Effects of long-term livestock grazing and habitat on understory vegetation

The herbaceous understory stratum contains most of the plant diversity in ponderosa pine ( Pinus ponderosa P. & C. Lawson var. scopulorum Engelm.) forests of the American Southwest and provides critical food and habitat for many wildlife species. During the last century, this stratum has been affected by livestock grazing and by increased dominance of overstory trees. We sampled a unique grazing exclosure to examine the relative importance of long-term livestock grazing (grazed or ungrazed) and habitat (park or tree) on the understory community. We sampled 3 plots of 192 contiguous quadrats (each quadrat 0.5 m 2 ) in each of the 4 treatment combinations, for a total of 2304 quadrats. Species-area curves were generated by aggregating quadrats into nonoverlapping areas at grain sizes of 0.5 to 576 m 2 . The effects of habitat and grazing on species density were evident at very different scales. Species density was higher in park than tree plots at scales ≤32 m 2 but did not differ between habitats at larger scales. Species density differed minimally between grazed and ungrazed treatments at small grains, but grazed plots contained more species than ungrazed plots at larger grains. Grazing treatments differed at smaller grains (to 4–8 m 2 ) than did habitats (to 32 m 2 ), with respect to density of native species and graminoids. Grazed plots had more exotic species than ungrazed plots at all grain sizes, though few exotics were present. Twenty-two species were identified as indicator species associated with habitats and/or grazing treatments. Evaluations of plant community response to treatments would be improved by accounting for the grain at which data have been collected and analyzed and by identifying indicator species associated with various treatments. These data would enable more-informed conservation and management decisions.     

Bi掺杂的直接甲醇燃料电池阳极催化剂研究进展

发展可替代能源对缓解全球能源问题具有重要意义。直接甲醇燃料电池(direct methanol fuel cell,DMFC) 因其工作温度低、能量密度高以及污染物排放少等特性,正逐渐成为最有发展前景的便携式能源技术之一。目前,其商业化进程主要取决于甲醇氧化反应(methanol oxidation reaction,MOR) 的动力学快慢、催化剂的成本和寿命。Bi元素的掺杂可以极大地提高甲醇电催化氧化的性能,并且可以提高阳极催化剂抵抗CO中毒的能力。介绍了掺杂Bi的贵金属和非贵金属阳极电催化剂,以及贵金属掺杂Bi₂O₃、Bi₂WO₆ 等光辅助电催化剂;综述了它们提高甲醇电催化氧化性能的机制,并展望了阳极Bi电催化剂在DMFC中所面临的机遇和挑战。     

WO3丙酮传感器及其多传感器融合技术的研究

高浓度丙酮会刺激黏膜和引起中毒,在医疗健康领域丙酮也可作为呼出气体标记物。为实现对丙酮气体的有效检测,利用WO₃纳米材料制备旁热式气敏传感器,系统测试该传感器的基本特性,并在此基础上与SnO₂传感器组成双气敏传感器阵列,结合反向传播神经网络方法对丙酮混合气体进行识别。在体积浓度10^-5丙酮气体测试中,WO₃丙酮传感器的响应值达到9.3,响应与恢复时间均为4 s,重复性误差小于10^-8;双气敏传感器阵列对混合气体中丙酮和乙醇的识别误差小于10^-8。结果表明,制备的WO₃传感器响应灵敏度、稳定性、选择性、瞬态响应等特性良好,采用多传感器融合技术可以实现混合气体的有效识别。     

基于离散正则模型的时变信道盲辨识算法

以CDMA时变信道离散正则模型为基础,提出了基于改进的矩阵外积分解的时变信道盲辨识算法。算法对传统的矩阵外积分解算法进行修正,使其适应离散正则模型两级盲辨识,并引入了精确的时延阶数估计,克服了现有算法需要预先知道信道时延实际阶数的局限,使其在仅知道信道阶数上界的条件下完成盲辨识。仿真结果表明了该算法的有效性。     

虚拟数控铣削表面粗糙度预测技术研究

分析了影响数控铣削加工表面粗糙度的主要因素,利用多元回归分析建立了基于球头铣刀的表面粗糙度关于其影响因素的数学模型.经显著性检验,该模型预测精度高,泛化能力强.该模型为虚拟数控铣削表面粗糙度预测的建模提供了理论依据.开发了基于DELPHI的可视化交互式的仿真系统,实现了虚拟数控铣床的表面粗糙度的在线预测.有助于理解表面粗糙度随铣削参数改变的规律,也为铣削参数优化提供理论依据.     

一种基于小波的互连线模型自适应约减算法

基于偏微分方程的谱分解理论,针对VLSI中的互连线系统,提出了一种基于小波的自适应模型约减算法,选取逼近空间的一组基函数,通过积分合同变换,从而得到一个能够逼近原无穷维系统的有限低阶无源模型.该算法在不需要对互连线进行离散处理的情况下,能直接对非均匀互连线系统进行约减;利用了小波的多尺度和紧支撑特性,并采用自适应机制来平衡逼近精度和计算效率的矛盾.理论分析和数值仿真说明了该方法的有效性.     

平面波对传输线耦合响应的仿真研究

在Agrawal传输线模型的基础上,以典型信号-电磁脉冲作为干扰源,详细分析了参数变化对传输线终端负载电压和电流响应的影响.仿真结果表明,负载、传输线间距、平面波入射角和平面波极化角等参数变化对负载端响应有不同的影响.在此基础上,又对双线传输线和单导体传输线受干扰特性进行了对比分析.仿真分析结果为实际系统中由于电缆耦合产生的干扰分析提供了理论依据,并具有一定的参考价值.     

一种不定长特征模式入侵检测模型及仿真分析

针对定长序列模式在刻画序列特征方面的不足,提出了一种不定长序列特征模式抽取算法,并以此为基础设计了一个新的入侵检测模型。模型引入了状态转移概念并将模式匹配检测和状态转移检测结合在一起,克服了已有模型只专注于单一特征检测的缺陷。实验证实,和单特征检测相比,该模型在保证检测率的同时降低了误报率,并且检测算法的效率也是可接受的。