USGS - Science for a changing world

Documentation regarding the development of the swan breeding habitat knowledge base.
 11 April 1999
Richard S. Sojda

This documentation describes the underlying ecological logic behind the development of the swan breeding habitat knowledge base.  The knowledge, itself, resulted from knowledge engineering sessions held with Mr. Todd Grant and Dr. Leigh Fredrickson.


Depth and Size Tree
The following table shows the criteria used to determine the "confidence" values that are assigned in the depth and size tree. E.g., row one of the table means:
    if
        ((>= 30% has a depth between 6" - 40" and >= 5 acres
        (or > few hundred acres and >= 10 contiguous acres are 6" - 40"))
    then
        we are 100 percent confident that the wetland is the correct depth
        and we are 95 percent confident that the wetland is large enough
    else
        wetland is neither the correct depth nor large enough
 
Confidence for Depth
6"-40"
Confidence for Size
>5 Acres
100
X
95
X
0
 
0
 
Prelaying Food Tree
The following table shows the criteria used to determine the "confidence" values that are assigned in the prelaying food tree. E.g., row one of the table means:
    if
        unfrozen three weeks preceding laying
        and >= 30% has vegetation with large, below-ground resources
    then
        we are 100 percent confident that prelaying food resources are sufficient
 
Confidence
Ice Free
30% Sago
50% Leafy 
Nearby Wetland Ice Free
Nearby Wetland 30% Sago
Nearby Wetland 50% Leafy 
Arrive Fat
100
X
X
         
100
X
 
X
       
100
     
X
X
   
100
     
X
 
X
 
80
           
X
0
X
           
0
     
X
     
Such a table can appear deceptively simple. It required a tree with 64 rules/end-nodes to implement.
 

Nest Sites Tree
The following table shows the criteria used to determine the "confidence" values that are assigned in the nest sites tree. E.g., row one of the table means:
    if
        robust emergents are present
        and interspersion can be classified as near the hemi-marsh stage
    then
        we are 95 percent confident that potential nest sites exist
 
Confidence
Emergents
Hemi-marsh
Muskrats or Beaver
Floating Vegetation
40 Meter Vegetation Band
95
X
X
     
95
X
 
X
   
90
X
   
X
 
90
       
X
0
X
       
0
         
Brood Habitat Tree
The following table shows the criteria used to determine the "confidence" values set in the brood habitat tree. E.g., row one of the table means:
    if
        some water remains available until typical fledging time
        and dry sites for brooding exist the first few weeks after hatch
        and emergents are growing at or just above the water's surface during brood rearing
    then
        we are 100 percent confident that brood habitat is sufficient
 
Confidence
Brood Water
Dry Sites - Brooding
Food: Emergents
Food: Submer-gents
Food: Treadling
100
X
X
X
   
100
X
X
 
X
 
90
X
X
   
X
85
X
 
X
   
85
X
   
X
 
75
X
     
X
0
X
X
     
Ice Free Period and Nest Flooding Trees
These two parameters go through some representational gymnastics to stay consistent, but eventually are interpreted as relatively straightforward probabilities as depicted in the "Final Assessment" table, below.
 

Calculating the Final Assessment of Breeding Habitat
Variables, created for each critical ecological parameter, are given values during a consultation. Then, the variables are combined with the following algorithm:
    if
        >= one parameter fails
    then
        final assessment = 0
    else
        final assessment = (Depth + Size + PreLay + Nest + Brood) * Ice * Flood
Internally, each ecological parameter is actually associated with two variables, one for the number of "points" to add to the total and one for the number to be used as a multiplier to the total number of points. Eighteen rules are actually needed to calculate the value for "Final Assessment", which is described to the user as :

This is a somewhat arbitrary assessment of the degree to which this wetland represents ideal habitat for breeding trumpeter swans, based on a synthesis of all information. On a scale of 0-100, the higher the number, the more confident one can be that satisfactory habitat exists. A zero indicates that at least one habitat component is severely compromised. The value of each additive and multiplier variable assigned for each value of an ecological parameter are provided in the following table.
 
Ecological Parameter
Parameter?s Value
"Points" to Be Added
Value of Multiplier
Depth
Fails
0
0
 
Passes*
25
1
Size
Fails
0
0
 
Passes*
25
1
Ice Free 
10-6*
0
1
[no. years out of 10]
5-3
0
0.4
 
0-2
0
0
Prelaying 
100*
15
1
[confidence value 
80
10
1
From tree]
Fails
0
0
Nest Sites
95*
20
1
[confidence value 
90
15
1
From tree]
Fails
0
0
Brood Habitat
90/100*
15
1
[confidence value 
75/85
10
1
From tree]
Fails
0
0
Nest Flooding
10-6*
0
0
[no. years out of 10]
5-3
0
0.6
 
0-2
0
1
* For the values of the ecological parameters marked with an asterisk, the system would calculate the value for "final assessment" to be 100, i.e., we are 100 percent confident that such a wetland represents satisfactory breeding habitat.


Return to the Swan Management Decision Support System homepage.




USGS - Science for a changing world
USDI - Geological Survey
Biological Resources Division
Northern Rocky Mountain Science Center
Maintainer: Rick Sojda (sojda@swan.msu.montana.edu)