REGULAR ARTICLE

Fertilization enhancing carbon sequestration as carbonatein arid cropland: assessments of long-term experimentsin northern China

X. J. Wang & M. G. Xu & J. P. Wang & W. J. Zhang &

X. Y. Yang & S. M. Huang & H. Liu

Received: 28 August 2013 /Accepted: 28 February 2014# Springer International Publishing Switzerland 2014

AbstractAims Soil inorganic carbon (SIC), primarily calciumcarbonate, is a major reservoir of carbon in arid lands.This study was designed to test the hypothesis thatcarbonate might be enhanced in arid cropland, in asso-ciation with soil fertility improvement via organicamendments.Methods We obtained two sets (65 each) of archivedsoil samples collected in the early and late 2000’s fromthree long-term experiment sites under wheat-corncropping with various fertilization treatments in north-ern China. Soil organic (SOC), SIC and their Stable 13C

compositions were determined over the range 0–100 cm.Results All sites showed an overall increase of SICcontent in soil profiles over time. Particularly, fertiliza-tions led to large SIC accumulation with a range of 101–202 g C m−2 y−1 in the 0–100 cm. Accumulation ofpedogenic carbonate under fertilization varied from 60to 179 g C m−2 y−1 in the 0–100 cm. Organic amend-ments significantly enhanced carbonate accumulation,in particular in the subsoil.Conclusions More carbon was sequestrated in the formof carbonate than as SOC in the arid cropland in north-

Plant SoilDOI 10.1007/s11104-014-2077-x

Responsible Editor: Eric Paterson.

X. J. Wang (*) : J. P. WangState Key Laboratory of Desert and Oasis Ecology, XinjiangInstitute of Ecology and Geography, Chinese Academy ofSciences, Urumqi, Xinjiang 830011, Chinae-mail: xjwang@ms.xjb.ac.cn

X. J. WangCollege of Global Change and Earth System Science, BeijingNormal University and Joint Center for Global ChangeStudies, Xinjiekouwai Street No.19, Haidian District,Beijing 100875, China

M. G. Xu :W. J. ZhangMinistry of Agriculture Key Laboratory of Crop Nutrition andFertilization, Institute of Agricultural Resources and RegionalPlanning, Chinese Academy of Agricultural Sciences,Beijing 100081, China

J. P. WangGraduate University of Chinese Academy of Sciences,Beijing 100049, China

X. Y. YangCollege of Natural Resources and Environment, NorthwestAgricultural and Forestry Science and Technology University,Yangling, Shaanxi 712100, China

S. M. HuangInstitute of Plant Nutrition, Resources and Environment,Henan Academy of Agricultural Sciences, Zhengzhou, Henan450002, China

H. LiuInstitute of Soil and Fertilizer and Agricultural Sparing Water,Xinjiang Academy of Agricultural Sciences, Urumqi 830091,China

ern China. Increasing SOC stock through long-termstraw incorporation and manure application in the aridand semi-arid regions also enhanced carbonate accumu-lation in soil profiles.

Keywords Soil carbonate . Stable 13C composition .

Carbon sequestration . Fertilization . Cropland . Aridregion

Introduction

Soil organic carbon (SOC) and inorganic carbon (SIC) areimportant reservoirs of carbon on the global lands. Soilorganic carbon, as a key index for soil fertility and carbonsequestration, has gained recognition (Lal 2001, 2004).The estimated global SOCpool for the upper 100 cm has arelatively narrow range, i.e., 1220–1576 Pg (Eswaranet al. 2000). In contrast, there is a large discrepancy inthe estimated global SIC pool, which ranges from lessthan 700 Pg to more than 1700 Pg (Eswaran et al. 2000),and SIC has received much less attention despite itspotential for carbon sequestration and climate mitigation(Eshel et al. 2007; Lal and Kimble 2000; Manning 2008).

Soil inorganic carbon, primarily calcium carbonate,is the most common form of carbon in soils of arid andsemi-arid regions. More than 35 % of Earth’s landsurface is characterized as either arid or semi-arid, whereSIC stock is 1–9 times higher than SOC stock(Scharpenseel et al. 2000; Schlesinger 1982).Therefore, accurately estimating SIC at all scales isessential to evaluate the role of soils in the global carboncycle. Moreover, attempts to decrease the atmosphericCO2 concentration require better understanding of car-bon sequestration as all forms in various systems, in-cluding carbonate in arid and semi-arid lands (Mongeret al. 2009; Zheng et al. 2011).

The SIC pool consists of two major components: thelithogenic carbonate (LIC) and pedogenic carbonate(PIC). The former originates as detritus from parentmaterials, mainly limestone, whereas the secondary car-bonate is formed by the dissolution and re-precipitationof LIC or through dissolution of carbon dioxide (CO2)into HCO3

−, then precipitation with Ca2+ and/or Mg2+

originating from non-LIC minerals (e.g., silicateweathering). Thus, PIC formation in soils could lead tocarbon sequestration (Monger and Gallegos 2000).

Recent reports (Wohlfahrt et al. 2008; Xie et al. 2009)of significant CO2 uptake (>100 g Cm−2 yr−1) in deserts

have raised some questions, e.g., where does the carbongo (Stone 2008). One hypothesis was made that theabsorbed CO2 would be precipitated as calcium carbon-ate. Affirmation of this hypothesis requires evidence ofPIC accumulation. However, field data necessary toquantify carbon sequestration as carbonate have beenlacking although limited studies seem to indicate thatPIC accumulation is extremely low in most parts of thearid and semi-arid regions (Schlesinger et al. 2009).

The Chinese National Soil Fertility and FertilizerEfficiency Monitoring Network was established in1990 across China, to study the effects of differentfertilization managements on agricultural productivityand soil fertility. There have been numerous studiesusing data collected from the network to evaluate carbonsequestration as SOC under various fertilization treat-ments (e.g., Cong et al. 2012; Shen et al. 2007; Zhanget al. 2010), which demonstrated widespread increasesin SOC stock as a results of fertilization in northernChina. Given that there was evidence of more carbonsequestrated in the form of PIC than as organic matter inarid regions (Landi et al. 2003), we hypothesized thatthere would be significant carbonate accumulation inarid croplands, in association with the increases of SOC.

To test this hypothesis, we selected three sites underarid and semi-arid climatic conditions across northernChina. Each had the same experiment design with sim-ilar fertilization treatments. We obtained 130 archivedsoil samples collected from different years under variousfertilization managements, and analyzed SOC and SICcontents and their Stable 13C compositions in the 0–100 cm soil profiles. The two objectives of this work areto study the soil carbonate dynamics in arid cropland,and to evaluate impacts of fertilization management oncarbonate accumulation in soil profile.

Materials and methods

Description of the long-term experiment sites

There are three long-term experiment (LTE) sites in thenetwork under arid (Urumqi) and semi-arid (Yanglingand Zhengzhou) climatic conditions in northern China(Table 1). Annual average temperature and precipitationare lower in Urumqi (7.7 °C and 299 mm) than inYangling and Zhengzhou (13–14.5 °C and 550–632 mm). Annual evaporation varies from 993 mm to2,570 mm. Irrigation is usually applied during the crop’s

Plant Soil

growing seasons. All three sites have calcareous soils,with illite as one of main minerals. Chlorite is alsodominant at Urumqi, and smectite at the other sites.Initial SOC content was in a range of 6.7–8.8 g kg−1,initial total nitrogen 0.67–0.87 g kg−1, and soil pH 8.1–8.6. Bulk density was measured for the 0–40 cm at theZhengzhou and Urumqi sites, and the 0–100 cm atYangling.

The crops at these sites are corn (Zea mays L.) andwheat (Triticum Aestivium L.). The Urumqi site has amono-cropping system whereas the other two sites havea double-cropping system. The mono-cropping systemhas a rotation of corn-wheat-wheat, with corn seededduring late April to early May and followed by springwheat (seeded in mid-April) then winter wheat (seededin late September in the same year). The doublecropping systems have a rotation of summer corn (seed-ed in late April to early May) and winter wheat (seededin October). While there are no replicates at all threesites (note: most LTEs in the network have no repli-cates), experiment plots are large enough (196–468 m2)to be representative. In addition, all sites had 2–3 yearsof land preparations (e.g., cultivation with tillage) toeliminate spatial variations in soil conditions prior toexperiments. A number of studies have carried out to

evaluate the impacts of fertilization treatments on SOCdynamics using data collected from the network (Conget al. 2012; Shen et al. 2007; Zhang et al. 2010).

We obtained two sets of archived soil samples: oneset collected in 2009 and the other set from the early2000s (2001 for the Urumqi site and 2002 for the othertwo sites). Five treatments were included: (1) no fertil-ization (control), (2) mixed mineral nitrogen-phosphorus-potassium fertilization (NPK), (3) NPK fer-tilization with straw incorporation (NPKS), (4) NPKfertilization with manure application (NPKM), and (5)50–100 % higher application rates of minerals and ma-nure than the NPKM treatment (hNPKM). Some soilsamples were missing for the NPKS treatment at theUrumqi site, and there were contamination in some soilscollected under hNPKM from the Zhengzhou site. Thus,we excluded the NPKS treatment for the Urumqi siteand hNPKM for the Zhengzhou site in this study. Themineral nitrogen, phosphorus and potassium fertilizerswere urea, calcium superphosphate, and potassium chlo-ride (or potassium sulfate), respectively. For the NPKStreatment, all above-groundmaterials (except the grains)from one crop (corn at the Yangling and Zhengzhousites) were incorporated into topsoil each year.Fertilization rates and manure types, summarized in

Table 1 The characteristics oflong-term experiment sites innorthern China

aBulk density

Properties Unit Urumqi Yangling Zhengzhou

Latitude N 43°59′26″ 34°17′51″ 34°47′25″

Longitude E 87°46′45″ 108°00′48″ 113°40′42″

Altitude m 600 525 59

Annual mean temp. °C 7.7 13.0 14.5

Annual precipitation mm 299 550 632

Annual evaporation mm 2,570 993 1,450

Soil classification (FAO) Haplic Calcisol CalcaricRegosol

Calcaric Cambisol

Parent material Limestone Loess River Alluvium

Soil pH 8.1 8.6 8.3

Initial SOC (0–20 cm) g kg−1 8.8 7.4 6.7

Initial TN (0–20 cm) g kg−1 0.87 0.83 0.67

BDa (0–20 cm) g cm−3 1.21 1.35 1.41

BD (20–40 cm) g cm−3 1.35 1.56 1.44

BD (40–60 cm) g cm−3 1.48

BD (60–80 cm) g cm−3 1.43

BD (80–100 cm) g cm−3 1.36

Cropping system Mono Double Double

Crop rotation Corn-wheat-wheat Wheat-corn Wheat-corn

Plot size m2 468 196 400

Plant Soil

Table 2, were based on local agricultural practices (Yanget al. 2012; Zhang et al. 2010).

Soil sampling and analyses

Soil samples were collected from five layers (0–20, 20–40, 40–60, 60–80, and 80–100 cm) during September–October in 2002 at Yangling and Zhengzhou (2001 atUrumqi) and 2009 at all three sites. There were 5–10cores (5-cm-diam) of soil in each layer, randomly sam-pled in each plot, air dried and thoroughly mixed.Representative sub-samples were crushed to 0.25 mmfor SOC and SIC measurements that were carried outusing a CNHS-O analyzer (Model EuroEA3000). ForSOC measurement, 20 mg soil was pretreated with 10drops of H3PO4 for 12 h to remove carbonate. Thepretreated sample was combusted at 1,020 °C with aconstant helium flow carrying pure oxygen to ensurecompleted oxidation of organic materials. Production of

CO2 was determined by a thermal conductivity detector.Total soil carbon was measured using the same proce-dure without pretreatment of H3PO4. Soil inorganiccarbon was calculated as the difference between totalsoil carbon and SOC. All the procedures were per-formed at the State Key Laboratory of Lake Scienceand Environment, Nanjing Institute of Geography andLimnology, Chinese Academy of Sciences (CAS).

Stable 13C isotopic compositions (δ13C) in SOC andSIC were determined by measuring the isotopic compo-sition of collected CO2 using a Finnigan MAT DeltaPlus XP Isotope Ratio Mass Spectrometer at theNanging Institute of Geology and Paleontology, CAS.For δ13C in SOC, CO2 was collected in the same way asthat for SOC. For δ13C in SIC, CO2 was collected duringthe reaction of pre-heated soil (at 375 °C for 17 h) withH3PO4. We reported isotopic data in delta notationrelative to the Vienna Pee Dee Belemnite (VPDB).

Estimation of PIC

Following Landi et al. (2003), we calculated the amountof PIC as:

PIC ¼ δ13CSIC − δ13CPM

δ13CPIC − δ13CPMSIC ð1Þ

where δ13CSIC, δ13CPM and δ13CPIC were the Stable 13C

in carbonate for the bulk SIC, parent material, and purePIC, respectively. For standard calculation, we setδ13CPM as zero for the Urumqi and Zhengzhou sites,but −1‰ for the Loess at Yangling (according to Liuet al. 2011). Based onMermut et al. (2000), δ13CPICwascalculated from the Stable 13C in SOC (δ13CSOC):

δ13CPIC ¼ δ13CSOC þ 14:9 ð2Þwhere the value of 14.9 represented an average differ-ence in Stable 13C composition between SOC and PIC,which included the isotopic fractionation of 4.4 for CO2

diffusion and 10.5 for carbonate precipitation (Cerling1984; Cerling et al. 1989, 1991).

Calculations and statistical analyses of SIC and PICchange rates

For each layer, SIC and PIC stocks were calculated frombulk density and carbon contents. For the Zhengzhou andUrumqi sites where bulk density was only measured forthe 0–20 and 20–40 cm layers (Table 1), we used the bulk

Table 2 Application rates of nitrogen (N), phosphorus (P), andpotassium (K) for each growing season under different treatments

Treatments Urumqi Yangling Zhengzhou

Maize/wheat Maize Wheat Maize Wheat

Nitrogen (kg N ha−1)

CK 0 0 0 0 0

NPK 242 188 165 188 165

NPKS – 188 165+43a 188 123+42a

NPKM 85+240a 188 50+115a 188 50+115a

hNPKM 152+360a 188 74+173a

Phosphorus (kg P ha−1)

CK 0 0 0 0 0

NPK 60 25 58 41 36

NPKS 25 58+4a 41 36+8a

NPKM 22+65a 25 58+95a 41 36+66a

hNPKM 39+98a 25 86+143a

Potassium (kg K ha−1)

CK 0 0 0 0 0

NPK 47 78 69 78 68

NPKS – 78 69+57a 78 68+86a

NPKM 9+160a 78 69+180a 78 68+92a

hNPKM 14+240a 78 103+271a – –

a The amount of N/P/K added by crop straw or manure. The manuretypes were horse manure from 1990 to 1998 and cattle manure from1999 to 2009 (no application in 2007) at the Zhengzhou site. Cattleand goat manure were applied at Yangling and Urumqi, respectively

Plant Soil

density values from the 20–40 cm layer for calculations ofSIC and PIC stocks below 40 cm. Accumulation rates ofSIC and PIC were estimated by the increases of the SICand PIC stocks from 2001/2002 to 2009. We applied one-way analyses of variance (ANOVA) and Fisher’sprotected least significant difference (LSD) to evaluatefertilization effects, using Microsoft Excel.

Results

Soil inorganic carbon and isotopic compositionat Urumqi

Figure 1 illustrates that SIC content was less than7 g C kg−1 at the Urmuqi site. There was a decrease in

SIC content with depth, particularly under fertilizationtreatments that revealed slightly higher SIC in the upper20 cm layer, but lower SIC below 60 cm. Applyingmanure caused a modest decrease by 0.4–1.2 g C kg−1

in SIC content in the soils collected in 2001, relative tothe NPK treatment. However, there was an increase inSIC over time in association with the long-term appli-cation of manure, with a greater increase found below40 cm (Fig. 1).

The δ13CSIC value varied from −1.2 to −4.4‰, withthe most negative values found under high rates ofmineral and manure application (hNPKM). There waslittle change in δ13CSIC over time except under thehNPKM treatment. On average, δ13CSIC was −2.4 and−2.8‰ in soils collected in 2001 and 2009, respectively.Similar to SIC, PIC revealed a weak decreasing trend

Fig. 1 Profiles of soil inorganic carbon (SIC, g kg−1: left column),Stable 13C isotopic composition in SIC (‰: middle column), andestimated pedogenic carbonate (PIC, g kg−1: right column) under

non-fertilization (a–c), NPK (d–f), NPKM (g–i) and hNPKM (j–l)for 2001 (white square) and 2009 (white triangle) at the Urumqisite

Plant Soil

with depth. Estimated PIC showed little difference in2001 with a similar range (1–2 g C kg−1) for all thetreatments, but a considerable difference in 2009. Whilethere was little increase in PIC over time under non-fertilization and NPK fertilization, there was a signifi-cant increase in PIC with manure application.Particularly, the hNPKM treatment resulted in approxi-mately 100 % increase in PIC (Table 3).

Soil inorganic carbon and isotopic compositionat Yangliang

The Yangling site showed a sharp decline in SIC withdepth, from 9 to 12 g C kg−1 in the 0–40 cm layer to nearzero below 80 cm (Fig. 2), particularly under thosewithout organic matter addition (such as control andNPK). Fertilization led to an increase of 1–2 g C kg−1

in SIC over time, with the greatest increase found underthe long-term application of manure.

The δ13CSIC value ranged from −5.6 to −8.8‰,showing an approximate value of −6‰ above 60 cmacross all the treatments. There was a depletion of 13C inSIC below 60 cm, particularly in the 80–100 cm layer.Estimated PIC showed a similar vertical distribution tothat of SIC, with much higher values in the topsoil (6–8 g C kg−1) than in the 80–100 cm layer (0–2 g C kg−1).Overall, application of organic materials (such as

NPKS, NPKM and hNPKM) resulted in an increase ofPIC content over time.

Soil inorganic carbon and isotopic compositionat Zhengzhou

Figure 3 shows a modest decrease in SIC with depth atthe Zhengzhou site, from approximately 8 g C kg−1 inthe 0–20 cm layer to 4–5 g C kg−1 below 80 cm. Ourdata showed little difference in the magnitude of SICbetween fertilization treatments except under theNPKM treatment that showed higher SIC contents(∼8 g C kg−1) in the 40–60 cm layer. There was anincrease in SIC content over time cross all thetreatments.

A narrow range of the δ13CSIC value (from −4.1 to−5.3‰) was found at the Zhengzhou site. Unlike theUrumqi and Yangliang sites, the δ13CSIC value was aslightly less negative in the subsoil relative to the surfaceand revealed little difference between 2002 and 2009.Estimated PIC showed a decline from 3 to 4 g C kg−1 inthe 0–20 cm layer to approximately 2 g C kg−1 below80 cm. Fertilization led to an increase of PIC in thewhole soil profile over time. Particularly, straw incorpo-ration and manure application in addition to chemicalfertilization resulted in increased PIC content by 0.1–1.2 g C kg−1 from 2002 to 2009.

Table 3 Change rates (g C m−2 y−1) of soil inorganic carbon (SIC) and pedogenic carbonate (PIC) over 0–20 and 20–100 cm under variousfertilization treatments

Treatment Urumqi Yangling Zhengzhou Meana Totalb

0–20 20–100 0–20 20–100 0–20 20–100 0–20 20–100 0–100

SIC

Control −12 29 31 5 12 65 10 33 a 43 (31)

NPK 3 60 31 144 27 39 20 81 ab 101 (64)

NPKS 137 6 160 2 125 4 142 ab 146 (28)

NPKM 35 85 41 197 30 46 35 109 ab 144 (84)

hNPKM 24 116 16 247 20 182 b 202 (87)

PIC

Control 0 5 4 17 13 20 6 a 14 a 20 (14)

NPK 10 26 25 61 35 22 23 ab 36 ab 60 (25)

NPKS 97 42 83 14 98 28 ab 91 c 119 (8)

NPKM 20 35 50 114 17 57 29 ab 69 bc 98 (58)

hNPKM 31 136 33 157 32 b 147 d 179 (16)

a Values followed by the same letter in each column are not significantly different, based on a LSD test (P<0.05)b Values in parentheses are standard deviations

Plant Soil

Carbonate change rates under different fertilizationmanagements

Table 3 illustrates that change rate of SIC varied widely,from −12 to 41 g C m−2 y−1 in the topsoil (0–20 cm) andfrom 5 to 247 g C m−2 y−1 in the subsoil (20–100 cm).Average SIC accumulation under non-fertilization was10 and 33 g C m−2 y−1 in the topsoil and subsoil,

respectively. NPK chemical fertilization led to an accu-mulat ion of SIC stock, with an average of20 g C m−2 y−1 in the 0–20 cm and 81 g C m−2 y−1 inthe 20–100 cm. Straw incorporation with NPK yielded alow accumulation rate in the topsoil (4 g C m−2 y−1) buta high rate in the subsoil (142 g C m−2 y−1). In general,manure application with NPK fertilizers resulted in agreater rate of SIC accumulation than NPK fertilization,

Fig. 2 Profiles of soil inorganic carbon (SIC, g kg−1: left column),Stable 13C isotopic composition in SIC (‰: middle column), andestimated pedogenic carbonate (PIC, g kg−1: right column) under

non-fertilization (a–c), NPK (d–f), NPKS (g–i), NPKM (j–l) andhNPKM (m–o) for 2002 (white square) and 2009 (white triangle)at the Yangling site

Plant Soil

particularly 46–247 g C m−2 y−1 in the subsoil. Onaverage, SIC accumulation rates over the 0–100 cmwere 43, 101, 146, 144 and 202 g C m−2 y−1 undernon-fertilization, NPK, NPKS, NPKM and hNPKMtreatments, respectively. Statistical analyses indicatedthat high rate of manure addition caused a significantlyhigher rate of SIC accumulation in the subsoil (P<0.05).

Our estimations showed that the average accumula-tion rates of PIC in the topsoil (6–32 g C m−2 y−1) weresimilar to those of SIC stock (4–35 g C m−2 y−1)(Table 3). However, average PIC accumulation rates inthe subsoil (14–147 g C m−2 y−1) were lower than thosefor SIC (33–182 g C m−2 y−1). Accumulation rates ofPIC were lowest without fertilization in both topsoil andsubsoil. Our analyses demonstrated that the PIC accu-mulation rates under organic amendments (i.e., NPKS,

NPKM and hNPKM) were significantly higher, partic-ularly in the subsoil. Average rates of PIC accumulationover the 0–100 cm were 20, 60, 119, 98 and179 g C m−2 y−1 for non-fertilization, NPK, NPKS,NPKM and hNPKM treatments (Table 3), respectively,which were 45, 59, 81, 68 and 89 % of the SIC stocks,respectively.

Discussion

Uncertainties in estimated PIC

Isotope technique is often used to differentiate betweenPIC and LIC because of their distinct isotopic signaturesthat are linked to different sources of carbonate (Eshel

Fig. 3 Profiles of soil inorganic carbon (SIC, g kg−1: left column),Stable 13C isotopic composition in SIC (‰: middle column), andestimated pedogenic carbonate (PIC, g kg−1: right column) under

non-fertilization (a–c), NPK (d–f), NPKS (g–i) and NPKM (j–l)for 2002 (white square) and 2009 (white triangle) at the Zheng-zhou site

Plant Soil

et al. 2007; Landi et al. 2003; Mermut et al. 2000).However, the isotope approach may introduce uncer-tainties in the estimates of PIC because of thedecoupling between CO2 production and carbonate pre-cipitation over time and space (Breecker et al. 2009;Mermut et al. 2000; Quade et al. 1989). On the otherhand, observations are limited in some relevant fields,

e.g., the value of δ13CPM and rates of CO2 production insoil profile. Thus, there are often assumptions madewhen applying the isotope approach, which may alsoproduce biases and/or uncertainties.

To assess the uncertainty in our estimates of PIC due

to the choices for δ13CPM , we carried out a sensitivitystudy by testing different values (the possible minimum

and maximum values) for δ13CPM (i.e., −2 and 1‰).Applying a smaller value for δ13CPM resulted in a sig-nificant decrease in the estimated rates of PIC accumu-lation (Table 4). While the differences between the twoestimates using δ13CPM values of −2 and 1‰ variedwidely, the average differences for all three sites weresimilar, i.e., from 19 g C m−2 y−1 at Zhengzhou to25 g C m−2 y−1 at Urumqi. Given that δ13C value ofSIC was generally more negative than −1.24‰ atUrumqi (Fig. 1), it would be reasonable to assume thatthe δ13CPM value was less negative than −1.24‰, whichimplied that the average rate of PIC accumulation waslikely greater than 43 g C m−2 y−1 under NPKfertilization.

Previous studies have indicated that atmosphericCO2 may enter upper soil pores under low rate of soilrespiration (Breecker et al. 2009; Cerling 1984;Stevenson et al. 2005), causing less negative δ13C valuein soil CO2, thus less negative δ

13C value in PIC. It isalso known that isotopic fractionation during carbonateprecipitation is negatively correlated with temperature(Breecker et al. 2009; Romanek et al. 1992). Our ap-proach did not take into account these processes, whichmight lead to underestimation of PIC particularly for

low fertility soils (e.g., under non-fertilization treat-ments) and under freezing conditions, and/or overesti-mation of PIC under warmer climates. Thus, it waslikely that PIC stock was underestimated at theUrumqi site because of the long cold winter, and ourapproach might overestimate the differences betweenUrumqi and other sites. Our simple calculation using areduced isotopic fractionation (by 2‰) in Eq. (2)yielded a modest decrease in estimated rates of PICaccumulation at Yangling (average 19 g C m−2 y−1)and Zhengzhou (average 16 g C m−2 y−1) (Table 5).

Given that PIC accumulation rate over 0–100 cmwaslower under NPK fertilization than under organicamendments (Table 3), that the rate of PIC accumulationunder the NPK would represent the minimum rate onfertilized cropland. Here, by taking into account of theuncertainties discussed above, we assumed that the min-imum rate of PIC accumulation under the NPKmight be36 g C m−2 y−1 (Table 3) at Urumqi. For the Yanglingand Zhengzhou sites, we took the values of 74 and40 g C m−2 y−1 (Table 4), then subtracted by 15 and17 g C m−2 y−1 (Table 5), respectively. Thus, we esti-mated that on average, fertilization might have led toPIC accumulation over 0–100 cm at a rate of greaterthan 38 g C m−2 y−1 in the arid and semi-arid croplandsin northern China.

Impacts of fertilization and cropping on soil carbondynamics

There is widespread evidence of land use impacts onsoil carbon dynamics. On the one hand, land uses foragricultural production have caused a decline in SOCstocks in many regions, e.g., in temperate and tropicalregions where forest or grassland were converted toagricultural land (Conant et al. 2001; Murty et al.2002; Ogle et al. 2005). On the other hand, convertingnative lands to croplands in arid and semi-arid regions

Table 4 Calculated PIC accu-mulation rates (g C m−2 y−1) over0–100 cm by setting δ13CPM to −2or 1‰ under different fertilizationtreatments

aValues in parentheses are stan-dard deviations

Treatment Urumqi Yangling Zhengzhou Meana

−2‰ 1‰ −2‰ 1‰ −2‰ 1‰ −2‰ 1‰

Control 0 6 17 27 10 40 9 (9) 24 (17)

NPK 15 40 74 103 40 61 43 (30) 68 (32)

NPKM 20 65 153 180 69 75 81 (67) 107 (64)

Mean 12 37 81 103 40 59 44 (35) 66 (34)

Plant Soil

results in a greater increase in both SOC and SIC thanother types of land use, e.g., in the middle of HexiCorridor, Gansu, China (Su et al. 2010) and in theRussian Chernozem (Mikhailova and Post 2006).

Our recent analyses of the LTE data demonstratedthat fertilization resulted in a general increase in SOC atmost sites in the arid and semi-arid regions of northernChina (Cong et al. 2012; Zhang et al. 2010). Table 6illustrates that accumulation rates of surface SOC (0–20 cm) were much higher under manure addition (64–126 g C m−2 y−1) relative to those under chemical NPKfertilization (0–43 g C m−2 y−1). While average SICaccumulation was generally lower than SOC accumula-tion in the topsoil, SIC accumulation rates over the 0–100 cm were significantly higher than those of surfaceSOC. In particular, continuous application of organiccarbon through manure led to an accumulation rate of76–238 g C m−2 y−1 in SIC (0–100 cm), implying thetransfer of organic carbon to carbonate (Manning 2008)and the formation of PIC (Pan and Guo 2000), andconfirming the hypothesis that “the rate of formationof secondary carbonates increases with increased addi-tion of organic matter to the soil” (Lal and Kimble2000). The increase of surface SOC on arid croplandwould be a result of intensive cropping through irriga-tion and fertilization that lead to enhanced plant growth

and subsequent increased organic carbon inputsinto the topsoil (Khan et al. 2009; Turner et al.2011; Zhang et al. 2010). The increase of carbon-ate might reflect increases of both CO2 productionand available Ca/Mg in soil profile. The formerwould be attributable to an increase in both SOCdecomposition (due to increased SOC) and rootrespiration (due to enhanced plant growth) whereasthe latter might be resulted from fertilization ofcalcium super-phosphate (Zhang et al. 2010), irri-gation with Ca-rich waters (Wu et al. 2008), and/or chemical weathering of calcium and magnesiumsilicate minerals (Baars et al. 2008; Moulton andBerner 1998).

Implications and future work

Our est imates of PIC accumulat ion rate (>38 g C m−2 y−1 over the 0–100 cm) in the cropland ofnorthern China are significantly higher than previouslyreported values (<3 g C m−2 y−1) for the arid and semi-arid regions in Canada, USA and New Zealand (Landiet al. 2003; Scharpenseel et al. 2000), but comparable tothose (10–40 g C m−2 y−1) for the surface soils ofAridisols in the northwest China (Pan and Guo 2000).While the large discrepancy is probably attributed to thedifferences in various factors between these regions, Caand Mg availability may be a key factor (Monger andGallegos 2000; Scharpenseel et al. 2000). Soils in theMojave Desert and semi-arid region of Canada may belimited by Ca andMg (Hirmas and Graham 2011; Landiet al. 2003; Monger and Gallegos 2000). On the otherhand, the cropland in northern China has been receivingextra Ca through fertilization (Zhang et al. 2010).However, application rate of Ca through fertilization(<6 g m−2 y−1) cannot explain the high rates of PICaccumulation, indicating other sources of Ca and/orMg.While little has been done to assess the sources of Ca

Table 5 Calculated PIC accumulation rate (g C m−2 y−1) over 0–100 cm under different fertilization treatments when setting theisotopic fractionation (IF) in Eq. (2) to be 14.9 and 12.9‰

Treatment Yangling Zhengzhou

IF=14.9

IF=12.9

Difference IF=14.9

IF=12.9

Difference

Control 21 13 8 33 20 13

NPK 86 71 15 57 40 17

NPKM 164 129 35 74 56 18

Mean 19 16

Table 6 Accumulation rates(g C m−2 y−1) of SOCa (0–20 cm)and SIC (0–100 cm) under dif-ferent fertilization treatments

aAccumulation rates of SOC werebased on those (Cong et al. 2012;Zhang et al. 2010), but using datafrom the period of 1990–2009

Treatment Urumqi Yangling Zhengzhou Mean

SOC SIC SOC SIC SOC SIC SOC SIC

Control −13 17 17 36 −5 77 0 43

NPK 0 63 43 175 20 66 21 101

NPKM 92 120 126 238 64 76 94 145

Mean 26 70 62 165 26 94 38 96

Plant Soil

and Mg at the LTE sites, we postulate that the possiblesources may include dust, irrigation water andweathering of calcium/magnesium silicate minerals.There is evidence of high concentrations of variousCa- or Mg-bearing salts in the ground water and riverwater (Aji et al. 2008), and high levels of Ca in aerosolsin arid and semi-arid regions across north China (Shenet al. 2009; Wang et al. 2005; Zhang and Iwasaka 1999).On the other hand, the minerals such as chlorite andsmectite at these LTE sites contain a considerableamount of Ca and/or Mg, which under high CO2 con-centration due to increased SOC and root respiration as aresult of fertilization, may be subject to enhanced chem-ical weathering (Baars et al. 2008). Further work isneeded to determine the sources and to quantify thefluxes of Ca and Mg in association with carbonateformation.

Apart from Ca and Mg availability, PIC formation isalso regulated by other factors, e.g., climatic condition(Breecker and Sharp 2008) and soil carbon content(Stevenson et al. 2005). Breecker et al. (2009) suggestedthat PIC formed onlywhen the soil was very dry and soilrespiration was limited due to moisture stress. On theother hand, Stevenson et al. (2005) reported that therewas a significant relationship between PIC and soilcarbon content along a bioclimatic gradient in thePalouse region, USA. Our study reveals greater increasein carbonate than in SOC, which is consistent with thefindings from the Brown and Grey soils of semi-aridregions, Canada (Landi et al. 2003), and Imperial soilfollowing nearly 90 years of irrigated farming in a semi-arid region in California, USA (Wu et al. 2008). Inaddition, our study shows a large increase of SIC stockunder long-term practice of straw incorporation andmanure addition. These findings suggest that increasingSOC stock in the arid and semi-arid regions would alsoenhance carbonate accumulation in soil profiles. Morestudies are needed to quantify the changes of majorcarbon pools (e.g., SOC, SIC and soil CO2) over varioustime scales, and the transformations and fluxes amongthese pools (e.g., transfer of SOC to SIC) in terrestrialecosystems, and to better understand the regulations ofphysical, biological and chemical processes from theregional to the global scales.

Acknowledgments This study is financially supported by theHundred Talented Program of the Chinese Academy of Sciences,the National Key Basic Research Program (2013CB956602) andthe Natural Science Foundation of China (41171239). We aregrateful for the reviewers’ constructive comments.

References

Aji K, Tang C, Song X, Kondoh A, Sakura Y, Yu J, Kaneko S(2008) Characteristics of chemistry and stable isotopes ingroundwater of Chaobai and Yongding River basin, NorthChina Plain. Hydrol Process 22:63–72

Baars C, Jones TH, Edwards D (2008) Microcosm studies of therole of land plants in elevating soil carbon dioxide andchemical weathering. Glob Biogeochem Cycles. doi:10.1029/2008gb003228

Breecker D, Sharp ZD (2008) A field and laboratory method formonitoring the concentration and isotopic composition ofsoil CO2. Rapid Commun Mass Spectrom 22:449–454

Breecker DO, Sharp ZD, McFadden LD (2009) Seasonal bias inthe formation and stable isotopic composition of pedogeniccarbonate in modem soils from central New Mexico, USA.Geol Soc Am Bull 121:630–640

Cerling TE (1984) The stable isotopic composition of modern soilcarbonate and its relationship to climate. Earth Planet Sci Lett71:229–240

Cerling TE, Quade J, Wang Y, Bowman JR (1989) Carbon iso-topes in soils and paleosols as ecology and paleoecologyindicators. Nature 341:138–139

Cerling TE, Solomon DK, Quade J, Bowman JR (1991) On theisotopic composition of carbon in soil carbon-dioxide.Geochim Cosmochim Acta 55:3403–3405

Conant RT, Paustian K, Elliott ET (2001) Grassland managementand conversion into grassland: Effects on soil carbon. EcolAppl 11:343–355

CongRH,WangXJ, XuMG, ZhangWJ, Xie LJ, YangXY, HuangSM, Wang BR (2012) Dynamics of soil carbon to nitrogenratio changes under long-term fertilizer addition in wheat-corn double cropping systems of China. Eur J Soil Sci 63:341–350

Eshel G, Fine P, Singer MJ (2007) Total soil carbon and waterquality: An implication for carbon sequestration. Soil Sci SocAm J 71:397–405

Eswaran H, Reich PF, Kimble JM, Beinroth FH, Padmanabhan E,Moncharoen P (2000) Global carbon stocks. In: Lal R,Kimble JM, Eswaran H, Stewart BA (eds) Global climatechange and pedogenic carbonates. Lewis Publishers, BocaRaton

Hirmas DR, Graham RC (2011) Pedogenesis and soil-geomorphicrelationships in an arid mountain range, Mojave Desert,California. Soil Sci Soc Am J 75:192–206

Khan S, Hanjra MA, Mu J (2009) Water management and cropproduction for food security in China: A review. Agric WaterManag 96:349–360

Lal R (2001) Potential of desertification control to sequestercarbon and mitigate the greenhouse effect. Clim Chang 51:35–72

Lal R (2004) Soil carbon sequestration impacts on global climatechange and food security. Science 304:1623–1627

Lal R, Kimble JM (2000) Pedogenic carbonate and the globalcarbon cycle. In: Lal R, Kimble JM, Eswaran H, StewartBA (eds) Global climate change and pedogenic carbonate.CRC Press, Boca Raton

Landi A, Mermut AR, Anderson DW (2003) Origin and rate ofpedogenic carbonate accumulation in Saskatchewan soils,Canada. Geoderma 117:143–156

Plant Soil

Liu WG, Yang H, Sun YB, Wang XL (2011) Delta C-13 values ofloess total carbonate: A sensitive proxy for Asian summermonsoon in arid northwestern margin of the Chinese loessplateau. Chem Geol 284:317–322

Manning DAC (2008) Biological enhancement of soil carbonateprecipitation: passive removal of atmospheric CO2. MineralMag 72:639–649

Mermut AR, Amundson R, Cerling TE (2000) The use of stableisotopes in studying carbonate dynamics in soils. In: Lal R,Kimble JM, Eswaran H, Stewart BA (eds) Global climatechange and pedogenic carbonates. CRC Press, Boca Raton

Mikhailova EA, Post CJ (2006) Effects of land use on soil inor-ganic carbon stocks in the Russian Chernozem. J EnvironQual 35:1384–1388

Monger HC, Gallegos RA (2000) Biotic and abiotic processes andrates of pedogenic carbonate accumulation in the southwest-ern United States-Relationship to atmospheric CO2 seques-tration. In: Lal R, Kimble JM, Eswaran H, Stewart BA (eds)Global climate change and pedogenic carbonate. CRC Press,Boca Raton

Monger HC, Cole DR, Buck BJ, Gallegos RA (2009) Scale andthe isotopic record of C4 plants in pedogenic carbonate: fromthe biome to the rhizosphere. Ecology 90:1498–1511

Moulton KL, Berner RA (1998) Quantification of the effect ofplants on weathering: Studies in Iceland. Geology 26:895–898

Murty D, KirschbaumMUF, McMurtrie RE, McGilvray H (2002)Does conversion of forest to agricultural land change soilcarbon and nitrogen? A review of the literature. Glob ChangBiol 8:105–123

Ogle SM, Breidt FJ, Paustian K (2005) Agricultural managementimpacts on soil organic carbon storage under moist and dryclimatic conditions of temperate and tropical regions.Biogeochemistry 72:87–121

Pan GX, Guo T (2000) Pedogenic carbonate of aridic soils inChina and its significance in carbon sequestration in terres-trial systems. In: Lal R, Kimble JM, Eswaran H, Stewart BA(eds) Global climate change and pedogenic carbonates. CRCPress, Boca Raton

Quade J, Cerling TE, Bowman JR (1989) Systematic vari-ations in the carbon and oxygen isotopic compositionof pedogenic carbonate along elevation transects in thesouthern Great-Basin, United-States. Geol Soc AmBull 101:464–475

Romanek CS, Grossman EL, Morse JW (1992) Carbon isotopicfractionation in synthetic aragonite and calcite—effects oftemperature and precipitation rate. Geochim CosmochimActa 56:419–430

Scharpenseel HW, Mtimet A, Freytag J (2000) Soil inorganiccarbon and global change. In: Lal R, Kimble JM, EswaranH, Stewart BA (eds) Global climate change and pedogeniccarbonates. CRC Press, Boca Raton

Schlesinger WH (1982) Carbon storage in the caliche of aridsoils—A case study from Arizona. Soil Sci 133:247–255

Schlesinger WH, Belnap J, Marion G (2009) On carbon seques-tration in desert ecosystems. Glob Chang Biol 15:1488–1490

Shen MX, Yang LZ, Yao YM, Wu DD, Wang JG, Guo RL, YinSX (2007) Long-term effects of fertilizer managements oncrop yields and organic carbon storage of a typical rice-wheatagroecosystem of China. Biol Fertil Soils 44:187–200

Shen Z, Caquineau S, Cao J, Zhang X, Han Y, Gaudichet A,Gomes L (2009) Mineralogical characteristics of soil dustfrom source regions in northern China. Particuology 7:507–512

Stevenson BA, Kelly EF, McDonald EV, Busacca AJ (2005) Thestable carbon isotope composition of soil organic carbon andpedogenic carbonates along a bioclimatic gradient in thePalouse region, Washington State, USA. Geoderma 124:37–47

Stone R (2008) Have desert researchers discovered a hidden loopin the carbon cycle? Science 320:1409–1410

Su YZ, Wang XF, Yang R, Lee J (2010) Effects of sandydesertified land rehabilitation on soil carbon sequestrationand aggregation in an arid region in China. J EnvironManag 91:2109–2116

Turner NC, Molyneux N, Yang S, Xiong Y-C, Siddique KHM(2011) Climate change in south-west Australia and north-west China: Challenges and opportunities for crop produc-tion. Crop Pasture Sci 62:445–456

Wang YQ, Zhang XY, Arimoto R, Cao JJ, Shen ZX (2005)Characteristics of carbonate content and carbon and oxygenisotopic composition of northern China soil and dust aerosoland its application to tracing dust sources. Atmos Environ 39:2631–2642

Wohlfahrt G, Fenstermaker LF, Arnone Iii JA (2008) Large annualnet ecosystem CO2 uptake of a Mojave Desert ecosystem.Glob Chang Biol 14:1475–1487

Wu LS, Wood Y, Jiang PP, Li LQ, Pan GX, Lu JH, Chang AC,Enloe HA (2008) Carbon sequestration and dynamics of twoirrigated agricultural soils in California. Soil Sci Soc Am J72:808–814

Xie JX, Li Y, Zhai CX, Li CH, Lan ZD (2009) CO2 absorption byalkaline soils and its implication to the global carbon cycle.Environ Geol 56:953–961

Yang XY, Ren WD, Sun BH, Zhang SL (2012) Effects of con-trasting soil management regimes on total and labile soilorganic carbon fractions in a loess soil in China. Geoderma177:49–56

Zhang D, Iwasaka Y (1999) Nitrate and sulfate in individual Asiandust-storm particles in Beijing, China in Spring of 1995 and1996. Atmos Environ 33:3213–3223

Zhang WJ, Wang XJ, Xu MG, Huang SM, Liu H, Peng C (2010)Soil organic carbon dynamics under long-term fertilizationsin arable land of northern China. Biogeosciences 7:409–425

Zheng J, Cheng K, Pan G, Pete S, Li L, Zhang X, Zheng J, Han X,Du Y (2011) Perspectives on studies on soil carbon stocksand the carbon sequestration potential of China. Chin SciBull 56:3748–3758

Plant Soil