Land use effects on the dynamics of soil C,N and microbes in the water-wind erosion crisscross region of the northern Loess Plateau,China

Yi WANG ,Chunyue LI and Shunjin HU

1State Key Laboratory of Loess and Quaternary Geology,Institute of Earth Environment,Chinese Academy of Sciences,Xi’an 710061(China)

2State Key Laboratory of Soil Erosion and Dryl and Farming on the Loess Plateau,Institute of Soil and Water Conservation,Northwest A&F University,Yangling 712100(China)

3School of Geography and Tourism,Shaanxi Normal University,Xi’an 710119(China)

4Department of Entomology&Plant Pathology,North Carolina State University,Raleigh 27695(USA)

ABSTRACT The water-wind erosion crisscross region of the northern Loess Plateau in China is under constant pressure from severe erosion due to its windy and dry climate and intensive human activities.Identifying sustainable land use patterns is key to maintaining ecosystem sustainability in the area.Our aim was to appraise the impacts of different land use regimes on the dynamics of soil total organic C(TOC),total N(TN),and microbes in a typical watershed in the northern Loess Plateau to identify suitable land use types that can maintain soil fertility and sustainability.A field experiment was performed in Liudaogou watershed in Shenmu City,Shaanxi Province,China,where the dynamics of soil TOC and TN,microbial biomass C and N,microbial respiration,and net N mineralization in six typical land use types,dam land,rainfed slope land,deciduous broadleaf forest,evergreen coniferous forest,shrubland,and grassland,were measured in three different growing seasons.Land use type and season significantly affected TOC,TN,and the dynamics of microbial biomass and activity.As the most anthropogenically disturbed land use pattern,dam land was an optimal land use pattern for TOC sequestration due to its higher TOC and TN,but lower microbial activity.Soil TOC,TN,and microbial properties demonstrated a decreasing trend after natural grassland was converted to shrubland,forest,and rainfed slope land.Shrubland with exotic N-fixing Korshinsk peashrub(Caragana korshinskii Kom.)can maintain TOC,TN,and microbial properties similar to those in grassland.Soil TOC,,TN,moisture,and extractable C were the principal indexes for soil microbial biomass and activity and explained 88.90%of the total variance.Thus,grassland was the optimal land use pattern in the water-wind erosion crisscross region of the northern Loess Plateau to maintain ecosystem stability and sustainability.

Key Words: grassland,microbial activity,microbial biomass,total organic C,total N,watershed

INTRODUCTION

Land is a crucial natural resource on Earth,providing ecosystem services for human survival and human well-being(Costanzaet al.,1997;Lambinet al.,2003).However,as the demand for food,feed,fiber,and fuel rapidly grows,ecosystem services such as soil erosion and accumulation,water cycling,and C sequestration are weakened or even inhibited by some land uses (Anacheet al.,2018;Liuetal.,2018;Tanget al.,2018).As land use is affected by human activities with subsequent impacts on soil ecosystem functions,researching soil processes in different land uses is critical for the sustainable use and management of land resources,especially in arid and semiarid ecologically fragile ecotones that cover 41%of the terrestrial surface,sustain over 38% of the population (Reynoldset al.,2007),and dominate the dynamics of land C sequestration to mitigate climate change(Post and Kwon,2000;Gamboa and Galicia,2011;Ahlströmet al.,2015;Liuet al.,2018).

Soil stocks the largest amounts of organic C(OC,1.5×103Pg)and play important roles in maintaining ecosystem functions and mitigating global warming(Lal,2004).Generally,maintenance of rootzone C at 1.5%-2.0%is essential for soil structure and aggregation,water retention and use efficiency,nutrient retention and use efficiency,rhizospheric processes,and gaseous emissions(Lal,2016b).It is estimated that increasing the C concentration in the agricultural soil in Iowa in USA by 100 kg C ha-1could decrease soil water erosion by 350 kg year-1,soil wind erosion by 0.4 kg year-1,and soil N runoffby 0.14 kg year-1(Kurkalovaet al.,2004).It is also estimated that increasing the global soil C pool to 3 m depth by 4%could decrease atmospheric CO2-C by 240 Pg or reduce CO2by more than 100 ppm(Lal,2016a).As enhanced C sequestration in soil has many benefits,the“4 per Thousand”initiative proposes to increase the global 0-40 cm soil total OC(TOC)content at 0.4%per year(Lal,2016a).However,soil TOC dynamics are mainly driven by erosion,and the removed soil(by either water or wind erosion)is 1.3-5 times full of organic matter than the left soil(Pimentelet al.,1995;Berheet al.,2018;Lal,2019).Soil C lost could be transported by runoffwater when attached to soil particles,dissolved OC(DOC),or CO2due to microbial heterotrophic respiration(Galyet al.,2015;Manninenet al.,2018).

Soil microbes make up over 95%of soil biomass and play fundamental roles as primary producers and decomposers in soil C and N cycling and storage(Doran,1987;Bardgett and Van Der Putten,2014).Soil microbes are mainly heterotrophic;they use plant (leaf,root litter,and root exudate) C as energy and emit 60 Pg annually from the TOC pool by respiration (Singhet al.,2010).As soil microbes are a reactive source of C and N,the activity of land management greatly impacts C dynamics and C sequestration.For example,no-tillage practices could enhance TOC and promote stable C pools in agricultural fields(Wanget al.,2011).Microbial biomass and activity are responsive to land use change.For example,soil microbial biomass decreased 4-5 times after forest and grassland changed to agricultural land,with decreased soil microbial activity(Van Leeuwenet al.,2017).Therefore,researching the dynamics of soil microbes under different land uses can shed light on C cycles.

The Loess Plateau of China is distinguished by deep loess deposits,complex terrain,drought conditions,severe soil erosion,and fragile ecosystems in a 6.4×105km2arid and semiarid ecotones.Extensive anthropogenic disturbances,such as long-term cultivation,grazing,and development of energy industries,have led to harsh soil and water loss as well as vegetation and ecosystem degradation(Yanet al.,2005;Yueet al.,2016).There is a typical hilly region named the“water-wind erosion crisscross region”that lies from northeast to southwest on the Loess Plateau (Tang,1996).This region covers 28.56%of the Loess Plateau,and the erosion that is more serious for both water and wind erosion can promote each other (Fanet al.,2010).The eroded soil from this region contributed to a majority of the sediment load to the Yellow River(China),although the load decreased in the last decade(Zhanget al.,2019).Since the 1950s,a series of environmental protection programs have been launched by the government to alleviate severe erosion and environmental degradation in this area.The land use was changed dramatically by the Three-North Shelter Forest Program and the Grain for Green Program,with large cropland areas being reconverted to forest,shrubland,and grassland(Lüet al.,2012b)

Establishing a long-term sustainable land use pattern is urgently needed to protect soil C storage and nutrient conservation and to minimize soil erosion.Several studies have investigated the variations in soil C and N in distinct land uses (Weiet al.,2009;Wanget al.,2016;Jiaet al.,2017b;Geet al.,2020),and some have investigated the microbial properties in different land uses(Liuet al.,2018;Renet al.,2018).However,the dynamics of microbialmediated soil C and N under different land use types remain poorly understood in the most fragile water-wind erosion crisscross region.In this study,we investigated soil TOC,N,and microbial biomass and activity in six typical land use patterns (cropland:dam land and rainfed slope land,forest:deciduous broadleaf forest and evergreen coniferous forest,shrubland,and grassland)in the water-wind erosion crisscross region’s center.We hypothesized that i)dam land is an optimal land use pattern to sequester more soil TOC and TN and ii)soil TOC,TN,and moisture were the principal indicators for the soil microbial biomass and activity in the water-wind erosion crisscross region.The results of this research could determine which land use type sustainably facilitates C sequestration and maintains soil nutrient fertility in land management in the northern Loess Plateau.

MATERIALS AND METHODS

Site description

The experiment was conducted in Liudaogou watershed,Shenmu City,Shaanxi Province,China(110°21′-110°23′E,38°46′-38°51′N,1 081-1 274 m altitude,and 6.89 km2).The watershed is located in the center of the water-wind erosion crisscross region with severe water erosion(in summer and autumn)and wind erosion(in winter and spring).The landscape is characterized by deep gullies and undulating slopes formed in the deep (up to 100 m) loess deposits.Liudaogou has a typical semiarid continental monsoon climate with more than 75% of the precipitation occurring from June to September,and the annual precipitation is 437 mm,while the potential evapotranspiration is 1 337 mm.The mean annual temperature is 8.4°C.The dominant soil type is Calcaric Regosols (Wanget al.,2015).Farmland,grassland,shrubland,and forest mosaic covered 8%,59%,22%,and 6%of the watershed,respectively.Bunge needlegrass(Stipa bungeanaTrin.)is the dominant native plant.Simon poplar(Populus simonii),Chinese pine(Pinus tabulaeformis),and Korshinsk peashrub(Caragana korshinskiiKom)were planted to prevent soil erosion in the late 1960s.

The history of the land use types was determined by interviews with village farmers and elders.Six main land use patterns were selected in the experiment:dam land(DL),established in 1964,can be irrigated and has been grown with maize(Zea maysLinn.);rainfed slope land(RL),opened up from natural grassland since the 1900s,cannot be irrigated and planted with potato (Solanum tuberosumL.);poplar forest(POF),with a Simon poplar(Populus simonii)plantation in natural grassland since the 1960s;pine forest(PIF),planted in natural grassland in the 1960s with Chinese pine(Pinus tabulaeformis);Korshinsk peashrub (KS),planted with Korshinsk peashrub(Caragana korshinskiiKom.)in natural grassland in the 1960s;and Stipa grassland (SG),natural grassland dominated by bunge needlegrass (StipabungeanaTrin.).Detailed land conditions are given in Table I.The landscape and vegetation conditions are presented as photographs in Fig.S1(see Supplementary Material for Fig.S1).

Soil sampling

Under each land use type,we randomly set up three replicate plots(10 m×10 m)at least 50 m away.Soil samples were randomly collected in April,July,and October 2013 by the five sampling point method from 0-10 cm depth.All samples in each plot were completely mixed and passed through a 2.0-mm sieve.Subsamples were kept at 4°C for microbiology analyses,and the rest were air-dried for physiochemical testing.

Soil properties

Physical properties.Soil pH was measured with a digital pH meter(PHS-3C,INESA Scientific Instrument Co.,Ltd,China)at a soil:water ratio of 1:2.5(weight:volume).Soil texture was analyzed by a Mastersizer 3000(Malvern Instruments,Malvern,UK) after pretreatment with 30%H2O2,(NaPO3)6,and sonication.The particle size was classed into clay (＜0.002 mm),silt (0.02-0.002 mm),fine sand (0.02-0.2 mm),and coarse sand (0.2-2 mm)according to the International Soil Science Society(ISSS)classification.Soil physical clay (SPA) was＜0.01 mm according to the Kachinsky classification(Russian system).Soil microaggregates (＜0.25 mm) were analyzed by a Mastersizer 3000 without any dispersion.

Total OC and TN.Soil TOC and TN were tested by an elemental analyzer(Elementar vario MACRO cube,Elementar,Germany)after soil milling to fine powder.

Extractable C and N.A soil sample(20.0 g dry weight)was digested with 0.5 mol L-1K2SO4(50 mL)to determine extractable C by a TOC analyzer (TOC-VWP,Shimadzu Corporation,Japan)(Huet al.,1997).To estimate extractable,20.0 g samples were mixed with 1 mol L-1KCl(200 mL)and then analyzed using an Auto Analyzer 3(SEAL Analytical,ltd.,Germany)to determine the concentration.

Microbial biomass C and N.Soil samples (20.0 g dry weight equivalent at 60%water-holding capacity)were fumigated for 48 h with chloroform(high-performance liquid chromatography grade)protected from light(Vanceet al.,1987;Ross,1992).The nonfumigated and fumigated soil samples were extracted with 50 mL of 0.5 mol L-1K2SO4.The difference in OC(tested by a TOC analyzer)between nonfumigated and fumigated soil was used to calculate microbial biomass C(MBC)with a factor of 0.33(Sparling and West,1988).The difference in N content(determined by an Auto Analyzer 3)between fumigated and nonfumigated soil was used to calculate microbial biomass N(MBN)with a factor of 0.45(Jenkinson,1988;Cabrera and Beare,1993).

Microbial respiration.Soil samples(20.0 g dry weight equivalent with 60%water-holding capacity)was placed in a 250-mL glass bottle with a beaker containing 5 mL of 0.5 mol L-1NaOH (Alef and Nannipieri,1995;Forster,1995;Hu and Van Bruggen,1997).The bottles were sealed and incubated at 25°C protected from light.After 7 d of incubation,beakers and NaOH were renewed and incubated for an additional 7 d.The absorbed CO2in NaOH was titrated by HCl(0.1 mol L-1)to calculate soil microbial respiration rate(MRR)(Colemanet al.,1977).

TABLE ISite conditions of the six selected land use types in Liudaogou watershed in Shenmu City,Shaanxi Province,China

Net N mineralization.Soil samples(10.0 g dry weight equivalent)was placed in flasks,covered with plastic wraps with a small hole and incubated for 28 d at 25°C protected from light.Soil moisture was adjusted to 60%water-holding capacity during the experiment.Soil inorganic N(and) was determined by an Auto Analyzer 3 after extraction with 1 mol L-1KCl (1:10,soil:solution,weight:volume).The difference in inorganic N contents before and after a 28-d incubation was the net N mineralized(NNM).

Data analysis

Soil TOC and TN were statistically analyzed by analysis of variance(ANOVA),as they were more stable and only tested in October 2013.Other parameters were tested in each sampling season and statistically analyzed by repeatedmeasures ANOVA.Pearson correlation analysis was used to evaluate the relationship between the tested indicators.Both ANOVA and repeated-measures ANOVA were conducted using SAS 9.0 (SAS Systems,USA) by least significant difference test at the 0.05 level.Redundancy analysis(RDA)was applied by CANOCO 5.0(Microcomputer Power,Inc.,USA).Detrended correspondence analysis was conducted before RDA,and the results showed that the largest gradient length was 0.32(fewer than 3),which certified that the linear ordination method was fit to test the microbial data.Monte Carlo analysis with 9 999 permutations was used to select the significance of environmental variables(P ＜0.05).All data were found on oven-dry soil weight and presented as the means±standard errors.

RESULTS

Soil TOC and TN

Land use patterns significantly influenced soil TOC(Fig.1).Specifically,DL,KS,and SG had significantly higher soil TOC than RL (by 45%-46%) and POF(by 111%-113%).Soil TOC in POFwas significantly lower than that under the other land use types(P ＜0.05).Differences were not statistically significant in TN between PIFand the other land use types,with the exception of DL and POF.The lowest TN occurred in POFand was 68% lower than the highest level in DL.The TN levels in KS,SG,RL,and PIF did not differ,but were all significantly(P ＜0.05)higher than that in POF.Soi C/N ratio ranged from 17 to 25 under different land use types.

Fig.1 Soil total organic C and N under different land use types in Liudaogou watershed in Shenmu City,Shaanxi Province,China.Values are means with standard errors of means shown by vertical bars(n=3).Bars with different letters indicate significant differences at P ＜0.05.DL=dam land;RL=rainfed slope land;POF=poplar forest;PIF=pine forest;KS=Korshinsk peashrub;SG=Stipa grassland.

Soil extractable C and N

Land use influenced soil extractable C(Fig.2,Table II).The highest soil extractable C was found in KS and was higher than that in DL,RL,POF,PIFand SG by 29%,40%,24%,32%and 26%,respectively(P ＜0.05).Conversely,the lowest level in RL did not differ from that in the other land use types with the exception of KS.Soil extractable C was significantly affected by season and decreased significantly(P=0.001)in the following order:April＞July＞October(Table II).The combined effect of land use type and season did not influence soil extractable C,despite soil extractable C in KS being significantly higher than that in SG in July and October.

Fig.2 Soil extractable C and N in April,July,and October under different land use types in Liudaogou watershed in Shenmu City,Shaanxi Province,China.Values are means with standard errors of means shown by vertical bars (n=3).DL=dam land;RL=rainfed slope land;POF=poplar forest;PIF=pine forest;KS=Korshinsk peashrub;SG=Stipa grassland.

Soil extractable N was also significantly affected by land use type(Fig.2,Table II).Soil extractable N was significantly(P ＜0.05)higher in DL and KS than that under the other land use types.The effect of season on soil extractable N was significant (Table II).July had the largest soil extractable N,which was significantly(P ＜0.001)higher than that in April and October by 55%and 38%,respectively.However,the difference between April and October was not significant.The combined influence of land use type and season on soil extractable N was significant.Soil extractable N in SG was significantly(P ＜0.05)lower than that in DL and KS in all seasons,lower than that in POFin April and July,and lower than that in RL in July and October.

Soil microbial biomass C and N

TABLE IIResults of repeated-measures analysis of variance

Land use type significantly influenced soil MBC(Fig.3,Table II).Grassland had the highest MBC,being 12%,29%,67%,93%,and 146%higher than that in PIF,KS,DL,RL,and POF,respectively.The difference in MBC between SG and PIFwas not significant.Soil MBC in DL was significantly higher than that in POF,but significantly lower than PIF,while MBC in RL and POFwere significantly lower than that in PIFand KS.Season had a significant influence on MBC(Table II).The lowest MBC was found in April,being 59%and 47%lower than that in July and October,respectively(P ＜0.001).Soil MBC in July was 30%higher than that in October.The combined effect of land use type and season on MBC was significant.In the three sampling seasons,MBC under all land use types were significantly different from MBC in SG,except that in PIFand KS in October.

Fig.3 Soil microbial biomass C and N in April,July,and October under different land use types in Liudaogou watershed in Shenmu City,Shaanxi Province,China.Values are means with standard errors of means shown by vertical bars(n=3).DL=dam land;RL=rainfed slope land;POF=poplar forest;PIF=pine forest;KS=Korshinsk peashrub;SG=Stipa grassland.

Land use type also strongly affected soil MBN(Fig.3,Table II).The highest MBN was observed in KS,which was 157%,135%,108%,93%,and 23%higher than that in PIF,POF,DL,RL,and SG,respectively.Soil MBN in KS and SG were significantly higher than that under the other land use types,but not different from each other.Soil MBN varied in the three seasons(Table II).July had the highest MBN,which was 135%and 112%higher than that in April and October,respectively(P ＜0.001).Soil MBN was not significantly different between April and October,although the latter was 11%higher than the former.The combined influence of land use type and season on MBN was primary.Soil MBN in SG was different from that under all the other five land use types in October and was also different from that in DL,RL,PIF,and KS in July and POFin April.

Soil MRR and NNM

Land use type significantly affected soil MRR(Fig.4,Table II),which could be classified into two groups under different land use types,one group including DL,RL,and POF,and the other group including PIF,SG,and KS.Within each group,MRR showed no significant difference among different land use types.Between groups,MRR in PIF,KS,and SG were higher than that in DL,RL,and POF.Soil MRR was influenced by season(Table II).The lowest MRR was found in April,which was significantly (P ＜0.001) different from that in July and October.However,the difference between July and October was not vital.The combined influence of land use type and season on MRR was significant.Soil MRR were lower in DL,RL,and POF than SG in July and October and was also lower in RL than SG in April,but was higher in KS than SG in April,July and October.

Fig.4 Soil microbial respiration rate(MRR)and net N mineralized(NNM)in April,July,and October under different land use types in Liudaogou watershed in Shenmu City,Shaanxi Province,China.Values are means with standard errors of means shown by vertical bars(n=3).DL=dam land;RL=rainfed slope land;POF=poplar forest;PIF=pine forest;KS=Korshinsk peashrub;SG=Stipa grassland.

Soil NNM was significantly influenced by land use type(Fig.4,Table II).The highest mean NNM in April,July,and October was found in KS and was 187%,87%,437%,and 99%larger than that in RL,POF,PIF,and SG,respectively.The PIFhad the lowest mean NNM,which was obviously lower than that in DL,POF,and SG.The mean PIFin DL was larger than that in RL and SG.Soil NNM was significantly affected by sampling season(Table II).The highest NNM was found in July,which was 38%and 115%larger than that in April and October,respectively.The difference between April and October was also statistically notable(P ＜0.001).The combined influence of land use type and season on NNM was significant.Compared with that in SG,NNM was significantly higher in KS in all three seasons,was higher in DL but lower in PIFin April and July,was obviously lower in RL in July,and was notably higher in POFin October.

Relationships between soil microbial properties and environmental indicators

Pearson correlation analysis results showed that TOC was significantly positively correlated with soil MBC,MBN,MRR,moisture,clay,and microaggregates but negatively correlated with(Fig.5).Microbial biomass C was positively correlated with soil moisture,clay,physical clay,microaggregates,TOC,and TN.Soil MRR was positively correlated with TOC,MBC,and MBN and negatively correlated with.Soil moisture was significantly positively correlated with soil pH,clay,physical clay,microaggregates,TOC,TN,and MBC and negatively correlated withand NNM.

Fig.5 Correlation coefficients(r)between soil properties.NNM=net N mineralized;Ext-C=extractable C;MBN and MBC=microbial biomass N and C,respectively;MRR=microbial respiration rate;TOC=total organic C;TN=total N;MA=microaggregates;MT=moisture;PyC=physical clay.

The RDA successfully modeled the environmental factors and soil microbial properties(Fig.6).Total variation was 18.40,explanatory variables accounted for 98.5%,and adjusted explained variation was 87.2%.All axes were significant atP=0.006 4,and axis 1 was significant atP=0.030 9.Axes 1 and 2 explained 60.8%and 28.7%of the total variance,respectively.Soil MBC,MRR,and MBN were negatively correlated with,but positively correlated with soil physical clay,moisture,clay,microaggregates,pH,TN,TOC,,and extractable C.Specifically,the selection procedure in the RDA model indicated that soil TOC,,TN,moisture,and extractable C were important predictors of soil microbial biomass and activity(P ＜0.05)and explained 88.90%of the variance.The RDA model also indicated that the SG,KS,PIF,and DL systems facilitated C sequestration and favored microbial properties.

Fig.6 Redundancy analysis plot showing the relationships between environmental factors and soil microbial properties.Red lines represent environmental factors,blue lines represent microbial properties,and green triangles represent different land use types.NNM=net N mineralized;Ext-C=extractable C;MBN and MBC=microbial biomass N and C,respectively;MRR=microbial respiration rate;TOC=total organic C;TN=total N;MA=microaggregates;MT=moisture;PyC=physical clay;DL=dam land;RL=rainfed slope land;POF=poplar forest;PIF=pine forest;KS=Korshinsk peashrub;SG=Stipa grassland.

DISCUSSION

Impact of land use type on soil TOC and TN

The results of the present research clearly support our first hypothesis that in the water-wind erosion crisscross region,land use type could significantly affect soil TOC and TN,and that dam land could sequester more TOC and TN.Soil TOC,TN,and extractable N in the exotic N-fixing Korshinsk peashrub were just below those in the dam land.Soil TOC and TN were slightly lower in the natural grassland than those in the dam land and Korshinsk peashrub land and higher than those in the other land use types.The rainfed slope land exhibited 32%lower TOC and TN than the dam land.The deciduous broadleaf Simon poplar forest showed the lowest TOC and TN.

Dam land,which was formed by the intercepted sediment in the upper reaches of the check dam,deposits C,nutrients,and water from upstream by runoff(Liuet al.,2017).Compared with other land use types,the agricultural conditions in dam land,for instance,the flat surface for easy agricultural management and the higher soil moisture for crop uptake,are more suitable for intensive agriculture(Xiaoet al.,2014).Soil moisture is a limiting factor for flora in arid and semiarid areas,and the dam land had higher soil moisture than the other scenarios during the whole experimental period in this study(Fig.S2,see Supplementary Material for Fig.S2).Moreover,with brighter sunlight and higher accumulative temperatures in the growing season,the agricultural management in dam land could in turn increase TOC and TN accumulation(Lüet al.,2012a;Liuet al.,2017).In contrast to dam land,residents in rainfed slope land were more easily eroded by runoffor blown away by the wind and had lower TOC and TN(Visser and Sterk,2007).The rate of soil erosion in areas affected by disturbance can be up to 5 times higher than natural erosion rates(Berheet al.,2018).Globally,erosion induced by agricultural practices lost 78±22 Pg C from 6000 BC to AD 2015(Wanget al.,2017).Natural grassland has a high density of fine roots and plays an important role in the feedback of TOC and TN to land use change(Weiet al.,2009).In the deep profile(0-300 cm),grassland had higher TOC and TN deposits when the mean annual precipitation was＜510 mm,while shrubland and forest facilitated TOC and TN when the annual precipitation was＞510 mm(Tuoet al.,2018).Pollen records show that grassland has been the historical landscape in this area since the last 130 000 years(Jianget al.,2013).

The broadleaf poplar has a high water consumption,similar to a“water pump”that accelerates water depletion from deep soil and forms a dry soil layer,which could cause popular trees to die in this watershed after being planted for nearly 30 years(Wanget al.,2010).Poplar is typically planted in sandy soil,which consists of nearly 90%sand and does not facilitate C and N absorbance and adherence to soil particles to preserve the levels in soil (Sixet al.,2002b).Although pine forest soil had relatively high clay and silt contents(8.30%and 31.18%)and the highest soil moisture in the 0-10 cm layer in summer,pine trees in this watershed showed a typical degradation trend as“small aged trees”,and the average 3.5 m height constituted only 20%of the southern Loess Plateau(Jiaet al.,2017a).One disadvantage of the exotic Korshinsk peashrub is its high water consumption,which can form a dry soil layer after being planted for 3 years and in severity with plant age and could reach 22.4 m(Wanget al.,2009,2010).However,in the transition from forest to grassland,shrubs are the best candidate for vegetation restoration(Denget al.,2016).

Impact of land use type on soil microbial biomass and activity

Consistent with the second hypothesis,land use affects microbes by inducing changes in C quantity and quality.Generally,it is believed that the stabilization of soil TOC is promoted by great microbial biomass and activity(Liang and Balser,2011).In this research,the percentage of C in microbial biomass and microbial respiration were the greatest in grassland,indicating that the microbial activities or turnover should be greater in grassland.This may be ascribed to the greater new C deposit,either by litter or root exudates.In addition,the secondary compounds produced by microbes may favor the formation of microaggregates,which can increase the stabilization of TOC(Tisdall,1994).This research clearly showed that grassland had higher microbial biomass and microbial activity,which indicates that grassland not only increases the quantity of TOC and TN but also increases the quality of TOC compared with other land use types.

Land use also affected microbial properties through its induced alterations in N resource availability.Generally,it is believed that plant growth is affected by N shortage,and N addition could facilitate plant growth,enhance C deposition,and favor TOC and TN maintenance(Huet al.,1998).The NNM is higher in dam land and Korshinsk peashrub land,as it has a high N input either from human activities such as fertilization (dam land) or plant N fixation (Korshinsk peashrub land)(Yuet al.,2020).The reasons may be that the chemical N fertilizer applied in dam land and the N fixed by Korshinsk peashrub can increase the N content.Without considering the water deficiency effect or the region having relatively higher precipitation,Korshinsk peashrub should be selected as the pioneer plant because it can increase soil N pool and greatly contribute to TOC storage.In addition,the difference in MRR and NNM between poplar forest and pine forest also reflected the influence of land use type on soil microbial decomposition,as the latter has a longer residence time and contains lower N and more lignin(Zhuet al.,2014).

Soil microbial biomass and activity were influenced by anthropogenic disturbance,such as land use change,and were higher in the less disturbed scenario.Being the most disturbed of the six land use types,cropland(dam land and rainfed slope land)had lower microbial biomass and activity than the less disturbed forest,shrubland,and grassland in this semiarid watershed.Interestingly,another experiment conducted in the humid and warm region of the mountains of western North Carolina,USA,also found that traditional high-input intensity agricultural fields had lower microbial biomass and respiration than natural grassland(Wanget al.,2011).In this research,tillage in dam land and rainfed slope land may disrupt soil macroaggregates and increase microaggregate exposure,which may decrease the protection and loose organic matter for microbial decomposition(Sixet al.,2002a).In addition,soil aggregates are easily destroyed by the changing topographic environment,such as rainfed slope land in this study.Our results indicated that dam land displayed considerable potential for soil TOC sequestrationviahigher TOC content but lower soil microbial activity.This result could be highlighted in the lateral C redistribution and human engineering project when assessing regional-and global-scale C cycles.

CONCLUSIONS

In accordance with our hypothesis,this study confirmed that in the water-wind erosion crisscross region of the Loess Plateau,land use significantly affected soil TOC,TN,and microbial properties.Dam land had the highest TOC and TN and sustainable production management practices,such as no-tillage plus organic farming,and needed to be applied in field management schemes.In addition,management to preserve the safety of dam land from erosion,salinization,and unfavorable hydrological conditions is urgently needed,as dam land may act as a source of global warming gases such as atmospheric CO2and nitrogen oxide.Rainfed slope land should be phased out because it was not an eco-friendly land use type and decreased soil TOC and TN.With fewer disturbances,PIF,KS,and SG had relatively high TOC,TN,and microbial biomass and activity.However,forest in this watershed was degraded by the high soil moisture consumption and formation of a dry soil layer.Grassland exhibited greater TOC,TN,and microbial properties than other land use types in the northern Loess Plateau.Specific grassland ecosystem responses to fire,grazing,drought,and global warming need to be investigated in future studies.Future land use planning in this area,such as the restoration of the opencast coal mine waste dump land,should involve grassland or plant species using less water,less thinning,and less labor-intensive management practices.This will not only favor sustainable soil and water conservation,but will also not affect critical water resources in arid and semiarid areas,thereby supporting sustainable development of the fragile water-wind erosion crisscross region.

ACKNOWLEDGEMENTS

The research was jointly supported by the National Natural Science Foundation of China(Nos.41201259 and 41671269),the Open Fund of the State Key Laboratory of Soil Erosion and Dryland Farming on Loess Plateau,China(No.10501-1207),the National Science Foundation of Shaanxi Province,China(No.2013JQ5001),and the CAS“Light of West China”Program.We express gratitude for the logistics provided by the Shaanxi Shenmu National Observation and Research Station of Erosion and Environment in China.

SUPPLEMENTARY MATERIAL

Supplementary material for this article can be found in the online version.