Development and use of a database of hydraulic properties of European soils

Many environmental studies on the protection of European soil and water resources make use of soil water simulation models. A major obstacle to the wider application of these models is the lack of easily accessible and representative soil hydraulic properties. In order to overcome this apparent lack of data, a project was initiated to bring together the available hydraulic data which resided within different institutions in Europe into one central database. This information was then used to derive a set of pedotransfer functions applicable to studies at a European scale. These pedotransfer functions predict the hydraulic properties from parameters collected during soil surveys and can be a good alternative for costly and time-consuming direct measurement of these properties. A total of 20 institutions from 12 European countries collaborated in establishing the database of draulic operties of uropean oils (HYPRES). This database has a flexible relational structure capable of holding a wide diversity of both soil pedological and hydraulic data. As these data were contributed by 20 different institutions it was necessary to standardise both the particle-size and the hydraulic data. A novel similarity interpolation procedure was successfully used to achieve standardization of particle-sizes according to the FAO clay, silt and sand particle-size ranges. Standardization of hydraulic data was achieved by fitting the Mualem-van Genuchten model parameters to the individual θ(h) and K(h) hydraulic properties stored in HYPRES. The HYPRES database contains information on a total of 5521 soil horizons (including replicates). Of these, 4030 horizons had sufficient data to be used in the derivation of pedotransfer functions. Information on both water retention and hydraulic conductivity was available for 1136 horizons whereas 2894 horizons had only information on water retention. Each soil horizon was allocated to one of 11 possible soil textural/pedological classes derived from the six FAO texture classes (five mineral and one organic) and the two pedological classes (topsoil and subsoil) recognised within the 1:1 000 000 scale Soil Geographical Data Base of Europe. Next, both class and continuous pedotransfer functions were developed. By using the class pedotransfer functions in combination with the 1:1 000 000 scale Soil Map of Europe, the spatial distribution of soil water availability within Europe was derived.     

Modelling spatial uncertainty of soil erodibility factor using joint stochastic simulation

Soil erosion varies greatly over space and is commonly estimated using the revised universal soil loss equation (RUSLE). Neglecting information about estimation uncertainty, however, may lead to improper decision‐making. One geostatistical approach to spatial analysis is joint stochastic simulation, which draws alternative, equally probable, joint realizations of a regionalized variable. Differences between the realizations provide a measure of spatial uncertainty and allow us to carry out an error propagation analysis. The objective of this paper was to assess spatial uncertainty of a soil erodibility factor (K) model resulting from the uncertainties in the input parameters (texture and organic matter). The 500 km² study area was located in central‐eastern Sardinia (Italy) and 152 samples were collected. A Monte Carlo analysis was performed where spatial cross‐correlation information through joint turning bands simulation was incorporated. A linear coregionalization model was fitted to all direct and cross‐variograms of the input variables, which included three different structures: a nugget effect, a spherical structure with a shorter range (3500 m) and a spherical structure with a longer range (10 000 m). The K factor was then estimated for each set of the 500 joint realizations of the input variables, and the ensemble of the model outputs was used to infer the soil erodibility probability distribution function. This approach permitted delineation of the areas characterized by greater uncertainty, to improve supplementary sampling strategies and K value predictions. Copyright © 2007 John Wiley & Sons, Ltd.     

Using soil knowledge for the evaluation of mid‐infrared diffuse reflectance spectroscopy for predicting soil physical and mechanical properties

Mid‐infrared diffuse reflectance spectroscopy can provide rapid, cheap and relatively accurate predictions for a number of soil properties. Most studies have found that it is possible to estimate chemical properties that are related to surface and solid material composition. This paper focuses on prediction of physical and mechanical properties, with emphasis on the elucidation of possible mechanisms of prediction. Soil physical properties that are based on pore‐space relationships such as bulk density, water retention and hydraulic conductivity cannot be predicted well using MIR spectroscopy. Hydraulic conductivity was measured using a tension‐disc permeameter, excluding the macropore effect, but MIR spectroscopy did not give a good prediction. Properties based on the soil solid composition and surfaces such as clay content and shrink‐swell potential can be predicted reasonably well. Macro‐aggregate stability in water can be predicted reasonably as it has a strong correlation with carbon content in the soil. We found that most of the physical and mechanical properties can be related back to the fundamental soil properties such as clay content, carbon content, cation exchange capacity and bulk density. These connections have been explored previously in pedotransfer functions studies. The concept of a spectral soil inference system is reiterated: linking the spectra to basic soil properties and connecting basic soil properties to other functional soil properties via pedotransfer functions.     

Using existing soil databases for estimating retention properties for soils of the Pianura Padano-Veneta region of North Italy

Data collected on benchmark soils from experimental sites in the Pianura Padano-Veneta, Northern Italy, stored in local soil data bases were used to test the reliability of existing pedotransfer functions to derive soil water retention properties, required as input to simulation models of pollutant transport in soils. Eight different algorithms were selected according to their principle of calculation, input variable requirements and in agreement with the different information currently existing in local soil databases. Results were validated against an experimental set of 139 retention curves. In order to assess the validity of the selected pedotransfer functions and to provide operative guidelines for their selection, quantitative error indices mean difference, and square root of the mean squared difference, were calculated and the results evaluated for the whole data set and for subsets of observations classed in terms of soil textural composition, bulk density, organic carbon content and matric potential. Non-parametric significance tests for unbalanced data were applied to assess the significance of the differences among classes. According to the kind of soil on which they were tested, the selected pedotransfer functions showed significantly different responses in terms of accuracy, providing therefore useful indications for their optimal applicability for different levels of available information.     

青海三江源地区土壤水分常数转换函数的建立与比较

利用土壤理化性质数据建立转换函数是间接获得土壤水力参数的重要手段之一。基于测定的土壤理化性质和土壤水分常数数据,本文采用回归分析、BP神经网络和基于BP神经网络的Rosetta模型3种方式分别建立了青海三江源地区土壤饱和含水量、毛管持水量和田间持水量的转换函数,并对其预测精度进行了比较。结果表明:(1)回归分析方法总体预测效果比较理想,特别是田间持水量的平均误差(ME)和均方根误差(RMSE)都在3.397%以下,决定系数(R2)高达0.868;(2)BP神经网络方法的预测效果非常理想,各土壤水分常数平均误差和均方根误差都在4.685%以下,并且决定系数均在0.857以上;(3)Rosetta模型的预测效果相对较差,特别是饱和含水量和毛管持水量,平均误差(ME)和均方根误差(RMSE)相对较大,决定系数(R2)相对较小。3种方式中,BP神经网络方法所建立的毛管持水量和饱和含水量转换函数均为最佳,回归方法所建立的田间持水量的转换函数要好于BP神经网络方法和Rosetta模型,Rosetta模型对土壤水分常数的预测效果不如其他两种方式。研究可为青海三江源地区土壤水力特性参数研究以及区域尺度上土壤水分估算提供科学依据。     

基于单一和集合土壤转换函数模型对土壤含水量的模拟性能分析

土壤转换函数(Soil pedotransfer function,PTF)是一种高效获取土壤水力参数的方法。由于土壤具有很强的空间异质性,确定最优PTF模型成为模拟土壤含水量的关键。为此,以海河流域3个实验场地(密云站、大兴站、馆陶站)为研究区,采用7种常用的单一PTF模型预测土壤水力参数作为HYDRUS-1D的模型参数,求解Richards方程获得土壤含水量,并与实测土壤含水量进行比较,评价了常用单一PTF模型预测的土壤水力参数对土壤含水量的模拟性能。此外,采用3种方法构建集合PTF模型,评价了集合PTF模型对土壤含水量的模拟性能。结果表明：基于van Genuchten方程构建的单一PTF作为模型参数模拟土壤含水量的均方根误差最小;而其中Rosetta3模型表现更优。在集合PTF模型中,基于遗传算法加权法构建的模型表现最好。集合PTF模型预测土壤水力参数可以较好的捕捉多个单一PTF预测土壤水力参数的整体趋势,弥补单一PTF在某些情况下模拟误差较大的不足。     

Estimating forest soil bulk density using boosted regression modelling

Soil bulk density (ρ) is an important physical property, but its measurement is frequently lacking in soil surveys due to the time‐consuming nature of making the measurement. As a result pedotransfer functions (PTFs) have been developed to predict ρ from other more easily available soil properties. These functions are generally derived from regression methods that aim to fit a single model. In this study, we use a technique called Generalized Boosted Regression Modelling (GBM; Ridgeway, 2006 ) which combines two algorithms: regression trees and boosting. We built two models and compared their predictive performance with published PTFs. All the functions were fitted based on the French forest soil dataset for the European demonstration Biosoil project. The two GBM models were Model G3 which involved the three most frequent quantitative predictors used to estimate soil bulk density (organic carbon, clay and silt), and Model G10, which included ten qualitative and quantitative input variables such as parent material or tree species. Based on the full dataset, Models G3 and G10 gave R² values of 0.45 and 0.86, respectively. Model G3 did not significantly outperform the best published model. Even when fitted from an external dataset, it explained only 29% of the variation of ρ with a root mean square error of 0.244 g/cm³. In contrast, the more complex Model G10 outperformed the other models during external validation, with a R² of 0.67 and a predictive deviation of ±0.168 g/cm³. The variation in forest soil bulk densities was mainly explained by five input variables: organic carbon content, tree species, the coarse fragment content, parent material and sampling depth.     

Neural network models to predict cation exchange capacity in arid regions of Iran

Design and analysis of land‐use management scenarios requires detailed soil data. When such data are needed on a large scale, pedotransfer functions (PTFs) could be used to estimate different soil properties. Because existing regression‐based PTFs for estimating cation exchange capacity (CEC) do not, in general, apply well to arid areas, this study was conducted (i) to evaluate the existing models and (ii) to develop neural network‐based PTFs for predicting CEC in Aridisols of Isfahan in central Iran. As most researches have found a significant correlation between CEC and soil organic matter content (OM) and clay content, we also used these two variables for modelling of CEC. We tested several published PTFs and developed two neural network algorithms using multilayer perceptron and general regression neural networks based on a set of 170 soil samples. The data set was divided into two subsets for calibration and testing of the models. In general, the neural network‐based models provided more reliable predictions than the regression‐based PTFs.     

Relationship between hydraulic and basic properties for irrigated soils in southeast Australia

This paper examines the potential of soil maps and spatial information on basic soil properties for predicting soil hydraulic properties in the Shepparton irrigation region (SE Australia). For this purpose, the relationship between locally measured soil hydraulic properties and basic soil properties, and soil categories was analyzed. Pedotransfer functions developed for Australian soil were tested. Furthermore, association of field‐scale final infiltration rates with basic soil properties was investigated. Water‐retention properties, and in particular subsoil water‐retention properties, were significantly correlated with readily available basic soil properties. Spearman's rank correlation coefficients were particularly high for clay content, bulk density, and the sum of exchangeable cations Ca²⁺, Mg²⁺, Na⁺, and K⁺. Water‐retention properties were adequately predicted using Australian pedotransfer functions. Water‐transmission properties such the saturated conductivity and the final infiltration rate were overall poorly correlated with basic physical and chemical properties. Generally, median water‐transmission properties did not significantly change with soil groups and “within‐paddock variability” accounted for over half of the “within‐soil‐type variability” for many soil types. We concluded that it is feasible to regionalize water‐retention properties for the Shepparton irrigation region using basic physical and chemical soil properties, whereas the information on basic soil properties and from soil maps was insufficient to reliably estimate water‐transmission properties. It is demonstrated why field‐scale estimates of final infiltration rates, obtained by fitting a model for surface irrigation to field measurements of advance, depletion, and recession, may be better correlated with basic soil properties.     

Sampling Designs for Validating Digital Soil Maps: A Review

Sampling design (SD) plays a crucial role in providing reliable input for digital soil mapping (DSM) and increasing its efficiency. Sampling design, with a predetermined sample size and consideration of budget and spatial variability, is a selection procedure for identifying a set of sample locations spread over a geographical space or with a good feature space coverage. A good feature space coverage ensures accurate estimation of regression parameters, while spatial coverage contributes to effective spatial interpolation. First, we review several statistical and geometric SDs that mainly optimize the sampling pattern in a geographical space and illustrate the strengths and weaknesses of these SDs by considering spatial coverage, simplicity, accuracy, and efficiency. Furthermore, Latin hypercube sampling, which obtains a full representation of multivariate distribution in geographical space, is described in detail for its development, improvement, and application. In addition, we discuss the fuzzy k-means sampling, response surface sampling, and Kennard-Stone sampling, which optimize sampling patterns in a feature space. We then discuss some practical applications that are mainly addressed by the conditioned Latin hypercube sampling with the flexibility and feasibility of adding multiple optimization criteria. We also discuss different methods of validation, an important stage of DSM, and conclude that an independent dataset selected from the probability sampling is superior for its free model assumptions. For future work, we recommend:1) exploring SDs with both good spatial coverage and feature space coverage; 2) uncovering the real impacts of an SD on the integral DSM procedure; and 3) testing the feasibility and contribution of SDs in three-dimensional (3D) DSM with variability for multiple layers.     

Modeling of maximum dry density and optimum moisture content of stabilized soil using artificial neural networks

This study considers the use of artificial neural networks (ANNs) to predict the maximum dry density (MDD) and optimum moisture content (OMC) of soil‐stabilizer mix. Multilayer perceptron (MLP), one of the most widely used ANN architectures in the literature, is utilized to construct comprehensive and accurate models relating the MDD and OMC of stabilized soil to the properties of natural soil such as particle‐size distribution, plasticity, linear shrinkage, and the type and quantity of stabilizing additives. Five ANN models are constructed using different combinations of the input parameters. Two separate sets of ANN prediction models, one for MDD and the other for OMC, and also a combined ANN model for multiple outputs are developed using the potentially influential input parameters. Relative‐importance values of various inputs of the models are calculated to determine the significance of each of the predictor variables to MDD and OMC. Inferring the most relevant input parameters based on Garson's algorithm, modified ANN models are separately developed for MDD and OMC. The modified ANN models are utilized to introduce explicit formulations of MDD and OMC. A parametric study is also conducted to evaluate the sensitivity of MDD and OMC due to the variation of the most influencing input parameters. A comprehensive set of data including a wide range of soil types obtained from the previously published stabilization test results is used for training and testing the prediction models. The performance of ANN‐based models is subsequently analyzed and compared in detail. The results demonstrate that the accuracy of the proposed models is satisfactory as compared to the experimental results.     

Soil structure and pedotransfer functions

Accurate estimates of soil hydraulic properties from other soil characteristics using pedotransfer functions (PTFs) are in demand in many applications, and soil structural characteristics are natural candidates for improving PTFs. Soil survey provides mostly categorical data about soil structure. Many available characteristics such as bulk density, aggregate distribution, and penetration resistance reflect not only structural but also other soil properties. Our objective here is to provoke a discussion of the value of structural information in modelling water transport in soils. Two case studies are presented. Data from the US National Pedon Characterization database are used to estimate soil water retention from categorical field‐determined structural and textural classes. Regression‐tree estimates have the same accuracy as those from textural class as determined in the laboratory. Grade of structure appears to be a strong predictor of water retention at ?33 kPa and ?1500 kPa. Data from the UNSODA database are used to compare field and laboratory soil water retention. The field‐measured retention is significantly less than that measured in the laboratory for soils with a sand content of less than 50%. This could be explained by Rieu and Sposito's theory of scaling in soil structure. Our results suggest a close relationship between structure observed at the soil horizon scale and structure at finer scales affecting water retention of soil clods. Finally we indicate research needs, including (i) quantitative characterization of the field soil structure, (ii) an across‐scale modelling of soil structure to use fine‐scale data for coarse‐scale PTFs, (iii) the need to understand the effects of soil structure on the performance of various methods available to measure soil hydraulic properties, and (iv) further studies of ways to use soil–landscape relationships to estimate variations of soil hydraulic properties across large areas of land.     

The use of lysimeter data for the test of two soil–water balance models: A case study

Two soil–water balance models were tested by a comparison of simulated with measured daily rates of actual evapotranspiration, soil water storage, groundwater recharge, and capillary rise. These rates were obtained from twelve weighable lysimeters with three different soils and two different lower boundary conditions for the time period from January 1, 1996 to December 31, 1998. In that period, grass vegetation was grown on all lysimeters. These lysimeters are located in Berlin‐Dahlem, Germany. One model calculated the soil water balance using the Richards equation. The other one used a capacitance approach. Both models used the same modified Penman formula for the estimation of potential evapotranspiration and the same simple empirical vegetation model for the calculation of transpiration, interception, and evaporation. The comparisons of simulated with measured model outputs were analyzed using the modeling‐efficiency index IA and the root mean squared error RMSE. At some lysimeters, the uncalibrated application of both models led to an underestimation of cumulative and annual rates of groundwater recharge and capillary rise, despite a good simulation quality in terms of IA and RMSE. A calibration of soil‐hydraulic and vegetation parameters such as maximum rooting depth resulted in a better fit between simulated and observed cumulative and annual rates of groundwater recharge and capillary rise, but in some cases also decreased the simulation quality of both models in terms of IA and RMSE. The results of this calibration indicated that, in addition to a precise determination of the soil water‐retention functions, vegetation parameters such as rooting depth should also be observed. Without such information, the rooting depth is a calibration parameter. However, in some cases, the uncalibrated application of both models also led to an acceptable fit between measured and simulated model outputs.     

基于BP神经网络的土壤水力学参数预测

为了获取区域土壤水分和溶质运移模拟所需的土壤水力学参数,利用黄淮海平原曲周县的试验资料建立基于BP神经网络的土壤转换函数模型。本文采用土壤粒径分布、容重、有机质含量等土壤基本理化性质,来预测土壤饱和导水率Ks、饱和含水量sθ、残余含水量θr、以及van Genuchten公式参数α、n的对数形式ln(α)和ln(n),并与多元线性逐步回归方法进行比较。t检验结果表明,BP神经网络训练和预测得到的模拟值与实测值之间吻合很好,该方法具有较高的预测精度。通过对平均相对误差的比较,得出在粒径分布的基础上增加容重、有机质含量等输入项目,可以提高部分土壤水力学参数的预测精度,而有些参数的预测精度反而降低。以误差平方和为标准的比较结果表明,BP神经网络模型的预测效果总的来看要优于多元线性回归法。     

Applicability of site-specific pedotransfer functions and rosetta model for the estimation of dynamic soil hydraulic properties under different vegetation covers

Background, Aims, and Scope  During the last decades, different methods have been developed to determine soil hydraulic properties in the field and laboratory. These methodologies are frequently time-consuming and/or expensive. An indirect method, named Pedotransfer Functions (PTFs), was developed to predict soil hydraulic properties using other easily measurable soil (physical and chemical) parameters. This work evaluates the use of the PTFs included in the Rosetta model (Schaap et al. 2001) and compares them with PTFs obtained specifically for soils under two different vegetation covers. Methods  Rosetta software includes two basic types of pedotransfer functions (Class PTF and Continuous PTF), allowing the estimation of van Genuchten water retention parameters using limited (textural classes only) or more extensive (texture, bulk density and one or two water retention measurements) input data. We obtained water retention curves from undisturbed samples using the ‘sand box’ method for potentials between saturation and 20 kPa, and the pressure membrane method for potentials between 100 and 1500 kPa. Physical properties of sampled soils were used as input variables for the Rosetta model and to determine site-specific PTFs. Results  The Rosetta model accurately predicts water content at field capacity, but clearly underestimates it at saturation. Poor agreement between observed and estimated values in terms of root mean square error were obtained for the Rosetta model in comparison with specific PTFs. Discrepancies between both methods are comparable to results obtained by other authors. Conclusions  Site-specific PTFs predicted the van Genuchten parameters better than Rosetta model. Pedotransfers functions have been a useful tool to solve the water retention capacity for soils located in the southern Pyrenees, where the fine particle size and organic matter content are higher. The Rosetta model showed good predictions for the curve parameters, even though the uncertainty of the data predicted was higher than for the site-specific PTFs. Recommendations and Perspectives  The Rosetta model accurately predicts the retention curve parameters when the use is related with wide soil types; nevertheless, if we want to obtain good predictors using a homogenous soil database, specific PTFs are required. ESS-Submission Editor: Prof. Zhihong Xu, PhD (zhihong.xu@griffith.edu.au)     

GIS‐based regionalization of soil profiles with Classification and Regression Trees (CART)

A map of soil texture profiles was derived from readily available spatial data in combination with information from soil profiles using CART (classification and regression trees). The primary purpose was to provide a regionalized predictor for the vertical hydraulic conductivity profiles to be used as an input variable to an evapo‐transpiration model. In contrast to former studies, the texture of 110 soil profiles taken in the 10 km² area was not averaged vertically but the profiles were grouped according to their hydraulic properties. Therefore, it was possible to include site specific profiles, e.g. with histic or argillic horizons. Despite of small sampling quantities (110 soil profiles grouped into 8 classes) a prediction probability of 60 to 70 % was achieved in most classes. The resulting map provides valuable information for the granulometric and hydrologic characterization of the study area.<?show $6#>     

Potential of multi‐objective models for risk‐based mapping of the resilience characteristics of soils: demonstration at a national level

Policy makers rely on risk‐based maps to make informed decisions on soil protection. Producing the maps, however, can often be confounded by a lack of data or appropriate methods to extrapolate using pedotransfer functions. In this paper, we applied multi‐objective regression tree analysis to map the resistance and resilience characteristics of soils onto stress. The analysis used a machine learning technique of multiple regression tree induction that was applied to a data set on the resistance and resilience characteristics of a range of soils across Scotland. Data included both biological and physical perturbations. The response to biological stress was measured as changes in substrate mineralization over time following a transient (heat) or persistent (copper) stress. The response to physical stress was measured from the resistance and recovery of pore structure following either compaction or waterlogging. We first determined underlying relationships between soil properties and its resistance and resilience capacity. This showed that the explanatory power of such models with multiple dependent variables (multi‐objective models) for the simultaneous prediction of interdependent resilience and resistance variables was much better than a piecewise approach using multiple regression analysis. We then used GIS techniques coupled with an existing, extensive soil data set to up‐scale the results of the models with multiple dependent variables to a national level (Scotland). The resulting maps indicate areas with low, moderate and high resistance and resilience to a range of biological and physical perturbations applied to soil. More data would be required to validate the maps, but the modelling approach is shown to be extremely valuable for up‐scaling soil processes for national‐level mapping.     

Non‐invasive characterization of pore‐related and elastic properties of soils in linear Biot–Stoll theory using acoustic‐to‐seismic coupling

Soil structure, moisture content and strength have profound effects on plant growth. Traditional methods for monitoring soil condition are invasive and therefore may affect the samples of interest. We have demonstrated the potential of a non‐invasive measurement technique for the in situ monitoring of soil physical properties in the field. When soils are regarded as porous and elastic media, sub‐surface wave propagation can be indicative of the soil status. Such propagation can be initiated by airborne sound through acoustic‐to‐seismic (A–S) coupling. Measurements of near‐surface sound pressure and acoustically induced soil particle motion can be exploited to estimate the pore‐related and elastic properties of soils. We have conducted laboratory measurements on dry and wet sand and field measurements on an arable soil growing wheat using a compression driver, microphones and a laser Doppler vibrometer. The excitation levels were chosen so as to reduce the influence of soil non‐linearity while still yielding sufficient signal‐to‐noise ratios. Measured data were compared with model predictions based on wave propagation in layered homogeneous isotropic poro‐elastic media described by linear Biot‐Stoll theory. Soil properties were estimated through an optimization process minimizing the differences between the measurements and predictions. Latin hypercube sampling was adopted to ensure uniform seeding for optimization throughout the multi‐dimensional search space. The fitted soil characteristics are air permeability, porosity, P‐/S‐wave speeds (related to bulk and rigidity moduli) and a loss factor. Layer depth was also estimated for multi‐layered samples. The current work has demonstrated that soil can be characterized non‐invasively by using A–S coupling. It is also shown that field soils can be represented adequately by multiple homogeneous layers.