Complementarity of the Repeat Hydrography and Argo Programs

1
Complementarity of the Repeat Hydrography and
Argo Programs

• Gregory Johnson
• NOAA/Pacific Marine Environmental Laboratory

2

• Floats
• (http//www.argo.ucsd.edu/)
• 10-day intervals
• 3 degree spacing
• Boundaries?
• Covers all basins

• Sections
• (http//www.clivar.org/carbon_hydro/hydro_table.ph
  p)
• 10-year intervals
• 0.5 degree spacing
• Boundary currents
• O(1) per basin

3
WOCE Section P16 (circa 1991)(http//www-pord.ucs
d.edu/whp_atlas/)

• This is what we get from a repeat section
• Full Water column
• Resolves currents fronts from boundary to
  boundary (except when weather doesnt permit)
• Snapshot every decade but only for a few
  longitudes (or latitudes)

• This is (sort of) what we get with Argo floats
• Sampling to 2000 dbar (shallower for some
  locations)
• Lower spatial resolution than pictured (currents,
  fronts, water mass boundaries are blurred)
• But we get temporal information . . .

• But for water mass properties other than
  temperature, salinity (and velocity). . .
• Repeat Hydrography is nearly our only data
  source.
• How might the Argo better complement repeat hydro
  for other parameters?

4
Argo Deployments from Repeat Sections

• January 27, 2005 at 46.5S, 150W aboard the R/V
  Revelle (CO2/CLIVAR line P16S)
• Photo by Chris Sabine

5
Repeat Hydro Data Verify Argo Data

• Argo vs. P16S Repeat Hydro CTDs
• Blue hydro vs. red float
• Usually co-located to 10 days and 50 km
• Float goes only to 2000 dbar
• Vertical spacing finestructure?
• Lateral Spacing fronts?

• Zoom in on the mid-depth water
• 12 floats deployed on P16S 2005
• Co-located 10 days and lt 50 km
• Float - Station S 0.000 ? 0.001
• 10 floats deployed on P2 2004
• Float - Station S -0.009 ? 0.002
• Useful for climate research

6
Repeat Sections Quantify Water Masses

• South Pacific Eastern
• WOCE P18 Section Data
• Along 103W in 1994
• Potential vorticity is low
• Subtropical Mode Water
• Hanawa Talley (2001)
• after Wong Johnson (2003)

• Formed in High E-P region
• Winter evaporation cooling
• Warm salty over cold fresh
• Turner angle is high
• (Tu gt 77 density ratio lt 1.6)
• Potential for double diffusion

7
Argo Data Allow Temporal Analysis

• Seasonal Mode Water Ventilation
• Float WMO 4900454
• Localized Near 20S, 112W
• Wintertime
• Latent cooling with . . .
• Strong evaporation
• Low PV Salty anomaly subducted

• Seasonal Anomaly Evolution
• High Spring Turner Angles
• Favors Salt-fingering
• Anomalies eroded
• Migrate downward
• after Johnson (2006)
• Interannual variations . . .

8
Argo Data Provide Large-Scale Context

• Willis et al. (2004)
• Theromosteric height anomalies to 750 m from situ
  temp profiles
• SSH anomaly maps from satelllite altimeters
• Correlate map
• Global, annual, mesoscale resolution
• Argo will allow expanded analysis
• Add halosteric component
• 2000 m depth
• Better correlations
• Improved maps
• Reduced errors
• (Work with John Lyman Josh Willis)

• Complementary to Repeat Hydrography
• Circulation shifts between occupations
• Interior structure between sections
• A16N repeats in 1993 vs. 2003 . . .

9
Perspective for A16N Repeat Differences(40N to
Iceland after Johnson Gruber, 2006)
SPWM
MOW
LSW
SPWM
MOW
LSW
10
Opportunity for Mutual Improvement?(figures from
Feely et al., 2005)

• Time Differences of AOU and DIC
• Big changes at SPMW base
• Changes are well correlated
• Similar large AOU changes from decadal repeat
  sections in the
• North Pacific
• South Indian
• Southern Ocean

• Repeat sections are sparse
• Only a few locations
• Separated by decades
• Could floats help?

11
Maps of North Atlantic C Distributions (C.
Sabine)

• Approach
• Make fits of C to physical parameters from 2003
  A16, A20, A22 repeat hydrographic sections (RMSE
  5.6 µmol/kg, better than DIC RMSE of 11.4
  µmol/kg)
• Apply fits to data from ARGO profiles from same
  year to map C values across basin.
• ARGO based C distributions agree well with data
  along sections and maps are consistent with
  expected patterns.
• Subtract WOCE maps from modern maps to determine
  anthropogenic accumulation.

Difficult to make good basin-scale maps of C (or
other parameters) from WOCE sections alone
12
Adding Oxygen Greatly Improves The Fit (C. Sabine)
Rmse 11.4 µmol/kg without oxygen
Rmse 3.7 µmol/kg with oxygen
Adding oxygen to ARGO floats globally will make
these data much more relevant for biogeochemical
studies.
13
Oxygen float project Körtzinger, IFM-GEOMAR,
Kiel/Germany
Also work by Riser, Freeland . . .
Körtzinger et al. (2004). The ocean takes a deep
breath. Science 306, 1337.
14
Conclusions

• Argo and Repeat Hydrography are highly
  complementary
• Sections
• resolve boundary currents, fronts, eddies
• sample to the bottom
• Floats
• fill the interior between sections
• fill in the years between repeat sections
• Adding oxygen to floats would increase usefulness
  for climate
• But there are jobs to do and funds to raise
• Demonstrate sensor compatibility (need to show
  they wont harm the floats)
• Provide and integrate well tested calibrated
  hardware
• Provide necessary increases in stored energy or
  float equivalents to offset energy use
• Finance increased telemetry costs
• Share in data management
• Perform scientific quality control of data
• Expectation of real-time data availability

15
Is it possible to use ARGO to estimate carbon
changes?
Approach Use MLR fits of shipboard data (A16N,
A22, A20) to derive functions to estimate carbon
from ARGO profiles.
Measured Parameters T, S, P, Lat., Lon.
2003 N. A. ARGO Profiles
http//www.argo.ucsd.edu/
16
Examined Fits for a Variety of Carbon Parameters
Rmse 5.6 µmol/kg
Rmse 4.4 µmol/kg
Rmse 11.4 µmol/kg
Measured Parameters Temperature, Salinity,
Pressure, Latitude, Longitude
Physical Parameters used in MLR Fits Potential
Temperature (?), Salinity (S), Potential Density
(s?), Spiciness (p), Brunt-Vaisala frequency (N2)
C DIC0.688oxy- 0.5(Talk-0.094oxy)
Dissolved Inorganic Carbon
Total Alkalinity
Rmse 5.6 µmol/kg
Rmse 4.4 µmol/kg
Rmse 11.4 µmol/kg
C DIC0.688oxy- 0.5(Talk-0.094oxy)
17
Another complication with C approach is the
assumption of constant Redfield ratios.
?AOU 0.688 (µmol/kg)
There appear to be large AOU changes on A16N that
are of comparable magnitude to DIC changes.
18
Technology is ready and being implemented by a
few investigators, but needs large-scale
implementation to be most effective
Phase 2 Add pH to ARGO and use MLR fits of alk.
to calculate full carbon system