A comparative study of extensive machine learning models for predicting long-term monthly rainfall with an ensemble of climatic and meteorological predictors

Rainfall prediction is of vital importance in water resources management. Accurate long-term rainfall prediction remains an open and challenging problem. Machine learning techniques, as an increasingly popular approach, provide an attractive alternative to traditional methods. The main objective of this study was to improve the prediction accuracy of machine learning-based methods for monthly rainfall, and to improve the understanding of the role of large-scale climatic variables and local meteorological variables in rainfall prediction. One regression model autoregressive integrated moving average model (ARIMA) and five state-of-the-art machine learning algorithms, including artificial neural networks, support vector machine, random forest (RF), gradient boosting regression, and dual-stage attention-based recurrent neural network, were implemented for monthly rainfall prediction over 25 stations in the East China region. The results showed that the ML models outperformed ARIMA model, and RF relatively outperformed other models. Local meteorological variables, humidity, and sunshine duration, were the most important predictors in improving prediction accuracy. 4-month lagged Western North Pacific Monsoon had higher importance than other large-scale climatic variables. The overall output of rainfall prediction was scalable and could be readily generalized to other regions.     

中国地震网格试验系统中资源管理与作业调度研究-资源管理和作业动态分配模型ProRMJS的建立me

通过分析与研究建立资源管理和作业动态分配ProRMJS模型,解决了中国地震灾害仿真网格试验系统（CEDAGrid）中网格科学计算平台在初期建设中存在的一些资源管理方面的问题. 针对网格科学计算平台默认各个计算节点均能提供计算服务的问题,ProRMJS通过ldquo;计算池rdquo;这一虚拟组织为调度器分发作业提供支撑;而调度器则根据计算池中各个节点的计算能力和状态,动态分配任务,这样就保证了计算平台的稳定性. 针对默认各个计算节点均能完成任务的问题, ProRMJS通过监控各个节点所负担作业的运行进度和设定作业时间阈值来管理作业运行. 对于各个节点计算能力大小的区别, ProRMJS通过对各个节点计算性能的权值量化,实行按ldquo;需rdquo;分配作业. 对各个节点计算性能的权值量化, ProRMJS是通过综合考虑各个因素对机器计算能力大小的影响后通过加权计算的方式完成的,从而提高了整个计算平台的工作效率. 最后通过实现地震应力触发科学计算实例验证了方案的有效性,为网格技术在地震领域内的进一步应用作了有意义的探索.      

Characteristics of water isotopes and hydrograph separation during the wet season in the Heishui River, China

Runoff generation and dynamics is an important issue in watershed and water resource management, but the mechanism in large scale is unclear and site-dependent. For this reason, spatial variations of δD and δ¹⁸O of river water and their sources within large-area of the Heishui Valley of the upper Yangtze River in western China were investigated during the wet season. A total 117 river water samples were collected at 13 sampling sites located at the junction of the principal river course and its tributaries. The results showed no spatial variations of either δD or δ¹⁸O values existed among tributary sampling sites A, B, E, F, H and I during the wet season, and significantly spatial variation occurred between tributary sampling sites A, B, E, F, H, I and site K; which indicated different proportions of rain entering river water should lead to spatial variation of water isotopes. The hydrograph separation analysis, based on the isotope data of river water, meltwater and rain water samples, showed the contribution of snow and glacier meltwater varied from 63.8% to 92.6%, and that of rain varied from 7.4% to 36.2%; which meant that snow and glacier meltwater was the main supplying water source of baseflow in the Heishui Valley. And the roles of glacier and snow meltwater should be significantly noticed in water resource management in this alpine valley at the rim of the Tibetan Plateau.     

Hydrogen isotope composition of natural gases from the Tarim Basin and its indication of depositional environments of the source rocks

By measuring carbon and hydrogen isotope compositions for C1, C2 and C3 of 74 gas samples, natural gases from the Tarim Basin can be divided into six groups on the basis of their origins: (1) coal-type gas derived from coal measures; (2) coal-type gas generated from the T-J lacustrine mudstones; (3) oil-type gas derived from the Cambrian and low Ordovician marine source rocks; (4) oil-type gas from the source rocks deposited in the marine-transitional facies; (5) mixing gas between gas derived from the Carboniferous transitional source rocks and the Mesozoic humic gas, and (6) mixing gases of thermal genetic gas and little deep gas in the Southwest depression of the Tarim Basin. The δ D values of methane in natural gases originating from different type kerogens are affected by both palaeo-environments of the source rock formation (kerogen types) and thermal maturity, with sedimentary environment (kerogen type) as the main controlling factor. Under the similar thermal maturity, the hydrogen isotope composition of methane is more enriched in deuterium in marine environments than lacustrine one. With the increase of thermal maturity and the increase of carbon atomic numbers of gaseous alkanes, the hydrogen isotopes become enriched in deuterium. The δ D values of ethane and propane (δ D2, δ D3) are controlled mainly by thermal maturity and to a lesser degree by sedimentary environment of the source rock formation. The partial reversal of hydrogen isotopes for gaseous alkanes would be related to the microbial oxidation, mixing of sapropelic and humic gases and / or mixing of gases from similar kerogen sources with various thermal maturities. In the oil-type gas, the sulfate reduction reaction would result in the reversed order of δ D1 and δ D2 (e.g. δ D1>δ D2).     

The formation evaluation tool and its application in offshore China

The Formation Evaluation Tool （FET） introduced in the paper represents a new generation of formation evaluation systems developed and manufactured by China Oilfield Services Limited （COSL）, CNOOC, using a FET technology transfer from Crocker Research, Australia. The system has been applied successfully in the Bohai Sea and South China Sea. For instance, a multilayered oil and water system has been confirmed with the aid of accurate formation pressure tests, even in very thin beds and edge water reservoirs, overcoming the difficulty of determining this kind of oil-water and gas-water contacts. Moreover, the FET pumping and real-time fluid monitoring function allows acquiring a true sample of formation fluid unpolluted by drilling mud which plays an important role in determining the fluid properties of the target stratum and analyzing the fluid component. The principles and purpose of the Formation Evaluation Tool （FET） will be briefly introduced and successful examples of the application of the technology will be described in detail in this paper.     

Variations of near-surface atmospheric CO2 and H2O concentrations during summer on Muztagata

Expeditions during the summers of 2002 and 2003 implemented continuous monitoring of near-surface (2 m height) atmospheric CO₂ and H₂O concentrations at the 4500 m elevation on Muztagata. The resultant data sets reveal a slight decrease of CO₂ concentrations (of about 5 μmol·mol^-1) and changes in the diurnal variations from the end of June to the middle August. The daily maximum CO₂ concentrations occur between 02:30-05:30 AM (local time) and the minimum levels occur between 12:00-15:30 PM. The atmospheric CO₂ concentrations in the summer of 2002 were around 5 μmol·mol^-1 lower than those during the same period of 2003, whereas the diurnal amplitude was higher. In contrast, we found that the daily mean atmospheric H₂O content in 2003 was much lower than that in 2002 and there exists a striking negative correlation between CO₂ and H₂O concentrations. We therefore suggest that the near-surface atmospheric CO₂ concentration is affected not only by photosynthesis and respiration, but also by the air H₂O content in the glaciated region around Muztagata.     

Spatiotemporal variations of internal P-loading and the related mechanisms in the large shallow Lake Chaohu

Spatiotemporal variations of P species and adsorption behavior in water column, interstitial water, and sediments were investigated in the large shallow eutrophic Lake Chaohu. Orthophosphate (Ortho-P) and total phosphorus (TP) concentrations were significantly higher in the western part than in the eastern part of the lake, due to different nutrient inputs from the surrounding rivers. Moreover, particulate phosphorus (PP) concentration was in a similar spatial pattern to Ortho-P and TP concentrations, and also showed significantly positive correlation with the biomass of Microcystis, indicating more uptake and store of phosphorus by Microcystis than by other algae. Increase of pH and intensive utilization of P by phytoplankton were the main factors promoting P (especially Fe-P) release from the sediment to interstitial water during the cyanobacterial blooms in Lake Chaohu. Spatial dynamics in TP concentration, P species and adsorption behavior of the sediment, coupled with the statistical analyses, suggested that the spatial heterogeneity of P contents in the sediment was influenced by various factors, e.g. human activities, soil geochemistry and mineral composition. In spite of similar TP contents in the sediments, increase in proportion of Fe-P concentration in the sediment may result in a high risk of P release.     

Distribution of the indigenous microorganisms and mechanisms of their orientational activation in Daqing Oilfield

The distribution of indigenous microorganisms was surveyed in Block 1 of Daqing Oilfield. Based on this survey, the indigenous microorganisms in the formation water were activated with different activator systems at the simulated stratum ecological environment. The changes of the number of bacteria of various physiological groups were determined during the process of activation. Also changes of pH value and composition of gas productions were analyzed at the end of culturing. The results showed that the selected block formation water contained a great number of saprophytic bacteria, hydrocarbon-oxidizing bacteria, fermentative bacteria, methane-producing bacteria and sulfate-reducing bacteria. Under the conditions that the growth of sulfate-reducing bacteria was controlled the block had the potential to enhance oil recovery by activating beneficial bacteria. The growth of sulfate-reducing bacteria can be inhibited through the activation of nitrate-reducing bacteria. The number of nitrate-reducing bacteria reached 10⁶–10⁷ cells/mL, but sulfate-reducing bacteria reached only 0–45 cells/mL in A system. Methane-producing bacteria can be activated by C, D activators. The relative content of biological methane in the light hydrocarbon gas reached 80% in C, D systems. B activator was conducive to the propagation of acid-producing bacteria, so that the pH value of the culture medium decreased from 7.5 to around 5.0. Hydrocarbon-oxidizing bacteria can be activated by various activator systems. There was low molecular light hydrocarbon in gas production according to the analysis of gas chromatograph. According to the content of methane and the number of methane-producing bacteria, methane only can be generated through activating methane-producing bacteria. By choosing different activator systems, various populations of indigenous microorganisms can be activated accordingly.     

Dipping-interface Mapping Using Mode-separated Rayleigh Waves

Multichannel analysis of surface waves (MASW) method is a non-invasive geophysical technique that uses the dispersive characteristic of Rayleigh waves to estimate a vertical shear (S)-wave velocity profile. A pseudo-2D S-wave velocity section is constructed by aligning 1D S-wave velocity profiles at the midpoint of each receiver spread that are contoured using a spatial interpolation scheme. The horizontal resolution of the section is therefore most influenced by the receiver spread length and the source interval. Based on the assumption that a dipping-layer model can be regarded as stepped flat layers, high-resolution linear Radon transform (LRT) has been proposed to image Rayleigh-wave dispersive energy and separate modes of Rayleigh waves from a multichannel record. With the mode-separation technique, therefore, a dispersion curve that possesses satisfactory accuracy can be calculated using a pair of consecutive traces within a mode-separated shot gather. In this study, using synthetic models containing a dipping layer with a slope of 5, 10, 15, 20, or 30 degrees and a real-world example, we assess the ability of using high-resolution LRT to image and separate fundamental-mode Rayleigh waves from raw surface-wave data and accuracy of dispersion curves generated by a pair of consecutive traces within a mode-separated shot gather. Results of synthetic and real-world examples demonstrate that a dipping interface with a slope smaller than 15 degrees can be successfully mapped by separated fundamental waves using high-resolution LRT.     

特大目标分类别的地震抢险方案模型设计研究

研究“十五”“地震应急指挥技术系统”中“地震应急指挥辅助决策系统”的一部分。主要研究“特大目标分类别的抢险设计方案和通用性建议模型”,它主要是为地震应急指挥提供辅助决策背景支持信息和特大目标分类别的抢险设计方案,通过检索和查询应急基础数据库,生成专题图表,显示特大目标的相关信息,并结合灾区实际情况,为指挥部成员提供提醒性的决策参考信息。