Decadal Ventilation and Mixing of Indian Ocean Waters

RANA A. FINEa*, WILLIAM M. SMETHIE, JRb, JOHN L. BULLISTERc, MONIKA RHEINd, DONG-HA MINe, MARK J. WARNERf, ALAIN POISSONg, AND RAY F. WEISSh

aRosenstiel School of Marine and Atmospheric Science, University of Miami, 4600 Rickenbacker Causeway, Miami, FL33149-1098, USA

bLamont Doherty Earth Observatory, ColumbiaUniversity, Palisades, NY10964, USA

cNational Oceanic and Atmospheric Administration, Pacific Marine Environmental Laboratory, 7600 Sand Point Way, NE, Seattle, WA98115, USA

dUniversity Bremen, Institute for Environmental Physics, Department of Oceanography, D28359 Bremen, Germany

eUniversity of Texas at Austin, Marine Science Institute, 750 Channel View Drive, Port Aransas, TX 78373-5015, USA

fUniversity of Washington, School of Oceanography, Seattle, WA98195-7940, USA

gLaboratoire de Biogeochimie et Chimie Marines, IPSL-CNRS, Universite Pierre et Marie Curie, case 134, 4 place Jussieu, 75252, Paris cedex 05, France

hScripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0220, USA

Finally Revised for Deep-Sea Research, September 2007

*Corresponding Author: Tel: 305-421-4722; E-mail:

Abstract

Chlorofluorocarbon (CFC) and hydrographic data from the World Ocean Circulation Experiment (WOCE) Indian Ocean expedition are used to evaluate contributions to decadal ventilation of water masses. At a given density, CFC-derived ages increase and concentrations decrease from the south to north, with lowest concentrations and oldest ages in Bay of Bengal. Average ages for thermocline water are zero to 40 years, and for intermediate water they are less than 10 years to more than 40 years. As compared with the MarginalSeas or throughflow, the most significant source of CFCs for the Indian Ocean south of 12ºN is the Southern Hemisphere. A simple calculation is used to show this is the case even at intermediate levels due to differences in gas solubilities and mixing of Antarctic Intermediate Water and Red Sea Water.Bottom Water in the Australia-AntarcticBasinis higher in CFC concentrations than to the west in the EnderbyBasin, due to the shorter distance of this water to the Adelie Land coast and RossSea sources. However, by 40ºS, CFC concentrations in the Bottom Waterof the CrozetBasin originating from the Weddell Sea are similar to those in theSouth AustraliaBasin. Independent observations, which show that Bottom Water undergoes elevated mixing between the Australia-Antarctic Basin and before entering the subtropics, are consistent withhigh CFC dilutions (3- to 14-fold) and a substantial concentration decrease (factor of 5) south to north of the Southeast Indian Ridge. CFC-bearing Bottom Water with ages of 30 years or more is transported into the subtropical South Indian Ocean by three western boundary currents, and highest concentrations are observed in the westernmost current. During WOCE, CFC-bearing Bottom Water reaches to about 30ºS in the PerthBasin, and to 20ºS in the MascareneBasin. Comparing subtropical Bottom Water CFC concentrations with those of the South Pacific and Atlantic oceans, at comparable latitudes, Indian Ocean Bottom Water CFC concentrations are lower, consistent with its high dissipation rates from tidal mixing and current fluctuationsas shown elsewhere. Thus, the generally high dilutions and low CFC concentrations in Bottom Water of the Indian Ocean are dueto distance to the water mass source regions and the relative effectiveness of mixing. While it is not surprising that at thermocline, intermediate, and bottom levels, the significant ventilation sources on decadal time scales are all from the south, the CFCs show how local sources and mixing within the ocean affect the ventilation.

Keywords: ventilation, CFCs, water masses, Indian Ocean, circulation, tracers

1. Introduction

Properties of water masses in the Indian Ocean are affected by exchange at their boundaries with local sources from Marginal Seas (Red Sea and Persian Gulf), by the input of river runoff to the surface primarily into the Bay of Bengal, by exchange with adjacent ocean basins via the Indonesian throughflow and the southern boundary, and by modification due to mixing within the ocean (Schott and McCreary, 2001). Previous studies have described chlorofluorocarbon (CFC) observations in these boundary regions. Olson et al. (1993) found relatively low dissolved gas concentrations in the source regions of Persian Gulf Water (PGW) due to the high temperatures and salinities and thus lower solubilities. The distributions of dissolved gases in Red Sea Water (RSW) should be similar, as it leaves the source region at relatively high temperatures and becomes strongly diluted (Mecking and Warner, 1999; Bower et al., 2000). Observations within the IndonesianSeas (Gordon and Fine, 1996) and in the region where throughflow water enters the Indian Ocean (Fieux et al., 1996) do not reveal a distinct signature of CFCsin the upper water column, and deep waters are CFC-free. In contrast, Fine (1993) concluded, based on CFC distributions along 32°S,that Subantarctic Mode Water (SAMW) from the southeast and Antarctic Intermediate Water (AAIW) from the southwest transport high CFC concentrations into the subtropical gyre. Deeper in the water column, cold, fresh Bottom Water carrying low-level CFC concentrations enters the southern Indian Ocean from both the east and west (e.g., Orsi et al., 2002). Based on earlier observations of dissolved oxygen (e.g., Wyrtki, 1971; Swallow, 1984), the importance of the southern sources in ventilating the Indian Ocean is not surprising.

The World Ocean Circulation Experiment (WOCE) CFC and hydrographic data are used to estimate decadal ventilation from tracer ages of thermocline, intermediate, and Bottom Waters. Estimates of "age" can be calculated from the partial pressures of dissolved CFCs 11 and 12 (pCFC-11 and pCFC-12) or from the ratio of the partial pressures (pCFC-11/pCFC-12). In both cases, the observed partial pressures or partial pressure ratios are compared to the atmospheric source function (Walker et al., 2000) to determine the date at which the dissolved CFCs in the water sample would have been in equilibrium with the atmosphere. The resulting age is the elapsed time from that date to the time of sampling.

Ages calculated using either the ratio or partial pressure method are appropriate for different circumstances. For thermocline ventilation, water subducted in a given year mixes with water subducted during prior years, so that a water parcel is a mixture that has left the surface over a several-year period. The average age of a water parcel can be approximated by the pCFC age. Studies using simple models (e.g., Doney et al., 1997; Sonnerup, 2001) have shown that pCFC and ideal ages, which are mean ages, agree most closely over periods when the atmospheric growth rates for the CFCs were roughly linear (1965-1990). A recent ocean modeling study found considerable interannual variability of pCFC ages in the North Pacific (Tsumune et al., submitted), but downstream of the formation region this variability is smoothed out by isopycnal mixing and the pCFC ages represent a mean age. For older waters,and in waters that are mixtures from sources with different temperatures, non-linearities in the atmospheric source functions and solubilities cause pCFC ages to diverge from ideal ages. In these cases,(e.g., intermediate waters), we use ratio ages. Ratio ages represent the age of the CFC-bearing component in a mixture of water parcels. Tracer ages are model-dependent, and the errors associated with pCFC or ratio ages will vary for some of the reasons discussed. Errors in CFC ages range from 10% in the most optimistic case, and may be off by factors of 2-3 in the most pessimistic case. Thus, the ages are presented for relative comparisons rather than for quantitative purposes.

2. Data

The quality of the one-time WOCE Indian Ocean CFC data are excellent and generally meet the relaxed WOCE standards (defined as precisions better than 3% of the concentrations or 0.015 pmol kg-1 whichever is greater). The station locations and dates are given in Fig. 1 and its caption. The CFC data are reported on the SIO-98 calibration scale (Prinn et al., 2000). Measurement groups used slightly-modified procedures of the purge-and-trap technique of Bullister and Weiss (1988). The WOCE CFC Indian Ocean data are available on DVDs (WOCE Data Products Committee, 2002). Vertical sections, maps of ages and concentrations on isopycnal surfaces, surface saturation maps, and a table of blank level corrections and precisions are available at gecko.rsmas.miami.edu.

Near the Persian Gulf, anomalously high dissolved CFC-12 concentrations were measured, far above those expected from normal air-sea gas exchange with background atmospheric CFC concentrations (Rhein et al., 1997; Plähn et al., 1999). The CFC-12 anomalies were probably from solvents and fire extinguishers related to the first Gulf War. These anomalous data are not included in figures presented here.

3. Discussion of Ventilation

During the WOCE Indian Ocean expedition, observed CFC concentrations poleward of 40S are at least several times higher than the detection limit (roughly 0.005 pmol kg-1) throughout the water column (e.g., Figs. 2a and 2b). Equatorward of 40S, CFC-bearing waters are found at depths from the thermocline through intermediate layers in the subtropics and tropics (e.g., Figs. 2c and 2d). A common pattern for thermocline, intermediate, and Bottom Water is a decrease in CFC concentrations from the south to north along isopycnal surfaces.

3.1. Thermocline Water

The CFC concentrations and ages vary considerably through Indian Central Water (ICW) (~26-27). In the North Indian Ocean, CFC concentrations are relatively lower and ages are older in the Bay of Bengal than in the Arabian Sea (Fig. 3a-f). There are little (if any) sources of recently ventilated water in the Bay of Bengal, except for low salinity runoff into the surface layers. Recently ventilated water enters the Arabian Sea directly from the Marginal Seas, but the resulting CFC and oxygen concentrations of PGW are relatively low (Olson et al., 1993) compared with water at comparable density that is colder (and fresher) at its outcrop. Within the South Indian subtropical gyre, ICW has pCFC-12 ages of 2-14 years (Figs. 3a-e). These ages are similar to those observed at the same densities in the South Pacific Ocean (Fine et al., 2001).

In the thermocline of the subtropical gyre south of 15S, there are zonal variations in CFC concentrations and ages. Subantarctic Mode Water in the southeast Indian Ocean has lower potential vorticity (McCartney, 1982), higher CFCs (Fine, 1993), and is denser (26.7 ) (Karstensen and Tomczak, 1998) than the SAMW observed in the southwest (26.5 ). Ages of the SAMW on 26.7  vary from 2-4 years in the southeast to more than 14 years in the southwest (Fig. 3e), and SAMW is the dominant water mass for ventilation of the thermocline (Karstensen and Quadfasel, 2002). High CFC concentrations observed in the southeast in the WOCE data (on 26.7  Fig. 3e and 26.5  not shown) lend support to earlier suggestions (e.g., Warren et al. 1966; Wyrtki, 1971; Swallow, 1984; Prunier, 1992; Stramma and Lutjeharms, 1997; Sloyan and Rintoul, 2001) that the southeast Indian Ocean is the source for recently-ventilated thermocline water transported into the North Indian Ocean primarily along the western boundary.

The contrast between low CFC concentrations of the thermocline and intermediate waters of the North Indian Ocean and their high concentrations of the subtropical gyre is dramatic (Figs. 2c and 2d). Largest meridional gradients of CFC-11 concentration are located between 10º-20ºS. These meridional CFC-11 gradients increase with depth through the thermocline. Gradients are also observed in CFC ages (Figs. 3a-h) and in other tracers (e.g., Wyrtki, 1971; Fine, 1985; Gordon, 1986; Gordon et al., 1997), and reflect differences in the sources and the spreading ratesof water masses. The Indonesian throughflow contributes to the meridional gradients of some properties including CFCs. Throughflow water has CFC concentrations that are lowerthan those of the subtropical gyre and higher than those in the Bay of Bengal.

3.2. Intermediate Water

There are several sources of intermediate water to the Indian Ocean. There is relatively high CFC AAIW (~27.1 ) entering in the southwest (Fine, 1993; also observed in the 2002 transect along 32ºS), and low CFC AAIW from the Pacific entering in the southeast (Fine, 1993). In addition, there are local sources of intermediate water in the southeast Indian Ocean (Schodlok et al., 1997) and fromthe MarginalSeas. The CFC ratio ages for intermediate water are less than 25 years (Fig. 3g) in the subtropical gyre. Ratio ages are younger than the pCFC ages (Fig. 3f) because the dilution of CFC concentration by mixing with very low or zero-CFC water has lowered the CFC concentration resulting in an older age, but has not changed the pCFC-11/pCFC-12 ratio. The oldest intermediate water is observed in the Bay of Bengal, where CFC ratio ages exceed 40 years, while CFC ages in the Arabian Sea lie between those of the South Indian and Bay of Bengal. At 27.3 (Figs. 3g and 3h), there is an influence of relatively younger RSW (Mecking and Warner, 1999) coincident with westward-intensified high salinity. Ratio ages of RSW in the western Gulf of Aden are 18-27 years. There is a ratio age difference of about 10 years between the western Arabian Sea and the northern Bay of Bengal.

A rough estimate can be made of the contribution of RSW, AAIW, and Indonesian throughflow to the measured CFC concentration at 27.1  and 5ºN along the western boundary by combining the fractional water mass contributions from the source regionsas estimated by You (1998, his Figs. 9 and 19) with WOCE data (Fig. 4) for three components. At 27.1  and 32ºS, 55ºE water is about 80% AAIW (You, 1998) with CFC-11 concentrations of ~1.2 pmol kg-1. (They imply 1.5 pmol kg-1 at 100% if the mixing occurs with CFC-free water). At 15ºN, 55ºE there is 100% RSW (You, 1998) with concentrations of approximately 0.18 pmol kg-1. [At RSW outcrop, CFC saturations are 20-50% (Mecking and Warner, 1999) with dilution factors of ~2.5 (Bower et al., 2000).] At 10ºS, 120ºE there is 100% Indonesian throughflow water with concentrations of 0.06 pmol kg-1. Along the western boundary at 5ºN, You (1998) estimated the water at 27.1  to be a mixture of 20% AAIW, 70% RSW, and 10% Indonesian throughflow. Because of lack of information, we assume that these three end-member waters take the same amount of time to get to the boundary at 5ºN, without their concentrations changing. Then we can sum the product of these percentages times their CFC-11 concentrations at the sources (0.2*1.5 + 0.7*0.18 + 0.1*0.06) to get a concentration of 0.43 pmol kg-1.It is nearly double the 0.25 pmol kg-1 on the map (Fig. 4), which suggests that the high concentration AAIW portion is probably even smaller than 20%. Still this simple calculation shows that even though there is a lower percentage of AAIW than RSW along the western boundary, AAIW contributes most of the CFCs to the mixture because of the differences in gas solubilities and dilution at the sources of these two water masses. At 27.3 , below the core of AAIW, RSW is also a relatively weak and localized CFC source. Thus, when the effect on the large scale is considered, although RSWis a significant source of salt to the thermohaline circulation (e.g., Beal et al., 2000), it is not a significant influence as a source of water recently ventilated with CFCs (and oxygen) relative to AAIW.

Schodlok et al. (1997) discussed the possibility of another localized source of intermediate water formed in the Indian Ocean near Australia. This water was observed at depths of 400-500 m just south of the subantarctic front at 47º-48ºS, 115ºE, and was characterized by oxygen concentrations greater than 280 umol kg-1 at potential temperatures of 6ºC, and salinities less than 34.3. Several months later, WOCE stations were occupied on the I9S line along 115ºE. Dissolved oxygen concentrations greater than 280 umol kg-1 at 6ºC are found at 48ºS-50ºS, but at depths of only 80-150 m. At one of these stations, CFC-11 concentration is greater than 4.8 pmol kg-1. At a few stations from 46º-48ºS along section I8S (93ºE, Fig. 2b), there are oxygen concentrations greater than 280 umol kg-1 and CFC-11 concentrations greater than 4.4 pmol kg-1 at depths near 150 m.Yet stations further south have lower CFC and oxygen concentrations at the same temperature of 6ºC. Thus, the WOCE data support the 1994 observations of Schodlok et al. (1997) of unusually high tracer concentrations at intermediate densities for several stations south of the subantarctic front near Australia. The spatial extent and temporal persistence of this high-tracer water cannot be determined from the available data. However, low CFC concentrations at intermediate levels in the WOCE data from the southeast Indian Ocean suggest that, the local source has at most a small-scale influence on ventilating the Indian Ocean. In summary, the major ventilation source of CFCs in intermediate water of the Indian Ocean is AAIW entering in the south, as opposed to local sources (southeast AAIW and RSW) and throughflow.

3.3. Bottom Water

Antarctic Bottom Water with low salinity is prominent in data from lines extending into the Southern Ocean(e.g., Fig. 2a and 2b). The WOCE data provide an opportunity for a basin scale examination of how CFC concentrations in Bottom Water entering the Southern Ocean Indian sector and subsequently the subtropical Indian Ocean are influenced by proximity to source regions and mixing.

There are two significant sources for Bottom Water entering the Indian Ocean. Bottom Water from the most productive source, the Weddell Sea (e.g., Worthington, 1981; Rintoul, 1998; Orsi et al., 1999), enters the EnderbyBasin (Fig. 1), then the Agulhas and Crozet Basins,and the Mozambique and Madagascar Basins (e.g., Ivanenkov and Gubin, 1959; Mantyla and Reid, 1995) after undergoing substantial dilution(Haine et al., 1998). The second significant source is water primarily from the Adelie Land coast (~143ºE), (Gordon and Tchernia, 1972; Rintoul, 1998) with a contribution from the Ross Sea (Mantyla and Reid, 1995) that enters the Australia-Antarctic Basin, then the South Australia Basin, and the Perth Basin. There is an additional source of Bottom Water from the Amery Ice Shelf (near 70ºE), (Jacobs and Georgi, 1977), which is not presently a significant source of CFCs directly to the Indian Ocean. The Australia-AntarcticBasin is nearer to its source of Bottom Water than the EnderbyBasinis to its source and therefore has higher CFC concentrations (Fig. 5) (Orsi et al., 1999; Rintoul and Bullister, 1999).