Atmospheric Ar/N2

1
Atmospheric Ar/N2  A "New" Tracer of Oceanic and
Atmospheric Circulation

• Mark Battle (Bowdoin College)
• Michael Bender (Princeton)
• Melissa B. Hendricks (Princeton)
• David T. Ho (Princeton/Columbia)
• Robert Mika (Princeton)
• Galen McKinley (MIT/INE Mexico)
• Song-Miao Fan (Princeton)
• Tegan Blaine (Scripps)
• Ralph Keeling (Scripps)
• Natalie Mahowald (NCAR)

LDEO 11/05/03
GRL Vol 30, 15 (2003)
Funding from NSF NOAA GCRP Ford Res. Labs NDSEGFP
2
On the agenda

• What makes a good tracer
• Why Ar/N2
• How (and where) we measure Ar/N2
• What we observe
• Comparison with models
• Dirty laundry
• Conclusions and future prospects

3
(No Transcript)
4
My perspective on transport modeling
5
Inferring fluxes
6
But
7
How do we assess our understanding of transport?
Choose a computer model Run a tracer with known
sources through the model Compare with model
predictions with the real world
8
Not all tests of transport are equal

• Different aspects of atmospheric transport are
  important for different species
• Ar/N2 is a good analog for CO2

9
The ideal tracer(one experimentalists
perspective)

• Conservative
• Known sources and sinks, globally distributed
• Seasonally varying over land and ocean
• Measurable with great signal to noise

10
Ar/N2 The almost ideal tracer(one
experimentalists perspective)

• Conservative
• Known sources and sinks, globally distributed
• Seasonally varying over land and ocean
• Measurable with great signal to noise

chemically and biologically inert
11
Ar/N2 The almost ideal tracer(one
experimentalists perspective)

• Conservative
• Known sources and sinks, globally distributed
• Seasonally varying over land and ocean
• Measurable with great signal to noise

chemically and biologically inert
oceanic sources driven by heat fluxes
12
Ar/N2 The almost ideal tracer(one
experimentalists perspective)

• Conservative
• Known sources and sinks, globally distributed
• Seasonally varying over land and ocean
• Measurable with great signal to noise

chemically and biologically inert
oceanic sources driven by heat fluxes
seasonal, but ocean only
13
Ar/N2 The almost ideal tracer(one
experimentalists perspective)

• Conservative
• Known sources and sinks, globally distributed
• Seasonally varying over land and ocean
• Measurable with great signal to noise

chemically and biologically inert
oceanic sources driven by heat fluxes
seasonal, but ocean only
well, maybe not great
14
The Ar/N2 source/sink
Atmosphere Ar 1.2 O2 26.8 N2 100
15
The Ar/N2 source/sink
Atmosphere Ar 1.2 O2 26.8 N2 100
Heat Fluxes ? ?Ar/N2
16
The Ar/N2 source/sink
Atmosphere Ar 1.2 O2 26.8 N2 100
Heat Fluxes ? ?Ar/N2
?Ar/N2 ? ?O2/N2 (thermal)
17
A quick word on units
Ar/N2 changes are small ?Ar/N2 per meg ?
(Ar/N2sa Ar/N2st)/(Ar/N2st) x106 1 per meg
0.001 per mil
18
Our measurement technique

• Paired 2-l glass flasks
• IRMS (Finnigan DeltaXL) 40/28 and 32/28
• Custom dual-inlet system
• Standards High pressure Al cylinder

For more details GRL paper or David Ho
19
Princetons custom inlet system
20
Princeton Ar/N2 cooperative flask sampling network
21
Climatology of Ar/N2 seasonal cycle

• Monthly average
• values shown
• Multiple years (3) stacked

22
Testing models with observations
Observed modeled heat fluxes ? Solubility
equations ? Atmospheric transport
model ? Predicted Ar/N2
ECMWF or MIT OGCM (NCEP/COADS)
TM2 or GCTM or MATCH
23
Data-Model comparison

• Overall agreement

24
Data-Model comparison

• Overall agreement
• Phase problems

25
Syowa
Transport Matters (tough to get right over
Ant- arctica)
26
MacQuarie
Heat fluxes Matter (probably ECMWF- NCEP
difference)
27
SST relaxation term in MIT OGCM
28
Cape Grim
Transport and heat fluxes matter
29
Barrow
Model grid-cell selection matters
30
Data-Model comparison

• Overall agreement
• Phase problems
• SYO Transport matters
• MAC Heat fluxes matter
• CGT Both terms matter
• BRW Gridsize matters

31
Climatology of Ar/N2 seasonal cycle

• Monthly average
• values shown
• Multiple years (3) stacked

32
What about that nasty scatter?

• Problems with analysis
• Problems with collection
• Real atmospheric variability

33
What about that nasty scatter?

• Problems with analysis
• IRMS precision (? on one aliquot 4.0)
• Transfer from flask to IRMS (? 8.6)
• Total analytic uncertainty (? on a single
  flask 6.7)
• Average two flasks.

34
What about that nasty scatter?

• Problems with collection
• Does bottle air ambient air?
• From one bottle to next Yes! (? 2.6)
• From one site to next No!

35
Improving collections New sampling hardware at
Cape Grim (and elsewhere)
36
What about that nasty scatter?

• Real atmospheric variability
• Oceanic (? 0.6 1.2)
• Atmospheric (? 0.8 2.1)
• Interannual vs. Synoptic

37
Interannual Variability Ocean Atmosphere
38
In summary

• Problems with analysis
• Not negligible (? 5.1 on a collection)
• Problems with collection
• Big deal site-to-site
• New hardware helps!
• Real atmospheric variability
• Doesnt look too big, but
• Synoptic?

39
Conclusions and the future

• Ar/N2 a promising new tracer
• General data-model agreement
• Better observations to come
• Continental interior sites?
• Need Ar/N2 as active tracer in OGCMs
• Working on variability with MATCH

40
Correlated variability in Ar/N2 and O2/N2