Climate field reconstructions for the North Atlantic region of annual and seasonal resolution spanning CE 1241–1970

Abstract. The research area of climate field reconstructions has developed strongly during the past 20 years, motivated by the need to understand the complex dynamics of the earth system in a changing climate. Climate field reconstructions aim to build a consistent gridded climate reconstruction of different variables, often from a range of climate proxies, using either statistical tools or a climate model to fill the gaps between the locations of the proxy data. Commonly, large-scale climate field reconstructions covering more than 500 years are of annual resolution. In this method study, we investigate the potential of seasonally resolved climate field reconstructions based on oxygen isotope records from Greenland ice cores and an isotope-enabled climate model. Our analogue-type method matches modeled isotope patterns in Greenland precipitation to the patterns of ice core data from up to 14 ice core sites. In a second step, the climate variables of the best-matching model years are extracted, with the mean of the best-matching years comprising the reconstruction. We test a range of climate reconstructions, varying the definition of the seasons and the number of ice cores used. Our findings show that the optimal definition of the seasons depends on the variability in the target season. For winter, the vigorous variability is best captured when defining the season as December–February due to the dominance of large-scale patterns. For summer, which has weaker variability, albeit more persistent in time, the variability is better captured using a longer season of May–October. Motivated by the scarcity of seasonal data, we also test the use of annual data where the year is divided during summer, that is, not following the calendar year. This means that the winter variability is not split and that the annual data then can be used to reconstruct the winter variability. In particularly when reconstructing the sea level pressure and the corresponding main modes of variability, it is important to take seasonality into account, because of changes in the spatial patterns of the modes throughout the year. Targeting the annual mean sea level pressure for the reconstruction lowers the skill simply due to the seasonal geographical shift of the circulation modes. Our reconstructions based on ice core data also show skill for the North Atlantic sea surface temperatures, in particularly during winter for latitudes higher than 50∘ N. In addition, the main modes of the sea surface temperature variability are qualitatively captured by the reconstructions. When testing the skill of the reconstructions using 19 ice cores compared to the ones using eight ice cores, we do not find a clear advantage of using a larger data set. This could be due to a more even spatial distribution of the eight ice cores. However, including European tree-ring data to further constrain the summer temperature reconstruction clearly improves the skill for this season, which otherwise is more difficult to capture than the winter season.

Abstract. Two statistical methods are tested to reconstruct the interannual variations in past sea surface temperatures (SSTs) of the North Atlantic (NA) Ocean over the past millennium based on annually resolved and absolutely dated marine proxy records of the bivalve mollusk Arctica islandica. The methods are tested in a pseudo-proxy experiment (PPE) setup using state-of-the-art climate models (CMIP5 Earth system models) and reanalysis data from the COBE2 SST data set. The methods were applied in the virtual reality provided by global climate simulations and reanalysis data to reconstruct the past NA SSTs using pseudo-proxy records that mimic the statistical characteristics and network of Arctica islandica. The multivariate linear regression methods evaluated here are principal component regression and canonical correlation analysis. Differences in the skill of the climate field reconstruction (CFR) are assessed according to different calibration periods and different proxy locations within the NA basin. The choice of the climate model used as a surrogate reality in the PPE has a more profound effect on the CFR skill than the calibration period and the statistical reconstruction method. The differences between the two methods are clearer for the MPI-ESM model due to its higher spatial resolution in the NA basin. The pseudo-proxy results of the CCSM4 model are closer to the pseudo-proxy results based on the reanalysis data set COBE2. Conducting PPEs using noise-contaminated pseudo-proxies instead of noise-free pseudo-proxies is important for the evaluation of the methods, as more spatial differences in the reconstruction skill are revealed. Both methods are appropriate for the reconstruction of the temporal evolution of the NA SSTs, even though they lead to a great loss of variance away from the proxy sites. Under reasonable assumptions about the characteristics of the non-climate noise in the proxy records, our results show that the marine network of Arctica islandica can be used to skillfully reconstruct the spatial patterns of SSTs at the eastern NA basin.

Chapter 4 - Globality and Optimality in Climate Field Reconstructions from Proxy Data

When a permanent ice cap developed on Antarctica during the Eocene–Oligocene transition (EOT; ~34.44 to 33.65 million years ago (Ma)), Earth witnessed a transition from a greenhouse towards a glacially driven climate. Evidence of high-latitude cooling and increased latitudinal temperature gradients across the EOT has been found in both marine and terrestrial environments. However, the timing and magnitude of temperature change in the North Atlantic remains poorly constrained.Here, we used two independent organic geochemical palaeothermometers derived from (i) alkenones and (ii) Glycerol Dialkyl Glycerol Tetraether (GDGT) lipids, to reconstruct sea surface temperature (SST) evolution across the EOT from the southern Labrador Sea (Sites: ODP 647 and DSDP 112). In the Labrador Sea alkenones do not appear until the earliest Oligocene (both sites) while GDGT lipids (analysed in Site 647 only) provides a well-constrained temperature record across the EOT.  Our SST records provide the most detailed record for the northern North Atlantic through the 1 Myr leading up to the EOT onset, and reveals a distinctive cooling step of ~3 ºC (from 27 to 24 ºC), between 34.9 and 34.3 Ma, ~500 kyr prior to Antarctic glaciation. This cooling step, when compared visually to other SST records, is asynchronous across North and South Atlantic sites. This illustrates a considerable spatiotemporal variability in SST evolution in the northern sector of the North Atlantic and the Norwegian-Greenland Sea. Overall, the cooling step fits within a phase of general SST cooling recorded across sites in the North Atlantic in the 5 Myr interval bracketing the EOT.We used a modelling study (GFDL CM2.1) to try and reconcile the observation of pre-EOT cooling with the hypothesis that Atlantic Meridional Overturning Circulation (AMOC) switched on or intensified on the lead up to the EOT, which would be expected to have warmed the North Atlantic region. Results suggest that a reduction in atmospheric CO2 from 800 to 400 ppm may be sufficient to counter warming from an AMOC start-up. In the model, the AMOC start-up is initiated during closure of the Arctic–Atlantic gateway.While the model simulations applied here are not yet in full equilibrium, and the experiments are idealized, the results, together with the proxy data, highlight the heterogeneity of basin-scale surface ocean responses to the EOT thermohaline changes, with sharp temperature contrasts expected across the northern North Atlantic as positions of the subtropical and subpolar gyre systems shift in response to climatic and oceanic adjustments.

<p>We simulate the response of Asian summer climate to AMO-like (Atlantic Multidecadal Oscillation) sea surface temperature (SST) anomalies using the Intermediate General Circulation Model version 4 (IGCM4). Separate AMO SST patterns are obtained from seven Coupled Model Intercomparison Project phase 5 (CMIP5)/Paleoclimate Model Intercomparison Project phase 3 (PMIP3) global climate models, to explore the sensitivity of the atmospheric response to the SST pattern. Experiments are performed with seven individual and composited AMO SST anomalies globally, and over the North Atlantic Ocean only, for both the positive and negative phases of the AMO. During the positive AMO phase, a Rossby wave train propagates eastward, causing a high pressure anomaly over eastern China/Japan region, associated with a low level anomalous anticyclonic circulation along with warm and dry anomalies. In contrast, the mid-latitude Rossby wave train is less robust in response to the cold phase of the AMO. The circulation response and the associated temperature and precipitation anomalies are sensitive to the AMO SST anomaly patterns. The comparison between global SST and N Atlantic SST experiments indicates that midlatitude East Asian climate anomalies are forced from the North Atlantic region. However, global SST anomaly experiments show that the SST anomalies outside the North Atlantic region, but still associated with AMO, strongly influence South Asian climate as they either strengthen or reduce the precipitation. Experiments with different amplitudes of negative and positive AMO anomalies test the linearity of the response. Over a large region of South and East Asia, temperature has a linear response to the amplitude of North Atlantic SST anomaly associated with both positive and negative AMO conditions, but the precipitation response is nonlinear.</p>

Abstract. The effect of external forcings on atmospheric circulation is debated. Due to the short observational period, the analysis of the role of external forcings is hampered, making it difficult to assess the sensitivity of atmospheric circulation to external forcings, as well as persistence of the effects. In observations, the average response to tropical volcanic eruptions is a positive North Atlantic Oscillation (NAO) during the following winter. However, past major tropical eruptions exceeding the magnitude of eruptions during the instrumental era could have had more lasting effects. Decadal NAO variability has been suggested to follow the 11-year solar cycle, and linkages have been made between grand solar minima and negative NAO. However, the solar link to NAO found by modeling studies is not unequivocally supported by reconstructions, and is not consistently present in observations for the 20th century. Here we present a reconstruction of atmospheric winter circulation for the North Atlantic region covering the period 1241–1970 CE. Based on seasonally resolved Greenland ice core records and a 1200-year-long simulation with an isotope-enabled climate model, we reconstruct sea level pressure and temperature by matching the spatiotemporal variability in the modeled isotopic composition to that of the ice cores. This method allows us to capture the primary (NAO) and secondary mode (Eastern Atlantic Pattern) of atmospheric circulation in the North Atlantic region, while, contrary to previous reconstructions, preserving the amplitude of observed year-to-year atmospheric variability. Our results show five winters of positive NAO on average following major tropical volcanic eruptions, which is more persistent than previously suggested. In response to decadal minima of solar activity we find a high-pressure anomaly over northern Europe, while a reinforced opposite response in pressure emerges with a 5-year time lag. On centennial timescales we observe a similar response of circulation as for the 5-year time-lagged response, with a high-pressure anomaly across North America and south of Greenland. This response to solar forcing is correlated to the second mode of atmospheric circulation, the Eastern Atlantic Pattern. The response could be due to an increase in blocking frequency, possibly linked to a weakening of the subpolar gyre. The long-term anomalies of temperature during solar minima shows cooling across Greenland, Iceland and western Europe, resembling the cooling pattern during the Little Ice Age (1450–1850 CE). While our results show significant correlation between solar forcing and the secondary circulation pattern on decadal (r=0.29, p<0.01) and centennial timescales (r=0.6, p<0.01), we find no consistent relationship between solar forcing and NAO. We conclude that solar and volcanic forcing impacts different modes of our reconstructed atmospheric circulation, which can aid in separating the regional effects of forcings and understanding the underlying mechanisms.

Evidence is presented for a distinct Pattern Of ocean-atmosphere relationship associated with interdecadal variability in the North Atlantic region. Using a century of surface marine observations it is shown that middle- and high-latitude sea surface temperature (SST) display a long-term fluctuation with negative anomalies before 1920, and during the 1970s and 1980s. Positive SST conditions prevailed from about 1930 to 1960. The pattern of interdecadal SST variability is constructed by subtracting the average field during 15 cold years from that during a similar interval of warm years. The early-century warming and the more recent cooling display a similar spatial pattern. In both cases the pattern is basin scale and largely of one polarity, with maxima in the vicinity of Iceland, in the Labrador Sea, and northeast of Bermuda. The corresponding differences in surface atmospheric conditions are determined by averaging and subtracting, in the same manner, fields of sea level pressure (SLP) and surface winds. The results display a circulation anomaly in the middle of the ocean basin, centered at about 45°N and 35°W. In this midocean area, an anomalous cyclonic circulation prevailed during years with warm SST, and an anticyclonic anomaly dominated during years with cold SST. These circulation anomalies are strongest during the winter months. To emphasize the distinct nature of interdecadal variability, short-term, interannual variability is examined in a similar analysis. The resulting patterns display bands of zonally elongated centers of action that are negatively correlated with one another. These anomalies underlay similarly banded features in the zonal wind distribution. The mutual relationship between interannual fluctuations of SST and surface wind conditions suggests that the former are maintained by the latter through a local thermodynamic interaction. The interdecadal patterns of SST on one hand and SLP and winds on the other lack a similar coherent relationship. This fact, and the unique horizontal distribution of interdecadal SST anomalies, suggest that interdecadal variability may be governed by a basin-scale dynamical interaction between the large-scale oceanic circulation and the atmosphere.

The Gulf Stream plays an important role in North Atlantic climate variability on a range of timescales. The North Atlantic is notable for large decadal variability in sea surface temperatures (SST). Whether this variability is driven by atmospheric or oceanic influences is a disputed point. Long time series of atmospheric and ocean variables, in particular long time series of Gulf Stream position, reveal differing sources of SST variability on quasi‐decadal and multidecadal timescales. On quasi‐decadal timescales, an oscillatory signal identified in the North Atlantic Oscillation (NAO) controls SST evolution directly via air‐sea heat fluxes. However, on multidecadal timescales, this relationship between the NAO and SST changes, while the relationship between the NAO and Gulf Stream position remains consistent in phase and resonant in amplitude. Recent changes in the Gulf Stream Extension show a weakening and broadening of the current, consistent with increased instability. We consider these changes in the context of a weakening Atlantic overturning circulation.

Impacts of the North Atlantic Oscillation on sea surface temperature on the Northeast US Continental Shelf

Understanding the nature and causes of the variability associated with past warm, high p CO 2 climates presents a significant challenge to palaeoclimate research. In this paper we investigate the early Eocene climatic response in the North Atlantic region to forcing from an indirect effect of atmospheric methane (via polar stratospheric clouds (PSCs)), and we investigate the response of the climate system to forcing from a combination of orbital insolation changes and high atmospheric p CO 2 concentration. We find that sea surface temperatures (SSTs), sea ice extent, net surface moisture, continental runoff and upwelling in the North Atlantic Ocean are all sensitive to those forcing factors, and that the degree of sensitivity is a function of location and season. Our results suggest that high-latitude SST values can vary by as much as 20 °C during the winter season in response to precessional and polar cloud forcing, whereas in contrast summer temperature varies by 4 °C or less. Model predictions of net surface moisture balance also vary substantially with our prescribed forcing. There is a large difference in variability between the localized net surface moisture results and the mean North Atlantic Ocean results, which suggests that large-scale assumptions about past surface ocean salinities and seawater δ 18 O may need to be reassessed. According to model results, the influx of terrigenous material via continental runoff to the North Atlantic Ocean should be highly seasonal, with greatest runoff occurring in spring. Our model results also indicate that changes in wind-driven upwelling and in continental runoff on a precessional time scale should be seen in regions of the central North Atlantic.

Abstract. We present a coupled global climate model (CGCM) simulation, integrated for 1500 yr to quasi-equilibrium, of a stadial (cold period) within Marine Isotope Stage 3 (MIS 3). The simulated Greenland stadial 12 (GS12; ~44 ka BP) annual global mean surface temperature (Ts) is 5.5 °C lower than in the simulated recent past (RP) climate and 1.3 °C higher than in the simulated Last Glacial Maximum (LGM; 21 ka BP) climate. The simulated GS12 is evaluated against proxy data and previous modelling studies of MIS3 stadial climate. We show that the simulated MIS 3 climate, and hence conclusions drawn regarding the dynamics of this climate, is highly model-dependent. The main findings are: (i) Proxy sea surface temperatures (SSTs) are higher than simulated SSTs in the central North Atlantic, in contrast to earlier simulations of MIS 3 stadial climate in which proxy SSTs were found to be lower than simulated SST. (ii) The Atlantic Meridional Overturning Circulation (AMOC) slows down by 50 % in the GS12 climate as compared to the RP climate. This slowdown is attained without freshwater forcing in the North Atlantic region, a method used in other studies to force an AMOC shutdown. (iii) El-Niño-Southern Oscillation (ENSO) teleconnections in mean sea level pressure (MSLP) are significantly modified by GS12 and LGM forcing and boundary conditions. (iv) Both the mean state and variability of the simulated GS12 is dependent on the equilibration. The annual global mean Ts only changes by 0.10 °C from model years 500–599 to the last century of the simulation, indicating that the climate system may be close to equilibrium already after 500 yr of integration. However, significant regional differences between the last century of the simulation and model years 500–599 exist. Further, the difference between simulated and proxy SST is reduced from model years 500–599 to the last century of the simulation. The results of the ENSO variability analysis is also shown to depend on the equilibration.

<p>Multiyear to decadal predictability of the North Atlantic sea surface temperature (SST) is commonly attributed to buoyancy-forced changes of the Atlantic Meridional Overturning Circulation and associated poleward heat transport. Here we investigate the role of the wind stress anomalies in decadal hindcasts for the prediction of annual extratropical North Atlantic SST anomalies. A global climate model is forced by ERA-interim wind stress anomalies over the period 1979-2017. The resulting climate states serve as initial conditions for the decadal hindcasts. We find significant skill in predicting annual SST anomalies over the central extratropical North Atlantic with anomaly correlation coefficients exceeding 0.6 at lead times of 4 to 7 years. The skill of annual SSTs is basically insensitive to the calendar month of initialization. This skill is potentially linked to a gyre-driven upper-ocean heat content anomaly that leads anomalous SSTs by several years.</p>

Massive, long‐lived Siderastrea and Diploria corals are species commonly used for sea surface temperature (SST) reconstructions in the North Atlantic. However, they are rarely found to exceed 200 years in age. Thus, it is imperative to continuously develop alternative taxa for paleoreconstructions. Colpophyllia natans, a highly populous tropical North Atlantic coral, are known to grow large colonies, potentially containing environmental records spanning several hundreds of years. However, its low density and complicated architecture poses a challenge in extracting climate signals from this coral. This study presents the first monthly‐resolved climate calibration of Colpophyllia natans and validates its utility as a new paleoarchive, relative to Siderastrea siderea. Linear regressions of monthly and interannual coral Sr/Ca with instrumental SST reveal robust, significant relationships (p < 0.05), indicating that microsampling along a single thecal wall of C. natans allowed for robust climate reconstructions. Additionally, both corals capture similar SST variations (t‐test, p ≥ 0.05), which allowed for the generation of a single, composite interspecies SST record that correlates with instrumental SST even more strongly (p < 0.0001) than the individual corals. Mean annual and boreal summer interspecies SST correlate significantly with North Atlantic SST indices (p < 0.05), demonstrating the ability to capture regional, long‐term SST trends in the North Atlantic. Spatial correlation maps of boreal winter interspecies SST to instrumental SST and geopotential height anomalies reveal coherent spatial patterns linked to the North Atlantic Oscillation. Our findings suggest that Colpophyllia natans has enormous potential as a new paleoclimate archive for reconstructing temporal and spatial SST variability in the tropical Atlantic.

We analyse patterns of North Atlantic air–sea interaction in observations and model output throughout the annual cycle with the dual aims of validating natural variability in climate models, and quantifying seasonal predictability of the atmosphere and of ocean surface temperatures.The Met Office's HadCM3 coupled ocean–atmosphere model is shown to represent well the atmospheric forcing of sea surface temperature (SST) in the North Atlantic region through the action of latent‐heat fluxes. Winter ocean temperature anomalies in both observations and model appear to be preserved below the shallow summer mixed layer and to re‐emerge in the following autumn.We find observational evidence from the last half of the twentieth century for SST forcing of the atmospheric circulation in the North Atlantic region. Results validate our atmospheric model (HadAM2b) and confirm the levels of potential predictability of the North Atlantic Oscillation (NAO) that have been found in recent modelling studies. We suggest that the ultimate correlation skill of a winter NAO hindcast in this period is in the range 0.45 to 0.63.Analysis of observational and atmospheric model data from the first half of the twentieth century identifies very little predictability of the winter NAO. We suggest that the change in seasonal predictability is genuine and may be related to the strength of decadal oscillations.The ocean–atmosphere model, which incorporates a different atmospheric component (HadAM3), fails to show a causal link between North Atlantic temperatures and the atmospheric circulation even during periods where low‐frequency variability is as strong as in the recent observations.A brief observational investigation suggests that any implied predictability of the winter NAO based on a knowledge of South Atlantic SSTs depends on the existence of secular trends in the data. Tropical Pacific SSTs appear to influence the winter climate of the North Atlantic region, but not via trends in the NAO. © Crown copyright, 2002.

Holocene climate modes are identified by the statistical analysis of reconstructed sea surface temperatures (SSTs) from the tropical and North Atlantic regions. The leading mode of Holocene SST variability in the tropical region indicates a rapid warming from the early to mid Holocene followed by a relatively weak warming during the late Holocene. The dominant mode of the North Atlantic region SST captures the transition from relatively warm (cold) conditions in the eastern North Atlantic and the western Mediterranean Sea (the northern Red Sea) to relatively cold (warm) conditions in these regions from the early to late Holocene. This pattern of Holocene SST variability resembles the signature of the Arctic Oscillation/North Atlantic Oscillation (AO/NAO). The second mode of both tropical and North Atlantic regions captures a warming towards the mid Holocene and a subsequent cooling. The dominant modes of Holocene SST variability emphasize enhanced variability around 2300 and 1000 years. The leading mode of the coupled tropical-North Atlantic Holocene SST variability shows that an increase of tropical SST is accompanied by a decrease of SST in the eastern North Atlantic. An analogy with the instrumental period as well as the analysis of a long-term integration of a coupled ocean-atmosphere general circulation model suggest that the AO/NAO is one dominant mode of climate variability at millennial time scales.