Ecopath with Ecosim 6: S3‐4136
New generation ecosystem modeling packagemodeling package
Villy Christensen, Joe Buszowski, Robyn Forrest, Fang Gao, y , , y , g ,Carie Hoover, Joe Hui, Sherman Lai, Jeroen Steenbeek,
William Walters and Carl Walters
PICES 16th Annual Meeting, Victoria, October 30, 2007
/CCCC/FIS Topic Session S3Towards ecosystem‐based management:
Recent developments and successes in multi‐species modeling
NOAA celebrates 200 yearhttp://celebrating200years.noaa.gov/breakthroughs/welcome.html
2
Prince William SoundNMFS, Bering Sea, GoAlaska Greenland
Fisheries Inst.Faroe Fisheries InstIMR, Bergen
DFOUBC
Swedish Fisheriesh
Stockholm Univ.FIMR, Helsinki
Four Fish. Commissions
DIFRES
CEFAS, Lowestoft
NOAA, Chesapeake Bay
Trop. Tuna Comm.Virginia IMS
NCEAS
UoWisconsin PolandEstonia, Latvia, Lithuania
Mote lab Santander
FL FWRI
S Atlantic MFC
Fish. Inst, LisboaLa Paz, Mexico
A
Venice/Trieste
NCEAS
NMFS, Galveston
NMFS, Honolulu Thetis, Greece
GoMexico FMC
IFREMER
Six West African Countries
ColombiaTrinidad, Tobago
Jamaica, BVI, …
Azores F.I.
G.o MexicoYucatan reefs
G.o.Guinea 19 countries
Dominican Rep.
GoMexico FMC
Charles Darwin Research Station, Galapagos
Venezuela = training courses / workshops
Angola
countries
and there areConcepcion, Chile
Namibia
Cape TownTongoy Gulf, Chile
Sao Paulo, BrazilAbrolhos, Brazil
>300 models>4500 reg. users
3
and there are also activities on
the other side
p ,Argentina
Ecopath project activities
155 countries
EwE data models and uses
EwE data models and usesdata, models, and usesdata, models, and uses
ParameterizationParameterizationEcopathEcopath
Abundance, mortality, Abundance, mortality, catches, feeding rates, catches, feeding rates,
diets, growth, interaction diets, growth, interaction i ii i
Environmental Environmental assessments, network assessments, network analysis, ecological analysis, ecological
h fi h ih fi h iTimeTime‐‐dynamicsdynamics
EcosimEcosim
terms, carrying capacity, terms, carrying capacity, persistent pollutants, persistent pollutants, habitats, nutrient habitats, nutrient loading climateloading climate
theory, fisheries managetheory, fisheries manage‐‐mentment, adaptive manage, adaptive manage‐‐mentment, culling, ecosystem , culling, ecosystem manipulation protectedmanipulation protected
SpatialSpatial‐‐dynamicsdynamicsEcospaceEcospace
loading, climate loading, climate indicators, migration, indicators, migration, dispersal, occurrence, dispersal, occurrence, distribution costsdistribution costs
manipulation, protected manipulation, protected areas, spatial zoning, areas, spatial zoning, policy exploration, policy exploration,
fisheries sectorfisheries sector optimioptimi‐‐Persistent Persistent pollutant pollutant dynamicsdynamics
distribution, costs, distribution, costs, subsidies, prices, values, subsidies, prices, values,
existence values, existence values, employmentemployment
fisheries sector fisheries sector optimioptimizationzation, conflict reduction, , conflict reduction,
climate scenarios, climate scenarios, gaming, ...gaming, ...dynamicsdynamicsp yp y g gg g
The new Ecopath with Ecosim (6)
• Four year project funded y p jthrough Lenfest Ocean Program
• Lenfest Ocean Futures Project:Lenfest Ocean Futures Project:– EwE6 released Sep 2007
– Ocean Summitsmethodology– Ocean Summits methodology 2009
5
• Customized versions facilitated
• User Ownership
• Customized versions facilitated
• User Ownership
New in EwE6• Object oriented implementation
• Usability improvements:• Usability improvements: – Navigation tree
I d i li ti t l– Improved visualization tools
• “More uncertainty”
• Plug‐ins
• Alternative interfaces
• Linkages
• Localization• Localization6
EwE6 overview
Scientific Interface Plug-ins
EwE Core
Ecopath Ecosim Ecospace
Data sourceData source
Source code (live) available for download through cvs 7
EwE6: Spatial modelingEwE6: Spatial modeling
• EwE5 short cuts for differentiation no longerEwE5 short cuts for differentiation no longer needed
• Ecospace now models multi stanza groups• Ecospace now models multi‐stanza groups properly
O i f IBM f l i• Option for IBM for multi‐stanza groups incorporated (with Kenny Rose)
• Threading to multiple processors
10
Gulf of M i FMC
Red drum SRAEcosimEcospace 2Ecospace IBM8
1012 Menhaden
150
200
250
Mexico FMC• Ecosim
02
46
050
100
Ecosim results similar to Red snapper
1012 Mullet
1.5
results from single species
d l (SRA)
24
68
00.
51.
0
model (SRA)
• Ecospace versions
0 0.0
Grouper1.2 Mackarel2.5
versions similar to Ecosim and
Grouper
0.4
0.6
0.8
1.0
Mackarel
1.0
1.5
2.0
2
Ecosim and SRA
Walters et al (MS)
0.0
0.2
1950 1960 1970 1980 1990 2000 2010
0.0
0.5
1950 1960 1970 1980 1990 2000 201013
Fitting to time series: Fitting to time series: learning from ecosystem historylearning from ecosystem history
• Numerous EwE models have in recent years produced credible fit to
• Numerous EwE models have in recent years produced credible fit torecent years produced credible fit to historical data, and made plausible
li di ti
recent years produced credible fit to historical data, and made plausible
li di tipolicy predictions– requires inclusion of environmental as
policy predictions– requires inclusion of environmental as well as fisheries impactwell as fisheries impact
14
Ecosystems where EwE models have been tested with historical trend data
Ecosystems where EwE models have been tested with historical trend datatested with historical trend data tested with historical trend data • E Bering Sea• Aleutian Islands• E Bering Sea• Aleutian Islands
• Bay of Quinte • Oneida Lake• Bay of Quinte • Oneida Lake• Aleutian Islands
• W&C GoAlaska• E GoAlaska
• Aleutian Islands• W&C GoAlaska• E GoAlaska
• Oneida Lake• Scotian Shelf• Chesapeake Bay
• Oneida Lake• Scotian Shelf• Chesapeake Bay
• W Vancouver Island• Hecate Strait
B i i h C l bi Sh lf
• W Vancouver Island• Hecate Strait
B i i h C l bi Sh lf
p y• Tampa Bay• US Gulf of Mexico
S B il Bi h
p y• Tampa Bay• US Gulf of Mexico
S B il Bi h• British Columbia Shelf• Strait of Georgia • NE Pacific
• British Columbia Shelf• Strait of Georgia • NE Pacific
• S Brazil Bight• Norwegian Sea• North Sea
• S Brazil Bight• Norwegian Sea• North SeaNE Pacific
• CN & ET Pacific • NWHI, Hawaii
NE Pacific• CN & ET Pacific • NWHI, Hawaii
North Sea • Baltic • S Benguela
North Sea • Baltic • S Benguela
15
• Gulf of California • Central Chile • Gulf of California • Central Chile
• Gulf of Thailand• South China Sea • Gulf of Thailand• South China Sea
Christensen & Walters 2005
Modeling process: fitting & driversFormal estimation
Fishing
Ecosystem model (predation, competition, mediation,
g
Predicted C, B, Z, W, diets
Log Likelihood
(Diet0)
(Z0)
Search for
minimi‐zationmediation,
age structured)
Climate Nutrientloading
ObservedC,B,Z,W, diets
( BCC/B0)
Habitat area
zation
Judgmental evaluation
Errorpattern
recognition
Choice of parametersto include in final
estimation (e.g., climate anomalies)
Judgmental evaluation
anomalies)
Christensen & Walters 2005
Are seals causing fishAre seals causing fishcausing fish declines in the Georgia
causing fish declines in the Georgia Strait?
I it fi hi ?
Strait?
I it fi hi ?Is it fishing?
Is it environ‐
Is it fishing?
Is it environ‐Is it environmental change?
Is it environmental change?g
Or, is it all
g
Or, is it all three?three?
1950 19502000 2000‘50 ‘50 ‘90‘90
17‘70 ‘70
PP anomaly and climate indicators
• Strait of Georgia– Race Rocks Salinity Feb‐May– Race Rocks Salinity Feb‐May
1.1
1.3
duct
ion
anom
aly
-20
-10
upwelling (m
3/10
PPA 10MJJ 10• BC shelf
– Upwelling, May‐July), 54°N
0.7
0.9
1950 1960 1970 1980 1990 2000
prod
0
10
00m/s)
p g y y
0 9
1.1
1.3
rodu
ctio
n an
omal
y
-0 5
-0.1
0.3
0.7
1.1
PDO
index
PPA 30AMJJ 30
• NE Pacific– PDO, April‐July
0.7
0.9
1950 1960 1970 1980 1990 2000
pr
-1.3
-0.9
0.5
Preikshot (UBC FCRR) 2007 18
What are models used for?What are models used for?• Lyne Morisette contacted registered EwE users:
325 models constructed or under construction
• Lyne Morisette contacted registered EwE users:
325 models constructed or under construction325 models constructed or under construction– 42% ecosystem structure;
325 models constructed or under construction– 42% ecosystem structure;
– 30% fisheries management;
11%
– 30% fisheries management;
11%– 11% theoretical ecology;
– 11% theoretical ecology;
– 6% protected area evaluations
– 6% protected area evaluationsevaluationsevaluations
19
Use of EwE for fisheries management• Evaluate impact of shrimp
trawling, GoCalifornia;• Evaluate impact of shrimp
trawling, GoCalifornia;• EIA of proposed fisheries
interventions, Bering Sea;
• Evaluate impact of bycatch, GoCalifornia;
• Shrimp bycatch issues Gulf of
• Evaluate impact of bycatch, GoCalifornia;
• Shrimp bycatch issues Gulf of
• EIA of alternative TAC’s, Bering Sea and GoAlaska;
• Target species response to• Shrimp bycatch issues, Gulf of Mexico FMC
• Evaluate impact of predators
• Shrimp bycatch issues, Gulf of Mexico FMC
• Evaluate impact of predators
• Target species response to TACs, Bering Sea
• Closed area sizing, Great Barrier on shrimp, GoMexico;
• Demonstrate ecological role of species GoMexico;
on shrimp, GoMexico;
• Demonstrate ecological role of species GoMexico;
Reef, Australia
• Valuation of cormorant impact, Ortobello Italyspecies, GoMexico;
• Impact of proposed fisheries interventions, Namibia
species, GoMexico;
• Impact of proposed fisheries interventions, Namibia
Ortobello, Italy
• Evaluation of cormorant impact, Ringkobing, Denmark
• South Africa pelagic fisheries• South Africa pelagic fisheries20
So why aren’t ecosystem models d f t?
So why aren’t ecosystem models d f t?used more for management?used more for management?
• Lack of experience using ecosystem models for• Lack of experience using ecosystem models for• Lack of experience using ecosystem models for predictive purposes;
E t d li i f t t i t
• Lack of experience using ecosystem models for predictive purposes;
E t d li i f t t i t• Ecosystem modeling is for strategic management, and supplements the tactical single species
t
• Ecosystem modeling is for strategic management, and supplements the tactical single species
tassessment;
• Fisheries management process is focused on
assessment;
• Fisheries management process is focused on tactical management;
• Strategic decisions are virtually non‐existing.
tactical management;
• Strategic decisions are virtually non‐existing.
21
Our empirical knowledge is limitedOur empirical knowledge is limitedp gp g
• Habitat and environmental changes (including• Habitat and environmental changes (including• Habitat and environmental changes (including those caused by fishing) and intensive fishery removals create novel situations
• Habitat and environmental changes (including those caused by fishing) and intensive fishery removals create novel situationsremovals create novel situations–We do not to understand the ‘mechanics’ of ecological response well enough to be able to
removals create novel situations–We do not to understand the ‘mechanics’ of ecological response well enough to be able toecological response well enough to be able to predict all important responses to novel situations;
ecological response well enough to be able to predict all important responses to novel situations;
–Make models one can play with;–Make models one can play with;
22
Current developmentsCurrent developments
• Make Your Own ModuleMake Your Own Module
• Model linkages …
23
Model linkages• ROMS/NPZ and climate models
NOAA GFDL/Princeton– NOAA GFDL/Princeton
– FLEM (2 layer hydrographic model)
ERSEM– ERSEM
• Tracking persistent pollutants & isotopes
• Spatial zoning (Ecoseed & MARXAN Bridge)
• Database‐driven model generation (GEF/LME)g ( / )
• Global scenario modeling (UNEP/OECD)
• Visualization and gaming: Ocean Summits• Visualization and gaming: Ocean Summits (Lenfest Ocean Futures Project) 24
Adaptive Management Procedure
M tO i d l
for Ocean Summits ‐ EwE
Managementprocedure
Operating model
Assessmentmodule
IndicatorsClimate model
module (ecological)(economical)
(social) StatusHydrographicmodel
Policy scenarios
Ecosystem / bio‐economic
model Visualization Strategy
The strategy is set by the participants; arrows indicate flow of information 25