A tree-ring-based hydroclimate anomaly detection method and system

Abstract

The present application relates to a kind of hydrological climate anomaly detection method and system based on tree ring, method includes: the current region of the ring image data to be detected is collected;The ring image data to be detected is input into hydrological climate anomaly detection model, and the hydrological climate anomaly event of target area is output, wherein the hydrological climate anomaly detection model is based on machine learning algorithm construction and is obtained by training set training, the training set includes preprocessed ring image data and corresponding hydrological climate anomaly mark;The hydrological climate anomaly event is analyzed in space-time mode, and the time lag degree of different regional anomaly event occurrence and spatial correlation are calculated;According to the time lag degree of anomaly event occurrence and spatial correlation, the influence range of hydrological climate anomaly event is divided, and regional influence partition map is obtained.The present application improves the accuracy and reliability of hydrological climate anomaly identification.

Description

Technical Field

[0001] This invention relates to the field of hydrological and climatic science and technology, and in particular to a method and system for detecting hydrological and climatic anomalies based on tree rings. Background Technology

[0002] Hydroclimate research, as an important branch of environmental science, has irreplaceable value in understanding the dynamic changes of regional ecosystems, predicting natural disasters, and formulating resource management strategies. This field provides crucial evidence for addressing global climate change by analyzing the long-term patterns of climate and hydrological factors. However, current research methods have significant limitations in data acquisition and analytical precision. Many traditional methods rely on short-term observational data or indirect inferences, making it difficult to accurately reflect the long-term evolutionary characteristics of hydroclimate, especially in remote areas where data is scarce, where the shortcomings of existing methods are even more pronounced.

[0003] Against this backdrop, hydrological and climate anomaly detection faces numerous challenges. Firstly, there are limitations in data sources. Tree-ring data, as a natural archive recording long-term environmental changes, undergoes complex collection and processing processes that are susceptible to interference, resulting in inconsistent data quality. This instability directly impacts the accuracy of feature extraction, making it difficult to extract key parameters highly correlated with hydrological and climate change from tree-ring width sequences. Furthermore, the inadequacy of feature extraction further restricts the reliability of model construction, rendering anomaly detection models based on historical data insufficient in both prediction accuracy and adaptability when facing complex environmental changes. This progressive problem, from data to features to models, forms a technological chain that urgently needs breakthroughs.

[0004] Therefore, optimizing the collection and preprocessing of tree-ring data, accurately extracting hydrological and climatic features, and constructing a highly reliable anomaly detection model have become key issues in improving the ability to identify hydrological and climatic anomalies in target areas. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for detecting hydrological and climatic anomalies based on tree rings, which improves the accuracy and reliability of hydrological and climatic anomaly identification.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] A method for detecting hydroclimatic anomalies based on tree rings includes:

[0008] Collect image data of the tree rings to be detected in the current area;

[0009] The tree-ring image data to be detected is input into the hydrological and climate anomaly detection model, and the hydrological and climate anomaly events in the target area are output. The hydrological and climate anomaly detection model is constructed based on machine learning algorithms and trained on a training set, which includes preprocessed tree-ring image data and corresponding hydrological and climate anomaly markers.

[0010] Spatiotemporal pattern analysis was performed on the aforementioned hydrological and climatic anomalies to calculate the temporal lag and spatial correlation of the anomalies in different regions.

[0011] Based on the time lag and spatial correlation of abnormal events, the impact range of hydrological and climatic anomalies is delineated, and a regional impact zoning map is obtained.

[0012] Optionally, obtaining the training set includes:

[0013] The tree ring samples within the target area are scanned to obtain raw image data;

[0014] The original image data is denoised using a multi-scale filtering algorithm to obtain pre-denoised image data.

[0015] The filter parameters are dynamically optimized using an adaptive weighting method to obtain the optimized filter parameter configuration.

[0016] The optimized filter parameters are used to perform secondary processing on the initially denoised image data to obtain the final denoised image data.

[0017] Data extraction is performed on the final denoised image data to obtain the annual ring width sequence data;

[0018] The training set is obtained based on the tree ring width sequence data and the corresponding hydrological and climatic anomaly markers.

[0019] Optionally, training the hydro-climate anomaly detection model using a training set includes:

[0020] The annual ring width sequence data is decomposed and reconstructed using wavelet analysis to obtain the annual ring width component sequence.

[0021] The tree ring width component sequence is input into the hydrological and climate anomaly detection model, which outputs predicted hydrological and climate anomaly events. The hydrological and climate anomaly detection model is trained based on the labels of hydrological and climate anomalies, wherein the hydrological and climate anomaly events include the anomaly occurrence time, duration, intensity level and impact range.

[0022] Optionally, multi-scale decomposition and reconstruction of the annual ring width sequence data using wavelet analysis includes:

[0023] The discrete wavelet transform algorithm is used to perform multi-scale decomposition on the annual ring width sequence data to extract wavelet coefficients containing different frequency information.

[0024] The annual ring width component sequence is obtained by reconstructing information of different frequencies using wavelet reconstruction technology.

[0025] Optionally, spatiotemporal pattern analysis is performed on the aforementioned hydrological and climatic anomalies to calculate the time lag and spatial correlation of the anomalies in different regions, including:

[0026] The dynamic time warping algorithm is used to calculate the time series similarity of the occurrence of the hydrological and climatic anomalies in different regions, and to obtain the time lag of the occurrence of the anomalies in different regions.

[0027] The spatial autocorrelation index of the hydrological and climatic anomalies is calculated based on geospatial analysis methods to determine whether the spatial distribution pattern of the anomalies shows clustering and to obtain the spatial correlation of the anomalies in different regions.

[0028] Optionally, delineating the impact range of hydrological and climate anomalies includes: grouping regions with similar time lag and spatial correlation through hierarchical cluster analysis, delineating the impact range of hydrological and climate anomalies, and obtaining the regional impact zoning map.

[0029] Optionally, the calculation of the time lag of abnormal events in different regions may also include:

[0030] An abnormal event propagation network is constructed using graph theory network analysis. Key propagation paths and affected areas are identified through network topology analysis. In this network, nodes represent regions, edges represent propagation relationships, and edge weights represent time lag.

[0031] The present invention also provides a hydro-climate anomaly detection system based on tree rings for implementing a hydro-climate anomaly detection method based on tree rings, comprising: a data acquisition module, a hydro-climate anomaly detection module, an analysis module, and a classification module;

[0032] The acquisition module is used to acquire the image data of the tree rings to be detected in the current area;

[0033] The hydrological and climate anomaly detection module is used to input the tree-ring image data to be detected into the hydrological and climate anomaly detection model and output hydrological and climate anomaly events in the target area. The hydrological and climate anomaly detection model is constructed based on a machine learning algorithm and trained on a training set, which includes preprocessed tree-ring image data and corresponding hydrological and climate anomaly markers.

[0034] The analysis module is used to perform spatiotemporal pattern analysis on the hydrological and climate anomalies and calculate the time lag and spatial correlation of the anomalies in different regions.

[0035] The partitioning module is used to partition the impact range of hydrological and climate anomalies based on the time lag and spatial correlation of the anomalies, and obtain a regional impact zoning map.

[0036] The beneficial effects of this invention are as follows: First, automated image recognition technology is used to scan tree ring samples. Multi-scale filtering and local enhancement algorithms are employed to denoise and preserve details, significantly improving image quality and ensuring the reliability of subsequent analysis. Then, the tree ring width sequence is obtained and optimized. Wavelet analysis is used for multi-scale decomposition and reconstruction of the tree ring width sequence, enabling fine processing of the high-frequency and low-frequency components of the signal. Wavelet reconstruction recombines these decomposed components to recover the original signal or a clearer signal, ensuring the effectiveness of model training. Finally, a regional influence partitioning map is obtained, optimizing the detection capability of abnormal events.

[0037] This invention enables accurate identification and regional impact analysis of regional hydrological and climatic anomalies, improving the accuracy and reliability of hydrological and climatic anomaly identification and providing an effective tool for water resources management and climate change research. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is a flowchart of a hydrological and climatic anomaly detection method based on tree rings, according to an embodiment of the present invention. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0042] Example 1:

[0043] like Figure 1 As shown, this embodiment provides a method for detecting hydro-climate anomalies based on tree rings, including:

[0044] Collect image data of the tree rings to be detected in the current area;

[0045] The tree-ring image data to be detected is input into the hydrological and climate anomaly detection model, which outputs hydrological and climate anomaly events in the target area. The hydrological and climate anomaly detection model is built based on machine learning algorithms and trained on a training set, which includes preprocessed tree-ring image data and corresponding hydrological and climate anomaly labels.

[0046] Spatiotemporal pattern analysis was performed on hydrological and climatic anomalies to calculate the time lag and spatial correlation of anomalies in different regions.

[0047] Based on the time lag and spatial correlation of abnormal events, the impact range of hydrological and climatic anomalies is delineated, and a regional impact zoning map is obtained.

[0048] Furthermore, obtaining the training set includes:

[0049] The tree ring samples within the target area are scanned to obtain raw image data;

[0050] A multi-scale filtering algorithm is used to denoise the original image data to obtain the image data after preliminary denoising.

[0051] The filter parameters are dynamically optimized using an adaptive weighting method to obtain the optimized filter parameter configuration.

[0052] The optimized filter parameters are used to perform secondary processing on the initially denoised image data to obtain the final denoised image data.

[0053] Data extraction is performed on the final denoised image data to obtain the annual ring width sequence data;

[0054] The training set is obtained based on the tree-ring width sequence data and the corresponding labels of hydrological and climatic anomalies.

[0055] Specifically, during the initial scanning of tree ring samples using automated technology, a high-precision scanner can be used to acquire raw image data. Assuming a scanning resolution of 1200 dpi, which can capture the subtle textures of the tree rings, the scanning process must ensure the sample is placed flat to avoid image distortion. This method effectively improves the accuracy of subsequent processing.

[0056] To analyze the distribution characteristics of tree ring images, we can begin by understanding the overall features of the image data and the regularity of its spatial distribution. Tree ring images typically exhibit a periodically changing ring structure, which reflects the years of tree growth. By observing the original image data, we can find that noise often interferes with the continuity of tree rings in the form of random spots or stripes; therefore, denoising processing is needed to improve the accuracy of subsequent analysis.

[0057] In this embodiment, a multi-scale filtering algorithm is used to denoise the original image data. Different filtering scales can be selected based on the characteristics of different regions of the image. For example, a smaller filtering scale can be used to preserve details in areas with clear tree ring boundaries, while a larger filtering scale is used to smooth noise in noisy background areas. Assuming an image has a resolution of 2000x2000 pixels, after initial denoising, approximately 15% of the area still has noise interference, requiring further optimization. Based on the image data after initial denoising, the characteristics of the areas where details are lost are analyzed, and the filtering parameters are dynamically optimized using an adaptive weighting method to obtain the optimized filtering parameter configuration. Using the optimized filtering parameter configuration, the image data after initial denoising is processed a second time to generate final denoised image data with a higher degree of detail preservation. When extracting the tree ring width sequence from the final denoised image data and analyzing its spatial distribution characteristics, the pixel width of each tree ring can be scanned to record its continuity and integrity indicators. If, for example, the continuity indicator in a certain tree ring width sequence is 85%, which is lower than the preset 90% threshold, local optimization processing is required for that area to ensure data integrity. It should be noted that if the level of detail preservation in a local area is lower than the threshold, when reapplying the multi-scale filtering algorithm, the focus can be on the noise characteristics of that area, and the filtering parameters can be adjusted.

[0058] When extracting the tree ring width sequence, the width data of each tree ring can be recorded using a pixel measurement tool. Let's assume the width sequence is 3mm, 4mm, 2mm, etc.

[0059] Furthermore: Training the hydro-climate anomaly detection model using the training set includes:

[0060] Wavelet analysis was used to perform multi-scale decomposition and reconstruction of the tree ring width sequence data to obtain the tree ring width component sequence.

[0061] The tree ring width component sequence is input into the hydrological and climate anomaly detection model, which outputs predicted hydrological and climate anomaly events. The model is then trained based on the labels of hydrological and climate anomalies, including the occurrence time, duration, intensity level, and impact range of the anomalies.

[0062] Furthermore, multi-scale decomposition and reconstruction of the annual ring width sequence data using wavelet analysis includes:

[0063] The discrete wavelet transform algorithm is used to perform multi-scale decomposition on the annual ring width sequence data to extract wavelet coefficients containing different frequency information.

[0064] By reconstructing information of different frequencies using wavelet reconstruction techniques, the sequence of annual ring width components is obtained.

[0065] Specifically, discrete wavelet transform is an effective tool for analyzing the periodicity of tree ring width. It should be noted that for coniferous tree ring data, the Daubechies4 (db4) wavelet basis function typically performs best because its shape is similar to the characteristics of tree ring width changes. Performing a 5-level wavelet decomposition on the spruce tree ring sequence yields wavelet coefficients reflecting climate information at different time scales. For example, the first level of detailed coefficients reflects short-term changes of 1-2 years (possibly corresponding to seasonal precipitation), the third and fourth levels reflect medium-term changes of 7-16 years (possibly related to ENSO events), and the fifth level of approximate coefficients reflects long-term climate trends. Preferably, by calculating the energy proportion of each frequency band, it was found that the spruce sample has an energy proportion of 42% in the 7-11 year cycle, significantly higher than the 95% confidence interval threshold (23%), indicating a significant quasi-decadal cycle. This cycle amplitude is approximately 0.25 mm, highly consistent with the 10-year cycle of regional precipitation records (correlation coefficient r = 0.76), revealing the periodic impact of precipitation on tree growth. In one embodiment, wavelet reconstruction is performed on the high-frequency coefficients (layers 1-3) and the low-frequency coefficients (layers 4-5) respectively to obtain two annual ring width component sequences that reflect short-term climate fluctuations and long-term climate trends.

[0066] Machine learning algorithms include, but are not limited to, decision tree algorithms, support vector machine algorithms, or neural network algorithms. Methods such as grid search, random search, and Bayesian optimization are used to fine-tune the model's hyperparameters to find the optimal parameter combination and improve model performance.

[0067] Hydrological and climatic anomalies include the timing of occurrence, duration, intensity level, and affected area. For example, an abnormal rainfall event lasts for 3 months, with a peak rainfall of 300 mm / month and a cumulative rainfall of 700 mm. A feature vector [3, 300, 700] is constructed to comprehensively describe the event characteristics. This feature extraction provides a quantitative basis for subsequent analysis, facilitating event classification and comparison. In one embodiment, extreme value analysis combined with K-means clustering is used to classify abnormal events. Assuming that the feature vectors are clustered into three categories: mild events with a peak intensity below 200 mm, moderate events with 200 to 400 mm, and severe events with an intensity above 400 mm, an event with a peak intensity of 300 mm is classified as moderate. This classification method is intuitive and clear, helping to quickly assess the degree of impact of events. For example, a historical hydrological and climatic anomaly database can record a moderate flood event that occurred from June to August 2015, lasting for 3 months, with a peak intensity of 300 mm, affecting the downstream area of ​​a certain watershed.

[0068] Furthermore, spatiotemporal pattern analysis was conducted on hydrological and climatic anomalies to calculate the time lag and spatial correlation of these anomalies in different regions, including:

[0069] The dynamic time warping algorithm is used to calculate the time series similarity of hydrological and climate anomalies in different regions and to obtain the time lag of anomalies in different regions.

[0070] Based on geospatial analysis methods, the spatial autocorrelation index of hydrological and climatic anomalies is calculated to determine whether the spatial distribution pattern of anomalies shows clustering and to obtain the spatial correlation of anomalies in different regions.

[0071] Specifically, principal component analysis is used to reduce the dimensionality of multidimensional hydrological and climate data, retaining key variables such as rainfall, river level, and soil moisture to generate a low-dimensional feature dataset. Next, a one-dimensional Gaussian mixture model is applied to identify outliers. For example, if a monitoring point experiences 500 mm of rainfall in July 2023, far exceeding the historical average of 300 mm plus two standard deviations, it can be marked as an anomalous event. Time series analysis determines that the anomalous event lasted from July 1st to July 5th, and spatial interpolation is used to create a spatial distribution map of the anomalous event, showing that the rainfall anomaly is concentrated in the upper reaches of the watershed. In one possible implementation, a dynamic time warping algorithm is used to quantify the similarity of the time series of anomalous events. Two monitoring points A and B within the watershed are selected. The anomalous event at point A occurs on July 1st, and at point B, it lags behind to July 3rd. By calculating the warped path distance between the two points' time series, the time lag is determined to be 2 days. Further analysis of multiple monitoring points constructs a time propagation matrix, where the matrix elements represent the time difference of the anomalous events between each point. For example, the time from A to B is 2 days, and from B to C is 1 day, reflecting the temporal pattern of the propagation of abnormal events from upstream to downstream. The spatial autocorrelation index can be calculated using Moran's index analysis. For instance, the rainfall anomaly value calculated using Moran's index for 10 monitoring points in a certain watershed is 0.8, which is greater than the threshold of 0.5, indicating that the spatial distribution of abnormal events shows significant clustering.

[0072] Furthermore, delineating the impact range of hydrological and climatic anomalies includes: grouping regions with similar time lag and spatial correlation through hierarchical cluster analysis to delineate the impact range of hydrological and climatic anomalies and obtain regional impact zoning maps.

[0073] Specifically, hierarchical clustering analysis is used to delineate the impact range of abnormal events. In one embodiment, based on time lag and spatial correlation characteristics, the watershed is divided into three regions: an upstream high-intensity anomaly zone, a midstream transition zone, and a downstream low-impact zone. For example, an upstream region with five consecutive days of abnormal rainfall and high peak intensity is designated as a high-risk zone, facilitating precise disaster prevention deployment.

[0074] Furthermore, the calculated time lag of abnormal events in different regions also includes:

[0075] An abnormal event propagation network is constructed using graph theory network analysis. Key propagation paths and affected areas are identified through network topology analysis. In this network, nodes represent regions, edges represent propagation relationships, and edge weights represent time lag.

[0076] Specifically, graph theory network analysis was used to construct an abnormal event propagation network, treating 10 monitoring points within the basin as nodes and time lag as edge weights. Analysis revealed that node A, as the upstream core node, connects multiple downstream nodes, with the propagation path from A to B and then to C, with time lags of 2 days and 1 day respectively. Network centrality analysis showed that A is a critical propagation node, indicating it is a key flood control monitoring area.

[0077] In this embodiment, early warning data for hydrological and climate anomalies is acquired through a Geographic Information System (GIS). Spatial interpolation is used to generate continuous spatial distribution features, resulting in spatial distribution data for hydrological and climate anomalies. A kernel density estimation algorithm is employed to analyze the spatial distribution data, determining the location and boundaries of high-risk areas and generating a high-risk area distribution map. If the kernel density value of a region in the high-risk area distribution map exceeds a preset threshold, it is marked as a high-risk area, and the spatial extent data of the high-risk area is obtained. Buffer analysis technology is used to generate potential impact ranges based on the spatial extent data of high-risk areas, obtaining the spatial boundary data of the impact range. Based on the spatial boundary data of high-risk areas and impact ranges, a risk zoning map is generated using a hierarchical color-coding method, determining the spatial visualization results of the risk zoning. For the risk levels in the risk zoning map, a color-coding method is used to set early warning signals, generating an intuitive information product containing the early warning signals. The intuitive information product is exported as an interactive map format through the GIS, obtaining the final early warning information product for hydrological and climate anomalies.

[0078] Example 2:

[0079] A hydro-climate anomaly detection system based on tree rings includes: a data acquisition module, a hydro-climate anomaly detection module, an analysis module, and a classification module;

[0080] The acquisition module is used to acquire the image data of the tree rings to be detected in the current area;

[0081] The hydrological and climate anomaly detection module is used to input the tree-ring image data to be detected into the hydrological and climate anomaly detection model and output the hydrological and climate anomaly events in the target area. The hydrological and climate anomaly detection model is built based on machine learning algorithms and trained on a training set, which includes preprocessed tree-ring image data and corresponding hydrological and climate anomaly labels.

[0082] The analysis module is used to perform spatiotemporal pattern analysis on hydrological and climatic anomalies and calculate the time lag and spatial correlation of anomalies in different regions.

[0083] The partitioning module is used to divide the impact range of hydrological and climate anomalies based on the time lag and spatial correlation of the anomalies, and to obtain a regional impact zoning map.

[0084] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for detecting hydro-climate anomalies based on tree rings, characterized in that, include: Collect image data of the tree rings to be detected in the current area; The tree-ring image data to be detected is input into the hydrological and climate anomaly detection model, and the hydrological and climate anomaly events in the target area are output. The hydrological and climate anomaly detection model is constructed based on machine learning algorithms and trained on a training set, which includes preprocessed tree-ring image data and corresponding hydrological and climate anomaly markers. Obtaining the training set includes: The tree ring samples within the target area are scanned to obtain raw image data; The original image data is denoised using a multi-scale filtering algorithm to obtain pre-denoised image data. The filter parameters are dynamically optimized using an adaptive weighting method to obtain the optimized filter parameter configuration. The optimized filter parameters are used to perform secondary processing on the initially denoised image data to obtain the final denoised image data. Data extraction is performed on the final denoised image data to obtain the annual ring width sequence data; The training set is obtained based on the tree-ring width sequence data and the corresponding hydrological and climatic anomaly markers; Training the hydro-climate anomaly detection model using the training set includes: The annual ring width sequence data is decomposed and reconstructed using wavelet analysis to obtain the annual ring width component sequence. The tree ring width component sequence is input into the hydrological and climate anomaly detection model, which outputs predicted hydrological and climate anomaly events. The hydrological and climate anomaly detection model is trained based on the labels of hydrological and climate anomalies, wherein the hydrological and climate anomaly events include the time of occurrence, duration, intensity level and impact range of the anomaly. Spatiotemporal pattern analysis was performed on the aforementioned hydrological and climatic anomalies to calculate the time lag and spatial correlation of the anomalies in different regions, including: The dynamic time warping algorithm is used to calculate the time series similarity of the occurrence of the hydrological and climatic anomalies in different regions, and to obtain the time lag of the occurrence of the anomalies in different regions. Based on geospatial analysis methods, the spatial autocorrelation index of the hydrological and climatic anomalies is calculated to determine whether the spatial distribution pattern of the anomalies shows clustering and to obtain the spatial correlation of the occurrence of anomalies in different regions. Based on the time lag and spatial correlation of abnormal events, the impact range of hydrological and climatic anomalies is delineated, and a regional impact zoning map is obtained.

2. The method for detecting hydro-climate anomalies based on tree rings according to claim 1, characterized in that, Multi-scale decomposition and reconstruction of the annual ring width sequence data using wavelet analysis includes: The discrete wavelet transform algorithm is used to perform multi-scale decomposition on the annual ring width sequence data to extract wavelet coefficients containing different frequency information. The annual ring width component sequence is obtained by reconstructing information of different frequencies using wavelet reconstruction technology.

3. The method for detecting hydro-climate anomalies based on tree rings according to claim 1, characterized in that, Delineating the impact range of hydrological and climate anomalies includes: grouping regions with similar time lag and spatial correlation through hierarchical cluster analysis, delineating the impact range of hydrological and climate anomalies, and obtaining the impact zoning map of the regions.

4. The method for detecting hydro-climate anomalies based on tree rings according to claim 1, characterized in that, The calculated time lag of abnormal events in different regions also includes: An abnormal event propagation network is constructed using graph theory network analysis. Key propagation paths and affected areas are identified through network topology analysis. In this network, nodes represent regions, edges represent propagation relationships, and edge weights represent time lag.

5. A tree-ring-based hydro-climate anomaly detection system for implementing the tree-ring-based hydro-climate anomaly detection method according to any one of claims 1-4, characterized in that, include: The system includes a data acquisition module, a hydrological and climate anomaly detection module, an analysis module, and a classification module. The acquisition module is used to acquire the image data of the tree rings to be detected in the current area; The hydrological and climate anomaly detection module is used to input the tree-ring image data to be detected into the hydrological and climate anomaly detection model and output hydrological and climate anomaly events in the target area. The hydrological and climate anomaly detection model is constructed based on a machine learning algorithm and trained on a training set, which includes preprocessed tree-ring image data and corresponding hydrological and climate anomaly markers. The analysis module is used to perform spatiotemporal pattern analysis on the hydrological and climate anomalies and calculate the time lag and spatial correlation of the anomalies in different regions. The partitioning module is used to partition the impact range of hydrological and climate anomalies based on the time lag and spatial correlation of the anomalies, and obtain a regional impact zoning map.

Images