光谱技术结合水分校正与样本增广的棉田土壤盐分精准反演

棉田土壤盐分的精准反演对于棉花的种植管理具有重要意义。水分和盐分作为主要环境因素，共同影响棉田土壤的波谱特征，两者之间的耦合关系直接影响土壤盐分的检测分析。为了提高基于光谱技术构建的模型对棉田土壤盐分信息解析的准确性与可靠性，该研究联用可见/短波近红外（400～1 000 nm）和长波近红外（960～1 693 nm）技术，采集不同含水率与含盐量的新疆地区土壤样本的光谱；结合外部参数正交法（external parameter orthogonalization，EPO），校正不同标样集与不同波段光谱中的土壤含水率干扰信息；引入基于不同卷积步幅的深度卷积对抗网络（deep convolutional generative adversarial networks，DCGAN），进行样本增广与质量评估；参考三层残差神经网络设计一维卷积神经网络RNet，最终构建基于EPO-DCGAN-RNet的优化模型，用于棉田土壤盐分的反演。结果表明，与传统机器学习方法和基于VGG或EfficientNet结构一维卷积神经网络相比，该研究提出的EPO-DCGAN-RNet方法能够有效地滤除水分对盐分反演的影响、提高模型对特征波段的挖掘能力、降低深度学习算法对样本量的依赖性，并能得到更优的模型预测性能。EPO-DCGAN-RNet的建模集R²和均方根误差分别为0.942、115.420 μS/cm，验证集R²和均方根误差分别为0.910和136.472 μS/cm。研究结果可为新疆棉田土壤盐分快速精准检测提供理论指导和技术支持，有助于促进盐碱地区棉花种植的水肥科学管理。     

Soil and sediments listservers

Internet listservers, a freely available resource on the Internet for anyone who can send and receive emails, provide a means for professionals to communicate and collaborate with each other in the field of soils and sediments. While moderated listservers are highly effective, in terms of quality and participation (activity) level, do not underestimate the knowledge base held within many of the soil and sediment-related listservers. Readers are encouraged to subscribe to those lists which match their own areas of the soil and sediment fields and to become active. Currently, the only listservers dedicated to sediments are of very low activity, with the lowest numbers of subscribers of all listservers surveyed. As a word of caution, Internet listservers are not a primary source of knowledge (in any scientific field), the quality of knowledge exchanged is highly variable, and if not properly managed, can waste time.     

Microbial and Enzyme Activities of Saline and Sodic Soils

Salinization and sodication are abiotic soil factors, important hazards to soil fertility and consequently affect the crop production. Soil salinization is of great concern for irrigated agriculture in arid and semi‐arid regions of the world; sodicity is characterized by an excessively high concentration of sodium (Na) in their cation exchange system. In recent times, attention has been turned to study the impacts of these factors (salinity and sodicity) on soil microbial activities. Microbial activities play central role in degradation and decomposition of soil organic matter, mineralization of nutrients and stabilization of soil aggregates. To understand the ecology of soil system, therefore, it is important to be conversant with the soil microbial activities, which show quick response to little change in the soil environment. Microbial activities (generally measured as C–N dynamics, soil respiration–basal respiration, or CO₂ emission), microbial abundance, microbial biomass, quotients (microbial and metabolic) and microbial community structure, and soil enzymes have been considered as potential indicators to assess the severity of the land degradation and the effectiveness of land use management. Therefore, it is important to synthesize the available information regarding microbial activities in use and management of salt‐affected soils. The reclamation and management of such soils and their physico‐chemical properties have been reviewed well in the literature. In this review, an attempt has been made to compile the current knowledge about the effects of soil salinization and sodication on microbial and enzyme activities and identify research gaps for future research. Copyright © 2015 John Wiley & Sons, Ltd.     

Tillage Effects on Crop Yield and Physicochemical Properties of Sodic Soils

Tillage modifies soil structure and has been suggested as a practice to improve physical, hydrological and chemical properties of compacted soils. But little is known about effect of long‐term tillage on physicochemical soil properties and crop yield on sodic soils in India. Our objective was to investigate the effect of different tillage regimes on crop yield (wheat and paddy rice) and physicochemical properties of sodic soils. Two sodic sites under conventional tillage for 5 (5‐YT; 5‐year tillage) and 9 (9‐YT; 9‐year tillage) years were selected for this study. Changes in crop yield and physicochemical soil properties were compared with a control, sodic land without any till history, that is, 0‐year tillage/untilled (0‐YT). Five replicated samples at 0‐ to 10‐cm and 10‐ to 20‐cm soils depths were analysed from each site. In the top, 0‐ to 10‐cm soil depth 5‐YT and 9‐YT sites had higher particle density (Pd), porosity, water holding capacity, hydraulic conductivity, organic carbon, total nitrogen (N_t), available nitrogen (N_avail), phosphorus (P_avail) and exchangeable calcium (Exch. Ca⁺⁺) than 0‐YT, whereas bulk density (Bd), C : N ratio and CaCO₃ were significantly lower. Bd, pH, EC and CaCO₃ increased significantly with depth in all the lands, whereas Pd, porosity, water holding capacity, hydraulic conductivity, organic carbon, N_t, N_avail, P_avail and Exch. Ca⁺⁺ decreased. We conclude that continuous tillage and cropping can be useful for physical and chemical restoration of sodic soils. Copyright © 2014 John Wiley & Sons, Ltd.     

陕西土壤脲酶热力学特征研究

根据土壤脲酶动力学参数与温度的关系,借助于经典酶学理论,较为系统地研究了陕西水稻土、褐土、土脲酶的热力学特征。结果表明,将整个酶促反应划分为酶-底物复合物形成和分解两个阶段分别计算其热力学函数是可行的,得到了更多酶促反应机理方面的信息;不同土样脲酶复合物形成和分解阶段的温度系数Q10、活化能Ea、活化焓ΔH、活化熵ΔS和活化自由能ΔG有明显差异;土壤脲酶的Q10范围为0.82~2.48;两阶段土壤脲酶除个别处理外Ea、ΔH1、ΔH2、ΔS均大于零,揭示出形成和分解阶段酶促反应的有序性较差,混乱度较大,需要外界提供更多能量来完成酶活性部位与反应物的互补契合以及解离,最终完成酶促反应;形成阶段ΔG1和整个反应的ΔG、ΔH均小于零,表明形成阶段和整个酶促反应为一可自发进行的放热反应,总体上不需要外界能量即可完成;各土样分解阶段的ΔG2均大于零,表明土壤类型、肥力状况等对酶促反应的复合物分解过程的可能性无显著影响,复合物分解是酶促反应的限速步骤。可见得到的热力学参数从多方面共同表征了各土样酶促反应机理上的细微差异。     

漳州市农业土壤重金属现状、分析及防治

调查漳州市农业土壤重金属Hg、As、Cr、Cd、Pb背景值及污染现状,调查结果,漳州市农业土壤重金属背景值相对较低。其现状为,Hg、As、Cd、Pb含量高于背景值1~4倍(Cr除外),特别是Hg和As为高污染元素。按土壤类型划分,水稻土土类重金属含量高于赤红壤土类(Hg除外);按土地利用方式划分,水田>菜地>果园;按流域划分,九龙江流域农业土壤重金属含量高于非九龙江流域(Cr除外)。污染源以农业污染为主,工业污染相对较轻。     

湖南旱地土壤对铵离子的矿物固定与固定态铵释放的动力学研究

采用振荡平衡法和好气培养-间歇淋洗法研究了湖南省主要旱地土壤对添加铵离子的矿物学固定和土壤固定态铵释放的动力学特性。结果表明,供试土壤对添加铵离子的矿物固定速度很快,反应12小时后基本上达到了最大固定量;而土壤固定态铵的释放则非常缓慢,反应42天后仍未达到最大释放量。这一动力学特征有利于土壤对溶液中NH4+的保持,防止NH4+的环境污染,有效地为作物供氮和提高氮肥利用率。一级动力学方程能较好地拟合供试土壤对铵离子的矿物学固定的动力学资料,一级动力学方程和抛物扩散率方程均能较好地拟合供试土壤固定态铵释放的动力学资料,表现为相关系数r值较高,标准误SE值较低。     

黑土区基准坡长和LS算法对地形因子的影响

该文结合小区实测坡长资料及Mc Cool数据,对地形因子算法中涉及的基准坡长影响和不同算法计算值的特征进行了讨论。研究表明,在坡长较小的情况下,20和22.13 m基准坡长计算得到的坡长因子相差不大,但随着坡度增加,在坡长指数的复合影响下,两者的差异渐趋增大。理论分析表明,坡长指数采用与坡度相关的变值更为合理。通过与实测数据对比发现,修正版通用土壤侵蚀模型(RUSLE)算法中的坡长因子与其更为接近,结合与Mc Cool数据的对比分析表明,研究区坡长指数更适合细沟和细沟间侵蚀比率中等的情形,或者采用RUSLE模型算法计算得到。同时研究发现,张宪奎算法、Moore算法、Desmet算法和B?hner算法计算得到的LS因子值都要比参照值小,而且4种算法与参照值的接近程度依次降低。空间分布特征而言,栅格累计算法总体较为破碎,并在平直坡面出现较多平行分布条带,而含有面积指标的其他3种算法则总体呈现光滑连续特性,但有无数据区域的存在。该结果对于研究区土壤侵蚀模型地形因子算法选择具有一定的现实意义。     

南方红壤区植被覆盖因子估算模型构建与验证

研究红壤区植被覆盖因子变化规律、构建植被覆盖因子计算模型可为该区域水土流失防治及动态监测提供科学依据。该研究基于广东省五华县水土保持试验推广站200余场降雨试验观测数据,分析了影响水土流失的各关键因子与次降雨土壤侵蚀模数(Individual Rainfall Event Soil Erosion Modulus,ISEM)间定量关系,基于中国土壤流失方程计算了系列次降雨事件不同植被覆盖度对应植被覆盖因子值,构建了植被覆盖因子值与植被覆盖度间数学模型,从点、面2个尺度对模型计算精度进行了验证。结果表明:1)ISEM随植被覆盖度增加而降低;ISEM与雨前土壤表层含水率间存在显著正相关对数关系;ISEM随坡度增加呈先增后减的变化规律;ISEM随次降雨量、次降雨侵蚀力和次降雨径流深呈显著正相关线性变化关系。2)构建了植被覆盖因子值与植被覆盖度间二阶指数衰减模型,该模型决定系数和纳什系数分别达0.947和0.876。点尺度验证结果表明90%样本模型计算值与观测值相对误差均小于0.30;面尺度验证结果表明,70%～80%的植被覆盖因子计算值相对误差不超过0.1。总体而言,该模型计算精度较为理想,但由于研究对象典型的时空尺度特征,仍需要更多观测数据对该模型进行完善和验证。研究成果可为深入理解红壤区土壤侵蚀规律、水土流失动态监测提供有益参考。     

Multi-element analysis of soils and sediments by wavelength-dispersive X-ray fluorescence spectrometry

Soil and sediment reference materials were used to calibrate and evaluate an analytical method for the determination of major (Si, Al, Fe, Mg, Ca, Na, K, Mn, P, Ti) and trace elements (As, Ba, Cd, Co, Cr, Cu, Ga, Mo, Mb, Ni, Pb, Rb, S, Sb, Sn, Sr, Th, U, V, Y, Zn, Zr) by sequential wavelength X-ray fluorescence spectrometry. Samples were prepared as pressed pellets and analysis was done with a total measuring time of thirty minutes per sample. Special attention was given to the selection of the thirty reference materials used for calibration of the spectrometer. Another set of eleven RM (reference materials) was analyzed for the evaluation of accuracy. Detection limits for trace elements (1-2 mg kg^-1) are adequate both for geochemical and environmental purposes, except for Cd and Sb. Accuracy for trace elements falls within the expected interval of certified or recommended values in most cases, but for some major elements, like SiO₂, some results showed discrepancies, evidencing difficulties associated with the determination of light elements in complex matrices. But when quality criteria proposed by mapping programs are applied to the results, their requirements are fulfilled. Both instrumental precision, obtained by twelve sample replicate analyses, and analytical precision, considering also sub-sampling and pellet preparation, lie between the limits of the Horwitz expression, except at concentrations close to the detection limits.