C54A-01
High Arctic Temperature Variations During the Past Five Millennia: a Varve Based Record From Lower Murray Lake, Nunavut, Canada
Sediments in Lower Murray Lake, northern Ellesmere Island, Nunavut Canada (81°21'N, 69°32'W) contain annual laminations (varves) which provide a record of sediment accumulation spanning the past 5000+ years. Annual mass accumulation rates in Lower Murray Lake were compared to instrumental climate data, long term records of climatic forcing mechanisms and other regional paleoclimate records which indicate that lake sedimentation is positively correlated with regional melt season temperatures driven by radiative forcing. Lower Murray Lake mass accumulation rates were positively correlated with mean July 600 m free air temperatures at the two nearest permanent weather stations at Alert and Eureka, producing r2 values of 0.61 and 0.50, respectively. Consequently, we suggest that sediment mass accumulation in Lower Murray Lake is dominantly influenced by July temperatures in the upper watershed which have a controlling influence on snowmelt, streamflow and sediment transport into the lake. The lowest rates of sediment accumulation and by inference the coldest periods of the record occurred around varve year 1800 AD and prior to ~4200 varve years ago. In contrast, periods of increased sedimentation, and by inference the warmest conditions, occurred in the 12th, 14th, and 20th centuries, and throughout the middle portion of the record, approximately 1000 to 4200 varve years ago. By calibrating the mass accumulation record in terms of July temperatures we were able to produce a quantitative estimate of the range of past temperature variations at Lower Murray Lake (standard error +/- 1.04 °C). The temperature reconstruction suggests: (1) recent temperatures are ~2.6 °C higher than temperature minima observed during the Little Ice Age, ca. 1800 AD, (2) maximum temperatures during the past 5200 years exceeded modern values by ~0.6 °C, (3) minimum temperatures observed approximately 4900 varve years before present were ~3.5 °C colder than recent conditions.
C54A-02
High Resolution Reconstructions of Arctic Sea Ice for the Last 2 Millennia
Polar sea ice, by its contributions to oceanic processes, is a key component of the Earth's climate system. As such, in the current context of climate change, it is essential to improve our knowledge of its historical fluctuations both at a high spatial and temporal resolution. In the current project, we are investigating the potential to use chemical biomarkers of sea ice associated diatoms to serve as a proxy of sea ice cover in Polar Regions. We have demonstrated the potential of a C25 mono-unsaturated Highly Branched Isoprenoid alkene (IP25) as a proxy of sea-ice cover in the Arctic and have performed a detailed analysis of sediment cores MD99-2275 and MD99-2263 collected from the North Icelandic Shelf. The high sedimentation rates together with well documented occurrences of volcanic tephras have enabled us to document sea ice cover for the last two millennia at an unprecedented sub-decadal resolution (2-5 yrs) and to make comparisons with historical data of the past sea ice extending back to the early days of Icelandic colonization (ca. 1080 BP). We provide evidence for a succession of abrupt climate changes in Iceland during the second part of the last millennium. Most recently, we carried out detailed analyses of short sediments cores collected across a wide range of Arctic locations and observed downcore IP25 concentrations entirely consistent with previous paleoclimatic reconstructions.
C54A-03
Differential Response Times of Climate Proxies During the last 2000 Years Reconstructed from Icelandic Lake Systems
Rapid sediment infill (ca. 4.5 m ka-1) of Haukadalsvatn, a lake in northwest Iceland, provides a high- resolution archive of climate and environmental change over the past 2 ka. Comparisons with instrumental climate data indicate that biogenic silica (BSi) in this record reflects warm April-May temperatures, whereas total organic carbon (TOC) peaks indicate increased influx of carbon from eolian derived soil carbon following periods of cold summers and dry, windy winters. The proxy-based temperature reconstructions show a broad interval of warmth through Medieval times, but this warmth is punctuated by multi-decadal cold intervals. The transition into the Little Ice Age (LIA) occurred in steps, with initial summer cooling 1250-1300 AD, and a larger drop in summer temperatures between 1450 and 1500 AD, followed by two cold spikes in the early 1600s and mid 1800s AD. Glaciofluvial sedimentation dominates Hvitarvatn, a glacial lake in central Iceland. Ice rafted detritus (IRD) exhibits two peaks during the LIA and both occur ca. 100 years after the cold spikes of the Haukadalsvatn record. Although there is an apparent lag in response time of the different proxies, both lakes record four cooling periods in the last 2000 years that coincided with periods of severe explosive volcanism. Longer records from both lakes indicate that the LIA was by far the coldest interval since the early Holocene. At Hvitarvatn glaciers attained their maximum LIA positions, and their most extensive since deglaciation, between 1600 AD and the mid 19th century. Although severe soil erosion in Iceland is frequently equated with settlement, our reconstructions indicate that soil erosion began several centuries before settlement, whereas for several centuries after settlement, when summer temperatures were relatively high, there was little or no soil erosion. Only during the transition into and during the LIA did soil erosion become a major feature of the record.
C54A-04
Late Holocene Water Mass Change in the Norwegian Sea Caused by Different Ocean- Atmosphere Circulation Patterns
There is common consensus that the Holocene climate history of the polar North was strongly tied to the insolation change on one the hand and the specific post-deglacial water mass evolution on the other. Using deep-sea sediment records we have investigated two crucial areas of the Norwegian Sea (Arctic Front; Voring Plateau) in order to understand the natural variability of oceanic-atmospheric change in this area since the middle Holocene. The information available from this longer time scale allows better insight for predictive purposes, since these records would then provide a longer time frame within which to evaluate any natural variability. We analyzed different foraminiferal species for O-isotope analyses and interpreted the planktic foraminiferal assemblage variations in combination with records of ice-rafted detritus (IRD) >150μm. It is shown that surface temperatures started to decrease at the Arctic Front after 6 ka, concomitant with the occurrence of IRD. This cooling trend continued into the Little Ice Age (LIA) when highest IRD input is noted. At the Voring Plateau, relatively stable and warm conditions are still recognized between 2.5 and 1 ka, in both planktic and benthic O-isotopes. Although variability among certain foraminiferal species would indicate some surface changes, the abundance of the polar species N. pachyderma (s) increased from 30% before 1 ka to 70% during the LIA. This increase is associated with highly variable isotope values through the entire water column (up to 1‰) and the sudden occurrence of basaltic IRD, presumably from Iceland. We interpret the records of the last 2.5 ka, and in particular the time of the LIA, to be the result of a major change in overall ocean-atmosphere circulation (from NAO+ to NAO-) which forced colder water masses and sea ice far into the eastern Norwegian Sea.
C54A-05
Evaluating the potential for forced signals in regional Arctic climate using simulations with the Community Climate System Model
In the context of the last millennium, volcanic eruptions and solar variability are the two leading causes of forced climate variability, and together these factors explain much of the decadal to centennial-scale variance of reconstructed Northern Hemisphere surface temperatures. A compilation of new and earlier published climate proxy data, however, shows that there are strong regional deviations in temperature from the hemispheric average, and that there is typically no simple relationship between timeseries of the forcing and timeseries of climate proxies at a given location. Several authors have proposed that systematic changes in the atmospheric circulation may explain a portion of such regional differences. In the Northern Hemisphere, the major centers of action are the Aleutian Low and the Icelandic Low. Here, we use the fully coupled version 3 of the Community Climate System Model to investigate the response of these centers of action to prescribed changes in solar and volcanic forcing. We use an ensemble approach that emphasizes the forced response, and further improves upon earlier studies by a treatment of the spectral character of solar radiation change with associated adjustments in stratospheric ozone concentrations in addition to a simple step change in the solar constant. We also consider the differing response to volcanism at high and low latitudes. Our results suggest that while there are systematic changes in circulation, the regional response in variables commonly of interest to proxy studies—surface temperature and precipitation—depends strongly on the state of sea-ice, which would be different in different epochs of the last millennium. These experiments will help to distinguish patterns of forced and internal climate system variability that can be used as benchmarks for comparison with syntheses of new Arctic and sub-Arctic proxy data, such as those underway for the Arctic 2K lakes project and the PAGES Arctic 2K project.
C54A-06
Unprecedented Faunal Changes Around 2000 Years ago in the Arctic, Evidence From the Shelf of Northern Svalbard
Our knowledge of the magnitude, timing, and spatial patterns of climate anomalies during the late Holocene in the Arctic is still poorly documented. Particularly, the lack of high-resolution marine proxy records prevent us from obtaining better knowledge of the spatial and temporal variability of the inflow of the warm and saline North Atlantic Current (NAC) into the Arctic, which presently is an important source for heat and moisture for the European Arctic region. Here we present a marine proxy record that presently underlies the subsurface flow of Atlantic water into the Arctic at the northern Svalbard margin (Hinlopen trough) and spans the last 17,500 years with a detailed record of the last 2000 years. In order to trace the variability of the subsurface flow of Atlantic water we investigated the content of bottom dwelling microfossils (benthic foraminifera). The Holocene climate variability is clearly shown by the changes in relative abundance of the two species Cassidulina reniforme and Elphidium excavatum. They show an early Holocene dominance of C. reniforme interpreted to reflect strong influence of Atlantic water followed by a gradual increase of E. excavatum that started at c. 7500 cal yr BP and reached maximum values at c. 4000 cal yr BP suggesting diminished influence of Atlantic water. Surprisingly, the abundance of E. excavatum starts to decrease around 2000 cal yr BP which continues until present time. As E. excavatum starts to decline other species reach their highest occurrences in the late Holocene compared to the last 17,500 years. These late Holocene species are associated with less harsh environmental conditions and Atlantic-derived waters. We therefore relate the faunal changes at c. 2000 cal yr BP to an enhanced influence of the subsurface inflow of Atlantic water into the Arctic region. Comparison to other marine proxy records from the North Atlantic region indicate by the use of biogenic and sedimentologic proxies that changes did occur in the Arctic Ocean, the Irminger Current and the East Greenland Current at about 2000 cal yr BP whereas proxy records underlying the southwestern branch of the North Atlantic Current flowing into the Nordic Seas show no pronounced changes. The reason for the differences in the spatial response between the regions in the North Atlantic and high-latitude records is, however, not clear.
C54A-07
Three Ice Core Records From Svalbard To Reveal Important Differences In Regional Climate Development During The Past 700 Years
The Svalbard Archipelago, being situated in a climatically sensitive area at the turning point of the North Atlantic current, experiences substantial climate variations both in the spatial and temporal domains. A detailed study of the complicated nature of interactions within the climatic system is however hampered by shortness of the existing instrumental records from the area. It promotes the use of records from ice cores for making paleoclimatic reconstructions. This study of the regional climate variability spanning some 700 years is based on analysis of δ18O data from three ice cores drilled in different locations on Svalbard during the past decade. Despite retrieved from the relatively low-altitude ice caps and therefore subject to summer melt, the cores from Lomonosovfonna (1255 m asl, drilled in 1997, covers some 800 years), Holtedahlfonna (1150 m asl, in 2005, 400 yrs) and Austfonna (750 m asl, in 1999, 800 yrs ) were shown to contain a variety of climate records. The annual snow accumulation rate at the drilling sites of the order of 0.4 m w.eq. ensures a subannual resolution of the cores back to the 16th century. All three considered records show similar variations on century scales, with a gradual cooling associated with the onset of the Little Ice Age in the 1700s and a gradual warming starting in the mid-1800s with the most rapid change in the beginning of the 20th century. The features of variability on the shorter time scales, as well as the magnitude of the LIA-associated decrease in δ18O series, differ between the cores. It highlights the regional difference in the local climate between the core sites which is especially pronounced in the latitudinal direction. The two δ18O records of Lomonosovfonna and Holtedahlfonna from the western Spitzbergen display nearly identical variations throughout the period of overlap. At the same time the Austfonna series from the eastern Svalbard shows much stronger LIA-associated cooling with a maximum offset between the records during 1780-1860. This is interpreted in terms of the extended sea ice cover during the LIA. This hypothesis is supported by a coincidence with a period of increased continentality on Svalbard during the LIA and generally colder North Atlantic, as evidenced by a deuterium excess record from Lomonosovfonna.
C54A-08
A hierarchical Bayesian analysis of a high latitude, high resolution multiproxy data set over the last 500 years
A number of different temperature sensitive proxy records (tree ring widths and densities, lake sediment cores, ice cores) are available at high latitudes. These proxy records extend back further in time than the modern instrumental data set, but are sparser in space, and noisier. Efforts to reconstruct the temporal evolution of the temperature field should make efficient use of all available information. To this end, we have developed a hierarchical Bayesian model to assimilate the incomplete (in space and time) instrumental and proxy data sets to arrive at estimates of the temperature field through time, as well as the uncertainty in these estimates. The Bayesian model has three layers: the process level describes the evolution of the true surface temperatures as a multivariate AR(1) process with spatially correlated innovations; the data level specifies the relationships between the measurements (proxies and instrumental) and the true field values; and the prior level specifies diffuse prior distributions for all unknown parameters to provide closure to the scheme. By specifying the parametric form of the spatial covariance, and then estimating the actual parameters from the data, our approach allows for spatial infilling – we can predict, with uncertainties, the temperature at locations where there are no observations. Multiple draws from the posterior result in spatially and temporally complete ensembles of the temperature evolution compatible with the data, and the model assumptions. Probability distributions for various statistics can be estimated from this ensemble, from simple measures like the time series of spatial means to more exotic quantities like the maximum decadal average over a specific region. The integrated, Bayesian approach facilitates the propagation of uncertainties, avoids regression dilution (the temporal variance of the proxy based portion of the reconstruction is not biased low), and allows for a thorough residual analysis to check that the model assumptions are met by the data. We present surface temperature reconstructions from the analysis of a Northern high latitudes multiproxy data set, using a simple formulation of the hierarchical model. In particular, we compare reconstructions based on each proxy type to that based on the full multiproxy dataset to investigate the relative strengths and weaknesses of the different data types. Finally, we indicate aspects of the analysis model that can be improved and expanded upon, and point to the general utility of hierarchical Bayesian approaches to climate reconstruction problems.