National Park Service

1
National Park Service Upper Columbia Basin Network
Modeling Species-Habitat Relationships Using
Inventory Data
2
National Park Service Upper Columbia Basin Network
Modeling Species-Habitat Relationships Using
Inventory Data

• Introduction why bother?
• Occupancy estimation a model-based approach
• Example 1 Pika at CRMO
• Example 2 Pinyon-juniper woodland rodent
  community at CIRO

3
National Park Service Upper Columbia Basin Network
The Distribution and Abundance of Organisms A
Fundamental Concern
Where do species occur? Why? How many?
4
National Park Service Upper Columbia Basin Network
Resource Availability and Selection

• Species Presence/Absence Occupancy

Occupancy The proportion of area, patches, or
sample units occupied or used by a species
Used vs. unused patches Used vs. available habitat
Mackenzie et al. 2006
5
National Park Service Upper Columbia Basin Network
Resource Availability and Selection

• Issues of scale at what scale is selection
  occurring and important?

• Multiple patterns can arise from a single
  process or mechanism - multiple processes can
  lead to the same pattern

6
National Park Service Upper Columbia Basin Network
7
National Park Service Upper Columbia Basin Network
Resource Availability and Selection

• Issues of scale at what scale is selection
  occurring and important?

• Multiple patterns can arise from a single
  process or mechanism - multiple processes can
  lead to the same pattern

• Selecting the best model the best may not be
  the goodinformation-theoretics, model
  validation, etc

• Adaptive management we can evaluate multiple
  hypotheses, use monitoring data to update
  predictions, adjust management, etc

8
National Park Service Upper Columbia Basin Network
Occupancy Estimation
Pr(h1 101) ?p1(1-p2)p3, Where ?
probability site is occupied (or used) and p
probability of detecting the species during that
survey.
9
National Park Service Upper Columbia Basin Network
American pika, Ochotona princeps
10
National Park Service Upper Columbia Basin Network
11
National Park Service Upper Columbia Basin Network
Craters of the Moon National Monument and Preserve
Huddles Hole
Methods
Monument

• 2007-2008 2-week September survey June 2008
  survey
• 12 m radius site search area 20 minute search
  time
• Established detection with direct and indirect
  sign
• Focused sampling frame w/in 1 km roads/trails
  120 sample units revisited 2x
• Site and sample-specific covariates (lava
  complexity, lava type, veg cover)

Laidlaw Park
Wapi Flow
Craters of the Moon National Monument and Preserve
12
National Park Service Upper Columbia Basin Network
Results
Overall occupancy site use (n72) was low and
patchily distributed
Occupancy (?) 0.2 Detectability (p) 0.92
13
National Park Service Upper Columbia Basin Network
14
National Park Service Upper Columbia Basin Network
Huddles Hole
Monument
!!!
?
Laidlaw Park
Wapi Flow
Craters of the Moon National Monument and Preserve
15
National Park Service Upper Columbia Basin Network
Habitat selection in a pinyon-juniper woodland
rodent community - City of Rocks NR and Castle
Rocks SP
16
Juniperus osteosperma
Pinus monophylla
Tamias dorsalis
City of Rocks
17
Peromyscus truei
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
Pinyon Ips (Ips confusus) infestation
22
National Park Service Upper Columbia Basin Network
Methods

• Sampled 44 sites in P-J woodland habitat. Used
  Principal Components Analysis (PCA) to distill
  the range of fundamental environmental
  characteristics down to 1 gradient
• Rocky P-J with bare ground and low understory veg
  cover to P-J far from rock with dense understory
  vegetation Axis 1

23
Heterogeneity!
24
- AXIS 1
Increasing bare ground and rock
Increasing understory vegetation
25
(No Transcript)
26
In Conclusion.

• P. truei DOES occur at City of Rocks in Cassia
  Co., Idaho

• Rodent occupancy/use patterns reflect dominant
  environmental gradients

• Both P. truei and T. dorsalis exhibit a
  propensity to use well-developed P-J sites with
  open understory and a significant rock component

• Together with species like juniper titmouse,
  represent a unique P-J fauna and reflect P-J
  structural conditions. P-J mgmt implications!

• How might these relationships change over time
  at the periphery?

• Occupancy modeling, with non-traditional
  design components (e.g. few traps per site)
  presents a viable long-term monitoring approach.