Using Ecopath Modeling for Analysis and Management of Estuarine Food Web

1
Using Ecopath Modeling for Analysis and
Management of Estuarine Food Web

• Tuesday, Nov 30, 2004 - Friday, Dec 3, 2004
• 900 a.m. 500 p.m. Friday close at 1 p.m.
• NOAA Fisheries Laboratory, Building 216,
  Galveston, Texas
• Sponsors Galveston Bay Estuary Program, Texas
  Sea Grant, NOAA Fisheries at Galveston, and
  Houston Advanced Research Center

2
Purpose

• To give participants an overview and basic
  knowledge of the Ecopath with Ecosim approach and
  software
• To introduce the newest developments in EwE
• To present methods, capabilities and limitations
  of the approach and software
• To improve models participants may already have
  constructed
• To explore management options as part of an
  ecosystem approach to fisheries.

3
Program (tentative)
4
Ecopath with Ecosim its use for ecosystem-based
management
Carl Walters, Villy Christensen, Sherman Lai,
UBC
Galveston, 30 Nov. 3 Dec. 2004
5
Fishing down the food web
Northeast Atlantic
1950
1994
We get less fish in spite of the lower trophic
level
Pauly et al., Science, 1998, 279
6
Ecosystem breakdown?

• Backward bending curves
• Feeding triangles A North Sea example

7
Norway pout in the North Sea
8
Feeding triangles North Sea
4
Other fish
1
2
Norwaypout
9
Feeding triangles North Sea
4
Other fish
1
2
Norwaypout
50
5
17
Krill
11
100
Copepods
10
Feeding triangles North Sea
4
Other fish
1
2
Norwaypout
50
5
17
Krill
11
100
Copepods
11
Ecosystem-based management

• Why?
• There are questions that cannot be addressed
  using our traditional methods
• The old questions still need to be answered
• Ecosystem approaches will not replace stock
  assessment.
• Adopting ecosystem-based approaches poses new
  challenges
• how are we going to manage at the ecosystem
  level, considering the problems we have now
  managing stocks?

12
Ecosystem model publications

• c

13
Ecopath with Ecosim

• Describe ecosystem resources and their
  interactions
• Evaluate ecosystem effects of fishing (incl.
  indirect effects, e.g., through habitat
  modifications)
• Evaluate effects of environmental change
• Predict bioaccumulation of persistent pollutants
• Evaluate impact and placement of marine protected
  areas
• Evaluate uncertainty in the management process
• Explore management policy options incorporating
  economic, social, legal, and ecological
  considerations
• EwE is widely and freely distributed 3000 users
  in 125 countries.

14
and there are also activities on the other side
NMFS, Bering Sea, GoAlaska
Greenland Fisheries Inst.
Prince William Sound
Faroe Fisheries Inst
UBC
IMR, Bergen
DFO
Four Fish. Commissions
Baltic Sea RP, GEF
DIFRES, Charlottenlund
UoWisconsin
NMFS, Chesapeake Bay
CEFAS, Lowestoft
Virginia IMS
Black Sea, Turkey
Trop. Tuna Comm.
Venice
NCEAS
Santander
S Atlantic Fish.Comm.
Mote lab
La Paz, Mexico
Fish. Inst, Lisboa
G.o Mexico
Azores F.I.
West Florida
Mote Lab.
Six West African Countries
Yucatan reefs
Jamaica, BVI,
Colombia
Trinidad
Venezuela
training courses / workshops
Charles Darwin Research Station, Galapagos
Namibia
Abrolhos, Brazil
Cape Town
Sao Paulo, Brazil
Tongoy Gulf, Chile
?
Concepcion, Chile
Argentina
EwE project activities
15
Application to ecosystem trophic modeling
Ecopath with Ecosim

• Ecopath is used to organize historical data on
  trophic interactions and population sizes
• Ecosim builds dynamic predictions by combining
  the data with foraging arena theory
• Ecospace for addressing spatial policy questions,
  notably re. protected areas.

16
Information for management from single-species
to ecosystem-based approaches
Biology
Ecology
Biodiversity
Environment
Abundance Growth Mortality Recruitment Catches
Fleet dyn.
Feeding rates Diets Interaction
terms Vulnerabilities Habitats
Occurrence Distribution Rebuilding
Prim. Prod, SST,
Migration Dispersal
Economics
Costs Prices Values
Single-species approaches
Social cultural considerations
Ecosystem approaches
Y/R VPA Surplus production .
Employment Conflict reduction, ...
EwE .
Tactical
Strategic
17
EwE an overview
Data
Model
Research
Application
Manual
Who eats whom? Network analysis
Biol. B, P/B, Q/B, diet. Fleet catches
Mass-balance (Ecopath)
Ecoranger
Academic (ecol. theory)
Automatic
Functionalresponse, etc.
Seaso-nality
Sensitivity analysis
Pedigree ? M.Carlo
Economics, social info.
Policy exploration
Fisheries vs. environment
Time-dynamic (Ecosim)
Vulnerability, mediation,
Biol. fishing time series
Fisheriesmanagement
Tracer- dynamic (Ecotrace)
Environmental time series
Persistent pollutants
Protected area dynamics. Spatial effort
allocation
Ocean zoning
Habitat preference, dispersal, migration etc.
Spatial-dynamic (Ecospace)
Spatial cost of fishing
Nutrient-O2 seagrass,
MPA size (Ecoseed)
Prim.prod.(SeaWIFS)
Legend
Facultative input
Runoff, nutri-ents, depth,
Optional input
18
Ecopath

• Software system
• Open source-code

• Trophic model
• Mass balance
• Time-invariant
• Not steady-state!

19
Ecopath can incorporate all ecosystem groups and
all fisheries
20
Master Equation (I)
Production predation fishery biomass
accumulation net migration other
mortality
21
Master Equation (II)
Consumption production respiration
unassimilated food
22
Mass balance cutting the pie
Other mortality
Harvest
Unassi- milated food
Predation
Harvest
Respi- ration
Respi- ration
Predation
Predation
Unassi- milated food
Other mortality
Consumption
Other mortality
Unassi- milated food
Predation
Predation
Respi- ration
23
Application to ecosystem trophic modeling
Ecopath with Ecosim

• Ecopath is used to organize historical data on
  trophic interactions and population sizes
• Ecosim builds dynamic predictions by combining
  the data with foraging arena theory
• Ecospace for addressing spatial policy questions,
  notably re. protected areas.

Sherman Lai
24
The guts of Ecosim Foraging arena
What if?
25
(No Transcript)
26
Foraging arena is a theoretical entity

• May be impossible to observe directly or describe
  precisely
• Useful as a logical device for constructing
  predictions and interpreting data.

27
Organisms are not chemicals!
Ecological interactions are highly organized
Reaction vat model
Foraging arena model
Prey behavior limits rate
Predator handling limits rate
Big effects from small changes in space/time scale
28
Prey vulnerability top-down/bottom up control
Predator, P
aVP
Available prey, V
v(B-V)
vV
Unavailable prey B-V
v predator-prey specific behavioral exchange
rate (vulnerability) Also includes
Environmental forcing, nutrient limitation,
mediation, handling time, seasonality, life stage
(age group) handling,
29
A critical parameter vulnerability
Its all about carrying capacity
30
Predation mortality effect of vulnerability
Predicted predation mortality
Traditional
Ecosim
0
Carrying capacity
Predator abundance
31
Limited prey vulnerability causes compensatory
(surplus) production response in predator biomass
dynamics
1.0
If predator biomass is halved
Predator Q/B response -- given fixed total prey
abundance
0.5
0.0
If predator biomass is doubled
-0.5
CarryingCapacity
0
Predator abundance
32
Time predictions from an ecosystem model of the
Georgia Strait, 1950-2000
With mass-action (Lotka-Volterra) interactions
only
With foraging arena interactions
33
Ecosim predicts ecosystem effects of changes in
fishing effort
34
Ecosim can use time series data
35
Time series data
Drivers
Validation

• Fishing mortality rates
• Fleet effort
• Biomass (force)
• Time forcing data (e.g., prim. prod., SST)

• Biomass (relative, absolute)
• Total mortality rates
• Catches
• Average weights
• Diets

Yes, lots of Monte Carlo
36
Time series fitting Strait of Georgia
37
Northwest Hawaiian Islands (French Frigate
Shoals)
Fishing effort
Initial model runs fishing trophic
interactions only did not explain monk
seal decline, predicted lobster recovery.
Polovinas insight satellite chlorophyll
data indicate persistent 40-50 decline in
primary production from 1990.
Lo Chl
38
How can we test complex ecosystem models?

• No model fully represents natural dynamics, and
  hence every model will fail if we ask the right
  questions
• A good model is one that correctly orders a set
  of policy choices, i.e. makes correct predictions
  about the relative values of variables that
  matter to policy choice
• No model can predict the response of every
  variable to every possible policy choice, unless
  that model is the system being managed
  (experimental management approach).

39
So how can we decide if a given model is likely
to correctly order a set of specific policy
choices?

• Can it reproduce the way the system has responded
  to similar choices/changes in the past (temporal
  challenges)?
• Can it reproduce spatial patterns over locations
  where there have been differences similar to
  those that policies will cause (spatial
  challenges)?
• Does it make credible extrapolations to entirely
  novel circumstances, (e.g., cultivation/depensatio
  n effects)?

40
Ecosystems where EwE has been tested using
historical trend data

• E Bering Sea
• Aleutian Islands
• WC GoAlaska
• E GoAlaska
• W Vancouver Island
• Strait of Georgia
• GoCalifornia
• NE Pacific
• Central N Pacific
• FF Shoals, Hawaii
• Central Chile

• Bay of Quinte
• Oneida Lake
• Scotian Shelf
• Cheasapeake Bay
• Tampa Bay
• S Brazil Bight
• North Sea
• Baltic
• S Benguela
• Gulf of Thailand

41
Time predictions from an ecosystem model of the
Georgia Strait, 1950-2000
42
Central Pacific Ocean 1952-1998
Other marlin
Large albacore
Large bigeye
Large yellowfin
CPUE
Ecosim
Single-species assessment
Skipjack
Small albacore
Small bigeye
Small yellowfin
Biomass (kg km-2)
Blue marlin
Swordfish
Other shark
Blue shark
Cox et al. CJFAS in press
43
Baltic Sea model vs. survey
Cod, age 2-8
Herring, age 0-1
Herring, age 2-8
Sprat, age 0-1
Sprat, age 2-8
Ecosim predictions (solid line) vs. survey
indices (dots), where an environmental forcing
function on phytoplankton production was tuned to
the survey indices.
44
Baltic environmental forcing
Comparison of the fitted environmental forcing
function (FF, monthly) and the spawning volume
index (Spawn, annual) for the Baltic.
45
The big Ecosim worries

• Predation vulnerability settings
• Thompson-Burkenroad debate fisheries trophic
  vs. environmental effects
• Overestimates of total other mortality
• Cultivation/depensation effects at extreme
  states
• ...

46
Experience so far

• Prim. productivity patterns are amplified rather
  than dampened up the food web
• Separates top-down (fisheries, predators) vs.
  bottom-up (primary production, nutrients)
  effects
• Requires more data but mainly data we should
  have at hand in any case the ecosystem
  history
• Supplements single species assessment, does not
  replace it

• When we have a modelthat can replicate
  development over time we can (with some
  confidence) use it for ecosystem-based policy
  exploration.

47
If ecosystems are so complex, why do models often
seem to work quite well?

• Maybe they dont (past experience not a useful
  guide to the future)
• Strong external forcing, (e.g., overfishing)
• Structural complexity does not imply complexity
  in dynamic behavior of aggregate measures (like
  biomasses)!

48
Are we finally able to develop useful predictive
models for ecosystem management?

• Its beginning to look like it
• Recent success in replicating population history
  for a series of ecosystems
• We can with some credibility describe agents of
  mortality and trophic interdependencies
• Now we need to test how models predict policy
  choices.

49
It is useful to test prospective management
strategies against ecosystem models if they
don't work on simple models why should they work
in reality Keith Sainsbury ICES/SCOR
Conference, Montpellier March 1999
Andrew Trites
50
Ecosystem-scale optimizationpolicy objectives

• Maximize fisheries profit
• Maximize social benefits
• Maximize mandated rebuilding
• Maximize ecosystem health.
• We can evaluate the fleet configuration and
  effort levels that optimize each of these
  objectives individually or jointly

51
Ecospace for addressing spatial policy questions

• Replicates Ecosim dynamics over spatial grid of
  homogeneous cells
• Links cells through dispersal of organisms and
  fishing effort movement/allocation
• Incorporates an advection model
• Accounts for spatial variation in productivity
  and cost of fishing
• Represents habitat preferences by differential
  dispersal, feeding and predation rates.

52
Spatial model Ecospace

• Links to
• GIS databases for PP, Depth, T, habitat
  structures,...
• FishBase for depth pref., ...

53
Ecospace 2D advection
54
Ecospace seasonal or full-time closures

• Is designed to
• Quantify effect of protected areas
• Predict spatial distributions
• Evaluate spatial effects, e.g., feeding related
  or of habitat changes,

55
Ecospace predicts spatial distributions and
impact of protected areas habitat changes
56
Observed effort, Faroe Island
57
Driving EwE with hydrography

• Ecosim and Ecospace are linked to a hydrographic
  model
• Initial application completed Tampa Bay
• Second in progress Chesapeake Bay.

Chlorophyll a
58
Florida Ecosystem Model (FLEM)

• Outflows from cells that receive direct rainfall
  and runoff inputs, using historical time series
  data on these inputs
• Small-scale mixing due to tidal and eddy
  diffusion processes, parameterized by fitting the
  model to time series salinity and nutrient data
• Larger-scale mixing due to wind-driven
  circulation, predicted from historical input data
  on wind speeds and directions
• Large-scale mixing due to regional pressure
  fields represented by sea surface height
  anomalies at the model grid boundaries (general
  flow across/along the grid due to regional
  coastal circulation processes)

59
Input Data for FLEM
Rainfall
Wind
Nutrient Loading
River Gauge Data
Detailed bathymetry
60
Approach

• Hydrographic model
• 1 month time step
• 2-layers (deep and shallow water)
• Calculates monthly concentrations
• Sea grass area
• Chlorophyll
• Oxygen conditions
• Salinity
• Used to force
• Ecosim
• Ecospace

61
FLEM Predictions

• Salinity
• Total nitrogen
• Active organic carbon measured as BOD

• Dissolved oxygen
• Chlorophyll concentration,
• seagrass biomass

62
Linkages between FLEM and Ecosim

• Use Chlorophyll time series to force primary
  production in Ecosim.
• Sea grass density mediation effects.
• Chl. a and sea grass
• Epiphytes
• Habitat area for some fishes

63
Chlorophyll and seagrass (1955)
Chlorophyll a
Seagrass biomass
64
Chlorophyll and seagrass (2000)
Chlorophyll a
Seagrass biomass
65
Effects of PP forcing
66
Ecotrace bioaccumulation

• Quantifies accumulation of persistent pollutants
  (or nutrients) through the food web
• Example shows accumulation through the pelagic
  part of the Northern Benguela system
• Here illustrated with a point source, may also be
  continuous.

67
Models are not like religion

• you can have more than one

• and you shouldnt believe them

Andrew Trites
68
End