Estimating leaching losses from sub-surface drained soils

Abstract. Leaching losses of solutes can be calculated if two variables, the amount of water passing through the soil and the concentration of solute in that water (a flux concentration), are known. Two simple approaches, soil extraction and suction cup sampling, were used to estimate the concentration of solutes in the water moving through a silt loam soil. The results were compared with actual concentrations measured in the drainage water from a sub-surface (mole-pipe) drained soil.
Seasonal leaching losses were calculated as the sum of the products of estimated monthly drainage and the estimated average monthly solute concentration in the soil solution. These results were compared with the leaching losses measured in drainage water from the mole-pipe system. For non-reactive solutes such as bromide (an applied solute) and chloride (a resident solute), the suction cup data provided better estimates of the leaching losses than did the soil extraction data. The leaching losses calculated using volume-averaged soil solution concentrations (obtained by soil extraction) overestimated the loss for the resident solute, but under-estimated the loss for the surface-applied solute. On the other hand, the data for non-reactive solutes suggest that measurements on suction cup samples may be representative of the flux concentration of a solute during leaching. For nitrate, a biologically reactive solute, there was no clear pattern in the differences between the estimated and measured leaching losses. The flux-averaged concentration in the drainage water was about midway between those measured in the suction cup samples and in the soil solution.     

The utility of Burns’s equation to describe solute movement through soil under various boundary and initial conditions

Burns’s equation for describing solute movement through soil is attractive because it is simple and predicts adequately in many instances. However, the assumptions implicit in it are not inconsistent with preferential solute flow. We have explored the consequences of this by leaching initially resident chloride and surface-applied tritium and nitrate through 250-mm-long intact cores of a silt loam soil. The applied flow rates of 3 and 5 mm h^?1 (realistic rainfall intensities) produced unsaturated soil conditions, except near the base where free water dripped out. Burns’s equation described the movement of the three solutes fairly successfully with the water content parameter having values between 0.29 and 0.48, similar to the actual volumetric water content of 0.47. The leaching of resident chloride to 450-mm-deep mole drains in the field was also successfully simulated using Burns’s equation. However, simulation of the leaching of bromide applied to the soil surface as a solid salt was problematic. This resulted from uncertainty as to whether to treat the application as a pulse input to the flux or resident concentration. The observed behaviour fell about midway between the simulations for these contrasting initial conditions.     

A usable mechanistic model of nitrate leaching I. The model

A mechanistic model to simulate the leaching of nitrate in soil is described. Three flow patterns are accommodated: conventional convective-dispersion; preferential flow, in which water and solute moving in the larger water-filled pores fails to equilibrate with that in the smaller pores; and by-pass flow, in which water (and solute) moves rapidly down cracks and fissures without displacing the soil water. The pattern of flow depends on the rainfall intensity and the hydraulic conductivity of the water-filled pore space. The performance of the model is examined in terms of its ability to reproduce solute profiles calculated from analytical solutions to transport equations, and in the sensitivity of the output to changes in the main parameters. A second paper (Barraclough, 1989) illustrates the use of the model to reproduce nitrate leaching patterns observed in the field.     

A usable mechanistic model of nitrate leaching II. Application

Simulations produced by a mechanistic model are compared with field observations of nitrate leaching from 0.35 ha hydrologically isolated experimental plots. The parameters used in the model are obtained in two ways. First by fitting the model to field observations in one year. These parameters are then used to simulate leaching in other years. Second, model parameters are obtained by fitting eluant profiles from pulse inputs of solute to undisturbed cores in the laboratory. When used with the field-derived parameters, the model simulates total leaching losses well in other years, although the pattern of loss is only approximately reproduced. The simulation suggests that water and solute flow in drained, structured soils is complex; preferential flow in the upper horizons resulting in 20% of the water-filled pore space carrying most of the solute flow, and by-pass flow in the subsoil causing rapid movement of water and solute to the drains. The result is that much of the nitrate in the upper horizons appears to be protected from leaching. When used with laboratory-derived parameters, the model was a poor predictor of both the pattern and quantity of nitrate leached.     

华北平原水浇玉米-小麦轮作农田硝态氮的淋失

Soil water deep drainage and nitrate (NO3^-) leaching losses below the root zone were investigated in a 1 ha wheatmaize rotation field under traditional agricultural management that local farmers generally follow in the North China Plain, using the soil water balance method and NO3-N concentration in suction samples. Water drainage, and NO3-N distribution and leaching losses exhibited pronounced spatial and temporal variability. Soil water deep drainage and NO3- N leaching loss mostly occurred during the summer maize growing season (rainy season), which coincided with irrigations and significant rainfall. On average, soil water deep drainage was 39 and 90 mm in the 1998/1999 and 1999/2000 cropping years, correspondingly, accounting for 10% and 19% of the total irrigation plus rainfall, respectively. The NO3-N leaching loss from soil and fertilizer N below the root zone ranged from 6 to 17 (averaging 12) and 30 to 84 (averaging 61) kg N ha^-1 in 1998/1999 and 1999/2000, correspondingly, equivalent to 1.4%-4.1% and 7.3%-20.3% of N fertilizer applied,respectively. The results indicated that water and fertilizer inputs could be greatly reduced, thus improving water and nutrient use efficiency in this region.     

Critical drainage and nitrate leaching losses from manures applied to freely draining soils in Great Britain

Abstract. Pig slurry was applied by open-slot injection to experimental plots on a sandy loam site at ADAS Gleadthorpe, Nottinghamshire. Volume and distribution of over-winter drainage were adjusted through the use of rainfall exclusion covers or irrigation. The resultant slurry N leaching over the range of drainage values tested (up to 300 mm) could be satisfactorily described by curve-fitting, using a quadratic or exponential function. Initial simulations of slurry N leaching using the manure nitrogen decision support system manner (v. 3.0) compared poorly with the experimental data, predicting both earlier and greater amounts of nitrate leaching. However, the lack of fit could be explained by consideration of the likely ammonia emissions following slurry injection, the actual volumetric soil moisture capacity at the experimental site and the likely time delay for the nitrification of slurry N following application. Good agreement between modelled and observed data was achieved when these factors were taken into account. The manner model was used to simulate nitrate leaching beyond the range of drainage treatments tested in the experiments and the anticipated sigmoidal relationship between nitrate leaching and drainage was observed. The model was then used to study the effects of manure application timing and the likely impact on nitrate leaching, across the range of rainfall conditions found in Great Britain. Simulations for a range of manure types were undertaken, with manures applied at rates up to the limit of permitted N loading on freely draining sandy loams. Rainfall inputs for these simulations were based on long-term average climatic data. Results are presented for two contrasting manure types, cattle slurry and poultry manure, both of which are subject to controls in Nitrate Vulnerable Zones (NVZs) in Great Britain.     

溶质施放部位对红壤坡面溶质迁移特征的影响

以非吸附性溴离子为示踪剂,采用室内模拟降雨方法研究间歇降雨和施肥部位对红壤坡面产流、产沙以及溴离子迁移特征的影响。结果表明,降雨特征相同条件下,3次降雨的总径流量无明显差异,但第3次降雨形成的产沙量分别是第1次和第2次的7倍和2.2倍,这表明因前期含水量的提高而降低了土壤抗冲性是红壤侵蚀的主要原因之一;径流溴离子浓度随时间变化均呈幂函数衰减,与施肥部位无关;施肥部位越靠近坡底,径流溴离子初始浓度越高且衰减速度越快;溴离子流失数量与施肥距离呈显著正线性关系。通过估算3次间歇降雨径流中溶质流失数量的来源发现,淋溶到坡面土壤中的溴离子再次参与径流流失的数量,其平均比重从81.61%提高到了93.76%,这表明施加在上坡部位的肥料被淋溶后对后期径流养分流失的贡献十分显著。     

Cadmium leaching from some New Zealand pasture soils

Cadmium (Cd) inputs and losses from agricultural soils are of great importance because of the potential adverse effects Cd can pose to food quality, soil health and the environment in general. One important pathway for Cd losses from soil systems is by leaching. We investigated loss of Cd from a range of contrasting New Zealand pasture soils that had received Cd predominantly from repeated applications of phosphate fertilizer. Annual leaching losses of Cd ranged between 0.27 and 0.86 g ha^–l, which are less than most losses recorded elsewhere. These losses equate to between 5 and 15% of the Cd added to soil through a typical annual application of single superphosphate, which in New Zealand contains on average 280 mg Cd kg^?1 P. It appears that Cd added to soil from phosphate fertilizer is fairly immobile and Cd tends to accumulate in the topsoil. The pH of the leachate and the total volume of drainage to some extent control the amount of Cd leached. Additional factors, such as the soil sorption capacity, are also important in controlling Cd movement in these pasture soils. The prediction of the amount of Cd leached using the measured concentrations of Cd in the soil solution and rainfall data resulted in an overestimation of Cd losses. Cadmium concentrations in drainage water are substantially less than the current maximum acceptable value of 3 µg l^?1 for drinking water in New Zealand set by the Ministry of Health.     

An amended functional leaching model applicable to structured soils. I. Model description

A functional model designed to simulate the transport of non-interactive solutes through macroporous soil is described. The concept of mobile and immobile water is used but the pore volume available for mobile water is partitioned to allow for flow through smaller water-filled pores and rapid preferential flow through larger macropores and fissures. The general performance of the model under steady infiltration following an injection of solute is presented. The sensitivity of the output to variations in the model parameters is also discussed. A second paper compares the model with bromide and chloride leaching data on two texturally contrasting soils.     

Calibration and validation of a model of non-interactive solute leaching in a clay-loam arable soil

A capacity-type approximate leaching model (Addiscott et al., 1986) with a simple treatment of soil matrix permeability was tested, using field tracer experiments with CaBr₂, on hydrologically isolated plots. The model predictions are sensitive to the value of the soil matrix permeability factor, a, and four methods of estimating this parameter were evaluated: (1) using a calibration based on soil texture; (2) least-squares fitting of the model to successive neutron probe measurements of the water content profile; (3) least-squares fitting to daily drainage outflow; (4) least-squares fitting to cumulative drainage outflow. The best independent method (method 3) led to slight (20–30%) under-prediction of leaching losses for two of the four experiments, but in one experiment leaching was much less than predicted. As a management model the approach seems promising but more attention needs to be paid to estimation of the value and variability of the permeability parameter, a. The convective-dispersion equation, using steady-state assumptions and a fitted dispersion length, gave as good a prediction of cumulative leaching losses as the approximate model studied here.     

A simplified analysis of soil water flow to a mole drain

An analysis is described which allows transient mole-drain flow, surface runoff and water table depth to be computed, given measured or assumed values for rainfall, evaporation, deep percolation, mole-drain spacing and depth, surface detention, the drainage coefficient, and the saturated hydraulic conductivity, total porosity and macroporosity of the topsoil and subsoil. Dupuit-Forchheimer flow below the water table and vertical hydraulic potential equilibrium above the water table are assumed. The analysis was successfully tested by simulating the hydrographs from a mole-pipe drained silt loam, and comparing the results with the measured hydrographs for three contrasting rainfall events.     

Prediction of solute leaching under field conditions: an appraisal of three methods

The analytical equation of Rose was more successful than the models of Burns and Addiscott in predicting movement of chloride through a soil on Upper Chalk under irrigation and winter rainfall. The two models required water inputs to be divided into 5 mm increments. When this was done the model of Addiscott was more successful than that of Burns although both then gave predictions of acceptable accuracy for practical purposes. The analytical equation of Rose requires a value for solute dispersivity. This was obtained from the experimental measurements, but the value used (3 cm) was found not to be critical.     

地面灌溉技术参数对氮素运移分布影响的研究进展

地面灌溉是应用最广泛的灌水方法，通过优化地面灌溉设计和管理减少水氮淋失，正在成为研究的热点问题，该文对有关成果进行了综述。由于地面灌溉的尺度较大，加之土壤水力参数和溶质运移参数的空间变异，使得地面灌溉施肥条件下溶质(N)运移分布的田间试验研究，无论在技术上还是在投入上都具有一定难度。有关沟灌、畦灌和水平畦田灌溉施肥时水氮分布的田间试验结果指出，肥料(N)分布没有入渗水深分布均匀，且与土壤初始含水率、流量、土壤入渗参数、田面糙率和施肥时机等因素密切相关。在模拟研究方面，文献中报道的模型有纯对流模型、对流-弥散模型、基于活塞流的水平衡模型和传递函数模型等。鉴于地面灌溉条件下水氮运移的复杂性，在水氮运移分布的数学模拟、灌水技术参数和施肥方式对水氮淋失的影响以及灌溉施肥质量评价指标体系等方面尚需要进一步研究。     

Modelling water and solute transport in macroporous soil. II. Chloride breakthrough under non-steady flow

A model of water and solute transport in macroporous soils (Jarvis et al., 1991) has been evaluated in column breakthrough experiments under field conditions. Hydraulic properties were first measured in replicate soil monolith lysimeters sampled from grass ley and continuous barley treatments in a clay soil. A pulse input of 0.05 M KCl was then supplied by drip irrigation and measurements made of the water discharge and chloride leaching resulting from the natural rainfall over a 1-month period. The results showed that the macropores constituted the dominant flow pathway (accounting for 80% of the total water outflow) and that diffusive exchange of chloride between the two flow domains was the main factor causing variability in leaching. Larger hydraulic conductivities and macroporosities in the lower topsoil and at plough depth in the grass ley monoliths were taken as evidence of structural amelioration. Less of the applied chloride was leached in the grass monoliths than in the barley (means of 20% and 31% respectively). This was mainly due to a smaller effective aggregate size and thus a more efficient diffusion-controlled retention.     

地下滴灌土壤水运动和溶质运移数学模型的应用

利用建立的地埋点源土壤水运动和溶质运移数学模型描述地下滴灌条件下土壤水、肥运动的分布规律,将土壤质地结构、滴头出流量、滴头埋深和单次灌水历时等因素对土壤水分布的影响进行模拟分析。结果表明,在确定的土壤质地条件下,滴头出流量和埋深是影响地下滴灌系统性能的两个最重要的灌水设计参数,应尽量采用增加滴头数量而不是选用大流量滴头的方法来满足作物的需水要求。此外,合理的灌溉施肥时机应依据当地的土壤质地条件予以确定     

An amended functional leaching model applicable to structured soils. II. Model application

The performance of a functional leaching model (Hall, 1993) is compared with leaching data from two lysimeter experiments with soils of contrasting texture using sodium bromide and potassium chloride as the non-reactive solutes. The model parameters are optimized using the solute elution curves as standards and compared with the physical properties of the soil. A good match with the measured discharge of both water and solute was achieved for both soils using the moisture release characteristics to define the pore volume available for mobile and immobile water. The results indicate that preferential flow takes place through even coarse-textured soils but that there is negligible diffusive exchange of solute between water passing through the macropores and the rest of the soil.     

Nitrogen losses from the field: denitrification and leaching in intensive winter cereal production in relation to tillage method of a clay soil

Abstract. Losses of soil and fertilizer nitrogen by leaching and denitritication from a clay soil in southern England have been measured over four years. Nitrate losses in drainage water from direct-drilled land averaged 20–30 kg N ha 'a' with wide seasonal variation. Ploughing and conventional cultivations increased this loss. Denitritication from direct-drilled land averaged 5–10 kg N ha 'a' with wide seasonal variation. Ploughing and drainage both diminished denitritication losses but cultivation had the greater effect. These nitrogen losses occurred mainly in autumn and spring.
Nitrogen losses, in drainage water or by denitritication after spring fertilizer applications, were related to the rainfall in the 28 days following top dressing. Approximately 40 mm rain was needed to cause a loss of 10% of the nitrogen applied but in practice losses were quite variable.     

Simulation of Water and Chloride Transport in Field Lysimeters

In a lysimeter study in the field with soil samples of a Fluvisol the transport of chloride was monitored under high frequency irrigation. The results of this investigation were used for verification of two computer models of different complexity. One was a transient state water and solute movement model the other a simplified steady state model. It was demonstrated that both models predicted nearly equally well the chloride concentrations in the drainage water. For many practical situations, the simple model may be all that is needed to predict the downward movement of solutes in field soils.     

Micronutrients Losses from Soil under Subsurface Drainage System

The drainage system, although it reclaims waterlogged soils, poses a potential threat to leaching of valuable nutrients. To assess the micronutrient losses from such system established more than two decades ago in the plains of northwestern Pakistan, drainage waters were periodically analyzed for selected micronutrients. The inputs of micronutrients in irrigation waters were also determined. The micronutrient removal in drainage water was 1.39 to 9.79 times greater than micronutrient inputs in irrigation waters and fertilizers. Among micronutrients, the removal of copper (Cu) and manganese (Mn) in drainage water was greater than iron (Fe) and zinc (Zn). The concentrations of micronutrients generally changed with cropping (sampling time), increased with soil depth, and were invariably negative on the nutrient balance sheet, suggesting that these nutrients are continuously depleted from the system in drainage water. These nutrients therefore must be replenished in soil through an appropriate means for optimum crop yields under the given drainage–irrigation–cropping system.     

Nitrate leaching losses and their control in a mixed farm system in the Cotswold Hills, England

Abstract. Nitrate leaching was measured for four years at the Royal Agricultural College 's Coates Farm in the Cotswolds, England. Coates is a typical Cotswold mixed farm with thin, well-drained calcareous soils especially prone to leaching. Over the duration of this study there were dairy, sheep and arable enterprises on the farm. A 'Farm Gate' nitrogen (N) budget was constructed. Small 120 m × 20 m 'farmlets' were sited in ten fields across the farm, covering all parts of the rotation, as the sites for detailed measurements. Each farmlet received the same management as the rest of the field in which they were situated. Using ceramic probes inserted to 60 cm, soil water was sampled every two weeks throughout the winter drainage season. The annual drainage varied from 135 mm under grassland in 1996/7 to 600 mm under cereals in 1998/9. Average N losses by leaching were determined mostly by rainfall and were 65 kg N ha^–1 yr^–1, accounting for 25% of the N inputs. Especially leaky parts of the rotation were the ploughing out of a lucerne ley and the grazing of stubble turnips with sheep, both typical Cotswold farm practices. The research highlights some of the difficulties in developing practicable, profitable management practices to decrease nitrate losses.