Transcript Design and Analysis of Inventory and Monitoring Studies

MODEL-BASED STRATIFICATIONS
FOR ENHANCING SURVEY
DETECTION RATES OF RARE
SPECIES
Thomas C. Edwards, Jr.
USGS Utah Cooperative Research Unit
Richard Cutler, Mathematics &
Statistics, Utah State University
Niklaus Zimmermann
Swiss Federal Research Institute WSL
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE
 Overview
A (Biased) Historical Perspective of the PNW Forest Plan
The Case of Survey and Manage Species as Rare Events
 Design and Sampling Issues
Detection of rare events
 Example Analyses
Sampling issues related to rare ecological events: lichens as an
example
 Some Final Thoughts
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE
Historical Overview

The Context:





Northern spotted owls like old
forest …
Timber companies like old forest …
A Socio-Economic, Political,
Ecological collision led to …
Listing under the ESA …
And the Northwest Forest Plan
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE
Historical Overview

More Context:



Northwest Forest Plan Record of Decision identified
>350 rare species to be surveyed for management,
including lichens, bryophytes, fungi, and a few token
vertebrates
These species are identified as Survey and Manage
They represent species for which little to no
information is known
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE

Objectives of survey and manage effort were to
obtain estimates of, and/or determine, for EACH
of the >350 S&M species:



Abundance: Is the species abundant at local and
regional scales?
Spatial distribution: Is the species well-distributed
across the area of the Northwest Forest Plan?
Persistence: Do management activities ensure longterm persistence?
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE
Objectives of Survey and Manage

Information to meet objectives comes from:





Existing data
New data
Expert opinion
All must be merged so that simple policy
decisions can be made for each species
Decision framework must be multi-faceted
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE

Objectives of Survey and Manage
Meeting these objectives required significant
exploration into issues of:





Sampling,
Estimation,
Non-Spatial Modelling, and
Spatial Modelling
Can we detect, model, and eventually estimate,
attributes of rare species at landscape scales?
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE





Some constraints affecting ability to meet
objectives:
Rare species are, well, rare!
Limited life history information
available
Some populations exhibit irruptive
behaviors, necessitating multiple
site visits through time
Efficient sample designs a must
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE

Analytical approach




Develop models for common lichens based on
topographic and weather (DAYMET) variables
Translate these models into spatially explicit maps
Use maps as basis of stratification for sampling
associated rare species
Evaluate with independent data and determine if the
models increase detection rates of rare species
RARE ECOLOGICAL EVENTS
IN TIME AND SPACE
Example Analysis: Lichens

Characteristics of data




Forest Service CVS/FIA plots were
basis of sample design
All plots visited; number of visits
variable
Only first visit considered in
subsequent analyses
All lichen species searched for at
each plot
Modeling Survey & Manage Data
Case Studies
Model Families applied
to common species:



Linear logistic regression
(GLM)
Additive logistic regression
(GAM)
Classification trees (CART)
Modeling Survey & Manage Data
Case Studies
Internal Validation:


10 fold cross-validation.
(delete-one jackknife for
logistic regression)
External Validation:

Pilot and other random grid
surveys
Training
Validation
Modeling Survey & Manage Data
Case Studies
Rare & Common overlap (%)
Rare
Common
LobaOreg LobaPulm PseuAnom PseuAnth
LobaScro
-
78.7
83.0
-
NephLaev
-
96.0
76.0
76.0
NephOccu
-
76.9
100.0
-
NephPari
-
77.4
87.1
-
PseuRain
88.9
88.9
77.8
-
Modeling Survey & Manage Data
Case Studies
Summary statistics for Lobaria oregana

Differences in mean values
for presences and absences
for:


Topographic: Elevation,
Easting, and Northing
Weather: Minimum
temperature, Relative
humidity, Rainfall
Modeling Survey & Manage Data
Case Studies
Classification tree for
Lobaria oregana
Measures of model fit



PCC = 94.5%.
PCCAbsent = 94.8%.
PCCPresent = 82.7%.
Modeling Survey & Manage Data
Case Studies
10-fold internal cross-validation of Lobaria
oregana model
Measures of model fit
PCC = 90.5%.
PCCAbsent = 95.9%.
PCCPresent = 72.0%.
Actual



Predicted
Absent
Present
Absent
608
26
Present
52
134
Modeling Survey & Manage Data
Case Studies
External validation of Lobaria oregana model
Predicted



PCC = 81.2%.
PCCAbsent = 90.0%.
PCCPresent = 51.1%.
Actual
Measures of model fit
Absent
Present
Absent
571
63
Present
91
95
Modeling Survey & Manage Data
Case Studies
Measures of error (%) for classification tree models for
three other common lichen species used to model rarer
species
Crossvalidation Prediction
Model

LobaPulm
15.2
18.3
19.3

PseuAnom
12.6
15.4
15.0

PseuAnth
10.2
13.2
15.3
Modeling Survey & Manage Data
Case Studies


Models of common species
applied to spatial data for
PNW and probability of
lichen occurrence estimated
for each location
Estimated number of
detections for each rare
species using stratifications
based on common species
Modeling Survey & Manage Data
Case Studies
Detection likelihoods for rare species LobaScro
5.2
Northing
5.1
5.0
4.9
4.8
4.7
4.0
4.5
5.0
Easting
5.5
6.0
6.5
Modeling Survey & Manage Data
Case Studies
Detection likelihoods for rare species PseuRain
5.2
Northing
5.1
5.0
4.9
4.8
4.7
4.0
4.5
5.0
Easting
5.5
6.0
6.5
Modeling Survey & Manage Data
Case Studies
Observed / Expected (Efficiency)
Rare
LobaOreg
LobaPulm PseuAnom
Common
PseuAnth
LobaScro
-
13/26 (2.0)
13/36 (2.8)
-
NephLaev
-
19/23 (1.2)
19/48 (2.5) 19/60 (3.2)
NephOccu
-
1/5 (5.0)
1/5 (5.0)
-
NephPari
-
7/14 (2.0)
7/16 (2.3)
-
PseuRain
2/1 (0.5)
2/5 (2.5)
2/5 (2.5)
-
Modeling Survey & Manage Data
Conclusions

Stratification applied to
independent region for
field validation


Expected detections for
rare species should be
apportioned across
likelihood bins
Ideal concordance
would be 45° line
Modeling Survey & Manage Data
Conclusions

Common problem when designing surveys for
rare species is sufficient detections for analysis



Design-based approaches provide least biased
estimates, but can lead to low detections
Model-based stratification using more common
species can improve probability of detecting
more rare species
2 to 5-fold gains in detection realized when
process applied to rare epiphytic lichens
Modeling Survey & Manage Data
Conclusions
Questions?

Edwards et al. Enhancing survey detection
rates of rare species using model-based
stratifications. In press, Ecology.

Download at:
ella/gis.usu.edu/~utcoop/tce