Illuminating Intcal During the Younger Dryas

ABSTRACT As the worldwide standard for radiocarbon (14C) dating over the past ca. 50,000 years, the International Calibration Curve (IntCal) is continuously improving towards higher resolution and replication. Tree-ring-based 14C measurements provide absolute dating throughout most of the Holocene, although high-precision data are limited for the Younger Dryas interval and farther back in time. Here, we describe the dendrochronological characteristics of 1448 new 14C dates, between ~11,950 and 13,160 cal BP, from 13 pines that were growing in Switzerland. Significantly enhancing the ongoing IntCal update (IntCal20), this Late Glacial (LG) compilation contains more annually precise 14C dates than any other contribution during any other period of time. Thus, our results now provide unique geochronological dating into the Younger Dryas, a pivotal period of climate and environmental change at the transition from LG into Early Holocene conditions.


INTRODUCTION
The 2013 International Calibration Curve (IntCal13; Reimer et al. 2013a) is the "gold standard" for geochronological 14 C dating over the past~50,000 years. Based on 14 C measurements from tree rings, plant macrofossils, speleothems, corals, and marine sediments, the IntCal dataset is regularly updated towards higher resolution and precision (Reimer et al. 2013a). Accounting for 14 C/ 12 C variations in various proxy archives of the Holocene (e.g., tree rings) and even well into the Pleistocene (e.g., plant macrofossils, speleothems, and corals), reservoir standardization and intercomparisons between hemispheres and archives are key factors for the establishment and maintenance of IntCal (Reimer et al. 2013b).
Due to their resolution and independent cross-dating (Büntgen et al. 2018), tree-ring-based, annual to subdecadal 14 C measurements form the backbone of the calibration curve throughout the Holocene. The "Holocene Oak Chronology" by Becker (1993), together with the "Preboreal Pine Chronology" from Central Europe represent the longest absolutely dated tree-ring record securely extending back to 12,325 cal BP (PPC; Spurk et al. 1998;Friedrich et al. 1999Friedrich et al. , 2004Reinig et al. 2018a). A lack of subfossil wood from across the Northern Hemisphere (Reinig et al. 2018a), however, challenges IntCal precision during the preceding Younger Dryas (YD), so that the "Swiss Late Glacial Master *Corresponding author. Email: frederick.reinig@wsl.ch.
In order to better understand IntCal during the Younger Dryas, this study details the dendrochronological characteristics of 1448 new 14 C dates between~11,950 and 13,160 cal BP from 13 subfossil Scots pine trees (Pinus sylvestris L.) that were growing in Zurich, Switzerland. The exceptional quality (resolution) and quantity (replication) of this new LG 14 C compilation significantly improves 14 C wiggle-matching and dendrochronological cross-dating during the YD.

MATERIAL AND METHODS
At various construction sites in Zurich more than 400 subfossil pines were discovered and excavated between 1973 and 2013 (Kaiser et al. 2012;Reinig et al. 2018a). Cellulose-based 14 C measurements supported tree-ring width (TRW) dating of the material betweeñ 14,000-11,500 cal BP ( Figure 1; Reinig et al. 2018a). After sanding down the samples with up to 400 grain size sandpaper, at least two TRW radii per disc were measured and cross-dated using a LINTAB device (precision of 0.01 mm) and TSAP-Win software (both RINNTECH, Heidelberg). The TRW measurements where visually and statistically crossdated considering t-values (t BP ) and Gleichläufigkeit (Glk) indices (Baillie and Pilcher 1973) in TSAP-Win. Chronologies were established and checked with COFECHA (Holmes 1983).
A subset of 13 YD trees from the Zurich sites Gaenziloo, Krankenheim Wiedikon, and Binz (Reinig et al. 2018a) was selected and individual tree rings separated (Figure 1b). Highprecision 14 C AMS measurements were performed at the Laboratory of Ion Beam Physics, ETH-Zurich, on a "MIniCArbonDatingSystem" (MICADAS, Sookdeo et al. 2020). Wood of 2-4 consecutive rings was pooled, if material from a single ring was <20 mg. Holocellulose was extracted by the base-acid-base-acid method (Nemec et al. 2010), bleached and graphitized with an AGE system . The 14 C blanks, standards and references were continuously cross-checked in accordance with the ETH quality protocol (Sookdeo et al. 2019), ensuring the accuracy and comparability of the resulting 1448 14 C dates ( Figure 1a). The floating 14 C records derived from the individual trees were wiggle-matched to the Southern Hemisphere (SH) 14 C Kauri record (Hogg et al. 2016) using a χ 2 test (Sookdeo et al. 2020).

RESULTS AND DISCUSSION
The combination of TRW cross-dating and high-resolution 14 C wiggle-matching of 13 trees from Zurich closes the geochronological gap in the YD (Figure 1b). The samples are grouped into three distinct YD periods from~11,950 to 13,160 cal BP, during which good agreement is found within and between the TRW and 14 C measurements. Low sample size, juvenile growth disturbance and enhanced wood decay in the outer section of the samples (Reinig et al. 2018b), however, still challenges the establishment of a continuous, absolutely dated tree-ring chronology throughout the entire YD. The youngest trees are securely linked by dendrochronology and 14 C to the absolute chronology (12,314-11,863 cal BP), whereas the older trees yet remain floating.

Link between YD Swiss Trees and the PPC
The 14 C record of PPC was recently extended back to 12,049 cal BP by including previously undated trees from Breitenthal in southern Germany (PPC-Brei, Sookdeo et al. 2019), which allows the tentative cross-dating of two trees from the Zurich collection against the now improved German record. These two Swiss trees indicate good comparative t-values (t BP = 4.6, GLK = 66; OVL = 231) and synchronicity in both high frequency and longterm trends (see SM2). The 14 C measured tree "LAND0062" cross-dates with an overlap (OVL) of 187 years with PPC-Brei at t BP = 3.1 (Glk = 59), while BINZ0087 cross-dates to PPC-Brei at t BP = 2.9 (Glk = 57) over 99 years (Table 1). The YD-D ring width chronology, compiled from three trees from the site Krankenheim-Wiedikon in Zurich, Figure 1 Temporal distribution of the Swiss YD tree-ring records and corresponding 1448 14 C measurements (Sookdeo et al. 2020). High-resolution 14 C measurements of the Northern Hemisphere GAENALCH (light blue dots), YD-A (dark blue) and ZHYD-1 (orange dots) chronology and single tentative linked trees (yellow dots) wiggle-matched against the SH 14 C Kauri record (black dots, Hogg et al. 2016) now bridge the YD (a). An absolute dendrochronological link between the corresponding 14 C-measured trees (dashed bars) and Swiss TRW chronologies (filled bars) throughout the YD has not yet been established (b). However, the overall potential window for a temporal shift of the TRW records is minimal (± 8 yr, 2σ). Dashed lines indicate the Northern Hemispheric TRW gap between the absolutely dated PPC (Friedrich et al. 2004) and floating SWILM chronology (Kaiser et al. 2012). Sample replication (n T ) and total amount of 14 C dates (n 14C ) of each chronology is provided in parentheses. (c) Table of high-resolution 14 C-measured trees, outlining the trees' total number of rings and 14 C dates. (Please see electronic version for color figure.) Illuminating IntCal during the Younger Dryas 885 including 14 C data of tree KHWI0028 (Reinig et al. 2018a;Sookdeo et al. 2020), shows good visual and statistical coherence with BINZ0087 (Table 1; SM2). Cross-correlation of BINZ0087 improves to t BP = 4.8, when excluding the 77 outermost and disturbed rings from YD-D (YD-Dcut). LAND0004, the oldest 14 C dated tree of this absolutely dated chronology, is securely connected to YD-D (see SM2) and extends the chronology back to 12,314 cal BP. All established dendrochronological placements were found to be in accordance with the overlapping 389 high-resolution 14 C dates of the three trees (Figure 1c; Sookdeo et al. 2020). While the dendrochronological link between the Swiss and German pines is tentative, as the underlying Swiss sample size is low (n = 6), the good agreement with 14 C wiggle-matching (± 8 years (2σ); Sookdeo et al. 2020) securely links these tree-ring chronologies. This Swiss record represents with its 14 C support an important YD anchor to the PPC, for both 14 C wiggle-matching and future dendrochronological cross-dating.

Floating Swiss YD Chronologies
Two floating TRW chronologies and corresponding high-resolution 14 C measurements have been produced for the YD. ZHYD-1 is a chronology of 15 trees from the Zurich sites Gaenziloo (n = 4), Krankenheim Wiedikon (n = 3) and Binz (n = 8), reaching a cumulative length of 384 years. The six trees selected for 14 C measurements indicate good agreement to establish a robust mean curve, independent of their site of origin (see SM3). Interseries correlations between these trees range from t BP = 3.1 to t BP = 7.0, and all trees overlap by at least 131 years within the ZHYD-1 chronology ( Table 1). The TRW-based dating is reinforced by 635 high-resolution 14 C measurements (Figure 1c; Sookdeo et al. 2020), supporting the chronology particularly during periods of relatively weak replication. When developing ZHYD-1, two trees from Birmensdorf, originally included in the YD-B  (Kaiser et al. 2012), were removed, as their revised TRW measurements and additional 14 C dates proved their dating to be incorrect. Instead, the Binz material discovered in 2013 serves as the link between the oldest Gaenziloo and youngest Krankenheim Wiedikon trees in the chronology. The 14 C dates from ZHYD-1, wigglematched against the SH Kauri 14 C record (Sookdeo et al. 2020), suggested an TRW overlap of~20 years to LAND004 from the absolutely dated Swiss trees (Figure 1a,b). However, no valid dendrochronological link could be established as decay in the outermost rings of KHWI0030 and KHWI0031, as well as juvenile growth disturbances in LAND0004, challenge the extension of the absolute record.
The revised YD-A chronology from Kaiser et al. (2012), excluding decayed and disturbed TRW sections, represents the second floating chronology from Switzerland in the YD comprising of a total of eight trees and reaching a length of 199 years. The chronology is well-replicated over the first 140 years, but sample depth rapidly decreases thereafter. The youngest 35 years are represented by only one tree (GAEN0008). Due to the low replication, the TRW chronology displays increased variability over the most recent 60 years, limiting a secure cross-dating to ZHYD-1 and leading the obtained dendrochronological link to remain tentative. However, within YD-A the two high-resolution 14 C dated trees show good agreement to the YD-A mean curve (Table 1; see SM3). GAEN0007 (28 14 C dates) cross-dates at t BP = 3.4 to the YD-A mean (Glk = 65; OVL = 130), while GAEN0008 (132 14 C dates) is securely integrated in this chronology (t BP = 6.6, Glk = 71; OVL = 152). These placements are coherent with the 14 C wiggle-matching results (Figure 1c; Sookdeo et al. 2020). Based on 14 C wiggle-matching of the individual Swiss chronologies to the SH 14 C Kauri record and their obtained dendrochronological linkage, the combined record was shifted by seven decades, compared to the former YD-B position obtained through calibration using IntCal13 (Sookdeo et al. 2020). Nonetheless, additional dendrochronological analyses are necessary to conclusively verify the link between YD-A and ZHYD-1. The improved sample size in the YD may allow the cross-dating of additional 14 C dated YD tree-ring samples from Switzerland and Central Europe, facilitating temporal and spatial extensions of the record. The investigation of potential climate signals throughout the YD will hopefully become feasible.
Younger Dryas Transition GAEN0005, with 245 rings, is the youngest tree included in the Gaenziloo chronology (GAENALCH; n = 55; Kaiser et al. 2012) and represents the final portion of SWILM (Kaiser et al. 2012). With an overlap of 207 years, GAEN0005 is securely cross-dated into GAENLACH (t BP = 6.3; Glk = 64, Table 1; see SM4), and the now new 164 14 C measurements (Sookdeo et al. 2020) cover the transition from the Allerød into the YD (Figure 1c). The striking 14 C increase of GAEN0005 is confirmed by high-resolution 14 C measurements on LG pines from southern France (Capano et al. 2018). Nonetheless, no dendrochronological link could be established between the southern and northern Alpine sites. The distance between the sites, the varying climatic forcing, and the changing micro-site growth conditions might all contribute to their disagreement in TRW.
The long overlap (436 years), and high correlation (t BP = 11.1; Glk = 68; Table 1), demonstrate a secure link between GAEN0071 and the well-replicated GAENALCH chronology (see SM4). The now newly obtained 100 additional 14 C dates at 3-yr resolution from GAEN0071 (Sookdeo et al. 2020) will further improve the calibration towards the end of the Allerød (Figure 1c). In addition, the confirmed decreasing 14 C structure between Illuminating IntCal during the Younger Dryas 887 ~12,800-12,550 cal BP serves as an important period for wiggle-matching LG 14 C records globally. Their alignment along this distinct 14 C feature, obtained through the new highresolution data, will support the correction of dating discrepancies among various 14 C records. Accordingly, as the SWILM chronology is now shifted by 35 ± 8 years (2σ) after wiggle-matching to the SH Kauri 14 C record (Sookdeo et al. 2020), it can serve as a basis for additional high-resolution 14 C measurements back to~14,226 cal BP. In comparison to the dating precision of IntCal13 (± 20, 1σ), this marks a significant improvement. Crossdating and 14 C wiggle-matching of additional European and global tree-ring records can now be performed at higher temporal precision, which will eventually also improve dating accuracy throughout the Bølling and Allerød.

CONCLUSIONS
This study provides dendrochronological insight into 1448 new 14 C dates from 13 pine trees that were growing in Switzerland between~11,950 and 13,160 cal BP. Coherency between TRW cross-dating and 14 C wiggle-matching substantially improves the dating accuracy during the transition from the LG into the Early Holocene. Compared to their previous placement in IntCal13, the Swiss YD chronologies and the SWILM were shifted older by 70 and 35 years, respectively.