Ocean records and responses to orbital forcing

Sea-surface-temperature reconstructions

Reconstructions of surface layer temperatures of the high-latitude North Atlantic and Nordic Seas through the Holocene have been obtained from various proxy methods. Transfer functions have been produced based on data from alkenones,

Figure 5.4 Annual mean response in a simulation with the Massachusetts Institute of Technology Earth System model of intermediate complexity (Dutkiewicz et al. 2005) using 6 ka boundary conditions. (a) 6 ka sea-surface temperature (SST) increase compared with pre-industrial (ad 1850) SSTs. (b) Annual mean insolation change relative to pre-industrial (upper panel), and annual mean SST change relative to pre-industrial.

Longitude

Longitude

Latitude
Latitude

coccolith species, diatom, and forminifers, as well as planktonic foraminiferal oxygen isotope records and some Mg/Ca records, for parts of the Holocene. Multiple proxies have been analyzed in some cores, enabling a direct comparison of the output from the different methods without the errors of temporal correlation. Core MD95-2011, obtained from beneath the main flow of warm Atlantic water towards the Arctic, is a good example (Figure 5.5). The Holocene temperature development reconstructed by the different proxy methods is very different, both in terms of the amplitude of millennial and shorter time-scale variability, and for overall Holocene trends. The proxies cluster into two distinct categories: those

Figure 5.5 Comparison between sea-surface temperature reconstructions from: (a) alkenones (Calvo et al. 2002); (b) diatoms in MD95-2011 (Andersen et al. 2004; Koc personal communication); (c) planktonic stable oxygen isotope data from NEAP4K (Hall et al. 2004); (d) stable oxygen isotope data and (e) planktonic census data from MD95-2011 (Andersson et al. 2003; Risebrobakken et al. 2003).

16 14 12108 6

MD95-2011

MD95-2011

1312 11

NEAP4K"
MD95-2011

T"

T"

MD95-2011

3.5 —i—i—i—|—i—i—i—|—i—i—i—|—i—i—i—|—i—i—i—|—i—i—i—| 12 000 10 000 8000 6000 4000 2000 0 Age years BP

with a distinct Holocene thermal maximum (diatom-based SSTs, alkenones); and foraminiferal-based proxies (transfer functions, species distribution, oxygen isotope data corrected for ice-volume effects), which do not show an early to mid-Holocene maximum. Instead, these proxies show the opposite, a warming trend towards the late Holocene and more high-amplitude changes at century-millennial time-scales after about 4 ka. The close correspondence between the independent foraminiferal-based proxies indicates that the general pattern of the foraminifer-based temperature series is robust. Likewise, there is no a priori reason to distrust the fidelity of the diatom- and alkenone-based reconstructions. They are consistent with summer season orbital forcing and consistent with other surface ocean data from the region sampled from different cores, from ice-core data, ice-core borehole data, and terrestrial records.

We must assume that both categories (i.e. with and without a pronounced early to mid-Holocene thermal optimum) of temperature histories are valid and reflect different aspects of the Holocene thermal evolution of the high-latitude ocean. To understand why the difference has developed, the seasonality of the forcing, the habitat of the biological proxy indicators, and the seasonality of the vertical structure of the upper layers of the high-latitude ocean must be considered. The main character of the orbital forcing of the early versus late Holocene is a strong high-latitude early to mid-Holocene positive thermal anomaly during the summer season, due to the combined effects of tilt and precession (see Crucifix this volume, Figure 4.4). A significant, yet smaller, negative anomaly occurred in the winter season, but is more confined to low and mid-latitudes.

The result is an enhanced seasonality in the forcing and in the expected ocean response as compared with the late Holocene. The discrepancy between the two categories of temperature series might be explained as a response to the seasonality of the forcing. Yet it remains to be argued why the foraminiferal-based records should respond to the winter-time or annual mean forcing and not the summertime forcing, when we know that the primary foraminifer production season is in the spring and summer. It is, however, well known that different species of foraminifers have their main habitat at different depths in the ocean surface layer. In contrast to diatoms and alkenone producing algae, which photosynthesize and live in the euphotic zone of the upper 50 m of the ocean, planktonic foraminifers are zooplankton and are often found at the depths of the main food sources near the thermocline at the base of the surface layer. Risebrobakken et al. (2003) showed that both the left- and right-coiling varieties of Neogloboquadrina pachyderma share this pattern, i.e. showing a lack of evidence for the Holocene thermal optimum, yet there is a difference between their O-isotope values, with the left-coiling variety indicating colder temperatures than the right-coiling variety.

The seasonal forcing of surface layer temperatures provides strong contrasts between the upper 50 m and deeper layers. Temperature data from Ocean Weather Station M in the Norwegian Sea (Figure 5.6) shows that the summer solar insolation warms and stratifies the surface layer, imposing a strong heating in the upper 50 m, whereas the temperature of the thermocline is unrelated to the summer season, and instead reflects the winter-time ventilation of the whole surface layer when the summer season stratification breaks down. Thus at 100 m, summer temperatures are set by the winter-time situation and reflect this or the annual mean temperatures. As can been seen in the hydrographic data (Figure 5.6), a temperature proxy derived from organisms that thrive below approximately 50 m in these waters will not register any spring or summer-time warming, even though they have their largest standing stock during the summer season. Enhanced solar

Figure 5.6 (a) Seasonal and annual water temperature averages for the last 50 years from Ocean Weather Ship Mike in the upper 100 m of the water column. The averages are based on the temperature records presented in (b). (From (Nyland et al. 2006.)

Temperature (°C)

Temperature (°C)

1 I I 1 jgL 1 I 1 I I I I

i i i I^I i i i

■ M

w m

44 f

' /

Annual

: - ■

Winter

: ▲

Spring

/ " '

Summer

-

1950 1960 1970 1980 1990 2000

1950 1960 1970 1980 1990 2000

1950

1960 1970 1980 1990 2000

1950

1960 1970 1980 1990 2000

1950 1960 1970 1980 1990

1950 1960 1970 1980 1990

2000

insolation, as was the case for the early to mid-Holocene, would exacerbate this feature, stabilize the surface layer in the warm season, and produce a stronger contrast between the euphotic zone proxies and those derived from thermocline species.

For the thermal maximum this implies that the strong solar insolation maximum of the early to mid-Holocene originating from summer season orbital forcing did not manifest itself below the euphotic zone, and was only captured by the temperature reconstructions derived from phytoplankton with an upper surface ocean habitat. The difference between the alkenone- and diatom-based SST reconstructions is most likely due to the calibration of the methodologies. The diatom transfer functions are calibrated to surface August temperatures, whereas alkenones are mainly produced in the spring season and calibrated to the mean annual temperatures of the uppermost surface layer. Those foraminiferal species (e.g. Globigerina bulloides) that normally calcify above the summer-time thermocline would be expected to show the same pattern as phytoplankton-based proxies. For foraminiferal-based transfer functions the higher degree of stratification will give a mixed response depending on the blend of species with surface or thermocline habitat preference. In the Nordic Seas, thermocline species dominate the

foraminiferal assemblage and the resulting transfer function temperatures are consistent with the O-isotope-based records (Andersson et al. 2003; Risebrobakken et al. 2003). As a consequence of these findings, the wide use of N. pachyderma, as well as foraminiferal transfer functions in paleoceanographic studies from high-latitude areas, should be qualified with the information that the reconstruction is primarily that of temperature in the thermocline, and thus reflects winter-time or annual mean temperatures.

Dynamical responses or radiative forcing only?

So far we have seen that the seasonal character of orbital forcing explains the thermal maximum and also the temperature reconstructions that do not display any early to mid-Holocene optimum. The difference between the data-sets reflects seasonality anomalies. It has also been inferred that the early to mid-Holocene thermal maximum reflects enhanced advection of heat, either in the form of advection of warmer waters from the North Atlantic, by stronger advection of Atlantic waters towards the Arctic or by enhanced activity of the Westerlies (e.g. a mechanism akin to that described by Blindheim et al. 2000).

The MD95-2011 data rule out the existence of an advection of North Atlantic warm water anomalies to the north. The core location lies underneath the main flow path of Atlantic waters towards the Arctic. At present, the thickness of the Atlantic water layer in this area is 400-600 m, much thicker than the suggested depth distribution of the proxy temperature recorders from the site (Figure 5.5). The advective time constant of the present Norwegian Sea from south to north is in the order of several years (Skagseth 2004). Therefore, an advective temperature anomaly would be expected to override seasonal changes and be manifested down to several hundred meters water depth. The thermal maximum is, however, only manifest in the upper 50 m, which implies that it is attributable to the radiative forcing from the orbital configuration. This is consistent with the model results of Liu et al. (2003) who found a difference between the surface layer and thermocline waters in the model's response to orbital forcing. It remains to be tested if this is a robust feature of model behavior in the Nordic Seas.

Ruling out advective transport of warm water anomalies as an explanation, it is possible that oceanic heat fluxes towards the Arctic, larger than later in the Holocene, could have occurred. One possible case is that a larger heat flux could have been contained in stronger along-slope currents, similar to the positive mode of the North Atlantic Oscillation (NOA) observed in recent decades (Blindheim et al. 2000; Skagseth 2004). The available data cannot register such a situation with stronger flow of warm waters towards the Arctic and without any change in the mean temperature of the Atlantic water masses. There is, at present, no reliable kinetic flow proxy data-set from which to evaluate this aspect. Nevertheless, it seems likely that the polar amplification of the thermal maximum is mainly due to the polar amplification of orbital forcing and to sea-ice albedo feedbacks due to the summer insolation anomaly, as also noted from the PMIP2 models depicted in

Crucifix (this volume, Figure 4.3). This is also consistent with the shorter duration and early timing of the maximum temperatures near the present sea-ice margin, which implies a connection to the time of the strongest insolation anomaly along the sea-ice margin. Further to the south the maximum temperatures occurred later, after 8 ka (Figure 5.5), and continued at times when temperatures close to the present sea-ice margin had started to fall (Sarnthein et al. 2003; Kim et al. 2004; Hald et al. personal communication). Thus the polar amplification does not appear to be mainly due to advection of oceanic heat.

0 0

Post a comment