Studies of the equatorial box with the UMD ocean model

1
Studies of the equatorial box with the UMD ocean
model
Jim Christian Fisheries and Oceans
Canada Canadian Centre for Climate Modelling and
Analysis University of Victoria
2
? a brief note on the sea-truth experiment ?
correlation of box mean concentrations with full
domain EOFs ? carbon and nitrogen budgets
3
Since 2005 I have been testing diferent variants
of the model against underway pCO2 data provided
by Dick Feely (NOAA) and Masao Ishii (JMA)
4
High friction
Low friction
High friction
Low friction
Modified biology
Original biology
Climatological forcing
W92 gas exchange
5
Most experiments show a fairly strong correlation
over the full domain because of the mean spatial
pattern, so I defined a core ENSO region where
the correlation of a climatology run with
observed pCO2 anomalies would be 0 (panel 8)
6
Time-mean section along equator is fairly good
except in experiments 1, 4, and 5
7
? best results with original (Christian et al.
2002) biology low friction (Expt 2), or with
the modified (Wang et al. 2005) biology high
friction (Expt 3) ? we can suppress variability
either within the biological model itself or in
the mixing parameters ? applying both of these
smooths out a lot of the variability (Expts 1,
5), but applying neither generates too much
variability, and extreme values that are not
observed (Expt 4) Therefore I have selected Expts
2 and 3 for analyses of the equatorial box.
8
EOF analysis
Q what is the correlation between the mean
concentration of e.g. DIC within the box, and the
first principal component for the whole model
domain?
9
Expt 2
10
Expt 3
11
EOF analysis
A if we can adequately sample what is happening
in the box we have a pretty good index of what is
happening across the whole domain
First EOF accounts for around 30-40 of total
variance for carbon system variables (DIC, nDIC,
pCO2) and is strongly correlated with ENSO in all
cases
12
C and N budgets (for the surface mixed layer)
13
For carbon, the primary mass balance is between
upwelling and meridional advection, i.e., most of
the carbon upwelled within the box is advected
out of the box to the north and south.
Zonal advection can be convergent or divergent,
and is basically a mirror image of the net gain
or loss from upwelling after correcting for the
poleward advection.
14
The residual excess upwelling of C into the box
is 1.4 Tmol/month and is lost to particle export
(80) and CO2 outgassing (20)
15
For nitrogen, the primary mass balance is between
upwelling and sedimentation, and meridional
advection is negligible.
16
Conclusions ? the box is a very good proxy for
the first (ENSO) mode of variability over the
full domain if the mean within the box is
adequately sampled ? most carbon upwelled within
the box is advected poleward and outgasses in the
subtropics, whereas nitrogen is almost entirely
removed by sedimentation within the box ? export
of C by sedimentation within the box is 4X
larger than outgassing