Time series study of CFC concentrations in the Labrador Sea during deep and shallow convection regimes (1991–2000)



[1] Annual measurements of concentrations of chlorofluorocarbons (CFCs) along the Labrador Sea WOCE repeat section (AR7W) during 1991–2000 revealed increasing CFC-12 concentrations in the surface layer (<100 m), the newly ventilated Labrador Seawater (LSW), North East Atlantic Deep Water (NEADW), and Denmark Strait Overflow Water (DSOW). In newly ventilated LSW, CFC concentrations were influenced less by CFC concentrations in the atmosphere than by convection regimes, including the deep convection years of 1991–1994 and the shallow convection years of 1995–2000. The average saturation levels for CFC-12 and CFC-11 in the newly ventilated LSW also showed large variations between two convection regimes, which would suggest using caution in estimating ventilation ages for the LSW away from the source. A time series of the CFC-12 inventory in the Labrador Sea demonstrated the enhanced storage of CFC-12 during the years of deep convection, while a lower rate of inventory increase was observed during the shallow convection period of late 1990s. This suggests that a periodic renewal of intermediate and deep waters by deep convection influences the uptake of anthropogenic gases such as CFC-12 and presumably anthropogenic carbon dioxide.

1. Introduction

[2] Deep convection in the Labrador Sea in late winter varies from 200 m to over 2000 m, depending on atmospheric conditions and stratification in the water column, and produces a relatively homogeneous water mass, Labrador Seawater (LSW). This well-ventilated water mass, characterized by low salinity and temperature, provides an important vehicle for the transport of atmospheric gases, including transient tracers such as chlorofluorocarbons (CFCs), to intermediate depths of the North Atlantic Ocean. At depth in the Labrador Sea, North East Atlantic Deep Water (NEADW) and Denmark Strait Overflow Water (DSOW) flow into the region from the Nordic seas where they were formed. Consequently, all the water masses in the Labrador Sea are relatively young (<20 years [Tait et al., 2000; K. Azetsu-Scott et al., Transient tracers in the Labrador Sea, submitted to Marine Chemistry, 2003 (hereinafter referred to as Azetsu-Scott et al., submitted manuscript, 2003)]). The rate of formation of these water masses is suggested to influence the global ocean circulation, and therefore, to modulate long-term global climate [Dickson, 1997; Sy et al., 1997]. The Labrador Sea is thus an opportune region to study variability in these major water masses.

[3] The circulation of the Labrador Sea is generally cyclonic [e.g., Lazier and Wright, 1993], with a recently observed anticyclonic flow at mid-depth [Lavender et al., 2000]. In the surface layer, the cold and fresh West Greenland Current flows northwestward along the west coast of Greenland and the Labrador Current flows southeastward along the Labrador Shelf and Slope (Figure 1). The warm and saline North Atlantic Current flows northeastward across the mouth of the Labrador Sea. At depth, NEADW and DSOW flow into the region from the southeast and circulate cyclonically in the basin [Clarke and Gascard, 1983; Lazier et al., 2002]. The Labrador Sea is a highly variable region. The salinity and temperature of LSW, as well as the depth to which the water mass is ventilated in the winter, have varied significantly over the last 5 decades. Minima in temperature and salinity were observed in the mid 1950s, 1970s, and most recently in the 1993–1994 winter when deep convection reached a depth of about 2400 m [Yashayaev et al., 2000]. In the following years, convection weakened, reaching only 1500 m and shallower. Variations in salinity and temperature influence the solubility of gases, and convection depths could influence their sequestration and transport processes.

Figure 1.

Map of the Labrador Sea with the WOCE AR7W line.

[4] CFCs have well-understood atmospheric sources and have been used to calculate ventilation ages of the water masses and to validate numerical models (see England and Maier-Reimer [2001] for a review). Their inert characteristics and presently available techniques for sensitive analysis make these transient tracers powerful tools to study air-sea interaction processes such as deepwater formation and subsequent transport of ventilated waters. CFCs have been used to trace LSW across large area of the North and Equatorial Atlantic [Weiss et al., 1985; Doney and Bullister, 1992; Sy et al., 1997; Lozier, 1999; Molinari et al., 1998; Smethie et al., 2000]. An understanding of the variability of boundary conditions at the source regions is fundamental for the arguments to be valid in remote areas. We will present a comprehensive interannual study of the variability of CFCs in the Labrador Sea over a 10-year period, thus providing the information needed for assessing transport of LSW away from the source.

2. Materials and Methods

[5] The fieldwork was conducted in spring (April–July) along WOCE AR7W line (Figure 1) from 1991 to 2000. Ice conditions in the Labrador Shelf and the East Greenland Shelf did not allow us to sample whole section in some years. In 1992, ship scheduling allowed only half of the section to be samples (Table 1).

Table 1. Annual Sampling in the Labrador Sea During 1991–2000
YearSampling PeriodNumber of StationsNumber of SamplesNotes
1991April 24 to June 415114 
1992May 27 to June 149160ship scheduling allowed only half of the section to be samples
1993June 14 to June 2823245heavy ice condition over the East Greenland Shelf
1994May 24 to June 1223322heavy ice condition over the Labrador Shelf
1995June 8 to July 423252 
1996May 12 to June 127312low CFC-12 values (see text)
1997May 9 to June 1123419 
1998June 22 to July 922374heavy ice condition over the East Greenland Shelf
1999June 27 to July 1327462 
2000May 20 to June 820416 

[6] CFC-12 was measured using a purge and trap capillary GC-ECD system. A full description of the method is described in detail elsewhere (Azetsu-Scott et al., submitted manuscript, 2003). Results for samples collected during 1992 to 2000 were calibrated using working standards prepared gravimetrically at Brookhaven National Laboratories, which in turn were calibrated against a standard air sample certified by CMDL/NOAA at Boulder, Colorado. For samples collected in 1991, CFC-12 data from NEADW were calibrated against those in 1992–2000, since interannual variability of CFC-12 concentrations in NEADW was most stable among water masses in the Labrador Sea. Analytical precision of the CFC measurement is in the range of 1–3%.

3. Results and Discussion

3.1. Salinity and Temperature of Water Masses in the Labrador Sea and Their Temporal Variability

[7] From temperature, salinity and density profiles in the Labrador Sea, we characterized physical and chemical properties of five water masses: (1) the surface layer (representing the upper 100 m), (2) shallow LSW, (3) deep LSW, (4) NEADW, and (5) DSOW. Salinity and potential temperature in these waters are shown as a function of time in Figure 2. To avoid the influence of geographically limited events such as eddies and lenses, we calculated averages of temperature and salinity of each water mass at each station and then took medians. We restricted the CTD stations used in this analysis to the central region of the Labrador Sea, where the bottom depth exceeds 3300 m (approximately from 300 km to 750 km along the distance axis in Figure 3) to eliminate the influence of slope waters.

Figure 2.

Time series of (a) salinity and (b) potential temperature from 1991 to 2000. The data are medians among average values at each station for the central Labrador Sea, where the depth exceeds 3300 m. Vertical bars indicate deviations.

Figure 3.

CFC-12 distribution along the WOCE AR7W line from 1991 to 2000. In 1992, measurements were limited to the western half of the AR7W line. Access to the shelf stations depended on the ice conditions and was limited in some years.

[8] The surface layer of the Labrador Sea in the spring and summer is occupied by low salinity, warm water (S < 34.8, >2.8°C), which is formed on the Labrador Shelf and Slope, spreads offshore, and creates sharp seasonal contrasts in the water properties between 100 and 200 m. High variability of salinity and temperature in this layer is partly due to the time of the year when each survey took place (Table 1) and partly due to the effects of temporally and geographically variable shallow meso-scale eddies. Underlying the surface layer is LSW, which is divided into a shallow (LSWshallow) and a deep part (LSWdeep). The severe winters in the early 1990s resulted in the production of extremely cold and dense LSW, whose thickness exceeded 2000 m in 1994 [Lazier et al., 2002]. Shallower convection (<1500 m) in following years resulted in the coexistence of LSW with different characteristics, namely LSW produced in winter of 1993–1994 at depth (LSWdeep) and LSW produced after 1995 (LSWshallow) above the LSWdeep layer. LSWdeep is defined as a layer between the constant σ2 surfaces 36.92 and 36.95 [Yashayaev et al., 2000] and coincides with the local minima of salinity and potential vorticity [Talley and McCartney, 1982; Dickson et al., 2001]. LSWdeep has been getting warmer, saltier, and less dense at an almost constant rate since it was formed in 1993–1994. LSWshallow is defined by the σθ density surfaces between 27.72 and 27.75. Both salinity and temperature of LSWshallow increased in the late 1990s.

[9] NEADW is the oldest component in the water column of the Labrador Sea. It originates in the Nordic seas behind the sills of the Scotland-Iceland ridge. The core of NEADW was defined as a layer between two σ2 surfaces, 36.985 and 37.015. Salinity and temperature in NEADW decreased steadily during the period of this study and in longer timescale, continuing a trend begun in 1960s [Dickson et al., 2002].

[10] DSOW descends from the cold and dense overflow from the Denmark Strait sill and fills the bottom layer of the Labrador Sea. It is the densest water in the northern North Atlantic. Here we define the bottom water, mostly containing DSOW, as a 200-m-thick layer near the seafloor [Yashayaev et al., 2000]. DSOW salinity and temperature varied more over time than did NEADW salinity and temperature. Relatively warm and salty states of DSOW were observed in 1993 and 1997. In 2000, DSOW was the freshest and coldest in the 50-year record [Yashayaev et al., 2000; Dickson et al., 2002].

3.2. Distribution and Interannual Variability of CFC Concentrations in Different Water Masses

[11] CFC-12 distributions along the AR7W section from 1991 to 2000 are shown in Figure 3. CFC-11 distributions (not shown) are generally similar to that of CFC-12. However, since CFC-11 concentration in the atmosphere was declining during the period of this study, CFC-12 is chosen for the purpose of investigating interannual variability. In the central Labrador Sea, the highest CFC concentration was observed in the surface layer (<100 m). High CFC-12 concentrations on the Labrador Shelf correspond with colder and fresher water transported by the Labrador Current. When the ice condition allowed us to collect data, high concentrations of CFC-12 in the surface layer were also observed over the Greenland Shelf, in the West Greenland Current. LSW including both LSWshallow and LSWdeep extends to a depth of 2000–2200 m with high CFC-12 concentrations. The penetration of these high concentrations to depths of 2200 m by 1994 is consistent with the increasing depth of winter convection evident in the annual temperature and salinity sections [Lazier et al., 2002]. During the second half of the 1990s, concentrations continued to rise in the upper 1000–1500 m as weaker winter convection formed a shallower LSW layer above the older LSW from winter of 1993–1994. The Irminger Current, with relatively high CFC-12 concentrations, was also observed at depths between 500 m and 1500 m over the East Greenland Slope. NEADW contains the lowest CFC-12 concentration. The deepest part of the basin occupied by DSOW is characterized by CFC-12 concentrations higher than in NEADW.

[12] Remnants of active convection were observed on the western side and central part of the basin, as shown by elevated concentrations of CFC-12, in 1995, 1999, and 2000 and to a lesser degree in 1997 and 1998. In 1998, an eddy was observed on the eastern side of the basin, resulting in high CFC-12 concentrations at intermediate depths. At depth, the inflow of DSOW on the eastern side of the basin was not always apparent. Although the number of samples is limited (about 20 samples) in the thin layer of DSOW, higher CFC-12 concentrations were observed in some years far from the source region on the western side of the basin. This suggests an intermittent inflow of DSOW in the Labrador Sea or variable water composition of DSOW at the time of formation [Olsson, 2001].

[13] Interannual variability of CFC-12 concentrations in the surface layer (<100 m), LSWshallow, LSWdeep, NEADW, and DSOW was calculated (Figure 4) in the same manner as salinity and temperature (Figure 2). CFC-12 concentrations during this study increased in the surface layer, NEADW and DSOW at rates of 0.08, 0.08, and 0.12 pmol/kg/year, respectively. No increase was observed for LSWdeep during 1994–2000. The newly ventilated LSW, which is not subject to the surface warming after LSW formed each winter, is represented by LSWdeep before 1994 and LSWshallow after 1995. CFC-12 concentrations in the newly ventilated LSW increased at the rate of 0.11 pmol/kg/year. CFC-12 concentrations in 1996 were generally low, as was the CFC-12/CFC-11 ratio. A calibration problem is suspected; however, we could not identify the exact cause. Therefore 1996 data were excluded from the regression analysis.

Figure 4.

Interannual variability of CFC-12 for (a) LSW, (b) NEADW, and (c) DSOW. LSW includes the surface layer (<100 m) (solid triangles), LSWshallow (shaded diamonds) and LSWdeep (shaded squares). The data indicate the medians with vertical bars for the maximum and the minimum. Estimated equilibrium CFC-12 concentrations with the atmosphere for the newly ventilated LSW, NEADW, and DSOW were shown as open circles, open squares, and open triangles, respectively. The regressions with the correlation coefficients and a significant level (p) for each water mass are as follows: LSW: equilibrium CFC-12 conc. with atmosphere: y = 0.02x − 36.21, (r2 = 0.78, p < 0.01). The surface layer (<100 m): y = 0.08x − 147.27, (r2 = 0.87, p < 0.005). The newly ventilated LSW: y = 0.11x − 211.17, (r2 = 0.90, p < 0.01). (b) NEADW: Equilibrium CFC-12 conc. with atmosphere: y = 0.09x − 185.21, (r2 = 0.97, p < 0.001). NEADW: y = 0.08x − 154.34, (r2 = 0.96, p < 0.001). (c) DSOW: Equilibrium CFC-12 conc. with atmosphere: y = 0.10x − 199.34, (r2 = 0.96, p < 0.001). DSOW: y = 0.12x − 236.71, (r2 = 0.89, p < 0.002).

[14] To evaluate the oceanic increase of CFC-12 in comparison with that in atmosphere, we calculate the CFC-12 concentrations in equilibrium with atmosphere for each water mass. The equilibrium CFC-12 concentrations in the newly ventilated LSW were calculated using its salinity, temperature and the reported CFC-12 concentration in the atmosphere [Walker et al., 2000] (open circles in Figure 4a). The newly ventilated LSW became more saline and warmer in the latter half of 1990s, and this caused a decreasing solubility of CFC-12 as LSW was formed. The rate of increase for CFC-12 concentrations in the atmosphere declined in the same period and, together with the decreasing solubility, this resulted in only a slight increase of equilibrium CFC-12 concentration with atmosphere (0.02 pmol/kg/year) for newly ventilated LSW. A higher rate of CFC-12 concentration increase in the newly ventilated LSW than that in the atmosphere can be explained by undersaturation of CFC-12 in the water during deep convection years. During the deep convection period, convection plumes reached to and mixed with NEADW that contained the lowest concentration of CFC-12 in the Labrador Sea. Although convection ventilates the intermediate and deep waters, the convection takes place within a few weeks over a small area during winter and is followed by lateral mixing [Lilly et al., 1999]. The resultant LSW is thus undersaturated with atmospheric gasses such as oxygen and CFCs [Clarke and Coote, 1988; Azetsu-Scott et al., submitted manuscript, 2003]. The surface layer gradually comes to equilibrium with the atmosphere in the following shallow convection years. Saturation levels of CFC-12 and CFC-11 for the newly ventilated LSW are discussed in detail in the following section. Ventilation ages of NEADW and DSOW in the Labrador Sea are estimated to be about 12 years and 6 years, respectively (Azetsu-Scott et al., submitted manuscript, 2003) and the corresponding equilibrium CFC-12 concentration increased at constant rate of about 0.1 pmol/kg during 1970–1995. Their rates of increase correspond to that in the atmosphere. The steady increase in CFC concentrations in the DSOW and NEADW may reflect the formation mechanisms for those water masses in the Nordic Seas, which has not been strongly driven by deep convection since 1970s.

[15] The apparent oxygen utilization (AOU) in LSWdeep increased at a rate of 1.55 μmol/kg/year after 1994 (data not shown). Constant CFC-12 concentrations and increasing AOU during 1995–2000 confirmed that LSWdeep was produced during deep convection in the winter of 1993–1994, and has slowly been eroded by mixing with ambient waters since then as suggested by Lazier et al. [2002].

[16] CFC-11 concentrations increased in the newly ventilated LSW, in NEADW, and in DSOW as did those for CFC-12. CFC-11 concentrations in the surface layer (<100 m), however, did not show any change (Table 2). The atmospheric concentrations of CFC-11 stopped increasing significantly in mid-1980s and started declining in mid-1990s [Walker et al., 2000]. This CFC-11 concentration history in the atmosphere has already made the annual increase insignificant in the surface layer in the Labrador Sea. However, the newly ventilated LSW had not reached equilibrium with the atmospheric CFC-11, because convection occurs regionally during winter and subsequent lateral mixing with older water during the rest of the year lowers saturation levels as observed in CFC-12. Therefore ocean uptake of CFC-11 will continue until the equilibrium with the atmosphere is achieved. CFC-11 concentrations in newly ventilated LSW are still increasing at a rate of 0.13 pmol/kg/year, and will approach equilibrium with atmosphere in 2007 assuming a steady increase rate.

Table 2. Interannual Variability of Salinity, θ, σ(θ,2,4), CFC-12 and CFC-11 Concentrationsa
YearSurface Layer (<100 m)LSWshallowLSWdeep
Salinityθ, °CσθCFC-12, pmol/kgCFC-11, pmol/kgSalinityθ, °CσθCFC-12, pmol/kgCFC-11, pmol/kgSalinityθ, °Cσ2CFC-12, pmol/kgCFC-11, pmol/kg
  • a

    N.A no data available. Data are medians of averages for each station (refer to the context for detailed description). CFCs are expressed in median ± a standard deviation. Note the notation for potential density anomary (σθ, σ2 and σ4).

199134.773.1527.67N.A.N.A.     34.832.8136.931.56 ± 0.243.12 ± 0.35
199234.793.1227.712.18 ± 0.234.53 ± 0.07     34.842.8136.931.62 ± 0.163.41 ± 0.34
199334.632.8327.612.56 ± 0.195.08 ± 0.31     34.842.7336.941.82 ± 0.123.42 ± 0.21
199434.732.9527.672.49 ± 0.265.12 ± 0.19     34.832.7136.941.89 ± 0.223.80 ± 0.24
199534.703.1727.602.57 ± 0.404.77 ± 0.1334.812.9627.742.13 ± 0.324.23 ± 0.1534.842.7236.941.84 ± 0.293.79 ± 0.06
199634.663.1427.622.35 ± 0.154.97 ± 0.2534.813.0227.742.11 ± 0.154.15 ± 0.2134.842.7536.941.77 ± 0.143.67 ± 0.13
199734.743.7427.632.73 ± 0.165.02 ± 0.1634.833.1927.742.43 ± 0.224.32 ± 0.2034.852.7936.941.93 ± 0.163.76 ± 0.24
199834.784.1227.612.80 ± 0.265.01 ± 0.2434.833.1527.742.48 ± 0.254.54 ± 0.3134.852.8236.941.88 ± 0.133.82 ± 3.90
199934.684.0527.522.88 ± 0.364.90 ± 0.3034.843.2327.742.35 ± 0.303.90 ± 0.3434.862.8536.941.87 ± 0.163.43 ± 0.16
200034.743.4127.642.86 ± 0.194.99 ± 0.2934.833.1627.742.40 ± 0.234.31 ± 0.3934.862.8936.941.78 ± 0.103.79 ± 0.17
Salinityθ, °Cσ2CFC-12, pmol/kgCFC-11, pmol/kgSalinityθ, °Cσ4CFC-12, pmol/kgCFC-11, pmol/kg     
199134.922.7537.000.58 ± 0.301.06 ± 0.1334.881.4645.960.80 ± 0.222.01 ± 0.22     
199234.922.7637.000.74 ± 0.161.55 ± 0.2434.891.5845.950.97 ± 0.202.22 ± 0.40     
199334.922.7437.000.75 ± 0.321.52 ± 0.6634.891.6845.931.22 ± 0.142.30 ± 0.28     
199434.912.7137.000.94 ± 0.121.87 ± 0.1634.881.5645.941.38 ± 0.182.79 ± 0.30     
199534.912.7137.001.01 ± 0.121.79 ± 0.1734.881.4845.961.71 ± 0.273.09 ± 0.35     
199634.912.6937.001.04 ± 0.152.10 ± 0.2034.871.4545.961.61 ± 0.163.27 ± 0.21     
199734.912.6937.001.11 ± 0.142.16 ± 0.2534.891.6245.941.61 ± 0.203.11 ± 0.37     
199834.902.6837.001.18 ± 0.172.44 ± 0.2934.891.5945.941.88 ± 0.293.82 ± 0.28     
199934.902.6637.001.31 ± 0.222.33 ± 0.2434.881.5145.951.84 ± 0.313.32 ± 0.32     
200034.902.6637.001.27 ± 0.142.41 ± 0.2434.871.3345.981.89 ± 0.183.41 ± 0.29     

3.3. Interannual Variability of CFC-12 and CFC-11 Saturation in LSW

[17] Ventilation ages estimated from transient tracers are strongly influenced by boundary conditions [Doney and Jenkins, 1988]. The saturation level of the CFCs is a crucial parameter in the calculation of the ventilation ages (Azetsu-Scott et al., submitted manuscript, 2003) and formation rates of water masses [Smethie and Fine, 2001]. Information on CFC saturation levels is generally scarce, and previous studies are limited to those of CFC-11. For example, Wallace and Lazier [1988] reported the CFC-11 saturation as 60% for LSW at the western part of the basin. Smethie et al. [2000] recorded 70% saturation of CFC-11 for the Upper Labrador Seawater, which is considered to be produced at the southwest margin of the Labrador Sea, possibly in the Labrador Current [Pickart, 1992]. For the DSOW, a saturation of 60–75% was estimated [Smethie et al., 2000].

[18] Here we report for the first time the interannual variability of CFC-12 and CFC-11 saturation for the newly ventilated LSW using the solubility coefficients by Warner and Weiss [1985] and the atmospheric concentrations by Walker et al. [2000] (Figure 5). Both CFC-12 and CFC-11 saturation levels were low during the deep convection period of the early 1990s because of mixing with NEADW with its low CFC concentrations, and they both increased during the shallow convection periods in the latter half of 1990s. Average saturation levels for CFC-12 and CFC-11 were 61 ± 4% and 58 ± 4% during 1991–1994 and 81 ± 4% and 74 ± 4% during 1995–2000, respectively. Insufficient time for the water to equilibrate with the atmosphere at the surface is considered to be the reason for the low saturation of CFCs in the deepwater formation sites. However, remnant water of the active convection regions (for example, observed in 1999 and 2000) showed 5–10% higher CFC-12 saturation levels than the nonconvection regions. The observed undersaturation in the newly ventilated LSW can be accounted for geographically constrained convection in the western basin over a short period in winter, with subsequent and varying lateral mixing with older water in following seasons.

Figure 5.

Variability of saturation levels for CFC-12 and CFC-11 in the newly ventilated LSW (LSWdeep during 1991–1994 and LSWshallow during 1995–2000). Medians for CFC-12 (solid triangles) and CFC-11 (solid diamonds) with ± maximum and minimum as vertical bars are shown.

[19] The temporal variability in the CFC-12 and CFC-11 saturation levels counsels caution for the estimation of ventilation ages of LSW away from the source region. The observation of 20% variation in CFC-12 and over 15% in CFC-11 saturation between shallow and deep convection regimes would result in large errors for estimates of ventilation ages. This is especially evident for the younger water, because the declining increase rate of CFC concentrations in the atmosphere in recent years makes the resolution of the ventilation ages lower. A 25% variation in CFC-12 saturation results in an error of 6–8 years in the ventilation ages for the water produced in 1990s. For older waters (>20 years old), it corresponds to errors of about 4 years.

3.4. CFC-12 Inventory Variability

[20] CFC-12 inventories in the Labrador Sea were estimated using data from the AR7W section. Integration of CFC-12 over the section was calibrated with the inventory estimate made by a basin wide survey in 1997, which includes five sections with over 1600 data (Azetsu-Scott et al., submitted manuscript, 2003). CFC-12 inventory in 1997 was 7.56 × 106 mol. The inventory of CFC-12 increased at a rate of 0.80 × 106 mol year−1 during the years of deep convection (1991–1994) and at a rate of 0.24 × 106 mol year−1 after the years of deep convection (1994–2000) (Figure 6). This indicates that in years of deep convection, storage of CFC-12 occurs at a much faster rate than in years of shallow convection. The overall increase of CFC-12 in the Labrador Sea was 0.35 × 106 mol year−1.

Figure 6.

Inventory variation for CFC-12 during 1991–2000. Vertical bars indicate ±1 standard deviation of the estimate. The regressions for CFC-12 during 1991–1994 and during 1994–2000 and the overall regression (1991–2000) are: 1991–1994: y = 0.80x − 1583.00, (r2 = 0.90, p < 0.1); 1993–2000: y = 0.21x − 421.86, (r2 = 0.83, p < 0.05); 1991–2000: y = 0.35x − 691.55, (r2 = 0.80, p < 0.01).

4. Conclusions

[21] A time series study (1991–2000) in the Labrador Sea illustrated the increased concentration of CFC-12 in the surface layer (<100 m), the newly ventilated LSW, NEADW and DSOW. In NEADW and DSOW, the rates of CFC-12 concentration increase reflected the atmospheric trend, while those in the surface layer and newly ventilated LSW were faster than that in the atmosphere. This is due to the shift of convection regimes, from deep convection in the early 1990s to subsequent shallow convection, influencing the saturation levels of CFC-12 in the water. Weakening of convection in the Labrador Sea after 1994 resulted in formation of LSWdeep at depths between 1500 m and 2200 m. Constant CFC-12 concentrations together with increased apparent oxygen utilization confirmed this water mass has been remained in place since 1994. Average saturation levels for CFC-12 and CFC-11 during the deep convection (1991–1994) are 61% and 58% and during shallow convection (1995–2000) 81% and 74%, respectively. The wide range of saturation levels raises concerns and the need for caution in estimating LSW ventilation ages downstream from the Labrador Sea. Interannual variability of the inventory demonstrated the relationship between the storage of CFC-12 and the convection regimes. The deep convection observed in the early 1990s enhanced the storage of CFC-12 in the Labrador Sea. Our study demonstrated that the intensity and frequency of deep convection significantly influences the uptake of anthropogenic gasses such as CFC-12. Since CFC concentrations in the water correlate well with anthropogenic CO2 concentration, the variability in deep convection may influence not only the global ocean circulation, but also the sequestration of the anthropogenic CO2.


[22] We are grateful to scientists of Bedford Institute of Oceanography and the officers and crew of the C.C.G.S. Hudson, who conducted the lengthy cruises over the years. Our special thanks go to Pierre Clement, Frank Zemlyak, Mike Hingston, Anthony Isenor, and Val Tait. Comments by Allyn Clarke and Ross Hendry on the earlier version of the manuscript improved this work. This work was supported by Canadian Climate Action Fund and Canada Panel on Energy Research Development.