Methods


Modules are capitalized

ARCRASTER: Images in ASCII format were imported into IDRISI.

From the Landsat images, band 1 was found to best show snow cover so only band 1 images were used in the analysis.

RECLASS: Each band 1 image was reclassed so that all values up to 250 were shown as zero and values 250-255 were shown as one. This was done because only the highest values represented pure snow cover. Wolverines are known to select only those denning sites in areas of 100% spring snow cover.

ASSIGN: Images were then assigned discreet values.

RECLASS: The CDEM image was reclassed so that all elevations up to 1100m above sea level (asl) were shown as one and all elevations above 1100m asl were shown as zero. This was done because wolverines are known to select denning sites only above this elevation.

ASSIGN: The CDEM reclassed image was assigned discreet values.

RECLASS/ASSIGN: Road Networks were reclassed and assigned discreet values as with the above images (1=roads, 0=no roads).

BUFFER: A buffer of 7.5km was created around road networks. This because wolverines are known to avoid denning sites any closer than 7.5km to public roads.

OVERLAY: The values in the CDEM assigned image were subtracted from the assigned Landsat snow images. This was done to eliminate areas of spring snow cover that were not high enough above sea level.

OVERLAY: The 2006 road network buffers were subtracted from the adjusted snow cover images for 1986, 1990, 1990 and 2006. The 2010 road network buffer was subtracted form the adjusted snow cover image for 2010.

ASSIGN: In the resultant images, the remaining snow cover was given a value of one and everything else was given a value of zero.