Holocene temperature trends in the Northern Hemisphere extratropics

Show full item record


Title: Holocene temperature trends in the Northern Hemisphere extratropics
Author: Zhang , Yurui
Contributor: University of Helsinki, Faculty of Science, Department of Geosciences and Geography
Department of Earth Sciences, VU University Amsterdam
Thesis level: Doctoral dissertation (article-based)
Belongs to series: URN:ISSN:1978-7911
Abstract: As the latest epoch of the Earth’s history, the Holocene is commonly defined as the last 11.7 ka BP (hereafter referred to as ka) and represents a new phase, encompassing the time span of human civilization. The last deglaciation lasted well into the Holocene, implying that the early Holocene was characterized by a large-scale reorganization with transitions in various components of the climate system. Studying the Holocene can provide insights into how the climate system functions, apart from the theoretical contributions to climate history itself. We first conducted sets of simulations with different combinations of climate forcings for 11.5 ka and for the entire Holocene to investigate the response of the climate–ocean system to the main climate forcings. In particular, two possible freshwater flux (FWF) scenarios were further tested considering the relatively large uncertainty in reconstructed ice-sheet melting. Moreover, we compared four Holocene simulations performed with the LOVECLIM, CCSM3, FAMOUS and HadCM3 models by identifying the regions where the multi-model simulations are consistent and where they are not, and analysing the reasons at the two levels (of the models’ variables and of the model principles and physics) where mismatches were found. After this, these multi-model simulations were systematically compared with data-based reconstructions in five regions of the Northern Hemisphere (NH) extratropics, namely Fennoscandia, Greenland, North Canada, Alaska and high-latitude Siberia. Potential uncertainty sources were also analysed in both model simulations and proxy data, and the most probable climate histories were identified with the aid of additional evidence when available. Additionally, the contribution of climate change, together with forest fires and human population size, to the variation in Holocene vegetation cover in Fennoscandia was assessed by employing the variation partitioning method. With effects of climate forcings, including variations in orbital-scale insolation (ORB), melting of the ice sheets and changes in greenhouse gas (GHG) concentrations, the climate shows spatial heterogeneity both at 11.5 ka and over the course of the Holocene. At 11.5 ka, the positive summer ORB forcing overwhelms the minor negative GHG anomaly and causes a higher summer temperatures of 2–4 °C in the extratropical continents than at 0 ka. The ice-sheet forcings primarily induce climatic cooling, and the underlying mechanisms include enhanced surface albedo over ice sheets, anomalous atmospheric circulation, reduced the Atlantic Meridional Overturning Circulation (AMOC) and relevant feedbacks. In particular, the most distinct feature is a thermally contrasting pattern over North America, with simulated temperatures being around 2 °C higher than those at 0 ka for Alaska, whereas over most of Canada, temperatures are more than 3 °C lower. The geographical variability of simulated temperatures is also reflected in Holocene temperature evolution, especially during the early Holocene, as constant Holocene cooling in Alaska contrasts with strong early-Holocene warming (warming rate over 1 °C kyr-1) in northern Canada. The early-Holocene climate is sensitive to the FWF forcings and a brief comparison with proxy records suggests that our updated FWF (FWF-v2, with a larger FWF release from the Greenland ice sheet and a faster FWF from the Fennoscandian Ice sheet (FIS)) represents a more realistic Holocene temperature scenario regarding the early-Holocene warming and Holocene temperature maximum (HTM). Comparison of multiple simulations suggests that the multi-model differences are spatially heterogeneous, despite overall consistent temperatures in the NH extratropics as a whole. On the one hand, reasonably consistent temperature trends (a temporal pattern with the early-Holocene warming, following a warm period and a gradual decrease toward 0 ka) are found over the regions where the climate is strongly influenced by the ice sheets, including Greenland, N Canada, N Europe and central-West Siberia. On the other hand, large inter-model variation exists in the regions over which the ice sheet effects on the climate were relatively weak via indirect influences, such as in Alaska, the Arctic, and E Siberia. In these three regions, the signals of multi-model simulations during the early Holocene are incompatible, especially in winter, when both positive and negative early-Holocene anomalies are suggested by different models. These divergent temperatures can be attributed to inconsistent responses of model variables. Southerly winds, surface albedo and sea ice can result in divergent temperature trends across models in Alaska, Siberia and the Arctic. Further comparisons reveal that divergent responses in these climate variables across the models can be partially caused by model differences (e.g. different model physics and resolution). For instance, the newly adopted formulation of the turbulent transfer coefficient in CCSM3 causes an overestimated albedo over Siberia at 0 ka, which leads to a stronger early-Holocene warmth than in other models. Moreover, the relatively simplified sea ice representation in FAMOUS probably leads to overestimated sea ice cover in the Arctic Ocean. The coarse vertical resolution in LOVECLIM might also introduce strong responses in atmospheric circulation over Alaska. From the perspective of climate features, the transient feature of the early-Holocene climate driven by the retreating ice sheets also influences the inter-model comparisons, as this transient feature induces a large degree of uncertainty into the FWF forcing. Comparisons of multiple model results with compiled proxy data at the sub-continental scale of NH high latitudes (i.e. Fennoscandia, Greenland, north Canada, Alaska and Siberia) reveal regionally-dependent consistencies in Holocene temperatures. In Fennoscandia, simulations and pollen data suggest a summer warming of 2 °C by 8 ka, although this is less expressed in chironomid data. In Canada, an early-Holocene warming of 4 °C in summer is suggested by both the simulations and pollen results. In Greenland, the magnitude of early-Holocene warming of annual mean ranges from 6 °C in simulations to 8 °C in δ18O-based temperatures. By contrast, simulated and reconstructed summer temperatures are mismatched in Alaska. Pollen data suggest 4 °C early-Holocene warming, while the simulations indicate 2 °C Holocene cooling, and chironomid data show a stable trend. Meanwhile, a high frequency of Alaskan peatland initiation before 9 ka can either reflect a high temperature, high soil moisture content or large seasonality. In high-latitude Siberia, simulations and proxy data depict high Holocene temperatures, although these signals are noisy owing to a large spread in the simulations and to a difference between pollen and chironomid results. On the whole, these comparisons of multi-model simulations with proxy reconstructions further confirm the Holocene climate evolution patterns in Fennoscandia, Greenland and North Canada. This implies that the Holocene temperatures in these regions have been relatively well established, with a reasonable representation of Holocene climate in the multiple simulations and a plausible explanation for the underlying mechanisms. However, the Holocene climate history and underlying mechanisms in the regions of Siberia and Alaska remain inconclusive. Variation partitioning revealed that climate was the main driver of vegetation dynamics in Fennoscandia during the Holocene as a whole and before the onset of farming. Forest fires and population size had relatively small contributions to vegetation change. However, the size of the human population became a more important driver of variation in vegetation composition than climate during the agricultural period, which can be estimated to have begun at 7–6 ka in Sweden and 4–3 ka in Finland. There is a clear region-dependent pattern of change caused by the human population: the impact of human activities on vegetation dynamics was notably higher in south Sweden and southwest Finland, where land use was more intensive, in comparison with central Sweden and southeast Finland. This thesis investigates the climate responses to the main forcings during the Holocene through various approaches, which has potential implications for the interactions between ice sheets and the climate, the Holocene climate history and current global change. The atmosphere-ocean system was sensitive to the FWF forcing during the early Holocene, implying that existing uncertainties in reconstructions of ice-sheet dynamics can be constrained by applying different freshwater scenarios via a comparison with proxy data. The Holocene climate history in most of the Northern Hemisphere extratropics is relatively well established, especially in regions that were strongly influenced by ice sheets. The implications of our investigation (on the transient early-Holocene) for the current global change are twofold. First, regional heterogeneity of the climate responses implies that regional differences should be taken into account when adapting to the current global change. Second, apart from the different scenarios of GHG forcing, inter-model comparison would be a good option to reduce model-dependency in estimation of the future climate.
URI: URN:ISBN:978-951-51-2929-1
Date: 2017-09-29
Rights: This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.

Files in this item

Total number of downloads: Loading...

Files Size Format View
Holocene.pdf 4.887Mb PDF View/Open

This item appears in the following Collection(s)

Show full item record