基于无人机遥感影像的大豆叶面积指数反演研究

作物叶面积指数的遥感反演是农业定量遥感研究热点之一,利用无人机遥感监测系统获取农作物光谱信息精确反演叶面积指数对精准农业生产与管理意义重大。本研究以山东省嘉祥县一带的大豆种植区为试验区,设计以多旋翼无人机为平台同步搭载Canon Power Shot G16数码相机和ADC-Lite多光谱传感器组成的无人机农情监测系统开展试验,分别获取大豆结荚期和鼓粒期的遥感影像。使用比值植被指数(RVI)、归一化植被指数(NDVI)、土壤调整植被指数(SAVI)、差值植被指数(DVI)、三角植被指数(TVI)5种植被指数,结合田间同步实测叶面积指数(leaf area index,LAI)数据,采用经验模型法分别构建了单变量和多变量LAI反演模型,通过决定系数(R2)、均方根误差(RMSE)和估测精度(EA)3个指标筛选出最佳模型。研究表明,有选择性地分时期进行农作物的叶面积指数反演是必要的,鼓粒期作为2个生育期中大豆LAI反演的最佳时期,其NDVI线性回归模型对大豆LAI的解释能力最强,R2=0.829,RMSE=0.301,反演大豆LAI最准确,EA=85.4%,生成的鼓粒期大豆LAI分布图反映了当地当时大豆真实长势情况。因此,以多旋翼无人机为平台同步搭载高清数码相机和多光谱传感器组成的无人机农情监测系统对研究大豆叶面积指数反演是可行性,可作为指导精准农业研究的一种新方法。     

基于无人机多光谱影像的病害棉田产量损失估算

该研究利用无人机多光谱遥感影像对棉花黄萎病造成的产量损失进行估算,为棉花黄萎病预防和防治提供依据。对病害棉田进行调查,获取无人机多光谱影像及地面产量损失数据,利用相关系数法及灰度值标准差法分别筛选识别病害棉株的最佳植被指数、最佳波段组合;基于筛选的结果建立棉田综合影像（最佳波段组合与差值植被指数综合影像）,利用支持向量机径向基核函数分类法对病害棉田原始影像和综合影像进行产量空间分布分析及产量损失估算。结果表明,无人机多光谱影像识别病害棉田的最佳植被指数、最佳波段组合分别是差值植被指数（相关系数为-0.86）、波段B3（550～10 nm）、B5（656～10 nm）、B8（800～10 nm）的波段组合(B3-B5-B8)（最佳指数因子为153.44）;综合影像较原始影像更能准确识别病害棉田产量空间分布情况（总体精度为96.64%,Kappa系数为95.61%）,不同病害严重度（健康b0、轻度b1、中度b2、重度b3、极严重b4）对应棉田面积比例分别为7.81%、23.78%、29.20%、13.92%、17.43%;综合影像对病害棉田产量损失量估算效果最好,不同病害严重度（b0、b1、b2、b3、b4）对应棉田产量损失率分别为0、22.80%、31.32%、49.02%、76.33%,预估籽棉损失量达4 260.01 kg,损失率为49.16%,皮棉损失量达2 267.18 kg,损失率为54.51%。与病害胁迫棉田产量损失估算值相比,实际棉田籽棉损失率高6.28%,皮棉损失率高4.48%。病害胁迫棉田产量估算值与实际棉田收获值差异不显著,能够准确实现病害棉田产量损失估算。研究结果可为无人机遥感监测作物病害造成的产量损失提供理论依据和参考。     

基于机器学习和无人机多光谱遥感的苜蓿产量预测

作物产量快速准确预估对规模化农业生产具有重要意义。该研究在河北省涿州市开展田间试验，设置5个灌水处理及雨养对照区，使用无人机搭载多光谱相机在苜蓿的分枝期、现蕾期和初花期进行遥感监测，将光谱参数与苜蓿产量做相关性分析，分别挑选各生育时期相关性较好的5种植被指数，以去除土壤噪声后的5种植被指数和作物表面模型提取的苜蓿株高为输入变量，通过支持向量回归算法建立各生育时期的苜蓿产量预测模型，并结合实测产量进行模型评价。结果表明，苜蓿产量预测模型精度由高到低的生育时期依次为初花期、现蕾期、分枝期，使用植被指数和株高组合作为输入变量可以提高产量模型的预测精度。在初花期使用植被指数和株高组合输入变量所构建的估产模型最优，其决定系数、均方根误差和标准均方根误差分别为0.90、500 kg/hm2和14.3%，可用于苜蓿产量的快速准确评估。研究结果可为苜蓿的规模化生产和精细化管理提供技术支持。     

基于无人机多光谱影像的土壤盐分反演模型

为探究不同作物覆盖下不同深度的土壤盐分快速反演模型,该研究采集苜蓿、玉米覆盖下0～15、>15～30、>30～50 cm层深度的土壤盐分含量,基于无人机多光谱影像数据,提取各地块采样点的光谱反射率,在此基础上引入红边波段计算光谱指数作为特征变量,采用支持向量机递归特征消除算法（Support Vector Machine-Recursive Feature Elimination,SVM-RFE）以筛选光谱指数及未经过筛选的全指数组作为模型输入组,共构建出36个基于随机森林（Random Forest,RF）、极限学习机（Extreme Learning Machine,ELM）、BP神经网络（Back Propagation Neural Network）等机器学习模型,确定不同作物覆盖下的最佳土壤盐分反演模型。结果表明：SVM-RFE算法筛选光谱指数构建模型精度优于未进行筛选构建的模型。对于苜蓿和玉米覆盖土壤,整体上,RF反演效果优于ELM模型和BPNN模型,反演结果能体现真实土壤盐分含量,在0～15和>30～50 cm土层上,RF模型反演效果优于其他模型,苜蓿样地验证集决定系数Rp2分别为0.71、0.58,验证集均方根误差RMSEp分别为0.026、0.033,玉米样地Rp2分别为0.67、0.64,RMSEp分别为0.111、0.094,在>15～30 cm土层上ELM反演效果较好,苜蓿样地Rp2为0.58,RMSEp为0.039,玉米样地Rp2为0.68,RMSEp为0.059。0～15 cm是作物覆盖下的土壤含盐量最佳反演深度,验证集平均决定系数R2为0.65,均方根误差RMSE为0.084。研究结果可为土壤盐分的快速反演提供理论依据。     

基于无人机多光谱影像的水稻氮营养监测

探究消费级无人机多光谱影像对不同生态点、不同品种水稻氮营养监测建模精确度和普适度的影响,对于实现区域水稻氮营养精确管理与应用有重要意义。该研究分别在云南省西双版纳勐遮镇(供试品种：云粳37)与重庆市北碚区(供试品种：极优6 135)2个试验点设置不同氮水平田间试验,利用大疆精灵4多光谱无人机于水稻分蘖期、拔节期和抽穗期采集水稻冠层多光谱图像,采用凯氏定氮法测定水稻植株冠层氮含量(canopy nitrogen content,CNC)并计算地上部氮累积量(plant nitrogen accumulation,PNA);分别利用植被指数、偏最小二乘回归(partial least squares regression,PLSR)、随机森林(random forest,RF)、反向传播神经网络(back-propagation neural network,BPNN)对单一试验点、单品种和不同试验点、多品种水稻建立氮营养监测模型并探究模型的迁移能力。拔节期和抽穗期的模型精度较高(归一化植被指数NDVI或近红外归一化植被指数NNVI,R²为0.68～0.88),而分蘖...     

无人机多光谱遥感反演冬小麦SPAD值

为研究无人机多光谱遥感5个波段光谱反射率反演冬小麦SPAD（Soil and Plant Analyzer Development）值的可行性,该研究采用六旋翼无人机搭载五波段多光谱相机,采集冬小麦拔节期、孕穗期、抽穗期、开花期的冠层光谱影像并提取反射率特征参数,建立SPAD值的反演模型。结果表明,当波长范围在蓝光、绿光和红光波段,冬小麦拔节期、孕穗期和开花期的无人机多光谱影像反射率参数与SPAD值呈负相关关系,而在抽穗期,二者呈正相关;当波长范围为红边及近红外波段,在整个生长期,二者均呈现正相关关系。该研究构建冬小麦SPAD值反演模型采用了主成分回归、逐步回归和岭回归法,经对比发现基于逐步回归法构建的模型效果最优,该模型的校正决定系数为0.77,主成分回归法次之,岭回归法较差。此外,冬小麦抽穗期多光谱反射率反演SPAD值效果最显著,主成分回归、岭回归和逐步回归3种回归模型的校正决定系数分别为0.72、0.74和0.77。该研究可为无人机多光谱遥感监测作物长势、实现精准农业生产管理提供技术依据。     

利用无人机多光谱影像的多品种玉米成熟度监测

基于遥感监测多品种玉米成熟度进而掌握最佳收获时机，对提高其产量和品质至关重要。该研究在玉米成熟阶段获取无人机多光谱影像，同步采集叶片叶绿素含量（chlorophyll content，C）、籽粒含水率（moisture content，M）、乳线占比（proportion of milk line，P）等地面实测数据，以此构建玉米成熟度指数（maize maturity index，MMI），从而定量表征玉米成熟度。通过MMI与植被指数构建回归模型和随机森林模型，验证MMI适用性，并分析无人机遥感对不同品种玉米成熟度的监测精度。结果表明：1）不同品种玉米的叶片叶绿素含量、籽粒含水率、乳线占比的变化速率均存在差异。2）MMI与所选植被指数的相关性均可达到0.01显著水平，其中与归一化植被指数（normalized difference vegetation index，NDVI）、转换叶绿素吸收率（transformed chlorophyll absorbtion ratio index，TCARI）相关性最高，相关系数均为0.87。3）该研究基于不同组合的数据集进行了模型验证，其中随机森林模型对MMI的估测精度最高，测试集决定系数（coefficient of determination，R²）为0.84，均方根误差（root mean squared error，RMSE）为8.77%，标准均方根误差（normalized root mean squared error，nRMSE）为12.05%。此外，随机森林模型对不同品种MMI的估测精度较好，京九青贮16精度最优，其中R²、RMSE、nRMSE为0.76、10.67%、15.88%，模型精度证明了可以利用无人机平台对不同品种玉米成熟度进行监测。研究结果可为多光谱无人机实时监测农田多品种玉米成熟度的动态变化提供参考。     

基于无人机载高光谱空间尺度优化的大豆育种产量估算

为探讨无人机载高光谱空间尺度对大豆产量预测精度的影响,该文以山东嘉祥圣丰大豆为研究对象,设计以多旋翼无人机为平台搭载Cubert UHD185成像高光谱传感器的无人机遥感农情监测系统,获取了大豆多个生育期的无人机高光谱数据。首先,该研究利用盛荚期-始粒期(R4-R5期)的高光谱影像,由21个不同光谱空间尺度提取的高光谱数据构建植被指数,通过植被指数方差分析结果可知所选冠层植被指数与不同品种大豆植株的生长状况密切相关,但是不同空间尺度下的F值仍存在较为明显的差异;其次,采用偏最小二乘回归建立产量与不同空间尺度的植被指数之间的回归模型,通过模型方程估算精度的曲线变化趋势进一步将最优空间尺度面积确认至9.03~10.13 m2,即当采样空间尺度区域长、宽与小区总长、宽比例介于4.25:5和4.5:5时,所得到的冠层光谱能够尽可能准确地估测大豆产量,此时估算产量和实测产量呈极显著相关(相关系数r=0.811 7,参与建模的样本个数270)。该研究可为使用高、低空高光谱影像进行作物表型信息解析和估产提供参考。     

基于无人机多光谱影像的寒地水稻品质估测

【目的】在稻谷收获前准确估测稻米品质,为改善水稻养分管理、实现优质优价提供参考。【方法】试验在黑龙江省佳木斯市大兴农场和青龙山农场进行,两个地点土壤肥力具有较大差异。每个农场选择一块稻田,以20 m×20 m间隔采集土壤样品分析碱解氮含量,通过变量施氮形成产量水平、长势差异明显的水稻群体。利用大疆精灵4四旋翼多光谱无人机获取水稻分蘖期、拔节期、抽穗期和成熟期冠层多光谱数据。成熟期采集长势具有明显差异的水稻样品和对应的土壤样品,测定土壤有机质和速效氮、磷、钾含量,水稻籽粒蛋白质和直链淀粉含量,产量和食味值,随机选择其中67%的数据通过逐步多元线性回归构建不同生育时期的品质估算模型,其余33%的数据用于模型的验证。【结果】大兴农场水稻成熟期土壤有机质、碱解氮、有效磷和速效钾含量的变异系数分别为11.65%、14.44%、37.66%和11.60%,青龙山农场分别为14.45%、14.32%、36.37%和28.51%。成熟期两个地点水稻产量和食味值的变异系数均大于10%,青龙山农场水稻蛋白质含量变异系数也超过10%,而两地稻米直链淀粉含量的变异系数仅为1.11%～1.83%,因此,水稻直链...     

大豆叶绿素含量高光谱反演模型研究

叶绿素是植物体进行光合作用、进行第一性生产的重要物质,能够间接反映植被的健康状况与光合能力,同时也能反映植被受环境胁迫后的生理状态。高光谱遥感为快速、大面积监测植被的叶绿素变化提供了可能。该研究实测了不同水肥耦合作用下,大豆冠层的高光谱反射率与叶绿素含量数据,对二者进行了相关分析;采用特定叶绿素敏感波段建立了植被指数叶绿素估算模型;最后采用相关系数较大的波段作为神经网络模型的输入变量进行了叶绿素含量的估算。经对比发现叶绿素A、B与光谱反射率在可见光与近红外波段的相关系数的变化趋势基本一致,在可见光谱波段呈负相关,近红外波段呈正相关,红边处相关系数由负变正。特定色素植被指数可以提高大豆叶绿素估算精度(R²>0.736),但是人工神经网络模型可以大大提高大豆叶绿素含量的估算水平,当隐藏层节点数为4时,R²大于0.94,随着隐藏层节点数的增加,R²可高达0.99,表明神经网络模型可以大大提升高光谱反演大豆叶绿素含量的能力。     

大豆灰斑病对大豆产量与品质的影响

采用不同病级病叶、荚、粒,研究了大豆灰斑病对大豆产量和品质的影响。结果表明,大豆灰斑病使大豆光合面积减少、呼吸消耗增加、植株净光合生产率下降,导致百粒重和产量降低;同时感病籽粒蛋白质相对含量随病级增加而上升,绝对含量则下降;脂肪的相对含量变化不大,绝对含量随病级增加而降低,并提出大豆灰斑病的防治方法。     

播期对优质小麦籽粒灌浆特性及旗叶光合特性的影响

采用裂区设计,研究了播期对强筋小麦"临优145"和中筋小麦"临优2018"籽粒灌浆进程中粒重、蛋白质含量、籽粒产量、蛋白质产量、旗叶叶绿素相对含量及净光合速率的影响,并探明了它们间的相关性.结果表明:不同播期下优质小麦灌浆进程及蛋白质含量符合一元三次方程,分别呈"慢 -快 -慢"的"S"型和"高 -低 - 高"的"V"型变化;开花后23 d蛋白质含量最低;蛋白质产量随灌浆进程呈持续上升趋势.随播期推迟,优质小麦最大粒重、最大灌浆速率、平均灌浆速率及起始生长势提高,灌浆持续期和有效灌浆持续期延长,旗叶叶面积、叶绿素相对含量和净光合速率提高,产量呈先升高后降低趋势,并以10月9日最高.相关性分析表明,灌浆速率与千粒重极显著正相关,"临优145"有效灌浆持续期与千粒重显著正相关,"临优2018"有效灌浆持续期与千粒重间相关性不显著.     

单作和间作对玉米和大豆群体辐射利用率 及产量的影响

通过两个生长季(2006～2007年)的大田试验,研究了1:3和2:3两种问作模式及单作种植对玉米和大豆群体辐射截获与利用以及产量的影响.结果表明,单作玉米、单作大豆、1:3间作和2:3间作的消光系数分别为0.45、0.60、0.53和0.52.播后第79 d前,间作和单作玉米的干物质量间差异不显著;此后,1:3同作玉米的干物质量极显著高于单作玉米,1:3和2:3间作模式之间玉米干物质量的差异显著;观测期内,大豆干物质量单作和问作间差异显著.玉米/大豆1:3间作群体的辐射利用率(3.51 g·MS-1)和2:3间作群体的辐射利用率(3.49g·MJ-1)约为单作大豆辐射利用率(1.24g·MJ-1)的2.8倍.但略低于单作玉米(3.56g·MJ-).虽然单作玉米和大豆的籽粒产量均高于间作群体内玉米和大豆的籽粒产量,但间作群体的总产量分别比单作玉米和大豆的产量高约6.0%和320%;间作种植收入比单作玉米高56%～60%.比单作大豆高70%～74%.本研究表明,间作种植能够通过更有效地利用光能来增加作物产量,进而提高农民收入.     

用冠层叶色偏态分布模式RGB模型预测大豆产量

为了探索加色混色(Red-Green-Bule,RGB)模型偏态分布模式在大豆产量预测上的可行性,并验证其在不同肥料运筹、不同品种上的通用性,该研究选用曲茎和徐豆18两个大豆品种,设计了不同种植密度和氮肥水平的大田裂区试验,以无人机搭载数码摄像机,在花期及以后的2个重要生殖生长期采集大豆冠层数据.研究证实了大豆冠层数码...     

伤根对春小麦光合特性及水分利用效率的影响研究

盆栽春小麦伤根处理试验结果表明 ,苗期春小麦伤根适度处理可提高叶片光合速率及水分利用效率 ,促进光合产物积累 ,增加春小麦产量。拔节期春小麦伤根处理则抑制春小麦生长 ,降低其产量。     

施用控释尿素对大豆吸氮量及产量的影响研究

试验研究施用控释尿素对大豆吸N量及其产量与品质的影响结果表明 ,控释尿素比普通尿素N肥利用效率提高 3 0 6 %～ 1 1 93% ,沙土中尿素N残留量增加 1 2 5 %～ 1 5 89% ,大豆产量提高 2 1 6 9%～ 4 9 2 2 % ,且改善大豆品质。     

Water transmission characteristics of a Vertisol and water use efficiency of rainfed soybean (Glycine max (L.) Merr.) under subsoiling and manuring

In Vertisols of central India erratic rainfall and prevalence of drought during crop growth, low infiltration rates and the consequent ponding of water at the surface during the critical growth stages are suggested as possible reasons responsible for poor yields (<1 t ha⁻¹) of soybean (Glycine max (L.) Merr.). Ameliorative tillage practices particularly deep tillage (subsoiling with chisel plough) can improve the water storage of soil by facilitating infiltration, which may help in minimizing water stress in this type of soil. In a 3-year field experiment (2000–2002) carried out in a Vertisol during wet seasons at Bhopal, Madhya Pradesh, India, we determined infiltration rate, root length and mass densities, water use efficiency and productivity of rainfed soybean under three tillage treatments consisting of conventional tillage (two tillage by sweep cultivator for topsoil tillage) (S₁), conventional tillage + subsoiling in alternate years using chisel plough (S₂), and conventional tillage + subsoiling in every year (S₃) as main plot. The subplot consisted of three nutrient treatments, viz., 0% NPK (N₀), 100% NPK (N₁) and 100% NPK + farmyard manure (FYM) at 4 t ha⁻¹ (N₂). S₃ registered a significantly lower soil penetration resistance by 22%, 28% and 20%, respectively, at the 17.5, 24.5 and 31.5 cm depths over S₁ and the corresponding decrease over S₂ were 17%, 19% and 13%, respectively. Bulk density after 15 days of tillage operation was significantly low in subsurface (15–30 cm depth) in S₃ (1.39 mg m⁻³) followed by S₂ (1.41 mg m⁻³) and S₁ (1.58 mg m⁻³). Root length density (RLD) and root mass density (RMD) of soybean at 0–15 cm soil depth were greater following subsoiling in every year. S₃ recorded significantly greater RLD (1.04 cm cm⁻³) over S₂ (0.92 cm cm⁻³) and S₁ (0.65 cm cm⁻³) at 15–30 cm depth under this study. The basic infiltration rate was greater after subsoiling in every year (5.65 cm h⁻¹) in relation to conventional tillage (1.84 cm h⁻¹). Similar trend was also observed in water storage characteristics (0–90 cm depth) of the soil profile. The faster infiltration rate and water storage of the profile facilitated higher grain yield and enhanced water use efficiency for soybean under subsoiling than conventional tillage. S₃ registered significantly higher water use efficiency (17 kg ha⁻¹ cm⁻¹) over S₂ (16 kg ha⁻¹ cm⁻¹) and S₁ (14 kg ha⁻¹ cm⁻¹). On an average subsoiling recorded 20% higher grain yield of soybean over conventional tillage but the yield did not vary significantly due to S₃ and S₂. Combined application of 100% NPK and 4 t farmyard manure (FYM) ha⁻¹ in N₂ resulted in a larger RLD, RMD, grain yield and water use efficiency than N₁ or the control (N₀). N₂ registered significantly higher yield of soybean (1517 kg ha⁻¹) over purely inorganic (N₁) (1392 kg ha⁻¹) and control (N₀) (898 kg ha⁻¹). The study indicated that in Vertisols, enhanced productivity of soybean can be achieved by subsoiling in alternate years and integrated with the use of 100% NPK (30 kg N, 26 kg P and 25 kg K) and 4 t FYM ha⁻¹.     

Influence of lime and phosphorus on soybean (Glycine max (L.) Merrill) yield,leaf and seed composition on a paleudult soil

Abstract

This research was undertaken on a paleudult soil in southern Brazil, 30° south latitude, to quantify lime and P effect upon soybean (Glycine max (L.) Merrill). A lime x P factorial experience with lime treatments of 0, 0.5, 1, and 2 times SMP interpretation to pH 6.5, and 0, 44, 88, 132, and 176 kg P/ha with 3 replications were installed. The experiment was conducted for 2 years (1973–74, 1974–75), with leaf‐N, P, and K; yield; seed‐N, P, and K; Bray P₂ (0.03N NH₄F + 0.1N HC1) avail‐able‐P and soil pH measurements completed each year. Data was evaluated with linear, quadratic, logarithmic, polynomial, segmented line, and multiple regression using the coefficient of determination as goodness of fit.

The best model fit between P treatment and Bray P₂ available‐P was a quadratic equation; the model between relative yield and Bray P_2‐P with 54% of the relative yield attributed to Bray P₂ available‐P, a sigmented line. This model indicated point of maximum yield (91% relative yield) was obtained at 7.4 ppm‐P, with no increase in relative yield with increasing levels of soil available‐P. To calculate the P fertilizer necessary to increase available soil‐P to the level of maximum yield of equation Y_p = [1639(7.4 ‐ x_s)]^1/2, where Y_p = kg P/ha fertilizer needed; and x_s = initial Bray P₂ soil available‐P in ppm's. The lime effect upon soil pH was best described as a linear relationship. Yield increase with lime at this site was not significant at the 5% level.

The leaf‐N, P, and K increased significantly with soil available‐P levels. A second degree polynomial with logarithmic function best defined these relationships. The calculated DRIS indices and sum proved useful to evaluate the plant‐N, P, and K balance of each treatment.

Only seed‐P level was directly related to soil available‐P. Both seed‐N and seed‐K were highly correlated with indirect effects of soil available‐P levels.

Results from this study suggest the segmented line model would best interpret soybean yield response to Bray P₂ available‐P for this soil. To obtain maximum yield using this model rather than the second degree polynomial would require less fertilizer P. Foliar analyses interpretation confirmed adequate plant‐P level would be supplied for maximum yield at this level of fertilization.     

Long-term yield trend and sustainability of rainfed soybean–wheat system through farmyard manure application in a sandy loam soil of the Indian Himalayas

A long-term (30 years) soybean–wheat experiment was conducted at Hawalbagh, Almora, India to study the effects of organic and inorganic sources of nutrients on grain yield trends of rainfed soybean (Glycine max)–wheat (Triticum aestivum) system and nutrient status (soil C, N, P and K) in a sandy loam soil (Typic Haplaquept). The unfertilized plot supported 0.56 Mg ha⁻¹ of soybean yield and 0.71 Mg ha⁻¹ of wheat yield (average yield of 30 years). Soybean responded to inorganic NPK application and the yield increased significantly to 0.87 Mg ha⁻¹ with NPK. Maximum yields of soybean (2.84 Mg ha⁻¹) and residual wheat (1.88 Mg ha⁻¹) were obtained in the plots under NPK + farmyard manure (FYM) treatment, which were significantly higher than yields observed under other treatments. Soybean yields in the plots under the unfertilized and the inorganic fertilizer treatments decreased with time, whereas yields increased significantly in the plots under N + FYM and NPK + FYM treatments. At the end of 30 years, total soil organic C (SOC) and total N concentrations increased in all the treatments. Soils under NPK + FYM-treated plots contained higher SOC and total N by 89 and 58% in the 0–45 cm soil layer, respectively, over that of the initial status. Hence, the decline in yields might be due to decline in available P and K status of soil. Combined use of NPK and FYM increased SOC, oxidizable SOC, total N, total P, Olsen P, and ammonium acetate exchangeable K by 37.8, 42.0, 20.8, 30.2, 25.0, and 52.7%, respectively, at 0–45 cm soil layer compared to application of NPK through inorganic fertilizers. However, the soil profiles under all the treatments had a net loss of nonexchangeable K, ranging from 172 kg ha⁻¹ under treatment NK to a maximum of 960 kg ha⁻¹ under NPK + FYM after 30 years of cropping. Depletion of available P and K might have contributed to the soybean yield decline in treatments where manure was not applied. The study also showed that although the combined NPK and FYM application sustained long-term productivity of the soybean–wheat system, increased K input is required to maintain soil nonexchangeable K level.     

Enhancement of the rootability of a structurally fragile tropical soil using gypsum and leguminous residues to increase the yield of maize

In humid tropical areas, a combination of factors related to soil conditions for root development contribute to decreased nutrient use efficiency. Our aim was to evaluate the combined effects of the application of gypsum and mulch from leguminous trees on rootability of root zone soil, root growth, phosphorus recovery efficiency and maize yield in a sandy loam soil prone to hardsetting. The experiment involved a randomized block design with four replications and six treatments: no residue or gypsum (control, C); leguminous residues (L); leguminous residues and 6 t/ha of gypsum (LG6); urea and 6 t/ha of gypsum (UG6); leguminous residues, urea and 6 t/ha of gypsum (LUG6); and leguminous residues, urea and 12 t/ha of gypsum (LUG12). Application of gypsum plus leguminous residues enhanced soil rootability, when applied to a tropical Arenic Hapludult, through the combined effects of increased porosity and decreased penetration resistance. These improvements facilitated root growth in deeper soil layers and greater nutrient uptake. Application of leguminous residues increased the input and retention of calcium in deeper layers. The use of large amounts of gypsum with leguminous residue, despite a significant increase in phosphorus uptake, did not result in a proportional increase in maize yield. Given the small differences in calcium content between the 6 and 12 t/ha gypsum treatments, applying a low dose of gypsum combined with leguminous residue, rather than administering a single large gypsum dose, is recommended for tropical sandy loam soils with low cation retention capacities.