Transcript Brief to the Scientific Advisory Board, October 20, 2005

Performers
Dr. Aaron Moody
University of North Carolina, Chapel Hill
Specialist in Landscape Ecology, Remote Sensing, GIS
Dr. Nick Haddad
North Carolina State University
Specialist in Landscape Ecology and Conservation Biology
Dr. Bill Morris
Duke University
Specialist in Population Ecology, Dispersal Modeling
Dr. Jeffrey Walters
Virginia Tech
Specialist in Avian and Population Ecology
Dr. Larry Crowder & Mr. Jeffery Priddy
Duke University
Specialist in Demographic Modeling
2
Problem Statement
• Two strategies drive land acquisition:
Conservation of high quality required habitats
Conservation of dispersal habitats
• Which land parcels to conserve to meet needs of
different species in context of military and nonmilitary land uses?
3
Technical Objectives
• Spatial decision support for land acquisition
• Behavioral and landscape approaches
• Habitat connectivity
• Multiple at-risk species
4
Technical Background
Landscape Connectivity
5
Landscape Connectivity
Our goal is to identify land units that provide the
greatest habitat connectivity for multiple at-risk
species.
6
Connectivity Around Bases
7
Multiple At-Risk Species
Ft. Bragg
8
What is Novel?
Current Approaches to Land Acquisition
• Decisions based on expert opinion
• Connectivity virtually ignored
• Dispersal considered only for focal species
New Approach
• Integrate behavioral and landscape approaches
9
High Likelihood of Success
•Spatial framework
•Environmental data
•Habitat-specific movement data
•Dispersal models
decision-support environment for
quantifying connectivity
known occupancy
predicted dispersal path
observed dispersal path
low
Dispersal Resistance
high
10
Technical Approach
Integrated Spatial Database
Movement
Data Collection
Spatial Data
Acquisition
Field Data
Collection
Spatial Data Development
Dispersal
Modeling
Spatial
Modeling
Evaluations &
Model Updates
Transition
Implement Decision Support System
for Habitat Management
11
Spatial Data: Acquisition and
Development
Acquisition & Integration
● Maps & Infrastructure
● Remote Sensing Data
● Ft. Bragg Datasets
Spatial Data Layers
Elevation
Hydroperiod
Development
● Field Data
● Land-Use
● Canopy Structure
● Hydroperiod
● Habitat Maps
Infrastructure
Soils
Land Use
Zoning
Site Data
Canopy Structure
Final Validations
12
Field Data Collection & Environmental
Data Development : Land Use
Spectral distribution functions
for each type
spectral band 2
Random samples stratified by
land-use class
spectral band 1
Pasture, Row Crop
Forest Plantation
Upland Forest
Longleaf Pine Woodland
Wetlands
13
Field Data Collection & Environmental
Data Development : Hydroperiod
Crest gauges in known
and potential amphibian habitats
vented cap
Statistical inundation model:
2” pipe
measuring
stick
perforated cup with
crushed cork
perforated
cap
• crest gauge data
• digital terrain data
• stream gauge data
Thermal remote sensing data
to extend evaluation of model
Inundation surfaces
~30 Gauges
gauges
surface depressions
from terrain data
stream channel
ground-water level
from stream gauge
data
14
Environmental Data Development:
Habitat Mapping
Spatial Data Layers
Elevation
Land use
Hydroperiod
Infrastructure
Soils
Habitat Models
+
Land Use
Zoning
Site Data
Canopy
Structure
Habitat Maps
Canopy Structure
Validate by visiting
predicted habitat
Hydroperiod
Occurrence
and dispersal
data
15
Collection of Movement Data:
St. Francis’ Satyr
• Visually track movement
behavior of naturally
occurring butterflies
• Experimentally release
butterflies in dominant
habitats and at their
boundaries
Habitat A Habitat B
x
• Monitor dispersal events and
occupancy in natural habitats
x
x
r
r = release point
16
Movement Data
Habitat A
Boundary
High Resistance
L1
A1
L2
Habitat B
High Conductivity
L1
L2
L3
A1
A2
17
Dispersal Modeling
Computer Simulation
Mean squared distance
Analytical Models
Habitat A
Habitat B
Turn angle
Turn angle
Move length
Habitat B
Move length
Habitat A
Number of moves
Kareiva and Shigesada 1983
Boundary Behavior
18
Spatial Modeling:
Landscape Resistance
Translate environmental
data to resistance surfaces
Distance Moved
for habitat k and species s:
Habitat B
rks = 1/slope
Habitat A
Number of moves
rks =
low
high Water Bldg
19
Spatial Modeling:
Connectivity
Create landscape network of
habitats and connecting paths
Determine least-cost paths
and least-cost networks
j
Costij = resistanceweighted distance
along path
i
rki =
lo
hi
Water
Bldg
Sensitivity Analyses: Costs/benefits of altering pathways
20
Testing Models
• Test against observed
dispersals and habitat
occupancy
Observed
dispersers
known occupancy
predicted dispersal path
observed dispersal path
1st
2nd
3rd
4th
Connectivity between
patch pairs (quartiles)
low
Dispersal Resistance
high
• Assess trade-off between
information value and data
requirements of methods
21
Implementation
Use our data and tools to:
Map value of land parcels for
habitat connectivity
Test connectivity impacts of
management alternatives
Determine optimal connectivity
scenarios for multiple species
To what extent does optimizing
connectivity for one species
benefit others?
22
Overall Project Plan
Year 1
Year 2
Year 1
Year 2
Year 3
Year 4
Year 3
Year 4
Year 5
Year 5
Spatial Data Acquisition & Dbase Dvlpmt
Collection of Movement Data
Environmental Field Data Collection
Spatial Data Development
Dispersal Modeling
Spatial Modeling
Assessments, Comparisons, Updates
Produce Final Deliverables
Start
Duration
Progress contingent upon quality assessment of field
data and derived products.
23
Deliverables
• A decision support system to identify habitat
with high conservation value for multiple
species
• Specific recommendations for conservation
in the vicinity of Ft. Bragg
• A set of modeling tools that can be applied to
other installations
• Training 4 PhD students and 3 Post-docs
• Peer reviewed publications
24
Backup Slides
25
What is known? What is not?
Red cockaded woodpecker:
Detailed Demographic Model
Dispersal – Straight-line, random direction
Saint Francis’ satyr:
Detailed movement data in breeding
habitat only
Eastern tiger salamander:
Carolina gopher frog:
Disperse between breeding and nonbreeding habitats
Habitat-specific and boundary behavior unknown for all species.
Helper
Wins
Vacancy
Competition
Age=1 year
Fledgling
Both
Breeders
Die
p=.23
p=.20
p=.50
p=.81
Breeder
Mate
Dies
Found by
Female or
Wins Vacancy
Competition
Death
Birth
p=0
p=.19
Season
to
Leave
Fledge
Disperser
p=.34
p=.38
Wins
Vacancy
Competition
Solitary
Loses
Competition
Finds Empty
Age=1 year
Flowchart of
Male Behavior
Dispersing Bird’s
Search Area
Floater
Loses
Competition
Finds Empty
Wins
Vacancy
Competition
Search Pattern of a Dispersing Bird
26
Dispersal Distance
Distributions for RCW Males
27
28
Collection of Movement Data
Red cockaded woodpecker
• Thirty-five juvenile females fitted with
tail-mounted radio-telemetry during
dispersal periods of years 1 and 2
• Daily recording of geographic
coordinates, hourly when moving.
• 2000+ existing dispersal records
within study area
29
Collection of Movement
Data ETS & CGF
• Radio telemetry to track movement
behaviors of naturally occurring
amphibians
• Monitor dispersal events in occupied
ponds using capture-recapture
• Radio telemetry to track movement
behaviors of experimentally released
amphibians in dominant habitats and
at boundaries
30
Field Data Collection:
Canopy Structure
Randomly located samples stratified by
woody vegetation class
Ten plots for each of 8 woody vegetation
classes; ten 7 m subplots per plot
● canopy height classes
● percent cover by class
● stem density by class
Train/validate visual and kernal-based
methods to extract variables from firstpulse LiDAR DEM
25 m
31
Similar Sets of At-Risk
Animals
32
Reviewer Comments
Comment: Your study area is extremely large and does not relate to the biology of species proposed for
study, except for Red Cockaded Woodpecker. Consideration should be given to reducing the size of study
area (for at least some of the species) which should reduce costs.
Response: We have reduced the study area by about half. This still encompasses areas of known habitat
for all of our study species. Since RCW is a long distance disperser, we have allowed the study area to
extend somewhat further to the west and the NE, which are areas of particular interest for RCW
management.
Comment: Look at the movement of species both on and off of the installation.
Response: It is in our research plan to study dispersal both on and off base for all species.
Comment: Explain any ecological relationship between the four species or the underlying reasons for
selecting these four species.
Response: We have a set of speccies with varying habitat requirements, life-histories, and dispersal
modes, thus reflecting the varied conservation needs of rare, threatened and endangered species found on
many military installations. This reflects the management challenges faced on numerou8s miltary
installations and provides a rigorous test case for the methods we propose. Our study species are not
strongly related ecologically, except that the amphibians are functionally similar and share habitat.
Comment: This project assumes that Dr. Jeffery Walters’ proposal will be funded, what is the implication for
this project if Dr. Walters’ is not funded?
Response: Funds to modify the RCW model based on movement data were moved from the other project
budget to ours in revision, so that our project is no longer dependent on the other one. The two are
complimentary however; their relationship is explained in detail in a document submitted to SERDP by Dr.
Walters.
33