区域地质调查文本中文命名实体识别

作为我国地质调查领域最重要的数据源之一,地质调查报告中蕴含着丰富的地学知识及地质体描述等关键信息,准确高质量地抽取地质命名实体为地学知识图谱构建、知识推理及知识演化提供基础。笔者等在阐述地质命名实体识别任务基础上,分析地质实体不仅包含大量专业术语,还存在实体嵌套、大量长实体等领域特性,进一步增加了地质命名实体识别难度。笔者等提出一种基于轻量级预训练模型(ALBERT)—双向长短时记忆网络(BiLSTM)—条件随机场(CRF)模型的地质命名实体识别方法。首先利用ALBERT对输入字符上下文特征进行建模,并采用BiLSTM对其进行进一步上下文特征表征,最后采用CRF实现标注序列预测。实验结果表明,在构建的地质命名实体识别数据集上,相比于主流的命名实体识别模型算法,本文所提出的方法具有更好的抽取性能,提出的命名实体识别模型能为领域实体识别提供借鉴,同时为地学领域实体关系抽取和地学知识图谱构建提供有力方法支撑。     

The GEOMON network of Czech catchments provides long-term insights into altered forest biogeochemistry: From acid atmospheric deposition to climate change

In 1994, a network of small catchments (GEOMON) was established in the Czech Republic to determine input–output element fluxes in semi-natural forest ecosystems recovering from anthropogenic acidification. The network consists from 16 catchments and the primary observations of elements fluxes were complemented by monitoring of biomass stock, element pools in soil and vegetation, and the main water balance components. Over last three decades, reductions of SO₂, NO_x and NH₃ emissions were followed by sulphur (S) and nitrogen (N) deposition reductions of 75% and 30%, respectively. Steeper declines of strong acid anion concentrations compared to cations (Ca, Mg, Na, K, NH₄) in precipitation resulted in precipitation pH increase from 4.5 to 5.2 in bulk precipitation and from 4.0 to 5.1 in spruce throughfall. Stream chemistry responded to changes in deposition: S leaching declined. However at majority of catchments soils acted as a net source of S to runoff, delaying recovery. Stream pH increased at acidic streams (pH < 6) and aluminium concentration decreased. Stream nitrate (NO₃) concentration declined by 60%, considerably more than N deposition. Stream NO₃ concentration was tightly positively related to stream total dissolved nitrogen to total phosphorus (P) ratio, suggesting the role of P availability on N retention. Trends in dissolved organic carbon fluxes responded to both acidification recovery and to runoff temporal variation. An exceptional drought occurred between 2014 and 2019. Over this recent period, streamflow decreased by ≈ 40% on average compared to 1990s, due to the increases of soil evaporation and vegetation transpiration by ≈ 30% and declines in precipitation by ≈ 15% on average across the elevational gradient. Sharp decreases of stream runoff at catchments <650 m a.s.l. corresponded to areas of recent forest decline caused by bark beetle infestation on drought stressed spruce forests. Understanding of the interactions among legacies of acidification and eutrophication, drought effects on the water cycle and forest disturbance dynamics is requisite for effective management of forested ecosystems under anthropogenic influence.     

Analysis and modelling of rainfall fields at different resolutions in southern Italy

Abstract

The spatial and temporal variability of the scaling properties and correlation structure of a data set of rainfall time series, aggregated over different temporal resolutions, and observed in 70 raingauges across the Basilicata and Calabria regions of southern Italy, is investigated. Two types of random cascade model, namely canonical and microcanonical models, were used for each raingauge and selected season. For both models, different hypotheses concerning dependency of parameters on time scale and rainfall height can be adopted. In particular, a new approach is proposed which consists of several combinations of models with a different scale dependence of parameters for different temporal resolutions. The goal is to improve the modelling of the main features of rainfall time series, especially for cases where the variability of rainfall changes irregularly with temporal aggregation. The results obtained with the new methodology showed good agreement with the observed data, in particular, for the summer months. In fact, during this season, rainfall heights aggregated at fine temporal resolutions (from 5 to 20 min) are more similar (relative to the winter season) to the values cumulated on 1 or 3 h (due to convective phenomena) and, consequently, the process of rainfall breakdown is nearly stationary for a range of finer temporal resolutions.

Editor D. Koutsoyiannis; Associate editor A. Montanari     

2050年前中国雾日变化趋势的预估

 利用1971-2005年中国591个气象台站的雾日资料以及逐日最低气温、相对湿度、平均风速资料,分析了35 a来中国各区域年雾日数与这些因子的相关关系,并利用IPCC第四次评估报告所提供的模式数据资料,针对3种不同的排放情景,对21世纪上半叶各区域年平均雾日进行预估。结果表明：对划分的9个雾区的年雾日数的回归方程的拟合效果较好,可以用来进行预估;未来50 a中国大部分地区雾日呈明显减少的变化趋势,在A1B,A2和B1情景下,雾日减少的平均幅度分别为16.2%,13.4%和12.9%。未来50 a中国雾日预估结果的空间分布显示：3种情景下未来中国大部分雾区雾日数都将减少,个别地区雾日数有增加趋势,其中A1B情景下雾日减少区的减少趋势最明显,而B1情景下雾日增加区的增加趋势最明显。     

A two-stage time series model for monthly hydrological projections under climate change in the Lim River basin (southeast Europe)

ABSTRACT

Climate change projections of precipitation and temperature suggest that Serbia could be one of the most affected regions in southeastern Europe. To prepare adaptation measures, the impact of climate changes on water resources needs to be assessed. Pilot research is carried out for the Lim River basin, in southeastern Europe, to predict monthly flows under different climate scenarios. For estimation of future water availability, an alternative approach of developing a deterministic-stochastic time series model is chosen. The proposed two-stage time series model consists of several components: trend, long-term periodicity, seasonality and the stochastic component. The latter is based on a transfer function model with two input variables, precipitation and temperature, as climatic drivers. The Nash-Sutcliffe model efficiency for the observed period 1950–2012 is 0.829. The model is applied for the long-term hydrological prediction under the representative concentration pathway (RCP) emissions scenarios for the future time frame 2013–2070.     

中国南海东部海域气候特征及风浪流极值参数的研究——LAGFD数值模式群的应用

本文采用先进的ＬＡＧＦＤ风、浪数值模式和ＰＯＭ（ＰｒｉｎｃｅｔｏｎＯｃｅａｎＭｏｄｅｌ）三维海流模式对自１９４５～１９９５年间发生并影响南海东部海域的２９９个历史最强热带气旋过程进行数值后报,给出了南海东部部分海域（１９°～２３°Ｎ,１１３°～１１８°Ｅ）中１０００ｍ等深线内６０个点的多年一遇风、浪、流和水位极值,并简要分析了南海东部（１５°～２７°Ｎ,１０８°～１２２°Ｅ）的气候特征,为该海域区域性海洋环境研究与工程开发提供了基础参考数据。     

半参数模型稳健估计及其应用

在半参数模型估计中,均假设观测误差服从正态分布。当观测量含有粗差时,粗差对参数和非参数估计的影响是不可忽略。基于此,首先在总结线性参数模型稳健估计基本理论的基础上,论述了M估计权因子的确定方法。然后提出了半参数模型稳健估计方法,并导出半参数模型(广义)补偿最小二乘稳健估计的基本公式。最后通过两个模拟算例验证了其估计方法的有效性。     

A review of research on seismic behaviour of irregular building structures since 2002

The paper presents an overview of the progress in research regarding seismic response of plan and vertically irregular building structures. Three areas of research are surveyed. The first is the study of the effects of plan-irregularity by means of single-storey and multi-storey building models. The second area encompasses passive control as a strategy to mitigate torsional effects, by means of base isolation and other types of devices. Lastly, the third area concerns vertically irregular structures and setback buildings. Although fewer papers have been published in this last area with respect to the former ones, this state-of-the-art reports extensively on research efforts and progress into the seismic behaviour of irregular buildings in elevation to show the growing interest among specialists in the field.     

Source identification of PM2.5 particles measured in Gwangju, Korea

The UNMIX and Chemical Mass Balance (CMB) receptor models were used to investigate sources of PM_2.5 aerosols measured between March 2001 and February 2002 in Gwangju, Korea. Measurements of PM_2.5 particles were used for the analysis of carbonaceous species (organic (OC) and elemental carbon (EC)) using the thermal manganese dioxide oxidation (TMO) method, the investigation of seven ionic species using ion chromatography (IC), and the analysis of twenty-four metal species using Inductively Coupled Plasma (ICP)-Atomic Emission Spectrometry (AES)/ICP-Mass Spectrometry (MS). According to annual average PM_2.5 source apportionment results obtained from CMB calculations, diesel vehicle exhaust was the major contributor, accounting for 33.4% of the measured PM_2.5 mass (21.5 μg m^− 3), followed by secondary sulfate (14.6%), meat cooking (11.7%), secondary organic carbon (8.9%), secondary nitrate (7.6%), urban dust (5.5%), Asian dust (4.4%), biomass burning (2.8%), sea salt (2.7%), residual oil combustion (2.6%), gasoline vehicle exhaust (1.9%), automobile lead (0.5%), and components of unknown sources (3.4%). Seven PM_2.5 sources including diesel vehicles (29.6%), secondary sulfate (17.4%), biomass burning (14.7%), secondary nitrate (12.6%), gasoline vehicles (12.4%), secondary organic carbon (5.8%) and Asian dust (1.9%) were identified from the UNMIX analysis. The annual average source apportionment results from the two models are compared and the reasons for differences are qualitatively discussed for better understanding of PM_2.5 sources.Additionally, the impact of air mass pathways on the PM_2.5 mass was evaluated using air mass trajectories calculated with the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) backward trajectory model. Source contributions to PM_2.5 collected during the four air mass patterns and two event periods were calculated with the CMB model and analyzed. Results of source apportionment revealed that the contribution of diesel traffic exhaust (47.0%) in stagnant conditions (S) was much higher than the average contribution of diesel vehicle exhaust (33.4%) during the sampling period. During Asian dust (AD) periods when the air mass passed over the Korean peninsula, Asian dust and secondary organic carbon accounted for 25.2 and 23.0% of the PM_2.5 mass, respectively, whereas Asian dust contributed only 10.8% to the PM_2.5 mass during the AD event when the air mass passed over the Yellow Sea. The contribution of biomass burning to the PM_2.5 mass during the biomass burning (BB) event equaled 63.8%.