Relative cost effectiveness of three yellow-eyed penguin conservation measures

advertisement
Relative cost effectiveness of three
yellow-eyed penguin
conservation measures
Jonah Busch and Ross Cullen
This work has been supported by the National Science Foundation under grant no. 0114437,
and by the NSF East Asia Pacific Summer Institute (EAPSI)
Outline
• Introduction: why look at
•
•
•
•
•
conservation costeffectiveness?
About the yellow-eyed
penguin
Research methods
Effectiveness of
conservation measures
Cost-effectiveness of
conservation measures
Discussion
Conservation is a grand investment…
• US$6 billion annually on nature reserves
worldwide (James et al, 1999)
• US$1.5 billion spent in 2002 by
international conservation organizations
(Halpern et al, 2006)
• NZ$106.5 million annually on natural
heritage (DOC, 2004)
Q: Are we getting our
money’s worth?
A: We have no idea!
• “Few well designed empirical analyses assess
even the most common biodiversity conservation
measures” (Millenium Ecosystem Assessment,
2005)
• “If any progress is to be made in stemming the
global decline of biodiversity, the field of
conservation policy must adopt state-of-the-art
program evaluation methods to determine what
works and when.” (Ferraro and Pattanayak,
2006)
Why should we care?
• Knowing which conservation measures are
most cost-effective will enable us to use
existing conservation resources most
effectively
• Demonstrating conservation effectiveness
will attract new resources (untested
assertion)
Previous analyses of conservation
effectiveness and cost-effectiveness
• Parks (Bruner et al, 2001)
• U.S. Endangered Species Act (Abbitt and Scott,
2001; Ferraro et al, in press)
• TNC conservation easements (Kiesecker et al,
2007)
• Predator control (Engeman et al, 2002; Shwiff et
al, 2005)
• NZ endangered species conservation programs
(Cullen et al, 2001; Cullen et al, 2005)
Possibilities for effectiveness
evaluation are limited
• No objective independent or dependent
variable (Abbitt and Scott, 2001)
• No data on biological outcome (Kiesecker
et al, 2007)
Golden opportunity to evaluate
conservation cost effectiveness
• Yellow-eyed penguin
conservation program
– Stationary, observable
species
– Nest counts across 48
sites, 15 years
– Three conservation
measures used at different
sites at different times:
• Trapping
• Revegetation
• Intensive Management
About the yellow-eyed penguin
(Megadyptes Antipodes)
•
•
•
•
•
•
•
•
Large, long-lived species
Endemic to New Zealand
Nest on land; feed at sea
Philopatric
Begin breeding at age
2-3(F) or 2-5(M)
Require sheltered nests in
visual isolation
One to two chicks a year
High juvenile mortality
(~80%)
Photo Credit: ?
A charismatic and valuable bird
• Penguins are highly appealing
•
•
to humans, though yelloweyed are among the least
aesthetic (Stokes, 2006)
Yellow-eyed penguins elicit
feelings of wonder, improve
mood, and increase
environmental awareness
(McIntosh, 2000)
Substantial NZ penguin
tourism industry—126,000
overseas visitors viewed
penguins in 2006-2007
(Ministry of Tourism, 2007)
Photo Credit: Penguin place
Threatened from all directions
• Terrestrial Predators (mustelids,
•
•
•
•
•
cats, dogs)
Marine Predators (sea lions,
sharks)
Trauma
Gill nets
Disease
Starvation
Photo Credit: tuxxie.org
(Hocken, 2005)
• Toxic algal blooms (Shumway,
•
2003)
Unmanaged Tourism (Ellenberg et
al, 2007)
Photo Credit: James D. Watt
Conservation status
• 5930-6970 penguins in 1997
•
•
•
•
Payoff Continuum
100.00
Not threatened
Sparse
90.00
Range restricted
80.00
Gradual decline
70.00
Payout value
•
(McKinlay, 2001);
IUCN redlist: “endangered”
(BirdLife International, 2005)
Most endangered of 17
penguin species
DOC: “nationally vulnerable”
(Hitchmough, 2005)
464 South Island nesting pairs
in 2006 (DOC unpubl.)
DOC goal: 1000 South Island
nesting pairs by 2025
Serious decline
60.00
50.00
Nationally vulnerable
40.00
30.00
Nationally endangered
20.00
10.00
Nationally critical
0.00
1
2
3
4
5
Species Status
6
7
8
The penguin conservation mosaic
• Conservation Actors
– DOC
– Yellow-eyed penguin trust
– Private landowners and
conservationists
• Conservation measures
– Trapping
– Revegetation
– Intensive management
Trapping
•
•
•
•
Baited traps for stoats and ferrets
Near nest sites, or set along lines
Year round, or during nesting season only
Rationale: reduce terrestrial mortality
Revegetation
• With native forest species, or with nest boxes
surrounded by flax
• Rationale: indirectly increase breeding success
Intensive Management
• Rehabilitation of sick, injured, and
underweight penguins
• Regular monitoring of nests and traps
• Rationale: directly increase juvenile and
adult survivorship
So what works?
Data
• Dependent variable:
– Nest counts across 48 South Island sites, from 19922006 (DOC unpubl.)
• Independent variables:
– Conservation measures used (compiled from
interviews: McKinlay, MacFarlane, Goldsworthy,
Sutherland, Ratz, Lalas)
• Control variables:
– Site area (McKinlay, 1997)
– Site location (MapToaster, 2007)
• Cost (Spencer, McKinlay pers.comm.)
Econometric Specification
it=nt/nt-1 – change in nests at site i between year t1 and year t
• X – conservation measures taken
•
 – nest density
y – year dummy
• Assumption: use of a particular conservation
measure is exogenous to intrinsic probability of its
success; E(Xitit)=0.
Summary Statistics
Statistic
n
min
max
median
mean
st. dev
Year
15
1992
2006
-
-
-
Site
48
-
-
-
-
-
Population(site i, year t)
646
0
55
8
10.74
10.24
Population>0(site i, year t)
573
1
55
10
12.1
10.09
lambda
519
0.1429
7
1
1.0974
0.5408
log lambda
519
-1.9459
1.9459
0
0.0057
0.4111
trapping
204
0
1
-
-
-
revegetation
140
0
1
-
-
-
intensive management
57
0
1
-
-
-
area (Ha)
46
1
25
5
6.9891
5.5981
density
506
0.1111
21
1.6667
2.4296
2.2436
Regression Results
1
2
3
519
506
506
Intercept
0.0017
(0.0232)
0.0463*
(0.0246)
0.1099
(0.0730)
Trapping
0.0209
(0.0419)
0.0221
(0.0413)
0.0347
(0.0391)
Revegetation
-0.0322
(0.0467)
-0.0095
(0.0461)
-0.0246
(0.0436)
Intensive Management
0.0425
(0.0615)
0.1340**
(0.0631)
0.1203**
(0.0595)
Log Density
no
-0.1104***
(0.0216)
-0.0998***
(0.0207)
Year Dummy
(intercept=2006)
no
no
yes
10 km site effects
no
no
no
3 year lag
no
no
no
n
Alternate Specifications
• (4) Spatial autocorrelation—includes a separate
regressor for number of sites within 10km at
which conservation measure was taken
• (5) Temporal lag—conservation measures
lagged by 3 years (time from chick to breeding
adult)
• (6) Sea lion—controls for single rogue Hooker’s
sea lion which has feasted massively on
penguins at two intensively managed sites
Regression Results (2)
1
2
3
4
5
6
519
506
506
506
506
506
Intercept
0.0017
(0.0232)
0.0463*
(0.0246)
0.1099
(0.0730)
0.1181
(0.0848)
0.1221*
(0.0725)
0.1093
(0.728)
Trapping
0.0209
(0.0419)
0.0221
(0.0413)
0.0347
(0.0391)
0.0581
(0.0451)
-0.0047
(0.0413)
0.0267
(0.0393)
Revegetation
-0.0322
(0.0467)
-0.0095
(0.0461)
-0.0246
(0.0436)
-0.0451
(0.0472)
-0.0025
(0.0457)
-0.0115
(0.0443)
Intensive Management
0.0425
(0.0615)
0.1340**
(0.0631)
0.1203**
(0.0595)
0.0959
(0.0641)
0.1198*
(0.0630)
0.1951***
(0.0742)
Log Density
no
-0.1104***
(0.0216)
-0.0998***
(0.0207)
-0.1039***
(0.0212)
-0.0976***
(0.0206)
-0.1009***
(0.0207)
Year Dummy
(intercept=2006)
no
no
yes
Yes
Yes
Yes
10 km site effects
no
no
no
Yes
No
No
3 year lag
no
no
no
No
Yes
No
Sealion
no
no
no
No
No
-0.1871*
(0.1111)
0.0023
0.0519
0.1898
0.1921
0.1865
0.1945
4.3%
14.3%
12.8%
10.1%
12.7%
21.5%
n
R2
IM growth rate equivalent
Cost effectiveness of conservation
measures
• CEx=(N2006-Nw/o x,2006)/Cx
• Need to know:
– Actual nest #’s
– Counterfactual nest #’s
– Cost of conservation measures
Actual and counterfactual
• Actual:
ln(it )   0   1 X it  4   2 ln  it 1   3 y   it
nit  nit 1e
( N (  0 , 0 )  N ( 1 , 1 ) X i ,t 1  N (  2 , 2 ) ln  i ,t 1  N (  3 , 3 ) y t )
• Counterfactual:
nˆ it  nˆ it 1e
(ln(
nit
)  N ( 1 , 1 ) X it 1 )
nit 1
Cost of trapping
First Year
Materials
Each Subsequent Year
$50
5
$12.50
5
$312.50
Labor
$10
1
52
Total
per trap
$50
per trap
traps per hectare
0.5
traps per hectare
bait and poison per trap
per year
traps per hectare
per hectare
per person-hour
hour per hectare per
week
weeks per year
$12.50
5
$87.50
bait and poison per trap
per year
traps per hectare
per hectare
$16
per person-hour
1
hours per week
52
weeks per year
$ 832
per hectare
$832
$1,144.50
per hectare
$919.50
per hectare
per hectare
Cost of revegetation
First Year
Materials
Years 2, 4, 6
$3
2500
$7,500
Labor
Total
per plant
plants per hectare
per hectare
$16
per person-hour
4.5
hours
50
people
$16
8
4.5
per person hour
hours
people
$3,600
per hectare
$576
per hectare
$11,100
per hectare
$576
per hectare
Cost of Intensive Management
Materials
Labor
Total
$50
$45,000
food and medicine, per nest
full time ranger, per site
$40,000+$50/nest
Average cost per site-year
Trapping
Revegetation
Intensive
Management
Total Cost
$1.456,667
$1,344,252
$2,619,350
$5,420,269
Number of site-years
204
140
57
272
Average cost per
site-year
$7,141
$9,602
$45,954
$19,927
Total
Average cost-effectiveness
(1)
Simple
Nests
Nests gained
Nests gained/site-year
Nests gained/$100,000
(2)
w/ density
(3)
w/ density,
year effects
(6)
w/ density,
year effects, sea lion
actual
468.1 (1.0)
461.8 (1.7)
470.3 (4.5)
464.0 (3.1)
w/o traps
419.4 (7.7)
409.1 (8.0)
390.9 (8.4)
400.2 (6.2)
w/o IM
437.8 (5.9)
396.3 (2.7)
407.0 (4.4)
387.9 (3.6)
w/o anything
438.1 (11.7)
394.0 (9.4)
369.3 (10.7)
350.7 (8.4)
w/ traps
49.0 (7.6)
53.0 (7.9)
78.7 (7.0)
63.5 (6.1)
w/ IM
30.5 (5.7)
65.4 (2.5)
62.3 (2.8)
76.0 (1.9)
w/ everything
29.9 (11.7)
97.8 (9.0)
101.0 (9.9)
113.3 (8.1)
w/ traps
0.24 (0.04)
0.26 (0.04)
0.39 (0.03)
0.31 (0.03)
w/ IM
0.54 (0.10)
1.15 (0.04)
1.09 (0.05)
1.33 (0.03)
w/ everything
0.11 (0.04)
0.36 (0.03)
0.37 (0.04)
0.42 (0.03)
w/ traps
3.36 (0.52)
3.64 (0.54)
5.40 (0.48)
4.36 (0.42)
w/ IM
1.16 (0.22)
2.50 (0.10)
2.38 (0.11)
2.90 (0.07)
w/ everything
0.55 (0.25)
1.80 (0.17)
1.86 (0.18)
2.09 (0.15)
Discussion: Implications for yelloweyed penguin management
• Intensive management is most effective
• Trapping is most cost-effective
• Revegetation may have other benefits
• Research what makes IM effective
• Research marine effects on penguins
• Trial intensive management at more sites,
while continuing trapping
Discussion: Broader implications for
conservation
• Monitoring biological output is essential
• Unless species is critically imperiled, leave
control sites and monitor these as well
• Keep track of costs
Take home message
• Evaluating the cost effectiveness of
conservation activities can maximize
penguins (or any species) provided per
dollar spent.
• Demonstrating conservation cost
effectiveness should attract new funding.
Thank you!
Photo Credit: Dean Schneider?
Are treatments exogenous?
• Add regional dummies: no change in significance of
•
•
•
•
explanatory variables
Add fixed effects: all explanatory variables become
insignificant
Difference in area between treated and untreated sites? Not
significant.
Difference in nest density between untreated sites which
will/won’t receive treatment later?
No for trapping, IM. Reveg occurs at sites with significantly
lower nest density.
Expected sign on likely bias: If treatments are more likely to
be put in place where they are expected to make a difference,
then coefficients on effect of treatment are overestimates.
Download