基于野外实测Vis-NIR光谱的土壤肥力估算研究

为探讨野外实测光谱数据对土壤肥力的估算能力,采集青海省湟水流域表层0 ~ 20 cm土壤样品220份,同步测量其采样位置的野外实测光谱数据,实验室对土壤养分、机械组成含量以及pH值进行分析。基于上述数据,对野外实测光谱反射率进行多元散射校正（Multiplicative scatter correction,MSC）、SG-一阶导数变换（SG - First Derivative,SG-1st）预处理,采用稳定性竞争自适应重加权采样法（stability competitive adaptive reweighted sampling,SCARS）提取不同土壤养分、机械组成含量以及pH值的特征波段,以偏最小二乘回归（partial least squares regression,PLSR）模型对土壤全碳（TC）、有机质（OM）、全氮（TN）、碱解氮（AN）、pH、黏粒（clay）、粉粒（silt）、砂粒（sand）含量进行估算并对比分析,构建土壤养分含量、pH值以及机械组成含量的最优野外实测光谱估算模型。结果表明:通过MSC校正和SG-1st变换能够有效增强野外光谱特征;经SCARS选取的特征波段主要集中于近红外波段。基于野外实测光谱数据建立的PLSR模型能够对研究区土壤TC、OM、TN、AN含量以及pH值进行粗略估算;其中,对于TC、OM、TN含量及pH值而言,最佳估算模型为经SG-1st处理后的SCARS-PLSR模型,RPD值均达到1.70以上（RPD_TC = 1.76; RPD_OM = 1.82;RPD_TN = 2.04;RPD_pH = 1.89）,RPIQ值均达到1.90以上（RPIQ_TC = 1.91;RPIQ_OM = 2.53;RPIQ_TN = 2.98;RPIQ_pH = 2.03）;对于土壤AN含量而言,经MSC处理后的SCARS-PLSR模型最佳,其RPD_AN值高达1.91,RPIQ值高达2.39。对土壤clay、silt以及sand含量野外光谱均无法估算,RPD值均在1.00左右,RPIQ值在1.20左右。     

黑土养分含量的航空高光谱遥感预测

为监测黑龙江省黑土典型区土壤的养分元素含量,综合利用统计理论与光谱分析方法,研究建三江农场黑土土壤的3类养分含量与土壤光谱之间的关系,建立土壤全氮、有效磷、速效钾含量高光谱反演模型,实现土壤养分元素含量定量预测。对黑土土壤航空高光谱数据进行处理,应用偏最小二乘回归（PLSR）和BP神经网络方法分别建立土壤养分元素含量的高光谱定量反演模型,结果表明：全氮PLSR和BP神经网络预测模型的RPIQ值（样本观测值第三和第一四分位数之差与均方根误差的比值）分别为2.42和2.80;有效磷PLSR和BP神经网络模预测型的RPIQ值分别为0.83和1.67;速效钾PLSR和BP神经网络模型的RPIQ值分别为2.00和2.33。试验证明土壤全氮和速效钾的光谱定量预测模型具备较好的精度和预测能力。但有效磷的预测效果不是特别理想,仅可达到近似定量预测的要求;全氮、有效磷和速效钾的预测精度,BP神经网络建模相比偏最小二乘建模有更好的精度和预测能力,预测精度分别提高6.5%、10.1%和6.6%。     

灌溉水中悬浮固体对土壤水分入渗性能的影响

为监测黑龙江省黑土典型区土壤的养分元素含量,综合利用统计理论与光谱分析方法,研究建三江农场黑土土壤的3类养分含量与土壤光谱之间的关系,建立土壤全氮、有效磷、速效钾含量高光谱反演模型,实现土壤养分元素含量定量预测。对黑土土壤航空高光谱数据进行处理,应用偏最小二乘回归（PLSR）和BP神经网络方法分别建立土壤养分元素含量的高光谱定量反演模型,结果表明：全氮PLSR和BP神经网络预测模型的RPIQ值（样本观测值第三和第一四分位数之差与均方根误差的比值）分别为2.42和2.80;有效磷PLSR和BP神经网络模预测型的RPIQ值分别为0.83和1.67;速效钾PLSR和BP神经网络模型的RPIQ值分别为2.00和2.33。试验证明土壤全氮和速效钾的光谱定量预测模型具备较好的精度和预测能力。但有效磷的预测效果不是特别理想,仅可达到近似定量预测的要求;BP神经网络建模相比偏最小二乘建模有更好的精度和预测能力,预测精度分别提高6.5%、10.1%和6.6%。     

苏北沿海滩涂地区土壤有机质含量的高光谱预测

基于反射高光谱快速、无损的检测优势,以苏北沿海滩涂地区不同成陆年代土壤作为光谱信息源,应用偏最小二乘回归（PLSR）方法,研究了原始反射光谱（REF）、微分光谱（FDR）、反射率倒数的对数（lg(1/R)）和波段深度（BD）对不同成陆年代土壤有机质含量的预测精度。结果表明,不同成陆年代土壤有机质含量预测的最佳光谱指标存在差异。REF是构建总体样本有机质含量PLSR预测模型的最佳光谱指标,均方根误差（RMSE）和相关系数（r）分别为2.7231和0.8701;FDR是预测成陆千年土壤样本有机质含量的最佳光谱指标,RMSE和r分别为2.0110和0.9436;BD所构建的成陆百年土壤有机质含量的PLSR预测模型为最优,RMSE和r分别为2.7051和0.8770。相关分析表明,可见光波段、以1 400 nm为中心及1 900～2 450 nm的红外波段是估算土壤有机质含量的最佳波段。     

基于近红外光谱的土壤全氮含量估算模型

土壤全氮是诊断土壤肥力水平和指导作物精确施肥所需的重要信息,建立土壤全氮的近红外光谱估测模型并对建模波段进行优化选择对于土壤养分信息快速获取和精确农业发展具有重要意义。该研究以中国中、东部地区5种主要类型土壤为研究对象,利用近红外光谱仪采集土壤样品的光谱信息,结合近红外区域分子振动特点选取全谱、合频、一倍频、二倍频和N-H基团及其组合的8个波段,采用多元散射校正等多种预处理方法组合进行处理,结合偏最小二乘法(PLS)对每个波谱区域进行定标建模。结果表明,利用4000～5500cm-1波谱区域结合附加散射校正处理过的原始光谱建立的模型精度表现最好,其内部互验证决定系数达到0.90,均方根误差为0.16。经不同类型土壤的观测资料检验,模型验证决定系数为0.91,均方根误差为0.15,相对分析误差RPD为3.40,表明模型具有极好的预测能力。因此,利用近红外光谱可以实现土壤全氮的快速估测,且以合频波段(4000～5500cm-1)为建模区域可以得到更好的预测效果。     

iPLS-SPA变量选择方法在螺旋藻粉无损检测中的应用

该文研究了基于可见-近红外光谱技术的螺旋藻粉类别无损检测方法。采用簇类独立软模式法（SIMCA）建立可见-近红外光谱模型。全波段光谱所建立的模型得到了93.33%的预测集正确率。文章提出了基于间隔偏最小二乘法（iPLS）和连续投影算法（SPA）的组合光谱变量选择方法进行有效波长的选择。该方法从全波段675个变量中选择了5个最优的有效波段,并且得到了96.67%的预测集正确率。和基于全波段光谱、可见光波段光谱和近红外波段光谱进行SPA运算相比,基于iPLS的SPA运算可以有效减少计算时间。研究表明可见-近红外光谱可以用于对螺旋藻粉类别进行无损检测,同时iPLS-SPA是一个有效的光谱变量选择方法。     

基于可见–近红外光谱的水稻土全磷反演研究

采用PLSR偏最小二乘法回归结合留一法交叉验证,利用长期定位试验田以及直湖港小流域面上的水稻土土壤样本建立最优模型,研究了不同光谱预处理方式对水稻土全磷可见-近红外高光谱反演精度的影响,探索水稻土全磷光谱反演的可行性;并结合简单相关系数法以及PLSR模型回归系数法分析了水稻土全磷光谱反演的重要波段。结果表明,光谱预处理方法对土壤全磷反演精度的影响不大;基于PLSR建立的水稻土全磷光谱反演模型的校正决定系数达0.85,交叉验证决定系数为0.70,RPD为1.8,有较好的模型精度;440~740 nm为土壤全磷光谱反演的重要波段。利用PLSR对水稻土全磷进行光谱预测是可行的。     

基于CARS算法的不同类型土壤有机质高光谱预测

不同土壤类型的理化性质和光谱性质存在差异,以往研究多以高光谱反射率或光谱吸收特征建立模型,输入变量类型结构单一,往往导致土壤有机质（Soil Organic Matter,SOM）预测模型的精度不高。为提高SOM高光谱预测模型精度,该研究以黑龙江省海伦市为研究区,将不同类型土壤分别以竞争自适应重加权采样（Competitive Adaptive Reweighted Sampling,CARS）筛选的特征波段、数字高程模型（Digital Elevation Model,DEM）数据和光谱指数作为输入变量,结合随机森林（Random Forest,RF）算法建立SOM预测模型。结果表明：1）通过CARS算法筛选后,各土壤类型特征波段压缩至全波段数目的16%以下,在很大程度上降低土壤高光谱变量维度和计算复杂程度,从而提高了模型的预测能力,说明CARS算法在提取特征关键波段变量、优化模型结构方面起到重要作用;2）不同类型土壤的SOM预测精度存在差异,沼泽土的预测精度最高为0.768,性能与四分位间隔距离的比率（Ratio of Performance to InterQuartile distance,RPIQ）为3.568;黑土次之,草甸土的预测精度最低,仅0.674,RPIQ为1.848。3类土壤的RPIQ均达到1.8以上,模型具有较好的预测能力;3）局部回归预测精度最优,验证集的调整后决定系数为0.777,均方根误差（Root Mean Square Error,RMSE）为0.581%,模型验证RPIQ为2.689,模型稳定性高。该试验筛选的预测因子通过RF模型可实现SOM含量的快速预测,简化了传统复杂的程序,可为中尺度区域不同类型土壤的SOM预测提供依据,为输入量的选择提供参考。     

基于可见-近红外光谱的滩羊肉多品质指标无损检测

针对近红外光谱技术在生鲜肉品质检测中预测模型适用范围窄、检测指标单一、模型稳定性差、难以有效应用于生产检测等问题,本研究采集不同月龄宁夏滩羊宰后3个时期4个部位肉的可见-近红外光谱信息,测定色泽、pH值、蒸煮损失、剪切力以及蛋白质、粗脂肪和水分含量,利用2个波段(370~1 050 nm、900~1 700 nm)的光谱数据分别构建各个指标的偏最小二乘回归(PLSR)预测模型以实现滩羊肉多品质指标同步无损检测。结果表明,两波段中各品质指标的PLSR预测模型相关系数(R)均大于0.80,第二波段中水分含量PLSR模型预测集R可达0.941;两波段中各品质指标预测模型的性能较好,其中370~1 050 nm波段的光谱数据对样品色泽参数预测效果更好。综上所述,可见-近红外光谱技术可实现滩羊肉7个品质指标的快速无损检测。本研究结果为滩羊肉品质控制和滩羊屠宰加工企业优质特色产品的生产提供了技术支撑。     

黄土的红外反射光谱与红外光声光谱特征及其差异研究

红外光谱法作为一种新的研究手段已经广泛应用于土壤分析,由其检测区域和手段的不同又可分为多种光谱类 型。本研究以第四纪黄土为例,系统地比较了近红外区和中红外区反射光谱和光声光谱的吸收特征及其差异。结果表明,中红外光谱比近红外光谱的信息更为丰富,且中红外光谱与样品中物质的特征吸收关系更加密切,从而更有利于土壤定性与定量分析。土壤的反射光谱和光声光谱表现出了明显不同的特征,在近红外区,反射光谱和光声光谱吸收明显不同,而在中红外区,反射光谱和光声光谱具有相对应的吸收,但相对吸收强度明显不同,且吸收峰的位置也发生改变,尤其在1 000 ~ 2 000 cm-1谱区,反射光谱相互干扰很强,而光声光谱的吸收特征更为明显。在黄土的分类鉴别上,反射光谱优于光声光谱。红外反射光谱和光声光谱在不同波段下具有不同的吸收灵敏度,在土壤定性与定量分析中各自都将具有其明显的优势。     

基于相似光谱匹配预测土壤有机质和阳离子交换量

土壤可见光-近红外波段光谱(350~2 500 nm)包含了大量的土壤属性信息,相同类型的土壤具有相似的光谱曲线特征,但相似光谱曲线是否具有相似的属性含量?探讨此问题可为土壤光谱库的应用提供依据,从而最终服务于快速获取土壤信息技术体系的构建。该研究以安徽宣城为研究区,根据母质、地形特征和土地利用等信息,采集91个典型土壤剖面,共含400个土壤发生层样品,测定了有机质(soil organic matter,SOM)和阳离子交换量(cation exchange capacity,CEC)含量,同时采用VARIAN公司的Cary 5000分光光度计测定了土壤光谱,并将光谱数据变换为反射率(R)、反射率一阶导数(FDR)和吸收度(Log(1/R))3种形式。该文采用光谱角(spectral angle mapper,SAM)、偏最小二乘回归(partial least square regression,PLSR)和SAM-PLSR(spectral angle mapper-partial least square regression,SAM-PLSR)3种方法预测土壤SOM和CEC。SAM方法是通过对测试集104个光谱曲线与参考集的296个光谱曲线进行相似性计算,并以此实现土壤SOM和CEC含量的预测。SAM-PLSR方法以SAM算法下的匹配结果作为建模样本建立PLSR模型和进行预测分析。结果表明,具有相似光谱曲线的土壤具有相似的SOM和CEC含量,SAM算法下相似光谱匹配可直接预测SOM(R2=0.78,RPD=2.17)和CEC(R2=0.82,RPD=2.41)。PLSR方法可很好地预测SOM(R2=0.87,RPD=2.77)和CEC(R2=0.87,RPD=2.59);相较之下,SAM-PLSR方法不仅可以更加准确预测SOM(R2=0.89,RPD=3.00)和CEC(R2=0.91,RPD=3.06),而且大大减少了建模样本的数量。该研究使可见光-近红外光谱可更加高效地用于土壤属性分析,并为土壤光谱数据库的建设及应用提供技术参考。     

Prediction of soil organic matter using a spatially constrained local partial least squares regression and the Chinese vis–NIR spectral library

We need to determine the best use of soil vis–NIR spectral libraries that are being developed at regional, national and global scales to predict soil properties from new spectral readings. To reduce the complexity of a calibration dataset derived from the Chinese vis–NIR soil spectral library (CSSL), we tested a local regression method that combined geographical sub‐setting with a local partial least squares regression (local‐PLSR) that uses a limited number of similar vis–NIR spectra (k‐nearest neighbours). The central idea of the local regression, and of other local statistical approaches, is to derive a local prediction model by identifying samples in the calibration dataset that are similar, in spectral variable space, to the samples used for prediction. Here, to derive our local regressions we used Euclidean distance in spectral space between the calibration dataset and prediction samples, and we also used soil geographical zoning to account for similarities in soil‐forming conditions. We tested this approach with the CSSL, which comprised 2732 soil samples collected from 20 provinces in the People's Republic of China to predict soil organic matter (SOM). Results showed that the prediction accuracy of our spatially constrained local‐PLSR method (R² = 0.74, RPIQ = 2.6) was better than that from local‐PLSR (R² = 0.69, RPIQ = 2.3) and PLSR alone (R² = 0.50, RPIQ = 1.5). The coupling of a local‐PLSR regression with soil geographical zoning can improve the accuracy of local SOM predictions using large, complex soil spectral libraries. The approach might be embedded into vis–NIR sensors for laboratory analysis or field estimation.     

Prediction of volumetric soil organic carbon from field‐moist intact soil cores

This study investigated the potential for visible–near‐infrared (vis–NIR) spectroscopy to predict locally volumetric soil organic carbon (SOC) from spectra recorded from field‐moist soil cores. One hundred cores were collected from a 71‐ha arable field. The vis–NIR spectra were collected every centimetre along the side of the cores to a depth of 0.3 m. Cores were then divided into 0.1‐m increments for laboratory analysis. Reference SOC measurements were used to calibrate three partial least‐squares regression (PLSR) models for bulk density (ρ_b), gravimetric SOC (SOC_g) and volumetric SOC (SOC_v). Accurate predictions were obtained from averages of spectra from those 0.1‐m increments for SOC_g (ratio of performance to inter‐quartile (RPIQ) = 5.15; root mean square error (RMSE) = 0.38%) and SOC_v (RPIQ = 5.25; RMSE = 4.33 kg m^?3). The PLSR model for ρ_b performed least well, but still produced accurate results (RPIQ = 3.76; RMSE = 0.11 Mg m^?3). Predictions for ρ_b and SOC_g were combined to compare indirect and direct predictions of SOC_v. No statistical difference in accuracy between these approaches was detected, suggesting that the direct prediction of SOC_v is possible. The PLSR models calibrated on the 10‐cm depth intervals were also applied to the spectra originally recorded on a 1‐cm depth increment. While a bigger bias was observed for 1‐cm than for 10‐cm predictions (1.13 and 0.19 kg m^?3, respectively), the two populations of estimates were not distinguishable statistically. The study showed the potential for using vis–NIR spectroscopy on field‐moist soil cores to predict SOC at high depth resolutions (1 cm) with locally derived calibrations.     

桂西北典型环境移民迁入区土壤养分特征分析

在桂西北典型环境移民迁入区,分析了5种土地利用方式表层(0-20 cm)土壤有机质(SOM)、全量养分和速效养分的变化特征.结果表明,根据变异系数(Cv)大小,除速效磷(AP)为强变异外Cv＞1.0),其余土壤养分都为中等变异(0.1＜Cv＜1.0).整体而言,移民迁入区SOM、全氮(TN)含量处于较高水平,速效氮(AN)、速效磷(AP)含量处于中等水平,全磷(TP)、速效钾(AK)处于偏低水平,而全钾(TK)处于低水平.土地利用方式是影响土壤养分的主要因素.灌丛和荒草地的SOM,TN,AN处于较高水平,果园和旱地则处于中等水平;次生林SOM处于较高水平,TN处于中等水平,而AN处于偏低水平.除旱地和果园AP含量处于中等水平外,不同利用方式土壤的TP,TK,AP,AK含量处于低或偏低水平.由于采取了较好的水土保持措施,移民后较大规模的土地利用变化没有导致明显的土壤退化,但需要增加磷钾肥的施用量.     

Optimised use of data fusion and memory-based learning with an Austrian soil library for predictions with infrared data

Infrared spectroscopy in the visible to near-infrared (vis–NIR) and mid-infrared (MIR) regions is a well-established approach for the prediction of soil properties. Different data fusion and training approaches exist, and the optimal procedures are yet undefined and may depend on the heterogeneity present in the set and on the considered scale. The objectives were to test the usefulness of partial least squares regressions (PLSRs) for soil organic carbon (SOC), total carbon (C_t), total nitrogen (N_t) and pH using vis–NIR and MIR spectroscopy for an independent validation after standard calibration (use of a general PLSR model) or using memory-based learning (MBL) with and without spiking for a national spectral database. Data fusion approaches were simple concatenation of spectra, outer product analysis (OPA) and model averaging. In total, 481 soils from an Austrian forest soil archive were measured in the vis–NIR and MIR regions, and regressions were calculated. Fivefold calibration-validation approaches were carried out with a region-related split of spectra to implement independent validations with n ranging from 47 to 99 soils in different folds. MIR predictions were generally superior over vis–NIR predictions. For all properties, optimal predictions were obtained with data fusion, with OPA and spectra concatenation outperforming model averaging. The greatest robustness of performance was found for OPA and MBL with spiking with R² ≥ 0.77 (N), 0.85 (SOC), 0.86 (pH) and 0.88 (C_t) in the validations of all folds. Overall, the results indicate that the combination of OPA for vis–NIR and MIR spectra with MBL and spiking has a high potential to accurately estimate properties when using large-scale soil spectral libraries as reference data. However, the reduction of cost-effectiveness using two spectrometers needs to be weighed against the potential increase in accuracy compared to a single MIR spectroscopy approach.     

Effects of climate factors and soil properties on soil nutrients and elemental stoichiometry across the Huang–Huai–Hai River Basin,China

Purpose

Soil nutrients, elemental stoichiometry, and their associated environmental control play important roles in nutrient cycling. The objectives of this study were (1) to investigate soil nutrients and elemental stoichiometry, especially potassium and its associative elemental stoichiometry with other nutrients under different land uses in terrestrial ecosystems; (2) to discuss the impacts of climate factors, soil texture, and soil physicochemical properties; and (3) to identify the key factors on soil nutrient levels and elemental stoichiometry.

Materials and methods

Soil data, including pH, bulk density (BD), cation exchange capacity (CEC), volumetric water content (VMC), clay, silt and sand contents, total carbon (TC), nitrogen (TN), phosphorous (TP) and potassium (TK), available nitrogen (AN), phosphorus (AP), potassium (AK), and soil organic matter (SOM) under different land-use types, were collected, and their elemental stoichiometry ratios were calculated. Climate data including temperature, precipitation, relative humidity, wind speed, and evapotranspiration were collected. The least significant difference test and one-way analysis of variance were applied to investigate the variability of soil nutrients and elemental stoichiometry among land-use types; the ordinary least squares method and the general linear model were used to illustrate the correlations between soil nutrients, elemental stoichiometry, and soil properties or climate factors and to identify the key influencing factors.

Results and discussion

Woodlands had the highest SOM, TN, AN, and AK contents, followed by grasslands, croplands, and shrublands, while the TP and TK contents only varied slightly among land-use types. SOM, TN, AN, N/P, and N/K were strongly negatively correlated to soil pH (p <?0.05) and were strongly positively correlated to soil CEC (p <?0.05). For soil texture, only C/N was moderately negatively correlated to silt content but moderately positively correlated to sand content (p <?0.05). For climate factors, SOM, TN, AN, N/P, and N/K were significantly negatively correlated to evapotranspiration and temperature (p <?0.05), and the correlations were usually moderate. Soil pH explained most of the total variation in soil nutrients, and climate factors explained 5.64–28.16% of soil nutrients and elemental stoichiometry (except for AP (0.0%) and TK (68.35%)).

Conclusions

The results suggest that climate factors and soil properties both affect soil nutrients and elemental stoichiometry, and soil properties generally contribute more than climate factors to soil nutrient levels. The findings will help to improve our knowledge of nutrient flux responses to climate change while also assisting in developing management measures related to soil nutrients under conditions of climate change.

     

Transferability and scalability of soil total carbon prediction models in Florida, USA

The applicability, transferability, and scalability of visible/near-infrared (VNIR)-derived soil total carbon (TC) models are still poorly understood. The objectives of this study were to: i) compare models of three multivariate statistical methods, partial least squares regression (PLSR), support vector machine (SVM), and random forest methods, to predict soil logarithm-transformed TC (logTC) using five fields (local scale) and a pooled (regional-scale) VNIR spectral dataset (a total of 560 TC spectral datasets), ii) assess the model transferability among fields, and iii) evaluate their up- and downscaling behaviors in Florida, USA. The transferability and up- and downscaling of the models were limited by the following factors: i) the spectral data domain, ii) soil attribute domain, iii) methods that describe the internal model structure of VNIR-TC relationships, and iv) environmental domain space of attributes that control soil carbon dynamics. All soil logTC models showed excellent performance based on all three methods with R² > 0.86, bias < 0.01%, root mean squared error (RMSE) = 0.09%, residual predication deviation (RPD) > 2.70%, and ratio of prediction error to interquartile range (RPIQ) > 4.54. The PLSR method performed substantially better than the SVM method to scale and transfer the TC models. This could be attributed to the tendency of SVM to overfit models, while the asset of the PLSR method was its robustness when the models were validated with independent datasets, transferred, and/or scaled. The upscaled soil TC models performed somewhat better in terms of model fit (R²), RPD, and RPIQ, whereas the downscaled models showed less bias and smaller RMSE based on PLSR. We found no universal trend indicating which of the four limiting factors mentioned above had the most impact that constrained the transferability and scalability of the models. Given that several factors can impinge on the empirically derived soil spectral prediction models, as demonstrated by this study, more focus on their applicability and scalability is needed.     

Investigation of soil surface organic and inorganic carbon contents in a low-intensity farming system using laboratory visible and near-infrared spectroscopy

ABSTRACT

In some regions of Italy, low-intensity farming systems, together with variable climate conditions, have lowered soil organic carbon (SOC) content and soil quality attributes. This work aims to investigate on some aspects of (1) total organic carbon (TOC) prediction using Vis-NIR reflectance spectroscopy in combination with partial least squares regression (PLSR); (2) the most appropriate pre-processing techniques of Vis-NIR absorbance spectra; (3) the composition of organic carbon using variable importance of prediction (VIP). The study area was an olive grove, located at Montecorvino Rovella (Salerno, southwestern Italy), characterized by a calcaric soil (Leptic Calcisols) and (Luvic Phaeozem), with a low content of TOC (mean 2.03 g kg^?1), caused by a low-intensity farming. Results of univariate PLSR analyses showed a good agreement between measured and predicted values both for TOC (R²: 0.66) and total carbonate content (R²: 0.93), when pH, electrical conductivity (EC) and absorbance spectra were used as predictors. The best results were obtained using as pre-treatments of the spectral data: 1) standard normal variate (SNV); 2) Savitzky-Golay algorithm; 3) first derivative. Variable Importance for Prediction (VIP) statistics showed to be a good tool to gain insights in TOC composition also when its content is low and influenced by carbonate.     

Using pXRF and vis-NIR spectra for predicting properties of soils developed in loess

Visible near-infrared (vis-NIR) and portable X-ray fluorescence (pXRF) spectrometers have been increasingly utilized for predicting soil properties worldwide. However, only a few studies have focused on splitting the predictive models by horizons to evaluate prediction performance and systematically compare prediction performance for A, B, and combined A+B horizons. Therefore, we investigated the performance of pXRF and vis-NIR spectra, as individual or combined, for predicting the clay, silt, sand, total carbon (TC), and pH of soils developed in loess, and compared their prediction performance for A, B, and A+B horizons. Soil samples (176 in A horizon and 172 in B horizon) were taken from Mollisols and Alfisols in 136 pedons in Wisconsin, USA and analyzed for clay, silt, sand, pH, and TC. The pXRF and vis-NIR spectrometers were used to measure the pXRF and vis-NIR soil spectra. Data were separated into calibration (n=244, 70%) and validation (n=104, 30%) datasets. The Savitzky-Golay filter was applied to preprocess the pXRF and vis-NIR spectra, and the first 10 principal components (PCs) were selected through principal component analysis (PCA). Five types of predictor, i.e., PCs from vis-NIR spectra, pXRF of beams at 0-40 and 0-10 keV (XRF40 and XRF10, respectively) spectra, combined XRF40 and XRF10 (XRF40+XRF10) spectra, and combined XRF40, XRF10, and vis-NIR (XRF40+XRF10+vis-NIR) spectra, were compared for predicting soil properties using a machine learning algorithm (Cubist model). A multiple linear regression (MLR) model was applied to predict clay, silt, sand, pH, and TC using pXRF elements. The results suggested that pXRF spectra had better prediction performance for clay, silt, and sand, whereas vis-NIR spectra produced better TC and pH predictions. The best prediction performance for sand (R²=0.97), silt (R²=0.95), and clay (R²=0.84) was achieved using vis-NIR+XRF40+XRF10 spectra in B horizon, whereas the best prediction performance for TC (R²=0.93) and pH (R²=0.79) was achieved using vis-NIR+XRF40+XRF10 spectra in A+B horizon. For all soil properties, the best MLR model had a lower prediction accuracy than the Cubist model. It was concluded that pXRF and vis-NIR spectra can be successfully applied for predicting clay, silt, sand, pH, and TC with high accuracy for soils developed in loess, and that spectral models should be developed for different horizons to achieve high prediction accuracy.