Holocene peat humification and carbon dynamics in the Westerlies-influenced Northwest China

Understanding peat carbon dynamics in the past is of significance, given the uncertainties as to whether there will be an increase or a reduction in carbon as a result of future climate change. Studies of peat carbon dynamics have primarily been conducted in monsoon-influenced China. However, data relating to carbon dynamics in peat deposits has not yet been investigated in Westerlies-influenced Northwest China (NWC). In this study, the Holocene carbon accumulation rate (CAR) is explored at the Tuolehaite peat core, with a mean rate of growth of 2.4 yr mm−1, 0.4 mm yr−1, in the high-elevation Altai Mountains within NWC. Its CAR shows a decreasing trend, ranging from 4.8 to 68.8 g C m−2 yr−1, with a mean of 28.0 g C m−2 yr−1 since the Holocene epoch. Comparisons of the CAR in the Westerlies-influenced NWC with that in the monsoon-influenced Qinghai-Tibetan Plateau, Northeast China, and South China, reveal that the Holocene CAR trend in relatively high-elevation peat deposits (e.g. Tuolehaite Peat and Hongyuan Peat) is opposite to that found in relatively low-elevation peat deposits (e.g. Hani Peat and Dahu Peat). Different driving factors (temperature and precipitation) of CAR could be responsible for these opposing trends. To be specific, temperature is the main driving factor influencing the CAR in relatively high-elevation peat deposits, whereas precipitation is the key driving factor controlling the CAR in relatively low-elevation peat deposits. Our work indicates that comprehensive investigations into peat CARs in both Westerlies-influenced and monsoon-influenced regions contribute to an understanding of the peat CAR pattern in China as a whole.


Introduction
Wetlands are a special type of ecosystem, occurring between the aquatic ecosystem and the terrestrial ecosystem, having the highest carbon density and the fastest carbon accumulation rate (CAR) (Liu et al 2006, Lv 2008. It covers only 6%-8% of the land area in the world, and its carbon storage has been estimated to be 500 ± 100 Pg C since the Last Glacial Maximum (Yu 2011). This is equivalent to approximately one-third of the total C pool on earth (Mitra et al 2005, Limpens et al 2008. The carbon accumulations of wetlands are dependent on the balance between primary production and decomposition, determined by temperature and precipitation (Charman 2002, Friedlingstein et al 2006, Cai et al 2010. In the context of global warming, the question of whether wetland will turn from carbon accumulation to carbon emission has attracted increased attention (Roulet et al 2007, Mcguire et al 2009, Oh et al 2020.
Large spatial-scale estimations of carbon dynamics have been performed in northern peatland, including western Siberia, Europe and North America (Vitt et al 2000, Jones and Yu 2010, Gorham et al 2012. The CAR from peat cores and their controlling factors have also been investigated in China (e.g. Wang et al 2014, Zhao et al 2014, Xing et al 2015, Wei et al 2018. These studies all focus on monsooninfluenced regions, including the Qinghai-Tibetan Plateau (QTP), Northeast China (NEC) and South China (SC). There is little information in existing research relating to CAR variations and their controlling factors in the arid areas of Northwest China (NWC). The climate of arid NWC is dominated by the prevailing Westerlies, and is characterized by an increasing trend towards moisture during the Holocene epoch, in contrast to the Holocene moisture trend in monsoon-dominated regions (i.e. QTP, NEC and SC) (Chen et al 2008, . The purpose of our study is to perform highresolution chronological analysis of Tuolehaite Peat in the Altai Mountains, in order to explore changes in CAR during the Holocene, and to detect its possible driving factors by comparing the information obtained with existing local and regional palaeoclimatic data. Our work attempts to (i) deepen our understanding of wetland ecosystems during the evolution of the Holocene westerlies, and (ii) provide valuable insights into carbon dynamics in the varied climatic regions of China.

Study area
The Altai Mountains are an important geographical division between the taiga forests and the steppe deserts (Chen et al 2008), and also represent important climatic conjunction between the Westerlies airflows and the Asian monsoonal airflows during the Holocene period (Blyakharchuk et al 2004, 2007, Rudaya et al 2009, Wang and Feng 2013, Feng et al 2017 (Zhang et al 2016, Feng et al 2017. This mountainous belt has a very wet and cool environment, as compared with the arid climate in Central Asia (Chen et al 2008). The mean annual temperature is −3.6-1.8 ºC. The coldestmonth temperature in January is −14 ºC, and the warmest-month temperature in July is <16 ºC. The water vapor along the Irtysh valley is transported by the Westerlies, and precipitation has an obvious gradient of elevation: 200-300 mm, 300-500 mm, and 500-600 mm in low, middle and high elevations, respectively Zhang 2019, Li et al 2020). The wetlands in the poorly-drained intermontane basins cover about 108 811.8 hm 2 , distributed at elevations from 1700-2500 m a.s.l. (Nurbayev et al 2008). They are primary reservoirs of terrestrial carbon, and are particularly sensitive to climate change (Zhang et al 2016, Feng et al 2017, Xu et al 2019, Rao et al 2019, Yang et al 2019.

Study site and peat coring
The Tuolehaite Peat deposit (48.44 • N, 87.54 • E, 1700 m a.s.l.; site 1 in figure 1(a)), which developed in a small intermontane basin in the middle part of the southern Altai Mountains, is an ideal site for studying peat CAR changes, owing to its integrated and continuous deposits, and limited human disturbance. Modern vegetation around the peat is dominated by meadow steppe (see figure 1(b)), where Carex pamirensis and Carex altai represent the dominant plant life (figure 1(c)). A 380 cm sediment core was collected using a Holland peat corer (diameter 5 cm and length 50 cm) at Tuolehaite Peat in 2018. From these 50 cm cores, we collected samples of known volume every 2 cm over the whole peat sequence. The upper 358 cm of the core is composed of pure plant residues, and the lower 22 cm (i.e. 358-380 cm) is composed of clayey sediments (see figure 2).

Chronology
Eight peat samples were sent to the Beta Laboratory (Miami City, USA) for radiocarbon AMS dating, to establish the depth-age patterns of peat development. The selected samples without rootlets were deemed to be in an appropriate condition for radiocarbon analysis (see Table 1 in Zhang et al 2020a). Bacon age-modelling software (Blaauw and Christen 2011) is applied to establish a depth-age model, based on all 14 C dates (figure 2). This model simulates temporal deposition by assuming piece-wise linear accumulation, with accumulation rates, and the associated variability between nearby depths, constrained by prior information. The model is based on more than 5000 stored MCMC iterations, and has a stable run (right-first panel of figure 2). All settings are in their default mode, and the section thickness is set at 5 cm (right-third panel of figure 2). Based on the depth-age model, the Tuolehaite Peat core covers the past ∼10 650 cal. yr BP, with a clayey lacustrine layer occurring between ∼10 650 and ∼10 200 cal. yr BP (figure 2).

Dry bulk density and loss on ignition (LOI)
All samples were dried at 105 • C overnight, weighed, and then combusted in a muffle furnace at 550 • C for 4 h before being reweighed to determine dry bulk density and LOI (Dean 1974, Heiri et al 2001. The dry bulk density (g cm −3 ) was calculated by dividing the dry peat weight (g) by the wet peat volume (v = 1/2 × π × 2.5 2 × 2 = 19.64 cm 3 ). The total organic content (TOC) was estimated by multiplying the LOI by 0.5, based on previous Chinese data in (Yin et al 1991, Xin et al 2015. The LOI and dry bulk density analyses were performed at the Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences.

Humification analyses
Humification was analyzed via the commonly-used alkali-extraction method (Blackford and Chambers 1993). The measured value was expressed in terms of the percentage of light absorbance, which is proportional to the degree of peat humification (i.e. a lower value shows a lower degree of humification,   Table 1 of Zhang et al (2020a). and vice versa) (Aaby and Tauber 1975). The relationship between light absorbance and humification can be distorted by minerogenic constituents. It is therefore necessary to correct light absorbance values for any distortion caused by minerogenic constituents. A simple linear correction can be made based on the minerogenic content, as determined by LOI (Blackford and Chambers 1993, Chambers et al 2011, Zhang et al 2017. To remove aberrant data and 'random errors' , the minerogenically-corrected humification data were smoothed via a 3-point moving-window smoothing method. In addition, humification has been proved to increase over time (Clymo 1984). To remove the time-dependency effect, the 3-point moving-window smoothed data for the humification sequence were detrended by applying a linear regression line (Newnham et

Peat growth and carbon accumulation rates
The peat growth rate (PGR; mm yr −1 ) at Tuolehaite Peat was calculated based on the depth-age model and the thickness of the peat samples. The CAR was calculated via the following equation (Tolonen and Turunen 1996), based on raw data including PGR, dry bulk density, and TOC measurements: where CAR is the carbon accumulation rate (g C m −2 yr −1 ), PGR denotes the peat growth rate (mm yr −1 ), d is the dry bulk density (g m −3 ), and c is the TOC content (g C g −1 dry weight). The related data including dry bulk density, TOC, humification and carbon accumulation rates are provided in supplementary materials (stacks.iop.org/ERL/15/124014/mmedia).

Humification, dry density and total organic carbon (TOC)
Having performed the minerogenical correction and the removal of the time-dependency-effect, humification (i.e. light absorbance) in the ombrotrophic phase experiences one lower-value interval, at ∼10 200 to ∼8200 cal. yr BP, one higher-value interval at ∼8200 to ∼4000 cal. yr BP, and then a decreasing trend at ∼4000 to ∼1200 cal. yr BP. There is an increasing trend in the past 1200 years (see figure 3(a)). The mean dry density (g cm −3 ) at the Tuolehaite profile is 0.2 ± 0.1 g cm −3 (figure 3(b)). A clear decrease in dry density is observed towards the surface of the profile, with a value of 1.1 g cm −3 in the bottom layer (380 cm), and 0.1 g cm −3 in the surface layer (2 cm).
The mean TOC at the Tuolehaite profile is 39.2%, with a maximum value of 46.3% in the ombrotrophic phase, and a minimum value of 8.7% in the minerogenic phase (figure 3(c)). The TOC sequence can be classified into three stages in the ombrotrophic phase, i.e., since ∼10 200 cal. yr BP: high TOC at ∼10 200 to ∼7600 cal. yr BP, with a mean of 37.0%, higher TOC at ∼7600 to ∼3500 cal. yr BP, with a mean of 42.9%, and then a decreasing TOC, ranging from 43.9% to 30.1%, since ∼3500 cal. yr BP.

Peat growth rate (PGR) and carbon accumulation rates (CAR)
The mean PGR at Tuolehaite Peat is 0.4 mm yr −1 (2.4 yr mm −1 ). The PGR sequence shows that the growth of peat has been continuous at variable rates (see figure 3(d)). Specifically, for the period prior to ∼9700 cal. yr BP, the PGR (0.5 mm yr −1 ) was relatively high, and was followed by relatively low values, with a mean of 0.2 mm yr −1 , between ∼9700 and ∼7600 cal. yr BP. The PGR slightly increased (0.3 mm yr −1 ) between ∼7600 and ∼4300 cal. yr BP, and then abruptly increased to a high level, with a mean of 0.6 mm yr −1 , at ∼4300 to ∼3500 cal. yr BP. The PGR decreased rapidly at ∼3500 cal. yr BP, dropping to 0.25 mm yr −1 , before switching to a high level (mean 0.7 mm yr −1 ) in the past 1000 years.
The mean CAR is 28.0 ± 10.5 g C m −2 yr −1 at the Tuolehaite Peat core (figure 3(e)). During the initial period of peat formation (∼10 200 to ∼9700 cal. yr BP), the CAR (40.0 g C m −2 yr −1 ) was higher than the mean value of the whole profile. In the following two millennia (∼9700 to ∼7500 cal. yr BP), the CAR (25.0 g C m −2 yr −1 ) reduced. The CAR recovered between ∼7500 and ∼5800 cal. yr BP, followed by a decreasing trend between ∼5800 and ∼4200 cal. yr BP, with a mean of 23.7 g C m −2 yr −1 . The largest increase in CAR (40.1 g C m −2 yr −1 ) throughout the entire Holocene was detected between ∼4200 and ∼3400 cal. yr BP. The CAR values in the past ∼3400 cal. yr BP showed a decreasing trend, with a mean of 18.4 g C m −2 yr −1 .

Climate control in carbon accumulation
The mean CAR at Tuolehaite Peat was 28.0 g C m −2 yr −1 (figure 3(d)  . Properties of Tuolehaite Peat core (from a to e): humification, bulk density (BD, g cm −3 ), total organic carbon (TOC), peat growth rate (PGR, mm yr −1 ), and carbon accumulation rate (CAR, g m −2 yr −1 ). Note: humification data is inferred from minerogenically-corrected and time-dependency-effect-removed light absorbance values. Positive humification data indicates high peat decay, and negative humification data indicates low peat decay.
The CAR of peat is primarily dependent on the balance between primary production and decomposition, which is generally modulated by temperature and precipitation (Svensson 1988, Laiho 2006). The CAR (4.8-68.8 g C m −2 yr −1 ) at Tuolehaite Peat shows an overall decreasing trend since the Holocene era (figure 4(a)). The humification data indicates one high-decomposition interval at ∼8200 to ∼4000 cal. yr BP, and two low-decomposition intervals: at ∼10 200 to ∼8200, and since ∼4000 cal. yr BP ( figure 4(b)). The pollen A/C ratio-expressed moisture inferred from the same Tuolehaite core reveals a wet early Holocene (∼10 600 to ∼8500 cal. yr BP), a considerably dry middle Holocene (∼8500 to ∼4000 cal. yr BP), and a further wet period in the late Holocene (∼4000-0 cal. yr BP) (figure 4(c)) (Zhang et al 2020a), with this analysis being supported by the recorded molecular markers, P aq (the proportion of aquatic components), and moisture variations inferred from the nearby Ganhaizi Peat region  (Zhao et al 2017). Combined with these data regarding changes in peat humification, temperature and precipitation, other potential forcing factors relevant to Holocene carbon dynamics at Tuoleheite Peat are investigated below.
One of the phases of high CAR (40.0 g C m −2 yr −1 ) occurs at the beginning of peat formation at Tuolehaite Peat (figure 4(a)), which coincides with the cold and wet climate at the beginning of the Holocene era (∼10 200 to ∼9700 cal. yr BP) (see figures 4(c)-(f)). The latter may result from the major forcing of the Laurentide ice sheet and other highlatitude ice sheets (Zhao et al 2017, Wang and. These climatic conditions limited peat decay in this period ( figure 5(b)), which was conducive to the preservation of peat plants, and a concomitant increase in CAR. In the following two millennia, i.e., from ∼9700 to ∼7500 cal. yr BP, the reduction in CAR (25.0 g C m −2 yr −1 ) (figure 5(a)) coincides with a relatively cold and drying climate (figures 4(c)-(f)). Compared with the former interval (i.e., ∼10 200 to ∼9700 cal. yr BP), increasing humification could be responsible for the reduction in CAR during this period. A phase of notable recovery in CAR at ∼7500 to ∼5800 cal. yr BP corresponds well with the warmest and driest climatic conditions. The warmest climate may have led to the highest level of plant growth at Tuoleheite Peat. However, the warm-dry climate also contributes to a high level of decay in peat-living plants, due to increasing microbial activity in an aerobic environment; therefore, no higher CAR was observed during this interval (Zhang et al 2020a). The CAR once again fell below the mean CAR of the profile as a whole during the period from ∼5800 to ∼4200 cal. yr BP ( figure 5(a)), which resulted from the combined effects of cold-dry climate, and high humification. Furthermore, the CAR exhibited a rapid increase at ∼4200 to ∼3400 cal. yr BP (figure 4(a)), which coincides with a marked decrease in humification, and a slightly wetter and warmer climatic conditions. These conditions may have contributed to the higher net primary productivity indicated by the higher TOC at Tuolehaite Peat (see figure 3(c)). The next significant drop in CAR at Tuolehaite Peat was identified in the period since ∼3400 cal. yr BP, which coincides with a period of cold and wet climatic conditions.
Overall, the long-term trend of CAR at Tuolehaite Peat is parallel with the temperature trend during the Holocene interval. This indicates that the CAR in the southern Altai Mountains is primarily modulated by changes in temperature, and is also partly influenced by peat decomposition. A warming climate would enhance net primary productivity, owing to an increase in growing-season length in the Altai Mountains, and vice versa (Charman et al 2013).

Comparison with other peat carbon accumulations in China
As discussed above, studies of peat CAR in China mainly focus on monsoon-influenced regions, including QTP (Wang et  As shown in figure 5, the Tuolehaite CAR in NWC, ranging from 4.8 to 68.8 g C m −2 yr −1 (28.0 g C m −2 yr −1 on average) shows a decreasing trend since the Holocene period, with two episodes of higher CAR, prior to ∼9600 cal. yr BP, and at ∼4200 to ∼3400 cal. yr BP (see figure 5(a)). In the QTP, the Hongyuan CAR ranges from 9.8 to 80.1 g C m −2 yr −1 , with a mean of 33.6 g C m −2 yr −1 (Cai et al  2014). The long-term CAR trend at Hongyuan Peat ( figure 5(b)) is basically consistent with that at Tuolehiate Peat ( figure 5(a)). The CAR inferred from Hani Peat in NEC ranges from 8.6 to 168.5 g C m −2 yr −1 (mean 41.2 g C m −2 yr −1 ) (Cai et al 2013). During the Holocene, the Hani CAR can be defined in four stages: (1) a relatively low and stable CAR before ∼8200 cal. yr BP; (2) a relatively high CAR between ∼8200 and ∼4000 cal. yr BP; (3) a further period of decreasing CAR between ∼4000 and ∼1500 cal. yr BP; and (4) rapidly increasing CAR in the past 1500 years  (Wang et al 2005, Zhao et al 2009, which favoured the incidence of widespread lacustrine deposits with low organic carbon content in shallow basins, and produced a low CAR in NEC and SC (Xing et al 2015, Wei et al 2018. In the past ∼4200 cal. yr BP, a weakening EASM and the related drying climate (figure 5(e)) has led to the accumulation of organiccarbon-rich sediments (i.e. higher CAR) in NEC and SC (figures 5(c)-(d)). Therefore, we propose that precipitation could be an important factor modulating the CAR in peat regions in NEC and SC. It should be noted that the increasing CAR at Hani Peat during the past 1500 years (figure 5(c)) is attributable to the rising moisture in the context of a cooling climate in NEC (Cai et al 2013).
Conversely, high summer insolation and strong climate seasonality (figure 5(g)) in the early-middle Holocene (Zhao andZ C 2012, Zhao et al 2014) resulted in higher CAR at the high-elevation Tuolehaite Peat in NWC (figure 5(a)) and at Hongyuan Peat in QTP ( figure 5(b)). An increased temperature would enhance primary productivity, due to the extension of the growing season of plants in these relatively high-elevation regions (Charman et al 2013, Zhao et al 2014. The warm climate could also improve peat decay by stimulating microbial activity (Ise et al 2008, Dorrepaal et al 2009. However, the decomposition process may only have a minor effect on primary productivity at high-elevation regions in QTP and NWC, due to strong seasonality at higher elevations, with increasing plant production in shortwarm summers, and the cessation of peat decomposition in long, cold winters (Wei et al 2018). Subsequent decreases in CAR in QTP and NWC during the middle and late Holocene are attributable to the subsequent summer-insolation-induced cooling climate (figure 5(g)) (Berger and M F 1991). The decreasing trends in CAR in the high-elevation areas of QTP and NWC are consistent with the northern peatland CAR variations during the Holocene . Huang et al (2018) and Zhang et al (2020b) confirm that vegetation changes in the high-elevation Altai Mountains are modulated by the insolationdriven temperature change in the context of a wetter Holocene trend (figure 5(f)) . A similar vegetation-climate relationship in QTP has previously been reported by Zhao et al (2009). We therefore propose that temperature is a key factor influencing peat CAR in relatively high-elevation peat deposits (e.g. Hongyuan Peat and Tuolehaite Peat), while precipitation is the key factor controlling the CAR in relatively low-elevation peat deposits (e.g. Hani Peat and Dahu Peat).

Conclusions
In this paper, we have investigated the carbon accumulation rates (CAR) at Tuolehaite Peat in the Westerlies-influenced Altai Mountains in NWC. The results indicate that the CAR at Tuolehaite Peat experienced a decreasing trend, with a mean of 28.0 g C m −2 yr −1 , during the Holocene interval. Compared with the CAR trend in the monsooninfluenced QTP, NEC and SC, the Holocene CAR trends at Tuolehaite Peat in the Altai Mountains, and at Hongyuan Peat in the QTP, are opposite to those found at Hani Peat in NEC and at Dahu Peat in SC. These opposing trends in relation to CAR appear to result from specific influencing factors (i.e., temperature and precipitation) at relatively high and low elevations.