With the implementation of the South-North Water Diversion Project and the implementation of groundwater suppression, the local groundwater levels in the North China Plain have gradually rebounded. However, there is a lack of systematic studies of the trends of groundwater balance elements and their impacts on the ecological environment. This paper takes the Baoding Plain, a typical area in the North China Plain, as an example, and uses the water balance method to calculate the groundwater recharge and discharge terms, applies the factor analysis method to analyze the causes of the changes in groundwater balance elements from 1975 to 2019, and calculates the amount of recoverable and suppressed groundwater resources by using the optimal exploitation coefficient method, which provides a basis for the development and utilization of groundwater resources in the study area. The results show that in the past 40 years, the groundwater recharge term was smaller than the discharge term in the Baoding Plain, which is in a negative equilibrium state, and the main change elements are canal irrigation infiltration, canal system seepage, well irrigation return, river seepage, rainfall infiltration and artificial exploitation. The main factor affecting the change of groundwater equilibrium elements is human activity, with a contribution rate of 77.2%. After the groundwater recharge and discharge imbalance is slowed down, the increase of groundwater level burial depth becomes smaller, the area of groundwater level depression cone gradually decreases, and the area of the Baiyangdian wetland gradually recovers. The optimal exploitation coefficient of groundwater resources in the Baoding Plain is determined to be 0.64, the exploitable groundwater resources range from 8.89×10⁸ to 11.35×10⁸ m³/a, and the amount of compression exploitation ranges from 2.68×10⁸ to 5.14×10⁸ m³/a. The research results can provide a scientific basis for the sustainable development of groundwater resources and ecological environment in similar areas.

The landslides induced by typhoons and rainstorms in Southern China are often shallow, fluidized and clustered. A large number of investigations have shown that vegetation infiltration has a significant impact on the formation of granite residual soil landslides. However, most of the current studies have focused on the effect of vegetation roots on soil infiltration. The influence mechanism of vegetation infiltration enhancement effect on shallow landslides has not been revealed. Therefore, the mass landslide disaster on September 9, 2019 in Longchuan County in Guangdong is taken as an example, and a large number of landslide site investigations are made to identify the geological environment conditions and vegetation development in the landslide area and analyze the infiltration effect of vegetation on shallow landslides. The “double-ring method” is used to measure infiltration rate of different vegetation types, and to analyze the infiltration process and infiltration law of different vegetation types. A typical landslide profile is selected to establish a geological model, and the Geo-Studio software is used to simulate the seepage law of shallow landslide and the stress-strain characteristics of the soil under heavy rainfall conditions (250 mm/d). Finally, combined with the simulation results and field investigations, the infiltration effects of vegetation and their response mechanisms to shallow landslides under heavy rainfall are analyzed. The results show that vegetation can effectively enhance the permeability of soil, and the order of permeability is coniferous forest land, shrub forest land and bare land. Under the influence of vegetation infiltration effect, rainwater infiltration to the bottom of the root-soil composite layer will cause water stagnation, the shallow soil tends to be saturated rapidly, the pore water pressure and seepage force in the soil increase instantaneously, the soil is saturated with water, the residual soil softens, and the weight of the slope increases, which eventually leads to the overall instability of the slope. The research results explain the formation mechanism of granite residual soil landslide in high vegetation covered area under heavy rainfall in Southern China, and provide scientific basis for early warning and prediction of such disasters, which are of great significance.

Tunnels in karst areas are prone to collapse during construction. There are many analyses on the mechanism of tunnel collapse in mechanical aspects, but the mechanism of tunnel collapse in karst weak fracture zones and other strata were seldom examined. In order to ensure the safety, economy and feasibility of tunnel construction, it is necessary to master the mechanism of collapse in the tunnel construction. Relying on a tunnel project in a karst broken stratum in Guizhou, where the collapse phenomenon occurred during the excavation process, the monitoring data of the tunnel are examined, and the construction principle of the BP neural network is used to invert the stratum parameters of the tunnel. The inversion soil mechanical parameters are input into different construction models constructed by using the FLAC3D finite element software, and the collapse failure mechanism and risk of typical sections are judged and analyzed. The results show that the construction method has a great influence on the stability of the tunnel excavation, and for the tunnel with the surrounding rock grade V , the three-step seven-step method and the single-side wall pilot pit method are safer for construction, and the tunnel collapse has no relationship with the simultaneous excavation of the tunnel in both directions. The predicted value of the tunnel vault displacement obtained by the inversion is 2.3 cm, and the predicted value of the surface displacement is 1.2 cm. The deviation from the monitoring data is about 13%, and the inversion result has certain reliability. The research results are of important guiding significance for the construction of tunnels and highways in weak and broken strata in karst areas.

Tight sandstone gas is an important unconventional natural gas resource. Mineral characteristics of the tight sandstone not only affect the occurrence of gas, but also have a significant impact on the hydraulic fracturing effectiveness. The potential influence of tight sandstone mineral compositions and microscopic morphology on gas reservoir during hydraulic fracturing was seldom examined. In this study, five tight sandstone gas wells in the Sulige gasfield located in the Ordos Basin were drilled. The core samples of the 8^th formation of the lower member of the Permian Shihezi Group (He 8) were collected. The mineral characteristics, microstructure and pore feature of the tight sandstone are systematically analyzed by using XRF, XRD, casting thin section, SEM and EDS technologies and the potential effects of the above factors on hydraulic fracturing are discussed. The results show that the tight sandstone is mainly composed of lithic sandstone and lithic quartz sandstone. The density of the tight sandstone ranges from 2.44 to 2.56 g/cm³, the porosity from 7.7% to 12.6%, and the permeability from 0.16 to 1.42 mD. The mineral compositions are mainly quartz and clay minerals (16.5%−47.4%), and feldspar and carbonate are absent. Kaolinite, illite and chlorite are the main clay minerals. Kaolinite is widely developed, filling the intergranular pores and surface in the forms of “booklets” and worm-like. The intergranular pores, intragranular solution voids, intergranular and intragranular fractures occur in the tight sandstone, which provide basic reservoir space for natural gas. The results of mineral analysis and fracturing fluid-tight sandstone interactions reveal that the stability of clay minerals, especially kaolinite and illite, are crucial for hydraulic fracturing in the Sulige gasfield. It is necessary to understand the mineral compositions and formation water of the tight reservoir and select the appropriate clay stabilizer to optimize hydraulic fracturing.

When using machine learning models for landslide susceptibility evaluation, the non-landslide sample points are usually selected randomly outside the landslide influence area, leading to a certain error. To improve the accuracy of landslide susceptibility evaluation, this paper couples the self-organizing map (SOM) neural network, information (I) model, and support vector machine (SVM) model, and proposes a SOM-I-SVM model-based method of landslide susceptibility evaluation, comparing with K-means clustering to verify the reliability of this model. The Maojian District of the city of Shiyan is taken as an example, and seven factors of the distance from water system, slope, rainfall, distance from structure, relative height difference, distance from road, stratigraphic lithology are selected by correlation and importance analyses of environmental factors to establish a landslide susceptibility evaluation system. Based on these, the graded information values of each factor are calculated and used as input variables for landslide susceptibility evaluation. The SOM neural network and K-means clustering are used to select non-landslide samples, and the sample data set is substituted into the I-SVM model to predict landslide susceptibility. The prediction accuracies of the four models, SVM, I-SVM, KMeans-I-SVM and SOM-I-SVM, are compared, and the area under the ROC curve (AUC values) are 0.82, 0.88, 0.90 and 0.91, indicating that the SOM-I-SVM model can effectively improve the accuracy of landslide susceptibility prediction.

Organic pollution of polycyclic aromatic hydrocarbons (PAHs) in topsoil has threatened human health and ecological environment. In order to understand the distribution and pollution characteristics of PAHs in topsoil in the Beiyun River Basin in Beijing, an investigation is carried out to examine the contents, distribution trends, spatial distribution and pollution sources of 16 optimal PAHs in topsoil of the study area by means of multivariate statistical methods including the Kriging interpolation, principal component analysis-multiple linear regression and the concentration ratio among certain components. The results are as follows (1) all 16 PAHs are detected, and most of them are high loop PAHs (4−6 rings). The total contents of PAHs in topsoil range from 10.5 to 19466.5 μg/kg, about 29.63% of the samples are polluted in the study area. (2) The PAHs contents in topsoil show a trend of higher in the middle and lower at ends in both east-west and south-north direction. In terms of spatial distribution, the PAHs contents in topsoil are higher in the northern region and the central urban area, while the PAHs contents are lower in other areas. Due to the accumulation of PAHs caused by human activities at some points, point source pollution or local pollution exist. (3)The ratio of specific PAHs components and principal component analysis indicate that the sources of 16 PAHs in the study area are mainly coal, biomass combustion and traffic combustion. Multiple linear regression shows that the contribution rates of the two are 89% and 11% respectively. The research results can provide strong support for pollution prevention and control, land quality evaluation and territorial space planning in the study area.

The Shuitoushang landslide in Ershe in Qijiagou Village of Xinchang Town in Tongjiang County is a typical red-bed terrace landslide in eastern Sichuan. There are few studies on the seepage analyses of such landslides in this area, but similar slopes are widely distributed in the area. In this paper, on the basis of a large number of field investigations, surveys and data collection of disaster sites, the satellite remote sensing, UAV aerial photography, airborne LiDAR, indoor tests and other technical means are comprehensively used to analysis the landslide. Based on the mathematical model of Fredlund & Xing soil-water characteristic curve, the SEEP/W module in Geo-Studio is used to conduct the seepage analysis, and the analysis results at different time stages are coupled with the SLOPE/W module. Then the dynamic relationship between the stability coefficient, rainfall and time is obtained, revealing the deformation process and formation mechanism of the landslide. The results show that: (1) the continuous rainfall causes the groundwater level to rise and the slope stability to decrease, eventually leading to the overall instability and failure of the Shuitoushang landslide. (2) The deformation process of the landslide and the characteristics of groundwater exposure are basically consistent with the results calculated by Geo-Studio, indicating that the results based on Fredlund & Xing model can provide more accurate simulation results in similar areas. (3) Reclamation of terraced fields (paddy fields) will reduce slope stability. The results can provide theoretical support for hidden danger investigation and active prevention of similar landslide disasters in mountainous areas of Sichuan, and provide a reference for disaster prevention and mitigation.

As one of the important ecosystem service functions, water conservation function is of great significance to the ecosystem and water safety in the Zhangjiakou and Chengde districts. Based on the integrated valuation of ecosystem services and tradeoffs (InVEST) model, this study quantitatively evaluates the water conservation function and analyses its driving factors of the Zhang-Cheng district from 2001 to 2020 using high-quality remote sensing data and other reanalysis products, in order to improve the problems of insufficient consideration of topography and soil permeability, as well as the inadequate analyses of data spatial heterogeneity in previous studies. The results show that the spatial distribution of the water conservation function in the Zhang-Cheng district from 2001 to 2020 was characterized by high values in the Baxia areas and low values in the Bashang areas. Despite some differences, the spatial distribution of the water conservation function in each year had some similarities to some extent. In terms of the trend, the water conservation depth in these 20 years showed a fluctuating downward trend with an average rate of −0.08 mm/a. Combined with the Sen+Mann-Kendall analysis, the trend of the water conservation function in this area was mainly classified as "No trend", "Slight increase" and "Slight decrease", accounting for nearly 98% of the total area. The results of correlation analysis show that precipitation had a strongly significantly positive correlation with water conservation function, temperature had a significantly negative correlation with water conservation function in some areas, and the correlation between vegetation and water conservation function was complex. In the Zhang-Cheng district in 2020, the woodland had the strongest water conservation function, and the water conservation depth reached 28.64 mm. On the other hand, the grassland water conservation function had the greatest contribution in the total amount, and the water conservation reached 1.12×10⁹ m³. During these 20 years, the water conservation of the farmland had the most obvious decrease, with a rate of −6.49×10⁶ m³/a. The spatio-temporal change characteristics of water conservation function in the Zhang-Cheng district in these 20 years were mainly controlled by precipitation and vegetated land use. This study provides an important decision-making basis for ecological construction and water resources management in the Zhangcheng district.

Landslide disasters occur frequently in Fujian Province, and early warning of landslide disasters on a regional scale is an important means of effective disaster prevention and mitigation. Due to the complex mechanism of landslide disasters, the traditional regional landslide early warning methods have such problems as insufficient accuracy. Deep learning mainly refers to the technology of feature extraction, abstraction, representation and learning by constructing the neural network model, which is a kind of machine learning. As a classical deep learning algorithm, convolutional neural network has more powerful classification and representation ability than traditional machine learning. Taking Fujian Province as the research area, this paper introduces the convolution neural network into the field of landslide disaster early warning and constructs a regional landslide early warning model of Fujian Province. The process is as follows: (1) The SMOTE optimization algorithm is used to optimize the sample database of landslide disasters in Fujian Province from 2010 to 2018, enlarging the number of positive samples and expanding the proportion of positive and negative samples from 1∶3.4 to 1∶2, and the total number of samples reaches 18040. (2) Construct a convolution neural network model structure, which includes an input layer, two convolution layers, two maximum pooling layers, a full connection layer and an output layer. (3) Use the convolution neural network to train the optimized samples (80% of the samples from 2010 to 2018 as the training set), and use the Bayesian optimization algorithm to optimize the model parameters to obtain the regional landslide early warning model of Fujian Province. (4) The model is tested with 20% of the samples from 2010 to 2018 as the test set, and the confusion matrix and ROC curve are used to test the model. The results show that the accuracy of the model ranges from 0.96 to 0.97, the AUC value is 0.977, indicating that the model accuracy and generalization ability are good. (5) The actual situation of the landslide disaster in the flood season of 2019 is taken as a positive sample, negative samples are collected through the method of time-space sampling, and the 2019 regional landslide sample verification set (603 samples) is constructed. The model is further verified by using the confusion matrix and ROC curve. The results show that the accuracy of the model ranges from 0.75 to 0.85, and the AUC value is 0.852. Although only the actual landslide samples in the flood season of 2019 is used for verification, good results is also achieved. In this paper, the convolution neural network algorithm is applied to the regional landslide early warning, which provides a new way to establish the regional landslide early warning model. The preliminary verification shows that the model is effective and will be further applied and verified in Fujian Province in the future.

With the development of urban engineering construction, the issue of construction engineering accidents has become more and more prominent. The geotechnical parameter interval obtained by using the traditional methods cannot meet the needs of actual engineering. Based on the idea of unsupervised learning, the peaty soil with the worst engineering properties is considered, and 8 physical indexes are selected as the input set. The principal component analysis (PCA) algorithm is used to realize the dimensionality reduction of multi-sample and multi-parameter decoupling, and the correlation and sensitivity of each physical index is obtained. Combined with its correlation and sensitivity, the comprehensive evaluation value of physical indexes of peat soil with different buried depths is given. The k-means clustering is used to analyze the relationship among physical index, and comprehensive evaluation value and engineering characteristics of peaty soil provide a theoretical basis for the selection of geotechnical parameters. The supervised learning method-BP neural network algorithm is used to analyze the unsupervised results and verify the accuracy of the (PCA—k-means) algorithm model. The normal samples obtained by clustering analysis are optimized by a variety of truncation methods to obtain a reliable value range, and the value results are compared with the actual engineering values to verify the rationality of the model engineering parameters. The algorithm model is of good engineering application value. The research results can provide references for engineering investigation, design and construction parameter values, and also provide a new analysis method for geotechnical parameter value analyses.

The active characteristics and genetic mechanism of coupled ground fissures mainly induced by fault activities and pumping are studied, and the Songzhuang Town in Tongzhou District of Beijing is taken as the research archetype. The damage of surface planes and stratigraphic profiles caused by ground fissure activities are clarified through field investigation, and the variation characteristics of the displacement field and stress field of the strata caused by different dislocation amounts and groundwater level drop are revealed. The response processes of the model stratum under the two conditions of fault misalignment and groundwater extraction are simulated and studied respectively by using the finite difference method. Finally, the relationship between this type of ground fissure and the main inducing factors is discussed. The results show that (1) the ground fissure is characterized by three-dimensional activities, which causes the vertical tension of the shallow stratum and wall to be 0.3−1.2 cm, and the vertical dislocation of the deep stratum gradually increases with the burial depth. (2) Stress changes caused by fracture activities are concentrated in the ground fissure development area and lead to significant vertical displacements in the hanging wall, the stratum located in the ground fissure area has large shear and traction deformation, and the vertical displacement difference between the two sides is the largest. The gradual increase of fault dislocations causes the hidden fractures to extend upward, and cause secondary cracks on the shallow surface of the hanging wall, resulting in the overall distribution of ground fissures with a certain width. (3) The vertical extension and horizontal expansion of ground fissures are aggravated by the lowering of groundwater levels, and the surface on both sides of the crack produces continuous settlement response, making the center of the subsidence funnel become a concentrated development area of ground fissures, with the maximum settlement of 10.2 cm in the model stratum at the fissure in its central area, and the settlement range of about 38 m in the hanging wall and about 16 m in the foot wall. (4) This type of ground fissures is obviously controlled by faults, but the increased activity in this period is mainly due to groundwater over-exploitation. This work will be of great theoretical and practical significance to understand ground fissure mechanism, establish quantitative relationship between formation and fault with groundwater, and prevent and reduce disasters.

Embedded piles act as an optimization structure compared with the traditional stabilizing pile. The determination of the thrust on the loading section is based mainly on the model test and numerical simulation, and there is a lack of in-depth theoretical analysis. For the bedrock-talus landslide reinforced by embedded piles, according to the potential overtop-sliding failure mode, the slide surface can be divided into top and bottom sections by the position of pile top, and the horizontal resultant force of the top section can be obtained by integration, which is the so-called thrust of the embedded section. Similarly, the force on the bottom section of the overtop-sliding surface can also be obtained. Based on the limit equilibrium theory, the force analysis of the sliding mass enclosed by the bottom sliding surface and the load section of the pile can be carried out, and the thrust on the loaded section can also be obtained. Example analyses show that the thrust of the embedded section and the loaded section obtained by the theoretical method are very consistent with the results of FLAC3D, the resultant force of the loading section decreases nonlinearly with the increase of the ratio, while the resultant force of the embedded section presents an opposite trend. With the increase of the embedded ratio from 0 to 0.67, the thrust ratio of embedded section and loading section increases slowly from 0 to 0.3−0.5. With the increase of embedded ratio from 0.67 to 0.8, the ratio increases sharply to 1.47−2.12. Generally, the thrust of embedded section is less than that of loaded section. The theoretical research of the thrust of the embedded pile is of great practical significance for the optimization of the pile internal force and the determination of the pile embedded depth, which will promote the further application of this structure.

Knowledge of the recharge from surface water to groundwater is the basement of the scientific understanding of water cycle and the sustainable management of groundwater resources. Meanwhile, the layered heterogeneity is the main structural feature of riverbed sediments (i.e., the lithologic difference between riverbed sediments and the underlying aquifer) and the main factor that controlling the recharge from surface water to groundwater. To reveal the influence mechanism of layered structure of pore media on the recharge from surface water to groundwater, a conceptual model of surface water and groundwater interaction is established based on the field test results of Henan reaches of the Yellow River, and the process of the recharge from surface water to groundwater interaction is described using flow path as the object. The results show that the exchange flux of surface water and groundwater is mainly affected by hydraulic conductivity of riverbed sediments, and the change of the thickness of riverbed sediments has little effect on the exchange flux between surface water and groundwater. That is, the increase of the ratio of the thickness of the sediments to that of the underlying aquifer (H_S/H) from 0 to 0.125 leads to the interaction flux decreased by 72%, indicating that the existence of the low permeability layer is the main reason that decreases the interaction flux between surface water and groundwater. The change of the permeability and the thickness of riverbed sediments has obviously changed the flow path from surface water to groundwater and the travel time. Specifically, the increase in K_U/K_L leads to a lager penetration depth of groundwater flow and lager travel times. The sensitivity of exchange flux between surface water and groundwater and groundwater travel time to the hydraulic conductivity of riverbed sediments increases with the decreasing hydraulic conductivity. At the same time, the groundwater travel time is more sensitive to the change of the thickness of the low permeability riverbed sediments, and the sensitivity increases with the increasing thickness. The research results can provide reference for groundwater resource management and sustainable development.

A large number of people live in areas with soluble rocks in China. Due to the existence of caves, fissures, holes, pipes and other erosion phenomena in soluble rocks, geological disasters such as mountain landslides and collapse, ground collapse, floods, etc, are often induced in the soluble rocks. It is important to study the chemical dissolution rate of soluble rocks to predict the development of karst and to guarantee the early identification and safety evaluation of karst geological disasters. Based on a large number of literature surveys, this paper comprehensively analyzes chemical dissolution methods obtained by scholars at home and abroad.(1)Chemical dissolution methods of soluble rocks are constantly maturing and improving, and the depth of research and the scope of application are continuously expanding.(2) There are various methods to study rock dissolution, which are mainly divided into indoor dissolution tests and field dissolution tests. Among them, the standard dissolution test strip method is widely used.(3) The principles and applicability of existing methods are different. It is necessary to use multiple test methods to gain relatively reliable test results.According to the research status of this scientific issue, the next step will focus on large-scale dissolution indoor test, establish model test systems to truly simulate the dissolution process of the soluble rock and further describe the dissolution process of soluble rocks in microscopic perspective. In addition, it is necessary to comprehensively use a variety of field test methods to carry out research to avoid or reduce the large errors that may be caused by a single test method, and establish correlations between field dissolution tests and indoor corrosion tests, so as to inferring results of large time scale tests in field through short-term indoor tests.

There are abundant phreatic water in loess tablelands in the Loess Plateau of northwest China. The vertical infiltration of atmospheric precipitation is the main source of the recharge, but the spatial channels of water occurrence and migration in loess are still ambiguous. Based on the geomorphology, strata structure and other hydrogeological conditions, groundwater utilization condition and related experimental data of the Dongzhiyuan and Weibei loess tablelands, combined with scanning electron microscope image processing and statistical analysis, the pore characteristics and the migration processes of water in the loess phreatic and aeration zones are discussed and zoned. Pores are considered to be the main space channel for the occurrence and migration of loess groundwater, and the minimum equivalent pore size of seepage is about 12 μm. The vertical structural combination of the loess aeration zone composed of the Malan loess with developed pores in the upper part is conducive to the infiltration of atmospheric precipitation, and the water migration in the pores of the aeration zone is very weak and slow, but is relatively continuous and uniform. The aeration zone can be divided into four zones for loess: the climate impact zone, storage regulation zone, slow runoff zone and capillary receiving zone. Except for the upper climate impact zone, the other three zones are in a basically dynamic equilibrium state of water transport. The research results can provide important references for comprehensively and systematically understanding the characteristics of water transport in the aeration zone of loess tablelands.

Efficient estimation of near-wellbore permeability is critical for evaluating fracturing effect and updating fracturing plan. However, due to technical or cost constraints, there is still a lack of methods for in-situ testing and estimating near-wellbore permeability in deep geothermal reservoirs. Considering that the hydraulic fracturing is often associated with injection breaking and back drainage to control the reservoir pressure, this study proposes a single-well-injection-withdraw-based tracer test approach and two permeability interpretation methods based on numerical and analytical solutions, which allow in-situ permeability estimation at low cost. Implementation of the proposed method at a realistic enhanced geothermal system indicates that the numerical interpretation method can still reasonably estimate near-wellbore permeability under the condition of incomplete tracer breakthrough curve in the single well injection and withdraw test, but the computational efficiency is low. Once the tracer breakout curve is relatively complete (i.e. the peak tracer concentration is monitored), the analytical method can be used to quickly estimate the permeability. However, the analytical method cannot accurately consider the long-distance tracer migration process inside wellbore and the influence of dispersion on the tracer breakthrough curve, hence the accuracy is relatively low. The numerical and analytical permeability estimations are at the same order of magnitude. The results suggest that in addition to the numerical method, the analytical method can still be used as an effective method for in-situ rapid permeability estimation. The proposed methodology may provide a new tool for in-situ permeability estimation in deep geothermal reservoirs.

Mine inflow threats mine safety production underground, and may trigger a decline in the groundwater level in the mine area, causing irreversible successional degradation of surface vegetation. In view of the key problems such as inaccurate generalization of boundary conditions and unreliable selection of hydrogeological parameters when constructing numerical models of water inflow, this study aims to accurately predict mine water inflow, ensure the safe mining of coal seams, and provide theoretical and data support for the protection of desert vegetation in the study area. The natural boundary is selected as the perimeter of the research area, and the model is repeatedly revised on the basis of fully collecting and analyzing the data of drilling, geophysical prospecting, pumping test, groundwater long-term monitering, and the scope of the mine goaf and its water inflow, and thus a more realistic three-dimensional unstable flow numerical model of groundwater is constructed. In addition, the model simulation and identification are carried out according to the expansion process of the mine goaf and its water inflow and groundwater monitoring data, which demonstrates the rationality and reliability of the model. The established numerical model is used to predict the mine inflow and submersible level depth drop under coal seam mining conditions, and then the influence of diving level decline on desert vegetation is analyzed based on the relationship between diving depth and desert vegetation. The results show that the predicted water inflow in the mine is 3.08×10⁴ m³/d, resulting in a decrease of 2.08−2.35 m in the diving level in the mine area, which will lead to the deterioration or even partial withering of the representative vegetation sand willow and poplar in the mine area, showing a succession trend from mesophytic vegetation type to xerophytic vegetation. The results can provide more accurate water inflow prediction in the study area, scientific and effective measures for the protection of desert vegetation in mining areas, and reliable treatment ideas for the construction of similar numerical models of groundwater flow.