基于人工神经网络的土壤有机质含量高光谱反演

研究了土壤有机质含量与土壤高光谱之间的关系,在对原始光谱进行了预处理分析后,运用多元线性逐步回归法(MLSR)和人工神经网络法(ANN)建立了土壤有机质含量的反演模型,并对模型进行了验证。结果表明:人工神经网络所建立的反演模型普遍优于回归模型,网络集成模型优于单个BP网络模型,网络集成是提高反演模型准确性与稳定性的有效途径。网络集成模型为最优模型,总均方根误差为1.31,可以用于土壤有机质含量的快速测算。     

Relationships of barley biomass and grain yields to soil properties within a field in the arid region: Use of factor analysis

Abstract

Understanding the variability of soil properties and their effects on crop yield is a critical component of site-specific management systems. The objective of this study was to employ factor and multiple regression analyses to determine major soil physical and chemical properties that influence barely biomass and grain yield within a field in the arid region of northern Iran. For this purpose, soil samples and crop-yield data were collected from 108 sites, at regular intervals (20×30 m) in a 5.6 ha field. Soil samples were analysed for total nitrogen (TN), available phosphorus (P_ava), available potassium (K_ava), cation-exchange capacity(CEC), electrical conductivity (EC), pH, mean weight diameter of aggregates (MWD), water-stable aggregates (WSA), field capacity volumetric (FC), available water-holding capacity (AWHC), bulk density (BD), and calcium carbonate equivalent (CCE). Results of the factor analysis, followed by regression of biomass and grain yield of barley with soil properties, showed that the regression equations developed accounted for 78 and 73% of the total variance in biomass and grain yield, respectively. Study of covariance analysis among soil variables using factor analysis indicated that some of the variation measured could be grouped to indicate a number of underlying common factors influencing barley biomass and grain yields. These common factors were salinity and sodicity, soil fertility, and water availability. The most effective soil variables to barley production in the study area identified as EC, SAR, pH, TN, P_ava, AWHC, and FC. In this study, factor analysis was effective to identify the groups of correlated soil variables that were significantly correlated with the within field variability in the yield of the barley crop. Our results also suggest that the approach can be applied to other crops under similar soil and agroclimatic conditions.     

Soil Particulate Organic Matter (POM) Prediction in a Mountainous Watershed using Artificial Neural Networks

Soil organic matter is a very important component of soil that supports the sustainability and quality in all ecosystems, especially in arid and semi-arid regions. A comparison study was carried out to verify when the artificial neural network (ANN) and multiple linear regression (MLR) models are appropriate for the prediction of soil organic matter (SOM) and particulate organic matter (POM). Discussions of advantages and disadvantages are given for both methods. Three different sets of easily available properties (soil properties alone, topographic and vegetation index, a combination of soil and topographic data) were used as inputs and the one affecting the model the most was determined. The smallest prediction errors were obtained by the ANN method; however, the prediction accuracies of the constructed MLR models using different data sets were closed to the ANN models in many cases.     

Barley yield response to soil organic matter and texture in the Pampas of Argentina

Soil organic matter (SOM) is known to play a major role in soil fertility due to its influence on physical, chemical and biological properties of soil; and it is closely related to particle size distribution. The ratio of SOM (g kg⁻¹) to clay + silt content (g kg⁻¹) was evaluated as an indicator of soil quality for barley (Hordeum vulgare) grain yield, reflecting N availability and soil physical conditions to which crop development is sensitive. Thirty-eight sites in the semiarid Pampa region of Argentina with a wide range of SOM and texture were evaluated for malting barley yield during three growing seasons. In control plots, 51% of grain yield could be explained by this indicator. The threshold value between high and low N-fertilization response was 4.4. Better yield prediction to almost 68% was achieved by combining the SOM to clay + silt indicator with initial nitrate content of the soil at seeding. This combined indicator was also able to explain a high proportion of water use efficiency, particularly in the early growth stages. The ratio of SOM to clay + silt content provided a better tool for estimating grain yield than nutrient availability or SOM alone.     

Organic matter in volcanic soils in Chile: Chemical and biochemical characterization

Abstract

Determinations were made of total soil organic matter (SOM), stable and labile organic fractions, biomass carbon (C), and chemical composition of several humus‐soil‐fractions in Chilean volcanic soils, Andosols and Ultisols. Their physico‐chemical properties and humification degree at different stages in edaphic evolution were also assessed. In addition, organic matter models were obtained by chemical and biological syntheses and the structures and properties of natural and synthetic humic materials were compared with SOM. Results indicate that Andosols have higher SOM levels than Ultisols, but the fraction distribution in the latter suggests a shift of the more stable fractions to the more labile ones. Moreover, contents of humines, and humic and fulvic acids suggest that Chilean volcanic soil SOM is highly humified. On the other hand, among the SOM labile fractions, carbohydrate and biomass are about 15% of the SOM which are one of the most important fractions in soil fertility.     

Soil-Quality Indicators for Predicting Sustainable Organic Rice Production

Soil physical, chemical, and biological properties in a rice field located at the Surin Rice Research Center, Thailand, were evaluated as indicators for predicting organic rice (Kao Dok Mali 105 variety) production and yield. Four treatments under different management practices were studied. They included (1) conventional farming (CF) receiving chemical fertilizer application; (2) organic plot receiving green manure (GM) addition; (3) organic plot receiving rice straw (RS) addition; and (4) control plot (CT) without any external plant nutrient source. Soil quality in the four treatments was assessed based upon selected physical, chemical, and biological parameters. Key findings are as follows: cation exchange capacity (CEC), electrical conductivity (EC), pH, soil organic matter (SOM), and essential macronutrients [nitrogen (N), phosphorus (P), and potassium (K)] were low in all plots. Soil biological properties including potential N mineralization (PMN), soil basal respiration (BR), microbial biomass carbon (MBC) and microbial biomass N (MBN) in all treatments were also low. Principal component analysis (PCA), using 15 soil properties, showed significant differences among farm management practices. Soil chemical and biological properties best related to soil quality included P, N, and SOM (for chemical properties) and MBC, MBN, and BR (for biological properties). Based on significant relationships between yield (r > 0.75) and the soil properties (r > 0.55), selected soil biological (MBC, MBN, and BR) and chemical (TOP [total organic phosphorus], TK [total potassium], TN [total nitrogen], SOC [soil organic carbon], and SOM) properties were determined to be suitable soil-quality indicators, respectively. A soil-quality indicator for predicting rice yield was computed using multiple regression analyses. The regression model (Y = ?1.685 + 0.333 (MBN) + 0.640 (TK) ? 0.282 (SOC), r² _adjusted = 0.962) was used for predicting yield. Grain yield of rice (RMSE = 0.046 t ha^?1, D index = 0.45) was obtained using this regression model.     

Effects of Different Organic Materials on Crop Production under a Rice–Corn Cropping Sequence

Soil organic matter (SOM) is an important index of soil quality because of its relationship with crop yield. The application of organic matter to soil is a significant method for increasing SOM. Different organic materials have varying effects in increasing SOM. This study investigates the effects of combining different sources of organic matter (i.e., compost, leguminous green manure, and peat) with a chemical nitrogen (N) fertilizer on the growth and N accumulation in corn and rice plants. This study examines seven treatments, including a no-fertilization check and a conventional chemical fertilizer treatment. Shoots of corn and rice were sampled at the tasseling (panicle initiation for rice) and maturity stages. The biomass yield was measured and the total N was analyzed. At the maturity stage, the soil samples were collected to determine the chemical properties. The results showed that a small percentage of the N in the compost and peat, after their application, was available to the crop during the growth season; the production of biomass and N absorption among rice and corn plants was minimal compared to that treated with chemical N fertilizer. The application of compost and peat resulted in SOM accumulation, particularly with peat. However, the application of compost combined with chemical fertilizer not only produced sufficient nutrients for crop growth but also resulted in an accumulation of SOM, which is vital for enhancing the soil quality. Most of the N in green manure (GM) was mineralized shortly after application, causing excessive growth of rice and corn plants during the early stage, but reducing their reproductive growth and grain yield.     

Spatial variability of selected soil properties and its impact on the grain yield of oats (Avena sativa L.) in small fields

Abstract

To investigate spatial variability in topsoil (0–20?cm) pH, available phosphorus (P), potassium (K), total nitrogen (N), and soil organic matter (SOM) of small fields (～2?ha), and to determine the impact of soil heterogeneity on the spatial variability of crop yield two fields were cropped with spring oats and one with winter wheat under humid-temperate conditions. In the two oat fields, some of the measured soil properties (P, K) and the grain yield varied considerably, and strong spatial trends were recorded for most of the soil traits. In the third field, soil properties showed only a moderate spatial variation, and no spatial trends were found. The spatial distribution of SOM and total N in the topsoil had some influence on the spatial pattern of the oat grain yield in the field of Gränichen; however, spatial relationships between soil chemical properties and grain yield were rather weak in our study.     

What are the primary factors controlling the light fraction and particulate soil organic matter content of agricultural soils?

Particulate organic matter (POM) and light fraction organic matter (LFOM) are the fractions of soil organic matter (SOM) considered most active in terms of nutrient cycling and maintenance of soil structure. They respond quickly to changes in management and may offer insights into the long-term effect of management on SOM. However, the literature provides contradictory evidence regarding the factors which influence the amount of POM and LFOM, and there is little evidence to differentiate the relative importance of factors. Utilising data from over 150 experiments reported in the literature, we employed multiple regression to produce separate models quantifying the effect of management factors and environmental variables on POM, LFOM and total SOM; 29.3 % of the variance in the response variables was explained for POM, 28.3 % for LFOM, and 29.3 % for total SOM. Climate, organic amendments and inclusion of fallow periods were significant terms for all fractions. Climate had a larger influence on total SOM than POM or LFOM, whilst POM and LFOM were more strongly influenced by factors related to the recent history of organic matter addition; organic amendments and inclusion of fallows. Factors that were not significant variables for any of the fractions included tillage and application of N fertiliser, whilst soil texture was only a significant factor for SOM. General agreement between the total SOM, POM and LFOM models on the most important factors supports the idea that both POM and LFOM are good predictors of long-term changes to total SOM.     

Interactions between soil organic matter level and soil tillage in a growing crop: N mineralization and yield response

Four levels of soil organic matter (SOM) had been established on a coarse sandy loam after application of four combinations of mineral fertilizer, animal manure, straw incorporation and catch crops for 12 years. Soil tillage was carried out in a growing spring barley crop (Hordeum vulgare) to examine the potential for improving the synchrony between soil N mineralization and crop N demand. Tillage raised soil nitrate concentrations temporarily but did not influence barley dry matter (DM) yield. At maturity, both grain DM yield and N uptake were largest on soil with the highest OM level. The previous OM applications had a pronounced influence on crop development and N availability, but soil tillage did not significantly improve the synchrony between soil N mineralization and crop N demand.     

Contributions from carbon and nitrogen in roots to closing the yield gap between conventional and organic cropping systems

This study investigates the effect of different crop rotation systems on carbon (C) and nitrogen (N) in root biomass as well as on soil organic carbon (SOC ). Soils under spring barley and spring barley/pea mixture were sampled both in organic and conventional crop rotations. The amounts of root biomass and SOC in fine (250–253 μ m), medium (425–250 μ m) and coarse (>425 μ m) soil particulate organic matter (POM ) were determined. Grain dry matter (DM ) and the amount of N in harvested grain were also quantified. Organic systems with varying use of manure and catch crops had lower spring barley grain DM yield compared to those in conventional systems, whereas barley/pea showed no differences. The largest benefits were observed for grain N yields and grain DM yields for spring barley, where grain N yield was positively correlated with root N. The inclusion of catch crops in organic rotations resulted in higher root N and SOC (g C/m²) in fine POM in soils under barley/pea. Our results suggest that manure application and inclusion of catch crops improve crop N supply and reduce the yield gap between conventional and organic rotations. The observed positive correlation between root N and grain N imply that management practices aimed at increasing grain N could also increase root N and thus enhance N supply for subsequent crops.     

基于三种空间预测模型的海南岛土壤有机质空间分布研究

为探索适合热带地形复杂区土壤有机质(SOM)含量的空间预测方法,以海南岛为研究区域,结合地形因子、归一化植被指数、土壤类型、土地利用类型变量,选用普通克里格法(OK)、回归克里格法(RK)、随机森林模型(RF)三种方法对训练集128个样点SOM含量的空间分布规律进行预测,并通过验证集32个验证点比较了三种方法的预测精度。结果表明:(1)0~5 cm土层三种方法的平均预测误差(ME)均接近于0,从均方根预测误差(RMSE)来看,RF(0.8867)RK(0.910 4)OK(0.9641),从决定系数(R~2)来看,RF(0.214 1)RK(0.171 5)OK(0.070 8)。综合以上三个参数,该土层最优拟合模型为RF。同理得出0~20、20~40、40~60 cm土层的最优拟合模型分别为RF、RF、OK。RK和RF能够更好地描述SOM含量局部变异信息;(2)四个土层SOM含量的均值分别为19.67、15.89、10.30、8.07 g kg~(-1),呈现出西南、东北高,西部、东南沿海地区低的空间分布趋势。     

Relationship Between Soil Properties and Nitrogen Mineralization in Undisturbed Soil Cores from California Agroecosystems

Soil nitrogen (N) mineralization rates from different agricultural regions in California were determined and related to soil properties. Undisturbed soil cores were sampled in spring from 57 fields under annual crop rotations and incubated at 25℃ for 10 weeks. Soil properties varied across and within regions, most notably those related to soil organic matter (SOM), with total soil carbon ranging from 6 to 198 g kg^?1. Multivariate linear regression was used to select soil properties that best predicted N mineralization rates. Regression models with a good fit differed between soils with high and low SOM contents, but generally included a measure of SOM quantity, its quality as well as soil texture or mineralogy. Adjusted R² values were 0.95 and 0.60 for high and low SOM soils, respectively. This study has shown that information on soil properties can contribute to better estimates of N mineralization in soils of contrasting characteristics.     

黄淮海平原集约种植条件下土壤有机质动态模拟

A modified CQESTR model, a simple yet useful model frequently used for estimating carbon sequestration in agricultural soils, was developed and applied to evaluate the effects of intensive cropping on soil organic matter （SOM） dynamics and mineralization as well as to estimate carbon dioxide emission from agricultural soils at seven sites on the Huang-Huai-Hai Plain of China. The model was modified using site-specific parameters from short- and mid-term buried organic material experiments at four stages of biomass decomposition. The predicted SOM results were validated using independent data from seven long-term （10- to 20-year） soil fertility experiments in this region. Regression analysis on 1 151 pairs of predicted and measured SOM data had an r2 of 0.91 （P≤0.01）. Therefore, the modified model was able to predict the mineralization of crop residues, organic amendments, and native SOM. Linear regression also showed that SOM mineralization rate （MR） in the plow layer increased by 0.22% when annual crop yield increased by 1 t ha^-1 （P ≤ 0.01）, suggesting an improvement in SOM quality. Apparently, not only did the annual soil respiration efftux merely reflect the intensity of soil organism and plant metabolism, but also the SOM MR in the plow layer. These results suggested that the modified model was simple yet valuable in predicting SOM trends at a single agricultural field and could be a powerful tool for estimating C-storage potential and reconstructing C storage on the Huang-Huai-Hai Plain of China.     

两种施肥处理下不同有机质含量农田黑土微生物群落结构PLFA分析

土壤有机质含量和施肥是影响黑土微生物群落结构的重要因素,但是受气候影响,很难单独明确有机质含量或施肥对土壤微生物群落的影响。本研究利用黑土生产力长期定位试验,将有机质含量不同的5个黑土(SOM1.7、SOM3、SOM5、SOM6、SOM11)置于相同气候条件下,通过分析磷脂脂肪酸,系统地研究了施肥与有机质含量对农田黑土微生物群落结构的影响。研究结果表明,5个有机质含量农田黑土中,土壤磷脂脂肪酸总量为10.6~31.5 nmol·g^-1,细菌磷脂脂肪酸含量为6.23~18.4 nmol·g^-1,真菌磷脂脂肪酸总量为1.78~4.57 nmol·g^-1。土壤有机质含量升高和施肥会显著提高土壤中总微生物量、细菌生物量和真菌生物量,但施肥和有机质含量对真菌/细菌比值无显著影响。非度量多维尺度分析(NMDS)分析表明,有机质含量和施肥是导致微生物群落结构差异的重要因素,但施肥可能会遮蔽有机质含量对微生物群落的影响。     

日本北海道地区森林、牧场、草原和玉米田温室气体排放及模型研究

Information on the most influential factors determining gas flux from soils is needed in predictive models for greenhouse gases emissions.We conducted an intensive soil and air sampling along a 2 000 m transect extending from a forest,pasture,grassland and corn field in Shizunai,Hokkaido (Japan),measured CO 2 ,CH 4 ,N 2 O and NO fluxes and calculated soil bulk density (ρ b ),air-filled porosity (f a ) and total porosity (Φ).Using diffusivity models based on either f a alone or on a combination of f a and Φ,we predicted two pore space indices: the relative gas diffusion coefficient (D s /D o ) and the pore tortuosity factor (τ).The relationships between pore space indices (D s /D o and τ) and CO 2 ,CH 4 ,N 2 O and NO fluxes were also studied.Results showed that the grassland had the highest ρ b while f a and Φ were the highest in the forest.CO 2 ,CH 4 ,N 2 O and NO fluxes were the highest in the grassland while N 2 O dominated in the corn field.Few correlations existed between f a ,Φ,ρ b and gases fluxes while all models predicted that D s /D o and τ significantly correlated with CO 2 and CH 4 with correlation coefficient (r) ranging from 0.20 to 0.80.Overall,diffusivity models based on f a alone gave higher D s /D o ,lower τ,and higher R 2 and better explained the relationship between pore space indices (D s /D o and τ) and gases fluxes.Inclusion of D s /D o and τ in predictive models will improve our understanding of the dynamics of greenhouse gas fluxes from soils.D s /D o and τ can be easily obtained by measurements of soil air and water and existing diffusivity models.     

黑河流域保护性耕作对农田土壤微生物量碳、氮和有机质的调控

为了探讨黑河流域保护性耕作对土壤生产力的影响,设计20cm留茬(NS20),20cm留茬压倒(NPS20),40cm留茬(NS40),40cm留茬压倒(NPS40)和传统耕作(CT)5个处理,研究了黑河流域保护性耕作对农田土壤有机质、土壤微生物量C、土壤微生物量N以及作物产量和水分利用效率的影响。结果表明,保护性耕作农田0—20cm土层土壤有机质、土壤微生物量C和N的含量均高于传统耕作,且其在剖面中的变化趋势基本一致,即随土层深度增加下降;土壤微生物量N有明显的"表聚现象";相关分析表明土壤有机质和土壤微生物量C之间显著正相关(r=0.85,p0.05),与土壤微生物量N之间无明显的相关关系(r=0.47,p0.05);保护性耕作提高了春小麦的产量,NPS20和NPS40增产效果最好,较CT分别增产53.08%和46.59%,与CT之间差异达到极显著水平;保护性耕作提高了春小麦的水分利用效率(WUE),NPS20,NS40,NPS40,NS20分别较CT的WUE提高了58.02%,43.40%,47.27%,23.78%。     

基于主成分回归分析的土壤有机质高光谱预测与模型验证

在室内条件下,利用ASD2500高光谱仪测定了风干土壤样品的光谱。通过相关分析对土壤有机质（SOM）光谱敏感波段进行了初步筛选;利用逐步回归分析和主成分回归(PCR)分析等统计方法进行了显著性变量筛选、共线性诊断、数据转换等处理;最终建立了东北黑土SOM回归预测模型。模型所选的波段为均位于近红外波段。经验证,模型预测值与实测值的决定系数R2=0.840,总均方根差RSME=0.226。     

滩涂土壤有机质含量的反射光谱估算

Rapid determination of soil organic matter (SOM) using regression models based on soil reflectance spectral data serves an important function in precision agriculture. “deviation of arch”(DOA)-based regression and partial least squares regression (PLSR) are two popular modeling approaches to predict SOM. However, few studies have explored the accuracy of the DOA-based regression and PLSR models. Therefore, the DOA-based regression and PLSR were applied to the visible near-infrared (VNIR) spectra to estimate SOM content in the case of various dataset divisions. A two-fold cross-validation scheme was adopted and repeated 10 000 times for rigorous evaluation of the DOA-based models in comparison with the widely used PLSR model. Soil samples were collected for SOM analysis in the coastal area of northern Jiangsu Province, China. The results indicated that both modelling methods provided reasonable estimates of SOM, with PLSR outperforming DOA-based regression in general. However, the performance of PLSR for the validation dataset decreased more noticeably. Among the four DOA-based models, the linear model of the DOA provided the best estimation of SOM and a cutoff of SOM content (19.76 g kg^-1), and the performance for calibration and validation datasets was consistent. As the SOM content exceeded 19.76 g kg^-1, SOM became more effective in masking the spectral features of other soil properties to a certain extent. This work confirmed that reflectance spectroscopy combined with PLSR could serve as a non-destructive and cost-efficient way for rapid determination of SOM when hyperspectral data were available. The DOA-based model, which requires only 3 bands in the visible spectra, also provided SOM estimation with acceptable accuracy.     

高产农田土壤有机质、养分的变化规律与作物产量的关系

以山东桓台县高产农田为研究对象 ,对全县 1982～ 1998年的平均作物产量、土壤有机质、速效养分、化肥施用量等进行了调查和分析。结果表明 ,从 1982年到 1998年 ,桓台县农田土壤有机质处于不断增加的状态 ,但增加的幅度随着秸秆还田的普及而逐渐变小。目前全县县域农田土壤有机质平均水平为 15.0 gkg-1,有可能是当前耕作方式、土壤类型和气候条件下土壤有机质的平衡点。研究还表明 ,高产条件下土壤有机质与作物产量之间存在显著正相关关系。在吨粮田发展过程中 ,土壤速效氮和速效钾对作物产量的贡献逐渐降低 ,而土壤速效磷的贡献呈增加趋势。