ALL DEGLACIAL MODEL-DATA COMPARISONS ARE WRONG

SOME MAY BE USEFUL

Andy RidgwellUniversity of Bristol / University of California, Riverside

0

Age (Ma)

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

10 20 30 40 50 60

18

d0 (

‰)

Zachos et al. [2001, 2008]

Outline of talkQuantifying ‘time’ in

models and data

0 50 100 150 200 2504000

4500

5000

5500

6000

6500

Age (kyr BP)

Depth

(m

)Outline of talk

Quantifying ‘time’ inmodels and data

One (or more) of the following:

This is not relevant at all.

Too late – this was relevant 15 years ago.

There are potentially important implications for bulk carbonate and low sedimentation rate records. But no-one in their right mind uses these any more.

There are important implications for data-data (wiggle matching) and model-data analysis.

Meh

­

­

­

­

­

­ There are important questions raised of where in the sediments, and what fraction, of carbonate dissolves.

forget about itdrink beerbe happy

keep going

Quantifying ‘time’ inmodels and dataAnticipated outcome of talk

meh

0

Age (Ma)

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

10 20 30 40 50 60

18

d0 (

‰)

Zachos et al. [2001, 2008]

Quantifying ‘time’ inmodels and dataA deep time perspective on shallow time, time

open oceansea ice coastal seas

terrestrialbiota

atmosphere

marine biota

sed

imen

ts

soils

land surface and rock weathering

icesheet

2D energy-moisture balance(no clouds, dynamics)

fully 3D (‘reducedphysics’)

ocean

surf

ace layer

bioturbatedzone of 1 cm

sedimentstack layers

partially-filledupper-most

layer

bio

turb

ati

on

al

mix

ing

non-bioturbated

zone ofburied layers

simplifiedthermo-dynamic

cGENIE

www.seao2.info/mycgenie.html

-0.5

Age relative to the PETM (Ma)0.5 1.0 1.5-1.0 0.0-1.5-2.0

4.0

3.0

2.0

1.0

0.0

13

dC

(‰)

PE

TM

(ET

M1)

Age model artifacts -- lessons from deeper-time (1)Quantifying ‘time’ in

models and data

Consider: An event characterized by a (severe) reduction in carbonate preservation

Age model artifacts -- lessons from deeper-time (1)Quantifying ‘time’ in

models and data

CaCO (wt%)3

13d C (‰)(CaCO3)

Depth

abo

ve r

efe

rence

leve

l (cm

)

0 50 100-70

-60

-40

-30

-20

-10

0

-50

0.0 4.52.25

?(model-data divide)

pCO (matm)2

low high

num

erica

l model t

ime-s

eries

or

ext

ern

al (

to o

cean)

reco

rd (

e.g

. ic

e c

ore

)

‘lysocline’

‘carbonatecompensation

depth’

(reefs)

marine biota

oceanic crust

de

ep

-se

as

ed

ime

nts

surfaceoceanlayer

Ca

CO

ra

in f

lux

3

dis

so

lutio

n f

lux

de

trita

l ra

in f

lux

bioturbatedzone of

sedimentstack

layers

stack top

bio

turb

ati

on

al

mix

ing

non-bioturbated

zone ofsediment

stacklayers

Age model artifacts -- lessons from deeper-time (1)Quantifying ‘time’ in

models and data

current (model) time(to machine precision)

18d OI/Ca environmental

proxies

dissolution(of all tracers,including time,proportionally)

Age model artifacts -- lessons from deeper-time (1)Quantifying ‘time’ in

models and data

CaCO (wt%)3

13d C (‰)(CaCO3)

Depth

abo

ve r

efe

rence

leve

l (cm

)

0 50 100-70

-60

-40

-30

-20

-10

0

-50

0.0 4.52.25 -85 0 0.0 2.01.0

pCO (matm)2

low high

num

erica

l model t

ime-s

eries

or

ext

ern

al (

to o

cean)

reco

rd (

e.g

. ic

e c

ore

)

Age (kyr) Sed rate-1(cm kyr )

-70

-60

-40

-30

-20

-10

0

-50

Age model artifacts -- lessons from deeper-time (1)Quantifying ‘time’ in

models and data

CaCO (wt%)3

13d C (‰)(CaCO3)

0 50 100 0.0 4.52.25 0.0 2.01.0

pCO (matm)2

low high

num

erica

l model t

ime-s

eries

or

ext

ern

al (

to o

cean)

reco

rd (

e.g

. ic

e c

ore

)

Depth (cm) Sed rate-1(cm kyr )

Mean C

aC

O a

ge r

ela

tive to r

efe

rence

leve

l (ky

r)3

20

10

-10

-20

-30

-40

-50

0

-70

-60

-40

-30

-20

-10

0

-50

Age model artifacts -- lessons from deeper-time (1)Quantifying ‘time’ in

models and data

CaCO (wt%)3

13d C (‰)(CaCO3)

0 50 100 0.0 4.52.25 0.0 2.01.0

pCO (matm)2

low high

Depth (cm) Sed rate-1(cm kyr )

Mean C

aC

O a

ge r

ela

tive to r

efe

rence

leve

l (ky

r)3

20

10

-10

-20

-30

-40

-50

0

20

10

-10

-20

-30

-40

-50

0

low high

13d C (‰)(pCO2)

-70

-60

-40

-30

-20

-10

0

-50

CaCO (wt%)3

0.15

0.10

0.05

0.0

-0.05

20 40 60 80 1000

Mo

de

l-g

en

era

ted

syn

the

tic

se

dim

en

t c

ore

re

spo

nse

[R

idg

we

ll, 2

00

7]

54.75

54.80

54.85

54.90

54.95

55.00

CaCO (wt%)3

55.05

Age (M

a)

1500

m

3600 m

2600 m

20 40 60 80 1000

1266C1265A1263C/D

1262A1267B

Bu

lk s

ed

ime

nt

wt%

Ca

CO

co

nte

nt

[Za

ch

os

et

al.,

20

05

]3

Quantifying ‘time’ inmodels and dataAge model artifacts -- lessons from deeper-time (1)

CaCO (wt%)3

0.15

0.10

0.05

0.0

-0.05

20 40 60 80 1000

Bu

lk s

ed

ime

nt

wt%

Ca

CO

co

nte

nt

[Za

ch

os

et

al.,

20

05

]3

54.75

54.80

54.85

54.90

54.95

55.00

CaCO (wt%)3

55.05

Age (M

a)

1500

m

3600 m

2600 m

20 40 60 80 1000

1266C1265A1263C/D

1262A1267B

CO perturbation22000 PgC

Mo

de

l-g

en

era

ted

syn

the

tic

se

dim

en

t c

ore

re

spo

nse

[R

idg

we

ll, 2

00

7]

Quantifying ‘time’ inmodels and dataAge model artifacts -- lessons from deeper-time (1)

CaCO (wt%)3

0.15

0.10

0.05

0.0

-0.05

20 40 60 80 1000

54.75

54.80

54.85

54.90

54.95

55.00

CaCO (wt%)3

55.05

Age (M

a)

1500

m

3600 m

2600 m

20 40 60 80 1000

1266C1265A1263C/D

1262A1267B

Bu

lk s

ed

ime

nt

wt%

Ca

CO

co

nte

nt

[Za

ch

os

et

al.,

20

05

]3

4000 PgC CO perturbation2

Mo

de

l-g

en

era

ted

syn

the

tic

se

dim

en

t c

ore

re

spo

nse

[R

idg

we

ll, 2

00

7]

Quantifying ‘time’ inmodels and dataAge model artifacts -- lessons from deeper-time (1)

CaCO (wt%)3

0.15

0.10

0.05

0.0

-0.05

20 40 60 80 1000

54.75

54.80

54.85

54.90

54.95

55.00

CaCO (wt%)3

55.05

Age (M

a)

1500

m

3600 m

2600 m

20 40 60 80 1000

1266C1265A1263C/D

1262A1267B

Bu

lk s

ed

ime

nt

wt%

Ca

CO

co

nte

nt

[Za

ch

os

et

al.,

20

05

]3

6000 PgC CO perturbation2

Mo

de

l-g

en

era

ted

syn

the

tic

se

dim

en

t c

ore

re

spo

nse

[R

idg

we

ll, 2

00

7]

Quantifying ‘time’ inmodels and dataAge model artifacts -- lessons from deeper-time (1)

-0.5

Age relative to the PETM (Ma)0.5 1.0 1.5-1.0 0.0-1.5-2.0

4.0

3.0

2.0

1.0

0.0

13

dC

(‰)

ELM

O(E

TM

2)

Age model artifact lessons from deeper-time (2)Quantifying ‘time’ in

models and data

Consider: An event characterized by a (mild) reduction in carbonate preservation

@elm

o

-0.5

Age relative to the PETM (Ma)0.5 1.0 1.5-1.0 0.0-1.5-2.0

4.0

3.0

2.0

1.0

0.0

13

dC

(‰)

ELM

O(E

TM

2)

0 50 100−20

0

20

40

60

80

100

wt%

Tim

e a

fte

r E

TM

2 o

nse

t (k

yr)

δ13C

-0.15 1.55

‰

Age model artifact lessons from deeper-time (2)Quantifying ‘time’ in

models and data

1262 (3500 m)

Quantifying ‘time’ inmodels and dataAge model artifact lessons from deeper-time (2)

0.0 2.01.0

Depth (cm) Sed rate-1(cm kyr )

Mean C

aC

O a

ge r

ela

tive to r

efe

rence

leve

l (ky

r)3

20

10

-10

-20

-30

-40

-50

0

‘golden spike’(normalized conc)

model experiment starts,surface sediments ‘tagged’(instantaneous pulse of inert, conservative, numerical tracer)

Quantifying ‘time’ inmodels and data

0.0 2.01.0

Depth (cm) Sed rate-1(cm kyr )

Mean C

aC

O a

ge r

ela

tive to r

efe

rence

leve

l (ky

r)3

20

10

-10

-20

-30

-40

-50

0

Age model artifact lessons from deeper-time (2)

‘golden spike’(normalized conc)

Depth above recorded ash maximum (cm)

020 10 -10 -2030

0.05

0.10

0.0

No

rma

lize

d a

sh c

on

cen

tra

tio

n

0.15

0.05

0.10

0.0

0.15-1

2.0-2.5 cm kyr

-10.5 cm kyr

RC17-126

E48-23

V29-39

V29-40

apparent offsetof event onset

apparent offsetof event onset

0 50 100−20

0

20

40

60

80

100

wt%

Co

nst

an

t d

etr

ital f

lux

ag

e r

ela

tive

to

re

f le

ve

l (k

yr)

CaCO3

δ13C

2 3.25 4.5

‰

1250 2000−20

0

20

40

60

80

100

matm

pCO2 −6.5 −5.5 −4.5

δ13C

Model a

tmosp

heric

forc

ing

‰

1262 (3500 m)

Quantifying ‘time’ inmodels and dataAge model artifact lessons from deeper-time (2)

0 50 100−20

0

20

40

60

80

100

wt%

Co

nst

an

t d

etr

ital f

lux

ag

e r

ela

tive

to

re

f le

ve

l (k

yr)

CaCO3

δ13C

2 3.25 4.5

‰

1262 (3500 m)

Quantifying ‘time’ inmodels and data‘Interface’ CaCO dissolution3

current (model) time(to machine precision)

18d OI/Ca environmental

proxies

1st, dissolution

2n

d,

bio

turb

ati

on

mix

ing

Quantifying ‘time’ inmodels and data‘Homogeneous’ CaCO dissolution3

current (model) time(to machine precision)

18d OI/Ca environmental

proxies

1st,

bio

turb

ati

on

mix

ing

2nd, dissolution

0 50 100−20

0

20

40

60

80

100

wt%

Co

nst

an

t d

etr

ital f

lux

ag

e r

ela

tive

to

re

f le

ve

l (k

yr)

CaCO3

δ13C

2 3.25 4.5

‰

1262 (3500 m)

Quantifying ‘time’ inmodels and data‘Homogeneous’ CaCO dissolution3

current (model) time(to machine precision)

18d OI/Ca environmental

proxies

1st,

bio

turb

ati

on

mix

ing

2nd, dissolution

0

Age (Ma)

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

10 20 30 40 50 60

18

d0 (

‰)

0.1

Ma

Shallow time, time (i.e. time in shallow time)Quantifying ‘time’ in

models and data

Quantifying ‘time’ inmodels and dataShallow time, time (i.e. time in shallow time)

Farrell and Prell [1989]

60% CaCO3

?

mixing(bioturbation)

dissolution(peservation)

Consider: Co-varying (glacial-interglacial) CaCO dissolution cycles and how 3

a (varying) stable isotope is recorded18Methodology: d O of planktic carbonate follows the LR04 stack, plus atmospheric pCO is 2

forced to follow the EPICA Dome C record, providing a varying preservation forcing on CaCO in marine sediments.3

Shallow time, time (i.e. time in shallow time)Quantifying ‘time’ in

models and data

Quantifying ‘time’ inmodels and dataShallow time, time (i.e. time in shallow time)

Consider: Co-varying (glacial-interglacial) CaCO dissolution cycles and how 3

a (varying) stable isotope is recorded18Methodology: d O of planktic carbonate follows the LR04 stack, plus atmospheric pCO is 2

forced to follow the EPICA Dome C record, providing a varying preservation forcing on CaCO in marine sediments.3

18Instead: d O of planktic carbonate follows SPECMAP while 500 PgC CO removed from 2

the atmosphere (to the terrestrial biosphere) across the deglacial transition (and then gradually added back again in a sawtooth shape).

Shallow time, time (i.e. time in shallow time)Quantifying ‘time’ in

models and data

0 50 1004000

4500

5000

5500

6000

6500

Age (kyr BP)

Depth

(m

)

150 200 250

color contours: wt% CaCO3

0 50 1004000

4500

5000

5500

6000

6500

Age (kyr BP)

Depth (m)

150 200 250

color contours: wt% CaCO3

0 50 100 150 200 2504000

4500

5000

5500

6000

6500

Age (kyr BP)

Depth (m)

18color contours: d O in bulk CaCO3

Shallow time, time (i.e. time in shallow time)Quantifying ‘time’ in

models and data

Implications

Once again, it matters ‘where’ CaCO dissolution occurs 3

(and what carbonate fraction) in accumulating sediments. Distortion of time-varying signals is likely minimized if a ‘homogeneous’ mode of dissolution dominates.

Use of multiple benthic individuals (even if single species) will avoid bulk sediment artifacts (other proxies will be differentially affected though), but give rise to an entertaining convolution of benthic foram population dynamics (driven by [O ] and Corg flux variability), with a 2

time-varying G-I environmental signal. (This would be an ‘interface’ like situation.)

Single foram analyses in which both age-scale and environmental proxy are simultaneously measured, is ideal.

18 14But e.g. d O would be adequate ( C not essential).

Modellers should learn some marine geology.

­

­

­

­

One (or more) of the following:

This is not relevant at all.

Too late – this was relevant 15 years ago.

There are potentially important implications for bulk carbonate and low sedimentation rate records. But no-one in their right mind uses these any more.

There are important implications for data-data (wiggle matching) and model-data analysis.

­ There are important questions raised of where in the sediments, and what fraction, of carbonate dissolves.

Meh

­

­

­

­

­

forget about itdrink beerbe happy

keep going

Quantifying ‘time’ inmodels and data

meh

Outcome

Answers to questionsQuantifying ‘time’ in

models and data