Published online 9 August 2007
Published in Soil Sci Soc Am J 71:1460-1468 (2007)
DOI: 10.2136/sssaj2006.0039
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
SOIL PHYSICS
Estimating the Influence of Nitrogen Transformations on Nitrate Leaching in Soils
Li Rena,
Junhua Maa and
Renduo Zhangb,*
a Dep. of Soil and Water Sciences, China Agricultural Univ., Key Lab. of Plant–Soil Interactions, MOE, Beijing 100094, P.R. China
b State Key Lab. of Water Resources and Hydropower Engineering Science, Wuhan Univ., Wuhan 430072, School of Environmental Science and Engineering, Sun Yat-Sen Univ., Guangzhou 510275, P.R. China
* Corresponding author (zhangrd{at}mail.sysu.edu.cn).
 |
ABSTRACT
|
|---|
It is challenging to estimate N leaching in soils attributable to complicated physical, chemical, and biological processes. In this study, a transfer function model was developed to simulate the outflow concentration of NO3 in the field, considering the influence of transient water flow, input of applied N, initial residual N in the soil, and main N transformations on the NO3–N leaching process. The N transformations in the model included immobilization, mineralization, volatilization, and plant uptake. In the probability density function of NO3–N, a weighting factor was introduced to quantify the leaching contributions from applied and residual N. A field experiment was conducted for 196 d during the growing seasons of winter wheat (Triticum aestivum L.) and summer maize (Zea mays L.). Soil water potential and NO3 concentrations were measured during the study period along two soil profiles to a depth of 2 m: at 0.10-m intervals from the soil surface to the 1-m depth, and at 0.20-m intervals from 1 to 2 m. A comparison between the experimental data and simulated results with the transfer function showed that the model provided reasonable predictions of the N leaching process as well as the total amount leached at the 2-m depth. Results also indicated that by considering the transient water condition and N transformations, the transfer function significantly increased the estimation accuracy. Compared with the measured data, relative errors of the estimated total N leached were 1 and 20% with and without considering the transient water condition and the N transformations, respectively. The transfer function with the weighting factor can be useful to estimate the contributions from the applied and residual N to the leaching process in the field.
Abbreviations: pdf, probability density function • TFM, transfer function model
|
INTRODUCTION
|
|---|
Increasing applications of N fertilizers for agricultural productivity have caused water quality problems for surface water and groundwater resources (Burkart et al., 1999). Nitrogen, especially in the NO3– form, represents one of the top factors in water quality degradation and nonpoint source contamination to the environment (Gardi, 2001). Therefore, it is of importance to quantify N transport and transformation processes in field soils. Our understanding of the area is still limited, however, because N transport and transformations involve interactions of various physical, chemical, and biological processes. Furthermore, these processes are complicated by variations of input factors to soils, such as rainfall, applications of fertilizers, and irrigation. Thus, it is challenging to estimate the leaching process of NO3–N and various associated N transformations in the soil.
Since the 1980s, transfer function models (TFMs) have been used to simulate solute transport processes in soils. Transfer function models have been used to study tracer chemicals (White, 1985) and reactive chemicals (Heng et al., 1994; Heng and White, 1996; White et al., 1998), and to simulate N transport processes in soil columns (Dyson and White, 1987, 1989) as well as in field settings (White et al., 1986; Magesan et al., 1994). To represent the process of chemical transport in soils, TFMs use the probability density function (pdf) of travel time of chemical molecules (Jury and Scotter, 1994). Transfer functions are applied to model complex systems in a simple way by characterizing the output flux as a function of the input flux. Jury and Roth (1990, p. 105–111) and Jury et al. (1990) developed TFMs for field-scale solute transport under transient water flow. In their derivation, it was assumed that the pdfs for steady-state and transient water flow have the same form with respect to cumulative water drainage. The pdf under transient flow was obtained based on that under steady-state flow with a correction for water content differences between the two states. Nevertheless, the transient TFM has not been used successfully to simulate N transport in dryland farming systems under natural conditions of rainfall and irrigation (White et al., 1998; Ren et al., 2003). White (1987) applied a TFM to predict NO3–N leaching in shallow well-drained fields, in which the studied time period was relatively short and NO3–N was treated as a conservative solute. Ren et al. (2003) developed a TFM to characterize NO3–N transport in deep soils with the main N transformations (mineralization and immobilization) and plant uptake; however, the TFM did not consider transient water flow and NH3 volatilization, which can be commonly encountered in long-term field studies.
To enhance fertilizer-use efficiency and environmental risk assessment, we need to identify various N sources in the soil, such as applied fertilizers, residual (or indigenous or initial) N, and different N transformations in the soil. Mallawatantri and Mulla (1996) washed out residual N from the soil and then added a new NO3–N solution as input to obtain its pdf from the output. The procedure was tedious, however, and not able to study N transport processes from combined sources. Currently, the 15N dilution technique is usually used to distinguish different N sources (Lyngstad, 1990; Martinez and Guiraud, 1990; Baker and Timmons, 1994). Nonetheless, using the 15N dilution technique in field experiments is costly and time consuming, which limits applications of the technique (Fox et al., 1996). In addition, the 15N dilution technique is not always successfully applied in the field (Russelle et al., 2001).
In this study, a more general TFM was used to simulate NO3–N leaching processes and the simulated results were compared with experimental data collected in a field with crops of winter wheat and summer maize during a growing period of 196 d. The main objectives of this study were to: (i) develop a TFM that can be used to simulate NO3–N transport in soils under transient water flow and with N transformations of immobilization, mineralization, plant uptake, and NH3 volatilization, based on the work of Jury and Roth (1990, p. 105–111) and Ren et al. (2003); (ii) introduce a weighting factor in the TFM to estimate the leached amount of NO3–N resulting from different sources; and (iii) evaluate the effect of transient water flow and N transformations on NO3–N leaching in a dryland area during long-term simulations.
|
MATERIALS AND METHODS
|
|---|
Theoretical Background
The chemical flux concentration (g m–3) at a specific depth L can be expressed using the transfer function (Jury and Roth, 1990, p. 105–111)
 | [1] |
where I = Jwt is the cumulative water flux at the depth of interest (m), in which Jw is the steady-state water flux (m d–1); I' is an integral variable; and f f(L,I) is the pdf of a conservative solute (m–1). Data from field and soil column experiments have shown that the following lognormal distribution can be used to characterize the pdf sufficiently (Jury et al., 1986; Dyson and White, 1987; Jury and Roth, 1990, p. 36):
 | [2] |
where µ and 
2 are the mean and variance, respectively, of lnI, which reflects the intrinsic characteristic of the pdf. Under transient water flow, the TFM is obtained based on Eq. [1] with a correction for soil water content (Jury et al., 1990):
 | [3a] |
 | [3b] |
 | [3c] |
 | [3d] |
in which 
W(t) is the water storage difference at time t, and W(t) = 
0L[
(z,t) – (z,0)]dz; (z,0) and (z,t) are the initial soil water content (m3 m–3) and soil water content distribution at time t, respectively; I(t) is the cumulative discharge (m); Jw(L,t) is the transient outflow water flux (m d–1); t' is an earlier time than t; and z is soil depth along the soil profile (m). Equation [1] is rewritten as
 | [4] |
The outflow water flux can be estimated by
 | [5] |
where K() is the hydraulic conductivity (m d–1) at L, and Hi and Hi–1 are the hydraulic heads (m) at depths zi and zi–1 (m), respectively. With a pulse input (a delta function) with a concentration of C0 (g m–3), Eq. [4] is simplified in the form of (Jury and Roth, 1990, p. 105–111)
 | [6] |
based on the assumption that the initial concentration in the soil is equal to zero.
Equation [1] was derived for a conservative solute without residual concentration in the soil. Under field conditions, however, N goes through physical, biological, and chemical processes, and NO3–N concentration data at the exit may include contributions from the applied N, the residual N, and source–sink terms in the soil. Therefore, we introduce a weighting (lumping) factor 
into the pdf to quantify the contributions from the applied N, the residual N, and source–sink terms in the soil. Thus Eq. [6] is changed as follows:
 | [7] |
where fmf(L,I) is the calculated pdf using measured data of NO3–N at the exit (at depth L; Ren et al., 2003). Different from f f in Eq. [1], fmf in Eq. [7] is calculated from NO3–N concentration data at the depth of interest, including possible contributions from the applied N at the soil surface, the indigenous N in the soil profile, and some biological and chemical processes. The value of is close to 1 at the soil surface because fmf is approximately the pdf of a conservative solute and decreases with depth. The estimation of is discussed below.
For NO3–N leaching processes with contributions from applied N, residual N, and transformed N (source–sink terms) in the soil, similar to White (1989), we have
 | [8] |
where Pa is the cumulative probability function (cpf) of NO3–N leaching from applied fertilizers, Pi is the cpf of leaching from residual N; and Ps is the cpf of leaching related to source–sink terms. In general, the leaching losses from the applied and residual N are positive during the study period. The contributions of source–sink terms, however, can be positive or negative; for example, mineralization increases the potential of NO3–N leaching, while uptake decreases leaching. Therefore, the total effect of source–sink terms on leaching may be negligible during a long-term simulation. Compared with leaching of residual N (reflecting the richness of the soil inorganic N pool), the effect of source–sink terms on NO3–N leaching is small, which results in a small probability of NO3–N leaching from the source–sink terms. Therefore, the contribution to the outflow chemical flux from the residual chemical with a uniform concentration of Ci (g m–3) is approximated by
 | [9] |
in which
 | [10a] |
 | [10b] |
where W is the change in water storage (m).
We consider the N transformations in the TFM using the following source–sink terms (Ren et al., 2003):
 | [11] |
where Cm is the net mineralization of organic N (g m–3) (i.e., the combination of immobilization and mineralization), Cd is the denitrification (g m–3), Cv is NH3 volatilization (g m–3), and Cu represents plant uptake of N (g m–3). In the North China Plain, the concentration of NH4–N in the root zone was usually <4 mg kg–1 and high concentrations of NH4–N only appeared in a short period after fertilization (Ju et al., 2003), indicating that NH4–N was rapidly transformed to NO3–N in the dryland area. Since nitrification of NH4–N is a much faster process than mineralization of organic N, it is assumed that NO3–N is the direct product of mineralization and nitrification is ignored in the calculation of source–sink terms under the conditions of our experiments. The contribution to the outflow chemical flux from the source–sink terms is approximated by
 | [12] |
in which Pm is the cumulative probability function estimated using measured data of NO3–N leaching from the exit (Ren et al., 2003).
Combining Eq. [7], [9], and [12] based on the superposition principle (Jury and Roth, 1990, p. 18; White et al., 1998), we obtain a general equation of the outflow chemical flux at the exit:
 | [13] |
This transfer function model accounts for influences of the applied and residual N sources and the source–sink terms on NO3–N leaching in the soil under transient water flow conditions.
The leaching concentration of NO3–N can be calculated at any time using Eq. [13]. With the water flux calculated from Eq. [5], we can estimate the total NO3–N amount leached using the following equation:
 | [14] |
where Ni is the NO3–N leached amount (kg ha–1) within the time interval ti = ti – ti–1, Ci–1 and Ci are NO3–N concentrations in the leachate at ti–1 and ti, Jw(i–1) and Jw(i) are water flux at ti–1 and ti, respectively, and 
and 
are the average values within the time interval ti.
Field Experiment
To study N transport in the field, an experiment was conducted at the Quzhou Experimental Station of China Agricultural University, Quzhou County (36°35'43''–36°57'56'' N, 114°50'22.3''–115°13'27.4'' E), Heibei Province. The mean annual air temperature is 13.1°C. The mean annual precipitation is 556.2 mm, with about 70% of the total rainfall in June to September. The total evaporation is 3.3 times the precipitation. The agricultural soil is an Aquic Cambisol and the cropping system has been mainly rotations of winter wheat and summer maize. The climate and soil at the experimental site are typical of the North China Plain.
Two soil profiles (called Profiles A and B in the following analysis) separated by 50 m were dug to 2-m depth. Soil properties, including bulk density, soil texture, saturated water content, saturated hydraulic conductivity, organic matter content, and total N were measured at different depths using soil samples collected along the profiles. Soil water retention curves were developed for the soil samples by the pressure plate method (Klute, 1986), with the sand box apparatus for lower suctions and the pressure membrane-plate system for higher suctions (Soilmoisture Equipment Corp., Santa Barbara, CA). The saturated hydraulic conductivities were measured with the constant-head method in the laboratory (Klute and Dirksen, 1982). Four soil layers were identified in the profiles and the corresponding soil properties are presented in Table 1. After the sampling along the soil profiles, the profiles were covered with concrete to prevent the collapse of bulk soil and some holes at different depths were retained to set up tensiometers and suction cups. Along the soil profile, 15 tensiometers and 15 suction cups were set up at 0.10-m intervals from the soil surface to the 1-m depth, and 0.20-m intervals from the 1- to 2-m depth. During the experiment, soil water matric potential values were recorded with the tensiometers and used in later analyses to estimate soil water contents based on the soil water retention curves. The suction cups were used to collect soil solution samples. The concentration of NO3–N was measured, using the soil solution samples, with a continuous-flow analytical system (TRAACS 2000, Bran+Luebbe, Norderstedt, Germany). Concentrations of NH4–N in the soil solution were too low to be detected by the apparatus.
Winter wheat was planted on 10 Oct. 1998. Before sowing, 172.5 kg N ha–1 of urea and 400 kg ha–1 of triple calcium superphosphate were applied as basal fertilizers. During the experimental period, 58.7 and 27.6 kg N ha–1 of urea were applied on 17 Mar. and 4 Apr. 1999, respectively, as topdressing fertilizer for winter wheat. Summer maize was planted on 15 June 1999. A one-time topdressing of 186.3 kg N ha–1 was applied for the maize on 26 July 1999. The fertilizer N amount was the conventional application of N fertilizer in this region. Since the experiment was to study the main N transformation processes and NO3–N leaching under the conventional management practice of fertilizer N in this region, a zero-N treatment was not considered. The field observation started during the growing period of winter wheat on 17 Mar. 1999 and ended on the harvest day of summer maize on 29 Sept. 1999, a total of 196 d. Ammonia volatilization was measured using the Bowen meter (CSI Bowen Ratio System, Campbell Scientific, Logan, UT) based on the diffusion principle of gas concentration gradient. Net mineralization was measured by the N-balance method (van Eerdt and Fong, 1998; Oenema et al., 2003; Ju et al., 2006). In this method, inorganic N surplus in the soil was considered as the sum of input components (including fertilizer, N from seed, wet deposition, N from irrigation, etc.) and output components (including NH3 volatilization, leaching, removal by aboveground plant parts, etc.). Through measuring the emission flux of N2O in the field, denitrification was estimated. Rainfall and irrigation distributions during the study period were recorded (Fig. 1
). In the North China Plain, the application of fertilizer through flood irrigation is a common and extensive practice to provide nutrients for growing crops and the irrigation amounts were close to the conventional amount in this region. The time periods of fertilization and irrigation were much shorter than the modeling period.
|
RESULTS AND DISCUSSION
|
|---|
In the TFM modeling process, each fertilization was treated as a pulse input and three different C0 values in Eq. [13] were used according to the fertilization time. Each C0 value was estimated using a ratio of the corresponding fertilization amount to the total water amount from irrigation and precipitation within 7 d after the fertilizer application. Therefore, the pdf corresponding to each pulse was obtained from the corresponding portion of the calculated pdf, fmf. The leaching of residual N was treated as a step change input (Jury et al., 1990) as Ci in Eq. [13] and Ci was estimated by the average concentration along the profile. The response of source–sink terms (Cb in Eq. [13]) was regarded as a kind of step change input (Heng and White, 1996; Ren et al., 2003).
The parameters of µ and 2 in Eq. [2] could be determined by pdf data at a steady-state water flow condition. According to the standard deviation of soil matric potential (Fig. 2
), the soil water flux in Profile A showed less variation with time than that in Profile B. Therefore, we approximated the soil water flux in Profile A as a quasi-steady state. Using Eq. [2], with the measured outflow NO3–N concentration vs. cumulative water drainage at L = 2 m in Profile A, we estimated the best-fit parameters of µ = 3.513 and 2 = 0.249. Figures 3A
and 3B show water content distributions in Profiles A and B, respectively, during the study period. The water content in Profile B showed larger variations than that in Profile A. The water content below 1.6 m in the two profiles was close to the saturated water content, which provided available water for NO3–N leaching.
The initial NO3–N concentration along the profiles and the NO3–N concentration at the end of the experiment are presented in Fig. 4A
and 4B, respectively. At the end of the experiment, a large amount of NO3–N from the applied fertilizer resided in the 0- to 80-cm soil layer, while NO3–N below 100 cm was leached out significantly. For an Aquic Cambisol, one of the main types of agricultural soils in Hebei Province, the clay-fixed NH4–N was between 125 and 239 mg kg–1, about 17 to 30% of the total N; the organic N was between 327 and 947 mg kg–1, about 67 to 79% of total N; the exchangeable NH4–N was between 10 and 21 mg kg–1, about 1.55 to 2.69% of total N (Meng et al., 2004). A long-term (13-yr) field experiment in northern China showed that the application of chemical fertilizers alone had little effect on the content of fixed NH4–N in the soil (Han et al., 1998). Since the content of clay-fixed NH4–N was relatively stable in the soil, the effect of clay-fixed NH4–N on N transformation was not considered in this study. Furthermore, some fertilization studies in the North China Plain showed that: (i) nitrification was complete within 7 d after urea was applied to the soil and NO3–N was the dominant inorganic N in the soil; (ii) relatively high concentrations of NH4–N only appeared for a short period after fertilization in topsoil and NH4–N concentrations retained in the soil were only 1 to 3 mg kg–1; (iii) NH4–N concentration in the subsoil was <4 mg kg–1 (Ju et al., 2003). Therefore, NO3–N leaching was the main concern of N issues in the North China Plain.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 4. Nitrate-N concentrations initially (17 Mar. 1999) and at the end of the experiment (29 Sept. 1999) through (A) Profile A and (B) Profile B.
|
|
To account for N transformations, we estimated the source–sink terms in Eq. [11]. As justified by Ren et al. (2003), denitrification was not considered in this study. In fact, the denitrification experiment in the field showed that the denitrification amount was <0.2 and 0.5% of the applied N during the growing seasons of winter wheat and summer maize, respectively. We considered NH3 volatilization under different conditions. For the winter wheat, the two applications of topdressing fertilizer were during the cold season, which limited NH3 volatilization. In addition, an immediate irrigation after the fertilizer application minimized NH3 volatilization (Fenn and Miyamoto, 1981). For the maize, however, after the application of topdressing fertilizer on 26 July 1999, the amount of NH3 volatilization was significant (Fig. 5
). Therefore, we used the date of 26 July 1999 (setting it as t0) to divide the study period into two subperiods. When t < t0, the source–sink term was expressed by
 | [15] |
When t 
t0, the source–sink term was in the form of
 | [16] |
The terms in Eq. [15] and [16] were determined using the same method as Ren et al. (2003). The value of Cu was estimated by assuming that the rate of plant uptake of N is proportional to the evapotranspiration rate of the crops (Heng et al., 1994). The cumulative evapotranspiration was calculated from the cumulative potential evapotranspiration, multiplied by a coefficient relating to water stress in the form of (Jensen et al., 1971)
 | [17] |
in which Av = [(Wt – Wr)/(Wf – Wr)]100%, Wt is the water storage at the field capacity (m), Wr is the actual water storage in the root zone (m) at time t (d), and Wf is the water storage at the wilting point (m). To avoid underestimating the cumulative evapotranspiration in this area, we used half of the saturated water content as field capacity and half of the field capacity as the wilting point (Or and Wraith, 2002). The terms of Wt, Wf, and Wr were determined from the water content distribution along the 2-m soil profile. The term Cm could be a sink or source, depending on whether immobilization or mineralization was dominant.
The amount of net mineralization during the study period is shown in Fig. 6
. Root systems of winter wheat and summer maize are well developed in this region. In the North China Plain, roots of winter wheat can reach 1.2 m after the turning-green stage and 1.8 m after flowering (Zhang et al., 1994; Feng and Liu, 1998). A 15N-labeled fertilizer study showed that winter wheat took up N from the soil at depths of 100 to 150 cm (Wu et al., 2005). Summer maize could take up 15N-labeled NO3– injected at a depth of 110 cm in the soil and the uptake amount was about 7 to 12% of the total uptake (Zhang et al., 2004). Because of the deep root systems and significant drainage volume, net mineralization was estimated using soil water storage in the whole soil profile. In other words, Cm is equal to the net mineralization amount divided by the water storage in the 2-m soil profile. Total amounts of net mineralization of N in the field were 84.1 and 23.8 kg ha–1 for the winter wheat and summer maize, respectively, which accounted for 52 and 15% of the total amount of uptake by the winter wheat (162.40 kg ha–1) and summer maize (158.60 kg ha–1), respectively.
During the growing season of summer maize, the total amount of NH3 volatilization was 45.41 kg ha–1, accounting for 24% of the topdressing fertilizer. Therefore, it was essential to consider the source–sink terms in the estimation of NO3–N leached in the soil. Volatilization mainly occurred in the topsoil because of fertilization. Since soil water content in the topsoil was affected by that in the subsoil, according to the relationship between soil moisture and NH3 volatilization of Al-Kanani et al. (1991), the volatilization term Cv was calculated as the amount of NH3 volatilization divided by the soil water storage in the soil profile.
To apply the transfer function model, we need to estimate the weighting factor in Eq. [13]. Values of the weighting factor may change under different conditions. Nevertheless, as shown below, within a certain range of the weighting factor, modeling results were not very sensitive to change in this factor, which can be treated as a constant in this range. In addition, fmf in Eq. [13] reflects fluctuation of the N leaching process to some degree under field conditions because of the characteristic of the pdf; that is, the pdf presents the intrinsic characteristic of the transport process of the applied chemical. Setting the water content change equal to zero in Eq. [13] and using fitted parameters of µ = 3.513 and 2 = 0.249 and other source–sink terms in Profile A, we estimated the value using an optimization procedure. Since the range of is between 0 and 1, we applied a bisectional method to select values until the sum of squared errors between the measured and estimated cumulative outflow chemical flux at L = 2 m reached the minimum (Fig. 7
). In the range of 0.28 to 0.32 for , the results of estimated cumulative outflow chemical flux approached the optimal. Since reflects the influence of surface conditions, such as rainfall, irrigation, and fertilizer applications, on outflow chemical flux, which were similar for the two soil profiles, we used = 0.30 to estimate the NO3–N leaching process in Profile B.
View larger version (10K):
[in this window]
[in a new window]
|
Fig. 7. Weighting factor () for estimating leaching from different sources, determined with the bisectional method.
|
|
Because of the same fertilization and irrigation conditions of Profiles A and B, we used the TFM (Eq. [13]) with the fitted parameters of µ and 2, the source–sink terms from Profile A, and a weighting factor = 0.30 to estimate the leaching process of NO3–N in Profile B. The exercise was also a way to verify the applicability of the TFM. We applied the TFM to estimate the leaching process and cumulative leached amount of NO3–N at L = 2 m with four methods: (i) considering both the transient water flow condition and the source–sink terms (Method 1); (ii) considering the transient water flow condition only (Method 2); (iii) considering neither the transient water flow condition nor the source–sink terms (Method 3); or (iv) considering the source–sink terms only (Method 4). Statistical analyses of the measured leaching process of NO3–N and estimated results using the different methods showed that Methods 1 and 4 provided better simulated leaching processes of N than Methods 2 and 3 (Table 2). The RMSEs between the measured leaching amount of N and simulated values using the four methods were 2.966, 5.260, 6.303, and 2.869 kg ha–1, respectively. Figure 8
presents the relative errors of the estimated leaching concentration of N vs. time. During the growing period of winter wheat, the four methods provided similar results. During the growing period of summer maize, however, NH3 volatilization became significant. The methods considering the source–sink terms (Methods 1 and 4) greatly improved the estimation accuracy. The maximum relative errors of Methods 2 and 3 reached 195 and 186%, respectively. The cumulative leaching processes of N estimated with the four methods are compared with the measured data in Fig. 9
. Again, the estimated results using Methods 1 and 4 agreed with the data well, whereas Methods 2 and 3 overestimated the process. Compared with the measured result of the total cumulative leaching amounts of N at the harvest time of summer maize (t = 196 d), the relative errors using Methods 1, 2, 3, and 4 were 1.18, 16.82, 19.71, and 3.21%, respectively. All the results above indicated that the TFM considering the transient water flow condition, applied and residual N sources, and source–sink terms provided more accurate estimations of N leaching processes. The results also indicated that considering the source–sink terms in the model had a greater influence than considering transient water flow conditions. It seems practical to consider the source–sink terms only and treat the water flow domain as a steady state when we simulate long-term chemical processes in dryland fields.
View this table:
[in this window]
[in a new window]
|
Table 2. Statistical analysis between different simulated values and measured leaching amounts of NO3–N at the depth of 2 m in Profile B.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 8. Relative error of estimated leaching concentration of NO3–N at the 2-m depth in Profile B, determined using four methods: Method 1, considering both the transient water flow condition and the source–sink terms; Method 2, considering the transient flow condition only; Method 3, considering neither the transient water flow condition nor the source–sink terms; and Method 4, considering the source–sink terms only.
|
|
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 9. Measured and estimated cumulative leaching amounts of NO3–N at the 2-m depth in Profile B, determined using four methods: Method 1, considering both the transient water flow condition and the source–sink terms; Method 2, considering the transient flow condition only; Method 3, considering neither the transient water flow condition nor the source–sink terms; and Method 4, considering the source–sink terms only.
|
|
We also calculated the leaching processes of NO3–N resulting from different sources. The leaching processes of NO3–N resulting from the applied N and residual N are shown in Fig. 10A
and 10B, respectively. The residual N included the initial N in the soil and N transformed from the source–sink terms. The peaks in Fig. 10A were mainly attributable to the irrigations, which resulted in relatively high water flux at the 2-m depth as well as high concentrations of NO3–N leached. The leaching process from the applied N started after about 60 d of the experiment, corresponding to the soil hydraulic properties and the conditions of applied water as well as fertilizers (Fig. 10A). At the earlier simulating period (see Fig. 10A from 0 to 60 d), little applied N leached out at the soil depth of 2 m because of slow unsaturated water flow and a high evapotranspiration rate. Therefore, during this stage, most of the leached amount was from the residual N in the soil (Fig. 10B). Because of a larger fertilization amount and water input, more applied N was leached during the summer maize growing stage. During the vegetative period of winter wheat, the crop consumed a large portion of the soil N, resulting in low leaching of N (see Fig. 10A from 90–120 d). The main contribution to leaching was from the residual N, with an initial high N concentration (22.47 g m–3 on average) and from mineralized N as well (Fig. 6). Because of strong mineralization in the soil during this period, NO3– tended to be leached easily with a high rate of irrigation. For example, the first leaching peak at 40 d corresponds with a 95-mm irrigation event (Fig. 10B).
View larger version (10K):
[in this window]
[in a new window]
|
Fig. 10. Estimated leaching concentration of NO3–N at the 2-m depth in Profile B attributable to (A) applied N and (B) residual N.
|
|
The leached NO3–N resulting from the applied N accounted for 0.6% of the total leached N and 0.15% of the applied N. Therefore, the leached NO3–N was mainly from the residual N, accounting for 99.4% of the total leached N. Calculations showed that the leached portions of NO3–N during the winter wheat and summer maize seasons were 0.01 and 0.22% of the applied N, respectively. During the winter wheat season, the amount of fertilizer from the two applications was not high. During the summer maize season, the large amount of topdressing fertilizer and the following high rates of rainfall and irrigation (total of 519 mm) resulted in higher NO3–N leaching rates. The above results are consistent with other studies. In a study of NO3–N leaching at a depth of 1.3 m in a lysimeter under different fertilizer rates in the North China Plain, difference methods showed that the amount of N leached was about 1 to 2% of the applied N for an N rate of 150 to 225 kg N ha–1 and the 15N dilution technique showed that labeled N was not leached in the spring and 1 to 2% of the applied N was leached in the summer (Yuan et al., 1995). Dowdell et al. (1984) used 15N-labeled fertilizer to measure the effect of applied N on NO3– leaching at a depth of 1.35 m in a lysimeter. Under various conditions with different soils and applied N rates, they found that the leached N resulting from the applied N accounted for 1.5 to 4.1% of the total leached N and 1.0 to 1.6% of the total applied amount. Lyngstad (1990) reported that the leached N resulting from applied N accounted for 0.5 to 1.3% of the applied N at a depth of 1 m in a sandy loam soil and <0.7% in a loam soil for different N fertilizer applications. Baker and Timmons (1994) found that the leached N at a depth of 1.37 m resulting from applied N accounted for 0.2 to 1.4% of the applied amount, depending on the method of application and fertilizer application rate.
The 15N dilution technique was used in the above studies and other similar research to identify the different sources of N leaching, which is costly and time consuming. Using the TFM with the weighting factor could be a useful solution to estimate the effect of different N sources on the leaching process. Although the selection of a proper value depends on a few factors, such as soil, vegetation cover, the amount of applied water and fertilizer, soil texture, and the depth of outflow, the applicable values were within an interval. For example, in our study area, a value of was selected from an interval from 0.28 to 0.32 and a constant of 0.30 gave reasonable simulation results.
For a further simplification, we also treated the three times of fertilization as one lumped pulse input in the TFM. We obtained almost identical results of the simulated cumulative leaching amount with the two treatments of fertilization inputs (i.e., treating the three times of fertilization separately as three pulse inputs as above and treating as one lumped pulse input). The results were expected because most of the leached N was from the indigenous N.
|
CONCLUSIONS
|
|---|
The TFMs of Jury and Roth (1990, p. 105–111) and Ren et al. (2003) were combined to develop a more general model, which can be used to simulate NO3–N transport in soils under transient water flow and with N transformations of immobilization, mineralization, plant uptake, and NH3 volatilization. A weighting factor was introduced in the model to quantify leached amounts of NO3–N resulting from different sources. In a field experiment, soil water and N concentrations were measured along two 2-m soil profiles under the condition of crop growth. The experiment lasted 196 d during the growing seasons of winter wheat and summer maize. Estimated results using the TFM were comparable with the experimental data of NO3–N leaching processes at the 2-m depth of soil profile. The TFM considering the transient water flow condition and the N transformations significantly improved the estimation accuracy. The results also showed that in long-term simulations of N leaching, considering N transformations improved model accuracy more significantly than considering the transient water flow condition. Practically, it may be sufficient to consider the source–sink terms only and treat the water flow domain as a steady state when we simulate long-term chemical processes in dry croplands. Within a certain range, a constant value of the weighting factor may be used to estimate the contributions between the applied N fertilizer and the residual N (including the transformed N from the source–sink terms) to the leaching processes. The estimated results were consistent with those of other studies measured using the 15N dilution technique and other methods.
|
ACKNOWLEDGMENTS
|
|---|
This research was partly supported by the National Science Foundation of China (Grant No. 50639040), the State Key Lab. of Water Resources and Hydropower Engineering Science, Wuhan University (Grant No. 2003B001), and by the Program for Changjiang Scholars and Innovative Research Team in the University (IRT0412). The field data were provided by the Department of Soil and Water Sciences and the Quzhou Experimental Station, China Agricultural University.
|
NOTES
|
|---|
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.
Received for publication January 25, 2006.
|
REFERENCES
|
|---|
- Al-Kanani, T., A.F. MacKenzie, and N.N. Barthakur. 1991. Soil water and ammonia volatilization relationships with surface-applied nitrogen fertilizer solutions. Soil Sci. Soc. Am. J. 55:1761–1766.[Abstract/Free Full Text]
- Baker, J.L., and D.R. Timmons. 1994. Fertilizer management effects on leaching of labeled nitrogen for no-till corn in field lysimeters. J. Environ. Qual. 23:305–310.[Abstract/Free Full Text]
- Burkart, M.R., D.W. Kolpin, and D.E. James. 1999. Assessing groundwater vulnerability to agrichemical contamination in the Midwest U.S. Water Sci. Technol. 39:103–112.
- Dowdell, R.J., C.P. Webster, D. Hill, and E.R. Mercer. 1984. A lysimeter study of the fate of fertilizer nitrogen in spring barley crops grown on shallow soil overlying chalk: Crop uptake and leaching losses. J. Soil Sci. 35:169–181.[CrossRef]
- Dyson, J.S., and R.E. White. 1987. A comparison of the convection–dispersion equation and transfer function model for predicting chloride leaching through an undisturbed structured clay soil. J. Soil Sci. 38:157–172.[CrossRef]
- Dyson, J.S., and R.E. White. 1989. A simple predictive approach to solute transport in layered soils. J. Soil Sci. 40:525–542.[CrossRef]
- Feng, G., and C. Liu. 1998. Analysis of root system growth in relation to soil water extraction pattern by winter wheat under water-limiting conditions. (In Chinese.) J. Nat. Resour. 13:234–241.
- Fenn, L.B., and S. Miyamoto. 1981. Ammonia loss and associated reactions of urea in calcareous soils. Soil Sci. Soc. Am. J. 45:537–540.[Abstract/Free Full Text]
- Fox, R.H., W.P. Piekielek, and K.E. Macneal. 1996. Estimating ammonia volatilization losses from urea fertilizers using a simplified micrometeorological sampler. Soil Sci. Soc. Am. J. 60:596–601.[Abstract/Free Full Text]
- Gardi, C. 2001. Land use, agronomic management and water quality in a small northern Italian watershed. Agric. Ecosyst. Environ. 87:1–12.[CrossRef]
- Han, X., P. Guo, E. Chen, D. Zhou, and C. Lou. 1998. Effect of long-term fertilization on content and availability of fixed ammonium in soils. (In Chinese.) Plant Nutr. Fert. Sci. 4:29–36.
- Heng, L.K., and R.E. White. 1996. A simple analytical transfer function approach to modeling the leaching of reactive solutes through field soil. Eur. J. Soil Sci. 47:33–42.[CrossRef]
- Heng, L.K., R.E. White, D.R. Scotter, and N.S. Bolan. 1994. A transfer function approach to modeling the leaching of solutes to subsurface drains: 1. Reactive solutes. Aust. J. Soil Res. 32:85–94.[CrossRef]
- Jensen, M.E., J.L. Wright, and B.J. Pratt. 1971. Estimating soil moisture depletion from climate, crop and soil data. Trans. ASAE 14:954–959.[Web of Science]
- Ju, X., C. Kou, F. Zhang, and P. Christie. 2006. Nitrogen balance and groundwater nitrate contamination: Comparison among three intensive cropping systems on the North China Plain. Environ. Pollut. 143:117–125.[CrossRef][Medline]
- Ju, X., X. Liu, and F. Zhang. 2003. Accumulation and movement of NO3–N in winter wheat–summer maize rotation system. (In Chinese.) Acta Pedol. Sin. 40:538–546.
- Jury, W.A., J.S. Dyson, and G.L. Butters. 1990. Transfer function model of field-scale solute transport under transient water flow. Soil Sci. Soc. Am. J. 54:327–332.[Abstract/Free Full Text]
- Jury, W.A., and K. Roth. 1990. Transfer functions and solute movement through soil: Theory and applications. Birkhäuser Verlag, Basel, Switzerland.
- Jury, W.A., and D.R. Scotter. 1994. A unified approach to stochastic–convective transport problems. Soil Sci. Soc. Am. J. 58:1327–1336.[Web of Science]
- Jury, W.A., G. Sposito, and R.E. White. 1986. A transfer function model of solute transport through soil: 1. Fundamental concepts. Water Resour. Res. 22:243–247.
- Klute, A. 1986. Water retention: Laboratory methods. p. 635–662. In A. Klute (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
- Klute, A., and C. Dirksen. 1982. Hydraulic conductivity and diffusivity: Laboratory methods. p. 687–734. In A. Klute (ed.) Methods of soil analysis. Part 1. 2nd ed. Agron. Monogr. 9. ASA and SSSA, Madison, WI.
- Lyngstad, I. 1990. A lysimeter study on the nitrogen balance in soil: I. Fate of 15N-labelled nitrate fertilizer applied to barley. Norw. J. Agric. Sci. 4:119–128.
- Magesan, G.N., D.R. Scotter, and R.E. White. 1994. A transfer function approach to modeling the leaching of solutes to subsurface drains: 1. Non-reactive solutes. Aust. J. Soil Res. 32:69–83.[CrossRef]
- Mallawatantri, A.P., and D.J. Mulla. 1996. Uncertainties in leaching risk assessments due to field averaged transfer function parameters. Soil Sci. Soc. Am. J. 60:722–726.[Abstract/Free Full Text]
- Martinez, J., and G. Guiraud. 1990. A lysimeter study of the effects of a ryegrass catch crop, during a winter wheat/maize rotation, on nitrate leaching and on the following crop. J. Soil Sci. 41:5–16.[CrossRef]
- Meng, C., S. Jia, and Y. Tang. 2004. Relationship between fixed ammonium and soil physical and chemical properties in Sinnamon and Fluvo-aquic soils. (In Chinese.) J. Hebei Agric. Sci. 8:39–42.
- Oenema, O., H. Kros, and W. de Vries. 2003. Approaches and uncertainties in nutrient budgets: Implications for nutrient management and environmental policies. Eur. J. Agron. 20:3–16.[CrossRef]
- Or, D., and J.M. Wraith. 2002. Soil water content and water potential relationships. p. 59. In A.W. Warrick (ed.) Soil physics companion. CRC Press, Boca Raton, FL.
- Ren, L., J. Ma, and R. Zhang. 2003. Estimating nitrate leaching with a transfer function model incorporating net mineralization and uptake of nitrogen. J. Environ. Qual. 32:1455–1463.[Abstract/Free Full Text]
- Russelle, M.P., J.F.S. Lamb, B.R. Montgomery, D.W. Elsenheimer, B.S. Miller, and C.P. Vance. 2001. Alfalfa rapidly remediates excess inorganic nitrogen at a fertilizer spill site. J. Environ. Qual. 30:30–36.[Web of Science]
- van Eerdt, M.M., and P.K.N. Fong. 1998. The monitoring of nitrogen surpluses from agriculture. Environ. Pollut. 102:227–233.[CrossRef]
- White, R.E. 1985. The transport of chloride and non-diffusible solutes through soil. Irrig. Sci. 6:3–10.
- White, R.E. 1987. A transfer function model for the prediction of nitrate leaching under field conditions. J. Hydrol. 92:207–222.[CrossRef]
- White, R.E. 1989. Prediction of nitrate leaching from a structured clay soil using transfer functions derived from externally applied or indigenous solute fluxes. J. Hydrol. 107:31–42.[CrossRef]
- White, R.E., J.S. Dyson, R.A. Haigh, W.A. Jury, and G. Sposito. 1986. A transfer function model of solute transport through soil: 2. Illustrative applications. Water Resour. Res. 22:248–254.
- White, R.E., L.K. Heng, and G.N. Magesan. 1998. Nitrate leaching from a drained, sheep-grazed pasture: 2. Modelling nitrate leaching losses. Aust. J. Soil Res. 36:963–977.[CrossRef]
- Wu, Y., S. Zhou, Z. Wang, and X. Zhang. 2005. Residual subsoil nitrogen utilization under water-saving cultivation in winter wheat. (In Chinese.) Acta Ecol. Sin. 25:1869–1873.
- Yuan, F., Z. Chen, Z. Yao, C. Zhou, G. Fu, Y. Song, and X. Li. 1995. NO3–N transformation accumulation and leaching loss in surface layer of Chao-soil in Beijing. (In Chinese.) Acta Ped. Sin. 32:388–399.
- Zhang, L., X. Ju, Q. Gao, and F. Zhang. 2004. Recovery of 15N-labeled nitrate injected into deep subsoil by maize in a calcaric cambisol in North China Plain. (In Chinese.) Plant Nutr. Fert. Sci. 10:455–461.
- Zhang, X., X. Yuan, L. Han, and H. Wang. 1994. Effects of soil conditions on root growth of winter wheat. (In Chinese.) Eco-Agric. Res. 2:62–68.
TOP
ABSTRACT
INTRODUCTION
