Modelling the export of biogenic particulates from upper ocean Philip Boyd
Jan 15, 2016
Modelling the export of biogenic particulates from upper ocean
Philip Boyd
Behrenfeld(OSU)
Outline Factors impacting export – selected models
• NPP and export (Suess, 1980)
• J100 (Martin et al., 1987)
• Algal cells and foodweb structure (Michaels & Silver, 1988); Boyd & Newton (1995)
• NPP and temperature – Laws et al. 2000 • Ballasting agents (Armstrong et al. 2001) • Mechanistic models – (Dunne et al., 2005)• Summary
FACTORS CONTROLLING EXPORTPresent status
Primary ProductionBallasting agentsAlgal cells – large versus smallParticle transformations – aggregationFoodweb structure – different grazersMicrobial processes - solubilisation
Boyd and Trull (2007)
Case study 1 – Suess (1980)a direct relationship between NPP, depth and export
Case study 1 – Suess (1980)a direct relationship between NPP, depth and export
From Bishop (1989)
From Bishop (1989)
Case study 2 – replacing NPP with J100 (Martin et al., 1987)
What does J100 represent?Why is it a better predictor of export?
What does J100 represent?Why is it a better predictor of export?
Case study 3 – Michaels and Silver (1988) - what sets J100?
Different foodweb structures result in
A range of export efficiencies (pe ratio)
pe ratio = particle export/NPP
Using Michaels & Silver - Comparison of NE Atlantic spring bloom signatures (Boyd & Newton, 1995)
• 1989• 2.7 µg chla L-1
• 16.1 g C m-2 NPP
• -27 mmol NO3 m-2
• 32.0 µmol kg-1 tCO2
• 80% diatoms• Microzoo grazing
• 1990• 3.6 µg chla L-1
• 14.7 g C m-2 NPP
• -33.5 mmol NO3 m-2
• -33.5 µmol kg-1 tCO2
• 70% diatoms• Microzoo grazing
Using Michaels & Silver - Comparison of NE Atlantic spring bloom signatures (Boyd & Newton, 1995)
• 1989• 2.7 µg chla L-1
• 16.1 g C m-2 NPP• -27 mmol NO3 m-2
• 32.0 µmol kg-1 tCO2
• 80% diatoms• Microzoo grazing• 720 mg POC m-2
export (3100 m)
• 1990• 3.6 µg chla L-1
• 14.7 g C m-2 NPP• -33.5 mmol NO3 m-2
• -33.5 µmol kg-1 tCO2
• 70% diatoms• Microzoo grazing• 410 mg POC m-2
export (3100 m)
Observed versus predicted POC export(mg C m-2 d-1)
• Predicted• 16.6 (Suess) 15.1• 41.8 (Betzer) 36.7• 19.2 (Berger) 17.5• 4.4 (Pace) 4.0• 9.5 (BN – Martin) 3.8
• Observed• 9.6
• 4.0
(1989 – black; 1990 – red)
Case study 4 Laws et al. (2000)Temperature effects on export fluxes
Calculated ef ratios (export/NPP) as a function of NPP and temperature
Nutrients
Inorganic nutrients
DOM
Small PP
Large PP Filter feeder
Bacteria
Flagellates Ciliates
Carnivore
Detrital POC
Export
Modelled ef ratio
0 0.6
0.4
0.7
Ross Sea *
Polyna *
NABE *
OSP * * Peru-normal
* Peru El Nino
* Arabian
HOT *
* BATS*EqPac-EN
* EqPac
NPP (mg N m-2 d-1)
0 500 1000
0.4
0.7
Obs.efratio
Ross Sea *
* Polyna
NABE *
* OSP Peru-normal *
Peru El Nino *
* HOT
* BATS*EqPac-EN
* Arabian* EqPac
Combining ef ratio with satellite NPP and SST – global export is 20% of NPP
Case study 5 – export and ballast – Armstrong et al. (2002)
Ballast revisited
0 60
AA
CARBLIPID
Plankton - EqPac
Export 1000 m
Export 3500 m
Weight %
Hedges et al. 2001
Non-selective preservation within theInorganic matrix of biominerals
The mineral matrix
8 nm
S
S
S
5
5
5
POC flux
EqPac
S
S
S5
5
5
Depth (km)
0
5
0
5
Fraction OC by weight
Fluxes normalised to mass flux (OC/M) are much less variable than POCfluxes alone
POC export here is based on quantitative association of POC with ballast minerals
Martin curve
POC flux
Protected POC
“Using ballast mineral data markedly increases the ability to predict organic carbon fluxes”
Dashed line = excess POC fluxi.e. POC not associated with ballast minerals
Case study 6 Dunne et al. (2005)Empirical and mechanistic models for the pe ratio
A synthesis of global field observations related to ecosystem size structure, NPP and particle export was used for model validation
Large phytoplankton augment small ones as production or biomass increases.
In this model, variability in NPP results in a biomass-modulatedswitch between small and large phytoplankton pathways
The empirical model captures 61% of the observed variance in the pe ratio of particle export using SST and chlorophyll concentrations (or NPP) as predictor variables.
The empirical model captures 61% of the observed variance in the ratio of particle export to primaryproduction (the pe ratio) using sea-surface temperature and chlorophyll concentrations(or primary productivity) as predictor variables.
Model NPP(Gt C yr-1)
Surface Export (Gt C yr-1)
Nutrient Inversion (of P, O2,
DIC, etc) Schlitzer (2002)na 9.6
Semi-prognostic (Temperature and e ratio) Laws et al. (2000)
52.1 12.9
Semi-prognostic (e ratio, NPP and SST – scaled using remote sensing data*) Laws et al. (2000)
52.1 11.1
Coupled Ocean Atmosphere Model (COAM) -LG (OPAICE – LMDS) Bopp et al. (2001)
na 13.1
COAM - LB (OPAICE – LMDS)! 64.7 11.1COAM - AG (OPAICE-ARPEGE)! Bopp et al. (2001)
na 9.5
Prognostic (COAM (NCAR) and offline ecosystem model) Moore et al. (2002)
45.2 7.9@
Prognostic (COAM (NCAR) and offline ecosystem model) Moore et al. (2002)
45.2 12.0$
13 models in OCMIP-II Doney et al (2004)
n.a. +40% range
Bo
Boyd and Trull (2007)
Similarity of global export estimates despite the diversity of approaches.
Either the problem has a relatively unique solution, or all models are making similar approximations.
No models have yet included sufficient complexity to capture the observed variability of export fluxes.
Determining which additional factors, beyond those of temperature, chlorophyll and NPP, are, most critical is a high priority task.
SUMMARY(Boyd & Trull)
Observed versus predicted POC export (% error of fit –
((100*(predicted-observed)/(observed))
BATS -234
BN
910 Suess
979 Berger
-406 Pace
HOT 71 597 932 -265
NABE -25 67 96 -56
ST Atlantic
-57 0 -5 -77
T Atlantic
10 666 654 15
Papa -20 391 477 31