Marine Biogeochemical and Ecosystem Modeling

1
Marine Biogeochemical and Ecosystem Modeling

• Michael Schulz
• MARUM -- Center for Marine Environmental Sciences
  
• and
• Faculty of Geosciences, University of Bremen

2

• 915 - 945
• 1. Introduction (Lecture)- The global carbon
  cycle, CO2 in seawater
• - Biological pumps
• - Reservoir or box models
• 2. Modeling Marine Nutrient and Carbon Cycles
  (Box-Model Exercise)- Global oceanic
  phosphate distribution
• - Nutrient productivity interactions
• - Oceanic carbon budget and large-scale ocean
  circulation
• 1045 - 1100 break

3

• 1100 1230
• 2. cont'd- Circulation-productivity feedback in
  the global ocean
• 3. State-of-the-art Biogeochemical Models
  (Lecture)- 2D and 3D Models- Included tracers
  and processes
• 4. Marine Ecosystem Models (Lecture)
• - Why ecosystem models?
• - Ecosystem models in paleoceanography

4
Course Material

• www.geo.uni-bremen.de/geomod/
• staff/mschulz/lehre/ECOLMAS_Modeling/
• This presentation
• Box-model exercises

5
Basic Literature

• Najjar, R. G., Marine biogeochemistry. in Climate
  system modeling, edited by Trenberth, K. E., pp.
  241-280, Cambridge University Press, Cambridge,
  1992.
• Rodhe, H., Modeling biogeochemical cycles. in
  Global biogeochemical cycles, edited by Butcher,
  S. S., R. J. Charlson, G. H. Orians and G. V.
  Wolfe, pp. 55-72, Academic Press, London, 1992.
• Sarmiento, J. L., and N. Gruber, Ocean
  biogeochemical dynamics, pp. 503, Princeton
  University Press, Princeton, 2006.
• Walker, J. C. G., Numerical adventures with
  geochemical cycles, 192 pp., Oxford University
  Press, New York, 1991.

6
For a climatologist biogeochemical cycles usually
translates into carbon cycle.
Ruddiman (2001)
7
Carbon-Cycle Characteristic Timescales
Reservoir Sizes in Gt C Fluxes in Gt C / yr
Sundquist (1993, Science)
8
Average surface- Water composition CO2 0.5
HCO3- 89.0 CO32- 10.5
Thurman Trujillo (2002)
9
Biological Productivity in the Ocean
Nutrients P, N, (Si, Fe)
Ruddiman (2001)
10
The Biological Pump
Atmosphere
CO
2
CO
2
Primary Production
Inorgan. C
Organ. C

Particle-Flux
Ocean
Remineralisation
Organ. C
CO

2
Sediments
Fig. courtesy of A. Körtzinger
11
Photic Zone
Aphotic Zone
Sediments
12
Biogenic Calcium Carbonate Production Raises
Dissolved CO2 Concentration
pH Reaction
(1) Biogenic carbonate uptake
(2) More bicarbonate dissociates
(3) More CO2 is formed
13
The Calcium Carbonate Pump
Atmosphere
CO
2
CO
2
Biogenic CaCO3
Formation
3
Lysocline
Ocean
CaCO3 Dissolution
CO
2-
3
Fig. courtesy of A. Körtzinger
14
Reservoir or Box Models

• Reservoir an amount of material defined by
  certain physical, chemical or biological
  characteristics that, under the particular
  consideration, can be regarded as homogeneous.
  (Examples CO2 in the atmosphere, Carbon in
  living organic matter in the oceanic surface
  layer)
• Flux the amount of material transferred from
  one reservoir to another per unit time

15
Single Reservoir Case
Reservoir (mass M)
Flux In
Flux Out
16
Basic Math of Box Models

• (Rate of change of mass in reservoir)
• (Flux in) (Flux out) Sources Sinks
• Or, for concentration (C mol/m3) and water flux
  (Q m3/s)

17
Numerical Solution of Box-Model Equations
Solution by finite-difference method
(approximation!) Euler Method
Initial Condition
Time (in steps of Dt)
18
Numerical Solution of Box-Model Equations
Solution by finite-difference method
(approximation!)
Initial Condition
19
Numerical Solution of Box-Model Equations
Solution by finite-difference method
(approximation!) Euler Method
Initial Condition
Time (in steps of Dt)
20
Euler Method

M
M(tn1)


Prediction

Slope Fi(tn) - Fo(tn) SMS(tn)
Error



True Value

M(tn)


Dt


tn

tn1


t


Assumption Slope at time tn remains constant
throughout time interval Dt
21
Coupled Reservoirs
F12
Reservoir 1 (mass M1)
Reservoir 2 (mass M2)
F21
Principle of mass-conservation requires M1 M2
const.
22
Large-Scale Ocean Circulation
(after Broecker, 1991)
23
(No Transcript)
24
Box-Model ofOceanic PO4 Distribution
Atlantic
Indo-Pacific
Southern Ocean
Surface (0-100 m)
AABW_P (20 Sv)
NADW (10 Sv)
20 Sv
20 Sv
10 Sv
Deep (gt 100 m)
AABW_A (4 Sv)
25
www.geo.uni-bremen.de/geomod/staff/mschulz/lehre/E
COLMAS_Modeling/ bm1_po4_only.gsp
26
Box-Model Experiment 1

• Vary the water transports and initial PO4
  concentration and observe the final PO4
  concentration and evolution (time series).
• Q1 How does the final PO4 distribution depend on
  these settings?
• Q2 How do these settings affect the time it
  takes to reach a steady state? (What
  characterizes the steady state?)

27
Inducing PO4 Gradients Biological Productivity

• Assume an average export production of
• 12 g C/m2/yr
• With a Redfield ratio of CP 1171 (molar
  ratio) and 1 mol C 12 g C
• ? Corresponding biological PO4 fixation is
• 1/117 mol P/m2/yr

28
Box-Model of Oceanic PO4 Distribution with
Productivity
Indo-Pacific
Southern Ocean
Atlantic
Surface (0-100 m)
AABW_P (20 Sv)
NADW (10 Sv)
Deep (gt 100 m)
AABW_A (4 Sv)
Assumption Biologically fixed PO4 sinks from the
surface layer to the underlying deep layer, where
the organic material is completely remineralized.
29
www.geo.uni-bremen.de/geomod/staff/mschulz/lehre/E
COLMAS_Modeling/ bm1_po4_fix_prod.gsp
30
Box-Model Experiment 2

• Q How does the inclusion of biological
  productivity affect the PO4-concentration
  difference between Atlantic and Indo-Pacific
  Oceans in the standard case?

31
10 m water depth
32
1750 m water depth
33
Box-Model Experiment 2

• Vary the water transports (try max. and small
  values) and observe how the PO4 distribution
  changes. Explain the changes.
• Q What happens if NADW 0 Sv? (Keep the
  remaining parameters at their default values.)
  Does this result make sense in the real world?
• Q For which initial PO4 concentration do no
  negative concentrations result (with NADW 0
  Sv)? Is this a reasonable increase for Late
  Pleistocene glacials?

34
Avoiding Negative PO4 Concentrations
Nutrient-Dependent Productivity

• Assume that productivity scales with the PO4
  availability in the surface layer (variety of
  relationships are possible linear, non-linear
  with saturation)
• PO4 fixation PO4sfc Volsfc / t mol/yr,
• where t is the residence time of PO4 in the
  surface due to biological productivity
• Assume tATL tIPAC 5 yr and tSOC 50 yr
  (Broecker and Peng, 1986)

35
www.geo.uni-bremen.de/geomod/staff/mschulz/lehre/E
COLMAS_Modeling/ bm1_po4_dyn_prod.gsp
36
Box-Model Experiment 3a

• Run the model for NADW of 0 and 10 Sv and write
  down the PO4 concentrations for the Atlantic
  boxes for each case.
• Calculate the difference between conc. in deep
  and surface box. What do you observe?

37
Box-Model Experiment 3a Atlantic
NADW (Sv) PO4 Surface (mmol/l) PO4 Deep (mmol/l) DPO4 (mmol/l)
10 0.24 0.69 0.45
0 0.18 0.88 0.70
Shift of PO4 content from surface to deep
Atlantic as NADW drops
38
Box-Model Experiment 3b

• Run the model for NADW 0, 5, 10, 15, 20 Sv
  and write down the final PO4 fixation in the
  Atlantic Ocean.
• Sketch NADW vs. PO4 fixation.
• QWhat is the paleoceanographic implication of
  this finding?

39
NADW and Productivity in the Atlantic Ocean
3.6
3.4
3.2
3
2.8
PO4 Fixation 1011 mol P /yr
2.6
2.4
2.2
2
0
5
10
15
20
NADW Flow Sv
40
Including the Marine Carbon-Cycle

• Tracers PO4 (? controls productivity)
• DIC (dissolved inorganic carbon)
• ALK (alkalinity)
• Aqueous CO2 partial pressure f(DIC, ALK)
• Redfield ratio of organic matter (CNP
  117161)
• Ratio between Corg and CaCO3 production (rain
  ratio) ? assumed to be temperature dependent (a
  crude parameterization of ecosystem dynamics)

41
Rain-Ratio Parameterization
42
(No Transcript)
43
Area-Weighted Average
Atmospheric pCO2 Mean Oceanic pCO2
www.geo.uni-bremen.de/geomod/staff/mschulz/lehre/E
COLMAS_Modeling/ bm1_c-cycle_fix_prod.gsp
44
Box-Model Experiment 4C-Cycle with Fixed
Productivity

• Run the model for the default setting. Identify
  the sources and sinks with respect to atmospheric
  CO2.
• Run the model for NADW of 0 and 10 Sv. Write down
  the final global mean pCO2 and the productivity
  in the Atlantic Ocean. (Neglect the negative PO4
  conc., identified in the previous exp.)

45
Box-Model Experiment 4C-Cycle with Fixed
Productivity
NADW (Sv) Prod. ATL (Pg C/yr) Prod. Glob. (Pg C/yr) Global pCO2 (ppm)
10 0.447 5.03 281
0 0.447 5.03 265
16 ppm Reduction
46
Box-Model Experiment 5C-Cycle with Dynamic
Productivity

• How will the response of the mean pCO2 change if
  productivity is no longer constant but a function
  of PO4?

47
www.geo.uni-bremen.de/geomod/staff/mschulz/lehre/E
COLMAS_Modeling/ bm1_c-cycle_dyn_prod.gsp
48
Box-Model Experiment 5C-Cycle with Dynamic
Productivity

• Run the model for again for NADW of 0 and 10 Sv.
  Write down the final global mean pCO2 and the
  productivity in the Atlantic Ocean.
• Interpret your results.

49
Box-Model Experiment 5C-Cycle with Dynamic
Productivity
NADW (Sv) Prod. ATL (Pg C/yr) Prod. Glob. (Pg C/yr) Global pCO2 (ppm)
10 0.447 5.03 281
0 0.350 4.83 275
Only 6 ppm Reduction
50
Box-Model Experiment 5C-Cycle with Dynamic
Productivity
NADW 0 ? DIC shifted from surface to deep
Atlantic ? pCO2 reduced BUT PO4 is shifted to
deep ocean too ? less nutrients in surface ?
productivity decreases ? biological pump weakens
? pCO2 increases ? Negative Feedback Mechanism
51
From Box-Models to 2D/3D-Models
Ruddiman (2001)
52
Structure of a Global Biogeo-chemical Model
Ridgwell (2001, Thesis)
53
Modeling Deep-Sea Sediments
Ridgwell (2001, Thesis)
54
Phosphate in the Atlantic Ocean mmol/l
2D-Model (Zonal Mean)
(Schulz and Paul, 2004)
3D-Model (N-S Section)
(Heinze et al., 1999)
55
Horizontal Resolution in a 2D-Biogeochemical Model
(Schulz and Paul, 2004)
56
Horizontal Resolution in a 3D-Biogeochemical Model
(Heinze et al., 1999)
57
A Modeled Sediment Stack in the North Atlantic
Heinze, C. et al., 1999 A global oceanic
sediment model for long-term climate studies.
Global Biogeochemical Cycles, 13, 221-250.
58
Modeled and Observed Modern CaCO3 Content of
Deep-Sea Sediments
Model
Observations
? Even the most sophisticated biogeochemical
models allow only for a crude approximation of
the real world. Discrepancies are largely due to
an inadequate resolution (e.g. MOR) and a lack of
knowledge of the processes being involved.
Heinze et al. (1999)
59
Marine Ecosystem Models Why?

• Productivity may depend on more than a single
  nutrient (N, P, Si, Fe)
• Export production controlled by ecosystem
  dynamics
• Understanding the preferential growth of
  different algae groups (e.g. diatoms vs.
  coccolithophores)
• Disentangling the seasonal imprint in biological
  proxy records

60
NPZD-Type Ecosystem Model

• 4 Compartments
• Coupled to carbon and alkalinity
• Nutrients are transported by ocean circulation
• Efficient in predicting seasonal patterns

(after Fasham et al., 1990)
61
Marine Ecosystem Model Components (Moore et al.,
2002)
62
Marine Ecosystem Model Forcing
Output from global OGCM
63
Global Foraminifera Model
Fraile et al. (subm.)
64
Fraile et al. (subm.)
65
Modeled / Observerd Distribution of N. pachyderma
(sin.)
Fraile et al. (subm.)
Brown University Foraminiferal Database (Prell et
al., 1999)
66
Modeled / Observerd Distribution of N. pachyderma
(dex.)
Fraile et al. (subm.)
Brown University Foraminiferal Database (Prell et
al., 1999)
67
Modeled / Observerd Distribution of G. bulloides
Fraile et al. (subm.)
Brown University Foraminiferal Database (Prell et
al., 1999)
68
Modeled / Observerd Distribution of G. ruber
(white)
Fraile et al. (subm.)
Brown University Foraminiferal Database (Prell et
al., 1999)
69
Modeled / Observerd Distribution of G. sacculifer
Fraile et al. (subm.)
Brown University Foraminiferal Database (Prell et
al., 1999)
70
Modeled LGM shift in seasonality of G. bulloides
Fraile et al. (subm.)
71
Benefits of Paleoecosystem Modeling

• To facilitate model-data comparison
• To obtain a mechanistic understanding of
  reconstructed shifts in species
• To assess the potential effect of altered
  plankton successions on proxy reconstructions
  based on organisms