially explicit concerning the location of individual ecosystems and, as such, simple polygons containing a single ecosystem unit were favored when sensitive ecosystems were present. Potential sensitive ecosystems vary among BEC subzones but typically include riparian areas, wetlands, avalanche paths, and grasslands. Site modifiers were added when conditions such as parent material and thin soils were known. Target minimum polygon size was set at ~0.5 ha and polygon mapping scale was 1:10,000 for all operations.  A primary difference between the pre-mining ecosystem mapping and standard TEM is that mapping accuracy was not supported by field data collected to meet TEM mapping standards. This gap was impossible to fill because pre-mining ecosystems had already been altered by mining and mapping therefore relied on legacy imagery and data. By applying historical field data, the final TEM products for each operation varied in survey intensity from 0 to 14% polygon visitation (i.e., TEM survey intensity R to 5) at a scale of 1:10,000 (RIC 1998). The final TEM included the following information for each mapped polygon: BEC zone, subzone and variant, site unit code, site modifier, structural stage, structural stage modifier, stand type, and full terrain description. The map outputs usefully described ecosystem conditions present prior to mining at each operation. For example, at FRO, 55 distinct site units were mapped within the study area, which includes a buffer around existing mining. Forested ecosystems dominated the study area, and the most common of these was the mesic Subalpine fir – Engelmann spruce – Azalea (101) ecosystem (Table 1). Less common ecosystems with high biodiversity value include grasslands, avalanche slopes, and wetlands (Table 1). Structural stage information indicates that the FRO study area was comprised of large amounts of old forest (Table 2). The pre-mining FRO study area showed the least amount of natural disturbance of all the operations in the Elk Valley, which explains the large areas of old forest, which were especially concentrated in the Kilmarnock Creek valley and the Shandley Creek area of the Greenhills Range.  FRO also had the most extensive complex of wetlands of the Elk Valley Teck operations, largely found along the upper Fording River and its tributaries. These wetlands were associated with wet forest and floodplain ecosystems units. Grassland units were commonly found in the Eagle Mountain area amongst open forest and avalanche ecological communities. The imagery allows reasonable confidence in identification of these sensitive ecosystems. The 1950’s imagery was useful in differentiating pre-mining anthropogenic disturbance from that associated with mining. There were signs of small patches of forest clearing and settlement on the 1950’s imagery but much of the FRO study area had little anthropogenic disturbance prior to the start of open pit coal mining circa 1971. The 1970’s forest cover data was indispensable for confirming forest type and structural stages. Of all the Elk Valley operation pre-mining maps, we were most confident in structural stage interpretation for FRO. We were also highly to moderately confident in FRO ecosystem interpretation.  Table 1: Pre-Mining Ecosystems in the Fording River Operations Study Area Site Code Site Description Area (ha) ESSFdk1 ESSFdkp ESSFdkw MSdw Total 102 Lodgepole pine – Douglas-fir  – Juniper – Douglas maple 1.6    1.6 102 Subalpine fir – Subalpine larch – Grouseberry   23.4  23.4 103 Subalpine fir – Lodgepole pine – Grouseberry 214    214 103 Subalpine fir – Engelmann spruce – Whitebark pine – Grouseberry   179.2  179.2 103 Lodgepole pine – Douglas-fir – Western larch – Pinegrass    24.9 24.9 101 Hybrid white spruce – Arnica – Pinegrass    215 215 101 Subalpine fir – Engelmann spruce – Grouseberry – Leafy liverwort   156.4  156.4 101 Subalpine fir – Engelmann spruce – Azalea 2,384.4    2384.4 104 Lodgepole pine – Western larch – Pinegrass – Twinflower    102.6 102.6 104 Subalpine fir –  Lodgepole pine  – Azalea – Grouseberry 1,012.3    1012.3 110 Subalpine fir – Engelmann spruce  – Grouseberry – Valerian   1.4  1.4 110 Subalpine fir – Engelmann spruce – Azalea – Foamflower 490.8    490.8 110 Hybrid white spruce – Subalpine fir – Azalea – Bunchberry    89.7 89.7 111 Hybrid white spruce – Horsetail    173.9 173.9 111 Engelmann spruce –  Horsetail – Bluejoint 140.6    140.6 Ag Alpine grassland 6.2 52.3 59.5  118 Am Alpine meadow  2.1 16.5  18.6 Cb Cutbank 5.3    5.3 Es Exposed soil 2 3.3 15.1  20.4 Fl Low bench flood class 109.5   32.7 142.2 Gb Brushland 109.9  2.4  112.3 Gg Grassland 92.7  11.1  103.8 Ow Open water 1.5    1.5 Gr Gravel bar 26.9   21 47.9 Ri River    11 11 Ro Rock outcrop 1.8 14.9 19.6  36.3 Rt Talus 1.5 10.2 24.9  36.6 Rz Road surface 5.1    5.1 Sc Shrub carr 8.3    8.3 Sk Krummholz association  37.7   37.7 Vh Avalanche herb meadow 85.2 5.7 159.2  250.1 Vs Avalanche shrub thicket 90.4 6 19.4  115.8 Vt Avalanche treed 107 15.3 26.9  149.2 Wf Fen wetland 62.5   22.7 85.2 Xh Herbaceous vegetation disclimax 16    16 Total 4975.5 147.5 715 693.5 6531.4  Table 2: Proportion of Pre-Mining Structural Stages in Fording River Operations Study Area Code Structural Stage Area (ha) Percent of Study Area 0 Non-Vegetated 12.5 0.2 1a Sparse 146.9 2.2 2a Forb 267.5 4.1 2b Graminoid 310.9 4.8 3 Shrub 218.1 3.3 3a Low Shrub 266.2 4.1 3b Tall Shrub 386.9 5.9 4 Pole/Sapling 1,224.8 18.8 5 Young Forest 1,502.7 23.0 5i Irregular Young Forest 89.8 1.4 6 Mature Forest 583.1 8.9 6i Irregular Mature Forest 1.2 <0.1 7 Old Forest 1,512.4 23.2 7i Irregular Old Forest 8.4 0.1 Total 6,531.4 100.0  WILDLIFE SPECIES Loss-gain calculations for species level EBE are commonly achieved using habitat models. By comparing modeled conditions in the pre-mining and closure landscapes, the residual habitat losses resulting from the coal mining operation, after incorporating the benefits of reclamation, can be calculated. The size and type of residual impact (if any) can guide identification of biodiversity offsets and other actions that can be undertaken in pursuit of NPI. Models representing habitat selection were developed or ten wildlife species: American badger (Taxidea taxus jeffersonii), American marten (Martes americana), bighorn sheep (Ovis canadensis), Canada lynx (Lynx canadensis), elk (Cervus elaphus), grizzly bear (Ursus arctos), northern goshawk (Accipiter gentilis), olive-sided flycatcher (Contopus cooperi), western toad (Anaxyrus boreas), and wolverine (Gulo gulo). Models have been previously produced for some of these species in the Elk Valley, including resource selection functions (RSF) and resource selection probability functions (RSPF) derived from GPS telemetry data (e.g., Apps and McLellan 2008, Teck Coal Limited 2015). However, the spatial data layers used to generate RSF and RSPF model outputs, such as indices derived from satellite image color bands, were not available for pre-mining conditions and could not easily be estimated for a closure landscape. Consequently, we developed Habitat Suitability Index (HSI) models using spatial layers available for the pre-mining condition to permit “apples to apples” comparisons for Teck’s loss-gain calculations. HSI models were based on RSF and RSPF models, where these were available.      We used the following spatial layers to develop our models: • Terrestrial Ecosystem Mapping – the TEM layer was the foundation of the wildlife HSI modeling. • Digital Elevation Model – A pre-mining DEM for the 1950s landscape in Teck’s terrestrial regional study area was created by Teck as part of a regional pre-development study (Golder 2014). • Disturbance map – mapped as part of TEM, with additional information about pre-mining disturbance types available through 1950 disturbance mapping (Ketcheson 2014a). • A stream layer – The stream layer was used in the grizzly bear and western toad models. The stream layer was derived from the 1950s DEM using the following list of ArcGIS tools: “Fill”, “Flow Direction”, “Flow Accumulation”, “Reclassify”, and “Raster to Polyline”. The first three tools were run sequentially. The output of the Flow Accumulation Tool was reclassified to 0 or 1 using a cutoff value to define areas with the highest flow accumulation (i.e., 1). Pixels with a score of 1 represented streams. The raster to polyline tool will create polylines from adjacent pixels with a value of 1. • An aspect layer – the aspect layer was used in several models and was derived from the DEM using the Aspect tool in ArcGIS. • Edge layers – edge layers are required for the elk, northern goshawk, and olive-sided flycatcher models. Different layers were produced for each species depending on their habitat requirements. Edge for elk consisted of areas where grasslands or shrub habitats occurred adjacent to forests, whereas edge for olive-sided flycatcher consisted of a 50m buffer around the intersection between forested and non-forested habitats. We developed models using a series of suitability indices (SI) derived from the spatial layers listed above. Each SI is expressed as a layer in a GIS and these layers are combined using a mathematical equation specific to each model to achieve a final HSI output. The TEM is the primary layer used to predict habitat conditions for each species. Site series are fine ecosystem divisions that occur in different BEC zones. Although BEC site series are the standard unit by which ecosystems are delineated in TEM, this level of detail results in hundreds of unique BEC zone, subzone, variant and site series combinations in the Elk Valley. This level of detail is often too specific for measuring and managing impacts to ecosystem function and wildlife habitat. In many cases, communities of plants and animals, and ecological functions are similar across groups of site series within different BEC zones. For example, wildlife species are rarely limited to a specific site series, but rather use a range of site series across the landscape as habitat. Consequently, we developed wildlife HSI models using broad ecosystem types that represent several individual site series (Teck Coal Limited, 2015). Where ecosystem types were used, all BEC site series included in that group received the same model score. For some wildlife species, variable scores for specific site series were warranted and those were applied in the HSI developed for that species. We also used the 8 primary structural stage classes (i.e., 0-7) applied to ecosystems in British Columbia as a modifier (Table 2, BC MOFR and BC MOE 2010). The mathematical equation developed for each model provides an output scaled between 0 and 1. Consequently, running the model will assign each polygon on the landscape a score between 0 and 1. Complex polygons are TEM polygons containing more than one site series, and can contain as many as three site series. To account for complex polygons, each distinct site series within each polygon was identified and we assigned suitability values according to the proportion of each site series in the polygon and then summed to achieve a single suitability score for the polygon.  For brevity, we present the model structure only for bighorn sheep as an example. We developed similar models for each of the 10 wildlife species. The bighorn sheep HSI represents winter habitat. Although some bighorn sheep populations migrate to mid- or low-elevation habitat during winter (Demarchi et al. 2000), the population in the Elk Valley typically spends winters on high-elevation grasslands that are kept free of snow by strong winds and solar radiation (Poole 2010). Therefore, highest quality sheep habitats near Teck’s operations in the Elk Valley are characterized by mid- to high-elevation grasslands located on steep slopes and warm aspects that provide escape terrain, solar radiation and improved foraging opportunities. In addition, reclamation practices on mines favour sheep forages that are less common off mine lands and are attractive to sheep. The bighorn sheep HSI model was developed using observations regarding winter habitat use derived from bighorn sheep GPS telemetry data collected in the Elk Valley (Poole 2010). The following responses were included in the model: • a positive relationship with herbaceous, shrub and exposed ecosystems or habitats; • a strong, positive relationship with grassland ecosystems and habitats at middle to high elevations (i.e., 1,700 m to 2,500 m); • a negative relationship with elevations above 2,500 m; • a positive relationship with warm aspects; • a positive relationship with terrain complexity; and • a positive relationship with mine lands during winter, presumably because reclamation practices favour sheep forage and mining activities expose mineral licks and produce steep, rocky terrains. The bighorn sheep winter habitat model has three indices: ecosystem type, elevation and aspect. Elevation is correlated with terrain complexity, so separate indices for elevation and terrain complexity were not used. The resulting model equation is: 𝐻𝑆𝐼𝑏𝑖𝑔ℎ𝑜𝑟𝑛 𝑠ℎ𝑒𝑒𝑝 = 𝑆𝐼(1) ∗ 𝑆𝐼(2) ∗ 𝑆𝐼(3) SI(1) represents the ecosystem type. Based on the available information about winter habitat preferences identified for bighorn sheep, suitable ecosystem types are assigned SI(1) values as indicated in Table 3. All other ecosystem types are assigned a suitability score of zero. Scores associated with reclaimed mine (Ry) assume previous reclamation practices, including substantial use of non-native grasses and forbs, as these areas currently account for most of the reclaimed areas. Reclaimed areas following recent updated prescriptions that aim for a variety of ecosystem conditions, including forests which would not benefit bighorn sheep as much, will need to be revisited in the future to confirm assumptions of reclamation area function.  Table 3: Bighorn Sheep Winter Habitat SI(1): Ecosystem Type Ecosystem Type Suitability Score Alpine dwarf shrub 1.0 Alpine grassland 1.0 Alpine late snowbed 1.0 Alpine meadow 1.0 Avalanche path 0.8 Brushland/Grassland 1.0 Herb meadow 0.8 Krummholz 0.6 Reclaimed mine 0.8 Rock/Talus 0.4 SI = suitability index. SI(2) represents elevation. Based on the available information about winter habitat preferences identified for bighorn sheep, suitable elevation ranges are assigned SI(2) values as indicated in Table 4. Elevation ranges should be derived from the DEM using the Reclassify Tool. Table 4: Bighorn Sheep Winter Habitat SI(2): Elevation Elevation Range [masl] Suitability Score below 1,200 0.1 1,200 to 1,299 0.2 1,300 to 1,399 0.3 1,400 to 1,499 0.4 1,500 to 1,599 0.5 1,600 to 1,699 0.6 1,700 to 1,799 0.7 1,800 to 1,999 0.8 2,000 to 2,499 1.0 2,500 to 2,599 0.8 2,600 to 2,699 0.6 2,700 to 2,799 0.4 2,800 to 2,899 0.2 2,900 and above 0.1 masl = metres above sea level; SI = suitability index. SI(3) represents aspect. Based on the available information about winter habitat preferences identified for bighorn sheep, suitable aspects were assigned SI(3) values as indicated in Table 5.   Table 5: Bighorn Sheep Winter Habitat SI(3): Aspect Degrees [°] Approximate Cardinal Directions Aspect Condition Suitability Score 135 to 285 SE to W warm 1.0 285 to 360, 0 to 135 WNW to ESE cool 0.7 SI = suitability index. For each species, we used the HSI to calculate the total number of habitat units in a polygon by multiplying the area of the polygon by its HSI score. Thus, a 10 ha polygon with an HSI score of 1 would result in 10 habitat units, whereas a 10 ha polygon with an HSI score of 0.5 would result in 5 habitat units. We then calculated the total number of habitat units for each species within the area impacted by an operation (i.e., disturbance area and zone of influence) by summing the habitat unit scores for the area. By incorporating the pre-mining TEM and TEM produced for approved and proposed mine expansions, we were able to calculate the total number of habitat units impacted by current and planned mining for each species at each Elk Valley Operation (Figure 1).    Figure 1: Habitat units impacted for each wildlife species at each of Teck’s five Elk Valley Operations.   DISCUSSION Using standard TEM methods and historical data can result in a reasonable approximation of pre-mining ecosystem conditions. Available large scale (1:10,000 – 1:32,000) panchromatic imagery provided useful pre-mining information for all of Teck’s Elk Valley Operations that started after 1969. Mapping in areas with little accurate historical plot data required a high degree of comparison with other similar areas that had plot data or more detailed current imagery. The methods we describe in this paper rely on a strong historical knowledge base, good knowledge of local ecosystems, and careful photo interpretation to make confident ecosystem calls about ecosystems present prior to mining at Teck’s Elk Valley Operations. Delineating historical ecosystem units relies on the availability of historical data and quality imagery. The main limitation of the methods we applied is the use of scanned panchromatic photographs. Panchromatic photographs have no associated reflectance data and therefore poorly distinguish between some vegetation types. While it is relatively easy to separate a fen wetland (Wf) association from a shrub carr (Sc) site using photographs alone, the texture and appearance of some associations like alpine meadow (Am) and grassland (Gg) types are very 020004000600080001000012000Habitat Units Wildlife Species LCOGHOFROEVOCMOsimilar and vegetation plots or infrared/ multi-spectral photographs are required to separate them with high confidence. In cases where plot data were not available, we resolved this for some sites by comparing more recent colour photographs and adjusting for temporal differences. Using air photos in a softcopy digital environment improved the ease of photo interpretation and greatly improved pre-mining mapping efficiency.  Another issue we encountered during mapping was the presence of shadows in steep terrain. Shadows sometimes hid vegetation and landform features and we found it important to have multiple sources of imagery to avoid site classification errors in polygons with shadows. Future applications of this approach should carefully consider whether sufficient number and quality of air photos are available to achieve a desired level of accuracy.   Subjectivity in manual ecosystem mapping was reduced by following a working legend identifying ecosystems and landform relationships to ensure consistency in mapping (RIC 1998). Manual mapping relies on the mappers familiarity with the landscape and the applicability of the classification systems used. Another source of error can be inaccurate imagery georectification due to anomalies in the DEM. Spatially accurate imagery relies on TRIM 1:20,000 scale mapping specifications provide by the Province of British Columbia; however, alignment errors can occur as historical imagery often lacks aerial triangulation points and DEM’s are not perfect. Limited field data and other challenges results in a confidence level for pre-mining ecosystem maps that is necessarily lower than regular modern TEM standards. Nevertheless, the pre-mining mapping we produced represents the best available depiction of ecosystems present before mining at each of Teck’s Elk Valley Operations. These map products form the basis for estimating biodiversity impacts at Teck’s Elk Valley operations and can be used to guide reclamation efforts undertaken as part of operation-specific biodiversity management plans developed in pursuit of NPI. The pre-mining TEM mapping also permitted us to develop wildlife HSI models that we used to estimate species level habitat losses associated with Teck’s Elk Valley Operations. The HSI models have not been validated and we may adjust them based on additional data moving forward, but these models already provide Teck with an initial understanding of total habitat losses for each species that can be used to focus rehabilitation and offsetting efforts in pursuit of NPI. These models also provide a means for Teck to calculate gains for wildlife as a result of reclamation and offsetting. Calculating residual impacts between pre-development and closure can be accomplished by subtracting the total number of habitat units in the pre-development case from the total number of habitat units in the closure case. A negative result indicates an adverse residual impact and the size of the impact represents the number of habitat units that Teck should seek to offset or otherwise mitigate to achieve NPI for the species. A positive result would indicate that NPI for the species has been achieved at closure. Loss-gain calculations derived from HSI models can be useful for identifying conservation actions that may be necessary to achieve NPI. However, the results of such calculations should be interpreted carefully. Actions that achieve apparent benefits in the models are not necessarily the only means by which NPI might be achieved and models have the potential to misdirect effort if not interpreted judiciously. For example, the model for Canada lynx, which are forest-dwelling animals, predicts habitat suitability based on amount of a particular forest type in a particular structural stage, adjusted for elevation, aspect, and slope. However, the importance of these forest conditions is as an indicator of the abundance of prey for Canada lynx (i.e., snowshoe hare). Consequently, any action that improves habitat for snowshoe hare should benefit lynx, and actions that increase forest but do not increase prey for lynx will not have the same benefit. Therefore, focusing strictly on model outputs may result in misdirected conservation actions if the mechanisms driving the model are not also considered. Loss-gain calculations derived from HSI models are also useful for identifying variation in species specific outcomes for wildlife on closure landscapes. Because of variation in habitat selection patterns among species, some reclamation practices benefit certain species over others and benefits change over time due to natural ecosystem succession. For example, American badgers and elk can benefit substantially from early seral conditions on reclaimed mines, whereas species that depend on mature forests such American martens and northern goshawks will see little benefit until decades after reclamation occurs. By the time forests become more prevalent on the reclaimed landscape, benefits to species like American badgers and elk decline. Carefully considering target reclamation prescriptions in this light can be helpful for achieving NPI overall, with the recognition that positive outcomes may not be achieved for all species at the same time. ACKNOWLEDGEMENTS  The authors wish to thank Scott Black, Steve Henstra, and Kristine Sare for input, analytical support, and discussion regarding this paper. We also wish to thank all of the people who work at Teck’s Elk Valley operations in southeastern British Columbia. Without their dedication to a sustainable approach in mining and their openness to implementing our biodiversity vision of NPI none of this could have happened.  REFERENCES  Apps, C. and B. McLellan.  2008.  Grizzly bear habitat selection and suitability across multiple scales in the Flathead and Lower Elk drainages, British Columbia. Aspen Wildlife Research Inc. and Matrix Solutions Inc., Calgary Alberta. Biogeoclimatic Ecosystem Classification Program of British Columbia.  2014.  BECMaster ecosystem plot database [VPro13/MSAccess 2007 format]. W.H. MacKenzie [editor]. B.C. Min. For., Lands, and Nat. Res. Ops., Smithers, British Columbia. Available at: Accessed November 2014. BC ENV (BC Ministry of Environment and Climate Change).  2014.  Policy for Mitigating Impacts on Environmental Values (Environmental Mitigation Policy). Available from: BC Department of Lands and Forests 1973-78. 1:15,000 Forest Cover Maps. Fernie P.S.Y.U. 1973, 1974, 1978. BC Forest Service Inventory Division  BC MOFR (BC Ministry of Forests and Range).  2008.  Six Krummholz Associations. Research Ecologist. Ministry of Forests, Lands and Natural Resource Operations, Nelson, BC. BC MOFR and BC MOE (British Columbia Ministry of Forests and Ranges and British Columbia Ministry of Environment).  2010 . Field Manual for Describing Terrestrial Ecosystems, 2nd Edition. Land Management Handbook. Province of British Columbia, Crown Publications Inc. 2010. PDF File. BBOP (Business and Biodiversity Offsets Programme).  2012.  Resource Paper: No Net Loss and Loss-Gain Calculations in Biodiversity Offsets. Forest Trends. Available from: BBOP (Business and Biodiversity Offsets Programme).  2017.  Biodiversity Offsets. Retrieved July 5, 2017, from Business and Biodiversity Offsets Programme: Boyle, B.L., R.E. Gullison, D.J. Vasiga, G.L. Luini, C.W. Franklin.  2018.  Vegetation Quality Assessment: Measuring Quality of Vegetation Communities to Support the Accounting Metrics of the Biodiversity Vision of Net Positive Impact for a Large-Scale Mining Operation. Symposium proceedings BC Technical and Research Committee on Reclamation. Williams Lake 2018. Demarchi, R.A., C.L. Hartwig and D.A. Demarchi.  2000.  Status of the Rocky Mountain bighorn sheep in British Columbia. British Columbia Ministry of Environment, Lands and Parks, Wildlife Bulletin No. B-99, Victoria, BC. Environmental Law Institute.  2002.  Banks and fees: the status of off-site wetland mitigation in the United States. Environmental Law Institute: Washington, DC, USA. EBA Engineering Consultants Ltd.  2005.  Characterization of High Elevation Grasslands in the Upper Elk Valley. Project No.: 0335-5330005. Submitted to Tembec. Fisheries and Oceans Canada.  2013.  Fisheries Protection Policy Statement. Available at Franklin, C.W., Hilts, S. and R. Gullison.  2018.  Recent technical and conceptual advances supporting Teck Coal Limited’s pursuit of Net Positive Impact at operations in the Elk River valley of southeaster British Columbia and near Hinton, Alberta. Submitted to TRCR.  Golder (Golder Associates Ltd). 2014. Baldy Ridge Extension Project. Annex L - Vegetation Baseline Report. Submitted to: Teck Coal Limited Elkview Operations. Report Number: 1213490013/R44 Howes, D.E. and E. Kenk.  1997.  Terrain classification system for British Columbia. British Columbia Ministry of Environment, Victoria, BC. IFC (International Finance Corporation).  2012a.  Performance Standard 6. Biodiversity Conservation and Sustainable Management of Natural Resources. January 1, 2012. 7pp IHS.  2004.  Watercourse and Waterbody data obtained from IHS Energy (Canada) Inc. Obtained July 2011. Ketcheson, M.V.  2014a.  Teck Regional Study Area (RSA) Predictive Ecosystem Models (PEM) Report: for 1950’s Models (bapid = pem_6433_6434_pro). JMJ Holdings Inc. Ketcheson, M.V.  2014b.  Teck Regional Study Area (RSA) Predictive Ecosystem Models (PEM) 1950 [shapefile]. JMJ Holdings Inc. MacKenzie, W.H.  2012.  Biogeoclimatic Ecosystem Classification of Non-forested Ecosystems in British Columbia. Prov. B.C., Victoria, B.C. Tech. Rep. 068. Available at: Accessed November 2014. MacKenzie, W.H. and J.R. Moran.  2004.  Wetlands of British Columbia: A Guide to Identification. Land Management Handbook No. 52. Research Branch, British Columbia Ministry of Forests, Victoria. 297pp. Mackillop, D.  2015.  Kootenay-Boundary Cranbrook Biogeoclimatic Units Revisions Draft Report. Unpublished draft report by Deb Mackillop, Regional Ecologist, Kootenay-Boundary Region. Meidinger, D. and J. Pojar.  1991.  Ecosystems of British Columbia. Research Branch, BC Ministry of Forests. xii + 330 pp. Available at: Accessed December 9, 2014. MFLNRO (Ministry of Forests, Lands and Natural Resource Operations).  2015a.  Fire Perimeters – Historical. Available at: Accessed: November 2015.  MFLNRO (Ministry of Forests, Lands and Natural Resource Operations).  2015b.  Freshwater Atlas 20K-50K Stream and Wetland layers Available at: Accessed October 2015. Poole, K.  2010.  Habitat use, seasonal movements, and population dynamics of bighorn sheep in the Elk Valley – interim report to March 2010. Prepared for BC Ministry of Environment and Teck Coal Limited. 23 pp. Province of British Columbia.  2015a.  British Columbia Soil Mapping Spatial Data (a compilation of digital soil mapping datasets from 1989). Data available from the British Columbia Ministry of Environment, Ecosystem Information Section. Available at: Accessed: November 2015.  Province of British Columbia.  2015b.  Elk Valley BEC Master Plots [various formats]. Data received from Kootenay/Boundary Regional Ecologist Deb MacKillop.  Quétier, F., Regnery, B. & Levrel, B.  2014.  No net loss of biodiversity or paper offsets? A critical review of the French no net loss policy. Environmental Science and Policy 38: 120-131. RIC (Resources Inventory Committee).  1998.  Standard for Terrestrial Ecosystem Mapping in British Columbia. Prepared by Ecosystems Working Group, Terrestrial Ecosystems Task Force, Resources Inventory Committee, Victoria, B.C. RISC (Resources Inventory Standards Committee).  2000.  Standard for Terrestrial Ecosystem Mapping (TEM) Digital Data Capture in British Columbia Version 3.0. Prepared by Ecosystems Working Group, Terrestrial Ecosystems Task Force, Resources Inventory Committee, Victoria, B.C. Teck Coal Limited.  2015.  Baldy Ridge Extension Project Environmental Assessment Certificate Application. Submitted to the British Columbia Environmental Assessment Office. Teck Resources Limited.  2015.  Sustainability Strategy. Retrieved June 19, 2017, from Teck: 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items