Using satellite remote sensing to monitor and assess ecosystem integrity and climate change in Canada’s National Parks

R. Fraser, I. Olthof, D. Pouliot, W. Chen, S. Wang, and A. Clouston, Natural Resources Canada, Canada Centre for Remote Sensing
J. Poitevin, D. McLennan, P. Zorn, and J. Quirouette, Parks Canada, National Parks Directorate,
J. Kerr and E. Young, University of Ottawa, Department of Biology,
M. Sawada and Z. Relijic, University of Ottawa, Department of Geography

Mosaic of Landsat images covering the area of Nahanni Park.

Figure 1: Six Landsat images covering the GPE of Nahanni National Park Reserve (left) were radiometrically normalized to produce a seamless mosaic (right) for land cover classification. The yellow outline shows the GPE boundary (left) and park boundary (right).

Larger image

text version

Figure 1

This illustration shows a mosaic of Landsat images covering the area of Nahanni Park. This image mosaic was first radiometrically adjusted to allow for classifying land cover. It indicates the park boundaries and the limits of the ecosystem of the park.

Small tower in the Prairies and an apparatus for evaluating net primary productivity

Figure 2: Measurements of the inter-annual variations of annual NPP in Prince Albert National Park has shown ranges from 0.4 to 0.7 kg carbon/m2.

text version

Figure 2

This photograph shows a small tower in the Prairies and an apparatus for evaluating net primary productivity, resulting in values of carbon per square meter. The photo was taken in the Prince Albert National Park.

The disruption of ecosystems in the region of Prince Albert National Park from four Landsat images acquired between 1985 and 2001.

Figure 3: Significant vegetation disturbance has occurred in the Prince Albert National Park Greater Park Ecosystem, resulting in local losses of closed canopy ecosystems and an increase in open canopy and grasslands.

text version

Figure 3

This figure illustrates the disruption of ecosystems in the region of Prince Albert National Park from four Landsat images acquired between 1985 and 2001. The analysis of the legend concludes a decrease in closed-canopy ecosystems including evergreen and deciduous classes, and an increase of open-canopy ecosystems including deciduous shrubland, grassland, burned and barren land classes.

The distribution pattern of America Kestrel falcon in the Nahanni National Park

Figure 4: Species distribution modelling is performed with DesktopGarp (Genetic Algorithm for Rule-set Production) and maximumentropy techniques based on species observations, land cover and elevation; shown here for the American Kestrel (Falco sparverius) in the Nahanni National Park Reserve. Green areas indicate the likely distribution of the Kestrel.

text version

Figure 4

This illustration shows the distribution pattern of America Kestrel falcon in the Nahanni National Park. This model incorporates a number of environmental variables, and combines this information with the biophysical characteristics of the Kestrel falcon. The intersection of these criteria is used to identify the ranges of Kestrel falcon. The model shows that the Kestrel falcon is found mainly in the south of the park. A scale indicates that the region covers about 200 km by 300 km.

Introduction

Canada's national parks system includes 42 parks – covering 3 percent (296,253 km2) of the country's landmass and representing the full diversity of its natural regions. Considering the vast and often remote areas under protection, Parks Canada Agency (PCA) envisions Earth observation (EO) technology to be an integral component of a national Ecological Integrity Monitoring and Reporting Program. Natural Resources Canada, PCA, and the University of Ottawa have been developing standardized EO-based methods for monitoring landscape and ecological change within and surrounding Canada's forested national parks. This work was supported by the Canadian Space Agency under the Government Related Initiatives Program (GRIP).

Land cover and land use change is often the major factor impacting the ecological integrity (EI) of terrestrial ecosystems. While changes occurring within national parks are generally smaller than those in the greater park ecosystems (GPE), both have strong effect on the species and processes that maintain EI. Land cover provides a key input to quantifying habitat fragmentation and its influence on an ecosystem, and is an important variable for modelling plant productivity and biodiversity.

NRCan scientists have developed a protocol called AMUSE (Automated Multi-temporal Updating by Signature Extension) that automates land cover change detection as much as possible while still employing expert analyst guidance and quality control. The procedure was developed using Landsat TM and ETM+ data and consists of seven major steps:

  1. Radiometrically normalize baseline (Master) imagery to 1-km imagery.
  2. Produce Master baseline land cover classification
  3. Remove haze and topographic effects
  4. Radiometrically normalize other dates to Master
  5. Identify changed pixels using Change Vector Analysis (CVA)
  6. Update land cover using constrained signature extension
  7. Validate baseline land cover and changes

Six national parks (Kejimkujik, La Mauricie, St. Lawrence Islands, Prince Albert, Nahanni, and Pacific Rim), representing a range of forested Bioregions across the Parks Canada system, served as pilot sites to develop, test, and demonstrate the change mapping methods. The resulting time series of land cover data can serve as a primary input for deriving various landscape-level EI indicators related to habitat fragmentation, succession and retrogression, net primary productivity (NPP), and focal species distributions. The methods developed in this collaborative project will be applied by PCA to provide information for future State of the Parks Reports, beginning with Pacific Rim National Park in 2008.

A joint PCA-CCRS follow-on project is now underway with funding from the Canadian Space Agency to develop satellite-based EI monitoring methods for Arctic and sub-Arctic parks. This ParkSPACE initiative will have a particular focus on detecting and quantifying the impacts of climate change on vegetation, permafrost, and wetlands.

For more information see:
R.H. Fraser, I. Olthof, and D. Pouliot, 2009. Monitoring land cover change and ecological integrity in Canada's national parks. Remote Sensing of Environment 113:1397-1409.