developing*small*pelagic*fish*deb** … · takeihome*messages* •...
TRANSCRIPT
Developing Small Pelagic Fish DEB (Dynamic Energy Budget) modules for IBMs
in spa?al popula?on dynamics studies
Why? How? Issues/Challenges?
Laure Pecquerie
W5 -‐ Victoria – 11 March 2017
Take-‐home messages
• Current situa+on: not sufficient data types to es?mate primary parameters of a DEB model for small pelagic fish. Several datasets provide similar fits
• Taking into account phylogeny in DEB parameter es?ma?on for small pelagic fish will reduce uncertainty in parameter es?mates. Mul+-‐species es+ma+on procedure under development – We can only tackle this problem as a community (mul?-‐species, mul?-‐popula?on approach)
• More data types by designing experimental protocols together – We can only tackle this problem as a community (experimenters, fielders and DEB modellers )
• Otoliths are key to improve and validate IBMs (…un?l we find a way to tag these fish) à Otolith data <-‐> Otolith model in IBMs
Take-‐home messages
• Current situa+on: not sufficient data types to es?mate primary parameters of a DEB model for small pelagic fish. Several datasets provide similar fits
• Taking into account phylogeny in DEB parameter es?ma?on for small pelagic fish will reduce uncertainty in parameter es?mates. Mul+-‐species es+ma+on procedure under development – We can only tackle this problem as a community (mul?-‐species, mul?-‐popula?on approach)
• More data types by designing experimental protocols together – We can only tackle this problem as a community (experimenters, fielders and DEB modellers )
• Otoliths are key to improve and validate IBMs (…un?l we find a way to tag these fish) à Otolith data <-‐> Otolith model in IBMs
Take-‐home messages
• Current situa+on: not sufficient data types to es?mate primary parameters of a DEB model for small pelagic fish. Several datasets provide similar fits
• Taking into account phylogeny in DEB parameter es?ma?on for small pelagic fish will reduce uncertainty in parameter es?mates. Mul+-‐species es+ma+on procedure under development – We can only tackle this problem as a community (mul?-‐species, mul?-‐popula?on approach)
• More data types by designing experimental protocols together – We can only tackle this problem as a community (experimenters and DEB modellers )
• Otoliths are key to improve and validate IBMs (…un?l we find a way to tag these fish) à Otolith data <-‐> Otolith model in IBMs
Take-‐home messages
• Current situa+on: not sufficient data types to es?mate primary parameters of a DEB model for small pelagic fish. Several datasets provide similar fits
• Taking into account phylogeny in DEB parameter es?ma?on for small pelagic fish will reduce uncertainty in parameter es?mates. Mul+-‐species es+ma+on procedure under development – We can only tackle this problem as a community (mul?-‐species, mul?-‐popula?on approach)
• More data types by designing experimental protocols together – We can only tackle this problem as a community (experimenters, fielders and DEB modellers )
• Otoliths are key to improve and validate IBMs (…un?l we find a way to tag these fish) à Otolith data <-‐> Otolith model in IBMs
Why using DEB models in IBMs
• Full life cycle model, no discon?nuity, (probably) less parameters
• Growth, Reproduc?on AND Development as a func?on of food and temperature condi?ons
• + other forcing func?ons and tracers: contaminant, parasites, δ15N, O2, …
• Individual parameters -‐> mul?-‐genera?on simula?ons, selec?on, eco-‐evo processes
Issues / Challenges?
We won’t be able to include all the processes we believe are at play
• Model structure depends on the research ques+on Ecotoxicology, Aquaculture, Ecology ques?ons may lead to different models within the same framework
à Spa?al popula?on dynamics – bioenerge?c module for IBM module – we need
the simplest model possible Do we need : • Spawning dates according to environmental condi?ons (seasonal / interannual
variability) ? • Age at which swim bladder is func?oning? (ver?cal migra?on) • Diges?on processes or constant assimila?on efficiency? Spa+al / +me scales of these processes vs Spa+al grid / +me step of hydro-‐biogeochemical models
A bit of history
• Bas Kooijman, in 1979 • Two ques?ons :
– How can we quan?fy the effect of toxic compounds on Daphnia reproduc?on?
– Which effect has a small decrease in individual reproduc?on on the dynamics of the popula?on?
…
Observa?ons : Asympto?c growth
Kooijman, 2010
maturity
1-κ maturity
maintenance
development
food faeces assimilation
reserve
structure
somatic maintenance
growth
κ
Life events in a standard DEB model
reproduction buffer
reproduction
maturity
1-κmaturity
maintenance
development
food faecesassimilation
reserve
structurestructure
somaticmaintenance
growth
κ
reproductionbuffer
reproduction
maturity
1-κmaturity
maintenance
development
maturitymaturity
1-κmaturity
maintenance
development
1-κmaturity
maintenance
development
food faecesassimilation
food faecesassimilation
reserve
structurestructure
somaticmaintenance
growth
κ
reproductionbuffer
reproduction
reproductionbuffer
reproduction
Model simula?ons
INPUTS DEB MODEL
Food density
Temperature
Flow
Weight
Fecundity / egg size
OUTPUTS
Length
• Oxygen consump?on – Empirical à R = aLb
Theore?cal à R = aL2 + bL3 (Kooijman, 2000)
Modelling approach
Daphnia pulex d’après Kooijman (2000)
R = aLb = 0.0516 L2.437 R = aL2 + bL3 = 0.0336 L2 + 0.01845 L3
Issues / Challenges?
• State variables do not correspond to observables/data
• DEB and tradi?onal bioenerge?c models have different emphases / processes, thus different data needs for calibra?on (but ways to link them)
• Specific to SPF: we hardly observe feeding and full reproduc?on investment (batch spawner / vitellogenesis de novo)
à Several parameter sets provide similar fits
Issues / Challenges?
• State variables do not correspond to observables/data
• DEB and tradi?onal bioenerge?c models have different emphases / processes, thus different data needs for calibra?on (but ways to link them)
• Specific to SPF: we hardly observe feeding and full reproduc?on investment (batch spawner / vitellogenesis de novo)
à Several parameter sets provide similar fits
1-κmaturity
maintenance
maturityoffspring
maturationreproduction
food faecesassimilation
reserve
structurestructure
somaticmaintenance
growth
κ 1-κmaturity
maintenance
maturityoffspring
maturationreproduction
1-κmaturity
maintenance
maturityoffspring
maturationreproduction
1-κmaturity
maintenance
maturityoffspring
maturationreproduction
food faecesassimilation
reserve
food faecesassimilation
reserve
structurestructure
somaticmaintenance
growth
κ
structurestructure
somaticmaintenance
growth
κ
FOOD ?
?
metabolic products in DEB models
Reserve
Structure
κ κ−1
Assimilation pA
Growth pG
Somaticmaintenance pM
Maturitymaintenance pJ
Reproduction pR
Food
pC
Reproductionbuffer
Assimilation Products
Dissipation Products
Growth Products
pDpA
pA
Faeces
CO2
pD pG
pG
(b)
(c)
(d)
(a)
Reserve
Structure
κ κ−1
Assimilation pA
Growth pG
Somaticmaintenance pM
Maturitymaintenance pJ
Reproduction pR
Food
pC
Reproductionbuffer
Assimilation Products
Dissipation Products
Growth Products
pDpA
pA
Faeces
CO2
pD pG
pG
(b)
(c)
(d)
(a)
Biocarbonate = metabolic product
Reserve
Structure
κ κ−1
Assimilation pA
Growth pG
Somaticmaintenance pM
Maturitymaintenance pJ
Reproduction pR
Food
pC
Reproductionbuffer
Assimilation Products
Dissipation Products
Growth Products
pDpA
pA
Faeces
CO2
pD pG
pG
(b)
(c)
(d)
(a)
Reserve
Structure
κ κ−1
Assimilation pA
Growth pG
Somaticmaintenance pM
Maturitymaintenance pJ
Reproduction pR
Food
pC
Reproductionbuffer
Assimilation Products
Dissipation Products
Growth Products
pDpA
pA
Faeces
CO2
pD pG
pG
(b)
(c)
(d)
(a)
How to go beyond?
• Taking into account phylogeny in DEB parameter es?ma?on for small pelagic fish will reduce uncertainty in parameter es?mates.
• Mul+-‐spcies es+ma+on procedure under development
Standard DEB model Engraulis encrasicolus
DEB model Bay of Biscay
Assump?ons
Parameter values Ini?al condi?ons
Environmental condi?ons
Fish Data
Classical approach
New approach
Standard DEB model
Small pelagic Fish DEB model
Engraulidae DEB model
Phylogen
y
Engraulis encrasicolus DEB model
Bay of Biscay Gulf of Lion Benguela
Assump?ons Parameter values
Environmental condi?ons
Clupeidae DEB model
Sardina pilchardus DEB model
Bay of Biscay Gulf of Lion
Sardinops sagax DEB model
Benguela
Amp Collec?on : currently 600 entries
hjp://www.bio.vu.nl/thb/deb/deblab/add_my_pet/species_list.html
Small pelagic fish entries : 6
Atlan?c herring entry
Atlan?c herring entry
How to go beyond?
• Taking into account phylogeny in DEB parameter es?ma?on for small pelagic fish will reduce uncertainty in parameter es?mates.
• Mul+-‐spcies es+ma+on procedure under development
• Modelling otolith growth, opacity, δ18O, δ13C in IBMs
Why modelling otolith forma+on?
Biocarbonates
DEB model Habitat
Temperature Growth
A) Calibra+on of proxies
Understand
B) Life history reconstruc+on
Improve
C) Environmental scenario studies (2D, 3D models)
Predict
(Phil. Trans. Roy. Soc. 2010 PLoS ONE 2011, MEPS 2012)
VIRTUAL LAB
Time
δ13 C, δ
18O
Food (Quan+ty, δ13C,…)
δ13C DIC δ18O w
Reproduc+on
Metabolism
è To embrace and take advantage of these mul+-‐factor interac+ons to improve our interpreta+on and use of these unique archives
Otolith modeling : individual history
Environment
Otolith growth and opacity
Length DEB model
(Pecquerie et al. 2012)