社会记忆视角下连州保安村民俗文化景观特色营造

乡村民俗文化景观是社会记忆传承的重要途径,承载了乡村生活、生产、文化实践等社会记忆活动。社会记忆视角乡村民俗文化景观营造研究对增强乡村文化的地方认同,维系良好的乡土社会秩序,实现乡村特色化发展具有重要意义。以粤北地区连州市保安村为例,在调研访谈的基础上,从山水记忆、建筑记忆、民俗记忆 3 方面分析乡村民俗文化景观特质。随着社会的变迁,保安村面临山水记忆意象片段化、建筑记忆空间特色趋同化、民俗记忆传承面临危机等问题。据此,提出以保安村社会记忆载体形成“营造、传承、促进”的内生机制,构建生态记忆、生活记忆和生长记忆的社会记忆体系,营造特色化乡村民俗文化景观的实施策略。     

Making Tropical Succession and Landscape Reforestation Successful

Forests are dynamic ecosystems, spatial mosaics of localized disturbances and regenerating vegetation. Following forest conversion to human land uses, expanding forest cover follows four different modes of reforestation: (a) spontaneous natural regeneration; (b) assisted natural regeneration; (c) agroforestry; and (d) commercial tree plantations. Natural regeneration fundamentally depends on the alignment of ecological and social factors at landscape scales. A deeper socioecological understanding of reforestation in the tropics will lead to the development of effective policies and governance structures that enhance forest cover and ecosystem services, protect biodiversity, and improve forest-dependent livelihoods.     

改进遗传算法求解作业车间提前/拖期调度问题

为解决作业车间提前/拖期调度问题,提出一个以最小惩罚为目标的改进遗传算法。该算法采用双染色体矩阵编码方式,利用部分映射交叉重排算子对父代个体进行交叉操作,提供了一种可以保留较高适应度个体的记忆功能,并利用爬山算法对记忆库进行更新,提高算法的局部搜索能力及收敛速度。仿真实验表明,与其他算法相比,该算法的搜索效率更高,收敛性能更好,求解的调度方案更优。     

Comparison of structural diversity of tree-crop associations in Peripheral and Buffer zones of Gachabari Sal forest area,Bangladesh

The structural diversity of different tree-crop associations were studied at Gachabari Sal forest area of Madhupur Garh on Buffer and Peripheral Zone during 2006. The total density, basal area of trees in the Buffer and Peripheral Zone were 155.5 trees·hm^-2, 795.4 trees·hm^-2 and 3.9 m^2·hm^-2, 5.8 m^2·hm^-2, respectively. No regeneration and natural trees were found in Peripheral Zone and the Zone is totally occupied by exotic species where the Buffer Zone comprised of both natural and exotic trees. The Peripheral Zone belonged to younger and smaller trees whereas the Buffer Zone belonged to mixture of smaller, taller, younger and mature trees simultaneously. For the practicing of different agroforestry systems both Zones have lost their original characters of Sal forest.     

光肩星天牛成虫寄主选择中的"记忆效应"

2004年和2005年,对羽化于5种树种的光肩星天牛成虫在5种相同寄主树种上的停留、取食、交配、刻槽、产卵等5个寄主选择指标进行了试验.从总体结果看,光肩星天牛成虫在进行寄主选择时,对原寄主具有一定“记忆效应”.具体来讲,受到嗜好树种复叶槭的影响,来源不同的光肩星天牛成虫在停留、取食和交配3个指标上没有表现出很强的“记忆效应”.但就刻槽和产卵两个指标而言,光肩星天牛成虫在其来源树种上的比率最高,“记忆效应”表现得十分明显.该结果为在多树种合理配置中发挥诱饵树作用提供了较好的理论依据.     

Forest resources variation along with the main rivers in typical forest region of Changbai Mountain

The restoration of the riparian vegetation disturbed by human activities is one of the hotspots of watershed ecology. Through interpreting the images of Remote Sensing in 1985 and 1999, the basic information of forest resources of Lushuhe Forest Bureau, which is a typical forest area of Chanbai Mountain was obtained with support of GIS. By dividing Land covers of Lushuihe area into 10 types (water body, residential land, stump land, farming land, wetland, mature conifer forest, midlife conifer forest, mature broadleaf forest, midlife broadleaf forest, and man-made young forest) and dividing the riparian zone into four buffers (in turn 1000, 2000, 3000, 4000 m away from the river), the changes of riparian forest resources during 1985–1999 were analyzed. The results showed that the deforestation intension has obviously decreased and the whole environment has been evidently improved, but the riparian ecosystem was still flimsy. In buffer 1, 2, 3 the area of midlife conifer forest increased largely, but the areas of other types of land covers all decreased. Midlife conifer forest had a comparatively good status in the three buffers. In buffer 4, midlife conifer forest, mature conifer forest, and mature broadleaf forest formed a forest-age rank that is helpful to stabilize the forest ecosystem and exert its functions. Area percentage of wetland decreased in buffer 1, buffer 2, and buffer 3, even in buffer 4 in which forest ecosystem rehabilitated comparatively well, so protecting and rehabilitating wetland is a very difficult task. Foundation item: This study is supported by major projects of Knowiedge Innovation Program. Chinese Academy of Sciences (No. KZCX2-SW-320-3) and Institute of Applied Ecology (a grant SCXZD010-01), CAS. Biography: HAO Zhan-qing (1962-), Male, Ph. Doctor, Professor in Institute of Applied Ecology, Chinese Academy of Science, Shenyang 1100016, P.R. China Responsible editor: Chai Ruihai     

邛海盆地土壤氮素空间变异特征与影响因素研究

以西昌市邛海盆地为研究区域,随机均匀布点采集386个耕作层土壤样点,分析化验土壤理化性质,运用地统计学方法和GIS技术研究邛海盆地表层土壤氮素含量的空间变异特征,利用方差分析、缓冲区分析和回归分析定量分析区域内土壤氮素空间变异的影响因素。研究结果表明,研究区表层土壤全氮和速效氮含量分别为(1.85±0.51)g/kg和(138.5±47.7)mg/kg,属较丰富水平,变异系数分别为27.52%和32.24%,属中等差异强度;土壤全氮和速效氮具有中等程度的空间相关性,空间变异均以随机变异为主,具有各向异性特征,在西北-东南方向上空间变异相对剧烈;研究区土壤全氮和速效氮含量总体上呈条带状或斑块状分布特征,高值区主要位于西昌市的东南郊区,并以此高值中心向南、向东和向北3个方向逐步递减,低值区主要出现在邛海北面近湖岸地带。影响因素分析结果表明,土地利用方式、土壤母质、土壤质地、土壤p H值、城镇、河流、邛海湖对土壤全氮和速效氮空间变异具有显著影响,而道路对土壤氮素空间变异影响不显著。     

基于WT-CNN-LSTM的溶解氧含量预测模型

溶解氧（Dissolved oxygen, DO）含量是影响水产养殖产量的重要因素之一,具有时序性、不稳定性和非线性等特点,且其影响因子过多、存在复杂的耦合关系,难以实现精准预测。针对传统长短时记忆神经网络（Long short term memory, LSTM）预测模型易引入冗余数据,且在训练过长序列时会出现梯度消失现象,从而不能捕捉因子间长期的依赖性问题,提出了基于小波变换（Wavelet transform, WT）、卷积神经网络（Convolutional neural network, CNN）和LSTM的溶解氧含量预测模型。首先,使用WT降低数据噪声;然后,使用CNN深度挖掘各变量之间的潜在关系;最后,利用LSTM的时序性预测2h后的水产养殖溶解氧含量。结果表明,本文提出的WT-CNN-LSTM模型预测效果良好,其平均绝对误差、均方根误差和决定系数分别为0.138、0.229和0.954,比传统LSTM模型分别优化了28.87%、21.03%和4.61%。     

基于部首嵌入和注意力机制的病虫害命名实体识别

为了解决农业病虫害命名实体识别过程中存在的内在语义信息缺失、局部上下文特征易被忽略和捕获长距离依赖能力不足等问题,以农业病虫害文本为研究对象,提出一种基于部首嵌入和注意力机制的农业病虫害命名实体识别模型(Chinese agricultural diseases and pests named entity recognition with joint radical embedding and self attention, RS-ADP)。首先,该模型将部首嵌入集成到字符嵌入中作为输入,用以丰富语义信息。其中,针对部首嵌入设计了3种特征提取策略,即卷积神经网络(Convolutional neural network, CNN)、双向长短时记忆网络(Bidirectional long short term memory network, BiLSTM) 和CNN-BiLSTM;其次,采用多层不同窗口尺寸的CNNs层提取不同尺度的局部上下文信息;然后,在BiLSTM提取全局序列特征的基础上,采用自注意力机制进一步增强模型提取更长距离依赖的能力;最后,采用条件随机场(Conditional random field, CRF)联合识别实体边界和划分实体类别。在包含11个类别和24715条标注样本的农业病虫害自制语料上进行了实验。结果表明,本文模型RS-ADP在该数据集上精确率、召回率和F1值分别为94.16%、94.47%和94.32%;在具体实体类别上,RS-ADP在作物、病害、虫害等易识别实体上F1值高达95.81%、97.76%和97.23%。同时,RS-ADP在草害、病原等难以识别实体上F1值仍保持86%以上。实验结果表明,本文所提模型能够有效识别农业病虫害命名实体,其识别精度优于其他模型,且具有一定的泛化性。     

职校生英语词汇的记忆策略

词汇是语言学习中最基本的组成部分,也是运用语言的先决条件,可以通过记忆方法和记忆规律的寻找,探索出一条适合职校生个体自身学习的有效途径,提升词汇记忆的效率.