Transcript w i

Habitat Evaluation Procedures
• 1969-1976 – an enlightened Congress
passes conservation legislation
• Affecting management of fish &
wildlife resources
• NEPA (National Environmental Policy Act)
• ESA
• Forest & Rangelands Renewable Resources
Planning Act
• Federal Land Policy & Management Act
Habitat Evaluation Procedures
• Stimulates federal & state agencies
to change management, thus:
1) simple, rapid, reliable methods to
determine & predict the species and
habitats present on lands;
2) expand database for T/E, rare species;
3) Predict effects of various land use
actions
Habitat Evaluation Procedures
• USFWS
• Habitat analysis models
• Goal = Assess impacts at a community
level (i.e., species
representative of all habitats
being studied)
• e.g., use guild of species?
Habitat Evaluation Procedures
• USFWS
• Habitat analysis models
• What is a model?
• Important points to consider relative
to models?
• What variables should be measured
and/or included in the model?
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
a) simple correlation models
e.g., vegetation type-species
matrix
Species habitat matrix
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
b) statistical models
i.e., prediction of distribution
and/or abundance
What types?
Carnivore Habitat Research at CMU
Spatial Ecology
• Overlay hexagon grid onto landcover map
• Compare bobcat habitat attributes to population of hexagon
core areas
Carnivore Habitat Research at CMU
Spatial Ecology
• Landscape metrics include:
• Composition
(e.g., proportion cover type)
• Configuration
(e.g., patch isolation, shape,
adjacency)
• Connectivity
(e.g., landscape permeability)
Carnivore Habitat Research at CMU
Spatial Ecology
Pij   ki  kj  / pVk
p
2
k 1
• Calculate and use Penrose distance to measure similarity
between more bobcat & non-bobcat hexagons
• Where:
• population i represent core areas of radio-collared bobcats
• population j represents NLP hexagons
• p is the number of landscape variables evaluated
• μ is the landscape variable value
• k is each observation
• V is variance for each landscape variable
after Manly (2005).
Penrose Model for Michigan Bobcats
Variable
Mean Vector bobcat
hexagons
NLP hexagons
% ag-openland
15.8
32.4
% low forest
51.4
10.4
% up forest
17.6
43.7
% non-for wetland
8.6
2.3
% stream
3.4
0.9
% transportation
3.0
5.2
Low for core
27.6
3.6
Mean A per disjunct
core
0.7
2.6
Dist ag
50.0
44.9
Dist up for
55.0
43.6
CV nonfor wet A
208.3
120.1
Carnivore Habitat Research at CMU
Spatial Ecology
• Each hexagon in NLP
then receives a
Penrose Distance
(PD) value
• Remap NLP using these
hexagons
• Determine mean PD for
bobcat-occupied
hexagons
Preuss 2005
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
b) statistical models
* modern statistical modeling &
model selection techniques
e.g., logistic regression & Resource
Selection Probability Functions
(RSF) & RSPF for determining
amount & dist. of favorable
habitat
Habitat Evaluation Procedures
Logistic regression:
Y = β0 + β1X1 + β2X2 + β3X3 = logit(p)
Pr(Y = 1 | the explanatory variables x) = π
π = e –logit(p) / [1+ e –logit(p)]
1
Y
0
X
Resource
Selection
Functions (RSF)
• Ciarniello et al. 2003
• Resource Selection
Function Model for
grizzly bear habitat
• landcover types,
landscape
greenness, dist to
roads
Resource
Selection
Probability
Functions (RSPF)
• Mladenoff et al. 1995
• Resource Selection
Probability
Function Model for
gray wolf habitat
• road density
Predicted American Woodcock
Abundance Map
Quantifying Habitat Use – Resource
Selection Ratios
Need:
1) Determine use (e.g., prop. Use)
2) Determine availability (e.g., prop avail.)
Selection ratio – for a given resource category i
wi = prop use / prop avail.
If wi = 1 , < 1, > 1
Quantifying Habitat Use – Resource
Selection Ratios
Selection ratio
wi = prop use / prop avail.
wi = (Ui /U+) / (Ai /A+)
Ui = # observations in habitat type i
U+ = total # observations (n)
Ai = # random points in habitat type i
A+ = total # of random points
Quantifying Habitat Use – Resource
Selection Ratios
Look at Neu et al. (1974) moose data
= 117 observations of moose tracks within 4
different vegetation [habitat] types
Quantifying Habitat Use – Resource
Selection Ratios
Veg. Type
Use
Avail
wi
Interior burn
25
0.340
(25/117)/0.340
= 0.628
Edge burn
22
0.101
Edge unburned
30
0.104
Interior
unburned
40
0.455
Totals
117
1.000
Quantifying Habitat Use – Resource
Selection Ratios
Veg. Type
Use
Avail
wi
Interior burn
25
0.340
Edge burn
22
0.101
(25/117)/0.340
= 0.628
(22/117)/0.101
= 1.862
Edge unburned
30
0.104
Interior
unburned
40
0.455
Totals
117
1.000
Quantifying Habitat Use – Resource
Selection Ratios
Veg. Type
Use
Avail
wi
Interior burn
25
0.340
Edge burn
22
0.101
(25/117)/0.340
= 0.628
(22/117)/0.101
= 1.862
Edge unburned
30
0.104
Interior
unburned
40
0.455
Totals
117
1.000
2.465
Quantifying Habitat Use – Resource
Selection Ratios
Veg. Type
Use
Avail
wi
Interior burn
25
0.340
Edge burn
22
0.101
(25/117)/0.340
= 0.628
(22/117)/0.101
= 1.862
Edge unburned
30
0.104
2.465
Interior
unburned
40
0.455
0.751
Totals
117
1.000
Quantifying Habitat Use – Resource
Selection Ratios
Selection ratio
* Generally standardize wi to 0-1 scale for
comparison among habitat types
std wi = wi / Σ (wi)
Quantifying Habitat Use – Resource
Selection Ratios
Veg. Type
wi
Std wi
Interior burn
0.628
Edge burn
1.862
0.628/5.706 =
0.110
1.862/5.706 =
0.326
Edge unburned
2.465
0.432
Interior
unburned
0.751
0.132
Totals
5.706
1.000
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
c) Habitat Suitability Index (HSI)
models
Habitat
Suitability
Index (HSI)
Habitat Suitability Index (HSI)
• Model (assess) habitat (physical &
biological attributes) for a wildlife
species, e.g., USFWS
- Habitat Units (HU) = (HSI) x
(Area of available habitat)
- Ratio value of interest divided by
std comparison
HSI = study area habitat conditions
optimum habitat conditions
Habitat Suitability Index (HSI)
• Model (assess) habitat (physical &
biological attributes) for a wildlife
species, e.g., USFWS
- HSI = index value (units?) of how
suitable habitat is
- 0 = unsuitable; 1= most suitable
- value assumed proportional to K
Habitat Suitability Index (HSI)
• include top environmental variables
related to a species’ presence,
distribution & abundance
Habitat Suitability Index (HSI)
• List of Habitat Suitability Index (HSI)
models
• http://el.erdc.usace.army.mil/emrrp/emris/emrishel
p3/list_of_habitat_suitability_index_hsi_models_p
ac.htm
e.g., HSI for red-tailed hawk
Habitat Suitability Index (HSI)
Red-tailed Hawk
Habitat Suitability Index (HSI)
Red-tailed Hawk
Habitat Suitability Index (HSI)
Red-tailed Hawk
Habitat Suitability Index (HSI)
Red-tailed Hawk
Habitat Suitability Index (HSI)
Red-tailed Hawk
Habitat Suitability Index (HSI)
Red-tailed Hawk
For Grassland:
Food Value HSI = (V12 x V2 x V3)1/4
For Deciduous Forest:
Food Value HSI = (V4 x 0.6)
Reproductive value HSI = V5
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
c) Habitat Capability (HC)
models
- USFS
- describe habitat conditions
associated with or necessary to
maintain different population
levels of a species ( compositions)
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
c) Habitat Capability (HC)
models
- uses weighted values based on
habitat capacity rates at each
successional stage of veg. for
reproduction, resting, and feeding
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
c) Habitat Capability (HC)
models
-
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
c) Pattern Recognition (PATREC)
models
- use conditional probabilities to
assess whether habitat is suitable
for a species
- must know what is suitable &
unsuitable habitat
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Single-species models
c) Pattern Recognition (PATREC)
models
- use series of habitat attributes
- must know relation of attributes to
population density
PATREC Models
Expected Habitat Suitability (EHS) =
[P(H) x P (I/H)] / [P(H) x P (I/H)] + [P (L) x P (I/L)]
P(H) = prop. high density habitat
P (I/H)] = prop. area has high population potential
P (L) = prop. low density habitat
P (I/L) = prop. area has low population potential
* Low & high population potential identified from surveys
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Multiple-species models
a) Integrated Habitat Inventory and
Classification System (IHICS)
- BLM
- system of data gathering,
classification, storage
- no capacity for predicting use or
how change affects species
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Multiple-species models
b) Life-form Model
- USFS
-
Habitat Evaluation Procedures
Three Categories of Techniques:
1) Multiple-species models
b) Community Guild Models
- can be used to estimate responses
of species to alteration of habitat
- (like Life-form model) clusters
species with similar habitat
requirements for feeding &
reproduction
Three Scales of Diversity
A = B = alpha () diversity – within habitat
C = beta () diversity – among habitat
D = gamma () diversity – geographic scale
Alpha & Gamma Species
Diversity Indices
• Shannon-Wiener Index – most used
- sensitive to change in status of rare
species
s
H '   ( pi )(ln pi )
i 1
H’ = diversity of species (range 0-1+)
s = # of species
pi = proportion of total sample
belonging to ith species
Alpha & Gamma Species
Diversity Indices
• Shannon-Wiener Index
s
H '   ( pi )(ln pi )
i 1
Alpha & Gamma Species
Diversity Indices
• Simpson Index – sensitive to changes
in most abundant species
s
D  1   ( pi )
2
i 1
D = diversity of species (range 0-1)
s = # of species
pi = proportion of total sample
belonging to ith species
Alpha & Gamma Species
Diversity Indices
• Simpson Index
s
D  1   ( pi )
i 1
2
Alpha & Gamma Species
Diversity Indices
• Species Evenness
H'
J
H 'max
H’max = maximum value of H’ = ln(s)
Beta Species Diversity Indices
• Sorensen’s Coefficient of Community
Similarity – weights species in
2a
common
S 
S
2a  b  c
Ss = coefficient of similarity
(range 0-1)
a = # species common to both samples
b = # species in sample 1
c = # species in sample 2
Beta Species Diversity Indices
• Sorensen’s Coefficient of Community
Similarity
Dissimilarity = DS = b + c / 2a + b + c
Or 1.0 - Ss
Species
1
2
3
4
5
6
7
8
9
10
11
12
Sample 1
1
1
1
0
1
0
0
1
1
0
1
0
Sample 2
1
0
1
0
1
0
0
0
1
0
1
0
Sorensen’s Coefficient
• Sample 1
– Total occurrences = b = 7
- # joint occurrences = a = 5
• Sample 2
– Total occurrences = c = 5
- # joint occurrences = a = 5
• 2*a/(2a+b+c)
• Ss = 2 * 5 / 10 + 7 + 5 = 0.45 (45%)
• Ds = 1 – 0.45 = 0.55 (55%)