Soil respiration in tropical seasonal rain forest in Xishuangbanna, SW China

With the static opaque chamber and gas chromatography technique, from January 2003 to January 2004 soil respiration was investigated in a tropical seasonal rain forest in Xishuangbanna, SW China. In this study three treatments were applied, each with three replicates: A (bare soil), B (soil+litter), and C (soil+litter+seedling). The results showed that soil respiration varied seasonally, low from December 2003 to February 2004, and high from June to July 2004. The annual average values of CO2 efflux from soil respiration differed among the treatments at 1% level, with the rank of C (14642 mgCO2· m-2. h-1)＞B (12807 mgCO2· m-2. h-1)＞A (9532 mgCO2· m-2. h-1). Diurnal variation in soil respiration was not apparent due to little diurnal temperate change in Xishuangbanna. There was a parabola relationship between soil respiration and soil moisture at 1% level. Soil respiration rates were higher when soil moisture ranged from 35% to 45%. There was an exponential relationship between soil respiration and soil temperature (at a depth of 5cm in mineral soil) at 1% level. The calculated Q1o values in this study,ranging from 2.03 to 2.36, were very near to those of tropical soil reported. The CO2 efflux in 2003was 5.34 kgCO2· m-2. a-1 from soil plus litter plus seedling, of them 3.48 kgCO2· m-2. a-1 from soil (accounting for 62.5%), 1.19 kgCO2· m-2. a-1 from litter (22.3%) and 0.67 kgCO2·m-2. a-1 from seedling (12.5%).     

Study of statistically correcting model CMAQ-MOS for forecasting regional air quality

Based on analysis of the air pollution observational data at 8 observation sites in Beijing including outer suburbs during the period from September 2004 to March 2005, this paper reveals synchronal and in-phase characteristics in the spatial and temporal variation of air pollutants on a city-proper scale at deferent sites; describes seasonal differences of the pollutant emission influence between the heating and non-heating periods, also significantly local differences of the pollutant emission influence between the urban district and outer suburbs, i.e. the spatial and temporal distribution of air pollutant is closely related with that of the pollutant emission intensity. This study shows that due to complexity of the spatial and temporal distribution of pollution emission sources, the new generation Community Multi-scale Air Quality (CMAQ) model developed by the EPA of USA produced forecasts, as other models did, with a systematic error of significantly lower than observations, albeit the model has better capability than previous models had in predicting the spatial distribution and variation tendency of multi-sort pollutants. The reason might be that the CMAQ adopts average amount of pollutant emission inventory, so that the model is difficult to objectively and finely describe the distribution and variation of pollution emission sources intensity on different spatial and temporal scales in the areas, in which the pollution is to be forecast. In order to correct the systematic prediction error resulting from the average pollutant emission inventory in CMAQ, this study proposes a new way of combining dynamics and statistics and establishes a statistically correcting model CMAQ-MOS for forecasts of regional air quality by utilizing the relationship of CMAQ outputs with corresponding observations, and tests the forecast capability. The investigation of experiments presents that CMAQ-MOS reduces the systematic errors of CMAQ because of the uncertainty of pollution emission inventory and improves the forecast level of air quality. Also this work employed a way of combining point and area forecasting, i.e. taking the products of CMAQ for a center site to forecast air pollution for other sites in vicinity with the scheme of model products "reanalysis" and average over the "area".     

三峡库区巫山县城新址岩质库岸再造预测

巫山县城新址位于西坪和大宁河两岸.组成该库段的基岩主要为泥岩、粉砂岩、泥灰岩等,岩性较软弱,且地质构造变动强烈,褶皱、断裂发育,岩层破碎,崩塌、滑坡分布较多,库岸再造问题严重;同时,岸坡再造又严重影响着新县城的规划和库岸的防护.因此把稳态坡形类比法应用于巫山县城新址基岩岸坡的再造预测,定量预测库岸的最终再造宽度,以指导新城的规划与建设.     

空间运动对象与时空数据类型研究

当前时空数据模型多以描述空间实体的离散变化为主。该文中对空间运动对象在抽象层次的无限连续空间、离散层次的有限离散空间上的数据类型进行分析和定义，将其分别划分为时间类型、空间类型和时态类型来研究，并提出支持空间运动对象的表示方法和操作方式。该方法既能表示空间实体的连续运动，也能表示其离散变化，为空间运动对象时空数据模型的建立奠定了基础。     

基于高通量测序的表层水体纤毛虫多样性评估方法比较与优化

基于扩增子的高通量测序技术广泛应用于微型生物多样性的研究, 不同的测序手段和数据分析流程, 影响着微型生物多样性与群落结构的分析结果。本研究以易于形态鉴定的砂壳纤毛虫作为研究对象, 比较DNA测序、RNA测序和形态学方法检获的多样性和物种组成等, 探究序列分析流程中关键步骤: 嵌合体处理、可操作分类单元分析方法选择、合并相似分类单元以及去除稀有类群等对多样性结果的影响。研究结果显示无论基于DNA还是RNA的分子手段与形态学方法检获的主要物种基本一致, 与DNA测序相比, RNA测序检获的物种数少, 但差异不显著。基于97%以上相似度聚类和单核苷酸变异所得群落结构相似, 无显著差异; 且所有分析方法都能在一定程度上反映出自然界中不同类群的相对丰度。相较于单核苷酸变异和其他相似度阈值, 99%相似度下聚类所得多样性更为接近形态学结果。去除嵌合体和稀有类群(去除阈值: DNA测序0.05%; RNA测序0.07%), 可明显改善分子多样性虚高的问题。本研究为纤毛虫等真核微生物的分子多样性研究提供了科学的数据分析流程, 对未来真核微生物多样性研究具有指导意义。     

L波段海水介电模型构建与误差分析

基于实验室测量数据,构建了盐度和温度多项式形式的L波段海水介电模型,并采用最小二乘法和奇异值分解技术求解模型系数。将新模型与目前广泛使用的Klein-Swift(KS)模型、Meissner-Wentz(MW)模型以及George Washington University(GWU)模型进行比较。结果表明：新模型计算的海水介电常数与实验室测量数据的RMSE误差为0.09（实部）和0.25（虚部）,均优于KS、MW、GWU模型。     

波浪条件下小型回转体高速入水动力学特性研究

采用光滑粒子流体动力学（SPH）方法,建立了刚体入水六自由度响应模型,实现了波浪环境下小型回转体高速入水过程的数值模拟。对波浪环境下小型回转体高速入水过程中的运动姿态、冲击载荷进行讨论分析,结果表明：波浪条件下入水点相位不同,导致回转体实际入水角不同,实际入水角越大,入水冲击载荷越大,弹道越稳定;实际入水角越小,入水冲击载荷越小,弹道越易发生失稳。本研究表明,入水点相位对小型回转体弹道特性和冲击载荷产生影响,进而为波浪条件下跨介质武器低载稳定入水提供基础性技术支撑。     

山东诸城地区下白垩统一新类型鸟类足迹

1931年,Mehl报道的来自Dakota Group的Ignotornis mcconnelli是第一个被广泛接受的中生代鸟类足迹。目前,亚洲、北美洲、欧洲、非洲和南美洲都有白垩纪鸟类足迹的报道,尤其以东亚地区最为丰富。山东诸城张祝河湾恐龙足迹点保存了一枚完整的鸟类足迹,本文对该鸟类足迹进行了详细描述与对比。研究认为,该鸟类足迹区别于以往发现的鸟类足迹,根据其形态建立了新种：玲珑诸城鸟足迹Zhuchengornipes exquisitus。玲珑诸城鸟足迹造迹者可能为滨鸟类,形成于干旱气候下的滨浅湖环境。     

基于A- CGAN的深反射地震数据随机噪声压制方法研究

基于深度学习的地震数据噪声压制方法是当前地震数据去噪处理的重要方向。深度学习方法突破了传统滤波处理的局限,在对常规地震数据的噪声压制中表现出效率高、信噪分离效果好的特点。但针对深部弱有效反射数据,当前的深度学习方法特征提取能力有限,难以取得较好的去噪效果。笔者等结合深反射地震数据特点,针对当前深度学习噪声压制方法在特征提取及对数据集依赖上的局限,提出了基于注意力循环生成对抗网络(Attention Cycle- Consistent Generative Adversarial Networks,A- CGAN)的深反射地震数据随机噪声压制方法。借助循环一致生成对抗网络(Cycle- Consistent Generative Adversarial Networks,Cycle- GAN)的域映射思想,降低对数据集的要求。为了构建适用于深反射地震数据的去噪网络,从3个方面对Cycle- GAN进行改进：在Cycle- GAN的生成器（去噪器）中加入残差结构和注意力机制,用于加深网络深度和提高其特征提取能力;在Cycle- GAN的鉴别器中使用块判决,提高鉴别精度和准确度;在损失函数部分加入感知一致性损失函数,提升网络模型恢复纹理细节信息的能力。通过合成地震数据和实际深反射地震数据测试,验证了优化算法的有效性,体现了良好的应用价值。