Crop Rotation and Nitrogen Fertilizer on Nitrate Leaching: Insights from a Low Rainfall Study

: The intensive use of agricultural fertilizers containing nitrogen (N) can increase the risk of nitrate (NO 3 − ) leaching. However, little information exists regarding its interaction with other factors that influence NO 3 − leaching, such as no-tillage, which is associated with different crop rotation schemes. The objective of this study was to quantify the leachate NO 3 − concentration and load below the root zone in two different crop rotations under no-tillage, with and without mineral N fertilizer. The experiment was conducted in a no-tillage area in Brazil between 2018 and 2020. The factors were two crop rotations (diversified and simplified) and two N fertilization managements (with and without N fertilizer). The soil solution was collected with suction lysimeters (1 m depth), the NO 3 − concentration (mg L − 1 ) was spectrophotometrically determined, and the NO 3 − load (kg ha − 1 ) was calculated from the volume of water drained and the NO 3 − concentration. The results were categorized into 24 evaluation periods. NO 3 − leaching was extremely low due to low rainfall throughout the experiment, with no significant differences between the factors and treatments. In the presence of N fertilization, leaching was substantially greater when rainfall increased, and vice versa. No significant difference was observed between the crop rotation schemes, except for one period in which the simplified soybean rotation exhibited high leaching. The evaluated treatments showed less NO 3 − leaching during the four periods when grass species were cultivated, indicating the importance of grasses in rotation systems.


Introduction
Nitrogen (N) is an essential element for any plant; therefore, N fertilizer (organic or mineral) is the main input in agricultural production.Excessive N fertilizer use can negatively impact the environment, resulting in decreased drinking water quality and surface water deterioration, eutrophication, and the subsequent occurrence of harmful cyanobacterial blooms, leading to the formation of "dead zones" in rivers, lakes, and oceans, as well as harm to fisheries and tourism [1][2][3][4].These effects are gaining increasing attention as fertilizer consumption increases annually.Global agriculture consumed more than 108 million metric tons in 2021, when the top consumers were China, India, the USA, and Brazil [5].Another relevant N input in agroecosystems is the symbiotic association between plants and N-fixing microorganisms [6], which can also result in the loss of this nutrient.In 2021, global agriculture received around 40 million metric tons of N through biological fixation [7], which is still relevant, despite being lower than the amount from fertilizer.
Estimates indicate that up to 70% of the total N available to plants is lost to the environment [8], with NO 3 − leaching being the major contributor [9].NO 3 − is the predominant mineral form of N in most soils and remains free in soil solutions owing to its reduced electrostatic adsorption, leading to easy leaching and quick passage to subsurface and surface waters [2,10,11], which ultimately impacts the environment.NO 3 − leaching in the soil is a complex process, and its extent depends on several factors, such as soil characteristics (e.g., anion exchange capacity, texture, and hydraulic conductivity), plant characteristics (e.g., species and growth stage), and N fertilizer application rates in agriculture [12].This complexity increases in no-tillage areas owing to different management schemes and cultivated species.Two major, simultaneous effects influence NO 3 − leaching in long-term no-tillage: the accumulation of organic matter over time and the increased number and continuity of macropores [13].
Brazilian no-tillage agriculture has developed substantially, reaching the equivalent of 61% of the total annual crop area in the country in 2017 [14,15].Especially in the southern, southeastern, and central-western regions, no-tillage is based on the sequential cultivation of soybeans in spring/summer and corn in autumn/winter [15].This management offers two commodities that are advantageous from a commercial perspective.In addition, corn provides a large amount of plant residues for soil cover, and soybean fixes atmospheric N. Regardless of these positive aspects, the continuous implementation of soybean-corn rotation subverts one of the fundamental principles of the no-tillage system, which is the diversification of plant species through crop rotation [16,17].Despite controversies and only a few conclusive studies on the subject, long-term no-tillage with diversified crop rotation (cover crops and commercial species) is considered a soil management system with lower NO 3 − loss through leaching than conventional management [18][19][20][21].However, studies on NO 3 − leaching in long-term no-tillage systems that compare diversified and simplified cropping rotations under subtropical agricultural conditions are scarce.Unlike other countries, Brazil, one of the major players in world agriculture [22], has conducted relatively few studies on NO 3 − leaching [23].The objective of this study was to investigate NO 3 − leaching in no-tillage systems with simplified (soybean-corn) and diversified (corn, soybean, wheat, and cover crops) crop rotations, with and without N fertilization.The specific aims of this study were: (1) to determine the effect of two crop rotations (simplified and diversified) on NO 3 − leaching, (2) to determine the effect of N fertilization on NO 3 − leaching, and (3) to investigate the possible interaction of crop rotation and N fertilization on NO 3 − leaching in a long-term no-tillage system.The study was based on the hypothesis that crop rotation and N fertilization influence NO 3 − leaching.It was expected that simplified crop rotation would lead to less leached nitrate, whereas under nitrogen fertilization, there would be greater leaching.

Experimental Area
The experiment was conducted between 2018 and 2020 at the Experimental Farm of the Institute of Rural Development of Paraná (Instituto de Desenvolvimento Rural do Paraná-IDR-Paraná), Londrina, Paraná, Brazil, located at a latitude of −23.37 • , a longitude of −51.17 • , and an altitude of 585 m.According to the Köppen-Geiger classification, the climate of Londrina is Cfa (humid subtropical with hot summers) [24].The city is located on the third plateau of Paraná and has soils formed from basic eruptive rocks [25].The soil is classified as oxisol [26,27], with an extremely clayey texture.
Soil chemical analyses were performed on composite samples (one sample per plot) collected at 0-20 cm depth before planting each crop.The following soil fertility characteristics were determined using the respective methods: active acidity (pH) using CaCl 2 (0.01 mol•L −1 ); carbon (C) using Walkley-Black; phosphorus (P) and potassium (K) using Mehlich I; calcium (Ca), magnesium (Mg), and aluminum (Al) using KCl M as extracting solution; potential acidity as H + Al; sum of bases (SB) = Ca + Mg + K; cation exchange capacity at pH 7.0 (CEC); and base saturation (V) = 100 × SB/CEC.The soil fertility analysis for each crop season is presented in Table 1.The average slope of the experimental area was 6%, which corresponds to the gently undulating relief [26].Surface runoff was not observed during the study because of the low slope, permanent vegetation cover, and soil type, which favor water infiltration.
The area has been cultivated in a no-tillage system with diversified crop rotation for 14 years.One year before the experiment, an intercrop of cover species (black oats + palisade grass + field peas) was established throughout the experimental area, which was chemically (glyphosate) and mechanically (knife roller) managed during spring.Approximately a month after soil management, soybean sowing was performed in the entire area.The evaluations were conducted after the soybean harvest in the fall of 2018, which concluded in the fall of 2020.

Experimental Design and Management
The experimental area was 4500 m 2 (150 m × 30 m), with 16 plots of 125 m 2 each (5 m × 25 m), and the remaining area was used as a border.The experiment was conducted using a factorial scheme with two factors and four treatments (combinations of factor levels) in a randomized complete block design with four replicates.These factors were the rotation system (simplified, with alternating corn and soybeans in autumn/winter and spring/summer, respectively; diversified, with soybean, corn, and different cover crop species in spring/summer and autumn/winter) and N fertilization management (with and without mineral fertilization).Factor combinations included the following treatments: simplified rotation with N fertilization, simplified rotation without N fertilization, diversified rotation with N fertilization, and diversified rotation without N fertilization.

332
The cultivation schemes for the production cycles of each planted species are presented in Table 2. 1 Black oats, field peas, and field radishes. 2Corn with palisade grass between rows. 3 Sowing was performed on 9/12/2018, but a hailstorm on 10/4/2018 destroyed the seedlings.Sowing was repeated on 10/29/2018. 4Plots corresponding to the management factor without N fertilization did not receive any source of N; each value corresponded to part of the total fertilizer applied in the fertilized plots, and the sum for each species was the total N application rate. 5Dates of N fertilization performed at sowing, first, and second coverage, successively.
Urea [CO(NH 2 ) 2 ] was used as a fertilizer in the N-fertilized plots at a rate established according to the cultivated species, the expected yield for a given crop, and, in the case of corn, whether it was the first or second crop season.Therefore, the N application rate varied among different species, harvests, and installment plans.Although N fertilization is not recommended for soybeans in Brazil, some farmers use it during sowing.Thus, the rate used for the fertilized soybean plots in this study was the mean rate used by farmers in the region.Information on sowing, fertilization, and harvesting is presented in Table 2.
During sowing, the plots were fertilized with P and K according to the chemical analyses of the soil and recommendations for each species [29].Phytosanitary control was maintained throughout the experimental period according to the technical recommendations for each species.
The plots were managed with no-tillage, mechanized sowing, 50 cm row spacing for soybean and corn, and 17-19 cm spacing for the other species.The central 10 m 2 of each plot was manually harvested when the grains of the commercial species attained 13% moisture.To estimate the grain yield, we extrapolated the quantity produced from 10 m 2 to 1 ha.The cover crops were managed with herbicides upon flowering.After desiccation, we collected 1 m 2 of plant mass per plot to estimate the dry biomass yield of the cover crops using the same extrapolation method.Knife roller management was then performed to lodge the cover crop residues in the soil.
Salvage irrigation with a cannon-type sprinkler was performed for all treatments to prevent plant death during periods of low rainfall.Irrigation was provided for only five months in 2018: April (76 mm), May (35 mm), June (36 mm), July (21 mm), and December (127 mm).In the other months, no irrigation was provided due to equipment failure, even during water deficits, which was a limitation of the study.

Soil Solution Collection
The soil solution was sampled using a suction lysimeter.Six months before the start of the experiment, two lysimeters were installed per plot, 9 m equidistant from each other and 8 m from the plot borders, for a total of 32 lysimeters.The center of the porous capsule of each lysimeter was positioned at a 1 m depth and, therefore, below the root zone of the cultivated species.A manual piston pump was used to generate a vacuum of −60 kPa in the lysimeter to collect soil solution samples.The lysimeter construction and operation were adopted from Bajracharya and Homagain [30].
A vacuum was generated in the lysimeters at the end of each day after a rain event of at least 20 mm, which resulted in some or no collection, depending on the dry period and the volume of rain during the previous days.Soil solution samples were collected in duly-identified 100 mL plastic containers the next morning.Each sample consisted of soil solutions collected from the two lysimeters in each plot.Samples were placed in a thermal box with ice and stored in a refrigerator (5 • C); NO 3 − concentrations could be determined either on the same day or up to a week later when stored in the freezer (−14 • C).
Soil solution samples were collected during the 2018/2019 and 2019/2020 evaluation seasons.Each evaluation season was a two-crop season: autumn/winter 2018 and spring/summer 2018/2019 crops in 2018/2019, with a fallow period between them; and autumn/winter 2019 and spring/summer 2019/2020 crops in 2019/2020 (Table 3).The results are presented considering the mean NO 3 − concentrations between two successive collection dates.This time interval was designated as the period, totaling 24 periods throughout the experiment.In the 2018/19 evaluation season, 13 soil solution samples were collected between 3/22/2018 and 1/10/2019, which correspond to periods 1-12, and in the 2019/20 evaluation season, 13 different soil solution samples were collected between 5/9/2019 and 1/15/2020, corresponding to periods 13-24 (Table 3).Although the experiment ended in March 2020, after the soybean harvest, the last soil solution extraction occurred in January 2020 because of low rainfall following that month (Figure 1a).Collections were also performed during the fallow period between August and October 2018, but not during buckwheat cultivation, nor at the beginning of the off-season corn cultivation in autumn/winter 2019 (March-April 2019) because of low soil moisture.

Meteorological Data
Monthly accumulated rainfall (Figure 1a) and average temperature (Figure 1b) values were calculated from daily data provided by the Automatic Meteorological Station of the Technology and Environmental Monitoring System of Paraná (Sistema de Tecnologia e Monitoramento Ambiental do Paraná), located in the IDR-PARANÁ experimental station.The same data were used to calculate the accumulated rainfall and average temperature for each of the 24 periods.Historical accumulated precipitation and average temperature values for 1976-2016 (Figure 1) were obtained from records of the IDR-Paraná agrometeorology sector of Londrina [31].The same source provided data to calculate actual evapotranspiration (ETR) (Figure 1a) using the Soil-Water-Atmosphere-Plant (SWAP) model [32].The ETR was calculated until 12 March 2020, when the final experimental sample was collected.

Drainage
The volume of drained water was estimated to calculate the NO 3 − load lost below the root zone.The SWAP model [32] was used for this estimation at a depth of 1 m.The SWAP model uses Richard's equation and includes root water extraction to simulate the movement of water, solutes, and heat in the unsaturated soil zone during plant growth stages.In this model, the transport processes are predominantly vertical (one-dimensional), and the vertical domain varies from a plane just above the canopy to a plane in shallow groundwater [33].The model considers environmental (e.g., maximum and minimum temperatures, solar radiation, relative humidity, and wind speed) and plant information (e.g., sowing and harvesting date, beginning of the vegetative and reproductive phases, effective depth of the root system, and leaf area).For calculations, this study considered only data related to the species used in simplified rotation (corn or soybean), and drainage was kept constant among all treatments, owing to difficulties in measuring the leaf area of plants cultivated in intercropped or mixed schemes.Hence, the drainage estimates were the same for the two rotation systems analyzed, which is a limitation of our study.The leaf area of soybean was calculated according to Richter et al. [34], and that of corn was calculated according to Guimarães et al. [35].More information about the model can be obtained from SWAP [32] and studies by Kroes et al. [33] and Pinto [36].

NO 3 − Concentration and Load
The NO 3 − concentration in the soil solution was analyzed in duplicate using the spectrophotometric method in the ultraviolet region at 220 and 275 nm, as described in the Standard Methods for the Examination of Water and Wastewater [37].
The leachate NO 3 − , expressed in kg ha −1 of NO 3 − , was also calculated for each period (Table 3).The load calculation considered the mean concentration and the sum of the volume of water drained below a 1 m depth between two collection dates using the following equation, according to the procedure adapted from Rosolem et al. [38]: where D (mm) is the sum of the drained water volumes between two consecutive soil solution collections; C (mg L −1 ) is the mean NO 3 − concentration between two consecutive collections; and 0.01 is the unit conversion factor.
The concentration results are expressed in mg L −1 of NO 3 − , not in mg L −1 of N-NO 3 − because the identified values were extremely low.To discuss the results, literature data were transformed from N-NO 3 − to NO 3 − by multiplying them by a conversion factor of 4.43.

Statistical Analysis
Statistical analyses were performed using Statistica ® software (version 14.0.015).Mean comparisons of nitrate leaching concentration and load for all 12 periods in each evaluation season and for each period separately in each evaluation season were performed using a two-way analysis of variance (ANOVA) using a randomized complete block design with two crop rotation levels and two N fertilization levels as independent variables and four blocks as a random factor.ANOVA was used to evaluate significant differences in crop yield between treatments and the interaction of crop rotation and N fertilization on NO 3 − leaching.Repeated measures ANOVA for means comparison between nitrate leaching concentration and load between periods in each evaluation season was performed with crop rotation (two levels), N fertilization (two levels), and period (12 levels) as independent variables and four blocks as a random factor.Tukey's test was used to compare the mean values of significant differences.The significance level was set at <5% probability.Before any statistical analyses, the Shapiro-Wilk test was used to check whether the residuals followed a normal distribution.
The SWAP model used for drainage estimation was calibrated from 2017 to 2019 using 13 samples of volumetric soil water content data (θ) at two depths (10 and 60 cm) and validated using new data.Leaf area indices for maize and soybeans were calculated during the two evaluation seasons [34,35].The SWAP crop file parameters were acquired from the online WOFOST dataset [39].The model effectively predicted soil water content in the validation phase with the following performing statistical indices: root mean square error (RMSE) of 0.017 cm 3 cm −3 and 0.014 cm 3 cm −3 for 10 cm and 60 cm depths, respectively; index of agreement (d) [40] of 0.86 for both depths; and the Nash-Sutcliffe efficiency (NSE) model [41] of 0.72 and 0.67 for 10 cm and 60 cm depths, respectively.The cited statistical metrics and their use in field-scale crop models, validation, and the calibration of soil water dynamics can be found in Groenendijk et al. [42].

Results
Table 4 presents the grain yield for commercial species and the dry mass for cover crops across crop seasons.Only the autumn/winter 2019 simplified rotation with corn showed a significant difference in yield between treatments with and without N fertilization (Table 4).In this case, the unfertilized plots produced almost 600 kg ha −1 more harvest than the fertilized plots, but both had lower yields than the national average of just over 5600 kg ha −1 .All plots presented lower yields owing to severe water deficits, which occurred from germination to harvest (periods 13-19; Figure 2).Analyzing the rainfall and irrigation data, ETR, accumulated drainage, and average atmospheric temperature for each of the 24 periods (Figure 2) indicated period 3 as having the highest accumulated rainfall and irrigation, as well as ETR, despite the mild average atmospheric temperature (19 °C).Period 3 spanned from 20 April to 28 August 2018, covering most of the irrigation events during the experiment, which explains the higher ETR.Analyzing the rainfall and irrigation data, ETR, accumulated drainage, and average atmospheric temperature for each of the 24 periods (Figure 2) indicated period 3 as having the highest accumulated rainfall and irrigation, as well as ETR, despite the mild average atmospheric temperature (19 • C).Period 3 spanned from 20 April to 28 August 2018, covering most of the irrigation events during the experiment, which explains the higher ETR.
Figure 3 shows the NO 3 − concentrations and load results for the entire experimental period.No significant differences were observed between the factors or treatments in terms of NO 3 − concentration and load during the two evaluation seasons.In the 2018/2019 evaluation season, the NO 3 − concentration among all treatments ranged between 5.2 and 7.2 mg L −1 for the simplified rotation, without and with N fertilizer, with a mean of 6.5 mg L −1 .In 2019/2020, the mean concentration was lower, 2.6 mg L −1 of NO 3 − , ranging from 2.0 to 3.1 mg L −1 for diversified rotation, with and without N fertilizer.The leachate NO 3 − load followed a concentration trend with means of 51 (range: 40-57 kg ha −1 ) and 8 kg ha −1 (range: 6-9 kg ha −1 ) for the 2018/2019 and 2019/2020 evaluation seasons, respectively.evaluation season, the NO3 − concentration among all treatments ranged between 5.2 and 7.2 mg L −1 for the simplified rotation, without and with N fertilizer, with a mean of 6.5 mg L −1 .In 2019/2020, the mean concentration was lower, 2.6 mg L −1 of NO3 − , ranging from 2.0 to 3.1 mg L −1 for diversified rotation, with and without N fertilizer.The leachate NO3 − load followed a concentration trend with means of 51 (range: 40-57 kg ha −1 ) and 8 kg ha −1 (range: 6-9 kg ha −1 ) for the 2018/2019 and 2019/2020 evaluation seasons, respectively.Comparisons of the mean NO 3 − concentration and load in the leached soil solution of all treatments over the 24 periods are presented in Figure 4.In 2018/2019, the highest NO 3 − concentrations occurred in period 2 (March-April 2018), followed by periods 1 (March 2018) and 3 (April-August 2018), with similar load trends.Both the NO 3 − concentration and load were extremely low during the 2019/2020 evaluation season.
Tables 5 and 6 present the mean leachate NO 3 − concentrations and loads for each period, for each level of the two factors, and their interactions, respectively.The mean leachate NO 3 − concentration and load indicated that plots with N fertilization had greater NO 3 − leaching during periods 4, 5, and 6, whereas the opposite occurred during periods 10, 11, 20, and 21 (Tables 5 and 6).These results indicated a tendency for greater NO 3 − leaching in plots that received previous fertilization after periods of high rainfall/irrigation (period 3, Figure 2) and under fallow conditions (periods 4, 5, and 6, Table 3).Under low rainfall conditions (periods 10, 11, 20, and 21, Figure 2), the plots without fertilization leached more than the fertilized plots.
A significant difference between cropping systems was only observed in period 12 (Tables 5 and 6), when leachate NO 3 − concentrations and loads were higher in plots with simplified soybean rotation.
A significant leaching difference was present among interaction factors in treatments only during the four periods (4, 10, 17, and 19) that shared a common characteristic of grass crops in lower NO 3 − leaching areas (Tables 5 and 6).Although period 4 corresponded to fallow, corn was previously grown in simplified rotation plots, and covered crops were grown in diversified rotation (Table 3).In period 4, the simplified rotation without N fertilizer showed the lowest leaching, and the simplified rotation with fertilizer showed the highest leaching (Tables 5 and 6).The diversified rotation treatment with N fertilization resulted in less leaching during periods 10, 17, and 19 (Tables 5 and 6).During period 10, the cultivated species in this treatment were corn intercropped with palisade grass and wheat during periods 17 and 19 (Table 3).Tables 5 and 6 present the mean leachate NO3 − concentrations and loads for each period, for each level of the two factors, and their interactions, respectively.The mean leachate NO3 − concentration and load indicated that plots with N fertilization had greater NO3 − leaching during periods 4, 5, and 6, whereas the opposite occurred during periods 10, 11, 20, and 21 (Tables 5 and 6).These results indicated a tendency for greater NO3 − leaching in plots that received previous fertilization after periods of high rainfall/irrigation (period 3, Figure 2) and under fallow conditions (periods 4, 5, and 6, Table 3).Under low rainfall conditions (periods 10, 11, 20, and 21, Figure 2), the plots without fertilization leached more than the fertilized plots.The values represent the means of four replicates, followed by standard errors in parentheses.Different vertical lowercase letters indicate significant differences between N fertilization management factors (0 and N) or between cropping system factors (S and R), Tukey's test, significance level p < 0.05.Vertical capital letters indicate a significant difference between interacting N fertilization and cropping system factors (S0, R0, SN, and RN; Tukey's test, significance level p < 0.05); ns indicates non-significant.The values represent the means of four replicates, followed by standard errors in parentheses.Different vertical lowercase letters indicate significant differences between N fertilization management factors (0 and N) or between cropping system factors (S and R), Tukey's test, significance level p < 0.05.Vertical capital letters indicate a significant difference between interacting N fertilization and cropping system factors (S0, R0, SN, and RN; Tukey's test, significance level p < 0.05); ns indicates non-significant.

Discussion
The corn, wheat, and soybean grain yields (Table 4) were consistent with the national average [43][44][45].Only the yield of intercropped corn and palisade grass in the spring/summer rotation of 2018/2019 (periods 8 to 12) was higher than the Brazilian average (6250 kg ha −1 ), even with a water deficit from the second month after sowing (Figures 1 and 2).For these crops, the corn yield was approximately 1000-1700 kg ha −1 above the national average yield for non-fertilized and fertilized plants, which did not differ significantly from each other.
In autumn/winter 2019, unfertilized plots in simplified rotation with corn produced significantly more than the fertilized plots (Table 4).However, both were below the national average, probably due to a water deficit in critical phases (Figure 1).Water deficit during the flowering stage leads to negative effects on phenological and yield traits, as it damages several corn production characteristics, such as the number and weight of grains [46], which was aggravated by the presence of fertilizer.However, this finding does not corroborate the results of some studies that specifically used N fertilizer to increase tolerance to water stress [47,48] but concurs with the results of Mansouri-Far et al. [49], who showed that a higher N dose under water deficit conditions imposed at the reproductive stage did not improve yield.Furthermore, the second corn crop was raised when the average air temperature was above the local historical average for all months (Figures 1 and 2).The combination of heat and water stress increases the negative effects of these factors on the yield compared to the results of each factor separately [50].
Period 3 had the highest rainfall and irrigation sum, decreasing soon after mainly as a result of reduced rainfall (Figure 2).In non-irrigated areas, NO 3 − leaching from layers below the root zone depends directly on rainfall because its formation via nitrification depends on soil moisture, in addition to ion transportation to deeper layers [51,52].Additionally, rainfall determines the NO 3 − concentration and load in the soil solution [53].The experimental period (January 2018 to March 2020) presented an atypical volume of accumulated monthly rainfall.Although rainfall in some months was higher than expected, the amount of rain in most months was below the historical average from 1976 to 2016 (Figure 1a).The annual historical average of accumulated rainfall was 1635 mm, whereas the expected accumulated rainfall for the first three months was 548 mm.The accumulated rainfall deficits were 117 and 398 mm in 2018 and 2019, respectively.In the first three months of 2020, this accumulated deficit was 272 mm, resulting in a total of 787 mm over the experiment (Figure 1a).
Atmospheric temperature also plays an important role in the nitrification process, and its optimal levels vary according to the type of environment [51].High temperatures are common in the tropics but rarely limit the nitrification process [12].The average atmospheric temperature in most months of the experimental period was above the historical average; it was approximately 1.8 • C higher in 2019 (Figure 1b).The temperature rise could have increased nitrification and, consequently, the concentration and load of NO 3 − in the leachate, which did not occur, however, due to the rainfall deficit (Figure 1a) [53][54][55].
Drainage was higher in period 8 owing to several rainfall events that accumulated to 176 mm.In addition, other factors contributed to the increased drainage during this period, such as low ETR and mild temperatures (Figure 2).The ETR directly affects the volume of water drained; thus, a higher value indicates lower availability for drainage.The drainage values were much lower than those found by Rosolem et al. [38], who obtained drainage sums between 533 and 595 mm per crop season in soil with a similar texture in Brazil.Drainage depends not only on the amount of rain but also on other factors, such as the type of soil [56]; however, we could not observe oxisol behavior similar to that of sandy soils in terms of drainage [12], probably due to the abnormal rainfall pattern and lack of irrigation (Figures 1 and 2).
The NO 3 − concentrations and load values were low during the experimental period (Figure 3) because of the water deficit resulting from scarce rainfall (Figures 1 and 2).This was also observed by Ghiberto et al. [57] in an oxisol area with conventional sugarcane planting in Brazil.They reported low N loss via leaching, with more than half in the form of NH 4 + and only 37% in the form of NO 3 − (2 kg ha −1 ).The greater amount of NH 4 + in that study indicated that the lack of soil moisture was responsible for the lower NO 3 − formation via nitrification.The authors attributed the low amount of leached N to a water deficit because the precipitation was 141 mm lower than the normal amount.In contrast, several studies have reported much higher NO 3 − concentrations and loads than those reported in this study.In a field study in Brazil, Rosolem et al. [38] observed leaching losses with NO 3 − concentrations in the range of 89-111 mg L −1 and NO 3 − loads of 68-144 kg ha −1 in no-tillage soybeans with cover crop rotation in clayey soil.Bortolotto et al. [58] found annual NO 3 − losses of 65 and 463 kg ha −1 for N rates of 400 and 800 kg ha −1 applied as fertilizer, respectively, in sandy soil cultivated with fertigated coffee trees in Brazil.Hussain et al. [10] found an NO 3 − concentration of 42 mg L −1 and leaching losses of up to 156 kg ha −1 of NO 3 − in a field experiment with corn in the US.Considering all treatments, the initial periods presented higher values of leachate NO 3 − concentration and load (Figure 4a,c), reflecting the presence of N compounds in the soil and the previous water regime.In period 2, for example, all N fertilization had already been performed in plots receiving fertilizer, which influenced the mean concentration and load.A study in China conducted by Xin-Qiang et al. [54] achieved a similar result, which showed high NO 3 − leaching in the presence of N fertilizer, with the load increasing at an increasing rate.In addition, a meta-analysis by Wang et al. [59] concluded that soil NO 3 − leaching emission factors increased with increasing N application rates.
Regarding the water regime, the rainfall deficit in February was compensated for in January, which had 176 mm more rain than the normal amount (Figures 1a and 2).Thus, in period 2, the water available in the soil was sufficient to promote nitrification, and excess NO 3 − leaching occurred (Figure 4a,c).Despite the excessive rainfall in August and October 2018, the rainfall deficit in the previous months (Figure 1a) did not allow the same process to be repeated in the evaluations from period 4, which showed even lower concentrations and loads (Figure 4).Another reason for the low NO 3 − concentrations and load values is that the no-tillage system provided less NO 3 − leaching than conventional management [60].This, associated with the lack of rainfall, likely caused the low NO 3 − concentrations and loads observed from period 4 onwards (Figure 4).
Tables 5 and 6 results suggest a propensity for increased leaching of NO 3 − in plots that had received prior fertilization following periods of heavy rainfall/irrigation (period 3, Figure 2) and during fallow conditions (periods 4, 5, and 6, Table 3).This direct relationship between N fertilization and increased NO 3 − leaching was also observed by Xin-Qiang et al. [54] in wheat crops and by Pasley et al. [61] in a modeling study of corn and soybeans.This finding also corroborates that of Meisinger et al. [62], who stated that the intensity of NO 3 − loss via leaching depends on the interaction of factors such as rainfall amount, timing, and N fertilization.Xiao et al. [63] similarly observed peak nitrate concentrations in surface waters (rivers, ponds, and reservoirs) in China, coinciding with the crop planting season.Some studies found that fallow is a factor that can increase NO 3 − leaching [64][65][66].It occurs because the NO 3 − not absorbed by cultivated plants is available for leaching, as well as that produced by the degradation of crop residues [66].
During periods with lower rainfall (10,11,20, and 21, Figure 2), the unfertilized plots exhibited significantly higher NO 3 − leaching than those fertilized (Tables 5 and 6).NO 3 − leaching in no-tillage areas without fertilization has already been reported by Rosolem et al. [38], who evaluated the phenomenon in these situations over a year without rainfall restrictions and found NO 3 − concentrations of up to 111 mg L −1 .However, in the present study, NO 3 − leaching was significantly higher in plots without fertilization (periods 20 and 21, Tables 5 and 6) with soybean crops in both the diversified and simplified rotation areas.This observation was unpredictable because adding N via fertilization to clayey soil could negatively affect soybean root growth, theoretically resulting in a higher proportion of available N for leaching [67], which was not the case in this study.
The significant difference between cropping systems in period 12 (Tables 5 and 6) indicates that a punctual event may have momentarily favored leaching in simplified rotation, such as the combination of irrigation events in previous periods (Figure 2) and the characteristics of the cultivated species itself, in this case, soybean.Legumes, such as soybeans, release large amounts of N compounds through rhizodeposition [68], which can be transformed into NO 3 − and thereby become susceptible to leaching.Similar results were reported by Rosolem et al. [38], who compared NO 3 − leaching in grasses and legumes; however, the authors attributed this result to the fact that most of the N absorbed by grasses originates from the soil, resulting in less NO 3 − available for leaching, whereas legumes obtain most of their N through biological fixation, which barely changes the soil solution NO 3 − concentrations.Significantly distinct NO 3 − leaching was observed among the interaction factors during periods 4, 10, 17, and 19 (Tables 5 and 6), which had in common the cultivation of grasses (Poaceae).The Poaceae family is one of the most efficient in reducing NO 3 − leaching [66].In period 4, for instance, the diversified rotation treatments consisted of a mixture of species from the Poaceae and Fabaceae families.Fabacea species (legumes) fix N from the atmosphere, which can be leached upon conversion to NO 3 − , whereas Poacea species reduce NO 3 − leaching directly.Thapa et al. [69] reported that non-legume cover crops could reduce leaching by up to 56%.
Regardless of the presence of fertilizer, treatments with diversified rotation (previously with cover crops) did not differ significantly from treatments with simplified rotation (previously with corn crops) (Tables 5 and 6).Therefore, cover crops had no influence on leaching, which is inconsistent with the meta-analyses by Nouri et al. [66] and Meyer et al. [70].The small influence of cover crops on NO 3 − leaching was due to the type of soil and management, since in sandier soils and conventional planting, they are more effective in reducing leaching.Furthermore, the use of legumes as cover crops does not reduce leaching.
The results from periods 10, 17, and 19, with less NO 3 − leaching in diversified rotation with N fertilization (Tables 5 and 6), corroborate those of Rosolem et al. [38], Pasley et al. [61], Thapa et al. [69], Zhou et al. [71], and Zhou and Butterbach-Bahl [72].These studies also showed that grass species tend to decrease N leaching because of their robust root systems and high N demand for growth.Corn has a more vigorous root system than other species, with higher biomass than soybean [73], providing higher N uptake.This characteristic of the corn root system resulted in a lower amount of available N for leaching into the soil solution, explaining the lower leaching in the plots with this crop.Comparing leaching in wheat and corn, Daryanto et al. [13] observed a higher concentration of NO 3 − in fields cultivated with wheat than with corn, despite higher corn fertilization.They attributed the result to a gap between fertilizer application and the wheat growing period.However, wheat can outperform corn NO 3 − efficiency because its roots can go deeper into the soil profile and may even recover NO 3 − leached in an area previously cultivated with corn [71].The size of the wheat root system, which varies according to genotype, affects the NO 3 − leaching potential.Genotypes that produce more roots absorb more N, resulting in less available NO 3 − for leaching [74].Thus, grasses can potentially support NO 3 − leaching reduction in no-tillage systems, at least in the context of the low rainfall in the present study.
While these findings may not directly apply to periods characterized by normal rainfall patterns, they hold significance within the framework of climate change.Projections for the upcoming years indicate an increase in extreme weather events, including enhanced occurrences of intense rainfall and prolonged droughts [75].Several studies link escalated NO 3 − leaching with periods of greater precipitation [76,77].Nevertheless, during drought phases, nitrification inhibition occurs, leading to N accumulation in the soil.The limited NO 3 − leaching, even in the fertilized plots of this study, suggests its accumulation.The N accumulated can be released with rainfall, as a result of the nitrification pulse that generates NO 3 − , which is susceptible to leaching during heavy rainfall [78].These extreme climate events are already occurring, as evidenced even by the climate data in our study (Figure 1), with consequences for NO 3 − leaching, as proved by our results.These findings highlight the need for a reconsideration of N fertilization guidelines to mitigate the potential for NO 3 − leaching in the predicted climate change scenarios.

Conclusions
This study did not confirm the working hypotheses, as the statistical analysis of nitrate leaching did not show marked differences between the treatments and their interactions.No significant differences existed between the treatments during most periods, but some grass treatments (corn, corn with palisade grass, and wheat) resulted in lower leaching rates.However, owing to atypical rainfall patterns, further studies are required to assess NO 3 − leaching in plots with corn or wheat vs. other cultivated species, primarily under normal or excess rainfall conditions.This study holds substantial value in guiding future research and expanding the understanding of the overall process of nitrate leaching in no-tillage, whereas weather extremes such as low rainfall conditions will become more frequent in the near future.

Note:
Values represent the means of four replicates; 1 0 = plots without N fertilization; N = plots with N fertilization; 2 Black oats, field peas, and field radish.3Corn intercropped with palisade grass showing corn yield.---=no treatment was applied.Tukey's test was used with a significance level of p < 0.05.represent the means of four replicates; 1 0 = plots without N fertilization; N = plots with N fertilization; 2 Black oats, field peas, and field radish.3Corn intercropped with palisade grass showing corn yield.---= no treatment was applied.Tukey's test was used with a significance level of p < 0.05.

Figure 3 .
Figure 3. Means of twelve periods of nitrate (NO3 − ) concentrations (mg L −1 ) and loads (kg ha −1 ) in leached soil solutions between cropping systems and nitrogen fertilization managements and their interactions in the 2018/2019 (a,c) and 2019/2020 (b,d) evaluation seasons (Londrina, PR, Brazil).Note: NR: diversified crop rotation with N fertilization; NS: simplified crop rotation with N fertilization; OR: diversified crop rotation without N fertilization; OS: simplified crop rotation without N fertilization; R: diversified crop rotation; simplified crop rotation; N: management with N fertilization; O: management without N fertilization.Tukey's test was used with a significance level of p < 0.05.Bars represent standard errors.Comparisons of the mean NO3 − concentration and load in the leached soil solution of all treatments over the 24 periods are presented in Figure4.In 2018/2019, the highest NO3 − concentrations occurred in period 2 (March-April 2018), followed by periods 1 (March 2018) and 3 (April-August 2018), with similar load trends.Both the NO3 − concentration and load were extremely low during the 2019/2020 evaluation season.

Figure 3 .
Figure 3. Means of twelve periods of nitrate (NO 3 − ) concentrations (mg L −1 ) and loads (kg ha −1 ) in leached soil solutions between cropping systems and nitrogen fertilization managements and their interactions in the 2018/2019 (a,c) and 2019/2020 (b,d) evaluation seasons (Londrina, PR, Brazil).Note: NR: diversified crop rotation with N fertilization; NS: simplified crop rotation with N fertilization; OR: diversified crop rotation without N fertilization; OS: simplified crop rotation without N fertilization; R: diversified crop rotation; S: simplified crop rotation; N: management with N fertilization; O: management without N fertilization.Tukey's test was used with a significance level of p < 0.05.Bars represent standard errors.

Nitrogen 2024, 5 , 11 Figure 4 .
Figure 4. Means of sixteen replicates (four treatments and four blocks) of leachate NO3 − concentrations (mg L −1 ) (a,b) and loads (kg ha −1 ) (c,d) in each period of the 2018/2019 and 2019/2020 evaluation seasons (Londrina, PR, Brazil).Note: different letters indicate significant differences between periods during each evaluation season.Tukey's test was used with a significance level of p < 0.05.Bars represent standard errors.

Figure 4 .
Figure 4. Means of sixteen replicates (four treatments and four blocks) of leachate NO 3 − concentrations (mg L −1 ) (a,b) and loads (kg ha −1 ) (c,d) in each period of the 2018/2019 and 2019/2020 evaluation seasons (Londrina, PR, Brazil).Note: different letters indicate significant differences between periods during each evaluation season.Tukey's test was used with a significance level of p < 0.05.Bars represent standard errors. )

Table 1 .
Chemical characteristics of the soil at a 0-20 cm depth before the start of the experiment.

Table 3 .
Evaluation seasons, crop seasons, type of rotation and species, periods, and intervals between the collection of soil solution samples leached below the root system.

Table 4 .
Grain yield for commercial species or dry mass for cover crops (kg ha −1 ).