Laboratory and numerical modelling of irrigation infiltration and nitrogen leaching in homogeneous soils

Lei WU ,Ruizhi LI ,Yan WANG ,Zongjun GUO ,Jiaheng LI ,Hang YANG and Xiaoyi MA

1Key Laboratory of Agricultural Soil and Water Engineering in Arid and Semiarid Areas,Ministry of Education,Northwest A&F University,Yangling 712100(China)

2Blackland Research and Extension Center,Texas A&M AgriLife Research,Texas A&M University,Temple TX 76502(USA)

3State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau,Northwest A&F University,Yangling 712100(China)

4College of Water Resources and Architectural Engineering,Northwest A&F University,Yangling 712100(China)

ABSTRACT Nitrogen(N)plays a key role in crop growth and production;however,data are lacking especially regarding the interaction of biochar,grass cover,and irrigation on N leaching in saturated soil profiles.Eighteen soil columns with 20-cm diameter and 60-cm height were designed to characterize the effects of different grass cover and biochar combinations,i.e.,bare soil+0%biochar(control,CK),perennial ryegrass+0%biochar(C1),Festuca arundinacea+0%biochar(C2),perennial ryegrass+1%biochar(C3),perennial ryegrass+2%biochar(C4),perennial ryegrass+3%biochar(C5),F.arundinacea+1% biochar (C6), F. arundinacea+2% biochar (C7),and F. arundinacea+3% biochar (C8),on periodic irrigation infiltration and N leaching in homogeneous loess soils from July to December 2020.Leachates in CK were 10.2%-35.3%higher than those in C1 and C2.Both perennial ryegrass and F.arundinacea decreased the volumes of leachates and delayed the leaching process in the 1%,2%,and 3%biochar treatments,and the vertical leaching rate decreased with biochar addition.The N leaching losses were concentrated in the first few leaching tests,and both total N(TN)and nitrateconcentrations in CK and C1-C8 decreased with increasing leaching test times.Biochar addition(1%,2%,and 3%)could further reduce the leaching risk of and the loss decreased with biochar addition.However,compared to 1%biochar,2%biochar promoted the leaching of TN under both grass cover types.The N leaching losses in CK,C1,C2,C3,C4,C6,and C7 were primarily in the form of .Among these treatments,CK,C1,and C2 had the highest cumulative leaching fractions (＞90%),followed by those in C3,C4,C6,and C7(＞80%).The cumulative leaching fraction of decreased with increasing leaching test times and biochar addition,and 3%biochar addition(i.e.,C5 and C8)reduced it to approximately 50%.The one-dimensional advective-dispersive-reactive transport equation can be used as an effective numerical approach to simulate and predict leaching in saturated homogeneous soils.Understanding the effects of different biochar and grass combinations on N leaching can help us design environmentally friendly interventions to manage irrigated farming ecosystems and reduce N leaching into groundwater.

Key Words: leaching loss,nitrate nitrogen,biochar,grass cover,analytical modelling

INTRODUCTION

Leaching is a primary transport mechanism for nitrate ions(Maet al.,2021).Nitrate leached below the crop-root zone often ends up as a pollutant in groundwater supplies (Teixeiraet al.,2021).Dryland,accounting for 54.4%of the total cultivated land area in China,is an important land use type regarding nitrogen (N) leaching because of the extensive distribution of irrigated and rain-fed agriculture(Huet al.,2011).Research has been performed on the effects of fertilizer application rates,tillage measures,and different soil amendments on N leaching(Yanget al.,2013;Huet al.,2019;Liu Det al.,2020;Xuet al.,2022).However,understanding the effects of agricultural interventions such as biochar addition and grass cover on N leaching processes associated with consecutive irrigation/rainfall is crucial in reducing N leaching losses in irrigated or rainfed agroecosystems(Wuet al.,2018)and is a key knowledge gap between the mechanisms of N leaching under consecutive irrigation/rainfall and the development of N management plans in agricultural or terrestrial systems of arid and semiarid regions.

In the last two decades there has been focused interest on the use of biochar for agricultural purposes in part due to the introduction of the term“ecological intensification”and biochar’s considerable abilities to adsorb nutrient(Wuet al.,2021,2022) and hence decrease nutrient leaching.Biochar is widely used as a soil amendment because its unique properties can improve the physical and chemical properties of soil and retain moisture and nutrients,thereby affecting soil N leaching(Yuan and Xu,2011;Li and Chen,2020).The effect of biochar amendments on the leaching of nutrients is a consequence of complex chemical,physical,and biological processes occurring within the soil(Luet al.,2021).Biochar can alter soil physical properties such as bulk density,porosity,permeability,and water retention.In a 45-week soil column leaching study(Lairdet al.,2010),addition of hardwood biochar to a typical Midwestern agricultural soil reduced netleaching from 60%to 40%of the N added with manure.Therefore,the main challenge for the risk assessment of N leaching in the dryland agroecosystem of the Loess Plateau is to accurately quantify and visualize the potential impacts of grass and biochar associated with consecutive irrigations.

The simulated soil column test,mostly used for the study of soil nutrient leaching,can overcome the complexity of the field environment and reduce the possible errors of over-or underestimating nutrient leaching by modelling methods to a certain extent (Zhouet al.,2006),largely because the conditions are easy to control (Wanget al.,2005).Research on the effects of biochar on soil N leaching has mainly focused on the amount of biochar addition,with different levels of biochar addition having different effects on N leaching in different soils(Xiaoet al.,2015).There are few studies on the combination of biochar with other factors,especially the different growth cycles of grass.Moreover,previous studies have shown contradictory results,with a noticeable effect of biochar on stimulating available N inputs and reducing N losses in short-term laboratory experiments(Ahmadet al.,2021).Reduced N leaching from biochar-amended soils has been documented in bare lands;however,different combinations of agricultural interventions may have better effects or overcome the disadvantages of single practice (Karhuet al.,2021).Thus,it is of great importance to contextualize lab-scale experiments based on biochar,grass cover,irrigation,and soil columns to provide a holistic approach for understanding the complex processes responsible for modulating N leaching in loess soil.Therefore,a set of laboratory-simulated columns filled with bare loess and grassed soils was selected to quantify N leaching processes with different levels of biochar addition.Moreover,the accurate modeling of solute transport in porous media is crucial for the remediation of contaminants in soils and aquifers.These requirements highlight the urgent need for reliable evaluations that can model rates of soil N leaching,especially,and assess the effects of different grass and biochar combinations on N lossviasubsurface hydrological pathways.

The objectives of this study were to:i)investigate the leachate characteristics of irrigation water through loess soils under different combinations of biochar and grass,ii)quantify the interactive effects of grass and biochar on irrigation leachates and N leaching processes,iii)develop a comprehensive advective-dispersive-reactive equation to characterize the migration and transformation processes of N transport in a porous medium,and iv)target agricultural management to explore effective practice options for tacklingleaching losses across the dryland agroecosystems in arid and semi-arid regions.The results are of great value for the design of environmentally friendly management practices focused on identifying and controlling hot spots of N leaching loss to reduce the risk of groundwater deterioration.

MATERIALS AND METHODS

Loess soil and biochar

Soil column leaching tests were conducted from July to December 2020 at the College of Water Conservancy and Architectural Engineering,Northwest A&F University,Yangling,China.After natural air-drying,the tested loess soil was sieved through a 0.5-cm sieve to remove stones,roots,and other debris.Biochar,with a total nitrogen(TN)content of 3.75 g kg-1and a particle size of 80-400 mesh,was made from tobacco straw by dry distillation at 500-600°C and purchased from Shaanxi Yixin Bioenergy Technology Development Co.,Ltd.(China).The texture and particle size distribution of the tested soil were sandy loam,6.28%clay(＜0.002 mm),36.82%silt(0.002-0.05 mm),and 56.89%sand(＞0.05 mm).The initial TN content was 0.23 g kg-1.

Leaching device

Three self-designed movable trolleys were used to simulate N leaching loss from the loess soil.Six soil columns and six Plexiglas Mahalanobis bottles were placed in each trolley,totaling 18 soil columns and 18 Mahalanobis bottles(Fig.1).The diameter and height of soil columns were 20 and 60 cm,respectively,and those of Mahalanobis bottles were 10 and 60 cm,respectively.Holes were evenly opened at the bottom of soil column,and a funnel was installed at the bottom to collect the leachates.

Fig.1 Schematic diagram of soil columns simulating irrigation infiltration and N leaching affected by different grass cover and biochar combinations:bare soil+0%biochar(control,CK),perennial ryegrass+0%biochar(C1),Festuca arundinacea+0%biochar(C2),perennial ryegrass+1%biochar(C3),perennial ryegrass+2%biochar(C4),perennial ryegrass+3%biochar(C5),F.arundinacea+1%biochar(C6),F.arundinacea+2%biochar(C7),and F.arundinacea+3%biochar(C8).There are two replicates for each treatment with the same soil conditions.

Design of leaching tests

Nine treatments including bare soil+0%biochar(control,CK),perennial ryegrass+0%biochar(C1),Festucaarundinacea+0%biochar(C2),perennial ryegrass+1%biochar(C3),perennial ryegrass+2%biochar(C4),perennial ryegrass+3% biochar (C5),F.arundinacea+1%biochar (C6),F.arundinacea+2% biochar (C7),andF.arundinacea+3%biochar(C8)were designed,each with two replicates.Prior to the experiment,the biochar and the tested air-dried soil were thoroughly mixed using a mechanical mixer according to the mass ratios of 1%,2%,and 3%.A layered filling method was employed to fill the soil column,loading the tested soil into the soil column to form a 5-cm soil layer each time,with a total of 10 layers.Each soil layer was weighed according to the designed bulk density,which was controlled to be about 1.3 g cm-3and compacted to ensure that there were no gaps between the layers.A layer of nylon net and a layer of 200-mesh nylon cloth were laid at the bottom of soil column,filled with 3 cm high quartz sand,and another layer of 200-mesh nylon cloth was laid on the quartz sand to prevent the upper soil loss.After the soil column was filled,perennial ryegrass(Lolium perenneL.) andF.arundinaceawith the same seed weight were sowed on the corresponding soil column.The leaching test started after a period of grass growth.In this experiment,the applied amount of N fertilizer(urea)was 120 kg ha-1according to the local fertilization level,and the fertilizer was evenly mixed into surface soil(0-10 cm depth)based on local fertilization methods.

Sampling test and analysis

Before the leaching test,soil column was saturated with water supplied by a Mahalanobis bottle under a constant water head(2 cm),and leaching tests were conducted after 1 d of standing.During the leaching test,500 mL distilled water was continuously added to soil column.After the water was initially discharged from the bottom,the water sample was collected in a polyethylene bottle,and the time required for irrigation for leachate generation was recorded(Fig.2).The continuous sampling mode lasted for 5 min each time with a total of six times.The remaining leachate was then collected in a polyethylene bottle.After site sampling,the water samples were taken back to the laboratory to measure the volume of the leachates and to determine the N concentration in the leachates.The water samples were filtered through a 0.45-μm microporous membrane to determine theN content.Then,washing was performed every 4 d using 500 mL distilled water each time with a total of eight times.The volumes of leachates were determined by measuring the cylinder,using ultraviolet spectrophotometry,and TN using ultraviolet spectrophotometry after alkaline potassium persulfate digestion.

Fig.2 Schematic diagram of N leaching sampling.

Advective-dispersive-reactive(ADR)transport equation

whereC1is theconcentration in soil solution(mg L-1),xis the soil depth (cm),tis the leaching time (h),Dis the solute diffusion-dispersion coefficient(cm2h-1),vis the actual seepage velocity in porous media(cm h-1),knis the first-order kinetic constant of nitrification(h-1),kbiois the zero-order kinetic constant of biological fixation adsorption (mg L-1h-1),kdenis the first-order kinetic constant of denitrification (h-1),C0is the initialconcentration in soil solution(mg L-1),Eis the evaporation rate (cm h-1),C01andC02are theandconcentrations,respectively,in soil solution at the upper boundary(x=0)(mg L-1),Cl1andCl2are theandconcentrations,respectively,in soil solution at the lower boundary (x=l) (mg L-1),C2is theconcentration in soil solution(mg L-1),is the initialN concentration in soil solution(mg L-1),ρis the soil bulk density (g cm-3),θis the soil saturated moisture content or effective porosity (cm3cm-3),kd2is the distribution coefficient ofin soil solution and solid phase(cm3g-1),kminis the zero-order kinetic constant of mineralization(mg L-1h-1),andkvis the zero-order kinetic constant of volatilization(mg L-1h-1).

Baseline conditions were defined using the parameters listed in Table I(Goltz and Huang,2017;Ercan,2020)and then analytical modelling was conducted by varying each parameter to observe how the contaminant fate and transport concentrations responded to such changes.

By solving the one-dimensional partial differential equations with the initial and boundary conditions listed in Table II,the exact solutions of Eqs.1 and 5 were obtained.

RUSULTS

Leaching rates and leachate volumes

The dynamic changes in the average leaching rates of irrigation water were different under the different treatments(Fig.3).The average leaching rates in CK,C1,C2,C3,and C6 were relatively stable at the early stage of leaching process and then decreased,whereas the leaching process of C4,C5,C7,and C8 showed a gentle trend generally.The average leaching rate within 20 min in CK was higher(1.11-1.55 times)than those in the other treatments.The 30-min average leaching rate in CK was 11.2 mL min-1,which was 5.6%-33.0% higher compared with the other treatments.Treatments C1 to C8 slowed the leaching rate compared with CK.The 30-min average leaching rate of CK was higher than that of C1 followed by C2;however,this order varied significantly at the end of the leaching test with the average leaching rate of CK decreasing to the lowest value(C2＞C1＞CK).That is,grass planting limited the leaching rate in the early stage of leaching process,but increased it in the late stage.

TABLE IBaseline parameters for advective-dispersive-reactive (ADR) transport equation

Fig.3 Changes in average leaching rates of irrigation water with leaching time under the treatments with different grass cover and biochar combinations.See Fig.1 for the detailed descriptions of treatments CK and C1-C8.

Regarding different grass types,the maximum leaching rate in C1 was 1.1 times that in C2.The 30-min average leaching rates in C1,C3,C4,and C5 with perennial ryegrass were in the order of C1＞C3＞C4＞C5.Compared with C3,C4,and C5,C1 increased leaching rate by 8.0%,14.4%,and 25.9%,respectively.The order of 30-min average leaching rate in C2,C6,C7,and C8 underF.arundinaceawas C2＞C6＞C7＞C8.The average leaching rate in C2 was 10.3%,13.2%,and 14.8% higher than those in C6,C7,and C8,respectively,indicating that biochar significantly reduced the leaching rate and that the average leaching rate decreased with increasing biochar addition.

The 30-min average leaching rates in C1-C8 were 10.6,10.4,9.9,9.3,8.5,9.4,9.2,and 9.1 mL min-1,respectively,and the order was C1＞C2＞C3＞C6＞C4＞C7＞C8＞C5.Treatment C1 increased leaching rate by 2.4%compared with C2,C3 increased it by 4.6% compared with C6,C4 increased it by 1.3%compared with C7,and C8 increased it by 7.1%compared with C5,indicating that the leaching rate underF.arundinaceatreated with 0%,1%,and 2%biochar was lower than that under perennial ryegrass,but the leaching rate underF.arundinaceatreated with 3%biochar was higher than that under perennial ryegrass.

The results of theF-test double-sample variance analysis showed that there were significant differences(P ＜0.05)between C5 and the other eight treatments,particularly between C5 and CK(P ＜0.01).There were also significant differences(P ＜0.05)between CK and C6,CK and C7,and CK and C8.However,no significant differences were observed between any of the other treatments,regardless of biochar addition or grass planting.

The cumulative volume of leachates can be used to characterize the differences in the effects of different agricultural interventions on the seepage of irrigation water.Different treatments showed no significant but only a small difference on the total volume of leachates (Fig.4).Compared with CK,the cumulative volumes of leachates in treatments C5,C7,and C8 increased by 0.76%,1.20%,and 6.55%,respectively,and those in C1,C2,C3,C4,and C6 decreased by 4.9%,2.4%,4.8%,0.2%,and 2.8%,respectively.Compared with CK,there was no significant interception effect on the leaching of irrigation water in the other treatments;the cumulative volume of leachates was approximately 2 900 mL under the different treatments,except for C8 where the volume was 3 149 mL.In other words,C8 not only had no water-holding effect,but also increased the volume of leachates by 193 mL compared with CK.The cumulative volumes of leachates in C1,C3,C4,and C5 followed the order of C1＜C3＜C4＜C5 and that in C1 was 0.1%,4.7%,and 5.6%lower than those in C3,C4,and C5,respectively.In contrast,the cumulative volumes of leachates in C2,C6,C7,and C8 followed the order of C6＜C2＜C7＜C8 and that in C2 was 3.6%and 8.4%lower than those in C7 and C8,respectively,and 0.4%higher than that in C6.Furthermore,the cumulative volume of leachates increased with increasing biochar addition under the same grass type.With the same amount of biochar addition,the cumulative volume of leachates was in the order of C1＜C2,C3＜C6,C4＜C7,and C5＜C8,indicating thatF.arundinaceaincreased the volume of leachates compared to perennial ryegrass.

Fig.4 Cumulative volumes of leachates after eight leaching tests under the treatments with different grass cover and biochar combinations.See Fig.1 for the detailed descriptions of treatments CK and C1-C8.

Total N leaching characteristics

The changing trends of TN concentration in leachates in the first,second,and eighth leaching tests are presented in Fig.5.In the first leaching test,the TN concentration in CK was much higher than that in the other treatments and its trend was relatively stable.Conversely,the TN concentration trends in C1-C8 decreased obviously with leaching time,and the decreasing rates of TN concentration in CK to C8 were 0.14-1.60 mg L-1min-1in the whole leaching process.In the second leaching test,the TN concentrations in CK to C8 decreased with increasing leaching time and the decreasing rates of TN concentration in CK to C8 were 1.21-2.30 mg L-1min-1in the whole leaching process,which were higher than those in the first leaching test.However,the TN concentrations in leachates in all treatments in the second leaching test were lower than those in the first leaching test,and the TN concentrations in C5 and C8 were obviously less than that of other treatments.In the eighth leaching test,the TN concentrations in the nine treatments were stable during the leaching process,with the average being 1.99(CK),2.45(C1),2.47(C2),2.02(C3),3.66(C4),7.05(C5),2.61(C6),3.88(C7),and 6.82(C8)mg L-1.Compared with CK,the TN concentrations in C5 and C8 increased by 254%and 242%,respectively.Interestingly,the TN concentration varied greatly in each leaching test,but the TN concentration in the treatments decreased with increasing leaching time.

Fig.5 Total N(TN)concentrations in leachates over leaching time in the first(a),second(b),and eighth(c)leaching tests under the treatments with different grass cover and biochar combinations.See Fig.1 for the detailed descriptions of treatments CK and C1-C8.

The TN concentrations in leachates and cumulative leaching losses in eight leaching tests varied considerably under the different treatments(Fig.6).Compared with CK,TN leaching losses in the other treatments decreased from 134.7%to 61.3%,indicating that agricultural interventions(C1-C8) effectively reduced TN leaching loss.Biochar addition further inhibited the leaching loss of TN based on grass cover.For instance,the TN leaching losses in C3,C4,and C5 were 13.9%,7.3%,and 30.2%,respectively,lower than that in C1 with perennial ryegrass and those in C6,C7,and C8 were 19.6%,13.6%,and 16.4%,respectively,lower than that in C2 withF.arundinacea.However,the relationship between TN leaching loss and the amount of biochar addition was not clear,following the order of C4＞C3＞C5 and C7＞C8＞C6.The TN leaching losses with 2%biochar addition for perennial ryegrass andF.arundinaceaincreased by 7.7%and 7.4%,respectively,compared with that with 1%biochar addition.Treatment C5(3%biochar addition)had the lowest TN leaching loss,indicating that 3%biochar addition under perennial ryegrass was the best practice for reducing TN leaching.

Fig.6 Total N(TN)concentrations in leachates in eight leaching tests and cumulative leaching losses after eight leaching tests under the treatments with different grass cover and biochar combinations.See Fig.1 for the detailed descriptions of treatments CK and C1-C8.

Fig.7 Nitrate-N()concentrations in leachates in eight leaching tests and cumulative leaching losses after eight leaching tests under the treatments with different grass cover and biochar combinations.See Fig.1 for the detailed descriptions of treatments CK and C1-C8.

Under the same grass cover,the order ofleaching losses in the different treatments was C1＞C3＞C4＞C5 and C2＞C6＞C7＞C8.Theleaching loss decreased with increasing biochar addition.Compared to C1,the cumulativelosses in C3,C4,and C5 decreased by 17.4%,20.1%,and 64.0%,respectively;whereas those in C6,C7,and C8 decreased by 24.3%,24.8%,and 50.7%,respectively,when compared with C2.There was no significant difference between 1%and 2%biochar addition,but 3%biochar addition strongly inhibitedleaching.Moreover,the cumulative losses ofin C3,C4,and C8 were 8.1%,5.2%,and 38.3%higher than those in C6,C7,and C5,respectively,with the same amount of biochar addition.The effect of grass cover types onleaching loss was much stronger under 3%biochar addition.

Fig.8 Variations in nitrate-N()fraction in leachates in eight leaching tests and cumulative leaching fractions of and ammonium-N()after eight leaching tests under the treatments with different grass cover and biochar combinations.See Fig.1 for the detailed descriptions of treatments CK and C1-C8.

The cumulative leaching fractions ofandafter eight leaching tests are presented in Fig.8.The cumulative leaching fractions ofin CK,C1,and C2 reached 93.75%,93.15%,and 92.51%,respectively,and there was no significant difference between them.The cumulative leaching fractions ofin C3,C4,and C5 were 89.34%,80.28%,and 47.97%,respectively,corresponding to a decrease of 4.09%,13.81%,and 48.5%compared with C1.The cumulative leaching fractions ofin C6,C7,and C8 were 87.01%,80.59%,and 54.52%,respectively,decreasing by 5.94%,12.89%,and 41.06%,respectively,compared to C2,indicating that the cumulative leaching fraction ofdecreased with increasing biochar addition.More importantly,the cumulative leaching fractions ofin C5 and C8 were 47.97%and 54.52%,respectively,indicating that 3% biochar addition significantly reduced the leaching fraction of.However,the cumulative leaching fraction ofin C1 was 44.7%,74.8%,and 86.5% lower than those in C3,C4,and C5,respectively,while that in C2 was 18.4%,56.4%,and 76.8%lower than those in C6,C7,and C8,respectively.This indicated that biochar addition increased the cumulative leaching fraction of,which increased with increasing biochar addition.However,the cumulative leaching fractions ofin all treatments were low(＜20%).

Overall,the cumulative leaching fraction ofwas over 80%in CK,C4,C6,and C7 and approximately 50%in C5 and C8,indicating thatwas the main form of N leaching.

DISCUSSION

Variations in leachate volume

The leaching rates of irrigation water in perennial ryegrass andF.arundinaceawith 0%biochar were lower than that in bare loess soil.These results are similar to those of Wu(2017),who found that the water leaching rate of grassed soil was lower than that of bare soil.This is mainly due to the fact that the physiological and biochemical functions of grass lead to the total volumes of leachates under perennial ryegrass andF.arundinacealower than that of bare soil(Li and Tan,2006).Moreover,soil voids are partly occupied by grass roots,which reduce the void ratio of grassed soil and thus the soil water infiltration rate(Peregoet al.,2012).The difference could also be explained by the fact that grass promotes the development of biological crusts on the soil surface.After absorbing water,the biological crust produces a water-resistant layer,which covers the soil surface with a layer of hydrophobic substances(Franzluebberset al.,2014).Therefore,the existence of a biological crust reduces the soil infiltration performance and slows water infiltration(Wanget al.,2015).In addition,compared to no biochar application,biochar addition significantly reduced the leaching rate of water under grass cover,and the higher the biochar addition,the lower the leaching rate of water.This is because biochar may also increase soil water-holding capacity owing to its large surface area and high porosity,thereby reducing soil water percolation and changing the water infiltration process(Liet al.,2016).The results of Xieet al.(2016) further confirmed the above view that biochar significantly inhibited the water infiltration of loess soil and that the higher the amount of biochar addition,the lower the water infiltration capacity.This may be attributed to the decrease in soil pores or effective poresviathe addition of biochar,which hindered water movement(Liu Jet al.,2020),resulting in the decline of soil infiltration performance.Therefore,the addition of biochar under grassy conditions can further inhibit soil water infiltration and slow the leaching of soil water.

The leaching rates of irrigation water under perennial ryegrass with 1%and 2%biochar addition were 4.6%and 1.3%,respectively,higher than those underF.arundinacea,whereas the water leaching rate under perennial ryegrass with 0%biochar was 2.4%higher than that underF.arundinacea.These results indicated that 1%biochar addition under perennial ryegrass may accelerate the leaching loss of soil water compared toF.arundinacea.However,the leaching rate of soil water underF.arundinaceawith 3%biochar was 7.1%higher than that under perennial ryegrass.This is because the inhibitory effect of high biochar addition on water leaching loss is uncertain under different grass covers,and its impact on grass growth and soil properties may be strengthened when the application rate of biochar reaches a certain level.

There was no obvious difference between 1%and 0%biochar addition in the total volume of leachates in grassed soil,but 2% and 3% biochar addition increased the total volume of leachates compared with 0% biochar addition.This may be attributed to the fact that excessive biochar has not only hydrophilicity but also hydrophobicity owing to the change in soil physicochemical properties,which leads to the promotion of water leaching loss(Kinneyet al.,2012).When compared with 0%biochar addition,C8(3%biochar addition underF.arundinacea)showed the largest difference(increased by 9.2%)in leaching loss of soil water in this test,but the difference was relatively small in the other treatments.

N leaching processes

The TN concentrations in leachates in the different treatments exhibited a decreasing trend with leaching time in the first and second leaching tests.However,the decreasing rate in the first leaching test was lower than that in the second leaching test.In the eighth leaching test,the TN concentrations in leachates in the different treatments tended to be stable and extremely low during the leaching process.The above results can be attributed to the following three key points:i)the retention of chemical fertilizer in soil is poor,and the N content in the bottom soil layer was higher than that in the top soil layer due to the downward movement of fertilizer N with leaching water;ii)the larger the basic content of N fertilizer in soil,the smaller the decreasing rate,e.g.,the higher N content in the tested soil in the first leaching test caused a lower leaching rate of TN,while the relatively low N content in the second leaching test led to an increase in the leaching rate of TN.Thus,soil N content in the soil column decreased as the leaching test progressed and the decreasing rate increased;and iii) when N fertilizer was completely lost,the background N contained in the soil and biochar itself was evenly lost with the leaching solution,making the final TN concentration in leachates stable.Moreover,theconcentration in leachates also showed a decreasing trend with increasing times of leaching test,and leaching loss mainly occurred in the previous leaching tests.The main reason may be thatgenerally exists in the form of negatively charged anions,and it has the characteristics of high solubility,mobility,and easy displacement by water.As so,it is not easily adsorbed by soil and is easily lost with leaching solution,thus leading to the widespread appearance ofin groundwater(Wanget al.,2018).

Generally,because of its unique structure and strong adsorption ability,biochar can effectively reduce N leaching(Li and Wei,2016;Wuet al.,2017).In the present study,it was found that 1%,2%,and 3%biochar addition significantly reduced the leaching loss ofcompared with 0%biochar addition under the two grass covers,and the higher the amount of biochar addition,the lower the leaching loss of.This is because the addition of biochar or other soil amendments generally increases soil microbial biomass(Bruunet al.,2012).The increased microbial biomass can entrap the added N fertilizer in the biomass and cause N immobilization,which help mitigate N leaching(Tammeorget al.,2012).The results also demonstrated that biochar can prevent N leaching by absorbing and fixing some of the(Liet al.,2019).Liuet al.(2018)conducted a metaanalysis of 36 related studies and found that biochar reduced N leaching by an average of 26%.However,there was no significant difference between the effects of 1% and 2%biochar addition on N leaching.Compared with 0%biochar addition,the decreasing rate of totalloss with 3%biochar addition was 2.1-3.7 times those with 1%and 2%biochar addition,and the inhibition effect of 3% biochar addition onloss was obviously enhanced.This may be because there is a critical value between the adsorption capacity and the amount of biochar addition.In other words,when the amount of biochar addition reaches a certain level,the adsorption capacity is significantly enhanced.The effect of biochar on N leaching was not obvious at relatively low addition,and only 4% addition significantly reduced the leaching loss of(Gaoet al.,2014).Similarly,both increasing and decreasing leaching rates are possible when polyacrylamide is added to soil at different concentrations(Yuanet al.,2005).

However,high biochar addition may aggravate the leaching loss of TN.It was found that TN loss was obviously higher with 3% biochar addition than in other treatments in the seventh and eighth leaching tests.This phenomenon can be mainly attributed to the fact that the effect of biochar on soil N loss is closely related to its type and amount and soil texture.Biochar is mostly made of biomass,and a large number of nutrients,such as N and phosphorus,are retained in biochar in the process of hydrolysis (Qinet al.,2021),resulting in an increased N content in biochar itself (Weiet al.,2020).Therefore,the application of excessive biochar to loess soil is risky,but the amount of biochar addition can be optimized according to the application purpose and scenario and biochar can be combined with plant measures based on scientific dosage.Plants need to absorb a large amount of available N during the growth process(Liuet al.,2019),which also allows perennial ryegrass andF.arundinaceato effectively reduce the leaching losses of TN andcompared to bare soil,indicating that the leaching loss risk was distinctly higher in bare soil than in grassed soil.In addition,the leaching losses ofand TN withF.arundinaceawas slightly higher(by 0.9%and 1.6%,respectively)than that with perennial ryegrass,indicating that there was no significant difference in the effect of the two grass covers on N leaching loss.Nitrate-N is a relatively stable compound that forms a major reservoir of biologically available N in the soil(Menget al.,2021).Whether nutrients,regardless of the source,are leached or retained within the soil depends on the extent to which they remain in the soil solution,are adsorbed on the surface of soil particles,precipitated as insoluble or sparingly soluble inorganic phases,retained in immobile soil water,or incorporated into soil organic matter.

Variations in leaching fraction

However,in the seventh leaching test,except for 3%biochar addition,the fraction ofin leachates in the other treatments increased to different degrees compared to the sixth leaching test(Fig.8).This may be attributed to the fact that the loss ofin the last few tests was very low,close to zero,whereas the loss of TN,especially in the seventh leaching test,was relatively slow owing to the accumulation and adsorption of different N forms,thus increasing the ratio ofto TN.Specifically,the fraction ofN in leachates decreased with biochar addition,while the treatments of perennial ryegrass andF.arundinaceashowed similar decay trends,although the decline inF.arundinaceatreatment was not obvious.This is because the leachates from biochar treatment have been proposed to have both positive and negative effects on plant growth(Thomas,2021).Regarding 3% biochar treatment,its adsorption capacity,especially of,was higher than those of 0%-2%biochar treatments.Compared to the sixth leaching test,the signs of decay in the two types of grass with the relatively low amount of biochar addition(0%-2%)reduced the ability ofabsorption and adsorption and increased the risk ofloss,thus increasing the fraction ofin leachates in the seventh leaching test.This phenomenon also indicated that the decay of grass would lead to lessleaching in the short term,but more intensive leaching in the long term(Zhang Wet al.,2021).

Notably,3%biochar addition significantly reduced the cumulative leaching fraction ofafter eight leaching tests compared to the other treatments.The cumulative leaching fractions ofwere approximately 50%under two grass covers with 3%biochar addition(i.e.,C5 and C8),indicating that the effect of biochar on nutrient leaching is complex.It can not only alter soil physical properties such as bulk density,porosity,permeability,and water retention(Hossainet al.,2020),but also soil chemical properties,such as pH,cation exchange capacity,anion exchange capacity,and the capacity to adsorb soluble organic and inorganic compounds.However,this is not representative,mainly because of the complexity caused by biochar application and possible calculation errors caused by the lowcontent.Nguyenet al.(2017)conducted a meta-analysis of changes in soil available inorganic N(SIN)under the influence of biochar and found that the variability in the influence of biochar on SIN was mainly affected by factors such as biochar residence time,pyrolysis temperature,application amount,fertilizer type,and soil pH,which also affected biochar in preventing N leaching.More importantly,the provision of multiple services by biochar depends on the biomass expression of functionally diverse species,because biochar can also influence microbial populations and their activity in soils and plant growth and differentiation.

Moreover,unreasonable irrigation causes serious N leaching because soil moisture is the basic driving force for N leaching(Zhu and Chen,2002).The main principle for preventing and controlling N leaching through agricultural water management is to reduce water leakage in the soil root layer,which can be primarily achieved by optimizing irrigation schedule,adjusting irrigation measures to meet the water demand of crops,and reducing the flow of excess water to the lower soil layer(Barton and Colmer,2006).

Modelling of contaminant transport in soils

The developed numerical equation was tested using a simple irrigated soil column,in which we simulatedleaching through the saturated zone for different treatments using eight tests.Assuming that the thickness of loess layer is 1000 cm,the unused N in bare and vegetated loess soils can be lost through leaching,denitrification,or volatilization.In the early stage of leaching modelling,theconcentration in leachates was high,and the difference in loss amounts in different soil layers(0-800 cm)was relatively small.However,with an increase in soil depth,theconcentration in the deep soil,especially below 850 cm,decreased sharply(Fig.9).Taking CK as an example,the 5-minconcentration in leachates in the first leaching process was 168.24 mg L-1,and the 50-minconcentration in leachates was 164.59 mg L-1,revealing a decrease of only 3.65 mg L-1;in the eighth leaching process,the 5-and 50-minconcentrations in leachates were only 0.815 and 0.257 mg L-1,respectively.This confirms thatanions are highly susceptible to leaching losses,particularly in irrigated or tile-drained agricultural fields.Moreover,the initialin soil solution was 1.984 mg L-1,which decreased to 0.962 mg L-1in the eighth leaching test.Practically,the factors affecting the analytical solutions of the equations used here include advection,dispersion,sorption,first-order degradation,and boundary conditions(Schulze-Makuch,2011).The effect of advection on the subsurface transport of a chemical is related to the pore velocity of the groundwater(De Filippiset al.,2021).The spread of the chemical concentration profiles in space and time may increase as the dispersion coefficient(D)increases.More importantly,the magnitude of leaching losses typically depends on management factors such as rate and timing of N application and on the timing,frequency,intensity,duration,and amount of infiltration(Zhang J Jet al.,2021),which affects soil hydrologic budget(irrigation,evapotranspiration,storage,etc.)and crop management practices(crop rotation and tillage).The leaching amount and rate are closely related to soil properties,primarily soil structure and texture(Wang and Li,2019).

Fig.9 Modelling the concentrations of and along soil profile over leaching time.

The plume of contaminated water in the upper soil moves mainly vertically downward and this is usually regarded as one-dimensional transport (Marjerisonet al.,2016).The flow rate of contaminants in soil and groundwater is much lower than that in river water,and the adsorption by soil particles further slows the moving rate of contaminants.Mineralization can supply large amounts offor leaching,particularly during fall,when soils are removed by precipitation after crops have been harvested and soils have been tilled(Addiscott,2000).Moreover,the depth of water table limits the impact of leaching on groundwater quality.The greatest risk of groundwater contamination occurs in coarse-textured soils over a shallow water table that receives high amounts of N from fertilizer and manure and high amounts of precipitation or irrigation.Most N losses occur through ammonia volatilization,leaching,and denitrification.Nitrogen loss can be significantly reduced using urease inhibitors(Dinneset al.,2002).In addition,synchronizing N application with N demand of plants is an important strategy for improving N use efficiency;however,crops may not be able to use N efficiently if water is a limiting factor for growth and production(Colombaniet al.,2020).This may result in increased residual N accumulation in the soil after crop harvest,which can degrade environmental quality through N leaching into groundwater and greenhouse gas emissions such as N2O(Wanget al.,2013).More importantly,numerical models can support the estimation ofleaching rates in space and time to support sustainable agricultural management practices(De Filippiset al.,2021).

CONCLUSIONS

A series of simulated soil column tests and numerical analysis were applied to study the comprehensive effects of different underlying surfaces(bare soil,perennial ryegrass,andF.arundinacea) and biochar addition (0%,1%,2%,and 3%) on the processes of irrigation infiltration and N leaching.Agricultural interventions such as biochar addition and grass cover in experimental soil had the potential to curtail N leaching.Biochar and grass were important factors affecting water movement into and through soil profile,because they may change soil properties,structure,surface crusting,and surface sealing.Biochar addition reduced the cumulative leaching fraction of,and this effect can be strengthened by increasing the amount of biochar added.The addition of 3%biochar significantly reduced the cumulative leaching fraction ofto approximately 50%.Although a certain amount of biochar inhibited the leaching ofit may increase the risk of TN leaching,which had a negative effect on the management of farmland N leaching.Notably,the laboratory results cannot be directly popularized in largescale regions because of the complexity and heterogeneity of field conditions.Using models to understand the impact of subsurface processes on contaminant transport is an excellent way to understand natural behaviors.Foreseeable upscaling research should also combine the characteristics of different N leaching resistance measures to realize their combination,optimization,and integration,so as to fully illustrate the advantages and avoid disadvantages of each technical measure,improve N use efficiency,and reduce groundwater pollution caused by N leaching in dryland agroecosystems.

ACKNOWLEDGEMENT

This study was supported by the National Natural Science Foundation of China(Nos.52070158,42277073,and 51679206) and the National Fund for Studying Abroad,China(CSC No.201706305014).