Apparent Electrical Conductivity Research Articles

Enhancing Irrigation Management: Unsupervised Machine Learning coupled with Geophysical and Multispectral Data for Informed Decision-Making in Rice Production

Integrating diverse data sources for site-specific management zones (SSMZ) in precision agriculture is a complex task. Soil surveys using apparent electrical conductivity (ECa) have proven effective in capturing field variability. However, relying solely on one sensing data type may not fully capture the intricate relationship between soil properties and crop physiological processes. This research assessed the effectiveness of applying multivariate geostatistical analysis and unsupervised machine learning (UML) to geophysical and multispectral data through ECa, NDWI and NDVI indices, for delineating and validating the SSMZ at different crop cycles in five rice field of Tolima department-Colombia. Lee´s correlation was conducted between ECa, NDWI and NDVI data. MULTISPATI-PCA was used to assess the autocorrelation in the input variables before SSMZ delineation. The correlation results revealed two field categories based on the direct or indirect relationship between ECa and the multispectral data. On one hand, ECa values showed an inverse correlation (p<0.01) with multispectral data (-0.25 to -0.45) for fields 1a, 1b and 4a. On the other hand, there was a direct correlation (p<0.01) between ECa and multispectral data (0.31 to 0.51) for fields 2b and 5a. For all rice fields, MULTISPATI-PCA explained a range between 75.01 % to 95.64 % of the total variability. UML results highlighted two, three and four SSMZ, with two SSMZ being the optimal management zones number. The validation suggests that the θ data showed a significant statistical difference for two SSMZs (p < 0.05), regardless of the water content ranges. This study emphasized the importance of integrating geophysical and multispectral datasets in precision agriculture for evaluating field variability and crop conditions.

Using Soil Sensors to Assess Soil Salinity

AbstractThis article is reproduced from Chapter 6 of the new book, Salinity and Sodicity: A Growing Global Challenge to Food Security, Environmental Quality and Soil Resilience, in a slightly modified form. The book is available for purchase from Wiley.com or Amazon.com. The article covers the use of electromagnetic (EM) sensors for measuring apparent electrical conductivity (ECa). In the field, electrical conductivity (EC) can be measured by two primary approaches: physical contact and electromagnetic (EM) induction. With physical contact, a current is injected into the soil, and the detector measures the resulting voltage. An EM meter does not make direct contact but uses a coil to produce an EM field. A sensor then measures the soil‐induced changes to the original EM field. Both types of sensors measure the ECa, which is different from laboratory‐derived EC values. When using ECa sensors, it is important to remember that they are sensitive to many factors, including salinity, soil moisture, bulk density, soil texture, and temperature. The purpose of this article is to provide an overview on the use of EM sensors to provide examples on the use of these sensors in the field. Earn 1.5 CEUs in Soil &amp; Water Management by reading this article and taking the quiz at https://web.sciencesocieties.org/Learning‐Center/Courses.

Integrating proximal geophysical sensing and machine learning for digital soil mapping: Spatial prediction and model evaluation using a small dataset

Geophysical methods support soil security by providing non-invasive tools to assess soil properties, monitor degradation, and guide sustainable management strategies. However, studies focusing the spatial prediction of geophysical data remain limited. In this research, we aimed to model and predict the spatial distribution of soil geophysical properties using parent material and terrain attributes with machine learning algorithms. In addition, we tested the nested leave-one-out cross validation (nested-LOOCV) method to deal with datasets with limited size. We performed a geophysical survey using three types of sensors (radiometric, magnetic and electric methods). The random forest (RF) and support vector machine (SVM) algorithms presented the best results, with RF showing higher performance for K40 and magnetic susceptibility, and SVM had higher performance for eU, eTh and apparent electrical conductivity. Parent materials and digital elevation model were the most significant variables for the modelling. The nested-LOOCV method proved to be adequate for small soil dataset. Machine learning techniques are potential tools for modelling soil geophysical variables. The combination with computational techniques shows the great relevance of geophysical measurements for the estimation of soil properties related to fertility and soil genesis.

On-farm experimentation of precision agriculture for differential seed and fertilizer management in semi-arid rainfed zones

Abstract Introduction This study explores the integration of precision agriculture technologies (PATs) in rainfed cereal production within semi-arid regions. Methods utilizing the Veris 3100 sensor for apparent soil electrical conductivity (ECa) mapping, differentiated management zones (MZs) were established in experimental plots in Valsalada, NE Spain. Site-specific variable dose technology was applied for seed and fertilizer applications, tailoring inputs to distinct fertility levels within each MZ. Emphasizing nitrogen (N) management, the study evaluated the impact of variable-rate applications on crop growth, yield, nitrogen use efficiency (NUE), and economic returns. For the 2021/2022 and 2022/2023 seasons, seeding rates ranged from 350 to 450 grains/m2, and basal fertilizer dosages varied between high and low levels. Additionally, the total nitrogen units were distributed differently between the two seasons, while maintaining a uniform topdressing fertilizer dose across all treatments. Results Results revealed a significant increase in yield in MZ 2 (higher fertility) compared to MZ 1 (lower fertility). NUE demonstrated notable improvement in MZ 2, emphasizing the effectiveness of variable-rate N applications. Economic returns, calculated as partial net income, showed a considerable advantage in MZ 2 over MZ 1, resulting in negative outcomes for low-fertility areas in several of the analyzed scenarios, and highlighting the financial benefits of tailored input management. Conclusion This research provides quantitative evidence supporting the viability and advantages of adopting PATs in rainfed cereal production. The study contributes valuable insights into optimizing input strategies, enhancing N management, and improving economic returns in semi-arid regions.

Relevance of NDVI, soil apparent electrical conductivity and topography for variable rate irrigation zoning in an olive grove

Relevance of NDVI, soil apparent electrical conductivity and topography for variable rate irrigation zoning in an olive grove

Evaluation of a Multivariate Calibration Model for the WET Sensor That Incorporates Apparent Dielectric Permittivity and Bulk Soil Electrical Conductivity

The measurement of apparent dielectric permittivity (εs) by low-frequency capacitance sensors and its conversion to the volumetric water content of soil (θ) through a factory calibration is a valuable tool in precision irrigation. Under certain soil conditions, however, εs readings are substantially affected by the bulk soil electrical conductivity (ECb) variability, which is omitted in default calibration, leading to inaccurate θ estimations. This poses a challenge to the reliability of the capacitance sensors that require soil-specific calibrations, considering the ECb impact to ensure the accuracy in θ measurements. In this work, a multivariate calibration equation (multivariate) incorporating both εs and ECb for the determination of θ by the capacitance WET sensor (Delta-T Devices Ltd., Cambridge, UK) is examined. The experiments were conducted in the laboratory using the WET sensor, which measured θ, εs, and ECb simultaneously over a range of soil types with a predetermined actual volumetric water content value (θm) ranging from θ = 0 to saturation, which were obtained by wetting the soils with four water solutions of different electrical conductivities (ECi). The multivariate model’s performance was evaluated against the univariate CAL and the manufacturer’s (Manuf) calibration methods with the Root Mean Square Error (RMSE). According to the results, the multivariate model provided the most accurate θ estimations, (RMSE ≤ 0.022 m3m−3) compared to CAL (RMSE ≤ 0.027 m3m−3) and Manuf (RMSE ≤ 0.042 m3m−3), across all the examined soils. This study validates the effects of ECb on θ for the WET and recommends the multivariate approach for improving the capacitance sensors’ accuracy in soil moisture measurements.

Remote and Proximal Sensors Data Fusion: Digital Twins in Irrigation Management Zoning.

The scientific field of precision agriculture employs increasingly innovative techniques to optimize inputs, maximize profitability, and reduce environmental impact. However, obtaining a high number of soil samples is challenging in order to make precision agriculture viable. There is a trade-off between the amount of data needed and the time and resources spent to obtain these data compared to the accuracy of the maps produced with more or fewer points. In the present study, the research was based on an exhaustive dataset of apparent electrical conductivity (aEC) containing 3906 points distributed along 26 transects with spacing between each of up to 40 m, measured by the proximal soil sensor EM38-MK2, for a grain-producing area of 72 ha in São Paulo, Brazil. A second sparse dataset was simulated, showing only four transects with a 400 m distance and, in the end, only 162 aEC points. The aEC map via ordinary kriging (OK) from the grid with 26 transects was considered the reference, and two other mapping approaches were used to map aEC via sparse grid: kriging with external drift (KED) and geographically weighted regression (GWR). These last two methods allow the increment of auxiliary variables, such as those obtained by remote sensors that present spatial resolution compatible with the pivot scale, such as data from the Landsat-8, Aster, and Sentinel-2 satellites, as well as ten terrain covariates derived from the Alos Palsar digital elevation model. The KED method, when used with the sparse dataset, showed a relatively good fit to the aEC data (R2 = 0.78), with moderate prediction accuracy (MAE = 1.26, RMSE = 1.62) and reasonable predictability (RPD = 1.76), outperforming the GWR method, which had the weakest performance (R2 = 0.57, MAE = 1.78, RMSE = 2.30, RPD = 0.81). The reference aEC map using the exhaustive dataset and OK showed the highest accuracy with an R2 of 0.97, no systematic bias (ME = 0), and excellent precision (RMSE = 0.56, RPD = 5.86). Management zones (MZs) derived from these maps were validated using soil texture data from clay samples measured at 0-10 cm depth in a grid of 72 points. The KED method demonstrated the highest potential for accurately defining MZs for irrigation, producing a map that closely resembled the reference MZ map, thereby providing reliable guidance for irrigation management.

Atmospheric modulation of apparent electrical conductivity in a metal−organic framework

Combining high surface area and efficient charge transport, electrically conductive metal−organic frameworks (MOFs) find wide applications in energy storage, sensing, and electrocatalysis. Reliable characterization of electrical conductivity, the key metric for assessing this class of materials, remains challenging due to its high sensitivity to the atmosphere. Herein, through electrical characterization of an exemplary MOF, Cd2(TTFTB) (TTFTB4− = tetrathiafulvalene tetrabenzoate), under various controlled atmospheres, we show that adsorption of water in humid air or N2 improves the apparent room-temperature electrical conductivity by one to two orders of magnitude compared to the values observed in dry atmospheres. This observation in conjunction with spectroscopic characterization, structural analysis, and band structure calculations indicates significant contribution of water-mediated proton conductivity and/or proton-electron coupling to the apparent electrical conductivity. Thus, controlling and reporting atmospheres in electrical conductivity measurements of MOFs is critical to improve their reproducibility and to gain insights into electrical conduction mechanisms.

Growth of α-Al2O3 layer involving abnormal oxides in FeCrAl alloy tube fabricated by WEDM process and electrical insulating performance in fusion reactor blanket

The tubular specimens of FeCrAl alloy APMT (Fe-22Cr-5Al-3Mo) were fabricated by a wire electrical discharge machining (WEDM) process. The oxidation characteristics of the machined surface were investigated by means of the oxidation tests performed in air atmosphere at the temperatures of 1273 K and 1373 K up to 100 h. The abnormal oxides such as CuAl2O4 and ZnAl2O4 were grew on the recast layer which was locally remained on the machined surface even after the mechanical polishing procedure. The α-Al2O3 layer involving the abnormal oxides was formed on the machined surface in the oxidation treatment at 1273 K for 10 h. The apparent electrical conductivity of the layer involving the abnormal oxides was larger than that without the abnormal oxides. Thick Al-rich oxide layer was formed under the abnormal oxides when the oxidation treatment was performed at 1373 K. The apparent electrical conductivity of the oxide layer became lower due to the presence of the thick Al-rich oxide. This oxide layer formed at 1373 K could function as an electrical insulating layer to reduce the magnetohydrodynamic pressure drop in liquid breeder blanket systems of magnetic confinement fusion reactors.

Forecasting Oligochaeta Abundance using the Apparent Electrical Conductivity of Irrigated Soils with Reclaimed Wastewater: Evidence from Corso, Algeria

Background: The main objective of this paper is to study the possibility of using the measurement of apparent electrical conductivity (ECa) to estimate Oligochaeta (earthworm) density in irrigated soils treated with reclaimed wastewater. Additionally, the study sought to analyze the impact of using reclaimed wastewater for irrigation on earthworm density in the study area. Methods: Earthworm sampling was carried out at thirty emplacements within the study area during two measurement campaigns. Soil ECa was measured by (EM38). Result: The results showed that irrigation with reclaimed wastewater promotes the abundance of earthworms (earthworm density). Indeed, sites irrigated with reclaimed wastewater have higher earthworm densities compared to sites irrigated with conventional water. The results also show a strong correlation between ECa and earthworm density (R2 greater than 0.7). The Wilcoxon test indicates a non-significant difference between measured earthworm densities and those predicted by ECa (p greater than 0.5). This result confirms the effectiveness of the EMI technique as a tool to evaluate the spatial variability of soil biological parameters. These results collectively suggest that ECa represents a real possibility of reducing the number of earthworm samplings, whil ensuring the reliability of the estimates of real values of earthworm density and their surface spatial distribution. ECa could therefore be a useful tool to spatialize and predict the presence of earthworms in soil irrigated by reclaimed wastewater.

Leafiness-LiDAR index and NDVI for identification of temporal patterns in super-intensive almond orchards as response to different management strategies

The use of super-intensive orchards is a growing trend in fruit production. The present study aims to improve management of these cropping systems by focusing on how agronomic decisions impact orchard dynamics in the short to medium term and by providing a decision-support approach based on stable temporal patterns from previous seasons. A multitemporal study using remote sensing and LiDAR was conducted in a commercial almond orchard over four growing seasons (2019–2022) to determine the optimal timing of image acquisition for variable pre-harvest treatments. A model-based clustering (mclust) was applied to optimal Sentinel-2 NDVI maps and apparent soil electrical conductivity (ECa) data, interpolated to the pixel centroids of Sentinel-2 image grids, to delineate potential management zones (PMZs). The leafiness-LiDAR index (LLI), a leaf area index (LAI) estimator, was obtained as ground truth after summer pruning and before harvesting, showing a significant influence of fertigation and pruning on the LAI, with summer pruning particularly influencing orchard dynamics. The optimal time for NDVI mapping was found to be two months after summer pruning in productive years and two weeks after in unproductive years. The delineated PMZs were consistent across seasons and corresponded to significant LAI differences. This method could contribute to improving resource management and sustainability in super-intensive commercial orchards.

Oxygen transport mechanism and the effect of microstructure on the oxygen response behavior of an oxygen-sensitive composite Ce0.9Gd0.1O2-ẟ – La0.6Sr0.4FeO3-ẟ at high temperature

Compositions like gadolinium doped cerium oxide Ce0.9Gd0.1O2-ẟ and strontium doped lanthanum ferrite La0.6Sr0.4FeO3-ẟ, or their composites are well known in the field of fuel cell electrolyte, oxygen transporting membrane and catalytic reactor applications. This work offers insight into composite material's oxygen response characteristics, taking Ce0.9Gd0.1O2-ẟ and La0.6Sr0.4FeO3-ẟ in 70: 30 wt percent. One pot sol-gel method was employed for the powder synthesis. The powder was characterized by TG-DSC, XRD, Raman spectroscopy, and X-ray photoelectron spectroscopy (XPS). FESEM-EDX and EDX-line profiling techniques were used to check the compositional homogeneity of sintered bodies. Oxygen response tests were conducted from ambient (20 % O2) to 100 % O2 concentration, using bulk samples sintered at three different temperatures, i.e., 1000, 1150, and 1250 °C at a fixed soaking time of 4 hours. The sintered specimens' apparent porosity, bulk density, surface morphology, and electrical conductivity were also examined. The sintering temperature significantly impacts response behavior as the porosity and grain size of the sintered body plays a vital role in sensing performance. The mechanism of response towards oxygen is also discussed. Sample 73–100 displayed the best performance among others. We achieved the lowest response and recovery time of 15 – 48 sec and 45 – 80 sec from a 0.60 mm thick bulk material at 950 °C. Harsh environments were intentionally created to check the reliability and reproducibility of the response data. This material could be a good candidate as an oxygen gas sensor at high temperatures owing to its good sensitivity, reliability, reversibility, and stability.

Assessing salinity and sodicity of irrigated soils using apparent electrical conductivity in the Pampean region

Assessing salinity and sodicity of irrigated soils using apparent electrical conductivity in the Pampean region

Spatial-temporal variability in nitrogen use efficiency: Insights from a long-term experiment and crop simulation modeling to support site specific nitrogen management

Within-field soil heterogeneity can lead to large variation in nitrogen use efficiency (NUE). Crop simulation models provide a multi-faceted approach to management considering both soil and plant interactions. However, research using crop models for investigating within field variation in NUE is limited, in part because of challenges quantifying spatially variable soil model parameters. Here soil apparent electrical conductivity (ECa) and measured soil properties were used to map spatial variations in soil characteristics across a Long-Term Experiment in Norfolk, England. The relationship between plot ECa across the 3 ha experiment and agronomic data across three different nitrogen rates (0, 110, and 220 kg N ha-1) over five wheat years (2010–2020) was quantified. The Sirius crop model was parameterized for two soils representing the extremes of ECa. Sirius was validated using recorded plot data. Site-specific optimal nitrogen and associated leaching risks were simulated across 29 years of weather data. Variation in soil properties had significant impact on measured NUE. At 220 kg N ha-1 mean observed yields across 5 years ranged from 9.0 to 10.7 t ha-1 and grain protein from 11.6% to 11% on the low EC and high EC plots, respectively. On average fertiliser grain N recovery was 19.7 kg N ha-1 lower on the low ECa plots. Sirius simulated the variation in yield, grain protein and grain N recovery to a good level of accuracy with RRMSE of 19.5%, 15.4% and 19.5%, respectively. Simulated optimal nitrogen on the low EC soils was on average 12 kg N ha-1 lower, with >1 in 4 years with optimal nitrogen <200 kg N ha-1. Our work demonstrated that using a combination of proximal soil EC scans and targeted soil sampling we can optimize the data requirements for model parameterisation to support site-specific N management.

Challenges in Mapping Soil Variability Using Apparent Soil Electrical Conductivity under Heterogeneous Topographic Conditions

Site-specific management requires the identification of treatment areas based on homogeneous characteristics. This study aimed to determine whether soil mapping based on apparent soil electrical conductivity (ECa) is suitable for mapping soil properties of fields with topographic heterogeneity. Research was conducted on two neighbouring fields in Fejér county, Hungary, with contrasting topographic heterogeneity. To characterise the spatial variability of soil attributes, ECa was measured and supplemented by obtaining soil samples and performing soil profile analysis. The relationship between ECa and soil physical and chemical properties was analysed using correlation, principal component, and regression analyses. The research revealed that the quality and strength of the relationship between ECa and soil remarkably differed in the two studied fields. In homogeneous topographic conditions, ECa was weakly correlated with elevation as determined by soil physical texture and nutrient content in a strong (R2 = 0.72) linear model. On the other hand, ECa was significantly determined by elevation in heterogeneous topographic conditions in a moderate (R2 = 0.47) linear model. Consequently, ECa-based soil mapping can only be used to characterise the soil, thus delineating management zones under homogeneous topographic conditions.

Restoration of degraded waterlogged saline soil through subsurface drainage: Spatiotemporal soil salinity assessment using electromagnetic induction (EMI) and geographic information system (GIS) techniques

AbstractReduction in soil salinity in the entire project area is a key indicator of proper functioning of the implemented subsurface drainage (SSD) projects. Electromagnetic induction (EMI) surveys were carried out to facilitate spatiotemporal assessment of soil salinity of a SSD project site in Kahni, Haryana (India). The surveys were conducted in two time frames: immediately after the start of SSD project in 2019 and after 2 years of successful operation in 2021. The collected data of 2 surveys was used to derive a depth‐specific calibrated correlation between the observed salinity (ECe) of collected soil samples and surveyed apparent electrical conductivity (ECa) observations in vertical (ECav) and horizontal (ECah) modes. Spatiotemporal analysis of soil salinity revealed notable changes in areas under different salinity classes over a period of 2 years, indicating the impact of SSD on soil desalinization. During the course of 2 years (2019–2021) of SSD operation, area under moderate salinity (4–8 dS m−1) in the top 0–0.15 m soil increased by 54.5% while it decreased by 40% and 6.6% under high (8–12 dS m−1) and very high (12–16 dS m−1) salinity classes. The rice and wheat yields were also enhanced by 24% and 7% respectively after 2 years operation of SSD in hitherto severally waterlogged saline soils. Overall findings of this study reinforce the scope of using EMI approach to quickly and effectively assess the efficacy of subsurface drainage. Creation of measurable soil salinity maps by proposed EMI and GIS techniques can help in planning and implementing site‐specific management strategies in saline soil reclamation projects.

Geo-spatial sensing of physical properties a leeway to agricultural soil assessment

The demand for economical means of evaluating soil nutrients’ unpredictability triggered the use of physical factors against the costlier, laborious, and time-consuming chemical approach. This drive led to resolving its capability in evaluating intricate soil properties as a productivity checker. This study aimed at assessing the physical parameters as a useful alternative to the conventional chemical examination of nutrient inconsistency. A petrographic examination was conducted on four rock samples for their classifications. Apparent Electrical Conductivity (ECa) measurements were seasonally executed in the wet (912-station) and dry (906-station). Ten cored soil samples were subjected to a permeability test. Twenty soil samples were examined for pH, Electrical Conductivity (EC), available phosphorus, acidity, Na, Mg, K, and Ca using standard soil science procedures. The mineralogical composition of six samples was determined with X-ray diffraction. The rock is biotite granite gneiss containing plagioclase (22%), microcline (24%), orthoclase (4%), quartz (25%), biotite (7%), and others (18%). The soils ECa were 10-344 µS/cm; categorised as low (1-49 µS/cm), moderate (50-99 µS/cm), and high (&gt;100 µS/cm). The ECa distribution varied from moderate (61%) to high (64%) suggesting a heterogeneous pattern of soil attributes. The infiltration rate was slow in high ECa (5.56x10-5-1.67x10-4 cm/s) signifying good retention capability whereas the low and moderate ECa (moderate-moderately rapid) sections promote nutrient leaching. The cation exchangeable capacity was low (2.99 cmol/kg) in the low ECa and moderate (3.30-4.85 cmol/kg) in the moderate and high ECa; with varying basic cation saturation in the high (81.38%), moderate (73.34%) and low (71.89%) ECa regions and high ECa had higher fertility status. The high ECa had low quartz (41.3%) and microcline (15.7%), but high kaolinite (31.1%) had an affinity to ads orb more cations compared to other ECa regions. ECa variability is practicable in predicting the spatial distribution of soil properties and delineating the management zones. Key words: Granite gneiss, electrical conductivity, permeability, soil composition, mineral assemblages

Time stability of soil volumetric water content and its optimal sampling design in contrasting forest catchments

Characterizing the temporal variation across a forest catchment soil volumetric water content (VWC) with different environmental conditions at different depths for different seasons requires a robust point-scale sampling strategy that balances monitoring cost with predicting accuracy. In the study, the most time-stable locations (MTSLs) using the time-stable concept and the minimum number of required samples (NRS) based on random selection method were studied to explore the optimal sampling design for catchment-mean VWC with an acceptable error (the relative bias (RBIAS) less than 10 %). VWC monitoring was conducted at approximately 25 plots throughout the year in five forest catchments with varying areas and VWC status. Results revealed that the number of MTSLs for mean-VWC estimation with RBIAS within 10 % was directly related to the spatial variability of VWC. Specifically, one MTSL was feasible for accurately estimating mean VWC at different depths in high-VWC catchments with area around or less than 100 ha; one MTSL was adequate for each depth in the catchment with moderate-level VWC; whereas more than one MTSL were necessary for each depth in low-VWC catchment with a larger are of 1800 ha. To identify the MTSL for each catchment, the location with mean VWC was typically situated at the mid-elevation or soil apparent electrical conductivity (ECa) at a certain spatial scale. The NRS for catchment-mean VWC could be determined based on the combination of VWC status, its temporal stability, and potential catchment size, and could be ranked based on the spatial variation of each catchment. If the MTSL cannot be guaranteed for a catchment, the NRS might be a better alternative for optimum sampling design. These finding could be useful in providing guidelines for optimizing VWC monitoring in forest catchments by sampling at a few selected locations representative of the catchment-scale behavior.

How do soil and topographic drivers determine tree diameter spatial distribution in even aged cork oak stands installed in average to high productivity areas

Local terrain or microsite conditions influence the development of trees, particularly at early ages. These conditions might be described by edaphic or topographic variables. We mapped soil and topographic variables from four even-aged and even-spaced cork oak plantations located in two climatically distinct Portuguese regions. The major goal of this research was to understand the relation between soil and topographic fine-scale conditions and tree growth expressed by diameter without cork annual growth (idu). The methodology consisted in (1) analysing the spatial variability and autocorrelation of idu; (2) modelling idu with ordinary least squares (OLS) regressions; (3) comparing with spatial modelling of idu, incorporating spatial autocorrelation. The driest stands A and B, exhibited weaker spatial autocorrelation, distributed in smaller clusters (R2 < 0.03, OLS models), while stands C (R2 = 0.18, OLS models) and D (R2 = 0.11, OLS models) showed higher predictive capacity. Spatial models increased R2 scores, keeping most variables from OLS models and accounting for spatial autocorrelation. A + B + C + D OLS model obtained an R2 = 0.34 and respective spatial model R2 = 0.58. Apparent electrical conductivity at 0.5 (ECa0.5) and 1 m of soil depth, slope, elevation and topography position index were included as predictors (OLS), but only ECa0.5, slope and elevation were selected in the spatial model. Models were fitted using average to high productivity stands and should be used cautiously outside this range. Local terrain conditions determine the growth of young cork oak trees. Mapping soil and topographic variables before establishing new plantations may identify limiting microsite conditions where using cork oak species is not suitable due to low growth rates expectations.

Predicting wetland soil properties using machine learning, geophysics, and soil measurement data

PurposeMachine learning models can improve the prediction of spatial variation of wetland soil properties, such as soil moisture content (SMC) and soil organic matter (SOM). Their performance, however, relies on the quantity of data used to train the model, limiting their use with insufficient data. In this study, we assessed the use of synthetic data constrained by limited field data for training an eXtreme Gradient Boosting (XGBoost) algorithm used to predict the distribution of soil properties based on geophysical measurements constrained by soil samples.Materials and methodsA spatial distribution of soil apparent electrical conductivity (ECa) and laboratory measurements of SOM and SMC from twenty-two core samples were acquired at the St. Michael restored wetland near Defiance, Ohio. The correlations between ECa, SOM, and SMC were explored for predicting the spatial distribution of SOM and SMC. We used a Beta Variational AutoEncoder (β-VAE) approach to synthetically generate over 70,000 training data from the original twenty-two data from soil cores. The training data samples were taken from the latent space. The XGBoost algorithm was then trained on the β-VAE generated data and used to predict the spatial distribution of SOM and SMC at the site. We also validated the accuracy of the XGBoost predictions using an original holdout model validation technique.Results and discussionsThe generated synthetic data using the β-VAE include both soil attributes and ECa, which are larger and more diverse than the original training set with an absolute mean reconstructed error for SMC and SOM ranging from 0.018 to 0.022 and 0.026 to 0.041, respectively. This indicates that the β-VAE successfully generated a realistic synthetic dataset and overcame the technical barrier of using limited datasets. In addition, using generated data to expand the original training data helps the XGBoost model make more accurate predictions compared to training on the original data. The XGBoost prediction performance yielded average Lin’s concordance correlation coefficient (LCCC) values of 0.82 and 0.85 for SOM and SMC and a ratio of performance to deviation (RPD) values of 1.92 and 2.22 respectively, indicating a good performance.ConclusionsThis study validated the use of β-VAE to successfully generate synthetic wetland soil datasets with attributes of the original field data that can be effectively used to train the machine learning XGBoost model. The proposed framework offers an efficient solution for mapping the spatial variability of soil property in data-scarce wetland soil environments.