A habitat model
Determine potential sighting locations of a bear species in a park

Factors
        1. Land cover types
        2. Species richness
        3. Species interspersion

Raw Data
        Land cover
        Source: satellite images, digital aerial photos, land use/land cover data, Gap data      

        Existing bear sighting data: 91 locations

Data layer preparation
        1. Species richness:
        The total number of unique land cover types in a 3x3 window

        2. Species interspersion:
        The number of cells with land cover type that is different from the center cell

        3. The window size can be 5x5, 7x7, 9x9, .....
        The optimal window size is the one that shows the greatest difference in richness or interspersion
        The "difference" can be absolute value range or variance for richness or interspersion

        4. Generate another set of 91 locations that are randomly located in the study area

Statistics analysis:
        Determine whether each of the three layers is relevant

        1. Do the %s of land cover types differ between the bear sighting sites and the entire area?
            Do the bear prefer certain types of land cover?
        Null H₁: % of each cover type used by birds = % of each type in the area
        C² test
        Accept or reject the null, % of each type used by bear ? % of each type in the area (r)

        2. Do the %s of land cover types differ between the random sites and the entire area?
        Null H₂: % of each cover type of random sites = % of each type in the area
        C² test
        Accept or reject the null, % types of random sites ? % of the area (a)

        3. Does species richness differ between the sighting sites and the random sites? 
        H₃: richness at bear sighting sites = richness of the random sites
        t-test, accept or reject the null (a)
        mean richness of sighting sites ? mean richness of random sites

        4. Does species interspersion differ between the sighting sites and the random sites?         
        H₄: interspersion at bear sighting sites = interspersion of the random sites
        t-test, accept or reject the null (r)
        interspersion of sighting sites ? interspersion of random sites

GIS overlay
        Keep the variables tested significantly different between the sighting sites and random sites
        cover type:   
        richness:       
        interspersion:     
              
        For each significant variable, prepare a data layer 

Results: potential sighting map
 
 
 
 
 
 

...Back to Ling Bian top page.