9:15 AM - 9:30 AM
[MIS14-12] Revision of pollen data from the Chiba Composite Section (CbCS) as proxy for terrestrial temperature around MIS 19
Keywords:Chiba Composite Section, pollen, MIS 19, temperature, proxy
The Chiba Composite Section (CbCS; 800-750 ka), which contains MIS 19, is socially known as the Global Stratotype Section and Point (GSSP) for the Chibanian Stage (774-129 ka), but has paleoclimatologically been focused as an analogue for the present interglacial (MIS 1). The present status of MIS 19 studies is summarized in Suganuma et al. (2021) and Head et al. (2021), which presented a preliminary pollen record (broadleaved/AP) together with paleomagnetic and planktonic foraminiferal δ18O records as the basis for ratification of the Chibanian GSSP.
As the broadleaved/AP (Arboreal Pollen) is usually equal to the ratio between thermophilic broadleaved trees and cold-loving conifers, it was used in Suganuma et al. (2018) as the proxy for terrestrial temperature. Basically, every stratum chosen as GSSP should be of marine origin. This means that fossil pollen tends to be the only method to know terrestrial environment from the GSSP sediments. However, any attempts to perform parallel pollen- δ18O analyses suffer from the following problems.
1) uneven resolution between the pollen and δ18O
2) disagreement between the pollen and δ18O due to local geographical reasons
The first problem was serious in the CbCS. As pollen analyses are highly human-powered technique, the resulting pollen records suffer from insufficient resolution compared with instrumental analyses.
Below we review contemporaneous records from the North Atlantic to southern Europe. The U1385 core, obtained from off the Iberian Peninsula (Goni et al. 2016), is located near the southern margin of IRD belt, which lies under similar environments to the NW Pacific that receives icebergs during stadials under changeable SST. The results from U1385 were consistent between their pollen and δ18O, and the variations in MIS 19a interstadials were also shared by C37:4-based freshwater pulse. The duration of MIS 19c was calculated as 13 kyr (787-774 ka), and some Bond Event-like signals were observed in their full interglacial interval.
On the contrary, the Ideale section of Montalbano (southern Italy) suffered from a significant time lag between their pollen and δ18O signals (Nomade et al. 2019). This phenomenon occurred in the Termination IX, and the lag of pollen behind their δ18O amounted to ~3000 years. A likely explanation for their pollen lag is forest migration. In the Mediterranean region, interglacial forest was replaced by open, dry steppe during glacial periods. A possible location of the forest refugia in full glacials is the Sicily Island, because it is known that the island was connected to the Italian Peninsula during MIS 2. Glacial refuge at the Sicily Island would produce a significant lag behind the forest recovery as Montalbano is distant from the Sicily Island.
These examples from southern Europe are suggestive when we consider the CbCS situation. Our new pollen record, obtained from the Chiba Composite Section, partly agrees with our δ18O in MIS 19a, but differs in MIS19c-b. In the new pollen record, the beginning of MIS 19c was as early as ‘790 ka’, and the end of MIS 19c was as late as 770 ka. Moreover, the main part of MIS 19c interglacial consists of a massive plateau. This means that the duration of MIS 19c is calculated to as long as 20,000 years.
We note that these results from pollen do not agree with the δ18O record published in Haneda et al. (2020). The disagreement cannot be explained by the migration lag because the forest expansion precedes the deglacial warming in the parallel δ18O. A likely explanation is that the broadleaved/AP may not be suitable for a temperature proxy.
It is possible that the broadleaved forest may contain both thermophilic trees and somewhat cool-loving elements (e.g., Cyclobalanopsis, Lepidobalanus, and/or Fagus sp). Subdivision of the ‘broadleaved’ forest to smaller botanical groups, or introducing a quantitative paleotemperature reconstruction would reduce the disagreement between terrestrial and marine realms.
As the broadleaved/AP (Arboreal Pollen) is usually equal to the ratio between thermophilic broadleaved trees and cold-loving conifers, it was used in Suganuma et al. (2018) as the proxy for terrestrial temperature. Basically, every stratum chosen as GSSP should be of marine origin. This means that fossil pollen tends to be the only method to know terrestrial environment from the GSSP sediments. However, any attempts to perform parallel pollen- δ18O analyses suffer from the following problems.
1) uneven resolution between the pollen and δ18O
2) disagreement between the pollen and δ18O due to local geographical reasons
The first problem was serious in the CbCS. As pollen analyses are highly human-powered technique, the resulting pollen records suffer from insufficient resolution compared with instrumental analyses.
Below we review contemporaneous records from the North Atlantic to southern Europe. The U1385 core, obtained from off the Iberian Peninsula (Goni et al. 2016), is located near the southern margin of IRD belt, which lies under similar environments to the NW Pacific that receives icebergs during stadials under changeable SST. The results from U1385 were consistent between their pollen and δ18O, and the variations in MIS 19a interstadials were also shared by C37:4-based freshwater pulse. The duration of MIS 19c was calculated as 13 kyr (787-774 ka), and some Bond Event-like signals were observed in their full interglacial interval.
On the contrary, the Ideale section of Montalbano (southern Italy) suffered from a significant time lag between their pollen and δ18O signals (Nomade et al. 2019). This phenomenon occurred in the Termination IX, and the lag of pollen behind their δ18O amounted to ~3000 years. A likely explanation for their pollen lag is forest migration. In the Mediterranean region, interglacial forest was replaced by open, dry steppe during glacial periods. A possible location of the forest refugia in full glacials is the Sicily Island, because it is known that the island was connected to the Italian Peninsula during MIS 2. Glacial refuge at the Sicily Island would produce a significant lag behind the forest recovery as Montalbano is distant from the Sicily Island.
These examples from southern Europe are suggestive when we consider the CbCS situation. Our new pollen record, obtained from the Chiba Composite Section, partly agrees with our δ18O in MIS 19a, but differs in MIS19c-b. In the new pollen record, the beginning of MIS 19c was as early as ‘790 ka’, and the end of MIS 19c was as late as 770 ka. Moreover, the main part of MIS 19c interglacial consists of a massive plateau. This means that the duration of MIS 19c is calculated to as long as 20,000 years.
We note that these results from pollen do not agree with the δ18O record published in Haneda et al. (2020). The disagreement cannot be explained by the migration lag because the forest expansion precedes the deglacial warming in the parallel δ18O. A likely explanation is that the broadleaved/AP may not be suitable for a temperature proxy.
It is possible that the broadleaved forest may contain both thermophilic trees and somewhat cool-loving elements (e.g., Cyclobalanopsis, Lepidobalanus, and/or Fagus sp). Subdivision of the ‘broadleaved’ forest to smaller botanical groups, or introducing a quantitative paleotemperature reconstruction would reduce the disagreement between terrestrial and marine realms.