Pedotransfer Functions Models Research Articles

Exchangeable acidity and pedotransfer functions for the soils of Ghana

AbstractSoil exchangeable acidity (EA) is an indicator of aluminium toxicity potential in acidic soils. Predicting the distribution and dynamics of EA is needed for the identification and management of acidic soils. In this study, we used datasets of 355 pedons from across Ghana and the Cubist rule‐based algorithm to generate pedotransfer functions (PTFs) of EA. Eight soil properties (pH, organic carbon, calcium, sodium, magnesium, total exchangeable bases, cation exchange capacity [CEC] to clay ratio and soil depth) were used to predict EA. We first used the whole dataset to construct generic PTFs and then stratified the dataset based on World Reference Base‐Reference Soil Groups (WRB‐RSGs) to generate soil‐specific PTFs. Goodness‐of‐fit statistics comprising the root mean squared error (RMSE), Lin's concordance correlation coefficient () and coefficient of determination (R2) were used to evaluate the prediction accuracy and reliability of the developed PTF models on both calibration and validation datasets. The Fluvisols EA‐PTF exhibited lower performance metrics in the validation (RMSE = 0.17 cmolc kg−1, = 0.19, R2 = 0.24), whereas the EA‐PTFs for all other WRB‐RSGs and whole dataset had above‐average performance metrics in the validation (0.05 ≤ RMSE ≤ 0.97 cmolc kg−1; 0.34 ≤ ≤ 0.94; 0.52 ≤ R2 ≤ 0.96). Soil pH sufficed for predicting EA in soils with pH above 5.0, but in soils with a pH < 5.0, the levels of exchangeable bases (e.g., Na+, K+, Mg2+, Ca2+), CEC to clay ratio (CCR) and soil depth improved the prediction of EA. The developed EA‐PTFs are useful for estimating the missing values of EA in soil databases.

A framework for recalibrating pedotransfer functions using nonlinear least squares and estimating uncertainty using quantile regression

Pedotransfer functions (PTFs) have been developed for many regions to estimate values missing from soil profile databases. However, globally there are many areas without existing PTFs, and it is not advisable to use PTFs outside their domain of development due to poor performance. Further, developed PTFs often lack accompanying uncertainty estimations. To address these issues, a framework is proposed where existing equation-based PTFs are recalibrated using a nonlinear least squares (NLS) approach and validated on two regions of Canada; this process is coupled with the use of quantile regression (QR) to generate uncertainty estimates. Many PTFs have been developed to predict soil bulk density, so this variable is used as a case study to evaluate the outcome of recalibration. New coefficients are generated for existing soil bulk density PTFs, and the performance of these PTFs is validated using three case study datasets, one from the Ottawa region of Ontario and two from the province of British Columbia, Canada. The improvement of the performance of the recalibrated PTFs is evaluated using root mean square error (RMSE) and the concordance correlation coefficient (CCC). Uncertainty estimates produced using QR are communicated through the mean prediction interval (MPI) and prediction interval coverage probability (PICP) graphs. This framework produces dataset-specific PTFs with improved accuracy and minimized uncertainty, and the method can be applied to other regional datasets to improve the estimations of existing PTF model forms. The methods are most successful with large datasets and PTFs with fewer variables and minimal transformations; further, PTFs with organic carbon (OC) as one of or the sole input variable resulted in the highest accuracy.

Predicting Soil Saturated Water Conductivity Using Pedo-Transfer Functions for Rocky Mountain Forests in Northern China

Soil physicochemical properties and macropore spatial structure affect saturated hydraulic conductivity (Ks). However, due to regional differences and long measurement time, Ks is tedious to quantify. Therefore, it is of great importance to find simplified but robust methods to predict Ks. One possibility is to use pedo-transfer functions (PTFs). Along this line, stratified sampling was carried out in six typical forestlands in the rocky mountain area of Northern China. Penetration experiments and industrial CT scanning were combined to explore the distribution characteristics of regional Ks and its influencing factors. Based on this, we compared three Ks PTF models by multiple linear regression for Ks prediction. The results indicated that: (1) Ks decreased with increasing soil depth, which followed the order coniferous forest < broad-leaved forest < mixed forest, and the change range of mixed forest was greater than that of homogeneous forest. (2) Soil bulk density, water content, sand, silt, organic matter, total nitrogen, total phosphorus, and total potassium were significantly correlated with Ks (p < 0.05). In addition, stand type and soil depth had a certain impact on soil physicochemical properties that affected Ks. (3) Soil macropore structure, such as number density, length density, surface area density, and volume density, all decreased with increasing soil depth. They were all significantly positively correlated with Ks (p < 0.001). (4) The best predictability and universality for PTFs was achieved for PTFs containing bulk density, organic matter content, and total phosphorus. Only PTFs containing parameters of macropore spatial structure did not yield high predictability of Ks. The findings of this study contribute to the understanding of forest hydrological infiltration processes in rocky mountain forests in Northern China, and provide theoretical support for the prediction and management of water loss and soil erosion and the enhancement of water conservation functions.

Investigation of Hillslope Vineyard Soil Water Dynamics Using Field Measurements and Numerical Modeling

Soil heterogeneities can impact hillslope hydropedological processes (e.g., portioning between infiltration and runoff), creating a need for in-depth knowledge of processes governing water dynamics and redistribution. The presented study was conducted at the SUPREHILL Critical Zone Observatory (CZO) (hillslope vineyard) in 2021. A combination of field investigation (soil sampling and monitoring campaign) and numerical modeling with hydrological simulator HYDRUS-1D was used to explore the water dynamics in conjunction with data from a sensor network (soil water content (SWC) and soil-water potential (SWP) sensors), along the hillslope (hilltop, backslope, and footslope). Soil hydraulic properties (SHP) were estimated based on (i) pedotransfer functions (PTFs), (ii) undisturbed soil cores, and (iii) sensor network data, and tested in HYDRUS. Additionally, a model ensemble mean from HYDRUS simulations was calculated with PTFs. The highest agreement of simulated with observed SWC for 40 cm soil depth was found with the combination of laboratory and field data, with the lowest average MAE, RMSE and MAPE (0.02, 0.02, and 5.34%, respectively), and highest average R2 (0.93), while at 80 cm soil depth, PTF model ensemble performed better (MAE = 0.03, RMSE = 0.03, MAPE = 7.55%, R2 = 0.81) than other datasets. Field observations indicated that heterogeneity and spatial variability regarding soil parameters were present at the site. Over the hillslope, SWC acted in a heterogeneous manner, which was most pronounced during soil rewetting. Model results suggested that the incorporation of field data expands model performance and that the PTF model ensemble is a feasible option in the absence of laboratory data.

Modeling Soil Water Retention Under Different Pressures Using Adaptive Neuro-Fuzzy Inference System

Soil Water Retention (SWR) is an important process in drainage, surface, and groundwater partitioning, hydrological modeling, water supply for irrigation, etc. Assessment of SWR characteristics is complex and difficult to conduct spatially in varied locations. Therefore, Pedotransfer Functions (PTF) which are empirical relations with easily available physical properties are commonly used. In the present study, the evaluation of soil moisture at different suction pressure using the adaptive neuro-fuzzy inference systems (ANFIS) approach based on soil texture (percentage of gravel, sand, silt, and clay) and compare with the PTF approach. The analysis was conducted for a total of eleven sites of two adjoining commands in India. The pressure plate apparatus along with coarse and fine sieve analysis, titration, and other tests were carried out to determine SWR, texture, organic carbon, and bulk density. The comparative analysis of Nash–Sutcliffe efficiencies of the best-fitted PTF models and ANFIS model confirmed that the ANFIS model can capture all variations of soil texture across all sites with Nash–Sutcliffe efficiency of nearly 1.0 indicative of an exact match, while no single PTF-based model can be used for all the sites. Therefore, the ANFIS model can be used to model soil water retention for the central India region using easily available texture properties of soils.

Prediction of saturated hydraulic conductivity Ks of agricultural soil using pedotransfer functions

The determination of the saturated hydraulic conductivity Ks on a field scale presents a challenge in which several variables have to be considered. As there is no benchmark or reference method for the Ks determination, the suitability of each available method has to be evaluated. This study is aimed at the functional evaluation of three publicly available types of pedotransfer functions (PTFs) with different levels of utilised predictors. In total, ten PTF models were applied to the 56 data sets including the measured Ks value and the required predictors (% sand, silt and clay particles, dry bulk density, and organic matter/organic carbon content). A single agricultural field with a relatively homogenous particle size distribution was selected for the study to evaluate the ability of the PTF to reflect the variability of Ks. The correlation coefficient, coefficient of determination, mean error, and root mean square error were determined to evaluate the Ks prediction quality. The results showed a high variability in Ks within the field; the measured Ks values ranged between 10 and 1261 cm/day. Although the tested PTF models are based on a robust background of soil databases, they could not provide estimates with satisfactory accuracy unless local soil data were incorporated into the PTF development.

Predictor Selection and Machine Learning Regression Methods to Predict Saturated Hydraulic Conductivity From a Large Public Soil Database

Highlights In this study, six machine learning (ML) models were developed using a large database of soils to predict saturated hydraulic conductivity of these soils using easily measured soil characteristics. Tree-based regression models outperformed all other ML models tested. Neural networks were not suitable for predicting saturated hydraulic conductivity. Clay content, followed by bulk density, explained the highest amount of variation in the data of the predictors examined. Abstract. One of the most important soil hydraulic properties for modeling water transport in the vadose zone is saturated hydraulic conductivity. However, it is challenging to measure it in the field. Pedotransfer Functions (PTFs) are mathematical models that can predict saturated hydraulic conductivity (Ks) from easily measured soil characteristics. Though the development of PTFs for predicting Ks is not new, the tools and methods used to predict Ks are continuously evolving. Model performance depends on choosing soil features that explain the largest amount of Ks variance with the fewest input variables. In addition, the lack of interpretability in most “black box” machine learning models makes it difficult to extract practical knowledge as the machine learning process obfuscates the relationship between inputs and outputs in the PTF models. The objective of this study was to develop a set of new PTFs for predicting Ks using machine learning algorithms and a large database of over 8000 soil samples (the Florida Soil Characterization Database) while incorporating statistical methods to inform predictor selection for the model inputs. Of the machine learning (ML) models tested, random forest regression (RF) and gradient-boosted regression (GB) gave the best performances, with R2 = 0.71 and RMSE = 0.47 cm h-1 on the test data for both. Using the permutation feature importance technique, the GB and RF regression models showed similar results, where clay content described the most variation in the data, followed by bulk density. The implication of this study is that, when predicting Ks using the Florida Soil Characterization Database, priority should be given to obtaining quality data on clay content and bulk density as they are the most influential predictors for estimating Ks. Keywords: Deep learning, Gradient boosted regression, Pedotransfer functions, Random forest regression, Soil database, Soil properties.

Impacts of Soil Properties, Topography, and Environmental Features on Soil Water Holding Capacities (SWHCs) and Their Interrelationships

Soil water holding capacities (SWHCs) are among the most important factors for understanding the water cycle in forested catchments because they control available plant water that supports evapotranspiration. The direct determination of SWHCs, however, is time consuming and expensive, so many pedotransfer functions (PTFs) and digital soil mapping (DSM) models have been developed for predicting SWHCs. Thus, it is important to select the correct soil properties, topographies, and environmental features when developing a prediction model, as well as to understand the interrelationships among variables. In this study, we collected soil samples at 971 forest sites and developed PTF and DSM models for predicting three kinds of SWHCs: saturated water content (θS) and water content at pF1.8 and pF2.7 (θ1.8 and θ2.7). Important explanatory variables for SWHC prediction were selected from two variable importance analyses. Correlation matrix and sensitivity analysis based on the developed models showed that, as the matric suction changed, the soil physical and chemical properties that influence the SWHCs changed, i.e., soil structure rather than soil particle distribution at θS, coarse soil particles at θ1.8, and finer soil particle at θ2.7. In addition, organic matter had a considerable influence on all SWHCs. Among the topographic features, elevation was the most influential, and it was closely related to the geological variability of bedrock and soil properties. Aspect was highly related to vegetation, confirming that it was an important variable for DSM modeling. Information about important variables and their interrelationship can be used to strengthen PTFs and DSM models for future research.

Use of Pedotransfer Functions in the Rosetta Model to Determine Saturated Hydraulic Conductivity (Ks) of Arable Soils: A Case Study

A key parameter for the design of soil drainage and irrigation facilities and for the modelling of surface runoff and erosion phenomena in land-formed areas is the saturated hydraulic conductivity (Ks). There are many methods for determining its value. In situ and laboratory measurements are commonly regarded as the most accurate and direct methods; however, they are costly and time-consuming. Alternatives can be found in the increasingly popular models of pedotransfer functions (PTFs), which can be used for rapid determination of soil hydrophysical parameters. This study presents an analysis of the Ks values obtained from in situ measurements conducted using a double-ring infiltrometer (DRI). The measurements were conducted using a laboratory permeability meter (LPM) and were estimated using five PTFs in the Rosetta program, based on easily accessible input data, i.e., the soil type, content of various grain sizes in %, density, and water content at 2.5 and 4.2 pF, respectively. The degrees of matching between the results from the PTF models and the values obtained from the in situ and laboratory measurements were investigated based on the root-mean-square deviation (RMSD), Nash–Sutcliffe efficiency (NSE), and determination coefficient (R2). The statistical relationships between the tested variables tested were confirmed using Spearman’s rank correlation coefficient (rho). Data analysis showed that in situ measurements of Ks were only significantly correlated with the laboratory tests conducted on intact samples; the values obtained in situ were much higher. The high sensitivity of Ks to biotic and abiotic factors, especially in the upper soil horizons, did not allow for a satisfactory match between the values from the in situ measurements and those obtained from the PTFs. In contrast, the laboratory measurements, showed a significant correlation with the Ks values, as estimated by the models PTF-2 to PTF-5; the best match was found for PTF-2.

Evaluating the Splintex model for estimating the soil water retention curve for a wide range of soils

The soil water retention curve (SWRC) describes how much water can be retained and is available for plants in soil under different matric potentials. SWRC plays an essential role in the modeling of soil hydraulic processes and can be estimated with pedotransfer functions (PTFs). Splintex 2.0 is a PTF model developed with a user-friendly computer interface that can estimate soil hydraulic functions' parameters. As opposed to common PTFs that need to be calibrated using empirical data, Splintex 2.0 was based on a semi mechanistic model that translates particle-size distribution data into solid mass fractions and pore-size distribution. The objective in this work was to widely test the performance of Splintex 2.0 against the performance of its previous version (Splintex 1.0) using a database of 1355 samples from various countries. This large number of samples allowed a detailed quantification of the univariate and bivariate probability distributions of the estimated parameters in different hydrogeology, climate, and soil types. The accuracy of Splintex 2.0 was examined by means of linear correlation analysis (r) for random errors and of mean error (ME), mean absolute error (MAE), and root mean square error (RMSE) for systematic errors. The Splintex 2.0 model overcame the estimation’s discrepancy of SWRCs from its previous version and performed similarly with two other published PTF models. In particular, Splintex 2.0 performed as well as these published PTFs in estimating water content at field capacity over a wide range of soil types. In summary, because of its mechanistic nature, Splintex 2.0 can be an alternative model for estimating SWRC and field capacity in new areas where there is a lack of calibration data.

Uncertainty analysis and ensemble bias-correction method for predicting nitrate leaching in tea garden soils

Pedotransfer functions were often applied to predict the soil water contents at field capacity (FC) and permanent wilting point (PWP), which are the key parameters used in the soil nitrogen (N) biogeochemical models for simulating the nitrate (NO3--N) leaching. However, the PTF prediction uncertainty was often ignored. In addition, the uncertainty of the N model structure (soil N cycling is described with a set of equations) can also be substantial. Based on the 12 classic pedotransfer functions (PTFs) (namely Baumer, Brakensiek/Rawls, British Soil Survey Topsoil, British Soil Survey Subsoil, EPIC, Hutson, Manrique, Rawls, Campbell, Mayr/Jarvis, Rawls/Brakensiek, and Vereecken) and 2 biogeochemical models (DayCent and DeNitrification-DeComposition (DNDC)), this study evaluated the PTF prediction and model structural uncertainty in soil NO3−-N leaching modelling on a tea garden hillslope in Taihu Lake Basin, China. The ensemble mean was then applied to combine the 12 outputs of each model and the 24 outputs of both models. Finally, the linear bias-correction combined with the ensemble mean, i.e., the ensemble bias-correction (EBC), was also applied for the prediction of the leachate NO3−-N concentrations. Data on basic soil properties were used to derive the FC and PWP by using the 12 PTFs. Results showed that both the PTF prediction and model structural uncertainty were equally important in soil NO3--N leaching modelling at four slope positions. The coefficients of variation of the NO3--N concentration forecasts obtained by different PTFs, representing the PTF prediction uncertainty, were positively related to the climate factors, especially when PTFs were used in DayCent. Ensemble mean was found to produce a very large bias in the prediction of the leachate NO3--N concentrations, which is due to the prediction bias of PTFs. The EBC can substantially improve the prediction of the soil NO3--N leaching, especially when the 24 outputs of both models were combined.

Using Machine Learning for Prediction of Saturated Hydraulic Conductivity and Its Sensitivity to Soil Structural Perturbations

AbstractSaturated hydraulic conductivity (Ks) is a fundamental soil property that regulates the fate of water in soils. Its measurement, however, is cumbersome and instead pedotransfer functions (PTFs) are routinely used to estimate it. Despite much progress over the years, the performance of current generic PTFs estimating Ks remains poor. Using machine learning, high‐performance computing, and a large database of over 18,000 soils, we developed new PTFs to predict Ks. We compared the performances of four machine learning algorithms and different predictor sets. We evaluated the relative importance of soil properties in explaining Ks. PTF models based on boosted regression tree algorithm produced the best models with root‐mean‐squared log‐transformed error in ranges of 0.4 to 0.3 (log10(cm/day)). The 10th percentile particle diameter (d10) was found to be the most important predictor followed by clay content, bulk density (ρb), and organic carbon content (C). The sensitivity of Ks to soil structure was investigated using ρb and C as proxies for soil structure. An inverse relationship was observed between ρb and Ks, with the highest sensitivity at around 1.8 g/cm3 for most textural classes. Soil C showed a complex relationship with Ks with an overall positive relation for fine‐textured and midtextured soils but an inverse relation for coarse‐textured soils. This study sought to maximize the extraction of information from a large database to develop generic machine learning‐based PTFs for estimating Ks. Models developed here have been made publicly available and can be readily used to predict Ks.

Estimating Soil Thermal Diffusivity Using Pedotransfer Functions with Nonlinear Regression

Background and Objective:Pedotransfer Functions (PTFs) are widely used for estimating soil thermal diffusivity. Some attempts have been made to indirectly predict soil thermal diffusivity from the easy available fundamental soil physics properties. The aim of the work was to validate usage PTFs with Nonlinear Regression (NLR) for estimating soil thermal diffusivity (KD), moreover was to select the best predictor variables used for determination of PTFs.Materials and Methods:Soil thermal diffusivity was measured at different values of water content using Kondratieff method. The parameters of the quadratic equation, which described the relation between thermal diffusivity and water content, were determined by the fitting curve and using PTFs (exponential equations) based on soil physical properties. The Combination of different soil physical properties used as PTF model’s independent variables was tested. Three classes of PTFs were proposed using NLR to estimate KDwere: KDPTF-1 (Sand+ Silt+ Clay), KDPTF-2 (Sand+ Silt+ Clay + Bulk density), and KDPTF-3 (Sand+ Silt+ Clay+ Bulk density + Organic matter).Results:The best class of PTF could be used for calculating the parameters of the quadratic equation and soil thermal diffusivity, was KDPTF-1 which taking into account the percentage of sand, silt and clay, RMSE=2.94×10-8m2/s, and GMER =1.05.Conclusion:The quadratic and exponential equations were representing the nonlinear regression equations, which could be used for estimating soil thermal diffusivity at different values of water content from easily available data on soil texture, bulk density, and organic matter content.

Modeling WEPP erodibility parameters in calcareous soils in northwest Iran

Modeling soil detachment rates at the regional scale is important for better understanding of the processes of erosion and the development of erosion models. Soil erodibility is an important factor for predicting soil loss, but its direct measurement at the watershed scale is difficult, time-consuming and costly. This study used stepwise multiple-linear regression (MLR) and artificial neural networks (ANNs) to model Water Erosion Prediction Project (WEPP) soil erodibility parameters, including the baseline inter-rill erodibility (Kib), baseline rill erodibility (Krb) and critical shear stress (τcb) of cropland conditions in calcareous soils of northwest Iran. Simulated inter-rill and rill erosion experiments were conducted at 100 locations with three replications. Kib, Krb and τcb and basic soil properties were measured at each location. Auxiliary variables related to soil erodibility were derived from a Landsat 7 satellite image and a 30m×30m digital elevation model (DEM). MLR and ANN models were employed to predict Kib, Krb and τcb using two groups of input variables: i) more easily measurable basic soil properties (pedo-transfer functions (PTFs)) and ii) more easily measurable basic soil properties and auxiliary data (soil spatial prediction functions (SSPFs)). The results indicated that the WEPP models performed poorly in comparison to the derived models. PTFs and SSPFs generated from ANN models provided more reliable predictions than the MLR models. ANN-based SSPF models yielded the best results (with the highest R2 and lowest RMSE values) for predicting Kib and Krb. ANN-based PTF model performed reasonably well for predicting τcb. These results show that information from terrain attributes and remote sensing data are potential auxiliary variables for improving prediction of soil erodibility parameters.

Assessment of the usefulness of particle size distribution measured by laser diffraction for soil water retention modelling

AbstractParticle size distribution (PSD) is a major soil characteristic, which is essential and commonly used for the development of pedotransfer functions (PTFs) to estimate the water retention of soils. The laser diffraction method (LDM) became a popular alternative to the standard sieve‐hydrometer method (SHM) of PSD measurement. Unfortunately, PSDs determined with LDM and SHM methods differ sometimes substantially. Moreover, it is claimed that the laser diffraction method underestimates finer fractions in favor of coarser fractions. Several authors have tried to elaborate on methods to recalculate LDM PSD into its SHM counterparts, but no universal methodology has been developed to this date. In this paper, we use PSD determined by LDM directly for PTF development and compare it with the classical PTF approach based on PSD measured by SHM. Four different PTF models based on LDM particle size distribution data were developed, with different PSD characteristics taken as the models' input variables. The possibility of using alternative PSD characteristics, such as deciles, area moment mean and volume moment mean, for PTF development was examined. The accuracy of PTF models constructed on the basis of LDM‐measured PSD was comparable with that of the developed models using texture data obtained from SHM, giving approximately the same RMSE and R2 values. Our study shows that LDM‐measured particle size distribution may be directly used for PTF developments without any recalculations to their sieve‐hydrometer counterparts.

Using Two Pedotransfer Functions to Estimate Soil Moisture Retention Curves From One Experimental Site of South Moravia

The soil moisture retention curves (SMRCs) were measured in laboratory conditions. The sand table and pressure extractor method were used to obtain a 9-point SMRCs for undisturbed soil samples taken from 20 cm depth. The data points of the individual retention curves were parameterized using the RETC computer program. For the same soil, the SMRCs were estimated by two models of pedotransfer functions (PTFs). In the first part of the study the Rosetta program with the model of artificial neural network (Schaap et al., 1998) was used. The PTFs derived by Wösten et al. (1998) were used in the second part of the study. The reliability of selected PTFs models were tested on local soils from one site of the South Moravia. The quality of estimated retention curves was evaluated using the coefficient for correlation R between the measured and the estimated soil water content at several pF values and the root mean squared error RMSE. Of the two models of PTFs, which have been tested in the study, Wösten’s model showed better agreement with the measured retention curves (RMSESch = 0.0383 (cm3.cm−3), RMSEW = 0.0264 (cm3.cm−3)).

Using different multimodel ensemble approaches to simulate soil moisture in a forest site with six traditional pedotransfer functions

Pedotransfer functions (PTFs) have routinely been used to estimate the soil hydraulic properties (SHPs) from easily measurable soil properties, such as particle-size distribution, organic matter content and bulk density. However, different PTFs often yielded different prediction results. In order to deal with the PTF selection problem, this study used multimodel ensemble approaches to simulate forest soil moisture based on the modelling results of different PTFs. A total of 300 days of observed soil moisture data at four depths (10-, 20-, 40- and 60-cm) were adopted to calibrate the Richards equation and obtain the SHPs by using the inverse option in HYDRUS-1D. Six published PTFs were selected to predict the SHPs, which were used to predict soil moisture temporal variations at these four different depths. Two multimodel ensemble methods, including the simple model average (SMA) and the multiple linear regression (MLR)-based superensemble, were used in this study. Under different selections of training periods (i.e. 50, 100 and 150 days), performances of these multimodel ensemble approaches were compared with those of the best single PTF model. The SMA always had worse performance than the best single model. However, the performances of the superensemble approach were better than those of the best single model, and even comparable to those of the calibrated soil water flow model. Results show that given the relatively long training period (>50 days), it is worthwhile to consider the superensemble method to simulate soil moisture contents in forestland.

Predicting Soil Bulk Density Using Advanced Pedotransfer Functions in an Arid Environment

<abstract><title><italic>Abstract.</italic></title> Soil bulk density (ρ_b) measurement is essential for many soil management and engineering applications in addition to the emerging need to estimate soil carbon reserves. Recently, pedotransfer functions (PTFs) have been gaining widespread recognition for their capability to predict soil bulk density using extractable soil available databases. However, these functions are site specific and might be hampered when used in different environments. The objective of this study is to develop and compare PTFs for estimating soil ρ_b in an arid environment from texture fractions and soil organic carbon (SOC). The examined PTFs were simple linear regression (SLR), multiple non-linear regression (NLR), stepwise multiple linear regression (SWR), partial least squares (PLS), and artificial neural networks (ANNs). A large data set of soil samples procured from the Amman-Zarqa basin, a typical arid environment in north Jordan, were used to calibrate and validate the PTFs. Soil organic carbon and sand contents had moderate correlations with soil ρ_b, and both were almost interchangeable predictors of soil ρ_b. Moreover, it has been found that increasing the number of predictor variables in the PTF models enhanced the prediction capabilities of soil ρ_b. Furthermore, ANNs had high capabilities in predicting soil ρ_b. As compared to SLR models, the ANNs were 57.5% less unbiased, 85% less precise, and 18.4% more accurate in predicting soil ρ_b in arid environments.

Functional evaluation of PTF prediction uncertainty: An application at hillslope scale

This study presents a methodology to assess uncertainties resulting from the use of pedotransfer functions (PTFs) when soil water budget is modelled at a hillslope scale. Two sources of uncertainty are examined: (i) errors in the assessment of the soil physical and chemical properties at unsampled locations, as related to the spatial spacing of the sample measurements across the entire study area, and (ii) errors associated to the PTF relations and parameters. This methodology is applied to an experimental hillslope in Southern Italy, where an intensive field campaign has been conducted to gather several soil physical and chemical properties and soil hydraulic properties. A sequential Gaussian simulation algorithm is used to generate multiple equally probable images of PTF input soil data, consistent with the estimated spatial structure and conditioned to the measured soil core properties. Two PTFs commonly used in Europe, Vereecken's PTF [Vereecken, H., Diels, J., van Orshoven, J., Feyen, J., Bouma, J., 1992. Functional evaluation of pedotransfer functions for the estimation of soil hydraulic properties. Soil Sci. Soc. Am. J. 56, 1371–1378] and HYPRES PTF [Wösten, J.H.M., Lilly, A., Nemes, A., Le Bas, C., 1999. Development and use of a database of hydraulic properties of European soils. Geoderma 90, 169–185] are applied to predict the soil hydraulic properties, which in turn are employed into a soil–vegetation–atmosphere model to evaluate the propagated uncertainty in simulated target variables, such as evaporation, transpiration and root zone soil water content variation during a wet-to-dry transition season. The outcomes of this case study suggest that accurate estimates of transpiration and soil water storage variation at the hillslope scale are obtained even when PTF input data are collected with a relatively coarse sampling strategy. The examined PTFs show worse level of performance with respect to the simulated evaporation. The simulated evaporation is much more affected by the PTF model error than by the input data error. A major implication of these results is that if one would reduce the prediction uncertainty in simulated evaporation, the PTF model structure has to be improved prior of reducing the uncertainty into the PTF input data.

TESTING SOME PEDO-TRANSFER FUNCTIONS (PTFS) IN APULIA REGION

The knowledge of soil water retention vs. soil water matric potential is used to study irrigation and drainage schedules, soil water storage capacity (plant available water), solute movement, plant growth and water stress. The hydraulic soil properties measuring is expensive, laborious and takes too long time, so, frequently, matemathic models, called pedo-transfer functions (PTFs) are utilized to estimate hydraulic soil properties through soil chimical and phisical characteristics. Six pedo-transfer functions have been evaluated (Gupta &amp;amp; Larson, 1979; Rawls et al., 1982; De Jong et al., 1983; Rawls &amp;amp; Brakensiek, 1985; Saxton et al., 1986; Vereecken et al., 1989) by comparing estimated with measured soil moisture values at soil water matric potential of –33 and –1500 kPa of 361 soil samples collected from 185 pedons of Apulia Region (South Italy), having various combinations of particle-size distribution, soil organic matter content and bulk density. Accuracy of the soil moisture predictions have been evaluated by statistic indexes such as Weighted stantard error (WSEE), Mean Deviation (MD), Root Mean Squared Deviation (RMSD) and the determination coefficient (R2) between estimated and measured water retention values. The Rawls PTF model demostrated to have the lowest values of WSEE, MD and RMSD indexes (0.044, -0.007 and 0.059 m3 H2O m-3 soil, respectively) at –33 Kpa soil water matric potential (Field Capacity), while for estimating soil moisture at the Wilting Point (-1500 kPa) Rawls &amp;amp; Brakensiek model is adequate (WSEE, MD and RMSD of 0.034, -0.016 and 0.046 m3 H2O m-3 soil). De Jong, Saxton and Rawls &amp;amp; Brakensiek models, at –33 kPa soil water matric potential and Gupta &amp;amp; Larson and De Jong models at –1500 kPa soil water matric potential, showed the highest statistic errors.