Abstract

We present a new global oxygen atlas. This atlas uses all of the available full water column profiles of oxygen, salinity, and temperature available as part of the World Ocean Database released in 2018. Instead of optimal interpolation, we use the Data Interpolating Variational Analysis (DIVA) approach to map the available profiles onto 108 depth levels between the surface and 6800 m, covering more than 99% of ocean volume. This 1/2° × 1/2° atlas covers the period 1955–2018 in 1-yr intervals. The DIVA method has significant benefits over traditional optimal interpolation. It allows the explicit inclusion of advection and boundary constraints, thus offering improvements in the representations of oxygen, salinity, and temperature in regions of strong flow and near coastal boundaries. We demonstrate these benefits of this mapping approach with some examples from this atlas. We can explore the regional and temporal variations of oxygen in the global oceans. Preliminary analyses confirm earlier analyses that the oxygen minimum zone in the eastern Pacific Ocean has expanded and intensified. Oxygen inventory changes between 1970 and 2010 are assessed and compared against prior studies. We find that the full ocean oxygen inventory decreased by 0.84% ± 0.42%. For this period, temperature-driven solubility changes explain about 21% of the oxygen decline over the full water column; in the upper 100 m, solubility changes can explain all of the oxygen decrease; for the 100–600 m depth range, it can explain only 29%, 19% between 600 and 1000 m, and just 11% in the deep ocean.

Abstract

This paper presents a modified objective mapping technique that takes advantage of the strong vertical correlations in ocean temperature profiles. This technique has been used here successfully to generate a uniformly gridded upper-ocean temperature dataset in the southwest Pacific Ocean region from most of the available bathythermograph casts collected between 0°–50°S and 140°E–180°, covering the period from 1955 to 1988. Important advantages of this technique over most previous objective methods are its (i) ability to deal with four-dimensional data (space and time), (ii) improved estimate of the first-guess (polynomial) mean, (iii) preservation of the vertical structure of the ocean temperature data, (iv) computational efficiency, and (v) objective error analysis.

The technique combines empirical orthogonal function (EOF) analysis, using the singular value decomposition, and objective mapping. In this application of the method, a digital “atlas” of upper-ocean temperatures has been created on a grid 2° × 2°, at 5-m depth intervals, and comprises a monthly climatology and two three-monthly time series (January, April, July, and October). The time series include a dataset from 1955 to 1988 to 100-m depth, and a shorter period, deeper dataset from 1973 to 1988 to 450-m depth. Only the first five vertical EOFs are needed to explain about 90% of the total variance in the data and to within the a priori noise estimates. The full four-dimensional temperature field was reconstructed using objective maps of the horizontal coefficients corresponding to each of the significant vertical EOFs. Although the method is statistically suboptimal, the final mapped temperature fields are unbiased and consistent with the a priori noise. In this application, the computing time is reduced by a factor of 36, making the mapping procedure feasible on modern workstations.

Abstract

Observed changes in temperature and salinity properties on isopycnals across hydrographic sections throughout the Indo-Pacific are compared with the changes modeled by the coupled climate model, HadCM3. Observations show cooling and freshening on isopycnals in midlatitudes, and there is quantitative agreement between modeled and observed water mass changes on five out of six zonal sections. The full Indo-Pacific pattern of change in the climate model is examined and it is discovered that the pattern of cooling and freshening on isopycnals in midlatitudes, with warming on isopycnals at high latitudes, may be thought of as a fingerprint of anthropogenic forcing. The water mass changes are related to changes in the surface fluxes and it is found that surface warming is the dominant factor in producing water mass changes, although changes in the freshwater cycle are important in the formation zone for Antarctic Intermediate Water. The coupled model has a low-amplitude, low-frequency (100-yr period) internal mode related to the anthropogenic fingerprint. Further observations are required to measure the amplitude of the internal mode as well as the anthropogenically forced mode.

Abstract

The spatial and temporal variability of the southwest Pacific Ocean is examined with the aim of describing the physical processes operating on interannual and decadal timescales. The study takes advantage of a new temperature atlas of the upper 450 m of the southwest Pacific Ocean, obtained from 40 000 bathythermograph profiles between 1955 and 1988. Rotated principal components analysis was used to filter the important spatial and temporal scales of temperature variability in the data. Three different analyses are presented. They include two intraocean analyses and a joint analysis of subsurface ocean temperature, sea level pressure, and surface winds.

The dominant El Niño mode describes the large vertical excursions of the thermocline in the western tropical Pacific in response to atmospheric forcing at a 3–6-month lag. More importantly, most of the retained modes, outside of the equatorial region, have time variations that correlate with El Niño. One ocean mode, with a spatial pattern representing sea surface temperature anomalies in the western Coral Sea (linked to the interannual migration of the South Pacific convergence zone), correlates significantly with (at the 99% level) and leads (by 3–6 months) the Southern Oscillation index (SOI), suggesting that sea surface temperature anomalies in this region may be a useful indicator for the onset of El Niño. A separate mode whose spatial pattern corresponds to the main oceanographic gyre also shows statistically significant temperature variations in phase with, or slightly leading, the SOI.

The main decadal variations occur in the midlatitudes, in the subtropical gyre, and in another mode associated with sub-Antarctic mode water (SAMW). The subtropical gyre warmed to a maximum in the mid-1970s and has been cooling since. In the SAMW a long-term warming of the upper 100 m of the southwest Tasman Sea is identified between 1955 and 1988. The depth-integrated warming in this region is found to be about 0.015°C yr⁻¹, representing a contribution to sea level rise, through thermal expansion, of about 0.3 mm yr⁻¹.

Abstract

Climatological monthly upper-ocean temperature anomalies from the annual mean in the subtropical southwest Pacific Ocean show a characteristic out-of-phase relationship between the mixed layer and the underlying water. The mixed layer temperature anomalies in the subtropical gyre and midlatitudes are consistent in the spatial distribution and phase expected from solar radiation. However, below the mixed layer, the temperature anomalies between 10°S and 30°S are coherent throughout the water column to 450-m depth and are almost 180° out of phase with the mixed layer temperatures. This pattern of temperature anomalies describes vertical movements of the thermocline more closely linked to the seasonal variations in the wind stress curl.

To test this hypothesis, a one-dimensional linear vorticity model was forced using the Hellerman and Rosenstein monthly wind stresses across the entire width of the South Pacific Ocean. This simple wind-driven model has considerable skill in predicting the gyre-scale pattern of change in the phase and amplitude associated with thermocline variations in the subtropical gyre. Experiments, varying the Rossby wave speed, showed that a better representation is achieved with speeds of 2 to 2.5 times that observed from altimeter observations. Overall, the inclusion of long Rossby waves appears to be a very important contribution to the amplitude of the thermocline depth variations in the southwest Pacific. Furthermore, this important Rossby wave contribution is supported by the large-scale anomaly patterns obtained from more sophisticated three-dimensional dynamical ocean models.

Abstract

Changes in atmospheric forcing can affect the subsurface water column of the ocean by three different mechanisms. First, warmed mixed-layer water that is subducted into the ocean interior will cause subsurface warming; second, the subducted surface water can be freshened through changes in evaporation and precipitation; and third, the properties at a given depth may be changed by the vertical displacement of isotherms and isohalines without changes of water masses. These vertical displacements of the water column can be caused either by changes in the rates of renewal of water masses or by dynamical changes (such as changes in wind stress). A method for analysing the subsurface temporal changes in hydrographic data is described in terms of these three processes: “pure warming,” “pure freshening,” and “pure heave.” Linear relations are derived for the relative strength of each process in terms of the observed changes of potential temperature and salinity in two different coordinate frames: (i) constant density surfaces, and (ii) isobaric surfaces.

Inverse methods are applied to three realizations of the SCORPIO section at 43°S in the Tasman Sea. These sections were obtained in 1967, and in the austral winter and summer of 1989 and 1990, respectively. This data is used to explore the relative strengths of surface warming, surface freshening, and heave of the water column. The six-month differences for this region show small changes in Sub-Antarctic mode water (SAMW) and are not characterized by any one process, whereas below the mode waters the observed differences are well described by the heave process. In contrast, the 23-year differences show significant changes in the properties of the water that flows into the Tasman Sea: SAMW (300-700 db) is well described by pure warming of near- surface waters, while the changes observed at the depth of the salinity minimum are consistent with pure freshening.

The observed changes in the interior of the ocean between adjacent seasons do not exhibit significant changes of water masses, consistent with the distance of this section from the outcrop region of the density surfaces of interest. For the 23-year differences, changed surface waters subducted into the ocean interior have sufficient time to influence the temperature-salinity correlations. The skill of our approach in discriminating between short-term changes (almost exclusively heave) and long-term changes associated with the subduction of changed surface waters is particularly encouraging. Although the observed changes could equally well be natural variability, they are qualitatively consistent with coupled numerical models of climate change in which surface waters are warmed and increased precipitation occurs south of the Sub-Antarctic Front.

Abstract

In the Indian Ocean subtropical gyre, historical temperature, salinity, and oxygen data with a median date of 1962 are compared with a hydrographic section taken at a mean latitude of 32°S in October–November 1987. Significant basinwide changes in all three hydrographic fields are observed below the mixed layer. On isobaric surfaces the main changes are (i) a warming of the upper 900 dbar of the water column with a maximum change in the sectional mean of 0.5°C, (ii) a freshening between 500 and 1500 dbar with a maximum freshening of 0.05 psu, and (iii) a pronounced decrease in oxygen concentration between 300 and 1000 dbar.

Examination of water mass properties shows that very significant water mass changes have occurred. On isopycnals subantarctic mode water (SAMW) and Antarctic Intermediate Water (AAIW) have freshened and cooled. Both of these water masses are on average deeper in 1987. Using the analysis of Bindoff and McDougall (1994), the changes of temperature at constant depth and at constant density are used to show that the water mass changes can most simply be explained by a surface warming in the source region of SAMW and by increased precipitation in the source region of AAIW.

The decrease in oxygen concentration can be explained simply by a slight slowing of the subtropical gyre allowing more time for biological consumption to decrease the oxygen concentration by water parcel translation from the formation area to the observation point. Estimates show that over the last 25 years there is an apparent decrease of the gyre spin rate of about 20% at the depth levels of SAMW; the estimated spin rate change decreases almost linearly with greater depth to zero at the oxygen minimum in Indian Deep Water (IDW). Below IDW the observed changes in oxygen concentration (and also the changes of temperature and salinity) are associated with the upward movement of isopycnals with no significant water mass change. The differences in temperature and salinity in the SAMW and AAIW are consistent with the relatively young age of these water masses inferred from CFC data.

Abstract

A near-global ocean model with resolution enhanced in the southern Indian Ocean has been spun up to seasonal equilibrium and then driven by NCEP–NCAR reanalysis 1 monthly mean forcings and Hadley SSTs over the period 1948–2002. The aim was to simulate changes in the subsurface properties observed in hydrographic surveys at 32°S in the Indian Ocean in 1965, 1987, and 2002. These surveys showed a zonally averaged cooling on isopycnals of 0.5° and 0.3°C in mode and intermediate waters between 1965 and 1987 and a warming of the mode water coupled with a continued cooling of the intermediate water between 1987 and 2002. The major changes in isopycnal depth and temperature modeled in this study were confined to the mode water and were qualitatively similar to those observed but concentrated in a lower density class and in the eastern half of the section. The dominant changes here were multidecadal, with maximum temperatures on the σ_θ = 26.7 kg m⁻³ isopycnal being reached in 1968 and minimum temperatures in 1990. The simulations showed a propagation of interannual anomalies toward the section from a region of deep late winter mixed layers in the southeast Indian Ocean within a period of several years. Surface temperatures in this region were lowest in the 1960s and highest in the late 1980s. Temperatures on isopycnals showed the opposite variation, consistent with SST having the controlling effect on mixed layer density and depth. Isopycnal depths within the mode water were strongly correlated with temperature, implying a redistribution of mode water density classes, the greatest volume of mode water being produced in a higher density class (σ_θ = 26.8–27.0 kg m^–3) during the period of cooler surface forcing in the 1960s and 1970s than during the warmer period following (σ_θ = 26.6–26.8 kg m^–3).

Abstract

Using optimal fingerprinting techniques, a detection analysis is performed to determine whether observed trends in Southern Ocean sea ice extent since 1979 are outside the expected range of natural variability. Consistent with previous studies, it is found that for the seasons of maximum sea ice cover (i.e., winter and early spring), the observed trends are not outside the range of natural variability and in some West Antarctic sectors they may be partially due to tropical variability. However, when information about the spatial pattern of trends is included in the analysis, the summer and autumn trends fall outside the range of internal variability. The detectable signal is dominated by strong and opposing trends in the Ross Sea and the Amundsen–Bellingshausen Sea regions. In contrast to the observed pattern, an ensemble of 20 CMIP5 coupled climate models shows that a decrease in Ross Sea ice cover would be expected in response to external forcings. The simulated decreases in the Ross, Bellingshausen, and Amundsen Seas for the autumn season are significantly different from unforced internal variability at the 95% confidence level. Unlike earlier work, the authors formally show that the simulated sea ice response to external forcing is different from both the observed trends and simulated internal variability and conclude that in general the CMIP5 models do not adequately represent the forced response of the Antarctic climate system.

Abstract

A multimodel comparison method is used to assess the sensitivity of Subantarctic Mode Water (SAMW) and Antarctic Intermediate Water (AAIW) formation to climate change. For the Intergovernmental Panel on Climate Change A2 emissions scenario (where atmospheric CO₂ is 860 ppm at 2100), the models show cooling and freshening on density surfaces less than about 27.4 kg m⁻³, a pattern that has been observed in the late twentieth century. SAMW (defined by the low potential vorticity layer) and AAIW (defined by the salinity minimum layer) warm and freshen as they shift to lighter density classes. Heat and freshwater fluxes at the ocean surface dominate the projected buoyancy gain at outcrop regions of SAMW and AAIW, whereas the net increase in the Ekman flux of heat and freshwater contributes to a lesser extent. This buoyancy gain, combined with shoaling of the winter mixed layer, reduces the volume of SAMW subducted into the ocean interior by a mean of 8 Sv (12%), and the subduction of AAIW decreases by a mean of 14 Sv (23%; 1 Sv ≡ 10⁶ m³ s⁻¹). Decreases in the projected subduction of the key Southern Ocean upper-water masses imply a slow down in the Southern Ocean circulation in the future, driven by surface warming and freshening. A reduction in the subduction of intermediate waters implies a likely future decrease in the capacity of the Southern Ocean to sequester CO₂.