Journal Paper Digests 2019 #5
- Evaluation of Parametric and Nonparametric Machine-Learning Techniques for Prediction of Saturated and Near-Saturated Hydraulic Conductivity
- Delineating Site-Specific Management Zones and Evaluating Soil Water Temporal Dynamics in a Farmer’s Field in Kentucky
- Evidences of soil geochemistry and mineralogy changes caused by eucalyptus rhizosphere
Evaluation of Parametric and Nonparametric Machine-Learning Techniques for Prediction of Saturated and Near-Saturated Hydraulic Conductivity
Authors: Kotlar, AM; Iversen, BV; Van Lier, QD
Source: VADOSE ZONE JOURNAL, 18 (1):80141-80141; FEB 21 2019
Abstract: Parametric and nonparametric supervised machine learning techniques were used to estimate saturated and near-saturated hydraulic conductivities (K-s and K-10, respectively) from easily measurable soil properties including the name of the pedological horizon (HOR), soil texture (sand, silt, and clay), organic matter (OM), bulk density (BD), and water contents (theta(pF1),theta(pF2),theta(pF3), and theta(pF4.2)) measured at four different matric heads (-10, -100, -1000, and -8,848 cm, respectively). Using a stepwise linear model (SWLM) and the Lasso regression as parametric methods with 316 data in training and 135 data in the testing phase, four pedotransfer functions (PTFs) were obtained in which water contents for both methods play an important role compared with other variables. The SWLM showed better performance than Lasso in the testing phase for log(K-s) and log(K-10) prediction, with RMSE values of 0.666 and 0.551 cm d(-1) and R-2 of 0.26 and 0.65. Nonparametric supervi sed machine learning methods trained and tested with a similar data set significantly improved the accuracy of K-s prediction, with R-2 of 0.52, 0.36, and 0.53 for Gaussian process regression (GPR), support vector machine (SVM), and ensemble (ENS) methods in the testing stage. These methods also described 74.9, 66.7, and 72.5% of the variation of log(K-10). Bootstrapping validated the strong performance of nonparametric techniques. The feature selection capability of GPR determined that instead of using a model with all predictors, HOR, silt, theta(pFl), and theta(pF3) are sufficient for the prediction of log(K-s), while HOR, silt, and OM can predict log(K-10) as accurate as the comprehensive model with all variables.
Delineating Site-Specific Management Zones and Evaluating Soil Water Temporal Dynamics in a Farmer’s Field in Kentucky
Authors: Reyes, J; Wendroth, O; Matocha, C; Zhu, JF
Source: VADOSE ZONE JOURNAL, 18 (1):80143-80143; FEB 21 2019
Abstract: Due to spatial variability of soil genesis, topography, and resulting soil properties in farmers’ fields, soil and crop processes vary in space and time. Therefore, optimum rates and timing of resource applications, such as nutrients and irrigation water, may vary as well. It remains a challenge to quantify the spatial variability of a field and to identify effective ways to manage fields in a site-specific manner. The objective of this study was to delineate management zones within a farmer’s field based on relatively easily obtainable information that is statistically integrated. Moreover, soil water temporal dynamics should be evaluated regarding their spatial differences in different zones. The set of direct and indirect observations included clay and silt content, apparent electrical conductivity, soil chemical properties (pH; organic matter; and total N, P, K, Ca, Mg, and Zn), satellite-based normalized difference vegetation index (NDVI), and lidar-based topographic var iables in a western Kentucky field. Several key variables and their capability to describe spatial crop yield variability were identified by using principal component analysis: soil clay content, slope, soil organic matter content, topographic wetness index, and NDVI. Two types of cluster analysis were applied to delineate management zones. The cluster analyses revealed that two to three zones was the optimal number of classes based on different criteria. Delineated zones were evaluated and revealed significant differences in corn (Zea mays L.) yield and temporally different soil moisture dynamics. The results demonstrate the ability of the proposed procedure to delineate a farmer’s field into zones based on spatially varying soil and crop properties that should be considered for irrigation management.
Evidences of soil geochemistry and mineralogy changes caused by eucalyptus rhizosphere
Authors: Korchagin, J; Bortoluzzi, EC; Moterle, DF; Petry, C; Caner, L
Source: CATENA, 175 132-143; APR 2019
Abstract: Eucalyptus trees grow in a variety of environmental conditions and cause contrasting effects on soils. Changes in soil geochemistry and mineralogy due to eucalyptus have not been clearly established to date. The objective of this study was to identify evidences of the effects of eucalyptus root system on geochemistry and clay mineralogy of a Ferralsol. Soil samples were collected in different positions of an old eucalyptus site (R1: rhizosphere of small roots; R2: coarse roots, and R3: bulk soil). The soil pH was very acid (3.8) and Al3+ (4.7 cmol(c) kg(-1)) and H+ Al3+ (40.2 cmol(c) kg(-1)) were higher in R1 compared to R3. Additionally, reduction of clay fraction content and an increase in the proportion of Al-interlayered clay minerals were also observed in the eucalyptus rhizosphere. This suggests the occlusion of Al3+ in the interlayer space of 2:1 clay minerals and explains the tendency of a low cation exchange capacity of the day fraction. Then, the acidification due t o eucalyptus did not alter the composition of the mineral assemblage (mineral species) but the trend was a decrease of the proportion of the fine clay fraction (< 0.1 mu m) as well as an increase of the proportion of 2:1 Al-interlayered clay mineral in the clay fraction. The results suggest that when alien tree species are introduced into natural lands worldwide, a comprehensive geochemical-mineralogical approach must be applied in order to provide the extension of constituent changes.