Journal of University of Chinese Academy of Sciences

Application of LSTM neural network for intermediate-term earthquake prediction: retrospective prediction of 2008 Wenchuan M_S8.0 Earthquake

SHI Yaolin, LI Linfang, CHENG Shu

2022, 39 (1): 1-12. DOI: 10.7523/j.ucas.2021.0063

Abstract ( 877 ) PDF (0KB) ( 0 )

Earthquake prediction is a difficult problem in contemporary science, and applications of machine learning methods in the prediction have drawn intensive attention. Large earthquakes can cause huge casualties and economic losses, and are the main goals of earthquake prediction. We studies intermediate-term (one-year) earthquake prediction in Sichuan and Yunnan provinces using the earthquake catalogue since 1970 by the sliding time-space window technique and LSTM(long short-term memory) neural networks. Sixteen earthquake prediction indexes that reflect the temporal and spatial features of earthquake sequences were used in the neural network. The neural network was trained using data sets from 1970 to 2004 (70% of all earthquake catalogues). Retrospective prediction tests were conducted on earthquakes after 2005, the accuracy rate (actual magnitude fell within ±0.5 of the predicted magnitude) was 70.2%, over-prediction rate was 18.7%, and under-prediction rate was 11.1%. The 2008 Wenchuan M_S8.0 earthquake was retrospectively predicted. In order to understand the robustness of the model, we have done some tests, such as to expand the study area, change the weights of large earthquakes in calculation of the mean square error, etc. The LSTM neural network model still performed well in the tests.

Molecular dynamics study of the effect of lithium on the tensile mechanical behavior of alpha-iron

WEI Wei, YU Xin'gang

2022, 39 (1): 13-20. DOI: 10.7523/j.ucas.2020.0024

Abstract ( 475 ) PDF (0KB) ( 0 )

Liquid lithium first wall is an advanced concept in the field of magnetic confinement fusion. Compared with the traditional solid wall materials, more and more experiments have shown that liquid lithium wall has unique advantages, which can not only effectively solve a series of problems faced by solid wall materials such as sputtering and tritium retention, but also significantly improve the constraint performance of core plasma. However, liquid lithium has brought severe corrosion problems to most metal materials, and the microstructure changes caused by corrosion will inevitably lead to the degradation of the macroscopic mechanical properties of materials. For structural materials, the degradation of mechanical properties will bring safety risks to the entire device. Based on this, the uniaxial tensile simulations of single crystal alpha-iron were carried out by using molecular dynamics, and the effects of lithium atoms on the plastic deformation behavior and yield stress were analyzed in this paper. The results show that lithium atoms can significantly suppress the phase transition, change the plastic deformation mode of alpha-iron, and affect the yield stress and its relationship with temperature.

Spatiotemporal variation of grassland aboveground biomass in Inner Mongolia from 2000 to 2019

YUE Rongwu, ZHANG Na, WANG Jingjie, LI Zhenyu, YAN Zhihui, FENG Yiming

2022, 39 (1): 21-33. DOI: 10.7523/j.ucas.2020.0047

Abstract ( 758 ) PDF (0KB) ( 23 )

Based on the acquisition of a large number of grassland aboveground biomass (AGB) field sampling data and the corresponding remote sensing indices, random forest (RF) models were constructed for different grassland types in Inner Mongolia. After validating the reliability of the models, the maximum AGB values from 2000 to 2019 were simulated. The results showed that the average annual maximum AGB was (82.74±56.34)g/m², gradually declining from northeast to southwest. The grassland types in the order of decreasing AGB:temperate meadow steppe, lowland meadow, temperate typical steppe, temperate desert steppe, temperate steppe desert, and temperate desert. From 2000 to 2019, the maximum AGB showed an overall significantly rising trend, and increased most significantly for lowland meadow and temperate typical steppe. The maximum AGB exhibited increasing trend for 70.62% of Inner Mongolia, and increased significantly for 16.28% of the area. The area with significantly increasing AGB was larger than that with significantly decreasing AGB for all the grassland types. However, the maximum AGB for all grassland types was on the fluctuant increase, and the temperature desert steppe had the largest interannual fluctuation while temperature desert the smallest. The interaction between summer air temperature and accumulated precipitation had the greatest positive effect on AGB, followed by the total precipitation from January to August, while summer air temperature had little negative effect on AGB.

Comparison of natural ¹⁵N abundance technique and ¹⁵N dilution technique in the determination of plant nitrogen fixation

LI Runfu, NIU Haishan, KONG Qian, LIU Qiang

2022, 39 (1): 34-42. DOI: 10.7523/j.ucas.2021.0008

Abstract ( 987 ) PDF (0KB) ( 0 )

Biological nitrogen fixation (BNF) is an important source of nitrogen in ecosystem, therefore, several experimental techniques and calculating methods have been developed to quantify the contribution of BNF to plants and vegetations. In this study, percentage of plant N derived from N₂-fixation (%Ndfa) was evaluated for a leguminous species, Astragalus arnoldii, in an alpine steppe in Qinghai-Tibet Plateau in order to compare two dominant in situ experimental techniques with different calculating approaches, reference plants, and sampling times. The δ¹⁵N values of four reference plants, i.e., Stipa purpurea, Kobresia pygmaea, Leontopodium nanum, and Carex moorcroftii, were significantly lower in late August than in late July (P<0.05). However, they were significantly higher in later August, 30 days after application of (¹⁵NH₄)₂SO₄ to soil, than late July which was 24h after labeling (P<0.05). In either site with or without (¹⁵NH₄)₂SO₄ application, δ¹⁵N of A. arnoldii did not differ in the two sampling times. Besides,%Ndfa of A. arnoldii did not differ between calculation based on ¹⁵N excess relative to atmospheric N₂ and that relative to unlabeled plants. Nevertheless,%Ndfa measured by natural abundance technique was significantly higher than that by isotope dilution technique (P<0.05) except for L. nanum as the reference species. The sampling time had a significant effect on estimated%Ndfa values in both the naturalabundance technique (F=89.906, P<0.01), and isotope dilution technique (F=496.712, P<0.01).

Transport pathways and potential source regions of PM_2.5 in Wuhai City of northwest arid area

YU Chuang, CHEN Wei, ZHANG Yuxiu

2022, 39 (1): 43-54. DOI: 10.7523/j.ucas.2020.0055

Abstract ( 502 ) PDF (0KB) ( 0 )

Wuhai was one of the major coal industry cities in northwest arid area of China, the concentration variation characteristics, transport pathways, and the potential source regions of PM_2.5 were unclear. The transport pathways and the potential source regions of PM_2.5 in Wuhai were discussed by cluster analysis, potential source contribution function (PSCF) and concentration-weighted trajectory (CWT) methods based on the hourly monitoring data of PM_2.5 mass concentration in Wuhai from 2016 to 2018. Results showed that the average annual concentration of PM_2.5 had a downward trend during 2016-2018, with the highest concentration in winter and the lowest in summer. Cluster analysis showed that the northwesterly pathways were the major transport pathway of PM_2.5 for all seasons. The long-distance transport airflow mostly occurred in spring, autumn, and winter, the concentration of PM_2.5 in four trajectories was about 97.96-151.33μg·m^-3, while the short-distance transport airflow was main pathway in summer, its concentration of PM_2.5 was about 87.11-96.88μg·m^-3. The PSCF and CWT analysis indicated that the potential source regions of PM_2.5 were the largest in winter, mainly occurred in Kumtag Desert, Qaidam Basin, Tengger Desert, Badain Jaran Desert, and Hexi Corridor Area. The main potential source regions in spring and autumn were located in Kumtag Desert, and Hexi Corridor Area. The potential source region was the smallest in summer, mainly came from Hexi Corridor Area. During the heavy pollution period, the main transport pathways of PM_2.5 came from northwest also, the potential source regions were mainly located in the border area of Qinghai and Gansu, parts of eastern Xinjiang and southern Wuhai. These results showed that the potential source regions of PM_2.5 in Wuhai were mainly located in the northwest arid desert area. Therefore, the implementation of wind prevention, sand fixation, and desertification control could effectively improve the air quality of Wuhai and the northwestern area of China.

Tracking hot spots of sustainable development research in domestic and foreign literature in the past twenty-five years: bibliometric analysis based on CiteSpace

ZHANG Xiaoping, ZHAO Yanyan, JIN Fengjun, SUN Wei

2022, 39 (1): 55-63. DOI: 10.7523/j.ucas.2020.0006

Abstract ( 1194 ) PDF (0KB) ( 0 )

Sustainable development is a theme concerned worldwide and has gained much attention from both domestic and foreign research fields. Based on the core databases of CNKI and WOS (Web of Science) and CiteSpace software, this paper tracks the research hotspots of sustainable development at home and abroad from 1994 to 2018. The results showed that:1) Research hotspots in common focused on multi-dimensional and multi-scale research on resources, environment and economy, sustainability evaluation, sustainable development strategy and policy, but the emphasis and perspective were different. 2) The research cooperation among international institutions was closely linked, while the cooperation between domestic institutions was relatively scattered and independent. 3) In contrast to Chinese literature focusing more on macro and regional scales, the research literature published in English from micro perspective on communities was relatively abundant. 4) Domestic studies tended to reveal problems in sustainable development, while foreign studies emphasized the role of science and technology, information and innovation in promoting regional sustainability. The comparative analysis of research hotspots at home and abroad is expected to shed new lights on future studies concerning sustainable development in China.

Regional differentiation and its influencing factors of traffic network in Jiangsu Province based on the expressway flow

ZHOU Jian, JIN Cheng, LI Pingxing

2022, 39 (1): 64-73. DOI: 10.7523/j.ucas.2020.0029

Abstract ( 480 ) PDF (0KB) ( 0 )

Based on the data of expressway traffic flow of Jiangsu Province, this paper builds the annual expressway traffic flow model, uses the method of social network analysis to explore the centrality and the agglomeration sub group of each county, and combines with the method of geographical detector to analyze the influencing factors from the perspective of the whole of Jiangsu Province, as well as the southern, central, and northern regions of Jiangsu. The results show that:1) On the whole, the spatial distribution of annual total outflow and inflow shows a decreasing trend from the south to the north; the areas with higher outflow rate are located at the junction of southern, central, and northern Jiangsu, the border areas of the whole province and the junction of Zhenjiang City and Changzhou City in southern Jiangsu. 2) The southern Jiangsu, with Suzhou, Nanjing, and Wuxi area as the core, has a strong traffic concentration and radiation effect on the whole province, and the cohesive subgroups show the characteristics of "four groups and eight zones" of the expressway network in Jiangsu Province in general.3) The results of detection factors show that economic level, industrial structure, car ownership, population scale, infrastructure, and purchasing power of residents all have a significant influence on the expressway flow in the province, and the explanatory power of factors are different in various regions.

Optimization of pretreatment parameters of produced water in coal series gas field with high salt, high hardness, and high turbidity

WANG Jingwei, LIU Yanping, JU Yiwen, JU Liting, LIU Xinchun

2022, 39 (1): 74-82. DOI: 10.7523/j.ucas.2021.0070

Abstract ( 402 ) PDF (0KB) ( 0 )

The produced water of coal-measure gas field in Linxing block, eastern margin of Ordos basin was characterized by high salt content, high hardness, and high turbidity. After experiment and engineering demonstration, the scheme of "pretreatment-reverse osmosis-MVR" was proposed to make the treated water meet standards for irrigation water quality. The pretreatment process of "hardness removal-flocculation-filtration-ultrafiltration" was simulated to explore the optimal process conditions for hardness and turbidity removal. The results showed that the hardness removal effect of Na₂CO₃ was better than that of NaOH. When the dosage of Na₂CO₃ was 15000mg·L^-1, the maximum hardness removal rate was 99.39%. The mathematical model established by response surface methodology with turbidity removal rate as response value was very significant (P<0.0001). The optimal process conditions were pH 8.85, PAC dosage 148.34mg·L^-1, PAM dosage 6.51mg·L^-1. Under these conditions, the predicted turbidity removal rate was 83.32%, the actual turbidity removal rate was 82.77%, indicating a high degree of fitting between the actual and predicted values. The optimization of pretreatment process parameters is conducive to the fine control of the project, so as to reduce the cost and provide a guarantee for the subsequent reverse osmosis desalination treatment.

Endmember optimization model of hyperspectral image based on constant constraint error

WANG Weijia, GENG Xiurui

2022, 39 (1): 83-90. DOI: 10.7523/j.ucas.2020.0008

Abstract ( 276 ) PDF (0KB) ( 0 )

The linear mixture model (LMM) plays a crucial role in endmember extraction. In general, under the assumption of LMM, the endmembers in an image can be obtained by minimizing the model reconstruction error. However, due to the existence of noise, the endmembers corresponding to the minimum model reconstruction error often deviate from the real ones. In order to balance the effects of reconstruction error and noise, the geometric optimization model is adopted to evaluate the reconstruction error in this paper and the reconstruction error is used as the constraint to further minimize the volume of the simplex. The presented method is called the error invariant constrained-optimal simplex volume method (EIC-OSV). The experiments with simulated and real hyperspectral data demonstrate that EIC-OSV can improve the overall accuracy of the popular endmember extraction methods.

Downstream response to the upstream water level variation and its application in flood early warning based on Sentinel-1A SAR images

GAO Long, YAN Fuli

2022, 39 (1): 91-101. DOI: 10.7523/j.ucas.2020.0007

Abstract ( 574 ) PDF (0KB) ( 0 )

For under-developed regions where the rivers have no or scarce hydrological gauging datasets, it is significant to explore the remote sensing techniques to determine the dynamic variation of up-/downstream water levels and to alert the potential flood inundation. In this work, the Nilwara Ganga in southern Sri Lanka, which is prone to floods, was taken as an example. A total of 14 scene Sentinel-1A SAR images from 2015 to 2017 were chosen to determine the up-/downstream flood peak levels. Based on the derived datasets, the prediction models of downstream flood peak levels were established, as well as the forecasting model on the maximum flood extent of the Nilwara Ganga. Consequently,the accuracy of the prediction model was evaluated, and an experiment of the predicted flood inundation was validated using 4 scene Sentinel-1A SAR images in 2018. The primary conclusions are summarized as follows:1) The fluctuation of the up-/downstream flood peak levels can be accurately and efficiently extracted by remote sensing technique;2) Among the established models, including quadratic polynomial, liner, power function, and exponential regression models, the exponential regression model under the ASTER GDEMV2 data is the optimal one, with R² of 0.79 and RMSE of 0.4, which means a consistent fluctuation between the upstream and downstream flood peak levels; 3) The validation results indicated that the overall accuracy of the predicted maximum flood extent is not less than 0.71. The method proposed in this paper aims to provide a new perspective for the flood early warning methodology using remote sensing techniques in the drainage area with less or no gauging datasets.

A change detection method by combining spectral-object-temporal features for remote sensing imagery

CAO Zhou, LIU Shibin, MA Yong, YAO Wutao, JIANG Liyuan

2022, 39 (1): 102-109. DOI: 10.7523/j.ucas.2020.0005

Abstract ( 410 ) PDF (0KB) ( 0 )

According to the fact that traditional change detection methods are difficult to meet the needs of practical applications in the era of big data of remote sensing, this paper proposes a method by combining spectral-object-temporal features to solve this problem. On the basis of extracting various features of remote sensing images, the Bi-LSTM network is used to extract the joint spectrum-temporal-object feature to obtain the effective information of changes in bi-temporal images. Experimental results based on bi-temporal medium-resolution remote sensing images show that the overall accuracy of this method is greater than 0.9, and the Kappa coefficient reaches 0.84. Compared to traditional methods, the proposed method effectively improves the accuracy and degree of automation of change detection.

Multi-target matching algorithm based on composite FMCW waveform

SONG Wenhao, ZHOU Bin, BU Zhiyong, ZHOU Zhigang

2022, 39 (1): 110-118. DOI: 10.7523/j.ucas.2021.0067

Abstract ( 522 ) PDF (0KB) ( 0 )

Millimeter-wave radar usually uses frequency modulated continuous wave(FMCW) to measure physical parameters such as distance and velocity of targets, but triangular waveform radar is difficult to accurately accomplish multi-target detection. As a result, an improved FMCW waveform and its corresponding multi-target matching algorithm are proposed in this paper, this composite waveform combines the characteristics of trapezoidal waveform and fast sawtooth waveform, through which multi-target detection can be finished more accurately. The simulation results verify that the method can accurately reject ghost targets and retain real targets.

Relay selection algorithm based on energy harvest in D2D communication

CHEN Guangzu, HUANG Xinchen, TAN Chong, BU Zhiyong

2022, 39 (1): 119-126. DOI: 10.7523/j.ucas.2020.0023

Abstract ( 431 ) PDF (0KB) ( 0 )

In order to select the relay node efficiently and energy-saving for D2D communication, we combine energy harvesting technology and D2D technology, build a D2D homogenous cell model based on energy harvesting, and propose a relay selection algorithm based on energy collection. Using energy harvesting technology to provide relay energy, the node is avoided from communication interruption or selection relay failure due to its low residual energy. Energy harvesting power and channel conditions are analyzed in the physical domain, and the appropriate relay in the alternative relay set is selected by the communication two-end node inspired by the human social campaign mechanism, so as to avoid the interruption of the entire communication caused by the relay meeting only one node requirement. The simulation results show that the algorithm proposed in this paper improves the probability of relay selection success, equipment survival time and system throughput compared with the prestige-based relay selection algorithm.

Distributed finite-time-consensus-based heavy-ball algorithm

QU Zhihai, LU Jie

2022, 39 (1): 127-133. DOI: 10.7523/j.ucas.2020.0009

Abstract ( 352 ) PDF (0KB) ( 0 )

Based on the finite-time-consensus algorithm and the heavy-ball algorithm which is a first order accelerate algorithm, a distributed optimization algorithm is proposed. The algorithm can achieve consensus after every periodic updates. The non-ergodic convergence rate is at the same order of the centralized heavy-ball algorithm. In addition, the numerical examples compare our algorithm with other state-of-art distributed optimization algorithms on machine learning problems and show the competitive performance.

Q-learning based QoS routing for high dynamic flying Ad Hoc networks

HUANG Xinchen, CHEN Guangzu, ZHENG Min, TAN Chong, LIU Hong

2022, 39 (1): 134-143. DOI: 10.7523/j.ucas.2020.0001

Abstract ( 559 ) PDF (0KB) ( 0 )

In high dynamic flying ad hoc networks (FANETs), such as UAV (unmanned aerial vehicle) ad hoc networks, the rapid change of network topology leads to the breakage of communication links and the frequent reconstruction of routes. To solve this problem, a QoS (quality of service) routing method based on Q-learning is studied. Based on the basic Q-learning framework, this method takes the number of neighbor nodes, link duration and link available bandwidth as routing metrics, and designs a Q-learning reward function to provide QoS guarantee. All nodes exchange local routing metrics information with neighbor nodes by broadcasting Hello messages and forwarding data packets. After receiving Hello packets or data packets, neighbor nodes calculate and update the Q value according to the reward function. Then one of neighbor nodes selects a next hop node to forward data packets intelligently according to the Q value table that it maintains. The simulation results in EXata simulator show that this method can provide stable and high QoS communication links for high dynamic flying ad hoc networks.

Current Issue