Modelling cation exchange in columns of disturbed and undisturbed subsoil

Cation exchange is often studied with disturbed and dried soils, but the applicability of the results to undisturbed soils is not straightforward. We investigated the value of exchange coefficients obtained from standard procedures for predicting cation exchange in soil. Columns of undisturbed and disturbed subsoil of a Luvisol (SBt horizon) were leached under saturated conditions with 0.4, 4, 20, 41, 102 and 205 mm BaCl₂ at a Darcy velocity of 1400 mm day^?1. The model PHREEQC was used to calculate one‐dimensional transport, inorganic complexation and multiple cation exchange. Two model variants were tested: m1 (exchangeable cations obtained by percolation with NH₄Cl) and m2 (exchangeable cations obtained by shaking the soil with BaCl₂). The exchange coefficients (Gaines–Thomas formalism) were calculated from the ion activities in solution and exchangeable cations obtained by NH₄Cl percolation (m1) or shaking with BaCl₂ (m2). Variant m1 predicted cation exchange of the disturbed (homogenized) soil for the entire BaCl₂ concentration range, whereas variant m2 resulted in a two‐fold overestimation of desorbed K for all experiments, which was related to large amounts of K released from the soil by shaking with BaCl₂. In experiments with undisturbed soil, variant m1 predicted the concentrations of Mg, Ca, K, and Na in the solution phase and the sum of cations released from exchange sites. However, variant m2 predicted changes in ion concentrations and exchangeable cations somewhat less well. This study suggests that the amounts of exchangeable cations and exchange coefficients obtained from experiments with homogenized soil by percolation are useful to predict cation concentrations in column experiments with undisturbed soils.     

Growth rates of Aporrectodea caliginosa (Oligochaetae: Lumbricidae) as influenced by soil temperature and moisture in disturbed and undisturbed soil columns

Earthworm growth is affected by fluctuations in soil temperature and moisture and hence, may be used as an indicator of earthworm activity under field conditions. There is no standard methodology for measuring earthworm growth and results obtained in the laboratory with a variety of food sources, soil quantities and container shapes cannot easily be compared or used to estimate earthworm growth in the field. The objective of this experiment was to determine growth rates of the endogeic earthworm Aporrectodea caliginosa (Savigny) over a range of temperatures (5–20 °C) and soil water potentials (−5 to−54 kPa) in disturbed and undisturbed soil columns in the laboratory. We used PVC cores (6 cm diameter, 15 cm height) containing undisturbed and disturbed soil, and 1 l cylindrical pots (11 cm diameter, 14 cm height) with disturbed soil. All containers contained about 500 g of moist soil. The growth rates of juvenile A. caliginosa were determined after 14–28 days. The instantaneous growth rate (IGR) was affected significantly by soil moisture, temperature, and the temperature×moisture interaction, ranging from −0.092 to 0.037 d⁻¹. Optimum growth conditions for A. caliginosa were at 20 °C and −5 kPa water potential, and they lost weight when the soil water potential was −54 kPa for all temperatures and also when the temperature was 5 °C for all water potentials. Growth rates were significantly greater in pots than in cores, but the growth rates of earthworms in cores with undisturbed or disturbed soil did not differ significantly. The feeding and burrowing habits of earthworms should be considered when choosing the container for growth experiments in order to improve our ability to extrapolate earthworm growth rates from the laboratory to the field.     

The variability of the apparent diffusion coefficient in undisturbed soil columns

In column studies in the laboratory the apparent diffusion coefficients(D)of chloride and TOH for undisturbed and unsaturated soils were determined as a function of the pore water velocity. Parallel the effective diffusion coefficients (D⁺) of Cl^? were measured as a function of water content. It was found that the values of D varied significantly between soil layers and soil types. This findings also held for the low flow velocities studied (0.3 and 1.0 cm/d). The impedance factors to calculate the effective diffusion coefficients (D⁺) from D values in the bulk water, decreased with decreasing bulk densities. In model calculations it was demonstrated that in spite of the differences of D values measured, for many purposes in the field good estimates of the solute distribution in the soils can be obtained by using simple relations between D and v_o.     

Physical and morphological characterization of undisturbed and disturbed ploughpans in a sandy loam soil

In a sandy loam soil the ploughpan had a lower porosity and a higher bulk density and mechanical resistance than ploughpans that had been disturbed by rota-digging two years previously. Hydraulic conductivities (K) for pressure heads (h) from 0 to ?20 cm were higher and moisture contents within the same h range were lower. Both values were identical for pressure heads lower than ?20 cm. According to the widely used physical capillary pore model, the volume of pores > 140 μm was 2% (v/v) in the undisturbed ploughpan and 8% (v/v) in the disturbed ploughpan. However, micromorphometric analyses showed that the undisturbed ploughpan contained the highest volume of large pores and many continuous root channels, while the disturbed ploughpan had no continuous large pores. Pore continuity was determined by using a staining test with methylene blue. Disturbance of the ploughpan in these soils results in formation of unstable fragments, which collapse, thereby disrupting the vertical continuity of the initially large packing voids between fragments.     

Seasonal pattern of total soil respiration in undisturbed and disturbed ecosystems of Central Himalaya

Summary The rates of CO₂ efflux were measured by an alkali absorption method (using 20 ml 0.5 N NaOH) from soils in four undisturbed sites [two evergreen oak forests, Quercus floribunda Lindl. (tilonj oak), Quercus leucotrichophora A Camus (banj oak), and two evergreen conifer forests, Cedrus deodara Loud. (deodar forest) and Pinus roxburghii Sarg. (chir pine forest)] and three disturbed sites. The sites were located between elevations of 1850 and 2360 m in the Central Himalaya. The seasonal pattern of soil respiration was similar in all the sites with a maximum during the rainy season, intermediate rates during the summer season and the lowest level of activity in winter. The rate of CO₂ efflux was higher in broadleaf than in conifer forests, and it was lowest in the disturbed sites. Among the edaphic conditions, soil moisture, N, organic C, pH, soil porosity, and root biomass positively affected total soil respiration. The proportion of root respiration to total soil respiration was higher in the disturbed sites than the undisturbed sites in winter. Conditions in the winter season were less favourable for microbial respiration than for root respiration.     

在土壤不受干扰条件下N矿化2/3规律研究

It is reported in literature that globally, only about two-thirds of N in nutrient sources is available to plants and microbes in soils under undisturbed conditions. The present study explores this phenomenon and suggests a new theory to explain this. Diffusion of NH4^＋ from microsite sources is considered here and analyzed for a capillary of the soil system. It has been found that 68% of NH4^＋ diffused from the nutrient source is initially preserved in the proximity of the nutrient source, whereas, the remaining 32% is diffused away from the nutrient source and possibly immobilized in the nonexchangeable abiotic pools or lost. The NH4^＋ that has been initially retained near the microsites is gradually released to the mineral pool. Thus, about two-thirds of NH4^＋ released from microsites is available only to plants and microbes in the soil. This is a universal, natural mechanism of nutrient conservation for plant and microbial availabilities. The theory has important implications and applications in recommending N fertilizations in the forms of organic or mineral sources to plants grown across various soil textures, for which further studies are needed. Any deviations from this theory can mainly be attributed to various disturbances to the soil; for example, mechanical disturbances, nutrient leaching, and variations in pH and soil faunal activities. This theory may apply to other elements mineralized in the soils under undisturbed conditions, which must be investigated in future studies.     

室外条件下香根草降低微污染物淋溶的结构性土柱研究

In Burkina Faso, significant amounts of endosulfan are applied to cotton fields; in addition, urban vegetable agriculture is often characterised by high fertiliser inputs, such as urban solid wastes containing heavy metals (e.g., Cu and Cd). Thus, the relevance of surrounding cotton and urban vegetable plots with vetiver (Vetiveria zizanioides) hedges to reduce environmental pollution by micropollutants was investigated using a leaching experiment, with outdoor lysimeters filled with two representative agricultural soils of Burkina Faso: Vertisol and Lixisol. After 6 months, little Cu was found in the leachates (< 0.010% of the applied amount) due to its high adsorption coefficient and its tendency to remain at the soil surface. Despite leachate and bromide recoveries being greater in soils planted with vetiver grass than in the bare soils, smaller amounts of endosulfan and Cd were found in the effluents from the planted soils (0.01% to 0.70% of the applied amount) than in those from the bare soils (0.01% to 1.48% of the applied amount), in agreement with their adsorption coefficients. These results may also be explained by a greater degradation of endosulfan in planted soils compared to bare soils and the absorption of Cd by vetiver. Thus, vetiver may decrease the risk of groundwater contamination, especially for Cd and endosulfan, which are more mobile than Cu. In addition, despite the smaller amounts of endosulfan and Cd measured in the Vertisol leachates (0.01% and 0.04% of the applied amount, respectively) compared to the Lixisol leachates, vetiver was more effective in decreasing the leaching of micropollutants if planted on Lixisol rather than on Vertisol. Further field monitoring is necessary to demonstrate the effectiveness of vetiver under the climatic conditions of Burkina Faso.     

Studies of 65Zn movement in soil columns under laboratory conditions

Zinc movement was studied in relation to moisture regime, element source, and method of application in soil columns constructed from samples of surface layers of Begu loamy sand and Gurgaon sandy loam. Maximum movement was only 10 mm laterally and 40 mm downward. The amount of zinc and the distance moved were greatest under the wettest regime. Movement was also greater in loamy sand than sandy loam columns, with Zn-EDTA rather than sulphate as the source, and when the amount of zinc was greater for a given volume of soil material.The effect of urea, monocalcium phosphate, calcium carbonate, and farmyard manure on the movement of zinc was determined at one moisture level in columns constructed from samples of surface layers of Begu loamy sand and Thana loam. The movement of zinc within the soil columns was reduced by all four treatments.     

Temporal and seasonal change in microbial community structure of an undisturbed,disturbed, and carbon-amended pasture soil

Disturbance and change to C inputs can alter microbial community structure and impact ecosystem function. Particularly in temperate regions, seasonal change also has an effect on microbial communities both directly through climate and indirectly through plant function. The temporal change in microbial communities of an undisturbed pasture, disturbed pasture (similar to a single tillage event) and pasture soil amended with two forms of particulate carbon were monitored over eight consecutive seasons after grass was reestablished. The soil microbial community was assessed by a DNA fingerprinting technique (terminal restriction fragment length polymorphism, TRFLP) of bacterial, fungal and archaeal communities, and also from phospholipid fatty acid (PLFA) analysis. The single disturbance had a significant effect on fungal microbial community structure (by TRFLP) and significantly decreased the fungal:bacterial ratio. Though the change was relatively small, it persisted throughout the sampling period. Nitrate was also higher on the disturbed treatment providing evidence for the theory that changes to fungal:bacterial ratios can alter nutrient cycling and retention. Fungal communities were the most altered by the C amendments, and while bacteria were also affected by the C amendments, seasonal change was a greater cause of variation. Correlation to soil and climatic variables explained more of the total variability for PLFA (78% for all treatments) than bacterial (50%), fungal (35%) and archaeal (14%) restriction fragments. Most climate and soil variables explained significant variation for seasonal patterns in the multivariate community structures but measurements of soil moisture were important for all communities while pH was relatively more important for bacteria, temperature for fungi, and soil C:N ratio for archaea. Autumn was particularly distinct from other seasons for bacteria (less so for the fungal community) and although there was seasonal change in pH suggesting pasture management was a factor, the significant correlation of other soil characteristics suggests that plant physiological changes (most probably root exudates) also played a significant role. The large change in the saprotrophic fungal community due to the particulate C addition but minor seasonal change would tend to suggest that the fungal community may be more responsive to changes in litter inputs rather than root exudates while the reverse is true for bacteria.     

Effect of rock fragment content on erosion processes of disturbed soil accumulation under field scouring conditions

Journal of Soils and Sediments - Soil and water erosion from high rock fragment content disturbed soil accumulation severely threatens ecological security. Evaluation of the quantitative influence...     

Dynamics of soil microbial biomass C, N and P in disturbed and undisturbed stands of a tropical wet-evergreen forest

To understand the spatial and temporal dynamics of soil microbial biomass and its role in soil organic matter and nutrient flux in disturbed tropical wet-evergreen forests, we determined soil microbial biomass C, N and P at two soil depths (0–15 and 15–30 cm), along a disturbance gradient in Arunachal Pradesh, northeastern India. Disturbance resulted in considerable increase in air temperature and light intensity in the forest and decline in the soil nutrients concentration, which affected the growth of microbial populations and soil microbial biomass. There were significant correlations between bacterial and fungal populations and microbial biomass C, N and P. Soil microbial population was higher in the undisturbed (UD) forest stand than the disturbed forest stands during post-monsoon and less during rainy season due to heavy rainfall. Greater demand for nutrients by plants during rainy season limited the availability of nutrients to soil microbes and therefore, low microbial biomass C, N and P. Microbial biomass was negatively correlated with soil temperature and pH in all the forest stands. However, there were significant positive relationships among microbial biomass C, N and P. Percentage contribution of microbial C to soil organic C was higher in UD forest, whereas percentage contribution of microbial biomass N and P to total N and total P was higher in the moderately disturbed site than in the highly disturbed (HD) site. These results reveal that the nutrient retention by soil microbial biomass was greater in the selective logged stand and would help in the regeneration of the forest upon protection. On the other hand, the cultivated site (HD) that had the lowest labile fractions of soil organic matter may recover at a slower phase. Further, minimum and maximum microbial biomass C, N and P during rainy and winter seasons suggest the synchronization between nutrient demand for plant growth and nutrient retention in microbial biomass that would help in ecosystem recovery following disturbance.     

Variation in 15N natural abundance of soil, humic fractions and plant materials in a disturbed and an undisturbed grassland

The natural abundance of δ¹⁵N in disturbed and undisturbed pasture soils was examined. From the disturbed soil, the top 10 cm of the profile was examined and the soil split into fractions based on particle size. Plant shoot and root material contained similar low enrichments in ¹⁵N, whereas recently deposited shoot residues were highly enriched. Differences between the soil fractions in observed total N did not reflect similar ¹⁵N variation. However, the enrichment of humic material extracted from the largest soil fraction was considerably lower in ¹⁵N relative to that from the smaller fractions. The complexity of the humic material from the larger fractions was less according to the E ₄ /E ₆ ratio. Analysis of the profile from the undisturbed soil showed increasing ¹⁵N enrichment with depth which corresponded well with visible soil horizons and showed an inverse relationship with total soil N. This ¹⁵N enrichment was mirrored by the enrichment in humic materials down the profile and also corresponded with an increasing chemical complexity as shown by the E ₄ /E ₆ ratio. Received: 15 March 1996     

基于BP神经网络的不同人为干扰强度下盐渍土SO42-定量分析

SO_4~(2-)是盐渍土阴离子中的主要离子,但目前针对不同人为干扰区域土壤中SO_4~(2-)反演研究却鲜有报道。土壤高光谱与土壤某元素间的关系表现为非线性,传统线性偏最小二乘模型(PLSR)对土壤元素的反演精度有限。本文以新疆昌吉回族自治州境内不同人为干扰区域的盐渍化土壤为研究对象,以土壤的野外高光谱和SO_4~(2-)含量为数据源,对原始(R)和对数(LogR)变换后的高光谱分别进行0阶、一阶和二阶微分预处理,选择通过0.05显著性水平的波段为敏感波段,将敏感波段对应的高光谱反射率作为非线性BP神经网络模型的输入变量,并设定BP的隐藏节点为300,学习速率为0.01,最大迭代次数为1 000,训练函数为trainscg。从SO_4~(2-)的真实值与预测值的散点图、拟合效果图和BP训练过程3个方面,定量分析无人为干扰(A区)和有人为干扰(B区)土壤SO_4~(2-)含量,并与PLSR对比预测精度。仿真显示, A区二阶微分后的BP预测精度优于一阶微分,而B区一阶微分后的BP预测精度优于二阶微分。且不论在A区还是B区, LogR光谱变换的反演精度均优于R。最佳BP模型的相对预测性能(RPD)、决定系数(R2)、均方根误差(RMSE)和迭代次数,在A区分别为3.309、0.906、0.253和8次,在B区分别为2.234、0.844、0.786和45次,表明BP对A区SO_4~(2-)的预测能力非常强(RPD2.5),对B区SO_4~(2-)的预测能力较强(RPD为2.0～2.5)。而在A区和B区两种光谱变换的一阶和二阶微分中, PLSR的RPD值均在1.4与1.8之间,其预测性能一般;在B区的0阶微分中, PLSR的RPD值均小于1.0,其不能对SO_4~(2-)进行预测。因此, BP模型能对不同人为干扰区域的SO_4~(2-)进行有效的定量分析。     

镉在中性和碱性土柱中不同深度处的运移模拟

Human health has been potentially threatened by cadmium (Cd) contained in sewage irrigation water.Previous studies of Cd transport in soils were mainly conducted using small soil cores with pH values less than 6.The objectives of this study were to determine the parameters of the convection-dispersion equation (CDE) for Cd transport in relatively larger columns with neutral and alkaline soils,and to investigate the parameters' variability with depth.The soil columns were 50 cm in length and 12.5 cm in diameter.Ceramic suction lysimeters were buried at depths of 2.5,7.5,17.5,27.5,and 37.5 cm to abstract soil solution.Cd concentration in the soil solution samples were subsequently analyzed to obtain breakthrough curves (BTCs).Equilibrium and nonequilibrium models in CXTFIT program were used to estimate parameters of the CDE.The results suggested that both equilibrium and non-equilibrium models performed well in modeling Cd transport.The hydrodynamic dispersion coefficient (D) ranged from 0.18 to 10.70 cm 2 h 1,showing large differences among different depths.The retardation factor (R d) ranged from 25.4 to 54.7 and the standard deviation of R d value was lower than 30% of the mean value.Precipitation coefficient (R p) decreased consistently with increasing depth,varying from 1.000 × 10 10 to 0.661 h 1.Sensitivity tests showed that D was less sensitive than R d.These results would be helpful in understanding the transport and retention of Cd in non-acidic soils.     

Fate of anthracene in contaminated soil: transport and biochemical transformation under unsaturated flow conditions

To obtain reliable estimates for the loss of polycyclic aromatic hydrocarbons (PAHs) from contaminated soils, one has to distinguish between (i) losses due to release and solute transport and (ii) losses resulting from degradation. We studied the interplay of these processes in a column experiment representing a typical soil contamination scenario: in the upper part of the column was a contaminated layer, spiked with 9‐¹³C‐labelled anthracene, and beneath it uncontaminated pristine soil. The experimental course comprised a steady‐state flow phase (constant irrigation for 4 months) followed by several periods during which flow was halted. The effects of varied residence time on anthracene biodegradation and on anthracene mass transfer were investigated. We monitored labelled anthracene and its transformation products, dissolved organic carbon, electric conductivity (EC), pH, and inorganic carbonate content in the column effluent, and the CO₂ evolved. Under steady‐state flow, pH, dissolved organic C, and EC approached steady states after 350 pore volumes. Concentrations of anthracene in the effluent, however, increased continuously and levelled off after 800 pore volumes. This marked retardation reflects the great affinity of anthracene to soil organic matter. The response to interruptions in the flow revealed that mass is transferred without equilibrium between solid and liquid phase for both anthracene and dissolved organic C. Thus, residence time is one factor controlling the concentration of anthracene in the effluent and therefore the export of contaminant to the aquifer. In the course of the experiment several labelled anthracene degradation products appeared in the effluent. At least three of them were identified as transformation products showing a dramatic increase in mobility relative to the parent compound. A third of the overall anthracene loss from the column was due to solute transport, and biodegradation was responsible for the remaining two thirds. The incomplete degradation of anthracene leads to the formation of highly mobile transformation products and thus promotes the export of carbon, derived from the contaminant.     

Measured and predicted transport of two S-triazine herbicides through soil columns

Effluent concentration of chloride and two pesticides (prometon and atrazine) were measured during column displacement experiments at two water flow rates. A constant suction of approximately 1300 Pa was maintained in the packed soil columns which were positioned vertically on top of a vacuum chamber enclosing an automatic fraction collector. Measured breakthrough curves (BTC's) were analyzed in terms of two solute transport models: the standard two-parameter convection-dispersion equation (CDE), and a four-parameter two-site/two-region nonequilibrium model (TRM). Calculations obtained with the TRM model were found to be in better agreement with measured BTC's than predictions using the CDE model. Column retardation factors for prometon and artrazine calculated from equilibrium batch sorption coefficients were comparable to those estimated from the observed BTC's only when the nonequilibrium TRM model was used.     

Role of a non-ionic surfactant and carboxylic acids on the leaching of aged DDT residues in undisturbed soil columns

Journal of Soils and Sediments - Irrigation or flooding events with riverine waters containing compounds such as surfactants, colloids, and dissolved organic carbon and plant growth will modify...     

Water transport under winter conditions

Winter as well as summer floods result in soil loss and sedimentation. Up to now the winter events cannot be adequately predicted. This paper focuses on the infiltration processes under frozen winter conditions in order to model soil erosion processes in winter by adapting the computer model EROSION 3D [Schmidt, J., Werner, M. v., 2000. Modeling Sediment and Heavy Metal Yields of Drinking Water Reservoirs in the Osterzgebirge Region of Saxony (Germany). In: Soil Erosion - Application of Physically Based Models, Schmidt, J.(Editor). Berlin, Heidelberg, New York., pp. 93-108.].A new snow accumulation and snow melt module has been implemented in order to estimate erosion rates during snowmelt events. Tests show that infiltration still occurs in frozen soils, however, infiltration rates are reduced compared to unfrozen soils [Weigert, A., Wenk, G., Ollesch, G., Fritz, H., 2003. Simulation of snowmelt erosion using the EROSION 3D model. Journal of Plant Nutrition and Soil Sciences, 1/2003.]. In order to improve the EROSION 3D model regarding partly frozen soils a physical based infiltration model extension has been developed and experimentally verified.Processes of infiltration into partly frozen soils are successfully quantified by a newly designed experimental set-up using a soil column (height 50 cm, diameter 21.5 cm). At the bottom of this column a negative pressure can be applied in order to establish unsaturated hydraulic conditions. The volume rate of the percolating water is constantly measured by an online balance. In addition the column is equipped with three TDR and temperature probes.The behaviour of two soil samples (sandy vs. loamy soil) are investigated under saturated, unsaturated and frozen conditions. The improved physical infiltration model based on the combination of Darcy's Law, Hagen-Poiseuille's Law, the capillary-rise equation and the van Genuchten θ(h) function determines with considerable accuracy both the unsaturated hydraulic conductivity and the effective saturated hydraulic conductivity of a partly frozen soil for rigid soil matrix conditions. This approach is compared with the Mualem concept for predicting unsaturated hydraulic conductivities. Fractures were observed due to freezing cracks in case of loamy material. For fractured soils the calibration with a skinfactor is found to be absolutely necessary to give reliable results.     

Water transport under winter conditions

Winter as well as summer floods result in soil loss and sedimentation. Up to now the winter events cannot be adequately predicted. This paper focuses on the infiltration processes under frozen winter conditions in order to model soil erosion processes in winter by adapting the computer model EROSION 3D [Schmidt, J., Werner, M. v., 2000. Modeling Sediment and Heavy Metal Yields of Drinking Water Reservoirs in the Osterzgebirge Region of Saxony (Germany). In: Soil Erosion - Application of Physically Based Models, Schmidt, J.(Editor). Berlin, Heidelberg, New York., pp. 93-108.].A new snow accumulation and snow melt module has been implemented in order to estimate erosion rates during snowmelt events. Tests show that infiltration still occurs in frozen soils, however, infiltration rates are reduced compared to unfrozen soils [Weigert, A., Wenk, G., Ollesch, G., Fritz, H., 2003. Simulation of snowmelt erosion using the EROSION 3D model. Journal of Plant Nutrition and Soil Sciences, 1/2003.]. In order to improve the EROSION 3D model regarding partly frozen soils a physical based infiltration model extension has been developed and experimentally verified.Processes of infiltration into partly frozen soils are successfully quantified by a newly designed experimental set-up using a soil column (height 50 cm, diameter 21.5 cm). At the bottom of this column a negative pressure can be applied in order to establish unsaturated hydraulic conditions. The volume rate of the percolating water is constantly measured by an online balance. In addition the column is equipped with three TDR and temperature probes.The behaviour of two soil samples (sandy vs. loamy soil) are investigated under saturated, unsaturated and frozen conditions. The improved physical infiltration model based on the combination of Darcy's Law, Hagen-Poiseuille's Law, the capillary-rise equation and the van Genuchten θ(h) function determines with considerable accuracy both the unsaturated hydraulic conductivity and the effective saturated hydraulic conductivity of a partly frozen soil for rigid soil matrix conditions. This approach is compared with the Mualem concept for predicting unsaturated hydraulic conductivities. Fractures were observed due to freezing cracks in case of loamy material. For fractured soils the calibration with a skinfactor is found to be absolutely necessary to give reliable results.