Extreme drought event in AD 1637‒1643 in North China: New insight from pollen records in Kaifeng City

Long-lasting droughts usually result in water shortage and famine and even hinder the progress of human civilization. This paper presents the first study to identify the “Chong Zhen drought” event during the Late Ming Dynasty (AD 1637‒1643) in a region with alluvial sediment. Using AMSC dating, grain size and historical documents, we determined the formation age of the sediment. Pollen records at 5‒7.8 m, with the highest values for mesoxerophyte and 10 xerophyte taxa such as Chenopodiaceae, Nitraria and Ephedra pollen, were found to provide evidence for the “Chong Zhen drought” event. Other pollen percentages were almost the lowest value of the entire core. However, sediment at 5‒6.7 m also had a very high value of Pinus, which was mainly caused by fluvial transport and depositional processes. Chenopodiaceae, Artemisia and Asteraceae pollen are always transported and deposited together with coarser sediment during the flood running stage, while pollen with thinner extine and air-sacs are always deposited in slow flow or hydrostatic conditions 15 during the flood falling stage. In addition, pollen transported by wind from local and regional areas deposited on the ancient ground are always mixed with the sediment during the flood running stage with coarser until it covered by next flood. Our work helps shed light on the interpretation of the ancient vegetation and past climate based on pollen in alluvium, but it is important to make allowances for flood processes and sedimentary facies.


Introduction
Understanding the process and mechanisms of climate change impacts and the resulting disaster events can enhance human response to global change (Costanza et al., 2007).Numerous studies have found close and complicated correlations between climate change and historical social events, including, amongst others, population variation and migration, economic fluctuation, social harmony and crisis, and even dynastic transition (Haug et al., 2003;Zhang et al., 2008;Buckley et al., 2010;Patterson et al., 2010;Tol and Wagner, 2010;Büntgen et al., 2011;Medina-Elizalde and Rohling, 2012;Pringle, 2012).In particular, drought is a recurring major natural hazard that has plagued civilizations through time and remains the "world's costliest natural disaster" (Herweijer et al., 2006).When drought conditions persist for a number of years and spread across vast areas, these impacts become devastating.According to historical records, the vicinity of Kaifeng was flooded many times.The earliest flood was recorded as early as 221 BC.Thereafter, floods were apparently not recorded for more than a thousand years until the beginning of the Song Dynasty, which was founded Kaifeng as its capital city.The Yellow River began to swing in the vicinity of Kaifeng in AD 1128, and floods became more prevalent.Six larger-scaled flood events happened respectively in the years of AD 1410, AD 1461, AD 1478, AD 1489, AD 1642 and AD 1841 inundated Kaifeng City during the Ming and Qing Dynasties, which not only damaged the city severely but also changed the surface sediment and natural landscape significantly (Cheng, 2003;Liu, 2009).
A 9-m-long sediment core (JM) was taken at 34°49′2.17″N, 114°18′29.31″E, elevation 74 m asl in April 2012 using a DPP-100 piston corer (9-cm in internal diameter).The sediment cores were transported back to the laboratory in sealed PVC pipes and described and subsampled shortly thereafter.A total of 78 samples for physico-chemical and palynological analysis were obtained at 10−20 cm intervals from the JM core.

Physico-chemical analysis
The grain size was measured at 10-cm intervals using a laser diffraction particle size analyzer (Mastersizer 3000, Malvern Co. Ltd., UK).Sample pretreatment included (1) adding H2O2 to remove organic matter and soluble salts, (2) using diluted 1 N HCl to remove carbonate and (3) using sodium hexametaphosphate to disperse aggregates.The wet color of the core was described using Munsell's soil color charts.

Pollen analysis
The 100−300 g samples (dry weight) were treated chemically according to the standard acetolysis procedure (Faegri and Iversen, 1989).Thirty-six percent HCl was used to dissolve calcareous minerals, 10% NaOH and 30% HF were used to remove the humic components and siliceous materials, respectively.Following the above, the pollen and spores were concentrated by heavy liquid (ZnCl2, density 2.0) to separate them from undigested minerals.Finally, cellulose and humic debris were removed by acetolysis.
Pollen residues were stored in glycerol, mounted on microscope slides, and then scanned at 200× or 400× magnification with a Leica DM5500 microscope.Identification was aided by Wang et al. (1995), with comparison to the modern pollen reference collections.For each sample, a minimum of five slides were examined, and more than 400 terrestrial pollen grains were counted.Pollen percentages were calculated based on the sum of trees, shrubs and terrestrial herbs.Pollen diagram was Clim.Past Discuss., doi:10.5194/cp-2016Discuss., doi:10.5194/cp- -122, 2016 Manuscript under review for journal Clim.Past Published: 13 December 2016 c Author(s) 2016.CC-BY 3.0 License.drawn using the TILIA software.Pollen zonation was performed using a stratigraphically constrained incremental sum of squares cluster analysis according to percentage data, facilitated by CONISS.

Radiocarbon dating
Radiocarbon dating took place using accelerator mass spectrometry (AMS) at the AMS Dating Laboratory at Peking University (Beijing, China).The 14 C date was calibrated to calendar years before the present with the program CALIB Rev. 5.0.1 using the IntCal04 calibration data set (Reimer et al., 2004).For this study, one sample (Lab number: BA130509) of plant macrofossil was selected and dated to 410 ± 20 BP, the calibrated age is AD 1460 ± 30 (7.85 m depth).

Sedimentary cycle and chronology
Grain size is a major indicator used to divide sedimentary cycles.The sand-clay ratio of a core can reflect hydrodynamic changes.The greater the ratio, the stronger the hydrodynamics (Lecce and Pavlowsky, 2004).Therefore, sand-clay ratio can be used to conduct the division of sedimentary cycles.In this paper, wavelet periodic analysis of the sand-clay ratio was carried out with MATLAB 7.0 software (Fig. 2).As can be seen in Fig. 2, the core exhibit cycle changes from three types of scales (i.e., 1 m, 3 m and 6 m).Specifically, there are 1.5 sedimentary cycles in the 6 m scale, wherein the depth of 0-6 m is a complete cycle and 6-9 m for a half cycle (negative anomalies section).There are 2.5 cycles in the 3 m scale, wherein the depth of 1.8-4 m and 4-7.8 m are complete cycles and 7.8-9 m is a half cycle (negative anomalies section).There are more cycles in the 1 m scale, but only the depth of 0.5-1.8m, 4-5.8 m and 5.8-7.5 m cycles are more stable.In addition, the other half of the cycle occurs at the depth of 0-0.5 m (positive anomalies section).
In March 2003, the Kaifeng Municipal Archaeological Team found a Song Dynasty cultural layer in the depth of 10-11.3m in northeastern of the JM core (Liu, 2009).This archaeological discovery indicated that below the depth of 10 m was the ground surface of the Song Dynasty, and the 9-m-long core of this research was deposited after the Song Dynasty.
According to historical records (Cheng, 2003), the flooding of the Yellow River has affected Kaifeng City six times since the early 15th century, including floods occurred in AD 1410, AD 1461, AD 1478, AD 1489, AD 1642 and AD 1841 respectively.Additionally, the latest two Yellow River floods in Kaifeng City occurred in AD 1642 and AD 1841 (Cheng, 2003;Liu, 2009), corresponding to the sediments below the surface of 4-7.8 m and 1.8-4 m sedimentary cycles, respectively (Fig. 2).Combined archaeological and dating data indicate that the 8 m depth, the grayblack paleosol, was the ancient ground surface for human activities after the Yellow River flood in AD 1489.
In the mid-20th century, the government introduced the Yellow River warping method for the purpose of controlling the hazards of sand storms around Kaifeng City.This formed a sludge layer of a certain thickness, which corresponded to the 0.5-1.8m sedimentary cycle.The 0-0.5 m interval of the core is the modern artificial accumulation layer.

Comprehensive analysis of JM core
A comprehensive information schematic of the JM core is shown in Fig. 3.The sediment texture is composed mainly of silt loam, silt and sandy loam, and the sediment color is primarily brown and yellow-brown.The dark-brown and dark-gray colored layers, which mainly appeared in 2.5-2.7 m and 4.8-5.2m depth of the core, indicate short-term hydrostatic reduction conditions.As seen from the vertical curve of sand-clay ratio (Fig. 3), the curve of cycle 2 fluctuated larger than cycle 1, probably due to temporal and spatial variation in the water flow during the flooding of the Yellow River in AD 1841 (Liu, 2009).However, an obvious secondary cycle is not found on the wavelet analysis chart because of the small sedimentary thickness.The flooding of the Yellow River in AD 1642 lasted three years, with widely varied hydrodynamic conditions, and only a mutation in 5.8 m depth was revealed on the wavelet analysis chart.
The core can be divided into four pollen assemblage zones and four subzones according to the changes in the relative abundances of terrestrial pollen and the CONISS cluster analysis results (Fig. 4).
Pollen Zone 2 (5-7.8m) corresponds to the sediment cycle s3-1, and the pollen assemblage is still dominated by herb component (38-97%, mean value 75%).It should be mentioned that mesoxerophytes taxa (e.g., Chenopodiaceae) percentages have the highest values of the entire core.This zone can be further divided into two subzones.
Pollen Zone 3 (2-5 m) corresponds roughly to the sediment cycle s3-2.Compared with subzone 2b, this zone is distinctively characterized by dramatic decreases in the values of Chenopodiaceae and Pinus and also by dramatic increases in Ranunculaceae (1-43%, mean value 24%) and Picea (up to 10%) pollen percentages.Other pollen taxa, such as Betula and Moraceae, are slightly lower, while Quercus, Gramineae and Cyperaceae are slightly higher than that of subzone 2b.
Artemisia percentages are almost unchanged from the previous subzone.

Pollen response to the "Chong Zhen drought" event during the Late Ming Dynasty
The pollen assemblage of the JM core reveals an absolute dominance of herbaceous plant pollen (e.g., Chenopodiaceae, Gramineae, Artemisia, Cyperaceae and Ranunculaceae).Studies on alluvial pollen suggest the existence of hydrodynamics and taphonomic process would have influenced the deposition of pollen into the alluvial system and its preservation in the sedimentary record, as has been inferred to occur in alluvial sedimentary deposits from other regions (Hall, 1985;Fall, 1987;Xu et al., 1996;Work et al., 2005;Rojo et al., 2012).According to these studies, pollen input and preservation in alluvial sedimentary archives could be controlled by pollen source, fluvial transport, depositional processes, sedimentary facies and post-depositional changes.
Chenopodiaceae and Asteraceae (including Artemisia) pollen, with relatively thicker pollen extine, are always transported and deposited together with coarse sand (Sangster and Dale, 1961;Hall, 1989).Thus, Chenopodiaceae pollen in the JM core was mainly derived from local areas and the middle and upper reaches of the Yellow River.Studies on modern pollen show that Chenopodiaceae pollen with over-represented and can be instructing regional ecology (Liu et al., 1999;Li et al., 2000;Cao et al., 2010;Zhao and Li., 2013).Studies of relationships between modern pollen and vegetation show that the Pinus pollen is over-represented in the pollen record, and pine trees are locally present when the percentages are greater than 30% (Li et al., 2000;Cao et al., 2010;Zhang et al., 2014).In the JM core, abundant Pinus pollen occurs in subzone 2b, with the Therefore, the vertical variation of pollen assemblage in JM core primarily reflects the different periods of vegetation changes in the areas near the middle and upper reaches of the Yellow River during the last 500 years.The pollen of the JM core is composed of local pollen, regional pollen and extra-regional pollen while regional pollen is the most essential component.Local pollen, which is deposited mainly during the intermittent period, is a reflection of local vegetation.
Human impacts on the natural vegetation, especially on arboreal taxa, were widespread and intensifying during the last 500 years (Li et al., 2006).As a result, the herb pollen assemblage in the JM core can better reflect dry and wet changes.Pollen zone 1 (7.8-9 m), which was mainly deposited in approximately AD 1460, has a high value of Chenopodiaceae pollen percentage (mean value 35%), suggesting an obvious drought climate during the Middle Ming Dynasty period.Numerous studies show that the AD 1440s may be the beginning of a drought climate during the Ming Dynasty, and the drought degree in AD 1450-1490 is next to the Late Ming Dynasty (Yang et al., 2014;Zheng et al., 2014;Lee et al., 2015;Hao et al., 2016).This is consistent with the pollen record in this study.
In pollen zone 2 (5-7.8m), the highest values of mesoxerophyte taxa such as Chenopodiaceae, Taraxacum and Asteraceae pollen percentage are found, accompanied by the appearance of xerophyte taxa such as Nitraria and Ephedra pollen (Fig. 5).
However, other pollen percentages are almost the lowest value of JM core (Fig. 4).Sediments of this interval are associated with the the Yellow River flood in AD 1642, and its high content of xerophyte and mesoxerophyte taxa correspond well with the "Chong Zhen drought" event during the late Ming Dynasty.

Reasons for the differences between pollen subzone 2a and 2b
Different from pollen subzone 2a (6.7-7.8 m), subzone 2b (5-6.7 m) is distinctively characterized by a high percentage of tree pollen, especially Pinus pollen, reaching the highest value of the entire core (Fig. 4).The reasons for this phenomenon may be the following two aspects.(1) The flood depositions of the Yellow River mainly brought herb pollen from the middle reaches and floodplains.During the "Chong Zhen drought" period, the decline in vegetation coverage in the middle reaches of the Yellow River may have led to an enhancement of wind erosion, largely as a result of a marked increase of surface pollen migration by the wind.In addition, the precipitation variability tends to increase when the climate is relatively dry, and more pollen were moved by water erosion from surface soil to the river.During a flood of the Yellow River there may be two flood sedimentary layers: one was deposited during flood running period, and another was deposited during flood falling stage.During the flood running stage, as the river water has higher flow velocity, only the pollen with thicker extine and greater gravity (e.g., Chenopodiaceae, Artemisia and Asteraceae) can be deposited with coarser sediment, however, during the flood falling stage, the river flow rate declined obviously, so the pollen with thinner extine (e.g., Moraceae) and air-sacs (e.g., Pinus) can be deposited with fine sediment (usually with clay) (Sangster and Dale, 1961;Traverse and Ginsburg, 1966;Hall, 1989;Xu et al., 1996;Brown et al., 2007;Albert, 2014).Pollen subzone 2b is precisely in the negative anomalies section of the cycle, also indicating the power of the flooding water diminished and even became hydrostatic deposition (Fig. 4).( 2) As the fine sediment (clay) was deposited during the flood falling stage, it will be the new ground surface after the flood.The pollen and sedimentary materials from local and regional areas transported by wind can be deposited on the ground until it was covered by next flood.In particular, Pinus pollen can be easily transported long distances by wind due to its air-sacs.Additionally, dry air caused by intense wind facilitates anther cracking and pollen dispersion (Shang et al., 2009).So, the pollen from local and regional areas transported by wind usually mixed with the sediment during the flood running stage with coarser sediment.

Conclusions
From our study of AMS 14 C dating, historical records and wavelet analysis of sand-clay ratio in the JM core, we found that the sediment in the 0.5-1.8m, 1.8-4 m and 4-7.8 m intervals in the core were formed by irrigation in the mid-20th century, flooding of the Yellow River in AD 1841 and AD 1642, respectively.Pollen records in the 5-7.8 m interval provide evidence for the "Chong Zhen drought" event, wherein mesoxerophyte taxa such as Chenopodiaceae, Taraxacum and Asteraceae in the 6.7-7.8 m interval reached the highest values, accompanied by the appearance of xerophyte taxa such as Nitraria and Ephedra pollen.Other pollen percentages are almost the lowest value of the entire core.However, sediment in the 5-6.7 m interval also has a very high value of Pinus, which was mainly caused by fluvial transport and depositional processes.Chenopodiaceae, Artemisia and Asteraceae pollen are always transported and deposited together with coarser sediment during the flood running stage, while pollen with thinner extine and air-sacs are always deposited in slow flow or hydrostatic conditions during the flood falling stage.Then, the surface sediment developed into ancient ground.Pollen transported by wind from local and regional areas deposited on the ground are always mixed with the sediment during the flood running stage with coarser until it covered by next flood.
Clim.Past Discuss., doi:10.5194/cp-2016-122,2016 Manuscript under review for journal Clim.Past Published: 13 December 2016 c Author(s) 2016.CC-BY 3.0 License.highest value reaching 27%, suggesting that Pinus pollen was mostly transported by floods from the middle and upper reaches of the Yellow River.

Figure 1 .
Figure 1.(a) Elevation map of the geographical location of the sampling site in China and (b) Henan province; (c) Map showing the location of the JM core.

Figure 2 .
Figure 2. Contour map of the sand-clay ratio by using wavelet analysis of the JM core.

Figure 3 .
Figure 3. Comprehensive information schematic of the JM core.In writing the notation of a wet color, the order is hue, value, and chrome.For instance, a color of 10YR in hue, 4/ in value, and /3 in chrome, the notation is 10YR 4/3.

Figure 4 .
Figure 4. Pollen percentage diagram of the JM core.

Figure 5 .
Figure 5. Pollen percentage of drought-tolerant taxa in the JM core.