作物水模型参数与时段数关系的研究

时段划分是影响作物水模型参数求解和应用的一个重要因素。依据山西水利职业技术学院试验基地2007和2008年度冬小麦田间试验资料,将冬小麦全生育期等间隔地划分为不同时段,用非线性优化方法求得了不同时段数条件下的模型参数,分析研究了模型参数与时段数的关系,据此在水分敏感指数累积函数中引入了时段数,并与模型1(未引入时段数)进行了比较分析。结果表明,模型2(引入时段数)模拟产量的相对误差随时段数的增加而减小,当时段数大于13时,相对误差平均值和最大值即分别减小到11%和20%以下;与模型1比较,参数个数未增加,模拟精度未降低;可用于任意时段数条件下的产量模拟计算,更精确地反映了水分胁迫时间对作物产量影响的信息。     

时段划分对冬小麦作物水模型影响的研究

引入逻辑斯蒂函数描述水分敏感指数随时间的变化过程,对Jensen模型(模型一)进行了改进,使得改进后的作物水模型(模型二)的参数固定为4个,避免了模型一参数随时段数增加而增加的缺陷。采用山西水利职业技术学院试验基地2007年和2008年冬小麦田间试验资料,将冬小麦全生育期等间隔地划分为23、21、19、17、……、3共11个时段,利用非线性规划的方法求得了相应的模型参数,进行了比较分析。结果表明,采用模型一时,相对腾发量划分的时段数以5左右为宜,不宜超过7;采用模型二时,则不受时段数的限制;模型二的修正复相关系数Ra随时段数的增加略有增大的趋势,均在0.84以上,F值均在16以上,大于F0.001=12.56,达到极显著水平,能够用于模拟供水对产量的影响;模型二的标准误随时段数的增加上下波动,变化于0.100~0.108之间,小于模型一的标准误,模拟精度高于模型一;采用模型二模拟产量时宜尽量使用较大时段数的参数,且腾发量划分的时段数与参数的时段数应尽可能一致。     

基于双参数和粒子滤波同化算法的夏玉米单产估测

为了提高河北省中部平原夏玉米的估产精度和进一步验证粒子滤波同化算法对农业作物估产的适用性,采用粒子滤波算法同化CERES-Maize模型模拟和MODIS数据反演的叶面积指数(Leaf area index,LAI)、条件植被温度指数(Vegetation temperature condition index,VTCI),应用随机森林回归算法确定夏玉米不同生育时期LAI和VTCI的权重,构建单产估测模型。结果表明,无论是单点尺度还是区域尺度,同化的LAI和VTCI均能较好地响应外部观测数据,同化LAI可减缓CERES-Maize模型模拟LAI的剧烈变化;同化VTCI结合模型模拟和遥感观测,更能反映夏玉米对水分胁迫的敏感性。利用2015年河北省中部平原各县(区)夏玉米产量对较优估产模型进行精度验证,结果表明,同化前后夏玉米产量模拟结果与统计产量间的归一化均方根误差由12.71%下降到10.50%,平均相对误差由12.57%下降到8.43%,说明基于同化LAI和VTCI构建的双参数单产估产模型可用于区域夏玉米单产估测。     

不同灌溉条件下保水剂对新疆棉花生长及产量的影响

为指导新疆灌溉棉区棉花节水灌溉和增产,通过大田试验,比较研究5种不同灌溉条件下保水剂对棉花生长、干物质积累与分配、水分吸收及产量的影响。结果表明,施用保水剂能不同程度地抑制棉花株高,使主茎叶数、单株蕾数和叶面积减小,但单株铃数会增加。保水剂能促进棉花根系和蕾铃发育,使干物质由营养器官向生殖器官的分配比增加13.8%～25.8%,同时增强了水分由根系向茎叶的运输能力。适宜减量灌溉（常规灌量的40%以上）条件下,施用保水剂在节水21.1%以上的前提下,仍能较对照（充足灌溉,且无保水剂）增产6.7%～22.0%;即使是以常规灌量的40%进行灌溉,施保水剂也能保障棉花不减产。在该试验条件下,常规灌量的80%为最适灌溉量,能使棉花产量较对照显著提高22.0%（有保水剂）和7.4%（无保水剂）。初步分析认为,引起棉花增产的直接原因是单株铃数和单铃质量增加,间接原因是保水剂调控了不同器官干物质的积累与分配,以及水分代谢状况。     

不确定性条件下分布式农业生产预警模型构建与应用

构建了基于梯形模糊数和分布式作物模拟模型的空间分布式农业生产预警模型来实现作物产量和水分生产力综合警情预警预报。模型采用空间分布式作物产量和水分生产力作为警情指标来计算系统警级,引入梯形模糊数来表征目标产量和水分生产力的不确定性,采用空间分布式作物模拟模型来模拟常规灌溉、0.8倍常规灌溉和0.6倍常规灌溉下的作物产量和水分生产力,进而对现状1976—2012年和未来RCP4.5情景下2026—2045年不同灌溉水平下进行农业生产风险预报预警,并衡量了未来20年产量和水分生产力的静态协调度和每5年4个周期的动态协调度。结果表明,同一作物在不同土壤类型和不同灌溉水平下预警等级不同,警级随着灌溉水平的降低呈现不规则变化规律,协调性随着灌溉水平的降低而减小。模型能够识别出未来气候变化不同节水灌溉水平下的空间异质性作物产量和水分生产力的警级,实现精准化农业生产风险预警预报,有利于实现高效率降警处理。     

基于AquaCrop模型的南疆无膜滴灌棉花灌溉制度优化

为探究AquaCrop模型对南疆地区无膜滴灌棉花种植的适用性、寻求最优灌溉制度,以2018—2019年田间实测数据对模型进行校验,并利用校准模型分别模拟2种不同灌水情景下的棉花冠层覆盖度、生物量及产量的变化规律。结果表明：采用AquaCrop模型模拟2019年冠层覆盖度的均方根误差（RMSE）、拟合指数（d）、标准均方根误差（NRMSE）及决定系数（R2）分别为6.03%、0.12、13.08%和0.97,生物量模拟的各参数分别为810kg/hm2、0.93、6.41%和0.80,产量模拟的各参数分别为751kg/hm2、0.84、14.02%和0.87,说明AquaCrop模型可以较好地模拟南疆地区无膜滴灌棉花的生长与产量。基于1960—2019年的气象数据,利用AquaCrop模型对无膜滴灌棉花进行情景模拟：灌水周期相同时,无膜滴灌棉花产量随灌溉定额的增加呈先增加、后减小的变化趋势,总灌水量为7200m3/hm2时产量达到最大值（5398kg/hm2）,水分利用效率（WUE）为0.75kg/m3;灌溉定额相同时,无膜滴灌棉花产量随灌水周期的增加而增加,总灌水量为5400m3/hm2、灌水周期为5d时,产量达到最大值（5315kg/hm2）、WUE为0.98kg/m3。研究表明,灌水周期5d、灌溉定额540mm的灌溉制度可保证无膜滴灌棉花具有较高的产量和水分利用率,可作为无膜滴灌棉花的参考灌溉制度在南疆地区推广应用。     

秸秆覆盖对冬小麦产量及水分生产效率的影响研究

为实现豫西半旱地区冬小麦节水高产栽培,提高水分利用效率,利用农田秸秆覆盖试验技术,研究了3种灌水模式下不同秸秆覆盖量对冬小麦生育期耗水量、作物产量以及水分利用效率（WUE）的影响。研究结果表明：秸秆覆盖对千粒重、穗粒数和产量的增长具有明显的促进作用,对公顷穗数的增长具有抑制作用。灌3水处理与灌4水处理相比较,产量与水分利用效率随覆盖量增加明显提高,产量和水分利用效率提高1.0%和1.6%,耗水量减少4%,实现了节水与高产的统一。灌2水处理下不同覆盖量的耗水量和水分利用效率都有所提高,但对产量影响不显著。试验最佳分配方案为灌3水处理与4 500 kg/hm2覆盖量。     

基于SIMDual_Kc模型估算非充分灌水条件下冬小麦蒸散量

为研究关中冬小麦植株蒸腾和土壤蒸发规律,利用2 a冬小麦小区控水试验实测数据,率定和验证了双作物系数SIMDual_Kc模型在关中地区的适用性.用大型称重式蒸渗仪的实测蒸散量值(或水量平衡法计算值)与模型模拟值进行对比.结果表明:SIMDualKc模型可较准确地模拟关中不同水分条件下冬小麦蒸散量,且模拟精度较高.模型估算的平均绝对误差为0.643 3 mm/d.模型估算的冬小麦初期、中期和后期的基础作物系数分别为0.35,1.30,0.20.另外,模型还可以较准确地估算不同水分供应条件下的土壤水分胁迫系数、土壤蒸发量和植株蒸散量.冬小麦整个生育期,土壤蒸发主要发生在作物生育前期,中期较低,后期略微增大;植株蒸腾主要发生在作物快速生长期和生长中期,整个生育期中呈先增大后减小的趋势.     

基于GLUE和PEST的CERES-Maize模型调参与验证研究

作物模型已逐渐成为干旱和半干旱地区优化农田水肥管理和实施节水灌溉的有力决策支持工具。为了探讨CERES-Maize模型模拟不同生育期受旱情况下夏玉米的生长发育、产量形成和土壤水分状况的模拟精度,进行了2013和2014年连续两季夏玉米田间分段受旱试验。试验将夏玉米整个生育期划分为苗期、拔节、抽雄和灌浆4个主要生长阶段,采用单个生育期受旱其他生育期灌水的方式,形成4个不同的受旱时段水平(D1~D4),又根据夏玉米多年生育期降雨量,设置了70和110 mm两个灌水水平(I1和I2),共形成8个处理,每个处理3次重复,在遮雨棚内按照裂区试验布设,此外设置1个各生育期均灌水110 mm的对照处理(CK)。利用两年试验数据,采用DSSAT-GLUE和PEST两种不同的模型参数估计工具,对CERES-Maize模型的遗传参数进行估计,并对该模型的模拟精度和可靠性进行验证,此外还使用交叉验证法对CERES-Maize模型的整体模拟精度进行评估。结果表明,GLUE和PEST两种调参工具所得的模型参数均有较好的稳定性和收敛性,但PEST调参工具耗时较少,效率较高;CERES-Maize模型能较好地模拟充分灌水条件下夏玉米的生长发育、产量和土壤水分变化,绝对相对误差(ARE)和相对均方根误差(RRMSE)均在6%~8%之间;但是现有CERES-Maize模型无法模拟由于不同生育期受旱造成的夏玉米物候期的差异。此外,交叉验证结果发现夏玉米生长前期(特别是拔节期)受旱处理的数据参与模型校正时,模型的总体平均模拟误差较大,精度较低。CERES-Maize模型模拟前期受旱对玉米籽粒产量的影响时结果不够准确,这可能是由于该模型低估了早期水分胁迫条件下的LAI值,进而使得ET模拟不准确所造成的。总之,CERES-Maize模型对生育期前期(特别是拔节期)受旱条件下夏玉米生长发育、产量形成和土壤水分变化的模拟还存在一定的不足,若将CERES-Maize模型应用于我国干旱和半干旱地区水分胁迫条件下玉米的生产管理和科学研究,应对模型进行相应的修正。     

基于概念模型的棉田土壤水分变化模拟研究

农田土壤水分模拟是农业用水管理的重要依据。以根系层土体水量平衡方程为依据,考虑根系层下界面水分通量,构建了农田土壤水分变化模拟模型,该模型由作物蒸发蒸腾量模型、根区下界面水分通量模型以及水量平衡方程等组成。依据山西水利职业技术学院试验基地2006-2008年3年棉花试验资料,确定了模型参数。结果表明,土壤储水量模拟计算值与实测值有较好的一致性,其相关系数达到0.928 7,F检验结果(F=96.44F0.001=3.27)达到高度显著水平,所建立的土壤水分变化模拟模型可用于棉花田间土壤水分的模拟计算,计算精度平均达到5.3%~15.8%;模型较好地反映了农田土壤水分转化过程以及降水、蒸发和深层土壤水分对作物蒸发蒸腾及产量的影响。     

Simulation of the soil-water dynamics and corn yields under deficit irrigation

Summary A simulation model capable of predicting the yield response of corn to a limited water supply was developed by combining two existing mathematical models. The resulting computer model was evaluated using experimental data taken under a wide range of soil moisture conditions. The soil profile water balances was simulated using SWATRE and SUCROS was used to model the crop growth in response to environmental conditions. In addition to the integration of the two existing models, some minor changes were made to each in an effort to improve the accuracy of the combined models. The model input parameters were derived entirely from published literature. The experimental data necessary for model validation were available from irrigation studies at the Sandhills Agricultural Laboratory of the University of Nebraska. These experiments not only provided the required input soil and climatic data, but also the observed irrigation levels, soil moisture distributions and crop yield required for model validation. Initial evaluation of the computer model indicates that the combined model adequately describes crop evapotranspiration, soil moisture extraction and crop yield under a fairly wide range of soil moisture stress. Additional modifications for the prediction of leaf area expansion and senescence, especially under moisture stress, are needed to improve the accuracy of the model.     

Predicting the effects of drainage systems on corn yields

Stress day index (SDI) models were incorporated in the water management simulation model, DRAINMOD, to quantify the effect of soil water stresses on corn yields. The effects of a combination of excessive and deficient soil water conditions were approximated by a simple first-order crop response model, YR = YRw × YRd, where YR is the overall relative yield, and YRw and YRd are the relative yields due to excessive and deficient soil water conditions, respectively.The accuracy of the modified water management model was evaluated by comparing predicted and measured corn yields for 16 plot years of experimental data on the Tidewater Research Station near Plymouth, NC. The predicted and measured results were in good agreement with the model describing 63% of the variation in yields for the 12-year period.Use of the modified water management model was demonstrated by simulating the performance of several drainage system designs for a Portsmouth sandy loam soil. The results of the simulation show that a maximum long-term relative yield of 80% of the potential corn yield can be obtained with a drain spacing of 40 m or less with good surface drainage. Higher yields could not be obtained without irrigation to reduce deficit soil water conditions. The response of long-term average corn yields to surface drainage varies inversely with the intensity of subsurface drainage. The 25-year average yield for 100 m spacing was only 47% of the potential yield when the surface drainage was poor as compared to 61% of potential yield for good surface drainage.     

A model for assessing crop response to salinity

CropSyst, a management-oriented crop growth model, was modified to assess crop response to salinity. The effect of salinity was included in the existing water uptake module by adding an osmotic component to the soil water potential and developing a function to account for salinity effects on root permeability. The effect of salinity on water uptake is the link to simulate crop growth reduction. A qualitative analysis showed that the model simulated expected trends of crop response to salinity as affected by cultivar tolerance, atmospheric vapor pressure deficit, and soil water availability. Comparisons with data from sprinkler line experiments were performed for barley grown at Zaragoza (Spain) in 1986 and 1989, and corn at Davis, Calif. and Fort Collins, Colo. in 1975. These experiments included different salinity and irrigation levels. At Davis, the model simulated well the effect of salinity/irrigation treatments on water use, biomass, and crop yield, with values for the Willmot index of agreement (d) generally better than 0.94 (a value of 1.0 implying perfect agreement). At Fort Collins, simulation of grain yield was less satisfactory (d fluctuated between 0.83 and 0.90), but the agreement was good for crop water use and biomass (d generally better than 0.96). The lower performance for grain yield was attributed to large and erratic variations in the observed harvest index. The agreement between simulated and observed values tended to be lower at Zaragoza, with d values fluctuating between 0.84 and 0.91 for biomass and yield in the 2 years included in this evaluation. Unusually high measured yields in 1989 and erratic variation in 1986 were attributed to small sample size. The small size (increased measurement error) of samples typically obtained in sprinkler line source experiments tends to limit their use for evaluation of simulation models.     

基于多时相多参数融合的麦玉轮作小麦产量估算方法

为了进一步提高冬小麦产量预测的准确性,针对麦玉轮作体系缺乏直接把前茬作物信息纳入到当季作物的产量估算及管理中的研究状况,利用前茬玉米季中长势遥感信息及产量信息,融合小麦拔节期、灌浆期及成熟期长势遥感信息、播前施肥信息及土壤特性信息等多时相多模态数据,基于GPR算法,建立多时相多模态参数融合的麦玉轮作体系小麦产量估算模型,结果显示：基于多生育期的产量估算模型较单生育期最优产量估算模型性能有所提升,R²提高0.01～0.03。其中基于拔节期产量估算模型精度略低于多生育期产量估算模型,但精度相近。基于多模态参数融合的产量估算模型中,除玉米作物信息与土壤特性信息融合构建的产量估算模型,多模态参数融合的产量估算模型精度较相应低模态参数融合的产量估算模型精度高。四模态参数融合的GPR模型决定系数R²为0.92,RMSE为213.75 kg/hm²,较其他模型,R²提高0.02～0.41。对于小麦产量估算模型,各模态参数影响由大到小依次为施肥信息、小麦遥感信息、土壤特性信息、玉米作物信息。玉米作物信息对于多模态参...     

HydroLOGIC: An irrigation management system for Australian cotton

The worldwide need to improve water use efficiency within irrigated agriculture has been recognised in response to environmental concerns and conflicts in resource use. Within the Australian cotton industry, the imperative to reduce water use and optimise irrigation management through the understanding of risk, using information generated by computerised decision aids was identified and subsequently developed into the HydroLOGIC irrigation management software. This paper summarises the attributes of the HydroLOGIC irrigation management software, with particular emphasis on functionality and its application to irrigation decisions within the Australian cotton industry. The software development process is documented to provide direction for future software application initiatives, with particular emphasis on a process of user feedback, evaluation and support requirements providing direction to software development. On-farm experiments throughout the development period allowed the validation of internal software logic, irrigator decision processes, and the OZCOT cotton growth model. The software demonstrated the ability to improve yield and water use efficiency by optimising strategic and tactical irrigation decisions in the Australian furrow irrigation cotton production system. In 7 of the 11 on-farm experiments conducted, the use of HydroLOGIC helped improve overall field water use efficiency by optimising the timing of irrigation events or by indicating further irrigations would not provide yield or maturity benefits. The paper also presents useful insights into the development of software targeted for irrigation utilising in-field measurements of soil water, crop growth and a crop growth simulation model.     

Simulated effect of drainage water management operational strategy on hydrology and crop yield for Drummer soil in the Midwestern United States

The hypothetical effects of drainage water management operational strategy on hydrology and crop yield at the Purdue University Water Quality Field Station (WQFS) were simulated using DRAINMOD, a field-scale hydrologic model. The WQFS has forty-eight cropping system treatment plots with 10 m drain spacing. Drain flow observations from a subset of the treatment plots with continuous corn (Zea mays L.) were used to calibrate the model, which was then used to develop an operational strategy for drainage water management. The chosen dates of raising and lowering the outlet during the crop period were 10 and 85 days after planting, respectively, with a control height of 50 cm above the drain (40 cm from the surface). The potential effects of this operational strategy on hydrology and corn yield were simulated over a period of 15 years from 1991 to 2005. On average, the predicted annual drain flows were reduced by 60% (statistically significant at 95% level). This is the most significant benefit of drainage water management since it may reduce the nitrate load to the receiving streams. About 68% of the reduced drain flow contributed to an increase in seepage. Drainage water management increased the average surface runoff by about 85% and slightly decreased the relative yield of corn crop by 0.5% (both are not statistically significant at 95% level). On average, the relative yield due to wet stress (RYw) decreased by 1.3% while relative yield due to dry stress (RYd) increased by 1%. Overall, the relative crop yield increased in 5 years (within a range of 0.8-6.9%), decreased in 8 years (within a range of 0.2-5.5%), and was not affected in the remaining 2 years. With simulated drainage water management, the water table rose above the conventional drainage level during both the winter and the crop periods in all years (except 2002 crop season). The annual maximum winter period rise ranged between 47 cm (1995) and 87 cm (1992), and the annual maximum crop period rise ranged between no effect (2002) and 47 cm (1993).     

Managing irrigation and drainage systems in arid areas in the presence of shallow groundwater: case studies

Two field studies were conducted on the west side of the San Joaquin Valley of California to demonstrate the potential for integrated management of irrigation and drainage systems. The first study used a modified cotton crop coefficient to calculate the irrigation schedule controlling the operation of a subsurface drip system irrigating cotton in an area with saline groundwater at a depth of 1.5 m. Use of the coefficient resulted in 40% of the crop water requirement coming from the groundwater without a loss in lint yield. The second study evaluated the impact of the installation of controls on a subsurface drainage system installed on a 65 hectare field. As a result of the drainage controls, 140 mm less water was applied to the tomato crop without a yield loss. A smaller relative weight of tomatoes classified as limited use, was found in the areas with the water table closest to the soil surface.     

基于SSA-RFR算法的采棉机测产传感器研究

随着棉花种植和收获的机械化程度提高,获取准确的产量图,分析田间产量数据,变得尤为重要,而采棉机作业时在输棉管道处监测产量是一种有效、可行的方法。现有光电对射式棉花测产传感器在作业中会有粘液遮挡检测通道、环境光影响等问题,面对复杂的田间作业环境,传感器标定普遍采用线性或多项式模型,精度和抗干扰性表现不够理想。针对上述现状,本文首先在传感器的结构和电路设计上做了抗干扰改进。然后在传感器标定过程中,尝试使用随机森林回归模型(Random forest regression, RFR),对实验样本进行训练、测试。在分析模型的表现后,提出了麻雀算法(Sparrow search algorithm, SSA)改进的随机森林回归模型,以均方误差作为适应度,对模型进行优化。经过验证,在相同验证集下,优化后的模型有更好的检测精确度。通过研究寻优上下界范围,平衡运行时间和检测精度,得到最优检测模型。该模型在验证集上表现良好,决定系数R²为0.99,平均绝对百分比误差(MAPE)为6.34%。台架实验结果表明,不同风速下最大误差为9.21%,平均误差为8.33%,改进后的传感器及检测...     

Using EPIC model to manage irrigated cotton and maize

Simulation models are becoming of interest as a decision support system for management and assessment of crop water use and of crop production. The Environmental Policy Integrated Climate (EPIC) model was used to evaluate its application as a decision support tool for irrigation management of cotton and maize under South Texas conditions. Simulation of the model was performed to determine crop yield, crop water use, and the relationships between the yield and crop water use parameters such as crop evapotranspiration (ETc) and water use efficiency (WUE). We measured actual ETc using a weighing lysimeter and crop yields by field sampling, and then calibrated the model. The measured variables were compared with simulated variables using EPIC. Simulated ETc agreed with the lysimeter, in general, but some simulated ETc were biased compared with measured ETc. EPIC also simulated the variability in crop yields at different irrigation regimes. Furthermore, EPIC was used to simulate yield responses at various irrigation regimes with farm fields’ data. Maize required ∼700 mm of water input and ∼650 mm of ETc to achieve a maximum yield of 8.5 Mg ha⁻¹ while cotton required between 700 and 900 mm of water input and between 650 and 750 mm of ETc to achieve a maximum yield of 2.0-2.5 Mg ha⁻¹. The simulation results demonstrate that the EPIC model can be used as a decision support tool for the crops under full and deficit irrigation conditions in South Texas. EPIC appears to be effective in making long-term and pre-season decisions for irrigation management of crops, while reference ET and phenologically based crop coefficients can be used for in-season irrigation management.