Recycling in the Puerto Rican mantle wedge, Greater Antilles Island Arc

Abstract The initial volcanic phase of Cretaceous island arc strata in central Puerto Rico, at the eastern end of the extinct Greater Antilles Arc, comprises a 6‐km thick pile of lava and volcanic breccia (Río Majada Group). Preserved within the sequence is a conspicuous shift in absolute abundances of the more incompatible elements, including Th, Nb, and the light rare earth elements (LREE: La, Ce, Pr and Nd). The compositional shift is marked by a decrease in La/Sm from averages of 2.11 in the lowest third of the pile (Formation A) to 1.48 at the top (Formation C), and by a distinctive flattening of LREE segments of chondrite‐normalized REE patterns. i⁸⁷Sr/⁸⁶Sr and ?_Nd average about 0.7035 and 8.2, respectively, in early Formation A basalts. These ranges normally overlap samples from later Formations B and C. Isotope compositions of the latter group are more variable, however, and several samples are considerably more radiogenic than Formation A basalts, such that i⁸⁷Sr/⁸⁶Sr averages almost 0.7042 while ?_Nd‐values decrease to 7.5 in Formation B and C basalts. Theoretical models of non‐modal melting processes in both amphibole peridotite and spinel lherzolite sources provide insight into the origin of depleted Th, Nb, and LREE abundances in Puerto Rican basalts. Low Nb concentrations less than normal mid‐oceanic ridge basalts in Formation A basalts indicate the wedge was slightly depleted by low‐volume decompression fusion due to induced convection in the back‐arc region prior to entry of the source into the arc melting zone. However, depleted patterns in Formation C basalts cannot be generated by relatively greater degrees of decompression fusion in the back‐arc, because addition of the La‐enriched slab‐derived component to more depleted source material invariably produces elevated rather than decreased La/Sm. Refluxing of Formation A harzburgitic residua is similarly precluded. In contrast, the observed patterns are readily reproduced by multistage melting models involving hybridized sources containing normal Formation A lherzolite source material blended with recycled, unrefluxed harzburgite residua. Successful models require hybrid sources containing large volumes of recycled harzburgite (up to 50%) during generation of Formation C basalts. Slightly elevated radiometric Sr and Nd isotopes in a few flows from Formation C are attributed to partial refluxing of the hybrid sources within the wedge.     

Statistical features of aftershock distribution size for moderate and large earthquakes in Chinese mainland

Based on data of earthquake sequences with MS≥5.0 in Chinese mainland from 1970 to 2004,for different se-quence types and different rupture modes of the main shock,the relationship between aftershock distribution size Rand the magnitude of the main shock M0 has been studied statistically.Considering the rupture mode of the mainshock,we give the quantitative statistical relationships between R and M0 under 95%confidence level for differentsequence types.Qualitatively,lgR,the logarithm of the aftershock distribution size,is positively correlative to theM0,but the data distribution is dispersed.Viewing from different sequence types,the correlation between R and M0is very weak for isolated earthquake type(IET)sequence,R distributes in the range from 5 to 60 km;For main-shock-aftershock type(MAT),lgR is positively correlative to M0;For multiple main shock type(MMT),the core-lation between lgR and M0 is not very obvious when M0≤6.2 and R distributes in the range from 5 to 70 km,whileit shows a linear correlation when M0≥6.3.The statistical results also show that the occupational ratios of differentsequence types for strike-slip and oblique slip are almost the same.But for dip-slip(mostly are thrust mechanisms),the ratio of MAT is higher than that of IET and MMT.Comparing with previous results,it indicates that,when M0is large enough,R is mainly determined by M0 and there is almost no relationship with the rupture mode of themain shock.     

基于粗细网格的有限元并行分析方法

并行计算己成为求解大规模岩土工程问题的一种强大趋势。探讨了粗细网格与预处理共轭梯度法结合的并行有限元算法。从多重网格刚度矩阵推得有效的预处理子。该算法在工作站机群上实现。用地基处理时土体强夯的数值模拟分析进行了数值测试,对其并行性能进行了详细分析。计算结果表明：该算法具有良好的并行加速比和效率,是一种有效的并行算法。     

放射虫硅质岩对华南古地理的启示

放射虫硅质岩在中国南方及邻区广泛分布,有重要的古地理意义。按其时空分布,可分为南区及北区和东带。南区包括滇西、滇东南、桂西和桂南,主要为晚古生代至中生代早期放射虫硅质岩。滇西硅质岩带南延至泰国和马来西亚,代表古特提斯主支。滇东南、桂西和桂南的硅质岩则指示古特提斯多岛洋的分支海盆。北区包括两广中北部、湘赣中南部和长江中下游一带,主要为二叠纪中晚期放射虫硅质岩,标志当时扬子台缘与古特提斯洋连通的深水盆地。东带沿南海和东海外侧的岛弧分布,从菲律宾北巴拉望经琉球到西南日本内带,主要为中二叠世至晚侏罗世放射虫硅质岩,与古地磁证据一起指示了该期间华南南方一个低纬度的远洋盆地,可称为“古南中国海”。它的张开可能是中晚二叠世云开地体和中国东南部其他地方造山事件的原因,它的随后发展对华南东部三叠纪和侏罗纪古地理演化也有重大影响。晚侏罗世太平洋伊泽奈崎板块的迅速北移,使“古南中国海"俯冲消减,导致东南沿海大规模的钙碱性岩浆活动。     

上思县油茶栽培气候生态条件分析

通过对上思县及全国其它适宜种植区的油茶栽培中油茶开花授粉、果实膨大、油脂积累等重要生长发育期的气象条件分析,找出上思县适宜发展油茶种植的科学依据,为当地发展油茶业提出建议,为当地林业可持续发展提供参考。     

中国农业气象模式(CAMM1.0)构建与应用

为发展适宜中国区域农业种植特点的农业气象模式,基于国外作物生长模拟方法,通过模式机理过程改进或重构以及应用方式革新,建立了中国农业气象模式（Chinese AgroMeteorological Model version 1.0,CAMM1.0）。CAMM1.0利用平均温度和土壤水分改进了作物发育进程模式,利用土壤水分改进了作物叶片光合作用、干物质分配和叶面积扩展过程模式,通过蒸发比法扩展了作物蒸散过程模式;自主建立了基于发育进程的冬小麦株高、基于遥感信息的作物灌溉、遥感数据同化、作物长势与灾害评价等模式。基于互联网技术构造了实时运转平台,主要功能包括作物生长过程实时常规模拟与用户个性化定制模拟。CAMM1.0的部分子模式采用多种方法构造,便于多模式集成。CAMM1.0对作物发育进程、光合过程、株高的模拟效果较好,但对土壤水分变化过程的拟合略差,模拟产量略偏低。CAMM1.0评价淮河流域夏玉米年际干旱减弱而涝渍增加的趋势与实际基本相符。     

阶段性干旱胁迫对小麦光合生理特性及产量结构的影响

在人工遮雨的条件下,采用盆栽的种植方式探究"皖麦68"营养生长期(返青期—开花期)及生殖生长期(开花期—成熟期)轻度干旱胁迫(土壤相对含水量为55%±5%)及复水(土壤相对含水量为70%±5%)对其光合生理特性及产量结构的影响。结果表明:返青期至成熟期充分供水(CK)的小麦旗叶光合参数和产量最高。开花至成熟期复水(DN)的小麦叶片在复水后光合能力迅速恢复,表现出了超补偿效应:光合速率(16.43μmol/(m~2·s))甚至超过了CK(15.01μmol/(m~2·s));采用非直角双曲线模型拟合小麦旗叶的光响应曲线,其中DN的曲角θ最大;DN产量较CK略有降低但千粒重为34.51 g,高于CK(34.44 g)。开花至成熟期轻度干旱(ND)及全生育期轻度干旱(DD)的小麦光合特征参数与产量均显著降低。DD产量最低、品质最差,但其收获指数I_H高于CK、仅次于DN。在小麦返青期—开花期进行水分管理适量减少灌溉,开花期—成熟期复水能够提升籽粒的干物质积累量,获得较高的产量及品质。     

VLAN技术在福建省地震信息网络中的应用

本文介绍了如何在福建省地震信息网络中应用VLAN技术规划网络并进行有效的管理,减少广播风暴的发生,提高网络安全性。     

黔北普宜地区晚三叠世二桥组砂岩碎屑锆石U-Pb年龄、重矿物分析及地质意义

普宜地区位于上扬子腹地——四川盆地东南缘,区内晚三叠世碎屑岩记录了上扬子前陆盆地演化信息,是研究盆地东缘盆山耦合的理想对象.二桥组主要由三角洲平原-三角洲前缘相砂岩组成,交错层理恢复的物源主要来自东侧,砂岩岩屑主要由岩浆岩岩屑和变质岩岩屑组成,碎屑重矿物主要由锆石、电气石、金红石、白钛石、黄铁矿、锐钛矿、石榴子石、铬铁...     

基于深度学习的天气雷达异常数据识别技术

天气雷达基数据中因观测设备故障或标定问题而产生的异常数据, 直接影响天气雷达数据质量、定量估测降水及天气系统的分析和判断。目前在中国气象局气象探测中心实时业务中, 通过人工勘误环节对异常数据进行处理。针对2020—2022年业务中勘误较多的、大面积故障异常和易与降水数据混合的局部电磁干扰或故障的两类异常数据, 分别构建和训练R-ResNet和R-LinkNet两种模型, 提取雷达硬件故障、电磁干扰等特征, 实现异常数据的识别和处理。评估结果表明:两种模型在提取异常数据特征方面均具有很强的学习能力, R-ResNet在分类判识异常数据与正常数据的准确率超过99%, R-LinkNet在分离电磁干扰杂波和降水回波的准确率超过98%。两种模型可用于实时业务中监控和勘误电磁干扰、故障等异常数据, 实现异常数据的自动勘误处理。