Disturbance density in woodland caribou (Rangifer tarandus caribou) territories throughout the South Peace Region
The abundance and distribution of woodland caribou (Rangifer tarandus caribou) has decreased across North America since the colonization and advancement of humans (Bergerud 1974, Seip 1992, Schaefer 2003, Vors et al. 2007). With continued population declines and increasing habitat fragmentation across Canada, the conservation and management of caribou is now a high priority (Cumming 1992, Wittmer 2004, Vors and Boyce 2009). All three ecotypes (northern, mountain, and boreal) of woodland caribou in B.C. are listed under the federal Species at Risk Act (SARA) as “threatened” throughout portions of the province. To accommodate the needs of declining caribou populations and the predators that historically co-existed with caribou, is important for planners, biologists, and resource managers to further understand how landscape change influences predator-prey dynamics.
Since the early 1990s, regions surrounding the Peace River in Northeastern British Columbia and Northwest Alberta have experienced rapid land-use change from resource extraction activities, such as oil and gas, in addition to large-scale commercial forestry, agriculture, and mineral development (Schneider et al. 2003). The cumulative effects resulting from industrial developments have produced forested landscapes that are progressively younger and increasingly fragmented (Schneider et al. 2003). Nitschke (2008) found that cumulative effects of landscape development accounted for an 89% increase in edge habitats, a 67% increase in areas containing early seral habitats, and a 47% increase in the amount of open landscapes. As edge and open landscape habitat structures have increased in area, various forms of predation have also increased (Nitschke 2008). Within 60 years, habitat available to woodland caribou is estimated to decline from 43% to 6% (Schneider et al. 2003).
Recent studies in B.C. and Alberta have demonstrated that industrial activities and features including roads, trails, geophysical exploration lines (seismic), pipelines, electrical right-of-ways, cutblocks, and oil and gas wells can negatively affect woodland caribou populations (Bradshaw et al. 1997, James and Stuart-Smith 2000, Smith et al. 2000, Dyer et al. 2001, Wittmer et al. 2007, Sorensen et al. 2008). These features can alter the movements, distributions, and population dynamics of both caribou and wolves (Jalkotzy et al. 1997). Timber harvesting is one of the primary agents of habitat change within the study area. Large-scale harvesting increases the area of early successional forests, and because this younger habitat favors moose and other non-caribou ungulates, historical predator-prey dynamics have changed (Fuller and Keith 1981, Rempel et al. 1997, Johnson et al 2004a, Wittmer et al. 2007, Nitschke 2008). Due to the above mentioned threats, this project focuses on the density of linear features and other human disturbances throughout the range of four herds of northern woodland caribou. This project was chosen because results have the ability to play a part in my MSc thesis, as well as influence management decisions in hopes to futher protect habitat for these declining caribou populations.
Images of woodland caribou: two young bulls sparing in winter, a female in summer, and a calf with a radio collared bull.
Study Area, Data Source and MethodsLocated on the eastern slopes of the Rocky Mountains in central B.C., the study area is approximately 12,000 km2 in size. Tumbler Ridge is located near the center of the study area, which then extends northwest towards the town of Mackenzie, northeast towards Dawson Creek and south along the Alberta border into the Redwillow boreal forest. This area supports a vast number of wildlife species that require mature, contiguous, late-seral stage forests (Nitschke 2008). Four biogeoclimatic zones occur within the study area: Boreal White and Black Spruce (BWBS), Sub-Boreal Spruce (SBS), Engelmann Spruce–Subalpine Fir (ESSF), and Alpine Tundra (AT; Meidinger and Pojar 1991).
My study will be premised on data collected from four northern woodland caribou herds (Fig. 1). The Moberly herd is found to the north of Hwy 97, a major highway in the area, whereas the Burnt Pine, Quintette, and Bearhole/Redwillow herds reside to the south. The Bearhole/Redwillow herd is the only study herd that remains in the low-elevation boreal forests during the winter. Two open-pit coal mines near Tumbler Ridge (the Wolverine and Trend mines) are found in the core winter range of the Quintette herd. The Quintette, Moberly, Burnt Pine, and Bearhole/Redwillow caribou herds are exposed to various levels of disturbance resulting from logging, mining, and extensive oil and gas exploration.
A total of 51 caribou within four herds (Bearhole/Redwillow = 6, Burnt Pine = 2, Moberly = 10, Quintette = 33) were captured between April 2002 and March 2009 by net-gunning from a helicopter. Caribou were fitted with either Televilt or ATS (Advanced Telemetry System) GPS collars equipped with VHF transmitters, as well as remote-release devices programmed to release at specified dates and times. ATS collars were programmed to take location fixes every 20 hours up until 2005; collars programmed after April 2005, acquired fixes between two and six times daily. In addition, two female caribou were captured in the study area and collared with Lotek ARGOS GPS collars (F900 and F901 of the Redwillow herd) in 2007. These two female caribou are part of a study affiliated with the University of Alberta (Edmonton, Alberta, Canada) and a data sharing agreement was established.
I created a 25m digital elevation model for use throughout my analysis (DEM; British Columbia Integrated Land Management Bureau, Land and Resource Data Warehouse, 2007). The DEM was available in BC Albers projection from the GIS lab at UNBC. I acquired two additional map sheets and converted them from ASCII into UTM and then into BC Albers in order to merge them with the initial DEM for the study area. I acquired the spatial data necessary for generating the linear features and cumulative disturbance layers from the British Columbia Integrated Land Management Bureau (British Columbia Integrated Land Management Bureau, Land and Resource Data Warehouse, 2007), West Fraser Timber Company (Chetwynd, BC), and the Oil and Gas Commission of British Columbia’s public access website (http://www.ogc.gov.bc.ca/GIS.asp, 2009). All analysis for this project was perfomed using associated extensions for ArcGIS (Hawth's Tools, Spatial Analysis, etc.) and MS Excel.
Photos of Disturbance Features across the landscape in the South Peace Region of B.C (all photos were taken by Libby while conducting field work). From left to right: cutblock in alpine, extinct coal mine just outside Tumbler Ridge, B.C., a wide powerline in the boreal forest, and seismic lines (natural gas exploration) in the boreal forest.
Data AnalysisI began analysis by creating minimum convex polygons (MCP, 100%) for each of the four caribou herds; each MCP was based off GPS collar locations collected from individual caribou (n = 47,750). Disturbance features were clipped to the study area boundry and further clipped and merged to each herd's MCP. Length and area fields were added to attribute tables for calculation of features in each herd territory.
Fig. 2 - Fig. 5. Four maps representing the different anthropogenic disturbances located across each caribou herd territory. Fig.4 is a larger-scale map to show the detail of different disturbance types in the Quintette and Redwillow/Bearhole territories.
Results indicate that cutblocks represent the most prominent disturbance feature type (n = 1453, total area = 53.97 km2) across the south Peace region (total area = 12,464 km2). The Quintette herd also experiences the largest amount of disturbance (0.94%). Large-scale industrial disturbances are sweeping into the region from the east, and this seems to be accurately represneted in my results. Due to the low density values for each herd, I am cautious about these results.
Limitations: There are many limitations to every study, and this one proved to be no different. The numbers represented in my final results do not seem to reflect the amount of disturbances that is present on the landscape. With such large territories, there could be many bias altering these results. One bias is that my calculation for road density is less accurate because there is no known width value associated with the road layer. Without the road width unknown, it is impossible to come up with a true density value. All other features contained an area value field, or it was created from the appropriate and availalble data.
My origional plan was to import all data as raster into IDRISI and do calculations in IDRISI. However, may complications led to a time constrant forcing me to conduct the analysis in ArcGIS. So, with this limitation in mind, I plan to extend this analysis by working through it this alternative raster software program. It will be fun to compare these results with the raster analysis results. Another limitation is the use of 100% MCPs. This polygon encompasses all locations, but does not weight areas where the caribou spend the majority of their time. This has the ability to give more surface area to each study area, but may not accurately depict true use of the territory. It may also be adventageous to create a buffer around the road layer to signify the effect roads have on wildlife populations. This could, however, lead to an additional bias if not conducting the buffer for each disturbance type. This project was a great start to understanding density calculations and limitations in a large study area.