Method for improving spatial resolution of groundwater reserves based on weighting and fusion of subsidence characteristics

Abstract

The present application belongs to the technical field of hydrology, and particularly relates to a method for improving spatial resolution of groundwater storage based on settlement characteristic weighted fusion. The technical scheme of the present application is based on extraction of settlement characteristics and derivation of weight algorithm, and the new settlement characteristic weighted fusion method fully considers the influence of two data characteristics of surface settlement and water level change on the change of groundwater storage. Through weighted fusion of the groundwater storage anomaly inverted by GRACE and InSAR data, a brand-new fusion image is obtained. The fusion result is more sensitive to the surface settlement characteristics, can better show the detail characteristics of water storage change, and can more accurately provide information services for the utilization of water resources.

Description

Technical Field

[0001] This invention belongs to the field of hydrology technology, specifically a method for improving the spatial resolution of groundwater storage based on weighted fusion of settlement characteristics. Background Technology

[0002] In many arid and semi-arid regions of the world, groundwater extraction exceeds natural recharge. Over-reliance on non-renewable groundwater resources leads to a significant drop in groundwater levels, causing a decrease in pore water levels in aquifers and weakly permeable layers. This stress transfer causes compaction and deformation of cohesive soils, resulting in land subsidence. Land subsidence can damage infrastructure such as buildings, highways, and bridges, and reduce flood discharge efficiency. Furthermore, inelastic settlement caused by aquifer system compaction can permanently damage aquifers' water storage capacity and affect the sustainable use of groundwater. Implementing effective groundwater management measures is crucial to addressing existing water supply and demand issues.

[0003] Currently, traditional groundwater monitoring methods include groundwater level monitoring wells and the groundwater balance method. However, these methods suffer from sparse monitoring points, high costs, time and labor consumption, and difficulty in maintaining good quality control. Therefore, well observations struggle to capture the spatial details of aquifer and groundwater changes, and also fail to achieve continuous real-time monitoring. The groundwater balance method is also a commonly used estimation method by management agencies, but it ignores the dynamic changes in aquifers and local water circulation that occur with changes in groundwater reserves. In recent years, remote sensing monitoring has shown great application potential in the field of hydrology and water resources. Remote sensing monitoring mainly uses optical or radar imagery to study groundwater distribution, and remote sensing models of groundwater level distribution have been established using soil moisture content. Luo obtained a map of groundwater depth distribution in the study area by analyzing the correlation between relative soil moisture content at different depths and groundwater depth.

[0004] Interferometric Synthetic Aperture Radar (InSAR) is increasingly used to monitor surface subsidence and infer hydrogeological properties due to its high spatial resolution for measuring ground deformation. Studies have shown that using InSAR technology to acquire time-series deformation information can estimate the elastic and inelastic storage coefficients of aquifers and monitor groundwater level changes. InSAR offers advantages such as high precision, high resolution, wide coverage, all-weather operation, and highly automated data processing, but it also has certain limitations in practical applications. This is because aquifers respond unevenly to pressure changes, and compaction thickness varies spatially. Even if pressure changes affect the structure of compressible aquifers, not all surface deformation is caused by aquifer depletion, especially in areas exceeding typical InSAR detection thresholds. In recent years, the Permanent scatterer (PS-InSAR) and small baseline set (SBAS-InSAR) technologies, developed based on differential interferometry, have been successfully applied to urban land subsidence monitoring. Their phase and amplitude remain relatively stable over extended periods, effectively reducing the impact of spatiotemporal decoherence, atmospheric delay, and noise, thus improving deformation monitoring accuracy. The later additions of ScanSAR and TOPSAR further facilitate the provision of large-area SAR imagery, enabling the mapping of surface deformation from urban to watershed / national scales.

[0005] Compared to InSAR methods, the GRACE gravity satellite can directly determine groundwater storage anomalies (GWSA) by incorporating hydrological models, assuming that other components (e.g., soil water, surface water, and snow water equivalent) have been subtracted from the total terrestrial water storage. However, GRACE has a relatively low spatial resolution (0.25°–3°) and is suitable for areas of 100,000 km². 2 In the aforementioned regions, typical GRACE resolution is insufficient to support groundwater remediation at small-scale scales. Currently, many new methods exist for downscaling GRACE estimates of water storage anomalies. Improved GRACE inversion algorithms demonstrate that considering the spatial variability of crustal mass at GRACE resolution is crucial for better estimating groundwater storage changes. Land subsidence rates and amounts are closely related to groundwater level decline rates and amounts; therefore, GRACE data can be downscaled using surface monitoring data. Castellazzi's research found that focusing the GWSA signal retrieved from GRACE using InSAR data as a spatial prior map of groundwater mass loss distribution can provide quantitative, high-resolution information on groundwater depletion. However, previous studies have only been conceptual introductions and have not yet proposed specific and effective fusion algorithms or case studies.

[0006] Unlike existing studies, GRACE, as an effective means of detecting changes in water storage, has limited application potential in small areas due to its coarse resolution, and it also exhibits inconsistencies with InSAR and official groundwater level data. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention proposes a method for improving the spatial resolution of groundwater storage based on weighted fusion of settlement features. By weighted fusion of groundwater storage anomalies retrieved from GRACE and InSAR data, a novel fused image is obtained, which better represents the detailed features of water storage changes.

[0008] This invention is based on a method to improve the spatial resolution of groundwater storage using a weighted fusion method of settlement features. It performs spatial downscaling of GRACE using this method, including:

[0009] S1, collect data from regional groundwater level monitoring wells, and calculate the time series variation trend of individual wells and the spatial distribution of the rate of change of the entire region;

[0010] S2, download Sentinel-1A imagery, high-precision DEM and POD precise orbit data, and use multi-temporal InSAR method to detect the spatiotemporal evolution of regional surface deformation;

[0011] S3: Obtain the JPL Mascon and CSR Mascon datasets, and extract terrestrial water storage anomalies from these two types of GRACE time-series data.

[0012] S4, combined with InSAR technology, detects the distribution of surface deformation in groundwater-depleted areas and groundwater storage anomalies retrieved from GRACE data.

[0013] Preferably, in step S2, the InSAR processing procedure includes determining the primary and secondary images for differential interferometry, terrain phase removal, phase unwrapping, phase filtering, and geocoding.

[0014] Preferably, the primary and secondary images are processed using 10×2 multi-view processing to suppress noise; after generating phase interferograms and amplitude intensity maps, terrain phases are removed from the interferometric phases using a DEM; buildings with high coherence and relatively stable scattering characteristics are selected as scattering points, and phase unwrapping is performed using a three-dimensional phase unwrapping algorithm.

[0015] Preferably, the Goldstein branching method is used for phase filtering, and the linear phase ramp of each interferogram is subtracted to eliminate orbital errors and atmospheric phase interference; the line-of-sight deformation time series of each pixel is obtained by least squares inversion, and the results are geocoded.

[0016] Preferably, in step S3, when GRACE inverts groundwater storage anomalies, it is necessary to remove the equivalents of soil water, surface water, vegetation water, and snow water.

[0017] Preferably, at the same resolution, the groundwater storage anomaly distribution map is obtained by subtracting soil water and surface water from terrestrial water storage, expressed by the following formula:

[0018]

[0019] Among them, GWSA indicates groundwater storage anomaly, TWSA indicates total terrestrial water storage anomaly, SMSA indicates soil water storage anomaly, and SWSA indicates surface water storage anomaly.

[0020] Preferably, the changes in terrestrial water storage and groundwater storage retrieved by GRACE, as well as the changes in soil water and surface water retrieved by NOAH, require interpolation and anomaly processing.

[0021] Preferably, in step S4, the GRACE and weights are combined into a new value based on the novel GRACE / InSAR settlement feature weighted fusion method, thereby improving the resolution of the GRACE signal.

[0022] Preferably, in step S1, more than 24 consecutive months are selected as time reference intervals to obtain GPS time series provided by CMONOC for seasonal variation analysis.

[0023] This invention proposes a novel weighted fusion method for settlement features. This method is based on the extraction of settlement features and the derivation of weighting algorithms. The novel weighted fusion method for settlement features fully considers the impact of two data features, surface subsidence and water level change, on changes in groundwater storage.

[0024] By weighted fusion of groundwater storage anomalies retrieved from GRACE and InSAR data, a novel fused image is obtained. The fused result is more sensitive to surface subsidence characteristics and can better represent the detailed features of water storage changes. This improves the spatial resolution of groundwater storage anomalies by ten times (from 0.5°×0.5° to 0.05°×0.05°), enabling more precise information services for water resource utilization. The technical solution of this invention is suitable for widespread application in related technical fields. Attached Figure Description

[0025] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0026] Figure 1 This is a geographical overview map of the Beijing area;

[0027] Figure 2 This is a flowchart illustrating the method of the present invention;

[0028] Figure 3 This is a schematic diagram of the rate of change of groundwater level in Beijing: (a) the rate of change of groundwater level from 2005 to 2011, (b) the rate of change of groundwater level from 2012 to 2014, and (c) the rate of change of groundwater level from 2015 to 2018.

[0029] Figure 4 This is a time series diagram of the average groundwater level in the Beijing plain from 2005 to 2018, and a schematic diagram of the rate of change of water level in different periods.

[0030] Figure 5 This is a schematic diagram of the groundwater levels of 10 monitoring wells showing a downward trend from 2015 to 2018.

[0031] Figure 6 This is a schematic diagram of the deformation of the Beijing Plain from May 2017 to July 2020, retrieved by InSAR. (a) Vector scatter plot, where A, B, C, and D are four settlement funnels. (b) 0.05°×0.05° raster plot.

[0032] Figure 7 This is a schematic diagram of the cumulative settlement and average settlement rate of the four settlement funnels, (a) cumulative settlement, (b) average settlement rate;

[0033] Figure 8 This is a schematic diagram of the settlement grid extracted based on InSAR results;

[0034] Figure 9 This is a schematic diagram comparing the time series of InSAR, GRACE, CLSM and ΔGWL from 2017 to 2020. (a)-(d) are the original time series comparison results of monitoring wells P17, P66, P68 and P86, respectively. (e)-(h) are the correlations between the InSAR subsidence trend term and the CLSM and ΔGWL trend terms of the four monitoring wells, respectively.

[0035] Figure 10 This is a schematic diagram of the annual trends of each component when GRACE / GLDAS inverts groundwater storage anomalies from 2017 to 2020. (a) 0.5° JPL-TWSA, (b) 0.25° CSR-TWSA, (c) 0.25° NOAH-SMSA, (d) 0.25° NOAH-SWSA, and (e)-(h) are the results after interpolating (a)-(d) to 0.05°.

[0036] Figure 11This is a schematic diagram of GWSA obtained based on a novel weighted fusion method of settlement characteristics.

[0037] Among them, (a) JPL-GWSA, (b) CSR-GWSA, (c) CLSM-GWSA, (d) JPL / InSAR fused inversion GWSA, (e) CSR / InSAR fused inversion GWSA, and (f) CLSM / InSAR fused inversion GWSA;

[0038] Figure 12 This is a schematic diagram illustrating the error assessment of the GWSA trend between the fused inverted GWSA and the estimated groundwater level;

[0039] Figure 13 This is a schematic diagram showing the relationship between GWSA, subsidence, and precipitation. (a) Comparison of regional average values ​​of GWSA and subsidence in the Beijing Plain; (b) Comparison of average InSAR subsidence in the northern Beijing Plain and GPS subsidence at station BJSH; (c) Comparison of seasonal components after decomposition of GWSA, subsidence, and precipitation.

[0040] Figure 14 This is a schematic diagram comparing the temporal subsidence of a single grid near the BJSH site with the average subsidence in the northern part of the plain;

[0041] Figure 15 This is a Google Earth map showing the terrain features at the subsidence funnel. Detailed Implementation

[0042] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0043] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0044] This invention takes the Beijing area as an example, such as... Figure 1As shown, Beijing is located at the northwestern end of the North China Plain, with geographical coordinates of 115°30' E~117°30' E and 39°00' N~41°00' N. The Beijing Plain comprises 13 districts, including the eight urban districts and parts of Miyun, Huairou, Shunyi, Tongzhou, and Pinggu districts. Beijing has a typical warm-temperate semi-humid continental monsoon climate, characterized by cold, dry winters and hot, rainy summers. Precipitation is highly unevenly distributed in time and space, with mountainous areas receiving more rainfall than the plains. Most of the annual precipitation is concentrated between June and September, accounting for 60%–85% of the total annual rainfall. Furthermore, there is significant interannual variation in precipitation, with frequent periods of consecutive droughts or abundant rainfall. The contradiction between water supply and demand is prominent, leading to continuous over-extraction of groundwater to ensure water supply. The average groundwater depth in the plain area decreased from 11.9 m at the end of 1998 to 24.3 m at the end of 2012.

[0045] Beijing has a complex topography, sloping from northwest to southeast. The urban area of ​​Beijing is situated on an alluvial plain sloping southeastward, formed by rivers such as the Yongding River and the Wenyu River. As an alluvial plain, the compressible material of the aquifer forms the overlying strata, with the thickness increasing from west (tens of meters) to east (hundreds of meters). Areas with larger subsidence gradients are located at the boundaries of Quaternary depressions. Based on Quaternary sedimentary patterns and ages, aquifer structure, and current groundwater extraction and utilization, the aquifers in the Beijing area can be divided into four groups. The first aquifer group mainly consists of unconfined and slightly confined water, with a depth of less than 50 m; the second aquifer is a confined zone with a depth of 80-120 m; the third aquifer is a confined zone with a depth of 150-180 m; and the fourth aquifer is a confined zone with a depth of 300 m. Among them, the lithology of the unconfined zone is mainly silt, while the lithology of the confined zone is silt, clay, fine sand, and coarse sand. The primary extraction layer for industrial and domestic water is the third confined aquifer, which is widely distributed throughout the plains. Worryingly, over-extraction of groundwater, coupled with the presence of compressible aquifers, has caused extremely severe subsidence. The second and third weakly permeable aquifers exhibit elastoplastic and creep characteristics with changes in the water level of the medium-deep confined aquifers, and are the main contributing layers to land subsidence in the Beijing area.

[0046] From the 1950s to the early 1970s, Beijing's groundwater development was in its initial stage, with extraction and replenishment basically balanced. Over-extraction occurred only in localized areas near the city's suburbs, and land subsidence also occurred only in these over-extraction areas. The increase in groundwater extraction in the suburbs was particularly significant between 1961 and 1970. From the mid-1970s to the early 1980s, the level of groundwater extraction in Beijing increased dramatically, with extraction volumes doubling those of the early 1960s. By 1978, severe over-extraction of groundwater had occurred in the suburbs, with groundwater reserves decreasing by 12.42 × 10⁻⁶ over the decade. 8 m 3Groundwater extraction in suburban districts and counties also increased significantly, but did not reach the level of serious over-extraction. From the early 1980s to the late 1990s, groundwater extraction in Beijing was relatively stable, and the groundwater level was slowly declining. Over-extraction areas gradually expanded from the near suburbs to the far suburbs and counties, accounting for 70% of the plain area.

[0047] However, the compression of groundwater extraction layers is the main cause of land subsidence, with the subsidence cone gradually expanding from the eastern suburbs towards areas such as Shunyi and Changping. In the first decade of the 21st century, the groundwater cone was mainly distributed from Huanggang and Changdian in Chaoyang District to Migezhuang in Shunyi District; the areas of severe land subsidence largely coincided with the distribution of the groundwater cone. By 2010, the area of ​​accumulated land subsidence greater than 50 mm in the Beijing plain reached 4288 km². 2 It is expanding at a rate of 30–60 mm / yr, seriously threatening urban safety. In the face of this problem, detailed monitoring of groundwater over a large area can help people understand water resource conditions and rationally exploit groundwater.

[0048] like Figure 2 As shown, this invention is a method for improving the spatial resolution of groundwater storage based on settlement feature weighted fusion. It performs spatial downscaling of GRACE using a settlement feature weighted fusion method, including:

[0049] S1, collect data from regional groundwater level monitoring wells, and calculate the time series variation trend of individual wells and the spatial distribution of the rate of change of the entire region;

[0050] S2, download Sentinel-1A imagery, high-precision DEM and POD precise orbit data, and use multi-temporal InSAR method to detect the spatiotemporal evolution of regional surface deformation;

[0051] S3: Obtain the JPL Mascon and CSR Mascon datasets, and extract terrestrial water storage anomalies from these two types of GRACE time-series data.

[0052] S4, combined with InSAR technology, detects the distribution of surface deformation in groundwater-depleted areas and groundwater storage anomalies retrieved from GRACE data.

[0053] Based on the Max-Min normalization, a statistical normalization algorithm is derived. This algorithm is used to apply deformation results as weights for spatial prior knowledge. A novel weighted fusion method based on GRACE / InSAR settlement features is then used to combine GRACE and the weights into a new value, thereby improving the resolution of the GRACE signal. Considering data quality and availability, as well as the actual conditions of the study area, the period from 2005 to 2018 was selected as the study period. Finally, the simulation accuracy is evaluated by comparing the absolute error index of the downscaling results and the measured values. The data processing flow is as follows: Figure 3 As shown.

[0054] Figure 3 and Figure 4 The spatial distribution map of the rate of change of groundwater level in the Beijing Plain and the time series curve of the regional average burial depth are shown respectively. Figure 3 (a)-(c) represent the rates of groundwater level change for the three periods: 2005-2011, 2012-2014, and 2015-2018, respectively. Figure 3 It can be seen that due to continuous over-extraction of groundwater, the groundwater level mostly showed a downward trend from 2005 to 2011; from 2012 to 2014, due to increased rainfall, the groundwater level gradually stabilized; from Figure 4 This shift can also be observed in the regional average burial depth. Linear fitting revealed that the average burial depth trends for the three periods were -0.91 m / yr, 0.16 m / yr, and 0.03 m / yr, respectively. However, some monitoring wells still show a continuously decreasing trend; a total of 22 such monitoring wells were identified. Ten of these wells exhibiting a decreasing trend were selected to illustrate the temporal changes in groundwater level during the third period, such as... Figure 5 As shown.

[0055] Figure 6 (a) shows the cumulative deformation of the Beijing Plain from 2017 to 2020. For subsequent fusion and inversion, the point map was converted into a raster map with a resolution of 0.05°, as shown below. Figure 6 As shown in (b), most areas are relatively stable with no significant subsidence. Spatially, with Yuyuantan Park as a reference point, the subsidence areas are mainly located at the junction of Chaoyang District and Tongzhou District, and at the junction of Changping District and Haidian District, forming four subsidence funnels. Temporally, using the main image as a reference value (May 20, 2017), the cumulative deformation range at the four funnels is -183.70 mm to -92.63 mm. Figure 7 (a)), the maximum annual average deformation rate reached -53.09 mm / yr ( Figure 7(b) This aligns with existing research on the distribution of land subsidence in Beijing. Previous studies mostly used ENVISAT data to extract deformation in Beijing during the 2007-2010 period, during which the deformation rate reached 150 mm / yr. Land subsidence in Beijing is primarily caused by over-extraction of groundwater. However, thanks to groundwater recharge and stricter groundwater extraction policies in recent years, land subsidence in most urban areas has gradually eased from 2017 to 2020.

[0056] Figure 8 The data presents 258 subsidence grids estimated by InSAR, with blank areas representing uplift zones. Four wells (P17, P66, P68, and P86) with relatively good data quality were selected from the subsidence zones for analysis of the relationship between groundwater level and surface subsidence. All four wells are located within the subsidence funnel area. Figure 9 (a)-(d) compare groundwater level changes, JPL-GWSA, CSR-GWSA, CLSM-GWSA, and InSAR subsidence at the four selected stations, showing relatively consistent trends. The GWSA values ​​retrieved from JPL and CSR are annual average changes, with a time series ranging from June 2018 to December 2020. The comparison between ΔGWL and InSAR subsidence shows that surface subsidence lags behind groundwater level changes by approximately 1-6 months. To reasonably analyze the correlation between the two, ΔGWL needs preprocessing. Using May 2017 as a baseline, the baseline value is subtracted from the original data, and the trend term is decomposed using the STL method before comparison. Figure 9 (e)-(h) show the correlation between InSAR subsidence at the four monitoring wells and CLSM-GWSA and ΔGWL, respectively, with correlation coefficients all greater than 0.85. During 2017–2020, the subsidence at the four monitoring wells showed good consistency with the groundwater level change trends. As the groundwater level decreased, the rate of change in cumulative subsidence increased; as the groundwater level rose, the corresponding rate of change in subsidence decreased.

[0057] Figure 10This paper presents annual trend plots of various variables for GRACE inversion of GWSA from 2017 to 2020, including JPL-TWSA, CSR-TWSA, NOAH-SMSA, and NOAH-SWSA. Figures 10(a)-(d) show the spatial trend plots of the four variables at their original resolutions; except for JPL (0.5°×0.5°), the other three variables are at 0.25°×0.25°. Figures 10(e)-(h) show the trend plots of the four variables after interpolation to 0.05°×0.05° using cubic spline interpolation. The overall spatial distribution before and after interpolation is basically consistent, so this invention does not consider the error caused by interpolation. Figures 10(a) and 10(b) show inconsistencies in the spatial distribution of terrestrial water storage anomalies estimated by the two GRACE data sets. This may be due to the different destriping strategies used in the two solutions, resulting in different effects of the striping residuals. GRACE spherical harmonic coefficient truncation, destriping, and Gaussian smoothing filtering all contribute to leakage errors in the GRACE inversion results. The most suitable GRACE solution for the study area has not yet been determined, so JPL, CSR and CLSM data will be used for GWSA inversion for comparative analysis.

[0058] Land subsidence rate and amount are closely related to groundwater level decline rate and amount. According to studies conducted by scholars in cities such as Cangzhou, Dezhou, and Langfang, the empirical relationship between land subsidence and groundwater level changes generally exhibits an exponential, quadratic, or linear correlation. This invention uses InSAR subsidence results as prior knowledge and reconstructs groundwater storage changes in the Beijing Plain based on a novel weighted fusion method of subsidence features. This method improves the spatial resolution of groundwater storage retrieval using GRACE. Figure 11 (a) and (b) show the 0.5°×0.5° GWSA trend from 2018 to 2020 obtained from two GRACE data based on the groundwater balance equation. Figure 11 (c) shows the CLSM-GWSA 0.5°×0.5° trend from 2017 to 2020, the time span, and the SAR image matching. Figure 11 (d)-(f) show the results of downscaling the three GWSAs to 0.05°×0.05° using a novel sedimentation feature weighted fusion method. Compared with the original GWSA, the fused GWSA has a significantly improved spatial resolution, which meets the needs of water resource management agencies.

[0059] The surface subsidence-weighted mass distribution provides a better map of groundwater storage changes, indicating that InSAR measurements not only provide the spatial location of groundwater storage loss but also offer important quantitative information. Furthermore, despite differences in lithology (aquifer confinement, thickness, and compressibility) within the study area, a significant correlation exists between groundwater depletion rates and subsidence rates. Similar inversion results were observed from different data sources. Figure 11(d)-(f)), which demonstrates the potential of combining the two. The novel GRACE / InSAR settlement feature weighted fusion inversion method proposed in this invention will be applied to future generations of GRACEMascon missions, and is expected to improve spatial resolution.

[0060] like Figure 12 As shown, the error distribution pattern reveals that most areas have relatively small errors, below 10 mm / yr, with the largest errors located in the northeast and western corners. The GWSA signals in these areas cannot be explained by the mass distribution map derived from InSAR, indicating that the subsidence signals in the northeast and western corners cannot be entirely attributed to changes in groundwater storage; the deformation in these areas may be significantly influenced by natural hydrological processes. Spatial correlation analysis of the remaining areas after excluding the northeast and western corners reveals that the estimation results from the novel subsidence feature weighted fusion method are largely consistent with those from groundwater level monitoring wells.

[0061] The error between the fusion results and the measured data can be explained from four aspects. First, due to leakage errors in GRACE, GRACE-GWSA does not fully reflect regional groundwater depletion. Second, in the original GWSA trend map at 0.25°, 36% of the pixels were not calculated ((406-258) / 406≈0.36), indicating that some grid aquifers did not subside or were compacted at rates below the InSAR detection threshold. Third, GRACE and the official groundwater budget differ in their measurement mechanisms for groundwater depletion. GRACE provides information on mass loss related to groundwater, including all dynamic effects within the aquifer, but it is contaminated by spatial leakage inherent in gravity field resolution. Fourth, the official groundwater budget ignores the dynamic effects of aquifer depletion, wastewater reinjection, and seepage in the water distribution system. Notably, changes in deep confined aquifer head have a greater impact on land subsidence than unconfined and shallow confined aquifers. In particular, the water level changes of the aquifers adjacent to the deepest part of the monitoring well, which contributes the most to the subsidence, are the main factor in the formation of land subsidence in this area, and our groundwater level data belongs to unconfined phreatic water.

[0062] Figure 13 It presents a comparison of GWSA, subsidence, and precipitation data in terms of long-term trends and seasonal variations. Analysis includes regional averages from the entire Beijing Plain and averages from the northern part of the plain (the plain is divided into three parts: north, central, and south). Figure 13(a) Comparing the average values ​​of the three types of data for the Beijing plain, it was found that the long-term subsidence trend of GRACE-GWSA and InSAR was relatively consistent, showing a downward trend; while CLSM-GWSA and GPS subsidence were relatively consistent, showing no obvious trend and remaining relatively stable. The average subsidence rate from 2015 to 2020 was higher than that of CLSM-GWSA, possibly because the CLSM hydrological model lacks human activity components in its structure. The difference between InSAR and GPS is because there are fewer GPS stations in Beijing, which cannot accurately represent the subsidence of the entire Beijing area. Figure 13 (b) The average InSAR subsidence in the northern part of the Beijing Plain was compared with the GPS subsidence at station BJSH. It can be seen that the long-term trend of InSAR time series and GPS vertical displacement and the magnitude of subsidence are in good agreement on a regional average, which also shows that InSAR can capture high-precision deformation. Figure 13 (c) indicates the seasonal signal after removing the trend term. The seasonal components show that (1) GWSA, settlement and rainfall all have obvious seasonal oscillations, but the amplitude of the oscillations varies greatly. The rainfall range is -62.22 to 87.27 mm, the GWSA range is -9.69 to 13.14 mm, and the settlement fluctuates within ±3 mm. (2) Settlement generally lags behind GWSA. This time difference is likely due to the low vertical conductivity of the aquifer system and the slow consolidation of the soil in humid weather. (3) Settlement is negatively correlated with rainfall, with an absolute correlation coefficient of about 0.45, while the correlation with GWSA is low (0.06), indicating that the oscillation is most likely caused by changes in rainfall.

[0063] Because the time series data on water storage changes and sedimentation are corrupted by noise, there are certain errors in estimating the fusion results and decomposing the seasonal term. The correlation coefficients also show that the seasonal term is much less correlated than the long-term trend term. Therefore, this invention averages the time series data of deformed pixels near the BJSH site to reduce random noise. Figure 14 The study presents a time-series subsidence comparison between a single grid near the BJSH site and the InSAR average value in the northern part of the plain. It finds that the time-series changes of the average value and the neighboring grid are similar, except that the fluctuation range of the average value is smaller. Therefore, the estimation at the plain scale is feasible.

[0064] Figure 15 Google Earth revealed numerous vegetable greenhouses, factories, and residential buildings at the subsidence cone site, suggesting that human activities such as agricultural production may also be a cause of subsidence. In summary, groundwater extraction, human activities, rainfall, and hydrogeological conditions all contribute to surface deformation.

[0065] The technical solution of this invention is based on the extraction of subsidence features and the derivation of a weighting algorithm. The novel subsidence feature weighted fusion method fully considers the impact of both surface subsidence and water level changes on groundwater storage changes. By weighted fusion of groundwater storage anomalies retrieved from GRACE and InSAR data, a completely new fused image is obtained. The fusion result is more sensitive to surface subsidence features, better represents the detailed characteristics of water storage changes, and can provide more accurate information services for water resource utilization.

[0066] Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the embodiments of the invention. Therefore, the embodiments of the invention are not to be limited to the embodiments shown herein, but are to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0067] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for improving the spatial resolution of groundwater storage based on weighted fusion of settlement characteristics, characterized in that, Spatial downscaling of GRACE was performed using a weighted fusion method based on settlement characteristics, including: S1, collect data from regional groundwater level monitoring wells, and calculate the time series variation trend of individual wells and the spatial distribution of the rate of change of the entire region; S2, download Sentinel-1A imagery, high-precision DEM and POD precise orbit data, and use multi-temporal InSAR method to detect the spatiotemporal evolution of regional surface deformation; S3: Obtain the JPL Mascon and CSR Mascon datasets, and extract terrestrial water storage anomalies from these two types of GRACE time-series data. S4, using InSAR technology to detect the distribution of surface deformation in groundwater-depleted areas and groundwater storage anomalies retrieved from GRACE data; In S2, the InSAR processing procedure includes determining the primary and secondary images for differential interferometry, terrain phase removal, phase unwrapping, phase filtering, and geocoding. The primary and secondary images are processed using 10×2 multi-view processing to suppress noise; after generating phase interferograms and amplitude intensity maps, terrain phases are removed from the interferometric phases using a DEM; buildings with high coherence and relatively stable scattering characteristics are selected as scattering points, and a three-dimensional phase unwrapping algorithm is used for phase unwrapping. The Goldstein branching method was used for phase filtering, and the linear phase ramp of each interferogram was subtracted to eliminate orbital errors and atmospheric phase interference. The line-of-sight deformation time series of each pixel was obtained by least squares inversion, and the results were geocoded. In S4, InSAR settlement results are used as prior knowledge. A statistical normalization algorithm is derived based on Max-Min normalization. The deformation results are used as weights of spatial prior knowledge using the statistical normalization algorithm. The GRACE and weights are combined into a new value based on the GRACE / InSAR settlement feature weighted fusion method, thereby improving the resolution of the GRACE signal.

2. The method for improving the spatial resolution of groundwater storage based on weighted fusion of settlement features according to claim 1, characterized in that, In S4, when GRACE inverts groundwater storage anomalies, it is necessary to remove the equivalents of soil water, surface water, vegetation water, and snow water.

3. The method for improving the spatial resolution of groundwater storage based on weighted fusion of settlement features according to claim 2, characterized in that, At the same resolution, the groundwater storage anomaly distribution map is obtained by subtracting soil water and surface water from terrestrial water storage, expressed by the following formula: Among them, GWSA indicates groundwater storage anomaly, TWSA indicates total terrestrial water storage anomaly, SMSA indicates soil water storage anomaly, and SWSA indicates surface water storage anomaly.

4. The method for improving the spatial resolution of groundwater storage based on weighted fusion of settlement features according to claim 2, characterized in that, The changes in terrestrial and groundwater storage retrieved by GRACE, as well as the changes in soil and surface water retrieved by NOAH, require interpolation and anomaly processing.

5. The method for improving the spatial resolution of groundwater storage based on weighted fusion of settlement features according to claim 1, characterized in that, In step S1, more than 24 consecutive months are selected as time reference intervals to obtain GPS time series provided by CMONOC for seasonal variation analysis.

Images