To investigate the tectonic evolution of Paleo--Asian Ocean, the authors select Late Paleozoic intrusive rocks exposed in Bilihe area of central Inner Mongolia as objects, and constrain their formation ages, petrogenesis, magma sources and tectonic settings through petrographic, geochronological and geochemical analyses. Zircon U--Pb dating yields ages of (263. 7 ± 2. 8) Ma for the potassium feldspar granite and (273. 8 ± 1. 2) Ma for the granitic porphyry, indicating an Early --Middle Permian emplacement. Their εHf ( t) values range from - 8. 5 to 4. 0 and - 0. 4 to 2. 3, respectively. Geochemical data reveal both rock types exhibit high silica [w ( SiO2 ) = 61. 98%--70. 65% ] , alkali-riched [w (Na2O) = 2. 78%--2. 94% ] and potassium-eriched [w (K2O) = 3. 47%-- 3. 60% ] characteristics. The A / NK ratios vary from 1. 30 to 1. 75, A / CNK ratios vary from 0. 86 to 1. 10, and Rittmann indices vary from 2. 07 to 2. 53, the granite and porphyry are belonging to high-K calc-alkaline series, being classified as metaluminous to weakly peraluminous rocks. Both rocks display enrichment in light rare earth elements (LREEs) and large-ion lithophile elements ( LILEs: Cs, Rb, Ba, Th) , depletion in heavy rare earth elements (HREEs) and high-field-strength elements (HFSEs: Nb, P, Ti), along with weak Eu negative anomalies (δEu = 0. 62--0. 67 and 0. 49--0. 54). Moderately fractionated REE patterns [ (La / Yb) N = 8. 10--8. 30 and 6. 80-- 6. 92] with REE right-inclined chondrite-normalized curves suggest characteristics of volcanic arc granites. Th / La (avg = 0. 44) , Rb / Sr ( avg = 0. 65) and Th / Nb ( avg = 1. 35) ratios align with crustal compositions, indicating derivation from partial melting of crustal materials. The potassium feldspar granite exhibits broader εHf (t) variations and higher Mg values [w (MgO) = 3. 05%--3. 33% ] , implying mantle-derived influence. Integrated results suggest these rocks formed during the subduction of the Paleo --Asian Ocean beneath the North China Block. In Early--Middle Permian, the Paleo--Asian Ocean was not closed, and the northern margin of the North China Block situated in an active continental margin setting, marked by intense regional magmatism. 

To further strengthen the research and development of Cu--Ni sulfide deposits in Jilin Province, the authors select the Shirengou gabbro in eastern Jilin Province as the study target. Through the analysis of whole-rock geochemistry, zircon Lu--Hf isotopes and platinum-group element ( PGE) of the gabbro body, the magma origin, evolution, petrogenesis and tectonic setting have been investigated. The Shirengou gabbro is primarily composed of pyroxene, amphibole, plagioclase and biotite. Major element compositions exhibit the characteristics of low silica [w( SiO2 ) = 45. 25%- 45. 75% ] , low titanium [w( TiO2 ) = 0. 98% - 1. 00% ] , high magnesium [w( MgO) = 19. 66% - 19. 90% , Mg# = 63. 52 --63. 97 ] , and low alkali contents [ w ( Na2O + K2O) = 2. 91%--2. 98% , Na2O / K2O = 2. 23 --2. 31 ]. The A / CNK ratios range from 0. 75 to 0. 76. In TAS and w ( SiO2 ) --w ( K2O) diagrams, the samples fall into the calc-alkaline field. Trace element compositions show enrichment in light rare earth elements (LREEs) and large-ion lithophile elements (LILEs, e. g. , Ba, Rb, K) , and depletion in heavy rare earth elements ( HREEs) and high-field-strength elements ( HFSEs, e. g. , Nb, Ta, Ti) , without obvious Eu anomaly. The m / f ratios vary from 1. 73 to 1. 76, indicating favorable conditions for Cu--Ni sulfide mineralization. Ratios such as La / Nb, Ce / Pb, Th / U, and Nb / U differ significantly from mantle or crustal values, suggesting crust-mantle contamination. In (Nb / La) -- (La / Yb) diagram, the samples fall into the lithospheric mantle field. Zircon 176Hf / 177Hf ratios range from 0. 282 674 to 0. 282 805, with εHf (t) values of + 0. 80 to + 5. 87, and the single-stage model ages ( TDM1 ) of 907-- 646 Ma, indicating a magma source derived from depleted mantle with crustal contamination. The Cu / Pd ratio (34. 8 × 103 ) exceeds the primitive mantle value, while the Pd / Ir ratio is 9. 07, suggesting that PGE characteristics are primarily related to magmatic processes. The primary magma originated from high--Mg basaltic melts in the lithospheric mantle. The PPGE (Pt, Pd) concentrations are significantly higher than IPGE, and the Ni / Cu ratio is 6. 31, further supporting the Shirengou gabbro as a favorable ore-forming intrusion. Combined with its Late Triassic emplacement age, the Shirengou gabbro likely formed in a post-collisional extensional setting following the closure of the Paleo--Asian Ocean. 

Dachebeigou--Fu􀆳erling area is an important gold deposit distribution area, which located in the north of Jiapigou gold deposit mining area in Jilin Province, and several gold and antimony minerals such as Xingfu and Datudingzi have been discovered, indicating a certain prospecting potential. Based on the detailed study of metallogenic geological background and typical deposit characteristics of gold deposits in Dachebeigou -- Fu􀆳erling area of Jilin Province, the authors systematically summarized the prediction factors of regional gold deposits, established a regional prediction model for gold deposits, and delineated the minimum prediction areas with prospecting potential to provide a basis for the further deployment of exploration work. The evidence weight method was used to carry out the prediction of gold ore by “ mineral resources evaluation system” ( MRAS) software, and on the basis of the contour map and color block map, the delineation of the minimum prediction area was carried out according to the principle of minimum area has maximum minerality and minimum ore leakage rate, and the predicted gold resources were estimated by using the geological volume parameter method. In this work, a total of 13 minimum prediction areas of gold deposit were delineated, including 5, 6 and 2 minimum prediction areas of class A, B and C, respectively, and the estimated predicted resources of gold deposit were 36 295. 24 kg, of which the predicted resources of gold deposit within 500 m were 21 604. 31 kg and the predicted resources of 500 m to 1 000 m were 14 690. 93 kg. It is recommended to deploy investigation and evaluation work in the minimum prediction areas of class A, such as Zhilin, Datudingzi, Erdaodianzi, Beishan, Sidaogou and Xingfutun through the comprehensive evaluation of geological condition, geophysical and geochemical anomalies, resource potential and prospecting prospects, in order to explore new mineral areas and achieve new breakthroughs in gold prospecting.

Gejiu tin mine in Yunnan is a super-large polymetallic ore concentration area dominated by tincopper, altered vein-type ore bodies within rock masses have been identified in recent years. To clarify the orecontrolling conditions and metallogenic regularities of these intra-rock altered vein-type ore bodies and to provide theoretical support for exploring new prospecting targets around Laochang ore field, the authors carry out detailed field investigation and measurement of the known ore bodies in Fengliushan ore section within the western depression zone of Laochang ore field, Gejiu mining area and three levels (1 800 m, 1 700 m and 1 600 m) of the joints and fractures (including ore veins) exposed by the internal tunnels through the comprehensive analysis of mine data. Statistical analyses were conducted on the orientation and dip angles of different joint sets, their ore-controlling and ore-hosting characteristics, and their relationship with regional fault tectonics. The results show that there are three major groups of well-developed joints and fractures in Fengliushan ore section, including NE-, NW- and nearly EW-trending, with most dip angles ranging from 60° to 80°. These features like the fracture distribution patterns in Gejiu mining area. Interactions among fractures from different orientations form favorable “X-type” tectonic structures conducive to mineralization. Among these, the nearly EW-trending fractures are the most prominent and commonly host tourmaline-quartz veins or chalcopyrite-pyrite-cassiterite-scheelite mineralization veins, maybe the radius tensional joints and fractures formed in the process of emplacement-condensation of granitic rock mass. In contrast, the NE- and NW-trending fractures primarily represent late-stage mineralization-disrupting structures that postdate and damage earlier EW-trending ore-controlling ( ore-hosting) fractures. A total of 284 ore-bearing fractures were measured, predominantly oriented nearly EW-trending, indicating that nearly EW-trending joints serve as the primary ore-controlling structures.

On the basis of field investigation and profiling, the authors divide the sedimentary phases of Middle Jurassic Xishanyao Formation in southern Changji by means of petrographic and petrological geochemical studies, and further explore provenance and basin-mountain coupling relationships. The results show that the main lithologies of the Xishanyao Formation in the study area are reddish-brown, greyish-green, greyish-white mudstones, siltstones, fine sandstones, and coarse sandstones, with local interbedded coal lines / seams and charcoal shales. The Xishanyao Formation in the study area is characterized by a subphase of deltaic plains and a subphase of deltaic frontal margins, and can be further divided into microphases such as branch channels, marshes and submerged branch channels. The obvious lateral variation of subphase of deltaic plains, and the discontinuity of the marshes in the space caused the poor comparability of seam and the thickness of the coal-forming. The average chemical index of alteration (CIA) of the mudstone samples is 79. 89, which indicates that the provenance area underwent moderate to strong chemical weathering. The average index of compositional variation ( ICV) of the mudstone samples is 0. 72, reflecting that the sediments underwent re-circling or experienced strong chemical weathering at the time of initial deposition, forming at a relatively stable tectonic environment. The granite and rhyolite clasts and the A-- CN--K diagram indicate that the source rocks were mainly derived from felsic igneous rocks in the upper crust. The occurrence of andesite and basalt clasts and the characteristics of La / Sc ratio indicate that there are minor additions of intermediate-basic source rocks. The La--Th--Sc, Th--Sc--Zr / 10, and Th--Co--Zr / 10 discriminant diagrams indicate that the tectonic background of the provenance area is in the continental island arc environment, and the source of detrital material is mainly from the Middle and South Tianshan Mountains, and there is no detrital material from the Bogda Mountains, which indicates that North Tianshan Mountains have not yet been uplifted at that time, and the height difference between the basin and mountain environments was small, with a large basin extent.

 

A typical example of three-dimensional composite hydrocarbon accumulation exists in southwest of Bozhong Sag and obvious seismic wave velocity anomalies have been revealed by drilling in the lower member of Minghuazhen Formation. In order to clarify the unclear distributions and causes of these seismic wave velocity anomalies in the lower member of Minghuazhen Formation in southwest of Bozhong Sag, the authors analyzed the relationships between seismic wave velocity with faults, lithology and structure in different oil-bearing units based on investigating the planar distribution of the velocity anomalies. It was found that the influencing factors of velocity anomalies vary among different oil-bearing units, and for the first time, it was discovered that structure has a regular impact on velocity anomalies. The average velocity anomalies are obtained by decompaction based on the average velocity field, which was derived from the stacking velocity field. Detailed analyses were conducted on the factors that cause velocity anomalies, such as faults and sand-to-strata ratios, to explore the geological causes of velocity anomalies in the lower member of Minghuazhen Formation. The results show that the main factors affecting the velocity anomalies are faults, structural amplitude, and sand-to-strata ratio in order of influence. However, these factors vary under different geological conditions and need to be considered comprehensively. In the area where faults are well-developed, the fault zones significantly increase the velocity anomalies, larger anomalies are closer to the faults. In the area without faults, structural amplitude is the main factor affecting velocity, higher velocities are corresponding to greater amplitudes. While, lithology is the main factor affecting velocity in low amplitude structure, higher content of sandstone results lower velocities.

 

Gravity and magnetotelluric sounding are commonly used techniques in geophysical exploration, with wide applications in mineral and geothermal resource exploration. However, a single geophysical inversion is often affected by the limitation of observational data and the ambiguity of inversion, making it difficult to provide detailed and comprehensive geological interpretations of actual underground anomalous structures. The gravity method can distinguish geological structures well in the horizontal direction, but it is difficult to distinguish deep or shallow anomalies in the vertical direction. Magnetotelluric sounding has a depth range of tens to hundreds of kilometers and strong deep resolution in the vertical direction, but its resolution in lateral geological structural changes is weak. In order to overcome the limitations of the two methods and combine their respective inversion advantages, the authors propose a new method for two-dimensional spatial coupled constrained inversion of magnetotellurics and gravity based on semantic segmentation ( spatial constrained inversion). This algorithm constrains the interrelationships between different physical property values by recognizing and fusing the features of the target area, thereby improving the inversion accuracy. Compared with traditional methods, this algorithm only performs inversion on the target area, keeping the non-target area unchanged, effectively reducing the dimensionality of the inversion solution, reducing the multiplicity of inversion solutions, and reducing computational complexity. The model test results show that the proposed algorithm can significantly improve the resolution of inversion and clearly reveal abnormal boundary information, providing more accurate results for geological structure analysis. Finally, this method successfully identified strong mineralization targets in Yongxin gold mining area of Duobaoshan, Heilongjiang Province, and finely delineated the boundaries of underground anomalous bodies, providing strong technical support for subsequent drilling and positioning.

In order to explore the electromagnetic wave detection response characteristics of shallow underground buried targets in a specific medium and meet the urgent needs of urban underground pipeline management and unknown target detection, a high-performance antenna for ground penetrating radar system was innovatively designed based on the in-depth analysis of the rectangular waveguide theory. Firstly, according to the propagation characteristics of electromagnetic waves in the rectangular waveguide, the basic model of the antenna was constructed, the cut-off frequency of the antenna was calculated, and the monopole probe was selected as the effective excitation source. Secondly, the high frequency structure simulator (HFSS) of Ansys electronics desktop (AEDT) simulation platform was used to carry out numerical simulation and optimization of the antenna model. By in-depth analysis of the influence of the installation position and insertion depth of the monopole probe in the rectangular waveguide on the transmission characteristics of the antenna, the S11 reflection coefficient, voltage standing wave ratio (VSWR) and radiation pattern of the antenna were accurately adjusted. Finally, in order to verify the accuracy of the theoretical analysis and simulation results, the antenna was fabricated with aluminum foil board material, and the ground penetrating radar system was built with vector network analyzer (VNA) in a controlled sand trough environment for field testing, and the detection height of the antenna was adjusted to cover underground targets at different depths. Experimental results show that the antenna can stably and accurately identify all preset targets. The ground penetrating radar antenna based on rectangular waveguide shows excellent directionality, ultra-wideband and high resolution after finely adjusting the insertion depth of the probe and the distance between the probe and the short-circuit surface of the rectangular waveguide to achieve the best matching of antenna impedance, which provides an efficient and accurate new tool for urban underground space management and unknown target detection.

In order to improve the accuracy of extracting erosion gully information based on remote sensing data, the erosion gully extraction model based on multi-source data fusion and object-based image analysis (OBIA) is proposed to extract the distribution information of erosion gully in Dehui City, Jilin Province, with a view to play an early warning role for the generation of erosion gully. The study fuses Sentinel--1 radar data, Sentinel--2 optical remote sensing images, and topographic data of the study area, and uses the multiresolution segmentation method to segment the features in the study area. Based on the results of multiresolution segmentation, intensity, texture, spectral, index, topographic and geometric features are extracted for constructing multi-dimensional sets, and the contribution of features is evaluated and optimized using the random forest Gini index to obtain the optimal set of features. The random forest extraction model based on OBIA was constructed to extract erosion gully information in the study area, and the extraction accuracy was evaluated using the method of generating random points and visually deciphering them. To further investigate the effect of multi-source data fusion on the accuracy of erosion gully extraction, based on the same sample dataset, the erosion gully in the study area was extracted for different data sources and its accuracy was compared. The results show that the overall accuracy of erosion gully extraction based on multi-source data fusion is 94. 03% , and the Kappa coefficient is 0. 839, indicating that the model has a good performance. Compared with a single data source, multi-source data fusion has higher extraction accuracy, and provides a large number of effective feature values for the random forest model. Erosion gullies in Dehui City are mainly distributed in the central and southwestern part of the city, and the erosion as a whole shows a northeastsouthwest belt direction, mostly distributed around the settlement and connected with certain water systems, and its erosion situation needs to be emphasized. 

Aming to the unique temperate forest vegetation within the Northeast Tiger and Leopard National Park, a study on multi-source remote sensing image coverage classification has been conducted, which aims to utilize modern remote sensing techniques and advanced machine learning methods to perform fine classification of complex vegetation types, thereby better understanding the ecological conditions of the region and providing scientific basis for the conservation efforts in the Northeast Tiger and Leopard National Park. Firstly, the authors employed multi-scale image segmentation technology to decompose the original remote sensing images into a series of small regions with similar texture and color features, which helps reduce data redundancy and highlight different characteristics of vegetation. Subsequently, these segmented images have been optimized by extracting feature parameters that effectively distinguish different types of vegetation. Then, the authors input these optimized feature parameters into three different machine learning classifiers for training and testing, including support vector machine ( SVM) , C5. 0 decision tree and random forest. The experimental results show that, in terms of classification accuracy, random forest achieving the highest accuracy of 91. 33% , and C5. 0 decision tree reaching 89. 59% , with the Kappa coefficients of 0. 90 and 0. 88, respectively. In comparison, the classification accuracy of support vector machine was lower, at only 58. 96% , and the Kappa coefficient is 0. 54. These results indicate that, for distinguishing vegetation cover types and identifying other land features, the random forest algorithm has significant advantages. It not only accurately distinguishes different vegetation types in complex backgrounds, but also effectively identifies other features such as roads and buildings. 

In order to study the importance of the spatial and temporal distribution characteristics of sea ice in the marine disaster warning, resource development and global climate change research, the authors use multi-tem- poral remote sensing data to extract and analyze the spatial distribution range and sea ice thickness information in Liaodong Bay of Bohai Sea during winter months based on the remote sensing inversion modeling of sea ice thickness. First, the sensitive wavebands that have a significant response to differences in sea ice thickness were screened based on the variance analysis method, and a sea ice thickness-sensitive band reflectance statistical correlation model was constructed based on the regression analysis method. The multi-temporal remote sensing data were used further to reflect the sea ice thickness in Bohai Sea and to analyze the spatial and temporal distribution characteristics of sea ice. Experimental results show that the red-light band data of Landsat satellites have high sensitivity in retrieving sea ice thickness, and the regression statistical model established on the red-light data has a linear fitting regression sum of squares as high as 0. 975 1, which indicates that the model has high accuracy and reliability. The result of Landsat red-light band data used for retrieval on the sea ice thickness and spatiotemporal distribution of sea ice in Liaodong Bay of Bohai Sea indicates that the sea ice begins in late December every year at Panjin and Yingkou in Bohai Sea, and reaches the thickest and largest distribution area by January. As the temperature gradually rises, sea ice gradually recedes in early February, and completely disappears by early March. In addition, the proportion of thin ice is generally high in each winter month, while that of thick ice is relatively low. The sea ice in Bohai Sea freezes earlier and thicker in higher latitude areas or closer to the coast, and retreats later. On the contrary, the ice in the low latitude areas freezes later and retreats earlier. 

Electrical resistivity tomography ( ERT) method is a geophysical exploration technique based on the differences in resistivity of media. Due to its characteristics of non-destructive and high efficiency, it is widely used for detecting and evaluating subsurface anomalies. The authors take a gold mine goaf as the subject, numerical simulations for analyzing potential anomalous characteristics in goaf areas, and field investigations to verify the applicability of ERT in detecting goaf collapse zones, which providing a scientific basis for goaf management. Focusing on Wenner, dipole-dipole, Wenner--Schlumberger, and pole-pole electrode arrays, RES2DMOD forward modeling software was used to construct models of a single isolated low-resistivity anomaly and a semi-high and semi-low resistivity anomaly to simulate fully water-filled and partially water-filled conditions of goafs. The data were processed by using RES2DINV inversion software with the least-squares method for inversion calculations. The results show that the dipole-dipole array exhibited the best performance for the fully water-filled anomaly, which accurately locate and describe the geometric shape and resistivity characteristics of low-resistivity anomalies with high resolution. For the partially water-filled semi-high resistivity and semi-low resistivity anomaly, the inversion accuracy of all arrays decreased, but the dipole-dipole array still outperformed the others, effectively reflected the distribution characteristics of the anomalies. In field investigations, the dipole-dipole array was employed for data acquisition. The processed data revealed a low-resistivity zone at the position approximately 115 m (30 m in depth) in the study area. Based on geological and hydrogeological conditions, the low-resistivity zone was inferred to be a goaf collapse zone. The findings indicate that the dipole-dipole array effectively reflects the extent and characteristics of goaf collapse zones, providing valuable insights for estimating the approximate range and developmental depth of anomalous areas.

As a complex rock mechanics medium, the plastic deformation of soft rock brings rock mechanics problems to hydropower engineering in the large and medium-sized hydropower projects, the creep behavior in water has a great influence on the stability of rock mass. In order to probe into the creep law of soft rock under waterbearing condition, the authors take argillaceous siltstone under different water-bearing states in a power station area in Haozhou as the research object. The effect of moisture content on the creep rate and creep stage of argillaceous siltstone is analyzed through tri-axial compression creep tests under three different water content states: dry, natural and saturated. According to the characteristics of steady-state creep and accelerated creep test curves of rock, the Nishihara creep model is used to construct the optimal creep model of argillaceous siltstone considering the effect of moisture content based on the accelerated creep test curve􀆳s characteristics. The result shows that the existence of water affects the creep deformation of argillaceous siltstone. With the increase of water content, the instantaneous strain and creep strain of rock show a continuous increasing trend. Under the same stress level, the higher the water content is, the lower the long-term strength of argillaceous siltstone is, and the longer it takes to enter the steadystate creep stage, the larger the creep deformation is. When the axial stress does not reach the long-term strength of the siltstone, the argillaceous siltstone of the rock will produce transient strain at first and then be in the steadystate creep stage. As the axial stress exceeds the long-term strength of the siltstone, the accelerated creep stage occurs in the rock. When the nonlinear viscoplastic element is introduced, and the traditional Nishihara creep model is connected with it in series, the nonlinear viscoelastic-plastic optimized creep model can describe the accelerated creep stage of soft rock more accurately, and have better applicability.