Transcript A Swomee - Swan Population Restoration Model

Searching for a good stocking policy
for Lake Michigan salmonines
Michael L. Jones and Iyob Tsehaye
Quantitative Fisheries Center, Fisheries and Wildlife
Michigan State University
Lake Michigan Decision Analysis - 2012
1
Decision Analysis
Structured, formal method for comparing
alternative management actions
Main components:
 Specify objectives
 Identify management options
 Assess knowledge and account for uncertainties
 Use model to forecast possible outcomes
• Consider the possible consequences of a decision, rather
than just predicting the most likely consequence
Lake Michigan Decision Analysis - 2012
2
The Big Question
How many salmon and trout should we stock into Lake
Michigan each year?
• more stocking leads to greater harvest, and thus benefits
- unless…
• too much stocking leads to poor feeding conditions and
increased mortality, but
• too little stocking may lead to negative effects of alewife
on other species
Lake Michigan Decision Analysis - 2012
3
What we need to know…
1. How many salmon and trout are out there?
2. How much do they eat?
3. How capable are the prey fish of meeting this demand?
4. What happens to salmon and trout feeding (and survival)
when prey numbers are low?
Lake Michigan Decision Analysis - 2012
4
Our approach
1. Analyze the past
Data we used
•
Salmonine abundance
•
Stocking and harvest
•
Salmonine
consumption
•
Growth and diet data
•
Prey fish survey data
•
Prey fish production
•
Supply vs demand
2. Forecast the future
•
Simulation model
Lake Michigan Decision Analysis - 2012
5
What does the past tell us?
Lake Michigan Decision Analysis - 2012
6
How many salmon and trout are out
there?
 Total salmonine numbers have remained relatively stable
since 1990
 Reduced Chinook stocking has been offset by increased wild
fish production
 More recently, improved survival of older Chinook salmon
has also offset reduced stocking
Lake Michigan Decision Analysis - 2012
7
How many salmon and trout are out
there?
Salmonine abundance
Lake Michigan Decision Analysis - 2012
Age-3 Chinook numbers
8
How much do they eat?
 Total consumption has remained fairly stable for last decade
 Chinook salmon have accounted for more than half of total
demand consistently since 1980
 Large alewife accounted for more than 40% of total prey
consumed since 1980, except in the late 1980s when small
alewife dominated
Lake Michigan Decision Analysis - 2012
9
How much do they eat?
Consumption by predator
type
Consumption by all species
of salmon and trout
100
80
60
1 KT = 2.2 million lbs
40
Consumption by prey type
20
0
1965
0.80
0.60
1975
1985
1995
2005
Lake Michigan Decision Analysis - 2012
Proportion
Consumption (KT)
120
Large alewife
Small alewife
Smelt and others
0.40
0.20
0.00
1965
1975
1985
1995
2005
10
How capable are the prey fish of
meeting this demand?
 Predation rates on alewife have ranged from 25%-45% per
year from the mid-1980’s to present
 Predation mortality peaked in mid-1980’s and has
approached peak levels again recently
 Alewife (and rainbow smelt) recruitment is variable and not
strongly related to adult abundance
Lake Michigan Decision Analysis - 2012
11
How capable are the prey fish of
meeting this demand?
Index of total predation mortality on alewife
Predation Mortality Index
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
1960
1970
1980
Lake Michigan Decision Analysis - 2012
1990
2000
2010
12
How capable are the prey fish of meeting
this demand?
Age-0 recruits (109)
60
50
Stock and
recruitment
40
30
20
10
0
0
10
20
30
Age 2+ numbers (109)
40
What happens to salmon and trout feeding
when prey numbers are low?
 Chinook salmon consumption has declined when alewife
abundance declined
 Similar, but weaker pattern for lake trout
Lake Michigan Decision Analysis - 2012
14
What happens to salmon and trout feeding
when prey numbers are low?
23
Chinook age-3 ration
Alewife age-3 abundance
Ration (kg)
20
22
15
21
10
20
5
0
1960
Alewife abundance index
25
19
1970
1980
Lake Michigan Decision Analysis - 2012
1990
2000
2010
15
What can we
say about the
future?
Lake Michigan Decision Analysis - 2012
16
Policy simulation model
Accounts for
uncertainties: key
uncertainties
concern prey
recruitment (supply)
and predator
feeding (demand)
What we
Know
Management
Decisions
LMDA
Prediction of
Outcome
y  x2  x  5
ƒ( x )
n
ˆ 
2
Lake Michigan Decision Analysis - 2012
 x  ˆ 
2
i
i 1
n 1
17
Multiple possible futures
Age-0 recruits (109)
60
Variability in actual
recruitment in a particular
year from the “average”
recruitment
50
40
Alternative relationships
that are consistent with
the data
30
20
10
0
0
10
20
Age 2+ numbers (109)
30
40
Model results
 The model forecasts possible future changes in fish
populations and harvest, given a stocking policy
 There are many possible futures, so we need to look at the
range of possible (likely) outcomes
 This range tells us what we think is most likely, but also what
might happen
 Mainly we’re interested in how likely it is that bad things will
happen
 Here’s how it works…
Lake Michigan Decision Analysis - 2012
19
Generating results:
First simulation
Average
biomass = 243 kT
Alewife Biomass (kt)
350
300
250
200
150
100
50
0
0
5
10
15
20
25
Year
Lake Michigan Decision Analysis - 2012
20
Generating results
Number of Simulations
10
First simulation:
average alewife
biomass = 243 kt
8
6
4
2
0
< 100
100-500
> 500
Biomass (kt)
Lake Michigan Decision Analysis - 2012
21
Generating results:
Second simulation
140
Average
biomass = 52 kT
Alewife Biomass
120
100
80
60
40
20
0
0
5
10
15
20
25
Year
Lake Michigan Decision Analysis - 2012
22
Generating results
Number of Simulations
10
8
6
Second simulation:
average alewife
biomass = 52 kt
4
2
0
< 100
100-500
> 500
Biomass (kT)
Lake Michigan Decision Analysis - 2012
23
Generating results
Number of Simulations
10
8
… and so on (e.g.,
results after 15
simulations)
6
4
2
0
< 100
100-500
> 500
Biomass (kt)
Lake Michigan Decision Analysis - 2012
24
An example result: Status quo policy
Number of simulations
50
In 26 of 100
cases alewife
biomass was
less than 100
kt: BAD
In 45 of 100 cases
alewife biomass
was between 100
and 500 kt: OK
40
30
20
10
0
<100
100-500
>500
Biomass (KT)
Lake Michigan Decision Analysis - 2012
25