Zhibo CHENG,Jingya WANG,William J.GALE,Haichang YANG and Fenghua ZHANG,∗College of Agronomy,Shihezi University,Xinjiang 832003(China)

2Key Laboratory of Oasis Ecology Agriculture of Xinjiang,Shihezi University,Xinjiang 832003(China)

ABSTRACTInformation on the effects of halophyte communities on soil organic carbon(SOC)is useful for sequestrating C in arid regions.In this study,we identified four typical natural halophyte communities in the Manasi River Basin in Xinjiang Province,Northeast China,namely,Karelinia caspia(Pall.)Less.,Bassia dasyphylla(Fisch.et C.A.Mey.)Kuntze,Haloxylon ammodendron(C.A.Mey.)Bunge,and Tamarix ramosissima Lour.We compared soil aggregation and aggregated-associated SOC under these communities.The aggregate fraction of 0.053-0.25 mm accounted for 47%-75%of the total soil mass,significantly more than the ＞0.25 and ＜0.053 mm fractions, under all the halophyte communities. Significant differences in soil aggregate size distribution were observed among the plant communities,with the H.ammodendron and B.dasyphylla communities showing the highest proportions of ＞0.25 mm aggregates(13.3%-43.8%)and T.ramosissima community having more ＜0.053 mm aggregates(14.1%-27.2%).Aggregate-associated SOC concentrations were generally the highest in the ＞0.25 mm fraction,followed by the ＜0.053 mm fraction,and were the lowest in the 0.053-0.25 mm fraction;however,because of their large mass,0.25-0.053 mm aggregates contributed significantly more to the total SOC.Total SOC concentrations(0-60 cm depth)decreased in the order of H.ammodendron(5.7 g kg-1)＞T.ramosissima(4.9 g kg-1)＞K.caspia(4.2 g kg-1)＞B.dasyphylla(3.4 g kg-1).The H.ammodendron community had the highest total SOC and aggregate-associated SOC,which was primarily because aggregate-associated SOC content at the 0-10 and 10-20 cm depths under this community were higher than those under other plant communities.The H.ammodendron community could be beneficial for increasing SOC in saline soils in the arid region.

Key Words: arid regions,C sequestration,soil aggregates,soil organic C,saline soils

INTRODUCTION

The soil carbon(C)pool is 3.3 times the size of the atmospheric C pool and 4.5 times the size of the biotic C pool(Lal, 2004). Soil organic carbon (SOC) sequestration has important influence on food production as well as soil,water,and air quality(Blanco-Canqui and Lal,2004;Jagadamma and Lal,2010).Soil aggregates are special organic-inorganic complexes and the basic unit of soil structure (Elliott et al., 1996; Six et al., 2000; Duiker et al., 2003). Several mechanistic processes of SOC stabilization have been proposed, including i) encapsulation of SOC within and between aggregates, ii) interaction of SOC with primary soil particles to form organic mineral complexes, and iii)selective preservation of organic compounds because of their inherent recalcitrance(Zinn et al.,2007;Wilson et al.,2009;Six and Paustian,2014).The distribution of aggregateassociated SOC determines the capacity of a soil to store and retain C(Mohammadi and Motaghian,2011).Soil aggregate distribution also affects soil aeration, soil erodibility, and water permeability. Therefore, soil aggregation and SOC distribution have an important influence on soil structure and fertility.

Tisdall and Oades(1980)observed that SOC concentrations were greater in macroaggregates than in microaggregates.They attributed the difference to the presence of decomposing roots and hyphae within macroaggregates.However,other studies have indicated that microaggregates have the highest SOC concentrations (Tisdall and Oades,1982;Six et al.,2000;Lorenz et al.,2008).In a vegetated black soil profile,the 0.053-0.25 mm fraction had the highest aggregate-associated SOC concentration among all the fractions(Yuan et al.,2012).Many researchers have studied the effects of land use and agroecosystems on soil aggregation and aggregate-associated SOC (Wright and Hons, 2005;Tan et al.,2007;Jagadamma and Lal,2010).However,few studies have investigated aggregate-associated SOC in saline soils undergoing natural vegetative succession.

The Manasi River Basin in the Xinjiang Uyghur Autonomous Region of Northeast China is the fourth largest irrigated agricultural area in China.Within the basin,soil salt concentrations are high in the alluvial fan area.A growing area of land is being converted for agricultural production.Land and water resources in the region are being used indiscriminately.Overall,these conditions have contributed to the abandonment of some farmlands because of secondary salinization. Halophyte vegetation often appears early in the reclamation process of such soils in study area and is considered a valuable step toward reclaiming the land for conventional crops because of its benefciial effects on soil properties (Knights and Phillips, 1979; Yi et al., 2007).Halophytes have the potential to directly sequester up to 0.7 Gt C,and they can play an indirect role in absorbing C from the atmosphere(Glenn et al.,1992).Therefore,the ecological benefti of halophytes has become an important research focus.Relatively few studies have examined soil aggregate distribution and aggregate-associated SOC under these halophyte communities,probably because of their low biomass and nutrients, in arid areas (Maestre and Cortina, 2003).More information on the distribution of aggregate-associated SOC in arid saline areas is required.The objectives of this study were to i) compare soil aggregate size distribution and aggregate-associated SOC concentrations under natural halophyte communities in the Manasi River Basin and ii)determine the contribution of aggregate-associated SOC to total SOC under each halophyte community.

MATERIALS AND METHODS

Study area

The Manasi River Basin is located at the center of Eurasia,along the southern edge of the Junggar Basin(44◦64′N,86◦03′E,354 m above sea level),in the Xinjiang Uyghur Autonomous Region of Northeast China.The area has an arid,continental climate with an average annual frost-free period of 169 d.The average annual precipitation is 153 mm;average annual evaporation is 2005 mm,which is 13 times greater than the amount of precipitation.The high evaporation rates can cause soil salinization and,consequently,abandonment of agricultural land.The watertable in the area is high(Fig.1).The soil is primarily classified as saline sierozem,and the area is covered by discontinuous shrub layers and relatively continuous herbaceous layers (Tu et al., 2011). Most of the vegetation consists of xerophytic,super-xerophytic,and saline-resistant species.

Experimental sites

Fig.1 Sketch map of study area and water table in the area.

We identified four typical natural halophyte communities at different sites(Sites I-IV),Karelinia caspia(Pall.)Less.(Site I), Bassia dasyphylla (Fisch. et C. A. Mey.) Kuntze(Site II),Haloxylon ammodendron(C.A.Mey.)Bunge(Site III), and Tamarix ramosissima Lour (Site IV), on salinealkaline land that had been abandoned for more than 20 years in the Manasi River Basin.The sites,located within 50 m of each other,were similar in slope,aspect,and elevation.The structural characteristics and species composition of each community were determined using standard survey methods.Three plots(20 m×20 m)were surveyed for each community to determine the species composition,canopy height,and canopy cover(Table I).

Soil sample collection

Soil samples were collected in mid-May in 2013 and 2014.The samples were collected by randomly digging three pits(30 cm wide×100 cm deep)at each site.Undisturbed soil samples were taken from depths of 0-10,10-20,20-40, and 40-60 cm with a flat spatula. The samples were immediately transferred to the laboratory,where they were combined to form one composite sample (approximately 1 kg) per depth per site. These composite samples were thoroughly mixed and then divided into quarters.The soil in three of the quarters was air dried and then stored at room temperature (25◦C). Three subsamples of this soil were separately analyzed as described below to determine aggregate size distribution at each depth of each site.Each subsample was considered a replicate.The remaining soil was air dried,passed sequentially through 1-and 0.25-mm soil sieves,and then stored for chemical analysis(Table II).

Soil sample analyses

Three subsamples from a depth of 0-20 cm(prepared by compositing soil samples of 0-10 and 10-20 cm depths)were analyzed to determine soil physicochemical properties at each site,including soil moisture,pH,electrical conductivity(EC), and contents of organic matter, total and available N,total and available P,and total and available K.Soil pH and EC were measured in a soil-water suspension(1:5 and 1:2.5, respectively), and SOC was determined using the Walkley-Black dichromate oxidation method(Díaz-Zorita,1999). Total N was measured using the semi-micro Kjeldahl method,and available soil N was determined using the alkaline hydrolysis diffusion method(Mulvaney and Khan,2001).Total P was measured using NaOH extraction,Mo-Sb colorimetry,and ultraviolet spectrophotometry,and available P was determined using the 0.5 mol L-1NaHCO3extraction method (Page et al., 1982). Total K was measured using atomic absorption spectrometry after digestion with solid natrium hydroxide, and available K was measured using the 1 mol L-1NH4Ac extraction method.Soil bulk density at a depth of 0-20 cm was determined using a soil core(100 cm3).

Aggregate fractionation was performed using a previously described dry-sieving method (Zhang et al., 2003).Three air-dried subsamples from each site were fractionated into ＞0.25, 0.25-0.053, and ＜0.053 mm aggregate size fractions with a sieve shaker. After weighing, the aggregate fractions were analyzed to determine SOC as described above.

Statistical analysis and calculations

All data were analyzed with Excel 2003 and SPSS version 11.5 software.Least significant difference tests were used to compare mean values. The contribution of aggregateassociated SOC to total SOC(CASOC,%)was calculated according to the following equation(Gao et al.,2010):

where ASOCiis the aggregate-associated SOC content in size class i(mg g-1),AMF is the aggregate mass fraction of the total soil sample(%)(i.e.,soil mass in size class i divided by the total weight of the soil), and TSOC is the SOC content of the total soil sample(mg g-1).

RESULTS AND DISCUSSION

Soil characterization

At the 0-20 cm depth,soil moisture and pH were the highest under the H.ammodendron community(Table II).Thegreatest levels of bulk density and EC were observed under the B.dasyphylla and K.caspia communities,respectively.Soil organic matter and total N contents were significantly greater under the H. ammodendron and T. ramosissima communities than under the K. caspia and B. dasyphylla communities(P ＜0.05).However,soil total P content did not differ significantly among the plant communities.Soil total K content was the lowest(21.6 g kg-1)under the K.caspia community;in contrast,the highest content of soil available N was observed under the K.caspia community.The highest soil available P and K contents were observed under the B.dasyphylla community.

TABLE IDescription of the natural plant community at each study site

Soil aggregate distribution

In arid and semi-arid regions, the distribution of soil aggregates is closely related to water use and wind erosion(Cantón et al.,2009;Li et al.,2014).Soil erosion resistance generally increases as the proportion of macroaggregates increases(Wang et al.,2010).The distribution of dry-sieved aggregates partly reflects soil quality in these regions.Soil aggregates are generally divided into macroaggregate (＞0.25 mm) and microaggregate (＜0.25 mm) fractions. In this study, the 0.25-0.053 mm aggregates accounted for 46.7%-74.6%of the total soil mass(Fig.2),significantly greater than the ＞0.25 and ＜0.053 mm fractions,which accounted for 10.6%-43.8%and 7.9%-27.2%of the total soil mass,respectively.Total SOC concentrations were low(＜10 g kg-1), and soil pH, EC, and bulk density were high (Table II). Overall, these findings indicated that soil structure and quality were poor in the study area.The soil was affected by low precipitation and sparse vegetation.Under these conditions,macroaggregates breakdown,and the proportion of microaggregates increases.

At all four soil depths, the highest percentage of ＞0.25 mm aggregates was generally observed under either H.ammodendron or B.dasyphylla community.The percentage of ＞0.25 mm aggregates at the 0-10 cm depth was highest under the H.ammodendron community.The percentage of＞0.25 mm aggregates decreased as the depth increased under the H.ammodendron community.In contrast,under the B.dasyphylla community,the percentage of ＞0.25 mm aggregates at the 0-10 cm depth was relatively low,but those at the 10-20,20-40,and 40-60 cm depths were higher.The same pattern was observed under the K.caspia community.Under the T. ramosissima community, the percentage of＞0.25 mm aggregates at the 0-10 and 20-40 cm depths were the second highest,but that at the 10-20 and 40-60 cm depths was the lowest among the four communities.

TABLE IISelected physicochemical propertiesa) of the soil(0-20 cm)at the study sites

Fig.2 Soil aggregate size(＞0.25,0.25-0.053,and ＜0.053 mm)distribution at different depths(0-10,10-20,20-40,and 40-60 cm)affected by the natural halophyte communities at the study sites.Site I:Karelinia caspia(Pall.)Less.community;Site II:Bassia dasyphylla(Fisch.et C.A.Mey.)Kuntze community;Site III:Haloxylon ammodendron(C.A.Mey.)Bunge community;Site IV:Tamarix ramosissima Lour.community.Values are means of three replicates with standard deviations shown by vertical bars.Bars with different lowercase letters indicate significant differences among the aggregate size classes within a depth in a site(P ＜0.05).Bars with different uppercase letters indicate significant differences among the sites within an aggregate size class in a depth(P ＜0.05).

Among the plant communities,the percentage of 0.25-0.053 mm aggregates at all depths except 40-60 cm was the highest under K.caspia community,and that the 40-60 cm was the highest under the T.ramosissima community.Under the K.caspia and B.dasyphylla communities,the percentage of 0.25-0.053 mm aggregates was greater in the surface soil(0-10 cm)than at the lower depths.In contrast,under the H.ammodendron and T.ramosissima communities,the percentage of 0.25-0.053 mm aggregates was the greatest at the 40-60 cm depth.

Under the T.ramosissima community,percentage of ＜0.053 mm aggregates at the 0-10, 10-20, and 20-40 cm depths was the highest among the plant communities.The differences among the K. caspia, B. dasyphylla, and H.ammodendron communities were not significant at these depths.At the 40-60 cm depth,＜0.053 mm aggregates were significantly more under the K.caspia,H.ammodendron,and T. ramosissima communities than under the B. dasyphylla community.The differences among the K.caspia,H.ammodendron,and T.ramosissima communities were not significant.

Total SOC

Some studies have shown that organic matter promotes aggregate formation(Six et al.,2004;Bronick and Lal,2005;Kögel-Knabner et al.,2008).Total SOC is mainly affected by plant litter,plant roots,animals,and microorganisms(Six et al.,2000).Total SOC content was the highest at the 0-10 cm depth,ranging from 5.7 to 7.6 g kg-1(Fig.3).Total SOC content declined significantly as the soil depths increased.These results are consistent with those of previous studies and can probably be attributed to the contribution of surface litter to SOC in the upper part of the soil profile(Jiménez et al., 2011). At the 40-60 cm depth, total SOC content ranged from 2.0 to 2.7 g kg-1. Total SOC content at the 0-10 and 10-20 cm depths was significantly higher under the H.ammodendron and T.ramosissima communities than under the B.dasyphylla and K.caspia communities.At of the 20-40 cm depth, total SOC content was significantly lower under the B.dasyphylla community than under all the other communities.When compared with other halophyte plant communities, H. ammodendron and T. ramosissima communities have a rich species diversity,very developed root system,and large crown.However,Site II is dominated by single shrub (B. dasyphylla) plants, and most of them are annual undersized herbs.Therefore,the Haloxylon and Tamarix communities input a higher amount of organic C into the soil (Li et al., 2007). No significant differences were observed among the H.ammodendron,T.ramosissima,and K. caspia communities. At the 40-60 cm depth, no significant difference in total SOC was found among the communities. Overall, total SOC content at the 0-60 cm depth decreased under the plant communities in the order of H. ammodendron (5.7 g kg-1) ＞T. ramosissima (4.9 g kg-1)＞K.caspia(4.2 g kg-1)＞B.dasyphylla(3.4 g kg-1).

Fig.3 Total soil organic C(SOC)content at different depths(0-10,10-20,20-40,and 40-60 cm)affected by the natural halophyte communities at the study sites.Site I:Karelinia caspia(Pall.)Less.community;Site II:Bassia dasyphylla (Fisch. et C. A. Mey.) Kuntze community; Site III:Haloxylon ammodendron(C.A.Mey.)Bunge community;Site IV:Tamarix ramosissima Lour.community.Values are means of three replicates with standard deviations shown by vertical bars.Bars with different lowercase letters indicate significant differences among the depths within a site(P ＜0.05).Bars with different uppercase letters indicate significant differences among the sites within a depth(P ＜0.05).

Aggregate-associated SOC

Aggregate-associated SOC content was generally significantly higher in the ＞0.25 mm fraction,followed by the ＜0.053 and 0.053-0.25 mm fractions(Fig.4).Some studies have indicated that aggregate-associated SOC content is higher in macroaggregates than in microaggregates(Elliott,1986;Xie et al.,2008).For example,macroaggregates accounted for the highest proportion of aggregate-associated SOC after revegetation(Zhao et al.,2006).In contrast,other studies have shown that aggregate-associated SOC is the highest in the ＜0.053 mm fraction, which accounted for 44%-88%of the total soil mass(Brodowski et al.,2006).Liu and Yu(2011)reported increasing SOC with decreasing aggregate diameter and the highest SOC in the ＜0.25 mm fraction. In the present study, aggregate-associated SOC content was significantly higher in the ＞0.25 and ＜0.053 mm fractions than in the 0.25-0.053 mm fraction.High SOC in the ＞0.25 mm aggregates is generally attributed to plant roots and fungal hyphae, which bind microaggregates to form macroaggregates(Tisdall and Oades,1980;Rillig et al., 2002). Some researchers have suggested that SOC is younger in macroaggregates than in microaggregates(Puget et al., 2000; Ma et al., 2013). Organic C is more readily mineralized in macroaggregates than in microaggregates.

Fig.4 Aggregate-associated soil organic C(SOC)in different aggregate size classes(＞0.25,0.25-0.053,and ＜0.053 mm)at different depths(0-10,10-20,20-40,and 40-60 cm)affected by the natural halophyte communities at the study sites.Site I:Karelinia caspia(Pall.)Less.community;Site II:Bassia dasyphylla(Fisch.et C.A.Mey.)Kuntze community;Site III:Haloxylon ammodendron(C.A.Mey.)Bunge community;Site IV:Tamarix ramosissima Lour.community.Values are means of three replicates with standard deviations shown by vertical bars.Bars with different lowercase letters indicate significant differences among the aggregate size classes in a depth within a site(P ＜0.05).Bars with different uppercase letters indicate significant differences among the sites within an aggregate size class in a depth(P ＜0.05).

The aggregate-associated SOC content generally decreased with depth under all four plant communities. At the 0-10 cm depth,aggregate-associated SOC content was the highest in all three fractions under the T.ramosissima community.At the 10-20 cm depth,the content of aggregateassociated SOC was the highest in all three fractions under the

H.ammodendron and T.ramosissima communities,whereas at the 20-40 cm depth,aggregate-associated SOC content was the highest under the K.caspia community.Plant roots in the K. caspia community are mainly distributed at the 20-40 cm depth.These roots and their secretions would have altered the soil physical,chemical,and biological properties related to aggregate formation(Rillig et al.,2002).

Pearson’s correlation between aggregate size and aggregate-associated SOC content was significant for the 0.25-0.053 mm fraction(r =-0.50),but not for the ＞0.25 mm fraction (r = -0.19) and ＜0.053 mm fraction (r =0.41).This was mainly because the content of the 0.25-0.053 mm fraction was significantly greater than that of the other fractions.

At the 0-60 cm depth, no significant difference in total aggregate-associated SOC was observed among the K.caspia (6.8 g kg-1), H. ammodendron (6.7 g kg-1), and T. ramosissima (6.9 g kg-1) communities (Fig. 5). The aggregate-associated SOC content was significantly lower under the B.dasyphylla community(4.8 g kg-1)than the other three communities.Some studies have shown that the half-spherical crown of T. ramosissima is conducive for capturing and protecting plant litter(Li et al.,2007).This increases litter input to soil. In the H. ammodendron and T.ramosissima communities,species richness is high,root systems are well developed,and vegetation crown is large.In contrast,the B.dasyphylla community consists primarily of short,annual herbs with relatively low biomass.Therefore,C input to soil is greater under the H.ammodendron and T.ramosissima communities than community the B.dasyphylla community.

Fig.5 Total aggregate-associated soil organic C(SOC)content affected by the halophyte communities at the study sites.Site I:Karelinia caspia(Pall.)Less.community;Site II:Bassia dasyphylla(Fisch.et C.A.Mey.)Kuntze community;Site III:Haloxylon ammodendron(C.A.Mey.)Bunge community;Site IV:Tamarix ramosissima Lour.community.Values are means of three replicates with standard deviations shown by vertical bars.Bars with different letters indicate significant differences among the sites(P ＜0.05).

Contribution of aggregate-associated SOC to total SOC

As described above,the 0.25-0.053 mm fraction generally had the lowest aggregate-associated SOC content,but the highest mass among the three size fractions. As a result,the 0.25-0.053 mm aggregates generally contributed significantly more to total SOC than the ＞0.25 and ＜0.53 mm aggregates (Table III). Jastrow et al. (1996) pointed out that microaggregate structure gives greater protection to SOC than macroaggregate structure.This was observed atall depths under the K.caspia,B.dasyphylla,and T.ramosissima communities.It should be noted that at the 10-20 cm depth under the B.dasyphylla community,no significant difference in percentage contribution was observed between the ＞0.25 and 0.25-0.053 mm aggregates. Under the H.ammodendron community,the contribution of ＞0.25 mm aggregates to total SOC was as large or larger than that of 0.25-0.053 mm aggregates at of the 0-10 and 10-20 cm depths. However, at the 20-40 and 40-60 cm depths, the 0.25-0.053 mm aggregates contributed significantly more to total SOC than the other aggregate classes.Among the size fractions,＜0.053 mm aggregates generally contributed the least to total SOC.Thus,microaggregates have the greatest ability to protect SOC in saline soils in arid areas.This is mainly because organic and inorganic colloids in microaggregates can closely integrate and hold C,which is not easily decomposed and released by microbes(Mao et al.,2007).The exception was at the 0-10 and 10-20 cm depths under the T.ramosissima community,where ＜0.053 mm aggregates contributed significantly more to total SOC than ＞0.25 mm aggregates.

TABLE IIIContribution of aggregate-associated soil organic C(SOC)to total SOC affected by the natural halophyte communities at the study sites

CONCLUSIONS

The halophyte communities had significant effects on the soil aggregate distribution and SOC content.Aggregateassociated SOC content was the highest in the ＞0.25 and＜0.53 mm fractions. The 0.053-0.25 mm fraction had the lowest SOC content.Overall,the ＜0.053 mm fraction contributed the least to total SOC.Thus,microaggregates had the greatest ability to protect SOC in saline soils in arid areas.The total SOC and aggregate-associated SOC contents were the highest under the H. ammodendron community.Therefore, of the halophyte communities in this study, H.ammodendron was the best for increasing SOC in arid areas.

ACKNOWLEDGEMENTS

This research was supported by the National Natural Science Foundation of China(No.31860360)and the National Key R&D Program of China(No.2016YFC0501406).We gratefully acknowledge the helpful comments by the anonymous referees,who inspired significant improvements in the analysis and writing.