ZHOU Lei ,XU Sheng-tao ,Carlos M.MONREAL ,Neil B.MCLAUGHLIN ,ZHAO Bao-pingLIU Jing-huiHAO Guo-cheng

1 Cereal Industry Collaborative Innovation Center,Inner Mongolia Agricultural University,Hohhot 010019,P.R.China

2 College of Agronomy,Inner Mongolia University for The Nationalities,Tongliao 028000,P.R.China

3 Ottawa Research and Development Centre,Agriculture and Agri-Food Canada,Ottawa,ON K1A 0C6,Canada

4 Agricultural Environment and Resources Institute,Yunnan Academy of Agricultural Sciences,Kunming 650205,P.R.China

5 Inner Mongolia Trirock Co.,Ltd.,Naiman Banner,Tongliao 028300,P.R.China

Abstract A bentonite-humic acid (B-HA) mixture added to degraded soils may improve soil physical and hydraulic properties,due to effects such as improved soil structure and increased water and nutrient retention,but its effect on soil physicochemical and biological properties,and grain quality is largely unknown. The effect of B-HA,added at 30 Mg ha-1,was studied at 1,3,5 and 7 years after its addition to a degraded sandy soil in a semi-arid region of China. The addition of B-HA significantly increased water-filled pore space and soil organic carbon,especially at 3 to 5 years after its soil addition to the soil. Amending the sandy soil with B-HA also increased the content of microbial biomass (MB)-carbon,-nitrogen and -phosphorus,and the activities of urease,invertase,catalase and alkaline phosphatase. The significant effect of maize (Zea mays L.) growth stage on soil MB and enzyme activities accounted for 58 and 84% of their total variation,respectively. In comparison,B-HA accounted for 8% of the total variability for each of the same two variables. B-HA significantly enhanced soil properties and the uptake of N and P by maize in semi-arid areas. The use of B-HA product would be an effective management strategy to reclaim degraded sandy soils and foster sustainable agriculture production in northeast China and regions of the world with similar soils and climate.

Keywords:bentonite-humic acid,soil organic carbon,microbial biomass,enzyme activity,grain quality,sandy soil

1.Introduction

Sandy soils are characterized by having low water holding capacity,water retention and fertility due to low clay and soil organic matter content,that restrict crop growth resulting in low productivity (Weber et al.2007). The application of amendments is a widely accepted strategy to improve the quality and productivity of soils (Hu et al.2015;Mohawesh and Durner 2019). Addition of soil humic acid (HA) improves soil quality by increasing macroaggregate formation,C pool,soil fertility,soil cation exchange capacity,mineral nutrition and microbial activity,all of which favor crop growth(Clemente and Bernal 2006;Puglisi et al.2009;Szczerski et al.2013). The combined effects of bentonite and humic acid (B-HA) addition to improve soil biotic properties are unknown. Bentonite is a clay mineral with three-layered(2:1) crystalline structure (Puma et al.2015),capable of considerable swelling and high water absorbing capacity(Huang et al.2013).

Soil microorganisms play an essential role in nutrient cycling,energy flow and plant nutrient acquisition,which are important processes for the long-term sustainability of agricultural systems (Liu et al.2010). Soil microbial biomass (MB) and its activity indicate soil health (Schloter et al.2003). The MB,a small labile portion of soil organic matter (SOM),influences nutrient cycling,conservation of soil carbon,and the supply of nitrogen and phosphorus to crops (Jia et al.2005;Niu et al.2015). Microbial biomass carbon (MBC),nitrogen (MBN) and phosphorus (MBP) pools are sensitive indicators of effects of agricultural management(Roberts et al.2015). Soil enzyme activities also serve as indicators of soil quality changes (Bowles et al.2014). Soil enzymes are derived from soil bacteria,fungi,plant roots and residues (Li et al.2010). Soil enzyme activities and MB are used as sensitive indicators to provide immediate and accurate information about small disturbances in soils(Hueso et al.2012). Understanding the potential effects of B-HA on soil MB and enzyme activities in sandy soil is important for formulating strategies for soil restoration and sustainable agricultural development in semi-arid regions.

We hypothesized that addition of B-HA to a degraded semi-arid sandy soil improves water-filled pore space(WFPS),soil organic carbon (SOC) content,MB and soil enzyme activities thereby increasing grain N and P content in maize (Zea mays L.). The objectives of this study were to determine the residual effects of one,three,five and seven years after a single B-HA soil application on:1) The dynamics of WFPS,SOC,MBC,MBN and MBP,invertase,catalase,urease and alkaline phosphatase activities at different growth stages of maize;2) potential relationships between the latter soil biological properties and the content of N and P in maize grain.

2.Materials and methods

2.1.Experimental site and climate

The experiment was performed on a sandy soil located in Naiman Banner (120°55′E,42°41′N),Inner Mongolia,China,which is part of the Horqin sandy land. Soil properties were as follows:pH (7.9,0.01 mol L-1CaCl2);SOM 4.8 g kg-1;soil texture (w/w) 96.3% sand,1.9% silt and 1.8% clay.This experimental area was originally a sand dune and was converted to a flat field with earth moving machinery.In 2009,soil was prepared for cultivation and in 2010,the B-HA plots were established (Zhou et al.2019).

The Horqin region has a typical continental semi-arid monsoon climate. The mean annual potential evaporation is＞1 935 mm,and the mean annual precipitation is 364.6 mm.The mean annual temperature is 6.4°C,and the mean annual wind velocity varies between 3.4 and 4.5 m s-1,depending on the location within the region.

2.2.Experimental design

The B-HA amendment applied to the sandy soil consisted of a mixture of calcium-bentonite,humic acid,Na2CO3and cellulose in a ratio of 91:6:2:1. The chemical properties and the process used to prepare the B-HA amendment were reported by Zhou et al.(2019).

Details of the experimental design and a schematic diagram of the research plot layout were provided by Zhou et al.(2019). Briefly,the experimental field was 3.36 ha(420 m×80 m). A control (CK,no B-HA addition) and the B-HA treatment sub-plots were established in 2010 in triplicate as a completely randomized design. The B-HA was applied to a new sub-plot each year starting in 2010,and the treated sub-plot was maintained with a maize crop grown each year before,the year of,and each year after the B-HA treatment until 2016 without additional B-HA,i.e.,each B-HA treated sub-plot received only one B-HA treatment from 2010 to 2016. The B-HA was applied at 30 Mg ha-1to the soil surface in May of each year prior to sowing and incorporated by tillage.

2.3.Maize growth and phenology

The maize variety was Zhengdan 958 and in the year of soil sampling (2016) it was sown on 1st of May and harvested on 28th of September. The plant density was 62 500 plants ha-1. Diammonium phosphate as starter fertilizer was applied to provide 49 kg N ha-1and 124 kg P2O5ha-1.Urea was applied at 138 kg N ha-1at 50 days after sowing,and second amount of urea (69 kg N ha-1) was added at 85 days after sowing. Flood irrigation was applied uniformly to all plots according to the local recommended irrigation practice. Different crop phenological stages were selected to characterize soil organic carbon content,and temporal changes in WFPS,MB and enzyme activity. The crop growth stages were defined at 30,50,70,90,110 and 130 days after sowing,which corresponded to the V3 (third leaf emerged with ligule visible),V6 (sixth leaf emerged with ligule visible),V12 (twelfth leaf emerged with ligule visible),R1 (silk stage),R3 (milk stage),R5 (dent stage),respectively.

2.4.Soil sampling

Soil samples were taken to determine WFPS,SOC content,soil enzyme activities and MB properties both in April of 2010 prior to sowing to examine soil heterogeneity prior to initiation of the experiment,and at each of the phenological growth stages in 2016 to examine B-HA effects. Samples were taken in 2016 from the control,and the sub-plots receiving B-HA treatments in 2010,2012,2014 and 2016 representing respectively 7,5,3 and 1 years after the B-HA treatment. Any sampling between 2010 and 2016 is not reported in this manuscript.

In each treatment replicate,five separate soil samples were taken with a 5 cm diameter soil auger from the 0-10 cm and 10-20 cm depth randomly along the rows of maize.Fresh moist soil samples from each depth were pooled,placed in a sterilized sealed plastic bag and brought to the lab. Each composite soil sample was thoroughly mixed and divided into three portions. One portion was used to determine soil moisture by oven-during at 105°C for 24 h.The second portion was used to measure SOC content,this sample portion was air-dried and sieved through 2 mm mesh to remove small stones and undecayed plant stems,leaves and roots,then sieved through 0.15 mm mesh to measure SOC content (Adnan et al.2020). A separate soil sample portion used to determine MB and enzyme activities was stored at 4°C for one week before analysis. Soil bulk density was determined following the method described by Zhao et al.(2006).

Soil organic carbonSoil organic matter was determined by the hot potassium dichromate digestion method,and the soil organic carbon content was estimated by applying the Van Bemmelen factor of 1.72 (Piper 1950).

Soil microbial biomassSoil MBC,MBN and MBP were determined using the chloroform fumigation-extraction method (Brookes et al.1982,1985;Cui et al.2018).Moist soil samples from each treatment and replicate were incubated in a sealed container in a dark cabinet for 10 days at room temperature. Both fumigated (with CHCL3for 24 h at 25°C) and unfumigated soil samples were extracted with 0.5 mol L-1K2SO4for MBC and MBN,and with 0.5 mol L-1NaHCO3(pH=8.5) for MBP. The content of MBC was determined by titrating the extract with FeSO4after hot digestion with the mixture of K2Cr2O7and H2SO4;MBN was determined calorimetrically using a spectrophotometer(570 nm) to calculate the NH3-N content with ninhydrin reagent;MBP was calculated using a spectrophotometer at 700 nm wavelength with mixed molybdenum antimony color reagent. The content of MB(C,N,P) was estimated as follows:where Fum and NFum were the values obtained from the fumigated and non-fumigated soil;k (the conversion factor)used for MBC determination was 0.38 (Vance et al.1987),for MBN was 0.45 (Joergensen and Brookes 1990) and for MBN was 0.4 (Brookes et al.1982).

Soil enzyme activitiesUrease activity (EC 3.5.1.5) was measured following the method of Guan et al.(1991). Five grams moist soil sample were mixed with 5 mL of 10%(w/v) urea solution,shaken for 2 min and then placed in an incubator at 37°C for 3 h. The concentration of NH3-N released was determined by measuring absorbance in a spectrophotometer at 578 nm wavelength. The urease activity was expressed as μg NH3-N g-1soil h-1.

Alkaline phosphatase activity (EC 3.1.3.1) was determined by using the p-nitrophenyl phosphate (pNPP) method. Five grams moist soil sample were incubated with 20 mL modified universal buffer (MUB) (Kumari et al.2017). A total of 5 mL of 25 μmol L-1p-Nitrophenyl phosphate (pNPP) solution was immediately added to the soil and MUB mixture,then swirled and incubated at 37°C for 1 h with 5 mL of 0.5 mol L-1CaCl2and 20 mL 0.5 mol L-1NaOH. The incubated soil suspension was then filtered through a phosphate-free filter paper. The amount of p-nitrophenol released from pNPP hydrolysis was calculated from a standard curve using absorbance values obtained from a spectrophotometer at 400 nm. The activity of alkaline phosphatase was expressed as μg p-nitrophenol g-1soil h-1(Tabatabai and Bremner 1969).

For measurement of invertase activity (EC 3.2.1.26),5 g moist soil sample was mixed with 15 mL of 8% (wt/v) sucrose and incubated at 37°C for 3 h (Guan et al.1991). The mixture was heated and diluted with distilled H2O. Invertase activity was determined after measuring glucose production by absorbance read at 508 nm in a spectrophotometer,and expressed as mg glucose released g-1soil h-1(Guan et al.1991).

Catalase activity (EC 1.11.1.6) was measured by mixing 2 g moist soil sample with 5 mL of 0.3% H2O2solution and 40 mL distilled water for 20 min. After incubation,5 mL of 3 mol L-1sulfuric acid was added to stabilize the unreacted H2O2and then filtered immediately (Liu et al.2008). Then 0.1 mol L-1KMnO4was used to titrate extra H2O2with 25 mL of the filtrate until the mixture color changed to pink.The catalase activity was expressed as KMnO4mL g-1soil 30 min-1(Liu et al.2008) .

Water-filled pore space (WFPS)The WFPS was calculated using the equation described by Linn and Doran (1984):

where θvis volumetric water content,TP is total soil porosity;θmis gravimetric water content (g g-1),ρpis soil particle density which assumed to be 2.65 g cm-3,ρbis soil bulk density (g cm-3).

Total nitrogen and phosphorus in maize grainAfter maize harvest,the grain was oven dried at 105°C until constant weight was achieved. Grain was ground to pass a 1 mm screen and a 0.2 g sample was used to determine N content using the Kjeldahl digestion method (Nelson and Sommers 1973). The P content was determined after digestion in H2SO4-H2O2and then determined by vanadium-molybdenum yellow colorimetry (Lu 2000).

2.5.Statistical analysis

Multivariate analysis of variance (MANOVA) was performed using R software (R Core Development Team 2018) to test for significant effects of B-HA application,years after B-HA addition,soil depth,and maize growth stage on WFPS,SOC,MBC,MBN,MBP and soil enzyme activities. Correlation analysis among the measured variables was performed by SAS ver.9.1 software (SAS Institute Inc.,Cary,NC,USA).Linear regression models were used to describe the quantitative relations between MBN,enzyme activity and grain N content,MBP,enzyme activity and maize grain N and P content. Principal component analysis(PCA) was conducted using R to show the linkage among variables and the year-effect after B-HA addition.For PCA and in order to preserve variance and eliminate effect of differences in the units,data were scaled using the formula (X-x)/SD (where x is the mean of the sample,and SD is the standard deviation).total variation in WFPS (Table 2;Fig.1). The year (Y)

Fig.1 Water-filled pore space (WFPS) in 0-10 cm depth at different days after sowing for control and 1,3,5 and 7 years after bentonite-humic acid (B-HA) addition. CK,control;B-HA-1,B-HA-3,B-HA-5 and B-HA-7 indicate 1,3,5 and 7 years after B-HA addition,respectively. Error bars indicate SD (n=3).

3.Results

3.1.Initial soil properties

One-way ANOVA on soil properties sampled in April 2010 prior to establishment of the experimental treatments showed that the treatment effect was not significant for all of the measured soil parameters at both depths indicating initial homogeneity of the experimental site in 2010 (Table 1). Here,treatment refers to the future B-HA treatment planned for each plot from 2010 to 2016.

3.2.Water-filled pore space

The addition of B-HA addition (T) increased WFPS significantly (P≤0.001) and it accounted for 49% of the after B-HA application significantly affected the WFPS but accounted for only 5% of the total data variability. Soil depth(D) and growth stage (GS) significantly (P≤0.001) increased WFPS and both variables were responsible for 11-12% of the variation of the total sum of squares. The double and triple interactions between variables showed significant effects on WFPS (Table 2).

3.3.Soil organic carbon (SOC)

The addition of B-HA significantly (P≤0.001) increased SOC content,especially between three and five years after its addition,but its content,remained unchanged from five years to seven years after B-HA addition (Table 2;Fig.2). The effects of soil depth and addition of B-HA were responsible for 30 and 23%,respectively,of the total variation in SOC content. Maize growth stage had a significant (P≤0.001)effect on SOC and accounted for 15% of the total annual variation in SOC content. The interaction between year and B-HA treatment was significant (P≤0.001) and contributed equally to the total variation (～8%) in SOC content.

Fig.2 Soil organic carbon (SOC) vertical changes in 2016 in 0-20 cm soil depth after bentonite-humic acid (B-HA) addition. A,SOC in 0-10 cm depth at different days after sowing for control and 1,3,5 and 7 years after B-HA addition. B,SOC in 0-20 cm depth plotted against number of years since B-HA addition. CK,control;B-HA-1,B-HA-3,B-HA-5 and B-HA-7 indicate 1,3,5 and 7 years after B-HA addition,respectively. Error bars indicate SD (n=3).

3.4.Soil microbial biomass

Relative to CK,the soil MB was increased significantly(P≤0.001) by both B-HA,treatment and maize growth stage(Table 3;Fig.3). The addition of B-HA increased MBC,MBN and MBP by 38,84 and 43% in the 0-10 cm depth;and by 40,93 and 45%,respectively,in the 10-20 cm depth (data not shown).

The MB continuously increased in the first 5 years after B-HA application,and its content remained unchanged from 5 to 7 years after B-HA addition. Growth stage explained,on average,58% of the total variation in soil MBC,MBN and MBP. The addition of B-HA also increased MB continuously during maize growth,from the 30th to 90th day of growth and it decreased thereafter. The maximum content in MBC,MBN and MBP occurred 90 days after sowing (Fig.3-A-C).Soil depth contributed～1% to the total variation in MB.The interactions of T×Y,D×GS and T×GS all significantly influenced MB (P≤0.001) (Table 3).

Fig.3 Temporal change in soil microbial biomass in 0-10 cm depth at different days after sowing for control and 1,3,5 and 7 years after bentonite-humic acid (B-HA) addition. MBC,microbial biomass C;MBN,microbial biomass N;MBP,microbial biomass P.CK,control;B-HA-1,B-HA-3,B-HA-5 and B-HA-7 indicate 1,3,5 and 7 years after B-HA addition,respectively. Error bars indicate SD (n=3).

Table 2 Summary of variance for WFPS and SOC (significant effects only,P≤0.05)1)

The addition of B-HA amendment significantly (P≤0.001)increased the activities of urease,alkaline phosphatase,invertase and catalase,and the amendment effect was responsible for an average of 8% of the total variability in enzyme activities (Table 4). Growth stage significantly(P≤0.001) influenced soil enzyme activity. The highest activity of urease occurred at 70 days after sowing (Fig.4-A)while alkaline phosphatase,invertase and catalase all showed the highest activity at 90 days after sowing(Fig.4-B-D). The effect of GS accounted,on average,for 84% of the total variation in soil enzyme activities. In comparison,year and soil depth only contributed ＜2% of the total data variability. The double and triple interactions of D×GS,T×Y,T×GS and T×Y×GS all significantly (P≤0.05)influenced the four enzyme activities.

Fig.4 Temporal change in soil four enzyme activities in 0-10 cm depth at different days after sowing for control and 1,3,5 and 7 years after bentonite-humic acid (B-HA) addition. Ure,urease;Int,invertase;Alkp,alkaline phosphatase;Cat,catalase.CK,control;B-HA-1,B-HA-3,B-HA-5 and B-HA-7 indicate 1,3,5 and 7 years after B-HA addition,respectively. Error bars indicate SD (n=3).

Table 3 Summary of variance for soil microbial biomass C,N and P (significant effects only,P≤0.05)1)

Table 4 Summary of variance for soil enzyme activities (significant effects only,P≤0.05) 1)

3.5.The content of N and P in maize grain

Relative to the CK,the addition of B-HA increased N content in maize grain by 16% and P by 43% after 5 years (Fig.5).The maximum concentration of grain-N was 10 g kg-1attained after 7 years of maize cropping since B-HA addition,and for grain-P was 2.2 g kg-1attained after 5 years.

Fig.5 Grain nutrients content after bentonite-humic acid (B-HA) addition. Grain N (A) and P (B) concentration in control (CK) and B-HA treated plots. CK,control;B-HA-1,B-HA-3,B-HA-5 and B-HA-7 indicate 1,3,5 and 7 years after B-HA addition,respectively.The columns with the same letters are not significantly different at P=0.05. Error bars indicate SD (n=3).

3.6.Association between biotic and abiotic soil properties

The MB and the activities of the four soil enzymes were significantly (P≤0.001) correlated with WFPS and SOC.Correlation coefficients r between MB and WFPS,and between MB and SOC ranged from 0.5 to 0.8 (Table 5).The values for r between MB and the activities of the four soil enzymes were higher and ranged from 0.75 to 0.89.

Results from our PCA explained 98% of the total data variability (Fig.6). The Principal Component 1 explained 96.8% of the experimental data variability and indicates a clear separation between the CK and B-HA treatments. All the measured soil parameters positively were positively correlated with B-HA addition 3,5 and 7 years after B-HA addition as shown in the first and fourth quadrants of Fig.6. The score plots associated with year 1 were clearly separated from those associated with years 3,5 and 7 after B-HA addition,statistically confirming the interannual changes caused by the soil amendment on the five biotic(MB and four enzyme activities) and one abiotic (WFPS)variables. The content of SOC was also associated with B-HA plots with 5 years after B-HA addition as shown in the fourth quadrant of Fig.6.

Fig.6 Principal Component Analysis (PCA) on selected variables of water-filled pore space (WFPS),soil organic carbon (SOC)and microbiological parameters in the control (CK) and bentonite-humic acid (B-HA) treatments. MBC,microbial biomass C;MBN,microbial biomass N;MBP,microbial biomass P;Ure,urease;Int,invertase;Alkp,alkaline phosphatase;Cat,catalase.CK,control;B-HA-1,B-HA-3,B-HA-5 and B-HA-7 indicate 1,3,5 and 7 years after B-HA addition,respectively.

3.7.Association between soil biotic properties and grain N and P content

The grain N and P contents were significantly correlated(P≤0.01) with soil enzyme activities (Table 6). Grain N (Y)content was closely associated with MBN (X) and described by the function Y=0.16X+5.7 (R2=0.98,P≤0.01). Grain-P(Y) content was closely associated with invertase activity(X) and described by the function Y=1.09X-2.18 (R2=0.97,P≤0.01) (Table 6).

Table 5 Pearson correlation coefficient (r) among WFPS,SOC and soil biotic properties1)

Table 6 Linear regression equations describing the amount of total N and P in the grain according to various soil biotic properties in a sandy degraded soil treated and untreated with B-HA

4.Discussion

4.1.Water-filled pore space and soil organic carbon content

Water is essential to crop growth and soil productivity,especially for sandy soils in arid and semi-arid areas (Yao et al.2013). With loose soil structure and low organic matter content,sandy soils retain little water from rainfall and irrigation (Weber et al.2007).

Our results showed that B-HA addition was an important factor to increase WFPS. The latter may be due to improvement of soil structure (i.e.,decreased soil bulk density and increased proportion of macro-aggregates) and water content in the presence of B-HA as reported earlier by Zhou et al.(2019). Too much or too little water in the pore space strongly affect the O2distribution and the balance of soil air and water has an effect on root growth and microbial activities and thus nutrient cycling (Skopp et al.1990). An earlier study showed that high soil water content and/or lower porosity helped to increase WFPS from 6 to 28% in the 0-15 cm depth of a silt loam soil (Mielke et al.1986). In addition,aerobic microbial activity decreases for a WFPS level ＞60% (Linn and Doran 1984). Our data showed that addition of B-HA increased WFPS from 12 to 20%,thereby contributing to greater MB size and enzyme activity in the sandy soil.

Soil organic carbon plays an important role in energy transfer,microbial metabolism and biological diversity to maintain a healthy agroecosystems health (Li et al.2016).Although the content of SOC in sandy soils is low,its importance should not be ignored as it provides energy and substrates for heterotrophic microorganism growth thereby helping to sustain soil fertility (Li et al.2016). Relative to CK,the addition of B-HA increased SOC content by 3-6% from 488 to 500 g C m-2in the top 20 cm of soil (Fig.2-B). Most of the increase in SOC content appears to be associated with the addition of HA in the B-HA mixture. In our experiment,the addition of 1 800 kg ha-1of HA (containing an average of～387 kg C) accounted for 3.7% of the total SOC content in the 0-25 cm depth. Therefore,the amount of C added in the B-HA accounted for 70% of the total increase in SOC content measured in the top 20 cm of soil. Humic acid is a recalcitrant form of biological C that forms stable organomineral complexes (Schnitzer and Monreal 2011). The presence of clay and the cation exchange capacity of B-HA help protect SOC from microbial decomposition,a physicochemical process known as ‘‘clay protection’’ (Mahieu et al.1999). The contribution of root exudates during maize growth,however,cannot be discarded as another important factor contributing to the additional 30% increase in soil C storage under the B-HA treatment.

4.2.Soil microbial biomass and enzyme activities

On average,the soil MBC accounts for 1 to 5% of the total SOC (Zhang et al.2011). Microbial characteristics of soil are frequently used as early indicators of changes in soil quality in agricultural systems. Soil MB size and activity are critical for plant growth and development,as they greatly affect the storage and transformation of soil nutrients (Glover et al.2000;Hsiao et al.2018).

The size of MBC,MBN and MBP showed similar temporal variation during maize growth,reaching a maximum pool size 90 days after sowing. Seasonal changes of soil temperature,rainfall,plant growth and root development have an important influence on microbial growth (Hu et al.2016;Xu et al.2016). In our study,root development and concomitant exudation of labile carbon likely helped increase MB pools during the growth of maize. We hypothesize that the improved WFPS and increased organic carbon supply from improved crop growth and exudation under B-HA directly supported soil MB growth. The maximum size of MB was attained during a period of fast crop growth (silk stage,at 90 days after sowing). Rapid crop growth stages are characterized by high supply and availability of labile carbon in the soil solution of crop rhizosphere to support microbial metabolism (Monreal et al.2018).

The activities of all soil enzymes increased significantly after B-HA addition (Table 1). Soil enzymes regulate soil biochemical processes (Bowles et al.2014) and they are important biotic components involved in soil nutrient cycling(Zhang et al.2011). Urease catalyzes urea into NH3,H2O and CO2(Li et al.2010). Urease exists both as intra and extra-cellular enzyme in soil,and it can continue to function when it is absorbed in the clay-organic matter surfaces(Fisher et al.2017). In general,urease activity is controlled by soil N content and microbial biomass (Allison et al.2007).In our experiment and relative to the other soil enzymes,the highest activity of urease occurred 70 days after sowing(V12) and unlike MB,not at 90 days after sowing and urea addition (R1). Lower urease activity at day 90 than at 70 days after initial urea addition appears to be associated with inhibition of soil urease after additional application of urea(69 kg N ha-1) at day 85. Soil urease is inhibited by its substrate at a concentration of 0.005 mol L-1as reported for a Gray Luvisol of Canada (Monreal et al.1986). In our sandy soil,and assuming a soil-water content 7% (w/w,or 225 m3water ha-1),the estimated concentration of urea was 0.011 mmol L-1in the soil solution of the 0 -20 cm soil depth after the addition of urea at day 85 which would be high enough to inhibit urease activity at the maize silk stage(90 days of growth).

Alkaline phosphatase is produced by plant roots,fungi and bacteria (Baker et al.2011),and its activity is used as an index to evaluate the ability of soils to mineralize organic P(Pascual et al.1998). In our study the maximum activity of alkaline phosphatase occurred at day 90 after sowing. The measured phosphatase activity increases under B-HA may be attributed to:1) the input of humic acid that may increase the availability of phosphate esters (Liu et al.2017);2) the diversity,abundance and activity of bacterial communities controlling P cycling (Tamilselvi et al.2015;Monreal and Zhang 2018);and 3) secretions of the enzyme by roots and microorganisms (Zhang et al.2011).

Catalase is an oxido-reductase enzyme that catalyzes H2O2into H2O and O2and oxidizes H donors thereby protecting soil organisms and plants from H2O2toxicity (Li et al.2017). Catalase activity is used as a proxy for soil biological oxygenation,that is closely related to the activity of aerobic microorganisms and soil fertility (Li et al.2010). The increase of catalase and the other enzyme activities may also be indirectly associated with the addition of the B-HA which improves soil physical properties (i.e.,aeration status)that ultimately result in enhanced MB pool size and crop growth (Zhou et al.2019). The increased catalase activity appears to reflect increased microbial production of protons associated primarily with the oxidation of labile-C pools.

Invertase catalyzes the hydrolysis of sucrose into glucose and fructose that constitute energy sources for microorganism metabolism. In soils,invertase originates from plant roots and microorganisms,such as Sacharomyces cerevisae (Paz-Ferreiro et al.2014;Li et al.2017). Addition of B-HA increased invertase activity indicating increased availability of SOC possibly enhanced through improved soil environment that favored microbial and root growth (Xiong et al.2008). Our results are consistent with the findings of Xu et al.(2016),who reported that application of HA into a sandy loam increased invertase activity.

Crop growth stage significantly increased the size of soil MB and enzyme activities in the sandy soil. Crop growth affects the supply of soluble organic substrates in the rhizosphere at different temperature and soil moisture(Hu et al.2016). The supply of labile soil C influences the dynamics,composition and distribution of soil bacterial communities and the size of soil MB pool (Tamilselvi et al.2015;Monreal et al.2018). In our experiment,the temperature and precipitation during the summer promoted plant growth,thereby creating a favorable environment for microbial activity in the rhizosphere and all four soil enzymes presented high activities during the active growth of maize.This correlation is consistent with the conjecture,that root growth and associated rhizodeposition stimulate microbial growth and the excretion and activity of the studied enzymes(Böhme and Böhme 2006). The increased supply of soluble carbon in canola rhizosphere was closely associated with crop growth stage and the relative abundance of soil bacterial communities (Monreal et al.2018). Root secretion is an important enzyme source contributing up to 90% of the extracellular soil enzymes pool (Gramss et al.1999;Li et al.2010). Amounts of root exudates are largely affected by the plant development stage (Li et al.2017). Tamilselvi et al.(2015) also observed high soil enzyme activities during the vegetative stages of maize growth (30 days after sowing) and flowering stage (60 days after sowing). These researchers indicated that crop growth stage influences soil enzyme activity through changing the temporal and spatial distribution of available substrates secreted by roots. Both temporal and spatial distribution of available substrates may have also played a role in increasing MB pool size. We also found that the addition of B-HA and years after its addition(T×Y),growth stage and soil depth (GS×D) and the addition of B-HA and growth stage (GS×T) interactions significantly affected MB pool size and the four enzyme activities(Table 3). These interactions appear to be a result of the improvements in soil structure that favored root growth,and increased MB and enzyme activities in the soil profile during the growing season and throughout the 7 years of the experiment (Zhou et al.2019).

Bentonite is an abiotic component that has a large adsorption capacity,thereby increasing the storage of soil water,adsorption of living cells and nutrients,thereby,facilitating the transformation and uptake of nutrients by crops (Filip 1973). The addition of B-HA influenced the size of soil MB and soil enzyme activities both directly and indirectly. Direct ways may include:1) B-HA addition directly supplied macro-elements (i.e.,N,P and K),mesoelements (i.e.,Ca,Mg and S) and micro-element (i.e.,Fe)to the soil (Zhou et al.2019). These elements,play a role in regulating soil microbial metabolism and likely contributed to the increase in soil MB. 2) The B-HA serves as a catalyst improving availability of other nutrient sources to maize by boosting rhizhodeposition,enzyme synthesis and activity in the rhizosphere (Kuzyakov and Cheng 2004). Similarly,Abbott and Hinz (2012) reported that incorporation of clay and organic matter into sandy agricultural soil in Australia increased soil MBC. 3) Bentonite also helps immobilize enzymes onto its surface thereby increasing the size of the extra-cellular enzyme pools. Enzymes found in organic materials added to soils also help increase soil enzyme activity (Tripathi et al.2007).

Indirect effects on increasing enzyme activity by the B-HA addition may include:1) A more suitable soil waterair environment such as improved soil water retention and oxygen level at the microsites in the rhizosphere that favors the microbial and crop growth was established after addition of bentonite (Zhou et al.2019). 2) Dead microbial and root cells and their released metabolites serve as substrates for increased growth of other microbial populations,and the synthesis and preservation of active enzymes in the presence of bentonite.

4.3.Soil microbial biomass,nutrient cycling and N and P content in grain

Our correlation analysis showed significant relationships between MB and the four soil enzyme activities (Table 5).One of the important ecological functions of the heterotrophic soil MB is to decompose organic materials thereby releasing nutrients which are subsequently used by crops (Monreal and Zhang 2018). During maize growth,increased root growth and exudation help increase soil MB pool size and associated metabolic activities that control the supply of soil nutrients for increased plant growth. Increased amounts of root exudates supply additional organic substrates to microorganisms,a process that also favors microbial growth and the turnover of nutrients for crop growth (Li et al.2016;Li et al.2017). Results from our study indicate that the addition of B-HA increased soil MB pool size and the activities of the four enzymes thereby promoting improved nutrient cycling in maize rhizosphere.

The content of soil MBN,MBP and enzyme activities were significantly correlated (P≤0.05) with grain N and P content.The increased MB and the enhanced activities of urease and alkaline phosphatase enhanced the supply of N and P for maize uptake and subsequent accumulation in the grain and improved crop growth (Appendix A).

5.Conclusion

The addition of Bentonite-humic acid (B-HA) significantly(P≤0.05) increased soil physical,chemical and biological properties in a semi-arid degraded sandy soil. The application of B-HA was the most important factor to increase WFPS and SOC content,especially from year 3 to 5 after its addition. In turn,maize growth stage and B-HA addition together increased MB pool size and the activities of the four soil enzymes explaining 86 and 84% of the total variation in MB and enzyme activities,respectively. Within the 7-yr time scale,the largest significant soil MB pool size and enzyme activity occurred at year 5 after B-HA addition.Maize grain N and P content increased under B-HA,and their contents were significantly correlated with MBN,MBP,and the activities of urease and alkaline phosphatase,catalase and invertase. The addition of B-HA improves soil health,and crop productivity and grain quality,and shows promise as a strategy to improve soil agriculture sustainability in sandy dryland soils which typically have low fertility and low productivity in semi-arid areas of the world.

Acknowledgements

We acknowledge the financial support provided by the National Special Fund for Agro-scientific Research in the Public Interest of China (201303126) and Agricultural Science and Technology Achievements Transformation Demonstration of Production and Application Technology and Popularization of Sandy Soil Amendment,Inner Mongolia,China (sq2012eca400008). We also thank the China Scholarship Council-Agriculture and Agri-Food Canada Joint Scholarship Program.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendixassociated with this paper is available on

http://www.ChinaAgriSci.com/V2/En/appendix.htm