金华市农业机械化投入机制初探

农业机械化是农业现代化的基础，是现代化的先决条件，发展农业机械化迫切需要加大投入。为此，分析了金华市现阶段农机化投入的现状及存在的问题；提出了通过建立政府、集体、信贷、农户、外资等“多位一体”的投入机制，加大农机化投入，促进农机化事业健康发展的思路。     

An evaluation of the Olsen test as a measure of plant-available phosphorus in grassland soils derived from basalt parent material

Abstract. Anecdotal and circumstantial evidence have suggested that the Olsen test underestimates plant-available phosphorus (P) in basaltic soils in Northern Ireland. Therefore, the ability of this test to predict plant-available P in basaltic (and non-basaltic) soils was investigated by regressing Olsen-P data against herbage P indices calculated from plant tissue test data using the diagnosis and recommendation integrated system. The average Olsen-P concentration for a range of fields situated on basaltic soils was considerably lower than the average Olsen-P concentration for a range of fields situated on non-basaltic soils, and yet mean sward P status, as given by the herbage P indices, was similar for both groups of fields. Herbage P indices were also much better correlated with Olsen-P measurements in non-basaltic soils than in basaltic soils. Furthermore, at low Olsen-P values (≶9mgPL⁻¹) some swards on basaltic soils were genuinely deficient in P, while others were sufficient or even in surplus for this nutrient. The results confirm that Olsen-P is inadequate as a predictor of plant-available P in basaltic soils. It is concluded that an alternative soil test is needed to provide a reliable assessment of plant-available P in basaltic soils, to prevent overuse of fertilizer and manure P and to minimize the amounts of P entering local watercourses.     

The Phytotoxicity of Various Graminicides in Durum Wheat in Greece

The effect of three herbicides, diclofop-methyl (DM), imazamethabenz-methyl (IM) and fenoxaprop-ethyl (FE), was investigated at the recommended rate (×) and double the recommended rate (2×) in a durum wheat crop. FE had the smallest effect on root system characteristics, while IM had the greatest effect. For all herbicides, the double rate treatment gave a greater statistical reduction than the single rate for root system characteristics. FE had the smallest effect on the activity of aerobic microorganisms, while DM gave the greatest reduction, 10 days after application, and all three herbicides reduced earthworm abundance. DM gave the smallest reduction in the number of tillers and in plant height, while the double rate of IM gave the greatest reduction. For all three herbicides, double the recommended rate resulted in a significant reduction in the weight of 1000 seeds, while no statistically significant deviations were noted at the recommended rate. Finally, the use of all three herbicides resulted in significantly lower statistical yields compared to the sample (control plot), although there were no significant statistical differences among them, for either recommended or double recommended rates.     

农机液压技术的应用与发展

液压技术的应用已成为现代化农业机械的典型标志。为此,介绍了农机液压技术的应用现状;并根据科学技术的发展,简要分析了农机液压技术的发展趋势。     

农机安全管理体制的变革和构想

结合《中华人民共和国道路交通安全法》有关规定和黑龙江省农机安全管理的实践,总结了农机安全管理体制焦点问题和体制的变革,并对实施道路交通安全法提出设想和建议。     

电算化会计信息系统下内部控制研究

本文对会计电算化发展中存在的一些问题提出了对策与建议，并针对财会工作的自身特点，从程序控制和制度控制两个方面阐述了电算化会计信息下的内部控制方法。     

供水工程影响下中国北方地区耕地后备资源开发潜力

灌溉水源是北方地区耕地后备资源能否有效开发的关键。当前,中国正在推动国家水网工程规划建设,为北方地区耕地后备资源开发提供了新的机遇。在以往研究基础上,该研究考虑北方地区已建、在建及规划的供水工程,基于自然适宜、利用高效、发展稳定三方面评价准则,结合三维魔方空间分类方法,开展了雨养情景和供水灌溉情景下耕地后备资源开发潜力分析。结果表明：雨养农业情景下,北方地区勉强适宜（Ⅱ级）、中度（Ⅲ级）和高度（Ⅳ级）适宜的耕地后备资源面积分别为2.9万、1.0万和0.6万km²。灌溉农业情景下,耕地后备资源面积明显增加,勉强适宜（Ⅱ级）、中度（Ⅲ级）和高度（Ⅳ级）适宜的耕地后备资源面积将分别达到5.4万、7.6万和6.5万km²,主要集中于新疆、内蒙古和甘肃3省（自治区）。该研究结果可为中国耕地后备资源开发利用及国家水网工程效益发挥提供参考。     

浅谈HACCP在机制针形名茶中的应用

对HACCP作简要介绍，根据HACCP体系的要求，分析了机制针形名茶过程中各个环节可能出现的危害因素，用HACCP的原理确立了关键控制点，并且建立了相应的控制措施。     

基于NGO-CNN-SVM的高标准农田灌溉工程施工成本预测

为提高高标准农田项目施工成本的预测精度,控制施工成本在合理范围,减少投资风险,该研究从单体灌溉工程施工成本预测角度出发,通过随机森林（random forest,RF）筛选出高标准农田灌溉工程施工成本的关键影响因素,结合卷积神经网络（convolutional neural networks,CNN）和支持向量机（support vector machine,SVM）两种模型的优点,通过北方苍鹰优化算法（northern goshawk optimization,NGO）对模型里的惩罚因子和核参数进行寻优,构建基于NGO-CNN-SVM的施工成本预测模型。通过辽宁省2018—2023年高标准农田工程中灌溉工程的施工成本数据,选取样本决定系数R²、平均绝对误差MAE、平均绝对百分比误差MAPE和均方根误差RMSE作为精度指标进行分析,结果表明：基于NGO-CNN-SVM的施工成本预测模型在渠道工程中MAE低于0.615万元,RMSE低于0.512万元,R²达到0.968以上,相对误差小于4.210%;在进水闸工程中MAE低于0.610万元,RMSE低于0.536万元,R²达到0.966以上,相对误差小于4.410%;在桥涵工程中MAE低于0.494万元,RMSE低于0.477万元,R²达到0.970以上,相对误差小于3.548%,并相比较于反向传播神经网络,CNN和CNN-SVM模型,NGO-CNN-SVM模型的预测结果均最优。通过特征选择、模型融合、算法优化以及不同模型对比表明NGO-CNN-SVM模型具有更高的预测准确率和泛化性,可为高标准农田灌溉工程施工成本预测提供理论依据。