基于EM38和WorldView-2影像的土壤盐渍化建模研究

在干旱半干旱地区,土壤盐渍化是常见的土地退化问题之一。本研究选取于田县克里雅河上游边缘典型盐渍化区域作为研究靶区,通过EM38大地电导率仪实测土壤表观电导率,提取不同系数下的土壤调节植被指数(SAVI),分析了SAVI指数与土壤电导率间的相关性,并利用同时期WorldView-2影像的敏感波段建立了基于高分辨率影像数据的土壤盐渍化偏最小二乘回归(PLSR)模型并进行了精度验证。结果表明:①从遥感影像提取SAVI指数时,在系数(L)调节范围内选取固定系数值,系数值(间隔为0.1)从0.1变化到1.0的过程中,相应提取的SAVI指数与土壤电导率的相关性明显提升,相关性系数(r)从0.30提高到0.50,并通过显著性检验(P0.01)。②选取的SAVI1.0、B6、B7、B8四种变量中,以SAVI1.0+B6+B8为变量组合所建立的PLSR模型为最优,该模型较其他变量组合建模的决定系数(R2p)提高了0.11,因此,在研究区该模型具有更好的预测能力,模型精度为RMSEC=0.77dS/m、RC2=0.68、RMSEP=0.79 dS/m、RP2=0.66、RPD=2.2。     

Effect of soil matric potential on sharp eyespot in germinating wheat following seed treatment

The influence of soil matric potential on sharp eyespot was examined in wheat seedlings following seed treatment with the fungicide cyproconazole. Emergence of wheat was tested at matric potentials between 0 and –850 hPa in non-infested soil and in soil infested with Rhizoctonia cerealis. Optimal shoot emergence was at –50 and –20 hPa in non-infested and infested soil, respectively. Disease severity was strongly affected by soil matric potential. It continuously increased as matric potential decreased from –5 to –200 hPa. In contrast, optimum growth conditions for the pathogen was at matric potentials between –50 and –200 hPa. With decreasing matric potential the drought stress for the plant seems to increase its predisposition to the pathogen. Seed treatment with cyproconazole reduced sharp eyespot although disease severity increased with decreasing soil matric potential.     

Application of VNIR and machine learning technologies to predict heavy metals in soil and pollution indices in mining areas

Purpose

Soil pollution indices are an effective tool in the computation of metal contamination in soil. They monitor soil quality and ensure future sustainability in agricultural systems. However, calculating a soil pollution index requires laboratory measurements of multiple soil heavy metals, which increases the cost and complexity of evaluating soil heavy metal pollution. Visible and near-infrared spectroscopy (VNIR, 350–2500 nm) has been widely used in predicting soil properties due to its advantages of a rapid analysis, non-destructiveness, and a low cost.

Methods

In this study, we evaluated the ability of the VNIR to predict soil heavy metals (As, Cu, Pb, Zn, and Cr) and two commonly used soil pollution indices (Nemerow integrated pollution index, NIPI; potential ecological risk index, RI). Three nonlinear machine learning techniques, including cubist regression tree (Cubist), Gaussian process regression (GPR), and support vector machine (SVM), were compared with partial least squares regression (PLSR) to determine the most suitable model for predicting the soil heavy metals and pollution indices.

Results

The results showed that the nonlinear machine learning models performed significantly better than the PLSR model in most cases. Overall, the SVM model showed a higher prediction accuracy and a stronger generalization for Zn (R²_V?=?0.95, RMSE_V?=?6.75 mg kg^?1), Cu (R²_V?=?0.95, RMSE_V?=?8.04 mg kg^?1), Cr (R²_V?=?0.90, RMSE_V?=?6.57 mg kg^?1), Pb (R²_V?=?0.86, RMSE_V?=?4.14 mg kg^?1), NIPI (R²_V?=?0.93, RMSE_V?=?0.31), and RI (R²_V?=?0.90, RMSE_V 3.88). In addition, the research results proved that the high prediction accuracy of the three heavy metal elements Cu, Pb, and Zn and their significant positive correlations with the soil pollution indices were the reason for the accurate prediction of NIPI and RI.

Conclusion

Using VNIR to obtain soil pollution indices quickly and accurately is of great significance for the comprehensive evaluation, prevention, and control of soil heavy metal pollution.

     

基于气象因子的长三角地区农田站点土壤水分时间序列预测

以长三角3省1市为研究区,旨在构建长三角地区土壤水分长时间序列,为农业生产和遥感算法提供数据支撑。研究基于空间匹配的站点土壤水分数据和气象数据,利用主成分分析得到4个有效主成分作为线性回归和BP神经网络模型的输入因子,建立土壤水分与气象因子间的定量关系,并评估所构建模型的精度。结果表明,基于全部站点数据建立的单一BP神经网络模型优于单一线性回归模型。单一线性回归模型的R 2=0.34,RMSE=0.046 m³/m³,MAE=3.67%;而单一BP神经网络模型的训练、验证和测试3个数据集的R 2均在0.64以上,RMSE<0.043 m³/m³,MAE低于3.4%。根据逐个站点分别构建分站点的BP神经网络模型,其总体精度高于基于全部站点数据构建的单一BP神经网络模型。分站点构建的BP神经网络模型的总体精度方面,3个数据集的R 2均值在0.75以上,RMSE<0.039 m³/m³,MAE低于3%。通过对逐个站点分别构建BP神经网络模型,获得了精度较高、较稳定的土壤水分拟合结果。     

Modeling Selected Properties of Extruded Rice Flour and Rice Starch by Neural Networks and Statistics

Rice flour and rice starch were single‐screw extruded and selected product properties were determined. Neural network (NN) models were developed for prediction of individual product properties, which performed better than the regression models. Multiple input and multiple output (MIMO) models were developed to simultaneously predict five product properties or three product properties from three input parameters; they were extremely efficient in predictions with values of R² > 0.95. All models were feedforward backpropagation NN with three‐layered networks with logistic activation function for the hidden layer and the output layers. Also, model parameters were very similar except for the number of neurons in the hidden layer. MIMO models for predicting product properties from three input parameters had the same architecture and parameters for both rice starch and rice flour.     

基于Logistic回归和RBF神经网络的土壤侵蚀模数预测

[目的]寻求估算土壤侵蚀模数的新方法,并通过GIS实现对土壤侵蚀空间分布情况的预测。[方法]采用土壤侵蚀模数作为判别条件,分别验证基于Logistic回归和RBF神经网络而建立的土壤侵蚀预报模型的适用性,进而构建并验证改进模型——LOG-RBF神经网络土壤侵蚀预测模型。[结果](1)Logistic回归模型判别目标土地是否发生土壤侵蚀的优势明显,未发生和发生土壤侵蚀的预测正确率分别为77.4%和97.9%,总预测正确率为94.9%。(2)RBF神经网络模型估计土壤侵蚀模数的能力较强,模拟结果的相对误差和平方和误差分别为0.612%和13.292,R2为0.57。(3)LOG-RBF神经网络土壤侵蚀预测模型预测结果的相对误差和平方和误差比RBF神经网络模型模拟结果分别降低了0.157%和2.601。R2为0.82,拟合程度上优于RBF神经网络模型。随着土壤侵蚀模数的增大,错估现象呈逐渐减少趋势。通过受试者工作特征曲线的判别,LOG-RBF神经网络模型的曲线下面积值比RBF神经网络模型大0.063,模型判断的准确性更高。[结论]利用LOG-RBF神经网络土壤侵蚀预测模型可更准确地估计土壤侵蚀模数,基于GIS能够预测土壤侵蚀的空间分布情况。     

基于1D-CNN的土壤全氮近红外光谱预测模型

摘要：【目的】传统的基于近红外光谱数据预测土壤全氮的方法需要对原始光谱数据做复杂的预处理，筛选出与土壤全氮含量相关性高的敏感波长之后进行模型的回归拟合。本文提出一种一维卷积神经网络（1D-CNN）模型，可以在对数据进行简单预处理甚至无处理的情况下达到非常理想的结果，实现用近红外光谱技术对土壤全氮含量的预测。【方法】于江苏无锡采集410个土壤样品，利用半微量开氏法(NY/T 53-1987)测定土壤的全氮含量，并利用NIR Quest 512光谱仪，在室内环境下对每份土壤样品做光谱检测，并用均值中心化(CT)、标准正态变换(SNV)、趋势校正(DT)对光谱进行预处理，运用偏最小二乘回归（PLS）、BP神经网络、1D-CNN方法建立土壤全氮含量的回归预测模型。每种模型在采用不同预处理方法的数据集上做十折交叉验证，记录预测模型的决定系数(R2)和均方根误差(RMSE)的平均值，并对比三种预处理方法对模型精度的影响。【结果】证明了本文提出的1D-CNN模型基于土壤近红外光谱数据预测土壤全氮含量的可靠性。使用原始数据与经均值中心化、标准正态变换、趋势校正预处理的数据训练得到的1D-CNN模型的决定系数分别为0.907、0.931、0.922、0.964，构建的PLS回归模型决定系数为0.856、0.863、0.861、0.880，训练的BP神经网络的决定系数为0.874、0.907、0.901、0.911。【结论】本文提出的1D-CNN模型在原始数据和经预处理的光谱数据上的表现都优于PLS和BP神经网络，且可以证明，对光谱数据进行预处理能够有效提高1D-CNN模型的性能，尤其是趋势校正对模型的提升效果最明显。研究表明，1D-CNN能更好地提取光谱特征并建立其与含氮量的映射关系，有效地避免过拟合，在未经过预处理的光谱数据上依然能够达到一定的精度。     

AquaCrop模型在西北胡麻生物量及产量模拟中的应用和验证

为了预测水分和养分对胡麻籽粒产量、生物量与水分生产率的影响,使用FAO研发的水分驱动作物模型AquaCrop对胡麻在不同灌溉与氮磷水平下的生长情况进行模拟和验证。试验分别于2011年、2012年在甘肃省榆中县良种场进行,试验设置4个灌溉水平,3个氮水平,3个磷水平。模型性能的评价采用模型效率(E)、决定系数(R2)、均方根误差(RMSE)和平均绝对误差(MAE)等统计指标。分析结果表明:AquaCrop模型校正的籽粒产量和生物量在不同灌溉与氮磷水平处理下的预测误差统计值为:0.97E0.99,0.11RMSE0.33,0.11 t·hm 2MAE0.42 t·hm 2,与2012年的试验观察数据(0.96E0.99,0.11RMSE0.42,0.11 t·hm 2MAE0.39 t·hm 2)基本一致;同时,群体覆盖(CC)与生物量的模拟结果与测定值也非常拟合。AquaCrop模型在充分灌溉处理下预测胡麻产量,比非充分灌溉处理下具更高的准确性。因而,水分驱动模型AquaCrop在西北胡麻区不同的灌溉与田间管理措施下有较高的模拟精确性,具有广阔的应用前景和价值。     

桂西北光皮桦人工林水源涵养功能

为了研究广西西北部不同林龄光皮桦人工林的水源涵养功能,选择具有代表性的11,16年生光皮桦人工林、16年生杉木林,从林冠层、枯枝落叶层和土壤层3个层次及综合性的水源涵养能力进行了定量分析。结果表明:(1)11,16年生光皮桦人工林林冠层、灌木层、草本层持水量范围分别为12.54～21.06,2.15～3.05,1.27～1.52 t/hm~2,凋落物总储量为4.54～7.42 t/hm~2,最大持水量为12.55～16.00 t/hm~2,16年生均显著大于11年生(P0.05),凋落物吸水速率与浸水时间存在良好的线性关系(R~20.86,P0.05)。(2)土壤的孔隙状况表现为16年生光皮桦林11年生光皮桦林,均大于对照的16年生杉木林,0—20 cm显著大于20—40,40—80 cm土层。(3)11年生光皮桦土壤最大持水量、毛管持水量、非毛管持水量的变化范围分别为28.97%～60.55%,25.35%～47.21%,3.71%～13.34%,16年生的为29.06%～63.45%,25.63%～48.70%,3.34%～14.75%,均随着土层的加深而减少;11,16年生光皮桦林0—80 cm土壤层自然含水量范围分别为27.46～30.16,28.12～30.22 g/cm~3;总蓄水量分别为3 813.4,3 732.2 t/hm~2,均大于16年生杉木林(3 659.2 t/hm~2)。总体上,林龄较大的光皮桦人工林表现出较强的水源涵养功能,且优于同林龄的杉木人工林。研究结果可为该地区光皮桦人工林的经营管理提供科学依据。     

Using a nitrogen mineralization index will improve soil productivity rating by artificial neural networks

ABSTRACT

In the Pampas, nitrogen fertilization rates are low and soil organic matter impacts crop yield. Wheat (Triticum aestivum L.) yield was related to total soil nitrogen (total N) and to nitrogen mineralization potential (mineralized N) to determine whether the effects of organic matter may be attributed to its capacity to act as a nitrogen source or to the improvement of the soil physical condition. Data of 386 sites from throughout the region comprised in a recent soil survey were used, in which climate and soil properties to 1 m depth were determined. Artificial neural networks were applied for total N and mineralized N estimation using climate and soil variables as inputs (R² = 0.59–0.70). The models allowed estimating total N and mineralizable N at county scale and related them to statistical yield information. Neural networks were also used for yield prediction. The best productivity model fitted (R² = 0.85) showed that wheat yield could be predicted by rainfall, the photothermal quotient, and mineralized N. The soil organic matter effect on crop yield seems to be mainly related to its nitrogen mineralization capacity. Using mineralized N as predictor would be a valuable tool for rating soil productivity.     

Evaluation of 1H NMR relaxometry for the assessment of pore‐size distribution in soil samples

¹H NMR relaxometry is used in earth science as a non‐destructive and time‐saving method to determine pore size distributions (PSD) in porous media with pore sizes ranging from nm to mm. This is a broader range than generally reported for results from X‐ray computed tomography (X‐ray CT) scanning, which is a slower method. For successful application of ¹H NMR relaxometry in soil science, it is necessary to compare PSD results with those determined from conventional methods. The PSD of six disturbed soil samples with various textures and soil organic matter (SOM) content were determined by conventional soil water retention at matric potentials between −3 and −390 kPa (pF 1.5–3.6). These PSD were compared with those estimated from transverse relaxation time (T₂) distributions of water in soil samples at pF 1.5 using two different approaches. In the first, pore sizes were estimated using a mean surface relaxivity of each soil sample determined from the specific surface area. In the second and new approach, two surface relaxivities for each soil sample, determined from the T₂ distributions of the soil samples at different matric potentials, were used. The T₂ distributions of water in the samples changed with increasing soil matric potential and consisted of two peaks at pF 1.5 and one at pF 3.6. The shape of the T₂ distributions at pF 1.5 was strongly affected by soil texture and SOM content (R² = 0.51 − 0.95). The second approach (R² = 0.98) resulted in good consistency between PSD, determined by soil water retention, and ¹H NMR relaxometry, whereas the first approach resulted in poor consistency. Pore sizes calculated from the NMR data ranged from 100 μm to 10 nm. Therefore, the new approach allows ¹H NMR relaxometry to be applied for the determination of PSD in soil samples and for studying swelling of SOM and clay and its effects on pore size in a fast and non‐destructive way. This is not, or only partly, possible by conventional soil water retention or X‐ray CT.     

Estimating Mo,Cu, Ni,Cd Contents in the Crop Leaves Growing on Small Land Plots Using Satellite Data

ABSTRACT

The main goal of this research was to estimate heavy metals (HMs) (molybdenum (Mo), copper (Cu), nickel (Ni), cadmium (Cd)) contents in crop leaves through multispectral satellite imagery. During the acquisition of a SPOT 7 satellite image (28 July 2017) in situ sampling (38 samples) was done from the leaves of potatoes and beans growing close to the mining town of Kajaran (Armenia). To estimate HMs contents, multivariate regression (multiple linear regression (MLR), partial least squares regression (PLSR)), and artificial neural network (ANN) were used. As input data for the models raw, atmospherically corrected (Dark Object Subtraction (DOS)) and hyperspherical direction cosines (HSDC) normalized values of SPOT 7 spectral data in combination with one or combined log10, multiplicative scatter correction (MSC), standard normal variate transform (SNV) preprocessing methods were utilized. The best results were obtained for Cu using MLR (R² _cal. = 0.79, R² _CV = 0.70, RMSE_cal. = 11.27, RMSE_CV = 13.47) and ANN (R² _Train ≈ 0.80, R² _Test ≈ 0.72, RMSE_Train ≈ 11, RMSE_Test ≈ 13) models in case of bean leaves. The results are quite optimistic, however, further research with the use of high spatial/spectral resolution satellite images is needed to improve the accuracy of models.     

Effect of biochar application rate on soil physical and hydraulic properties of a sandy loam

Biochar is used as a soil amendment for improving soil quality and enhancing carbon sequestration. In this study, a loamy sand soil was amended at different rates (0%, 25%, 50%, 75%, and 100% v/v) of biochar, and its physical and hydraulic properties were analyzed, including particle density, bulk density, porosity, infiltration, saturated hydraulic conductivity, and volumetric water content. The wilting rate of tomato (Solanum lycopersicum) grown in soil amended with various levels of biochar was evaluated on a scale of 0–10. Statistical analyses were conducted using linear regression. The results showed that bulk density decreased linearly (R² = 0.997) from 1.325 to 0.363 g cm^?3 while the particle density decreased (R² = 0.915) from 2.65 to 1.60 g cm^?3 with increased biochar amendment, with porosity increasing (R² = 0.994) from 0.500 to 0.773 cm³ cm^?3. The mean volumetric water content ranged from 3.90 to 14.00 cm³ cm^?3, while the wilting rate of tomato ranged from 4.67 to 9.50, respectively, for the non-amended soil and 100% biochar-amended soil. These results strongly suggest positive improvement of soil physical and hydraulic properties following addition of biochar amendment.     

Estimating the bulk density in 0– 20 cm of tilled soils in China’s Loess Plateau using support vector machine modeling

ABSTRACT

Pedotransfer functions (PTFs), as an indirect forecasting method, offer an alternative for labor-intensive bulk density (BD) measurements. In order to improve the forecasting accuracies, support vector machine (SVM) method was first used to develop PTFs for predicting BD. Cross-validation and grid-search methods were used to automatically determine the SVM parameters in the forecasting process. Soil texture and organic matter content were selected as input variables based on results of predecessors, coupled with gray correlation theory. And additional properties were added as inputs for improving PTF's accuracy and reliability. The performance of the PTF established by SVM method was compared with artificial neural network (ANN) method and published PTFs using two indexes: root-mean-square error (RMSE) and coefficient of determination(R²). Results showed that the average RMSE of published PTFs was 0.1053, and the R² was 0.4558. The RMSE of ANN–PTF was 0.0638, and the R² was 0.7235. The RMSE of SVM–PTF was 0.0558, and the R² was 0.7658. Apparently, the SVM–PTF had better performance, followed by ANN–PTF. Additionally, performances could be improved when accumulated receiving water was added as predictor variable. Therefore, the first application of SVM data mining techniques in the prediction of soil BD was successful, improved the accuracy of predictions, and enhanced the function of soil PTFs. The idea of developing PTFs using SVM method for predicting soil BD in the study area could provide a reference for other areas.     

Prediction of macronutrients in plant leaves using chemometric analysis and wavelength selection

Purpose

Fast and real-time prediction of leaf nutrient concentrations can facilitate decision-making for fertilisation regimes on farms and address issues raised with over-fertilisation. Cacao (Theobroma cacao L.) is an important cash crop and requires nutrient supply to maintain yield. This project aimed to use chemometric analysis and wavelength selection to improve the accuracy of foliar nutrient prediction.

Materials and methods

We used a visible-near infrared (400–1000 nm) hyperspectral imaging (HSI) system to predict foliar calcium (Ca), potassium (K), phosphorus (P) and nitrogen (N) of cacao trees. Images were captured from 95 leaf samples. Partial least square regression (PLSR) models were developed to predict leaf nutrient concentrations and wavelength selection was undertaken.

Results and discussion

Using all wavelengths, Ca (R²_CV?=?0.76, RMSE_CV?=?0.28), K (R²_CV?=?0.35, RMSE_CV?=?0.46), P (R²_CV?=?0.75, RMSE_CV?=?0.019) and N (R²_CV?=?0.73, RMSE_CV?=?0.17) were predicted. Wavelength selection increased the prediction accuracy of Ca (R²_CV?=?0.79, RMSE_CV?=?0.27) and N (R²_CV?=?0.74, RMSE_CV?=?0.16), while did not affect prediction accuracy of foliar K (R²_CV?=?0.35, RMSE_CV?=?0.46) and P (R²_CV?=?0.75, RMSE_CV?=?0.019).

Conclusions

Visible-near infrared HSI has a good potential to predict Ca, P and N concentrations in cacao leaf samples, but K concentrations could not be predicted reliably. Wavelength selection increased the prediction accuracy of foliar Ca and N leading to a reduced number of wavelengths involved in developed models.

     

Pedotransfer functions for solute transport parameters of Portuguese soils

The purpose of this study is to quantify solute transport parameters of fine‐textured soils in an irrigation district in southern Portugal and to investigate their prediction from basic soil properties and unsaturated hydraulic parameters. Solute displacement experiments were carried out on 24 undisturbed soil samples by applying a 0.05 m KCl pulse during steady flow. The chloride breakthrough curves (BTCs) were asymmetric, with early breakthrough and considerable tailing characteristic of non‐equilibrium transport. The retardation factor (R), dispersion coefficient (D), partitioning coefficient (β), and mass transfer coefficient (ω) were estimated by optimizing the solution of the non‐equilibrium convection–dispersion equation (CDE) to the breakthrough data. The solution could adequately describe the observed data as proved by a median of 0.972 for the coefficient of determination (r²) and a median for the mean squared error (MSE) of 5.1 × 10^?6. The median value for R of 0.587 suggests that Cl^– was excluded from a substantial part of the liquid phase. The value for β was typically less than 0.5, but the non‐equilibrium effects were mitigated by a large mass transfer coefficient (ω > 1). Pedotransfer functions (PTFs) were developed with regression and neural network analyses to predict R, D, β and ω from basic soil properties and unsaturated hydraulic parameters. Fairly accurate predictions could be obtained for logD (r² ≈ 0.9) and β (r² ≈ 0.8). Prediction for R and logω were relatively poor (r² ≈ 0.5). The artificial neural networks were all somewhat more accurate than the regression equations. The networks are also more suitable for predicting transport parameters because they require only three input variables, whereas the regression equations contain many predictor variables.     

北京市松山不同海拔油松林枯落物及土壤水文效应

以北京市松山4个海拔梯度(751,890,1 012和1 211m)的油松(Pinus tabuliformis Carr)天然林为对象,对其枯落物层及土壤层水文效应进行研究。结果表明:(1)枯落物总蓄积量、枯落物最大持水量和最大持水率均随海拔的升高先增大后减小;(2)枯落物的总储量为9.03~27.75t/hm2,最大持水量为26.66~90.54t/hm2,与浸泡时间呈明显的对数关系(R0.93);最大持水率为287.62%~296.73%,与浸泡时间呈明显的幂函数关系(R0.99);(3)土壤容重随海拔升高而减小,其变化范围为1.38~1.66g/cm3,总孔隙度随海拔升高先减小后增大;(4)土壤初渗速率相差较大,稳渗速率为1.95~7.06mm/min,入渗速率与入渗时间呈幂函数关系(R0.70)。综合分析得出,低海拔油松天然林水源涵养功能较强。     

Using support vector machines to predict cation exchange capacity of different soil horizons in Qingdao City,China

Agricultural, environmental and ecological modeling requires soil cation exchange capacity (CEC) that is difficult to measure. Pedotransfer functions (PTFs) are thus routinely applied to predict CEC from easily measured physicochemical properties (e.g., texture, soil organic matter, pH). This study developed the support vector machines (SVM)‐based PTFs to predict soil CEC based on 208 soil samples collected from A and B horizons in Qingdao City, Shandong Province, China. The database was randomly split into calibration and validation datasets in proportions of 3:1 using the bootstrap method. The optimal SVM parameters were searched by applying the genetic algorithm (GA). The performance of SVM models was compared to those of multiple stepwise regression (MSR) and artificial neural network (ANN) models. Results show that the accuracy of CEC predicted by SVM improves considerably over those predicted by MSR and ANN. The performance of SVM for B horizon (R² = 0.85) is slightly better than that for A horizon (R² = 0.81). The SVM is a powerful approach in the simulation of nonlinear relationship between CEC and physicochemical properties of widely distributed samples from different soil horizons. Sensitivity analysis was also conducted to explore the influence of each input parameter on the CEC predictions by SVM. The clay content is the most sensitive parameter, followed by soil organic matter and pH, while sand content has the weakest influence. This suggests that clay is the most important predictor for predicting CEC of both soil horizons.     

Estimation of bulk density of waterlogged soils from basic properties

Pedotransfer functions (PTFs) to predict bulk density (BD) from basic soil data are presented. Available data pertaining to seasonally impounded shrink–swell soils of Jabalpur district in the Madhya Pradesh state of India were used for the study. The data included horizon-wise information of 41 soil profiles in the study area covering nearly 5 million ha. Six independent variables, namely textural data (sand, silt and clay), field capacity (FC), permanent wilting point (PWP) and organic carbon content (OC) were used as input in hierarchical steps to establish dependencies, with bulk density as the dependent variable, using statistical regression and artificial neural networks. The PTFs derived using neural networks [average root mean square error (RMSE) 0.05] were relatively better than statistical regression PTFs (average RMSE > 0.1). The best-performing PTFs required input data on sand, silt content, FC and PWP, with lowest prediction errors (RMSE 0.01, maximum absolute error (MAE) 0.01) and highest values of index of agreement (d, 0.95) and R ² (0.65). Use of measures of structure, as well as information on pore structure, was found to be essential to derive acceptable PTFs. Inclusion of OC as an input variable showed relatively better fitting to the training data set, implying an underlying relationship between OC and BD, but the neural networks could not mimic the relationship when tested against subset.     

Rapid determination of lime requirement by mid‐infrared spectroscopy: A promising approach for precision agriculture

Mid‐infrared spectroscopy (MIRS) has proven to be a cost‐effective, high throughput measurement technique for soil analysis. After multivariate calibration mid‐infrared spectra can be used to predict various soil properties, some of which are related to lime requirement (LR). The objective of this study was to test the performance of MIRS for recommending variable rate liming on typical Central European soils in view of precision agriculture applications. In Germany, LR of arable topsoils is commonly derived from the parameters organic matter content (SOM), clay content, and soil pH (CaCl₂) as recommended by the Association of German Agricultural Analytical and Research Institutes (VDLUFA). We analysed a total of 458 samples from six locations across Germany, which all revealed large within‐field soil heterogeneity. Calcareous topsoils were observed at some positions of three locations (79 samples). To exclude such samples from LR determination, peak height at 2513 cm^?1 of the MIR spectrum was used for identification. Spectra‐based identification was accurate for carbonate contents > 0.5%. Subsequent LR derivation (LR_SPP) from MIRS‐PLSR predictions of SOM, clay, and pH (CaCl₂) for non‐calcareous soil samples using the VDLUFA look‐up tables was successful for all locations (R² = 0.54–0.82; RMSE = 857–1414 kg CaO ha^?1). Alternatively, we tested direct LR prediction (LR_DP) by MIRS‐PLSR and also achieved satisfactory performance (R² = 0.52–0.77; RMSE = 811–1420 kg CaO ha^?1; RPD = 1.44–2.08). Further improvement was achieved by refining the VDLUFA tables towards a stepless algorithm. It can be concluded that MIRS provides a promising approach for precise LR estimation on heterogeneous arable fields. Large sample numbers can be processed with low effort which is an essential prerequisite for variable rate liming in precision agriculture.