A method and system for joint determination of tree ring width and density at multiple sites in a watershed

Abstract

This invention discloses a method and system for joint determination of tree ring width and density at multiple sites in a watershed, belonging to the field of tree ring determination technology. The method includes six steps: sample collection and preprocessing, optical determination of tree ring width, joint determination of tree ring density using X-rays and blue light reflectance, multi-parameter cross-dating, measurement data quality assessment, and spatial distribution analysis. The system includes six corresponding functional modules. This invention achieves the acquisition of spatial distribution information of tree ring parameters at the watershed scale through batch determination at multiple sites, joint cross-dating of width and density, and three-dimensional quality assessment, and outputs a spatial distribution map of climate sensitivity, providing systematic data support for watershed paleoclimate reconstruction.

Description

Technical Field

[0001] This invention relates to the field of tree ring measurement technology, and more particularly to a method and system for the joint measurement of tree ring width and density at multiple sites in a watershed. Background Technology

[0002] Dendrology is an important discipline that studies historical climate change, ecological environment evolution, and tree growth patterns by analyzing the physical characteristics of tree rings. Ring width and ring density are two of the most crucial parameters in dendrology research; the former reflects the radial growth of the tree, while the latter reflects the combined characteristics of wood cell wall thickness and cell lumen size. Accurate determination of tree ring parameters has significant scientific and practical value in applications such as paleoclimate reconstruction, ecological environment monitoring, and ancient tree age determination.

[0003] Chinese invention patent application CN106093048A discloses a method and apparatus for tree ring dating. This invention addresses the limitations of traditional cross-dating methods based on tree species and climate, proposing a technical solution that establishes a standard curve based on the relationship between annual precipitation, annual accumulated temperature, and tree ring width. Specifically, the solution includes five steps: sample collection, sample processing, tree ring interpretation, tree ring dating, and primary age estimation. In the sample collection stage, a stress wave 3D imager is used to detect areas of the tree trunk free of decay, and a growth cone or saw is used to obtain heartwood samples or tree ring disc samples. In the sample processing stage, sandpaper of different coarse to fine grits is used to polish the tree until cell outlines are visible under an 80x stereomicroscope. In the tree ring interpretation stage, the locations of missing rings, broken rings, and pseudo-rings are determined using a stereomicroscope, and the number of rings is counted. In the tree ring dating stage, the tree ring width data is measured and standardized, and the age of each ring is determined through binary regression analysis with the tree ring width standard curve. This scheme has, to some extent, solved the problem of establishing accurate standard curves in urban microclimates, and enabled the dating of tree rings for open-field tree species within the Fifth Ring Road of Beijing.

[0004] However, the aforementioned existing technologies have the following limitations: First, this method only measures single trees or a small number of trees at a single site, failing to obtain spatial distribution information of tree-ring parameters at the watershed scale, thus making it difficult to meet the needs for spatial continuity data in watershed hydro-climate reconstruction and ecosystem assessment. Second, this method only measures the single parameter of tree-ring width, neglecting the measurement of tree-ring density. Tree-ring density, especially maximum latewood density, typically correlates more strongly with growing season temperature than tree-ring width, and is irreplaceable for paleoclimate reconstruction in temperature-sensitive areas. Third, this method relies on manual interpretation using stereomicroscopes, resulting in low efficiency and high subjectivity, making it difficult to achieve efficient measurement of large batches of samples from multiple sites. Fourth, this method lacks a systematic data quality assessment system, failing to quantitatively evaluate the statistical quality and climate signal intensity of the measurement sequences, thus affecting the reliable application of the data in paleoclimate reconstruction. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for the joint determination of tree ring width and density at multiple sites in a watershed, so as to solve the technical problem that existing tree ring measurement technology is limited to single site and single parameter and is difficult to obtain spatial distribution information of tree ring physical characteristics at the watershed scale.

[0006] To achieve the above objectives, this invention provides a method for joint determination of tree ring width and density at multiple sites within a watershed. The method includes: a sample collection and preprocessing step, acquiring watershed sampling point location data from multiple sampling sites within the watershed, and performing surface polishing and fixation on the core samples collected from each sampling site to generate processed sample data; an optical measurement step for tree ring width, using a high-resolution linear array scanner to scan the surface of the processed samples at each site, acquiring sample surface image data, automatically identifying the tree ring boundary positions using an image grayscale gradient analysis algorithm, and measuring and outputting the tree ring width sequence data ring by ring; and a tree ring density measurement step, using a microfocus X-ray scanning system to perform transmission scanning on the processed samples at each site, acquiring X-ray attenuation profile data, calculating the earlywood density, latewood density, and maximum density based on the quantitative relationship between the X-ray attenuation coefficient and wood density, and simultaneously acquiring blue light using a blue light reflectance measurement mode. The data is processed as follows: Reflectance intensity data is converted into density sequences according to a preset calibration relationship, and the resulting annual ring density sequence data is integrated and output. A multi-parameter cross-dating step utilizes the complementary temporal variations of annual ring width and density sequences for cross-validation. An anomalous annual ring detection algorithm identifies missing, pseudo-rings, and damaged rings and generates anomalous annual ring marker data. Based on the collaborative matching of the width and density sequences, the absolute age of each annual ring is determined, and the dating results are output. A data quality assessment step calculates the signal-to-noise ratio, inter-sample correlation coefficient, and population signal intensity for each station's dating results, comprehensively assesses the statistical quality of the measurement sequence, and outputs a quality assessment report. A spatial distribution analysis step performs spatial interpolation on the multi-station measurement data within the watershed to generate continuous distribution field data for the watershed. Climate sensitivity coefficients for annual ring parameters and climate factors at each station are calculated, and a spatial distribution map of climate sensitivity is output.

[0007] This invention also provides a system for implementing the above method, the system comprising: a sample collection and preprocessing module, used to acquire watershed sampling point location data from multiple sampling stations within the watershed, and to perform surface polishing and fixation processing on tree core samples collected from each sampling station; an annual ring width optical measurement module, including a high-resolution linear array scanner and an image processing unit, used to perform surface scanning on the processed samples to acquire sample surface image data, identify annual ring boundaries through an image grayscale gradient analysis algorithm, and output annual ring width sequence data; and an annual ring density measurement module, including a microfocus X-ray scanning unit and a blue light reflectance measurement unit, used to acquire X-ray attenuation profile data and blue light reflectance intensity data, and to calculate and output data including earlywood density, latewood density, and most abundant wood density. High-density tree-ring density sequence data; a multi-parameter cross-dating module, connected to the tree-ring width optical measurement module and the tree-ring density measurement module, is used to perform cross-validation, anomaly tree-ring detection, and absolute age determination based on tree-ring width sequence data and tree-ring density sequence data, and outputs dating results data and anomaly tree-ring marker data; a measurement data quality assessment module, connected to the multi-parameter cross-dating module, is used to calculate the signal-to-noise ratio index, inter-sample correlation coefficient, and expression population signal strength, and output quality assessment report data; a spatial distribution analysis module, connected to the measurement data quality assessment module, is used to perform spatial interpolation on multi-site measurement data to generate continuous watershed distribution field data, calculate climate sensitivity coefficients, and output climate sensitivity spatial distribution maps.

[0008] The beneficial effects of this invention are as follows: First, by constructing a multi-site batch measurement framework for the watershed, it achieves a leap from single-site single-parameter measurement to watershed-scale multi-parameter joint measurement, enabling the acquisition of spatial distribution information of tree-ring physical characteristics. Second, through joint measurement and cross-dating of width and density parameters, the dating accuracy is significantly improved, and the anomalous tree-ring detection algorithm can effectively identify missing rings, pseudo-rings, and damaged rings. Third, through the dual-modal configuration of X-ray density measurement and blue light reflectance measurement, both the absolute accuracy of density measurement and a rapid alternative solution are ensured. Fourth, through a three-dimensional quality assessment system of signal-to-noise ratio, inter-sample correlation coefficient, and expression population signal intensity, quantitative evaluation of measurement data quality is achieved. Fifth, through spatial interpolation and climate sensitivity analysis, a continuous distribution field and a spatial distribution map of climate sensitivity in the watershed are output, providing a systematic data product for watershed paleoclimate reconstruction. Attached Figure Description

[0009] Figure 1 This is a flowchart of the method for jointly measuring tree ring width and density at multiple stations in a watershed, as described in an embodiment of the present invention.

[0010] Figure 2 This is an architecture diagram of the watershed multi-site tree ring width and density joint measurement system described in this embodiment of the invention. Detailed Implementation

[0011] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments. The following embodiments are only used to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0012] like Figure 1 As shown, this embodiment provides a method for joint determination of tree ring width and density at multiple stations in a watershed. This method addresses the limitation of traditional tree ring determination techniques to single-site, single-parameter measurements by designing a framework that combines multi-site batch measurement, multi-parameter joint detection, and spatial characteristic characterization. In one embodiment of this invention, a paleoclimate reconstruction project in a watershed with an area of ​​approximately 5000 km² is used as the application scenario. 2 Eight sampling stations were set up, with 20 core samples collected at each station. The target tree species was Pinus tabuliformis. The entire measurement method includes the following six core steps.

[0013] Step S1: Sample collection and preprocessing steps.

[0014] Sample collection and pretreatment are fundamental steps in the entire measurement method. Their purpose is to obtain the geographical location information of multiple sampling sites within the watershed and to standardize the collected tree core samples, providing high-quality sample materials for subsequent width and density measurements.

[0015] Regarding the site layout, the sampling sites are first scientifically laid out based on the topographic features, vegetation distribution, and climate gradient differentiation patterns of the watershed. Preferably, the sampling sites should cover different elevation gradients, slope aspects, and hydrothermal conditions within the watershed to ensure that the acquired annual ring data can represent the overall climate response characteristics of the watershed. In one embodiment of the invention, eight sampling sites are deployed within the watershed, with elevations ranging from 1200m to 2600m, covering three slope aspects: shady, sunny, and semi-shady. At each sampling site, 15-25 healthy adult individuals of the target tree species are sampled, preferably those with a diameter at breast height (DBH) greater than 25cm, intact crowns, and no obvious pests, diseases, or mechanical damage. GPS equipment is used to record the latitude and longitude coordinates and elevation of each sampling site, forming watershed sampling point location data.

[0016] For sample collection, Swedish-made Haglof growth cones were used to collect core samples from the target trees. The growth cones had an inner diameter of 5.15 mm, and the sampling height was uniformly set at breast height (DBH), i.e., 1.3 m above the ground. The sampling direction was either perpendicular to the slope or north-south. Two core samples were collected from each tree for cross-validation, with the two core samples taken perpendicular to each other. During sampling, it was ensured that the growth cone penetrated the heartwood or was as close to the heartwood as possible to obtain a complete tree-ring sequence. The collected core samples were protected using plastic tubes or paper pipettes, and labeled with basic information such as sample number, sampling date, tree species, DBH, and tree height, forming the original core sample dataset.

[0017] For sample pretreatment, the collected tree core samples are transported back to the laboratory for drying, fixation, and polishing. First, the tree core samples are air-dried naturally at room temperature for 7-14 days until the moisture content drops below 12%. Direct sunlight and drastic temperature changes should be avoided during drying to prevent cracking or deformation. Preferably, the drying environment temperature is controlled within the range of 18-25 degrees Celsius, and the relative humidity is controlled within the range of 45-65%. For fresh samples with high moisture content, a gradual dehumidification method can be used for pre-drying, reducing the relative humidity by 5-10 percentage points every 24 hours until the target moisture content is reached.

[0018] The air-dried tree core sample is then fixed to a specialized wooden trough using wood glue. The trough should be made of hardwood or plastic that is not easily deformed. When fixing the sample, ensure the cross-section of the tree core faces upwards and the growth rings are parallel to the long axis of the trough. The width of the trough should match the diameter of the tree core, typically using a standard 6mm wide trough, with a depth approximately two-thirds of the tree core diameter. A fast-drying white wood glue or polyvinyl acetate emulsion adhesive is used. The amount of glue applied should be moderate, just enough to completely cover the contact surface between the tree core and the trough, avoiding glue overflow and contamination of the sample's measurement surface. After the glue has completely cured, it usually needs to stand for 4-8 hours before subsequent sanding.

[0019] The core surface is polished using a decreasing grit sandpaper sequence. Preferably, the grit sequence is 120 grit, 240 grit, 400 grit, 600 grit, 800 grit, and 1200 grit, with each grit lasting approximately 2-3 minutes. The sanding direction is back-and-forth along the long axis of the core. Uniform pressure should be maintained during sanding to avoid localized over-sanding that could distort the measurement of annual ring width. For harder wood species such as Pinus tabuliformis and Platycladus orientalis, the sanding time with lower grit sandpaper can be increased appropriately; for softer wood species such as spruce and fir, the pressure should be reduced to prevent surface damage.

[0020] After sanding, use compressed air or a soft brush to remove sawdust from the sample surface. If necessary, apply a small amount of chalk or talcum powder to the sample surface to enhance the visibility of the growth ring boundaries. The treated core surface should be smooth and flat, and the outline structure of wood cells should be clearly observed under a 40-80x stereomicroscope. For coniferous species, the boundary between earlywood and latewood should be clearly distinguishable, with latewood cells exhibiting a distinct dark banded characteristic due to the thickened cell walls. The growth ring boundary typically shows an abrupt transition from dark latewood cells to light earlywood cells; this boundary line should be continuous, complete, and without blurring or breaks.

[0021] For coniferous species with resin ducts or resin sacs, resin exudation may occur after polishing, affecting the quality of subsequent image scanning. Preferably, such samples are degreased by immersion in acetone for 24-48 hours, followed by air drying to constant weight under ventilated conditions. This treatment generates processed sample data, laying the foundation for subsequent high-precision annual ring parameter determination. The processed sample data includes metadata information such as sample number, site information, tree species information, sample length, estimated number of annual rings, and processing quality level.

[0022] Step S2: Optical measurement of annual ring width.

[0023] The optical measurement of tree ring width is the core step in obtaining information on the radial growth of tree rings. High-resolution linear array scanning technology and automatic image recognition algorithms are used to achieve efficient and accurate measurement of tree ring width.

[0024] For image acquisition, a high-resolution linear scanner is used to scan and image the surface of the processed tree core samples. In one embodiment of the invention, the scanner uses an EPSON Perfection V850 Pro or a professional scanning device with equivalent performance, and the scanning resolution is set to 2400 dpi, corresponding to a pixel size of approximately 10.6 μm. During scanning, the tree core sample is placed flat on the scanner's glass panel, with the sample's long axis parallel to the scanning direction, ensuring that the annual ring boundaries are perpendicular to the scanning line direction. Scanning is performed in RGB true color mode, with an image bit depth of 24 bits, and the output format is lossless TIFF. To ensure consistent image quality, the scanner's brightness and color are calibrated using a standard grayscale card before each batch of scanning. After scanning, sample surface image data for all samples at each site are obtained, and the image resolution and color characteristics meet the input requirements of subsequent automatic recognition algorithms.

[0025] For annual ring boundary identification, an automatic identification algorithm based on grayscale gradient analysis is used to process the sample surface image and automatically locate the annual ring boundary position. First, the RGB color image is converted to grayscale, and the blue channel grayscale value is extracted as the analysis object, because the absorption characteristics of lignin in the blue light band can better distinguish earlywood and latewood. Then, a grayscale profile curve is extracted along the radial direction of the annual ring, and the grayscale profile is smoothed and filtered to eliminate noise interference. Preferably, a Savitzky-Golay filter is used for smoothing, with a window width of 11 pixels and a polynomial order of 3. On the smoothed grayscale profile, the first derivative is calculated to obtain the grayscale gradient curve. The annual ring boundary corresponds to the negative extreme point of the grayscale gradient, that is, the position of the maximum rate of decrease in grayscale value when transitioning from latewood to earlywood.

[0026] In one embodiment of the present invention, the specific implementation of the automatic annual ring boundary recognition algorithm is as follows: Let the grayscale profile be... ,in The coordinates are radial position coordinates, in pixels. A smooth profile is obtained by applying Savitzky-Golay filtering to the grayscale profile. Calculate the first derivative:

[0027] ,

[0028] in: For position The smoothed grayscale value at the location is dimensionless and ranges from 0 to 255. This represents the distance between adjacent pixels, in pixels, which corresponds to an actual distance of approximately 10.6μm at a resolution of 2400dpi. This represents the grayscale gradient value, expressed in grayscale values ​​per pixel. (Location of tree ring boundaries) The following conditions must be met: grayscale gradient It is a local negative extreme value and its absolute value is greater than the preset gradient threshold. Preferably, a preset gradient threshold is used. Set to 5-15 grayscale values ​​per pixel, and adjust the specific value according to the tree species and image quality.

[0029] In determining the width of tree rings, based on the identified sequence of tree ring boundary positions, the pixel distance between adjacent boundaries is calculated ring by ring, and then converted into an actual width value by combining the scanning resolution. Let the first ring be... The inner boundary of the annual rings is located at... The outer boundary position is Then the first Annual ring width The calculation formula is:

[0030] ,

[0031] in: For the first The width of the annual rings, in mm; and The first The pixel positions of the inner and outer boundaries of the annual rings, in pixels; The scanning resolution is expressed in dpi, and in this embodiment, it is set to 2400 dpi; the constant 25.4 is the conversion factor between inches and millimeters. Following the above method, measurements are taken ring by ring from the heartwood towards the bark, ultimately outputting a complete sequence of annual ring widths. The measurement accuracy can reach 0.01 mm, meeting the requirements for high-precision annual ring studies.

[0032] To improve measurement efficiency and ensure measurement quality, this invention also includes a human-computer interaction verification mechanism. After the automatic recognition algorithm completes, a sample image with superimposed growth ring boundary markers is displayed on the graphical user interface. Operators can check the recognition results one by one and manually add or delete any missed or misidentified growth ring boundaries. Preferably, the software interface provides image browsing functions such as zooming in, zooming out, and panning, as well as a boundary position fine-tuning function, supporting precise adjustments at the pixel level. The interface also provides auxiliary functions such as automatic growth ring numbering, real-time display of measurement data, and historical operation undo / redo, enhancing the convenience of human-computer interaction and operational efficiency.

[0033] Regarding data output, the manually verified annual ring width sequence data is output as the final measurement result. The data format adopts the Tucson or Heidelberg format commonly used in international annual ring databases, facilitating comparison and integration with other annual ring data. Output data includes sample number, annual ring width values, measurement date, measurement personnel, and quality markings. The number and percentage of manually corrected annual rings are also recorded as evaluation indicators of the automatic identification accuracy. In one embodiment of this invention, for Pinus tabuliformis samples, the accuracy of the automatic identification algorithm can reach over 92%, the workload of manual verification is relatively small, and the complete measurement time for a single sample is approximately 10-15 minutes.

[0034] Furthermore, this invention supports a batch measurement mode, which allows multiple tree core samples to be arranged sequentially within the scanner's scanning area. Image data from multiple samples are acquired in a single scan, and the software automatically segments each sample region for ring identification and width measurement. This batch measurement mode significantly improves the efficiency of multi-site, large-sample-volume measurement projects, processing 8-12 standard-length tree core samples in a single batch scan.

[0035] Step S3: Annual ring density determination procedure.

[0036] The annual ring density measurement step is the core step in obtaining annual ring wood density information. It uses two modes of measurement in combination: X-ray micro-density measurement technology and blue light reflectance measurement technology. The former provides an absolute density value, while the latter serves as a rapid alternative.

[0037] In X-ray density determination, a microfocus X-ray scanning system is used to perform transmission scanning on tree core samples to obtain X-ray attenuation profiles and calculate wood density distribution. In one embodiment of the invention, the X-ray scanning system uses an ITRAX multifunction scanner or a device with equivalent performance, a tube voltage range of 20-60kV, a tube current range of 0.1-1.0mA, preferably a tube voltage of 30kV, a tube current of 0.3mA, a focal spot size of no more than 20μm, and a scanning step distance of 10μm to obtain a density profile with high spatial resolution. Before scanning, the tree core sample is cut into slices 1.0-2.0mm thick, with the cut surface parallel to the radial direction of the annual rings. To eliminate interference from resin and other extracts on density determination, the slices are preferably soaked in acetone for 48 hours and then air-dried at room temperature to constant weight.

[0038] The basic principle of X-ray transmission scanning is to estimate the density of wood by utilizing the attenuation characteristics of X-rays as they penetrate a wood sample. Let the intensity of the incident X-rays be... The intensity of the X-rays after transmission is The sample thickness is Then the X-ray attenuation coefficient The relationship between transmission intensity and Beer-Lambert law follows:

[0039] ,

[0040] in: The intensity of transmitted X-rays is expressed in counts per second. The intensity of the incident X-rays is expressed in counts per second. The linear attenuation coefficient for X-rays is expressed in cm⁻¹. Where is the sample thickness, in cm. Taking the logarithm of the above equation yields:

[0041] ,

[0042] X-ray attenuation coefficient and wood density There is a linear relationship between them:

[0043] ,

[0044] in: This refers to the density of wood, expressed in g / cm³, typically ranging from 0.3 to 0.9 for coniferous species. and The calibration coefficients are obtained through calibration using standard samples of known density. In this embodiment, a polymethyl methacrylate (PMMA) stepped wedge is used as the density calibration standard, with a wedge thickness gradient of 0.5 mm and a density of 1.19 g / cm³. A calibration curve of the attenuation coefficient versus density is established through linear regression.

[0045] Based on the above principles, X-ray attenuation profile data along the radial direction are obtained by scanning the tree core thin section sample. Applying a calibration formula to the attenuation profile data yields a continuous density profile curve. Annual ring density parameters, including earlywood density, are extracted from the density profile. Latewood density and maximum density Earlywood density is defined as the average density value of the first half of the annual rings (from the boundary of the annual ring to the point of lowest density), latewood density is defined as the average density value of the second half of the annual rings (from the point of lowest density to the boundary of the next annual ring), and maximum density is defined as the maximum density value within the annual ring. In one embodiment of the present invention, the typical earlywood density of the coniferous species *Pinus tabuliformis* is 0.35-0.45 g / cm³, the typical latewood density is 0.55-0.75 g / cm³, and the typical maximum density is 0.65-0.85 g / cm³.

[0046] In determining blue light reflectance density, the blue light channel reflectance intensity of the optically scanned image is used as a proxy for density. The basic principle of blue light reflectance measurement is that lignin in wood cell walls has a strong absorption characteristic for the blue light band (wavelength approximately 450 nm). The thicker and denser the cell wall, the weaker the blue light reflection; therefore, blue light reflectance intensity is negatively correlated with wood density. The blue channel is extracted from the RGB scanned image obtained in step S2, and a blue light reflectance intensity profile is extracted along the radial direction of the growth rings. To ensure a positive correlation between blue light reflectance intensity and density, the original blue light reflectance values ​​are inverted.

[0047] ,

[0048] in: The original blue light reflection intensity is dimensionless and ranges from 0 to 255. The value represents the inverted blue light intensity; a higher value indicates a higher density. A linear regression was used to establish a calibration relationship between the inverted blue light intensity and the density measured by X-rays.

[0049] ,

[0050] in: Density values ​​estimated by the blue light reflectance method, in g / cm³; and The regression coefficients were obtained by comparing X-ray density with blue light intensity; the regression determination coefficients... A value of at least 0.85 is required to ensure the reliability of the calibration relationship.

[0051] In one embodiment of the present invention, X-ray density and blue light reflectance were simultaneously measured using 20 tree core samples, and a regression equation was established between blue light reflectance intensity and X-ray density. The calibration results showed that the maximum blue light intensity in the latewood was... With maximum X-ray density The correlation coefficient was 0.94, and the regression determination coefficient was... The earlywood blue light intensity is 0.88. With X-ray early wood density The correlation coefficient was 0.93, and the regression determination coefficient was... The value is 0.86. The above calibration relationship indicates that the blue light reflectance method can serve as an effective alternative to X-ray density determination, especially suitable for rapid density estimation of large batches of samples.

[0052] The results of X-ray density measurement and blue light reflectance density measurement are combined to finally output the annual ring density sequence data. This data includes the earlywood density value, latewood density value, and maximum density value of each annual ring, as well as the measurement method identification (X-ray or blue light reflectance). When both methods are used to measure the same sample, the X-ray measurement result is preferred, and the blue light reflectance measurement result is used as a backup and verification.

[0053] Step S4: Multi-parameter cross-dating step.

[0054] The multi-parameter cross-dating step is the core step in determining the absolute age of each tree ring. It utilizes the complementary characteristics of the temporal changes in the tree ring width sequence and density sequence for cross-validation, and identifies and processes missing rings, pseudo-rings, and damaged rings through an abnormal tree ring detection algorithm, which significantly improves the accuracy and reliability of dating.

[0055] Regarding the principle of cross-dating, traditional tree-ring cross-dating mainly relies on the temporal pattern matching of tree-ring width sequences. This involves comparing the target sequence with the already dated master sequence, utilizing the synchronous response characteristics of tree-ring width to climate change in the same region and for the same tree species. However, single-parameter dating has limitations; misjudgments are prone to occur when tree-ring width changes are not significant or when abnormal tree-rings appear. This invention introduces a joint cross-dating strategy using tree-ring density and width parameters. This strategy leverages the differences and complementarities in the response characteristics of these two parameters to climatic factors to improve dating accuracy. Specifically, tree-ring width is primarily controlled by moisture conditions in the early growing season, while maximum tree-ring density is mainly controlled by temperature conditions in the late growing season. The climate-sensitive periods of these two parameters differ, and the distribution of characteristic years (wide or narrow years, high-density or low-density years) in their time series also differs. Using them together provides more time anchors for sequence matching.

[0056] In terms of cross-dating, the annual ring width and density sequences of each sample at each site are first standardized to remove the influence of tree growth trends and retain high-frequency climate signals. Standardization methods employ spline function detrending or negative exponential curve fitting for detrending. Then, the correlation coefficient and t-statistic between the sample sequence and the main sequence at each site are calculated to evaluate the matching quality. In one embodiment of this invention, the standardized width sequence is set as follows: The normalized maximum density sequence is The parameters corresponding to the main sequence are and Then the overall matching coefficient The calculation formula is:

[0057] ,

[0058] in: is the width series correlation coefficient, dimensionless, with a value range of -1 to 1; is the density sequence correlation coefficient, dimensionless, with a value range of -1 to 1; This is a weighting coefficient, dimensionless, ranging from 0 to 1, with a preferred value of 0.5 indicating equal weighted contributions from width and density; This is a comprehensive matching coefficient used to evaluate the overall matching quality. When If the value is greater than the preset matching threshold (e.g., 0.4) and both sub-coefficients are positive, the match is considered successful, and the absolute age of the sample sequence is determined.

[0059] In the detection of abnormal tree rings, a specialized detection algorithm was designed to identify three types of abnormal tree rings: missing rings, pseudo-rings, and damaged rings. Missing rings refer to the absence of rings in some or all directions in certain years due to extremely limited or stopped tree growth. Pseudo-rings refer to false ring boundaries caused by short-term stress (such as drought) during the growing season. Damaged rings refer to abnormal ring structures caused by mechanical damage, pests, or diseases. The core idea of ​​the abnormal tree ring detection algorithm is to utilize the collaborative information of two parameters: width and density. When the change patterns of these two parameters are inconsistent, an anomaly marker is triggered.

[0060] Specifically, the detection condition for suspected pseudo-rings is: the rate of change in width between adjacent annual rings. Exceeding the width change threshold (e.g., 0.5), but the corresponding density change rate Density change threshold not exceeded (e.g., 0.3). This situation indicates an abnormal abrupt change in the width of the growth rings while the density remains continuous, which may be due to errors in growth ring division caused by pseudo-rings.

[0061] The suspected missing ring detection condition is: the width sequence is discontinuous at a certain position (measured value of 0 or unmeasurable), but the density sequence has a valid density value at the corresponding position and the density value changes continuously. This situation indicates that the growth ring may be missing in the width measurement direction, but can still be identified in the density measurement direction, suggesting that supplementary measurements should be made in other directions or that the density sequence should be used to determine the age of this year.

[0062] The conditions for detecting damaged rings are: both the width and density measurements show significant abnormal deviations at the same growth ring location, specifically, the width or density value exceeds three standard deviations above the mean of the sample sequence, and the abnormal location is not synchronous in adjacent samples. This indicates that the growth ring may be affected by local damage or measurement error, requiring labeling and special handling in subsequent analysis.

[0063] For detected abnormal tree rings, abnormal tree ring labeling data is generated, including the abnormality type (missing ring, pseudo-ring, or damaged ring), the abnormality location (tree ring number), and processing suggestions (supplementary measurement, manual verification, or removal). The format design of the abnormal tree ring labeling data takes into account the correlation with the dating results data, and uses the tree ring number as the primary key for indexing, which facilitates the rapid location and processing of abnormal tree rings in subsequent analysis.

[0064] Regarding the specific implementation process of cross-dating, this invention adopts a step-by-step approach. First, preliminary cross-dating is performed between samples within the same site to establish a relative age sequence within the site. Then, the site sequence is compared and matched with the regional master sequence to determine the absolute age. Finally, the density sequence is used to verify and correct the dating results of the width sequence. This step-by-step approach effectively reduces the cumulative effect of dating errors and improves the reliability of the final dating results.

[0065] In one embodiment of the present invention, a cross-dating test was conducted on 20 Pinus tabuliformis samples from a certain site. The results showed that when using only width sequence dating, the initial dating success rate was 85%, and 15% of the samples required manual intervention. After adopting combined width and density dating, the initial dating success rate increased to 95%, and the number of samples requiring manual intervention decreased to 5%. This test result fully verifies the technical effectiveness of the multi-parameter cross-dating strategy in improving dating accuracy.

[0066] The dating results are corrected based on anomalous tree-ring markers, and the final dating data is output. This data includes the absolute age of each tree ring, the dating confidence level, and anomalous marker information. The dating confidence level is comprehensively evaluated based on the overall matching coefficient and the number of anomalous tree rings, and is divided into three levels: high, medium, and low. High confidence level indicates that the overall matching coefficient is greater than 0.6 and there are no unprocessed anomalous tree rings; medium confidence level indicates that the overall matching coefficient is between 0.4 and 0.6 or that there are processed anomalous tree rings; low confidence level indicates that the overall matching coefficient is less than 0.4 or that there are unprocessed anomalous tree rings, requiring further manual verification.

[0067] Step S5: Measurement data quality assessment steps.

[0068] The measurement data quality assessment step is a key step in ensuring the reliability of measurement data. The statistical quality of the measurement sequence is quantitatively evaluated by calculating three indicators: signal-to-noise ratio, inter-sample correlation coefficient, and expression population signal intensity.

[0069] The signal-to-noise ratio (SNR) is used to evaluate the relative intensity of effective climate signals and noise in a single sequence. It is calculated as the ratio of the portion of the sequence variance explained by common climate signals to the total variance. Suppose a certain station has... There are 10 samples, and the standardized tree-ring parameter sequence for each sample is as follows: ( The station mean sequence is Then the signal-to-noise ratio The calculation formula is:

[0070] ,

[0071] in: Signal-to-noise ratio (SNR) is dimensionless and ranges from 0 to positive infinity. In practical applications, a value in the range of 0.7-1.0 is generally considered good. The number of samples is expressed in units of individuals; in this embodiment, each site has 15-25 samples. The average correlation coefficient between samples is given below. A higher signal-to-noise ratio indicates a stronger common climate signal within the sequence and better data quality.

[0072] Correlation coefficient between samples ( The correlation index is used to evaluate the degree of correlation among samples at the same site, reflecting the consistency of the samples' responses to common environmental signals. It is calculated as the average of the pairwise correlation coefficients among all samples.

[0073] ,

[0074] in: The average correlation coefficient between samples is dimensionless and ranges from 0 to 1. For the first The sample and the first Pearson correlation coefficient between samples; This represents the number of samples. A higher value indicates a more consistent response among samples to common environmental signals, resulting in better repeatability of the measurement results. Preferably, A value greater than 0.3 is required to be considered as indicating good inter-sample consistency of the site sequence.

[0075] The Expressed Population Signal Strength (EPS) index is the most commonly used quality assessment index in dendroclimatology research. It is used to evaluate the degree to which a finite sample size of station sequences can represent a hypothetical infinitely large population sequence. The formula for calculating EPS is:

[0076] ,

[0077] in: To express the intensity of the group signal, it is dimensionless and ranges from 0 to 1; and The meaning is the same as before. When At that time, it is generally considered that the station sequence meets the quality requirements for paleoclimate reconstruction applications and can reliably represent regional climate signals. In one embodiment of the present invention, a configuration is set... The threshold is 0.85. When a certain site's If the sample size is below this threshold, it will be marked as insufficient in the quality assessment report, and it will be recommended to increase the number of samples.

[0078] Based on the above three indicators, the measurement sequences of each parameter at each station are assessed for quality, and a quality assessment report is generated. The report includes the sample size, sequence length, values ​​of each quality indicator, quality level determination, and improvement suggestions for each station. The quality level is divided into three levels: Level A indicates that all three indicators meet the threshold requirements, and the data can be directly used for climate reconstruction; Level B indicates that some indicators meet the threshold, and the data can be used for auxiliary analysis but should be interpreted with caution; Level C indicates that key indicators do not meet the threshold, and the data requires further processing or an increase in sample size.

[0079] Regarding the time window analysis for quality assessment, this invention also designs a sliding window assessment mechanism. Since the quality of tree-ring sequences may change over time, quality indicators are typically poor in the early stages of the sequence when sample depth is low, while gradually improving as sample depth increases in the later stages. The sliding window assessment mechanism uses a fixed window width (e.g., 50 years) sliding across the entire sequence length, calculating each quality indicator in segments to identify reliable periods where quality indicators consistently meet standards. Preferably, only tree-ring parameter data within these reliable periods are included in subsequent climate reconstruction analyses to ensure the reliability of the reconstruction results.

[0080] In one embodiment of this invention, the quality of tree-ring width and maximum density sequences from eight stations was assessed. The assessment results showed that the tree-ring width sequences from six stations met the Class A quality standard, and two stations met the Class B standard; the maximum density sequences from seven stations met the Class A quality standard, and one station met the Class B standard. Overall, the density sequence quality was slightly better than the width sequence, which is consistent with the characteristic that density parameters have a stronger common climate signal. For stations with Class B quality, it is recommended to improve data quality through supplementary sampling or optimized sample selection.

[0081] The quality assessment report also includes a comparative analysis of quality among stations and a spatiotemporal quality distribution map. The comparative analysis among stations displays the ranking of quality indicators for each parameter at each station in tabular form, facilitating the identification of stations with poor quality for targeted improvement. The spatiotemporal quality distribution map shows the quality level of each station in a spatial distribution format, revealing the spatial differentiation pattern of annual ring data quality within the watershed and providing a weighting basis for subsequent spatial interpolation analysis.

[0082] Step S6: Spatial distribution analysis step.

[0083] The spatial distribution analysis step is a key step in expanding the measurement data of discrete sampling points into a continuous distribution field of the watershed. Through spatial interpolation and climate sensitivity analysis, the watershed annual ring parameter distribution map and climate sensitivity spatial distribution map are output.

[0084] For spatial interpolation, geostatistical methods are used to interpolate the annual ring parameter measurements from each sampling station to the overall spatial range of the watershed. Preferably, the spatial interpolation method employs either Ordinary Kriging or Inverse Distance Weighted (IDW) interpolation. Kriging interpolation has the advantage of providing an estimate of interpolation uncertainty and is suitable for situations where stations are relatively evenly distributed; IDW interpolation is computationally simple and provides intuitive results, making it suitable for situations where stations are unevenly distributed.

[0085] Taking Kriging interpolation as an example, suppose there are [missing information] within the watershed. The sampling site, the first The location coordinates of each station are The corresponding average annual ring parameter is For any point to be interpolated within the watershed. Its estimated annual ring parameters The calculation formula is:

[0086] ,

[0087] in: These are the estimated values ​​of the annual ring parameters for the interpolation points, with the same units as the original parameters (mm for width, g / cm³ for density). For the first Measured parameter values ​​for each site; For the first The interpolation weights for each station are obtained by solving the variation function model and satisfy the constraints. The variogram describes how spatial correlation varies with distance. Commonly used models include the spherical model, the exponential model, and the Gaussian model. The optimal model and parameters are determined by analyzing the variogram of measured data.

[0088] The interpolation grid resolution is adaptively set based on the watershed area and sampling station density. In one embodiment of the invention, for an area of ​​approximately 5000 km²... 2 The watershed was sampled at 8 stations, with an interpolation grid resolution of 1km × 1km, generating approximately 5000 grid points. After interpolation, continuous distribution field data of the watershed was obtained, including the location coordinates of each grid point, the average annual ring width, the average annual ring density, and the interpolation standard error.

[0089] In climate sensitivity analysis, the correlation coefficients between the tree-ring parameter sequences and climate factor time series at each sampling station are calculated to evaluate the sensitivity of the tree-ring parameters to different climate factors. Climate factors include annual precipitation (P), growing season mean temperature (T), and growing season accumulated temperature (GDD), etc. Let the standardized tree-ring parameter sequence of a certain station be... The climate factor sequence is The sequence length is In which year, the climate sensitivity coefficient The calculation formula is:

[0090] ,

[0091] in: Pearson correlation coefficient, dimensionless, ranging from -1 to 1; positive values ​​indicate positive correlation (the parameter increases with increasing climate factors), and negative values ​​indicate negative correlation; and These are the mean values ​​of the annual ring parameter series and the climate factor series, respectively. Sensitivity coefficients were calculated for each climate factor at each station to obtain climate sensitivity coefficient data.

[0092] Based on the climate sensitivity coefficients of each station, a spatial distribution map of climate sensitivity was generated using the same spatial interpolation method as that used for tree-ring parameters. This map, presented as contour lines or color scales, displays the spatial distribution pattern of tree-ring parameters' sensitivity to specific climate factors within the watershed, revealing the spatial differentiation patterns of climate sensitivity within the watershed. The spatial distribution map of climate sensitivity can visually demonstrate which areas within the watershed are more sensitive to precipitation and which are more sensitive to temperature, providing a scientific basis for regionally differentiated paleoclimate reconstruction strategies.

[0093] In terms of visualizing the distribution map, this invention employs a multi-layered approach. The base layer displays the topographic elevation data of the watershed, represented by mountain shadows or contour lines. The data layer displays the interpolation results with a semi-transparent color gradient overlay, using warm colors (red-orange-yellow) for positively correlated areas, cool colors (blue-cyan-green) for negatively correlated areas, and neutral gray for areas with insignificant correlations. The label layer displays the location and correlation coefficient values ​​of each sampling station, as well as cartographic elements such as legends, scale bars, and direction indicators. This multi-layered visualization approach ensures both the complete presentation of data information and an intuitive display of the geospatial background.

[0094] In one embodiment of this invention, the final output includes a series of products such as a tree ring width-precipitation sensitivity distribution map, a tree ring width-temperature sensitivity distribution map, and a maximum density-temperature sensitivity distribution map. Analysis results show that the sensitivity of tree ring width to precipitation within the watershed exhibits a significant spatial gradient, with stronger sensitivity in low-altitude arid areas (correlation coefficient 0.4-0.6) and weaker sensitivity in high-altitude humid areas (correlation coefficient 0.1-0.3). The maximum density's sensitivity to temperature is strongest in high-altitude areas (correlation coefficient 0.5-0.7) and relatively weaker in low-altitude areas (correlation coefficient 0.3-0.4). These spatial distribution characteristics are consistent with the spatial variation patterns of regional climate gradients and tree growth limiting factors, verifying the scientific rationality of the method of this invention.

[0095] Furthermore, this invention supports comprehensive sensitivity analysis of multiple climate factors. By calculating the partial correlation coefficients between tree-ring parameters and multiple climate factors, the independent impact of a single factor can be assessed while controlling for the influence of other factors. The comprehensive sensitivity analysis results are output in the form of a multivariate climate sensitivity composite map, using pie charts or radar charts to display the relative contributions of each climate factor at each site location, providing data support for multifactor-driven paleoclimate reconstruction. These systematic data products provide a rich source of information and an analytical foundation for watershed paleoclimate reconstruction and ecological environment assessment.

[0096] like Figure 2 As shown, this embodiment provides a watershed multi-site tree ring width and density joint measurement system for implementing the above method. The system includes six functional modules: sample collection and preprocessing module 1, tree ring width optical measurement module 2, tree ring density measurement module 3, multi-parameter cross-dating module 4, measurement data quality assessment module 5, and spatial distribution analysis module 6. The modules are connected through data interfaces to form a complete measurement data processing flow.

[0097] The sample collection and preprocessing module 1 is used to acquire the location data of multiple sampling stations within the watershed and to perform surface polishing and fixation processing on the tree core samples collected from each sampling station. This module includes a GPS positioning unit, a growth cone sampling tool set, a sample fixing device, and a sandpaper polishing assembly. The GPS positioning unit uses a high-precision differential GPS device with a positioning accuracy better than 1m, used to record the latitude, longitude, and elevation coordinates of each sampling station. The growth cone sampling tool set includes growth cones of different specifications and matching wrenches, suitable for tree species with different diameters at breast height (DBH) and hardness. The sample fixing device includes a dedicated wooden groove, positioning clamps, and wood glue to ensure the stability of the tree core samples during polishing and measurement. The sandpaper polishing assembly includes a multi-grade sandpaper sequence and a polishing operating table, supporting both manual and electric polishing methods. The output of this module is the processed tree core sample and its location information data.

[0098] The optical measurement module 2 for tree ring width includes a high-resolution linear array scanner and an image processing unit. It performs surface scanning on the processed sample to acquire surface image data, identifies tree ring boundaries using an image grayscale gradient analysis algorithm, and outputs a sequence of tree ring width data. The high-resolution linear array scanner uses a CCD or CIS linear array sensor with a scanning resolution of at least 2400 dpi, and the scanning area covers the length of a standard tree core sample. The image processing unit runs dedicated image analysis software to perform functions such as grayscale conversion, filtering and smoothing, gradient calculation, boundary detection, and width measurement. It supports both automatic recognition and human-computer interactive verification modes. The input to this module is the processed sample data, and the output is the sample surface image data and the tree ring width sequence data.

[0099] The tree ring density measurement module 3 includes a microfocus X-ray scanning unit and a blue light reflectance measurement unit. These units acquire X-ray attenuation profile data and blue light reflectance intensity data, and calculate and output tree ring density sequence data including earlywood density, latewood density, and maximum density. The microfocus X-ray scanning unit uses a microfocus X-ray tube and a high-sensitivity detector array, with a focal size no larger than 20 μm, supporting high-resolution scanning in 10 μm steps. The blue light reflectance measurement unit reuses the scanning image data from the tree ring width optical measurement module, extracting the blue channel and performing density calibration conversion via software. This module is equipped with calibration wedges and a calibration coefficient database, supporting density calibration for different tree species and equipment. The output of this module is tree ring density sequence data, with a data format compatible with tree ring width sequence data, facilitating subsequent joint analysis.

[0100] The multi-parameter cross-dating module 4 is connected to the optical measurement module 2 for tree ring width and the measurement module 3 for tree ring density. It performs cross-validation, anomaly detection, and absolute dating based on tree ring width and density sequence data, outputting dating results and anomaly marker data. This module runs cross-dating algorithm software, performing functions such as sequence standardization, correlation coefficient calculation, comprehensive matching coefficient calculation, anomaly detection, and dating. The software includes a built-in regional master sequence database, supporting batch comparisons with multiple dated master sequences. The module also provides a graphical user interface, displaying the matching status of the sequence to be tested and the master sequences in line graph form, supporting manual judgment and correction by operators.

[0101] The measurement data quality assessment module 5 is connected to the multi-parameter cross-dating module 4 and is used to calculate the signal-to-noise ratio, inter-sample correlation coefficient, and expressed population signal strength, outputting quality assessment report data. This module runs a quality assessment algorithm program, calculates various quality indicators according to the formulas described in the method embodiment, and determines the quality level based on preset thresholds. The assessment report is presented in tabular and graphical form, including the quality indicator values ​​for each parameter at each site, the level determination results, and improvement suggestions. The report supports exporting to PDF or Excel format for easy archiving and sharing.

[0102] The spatial distribution analysis module 6 is connected to the measurement data quality assessment module 5. It is used to perform spatial interpolation on multi-site measurement data to generate continuous watershed distribution field data, calculate climate sensitivity coefficients, and output a spatial distribution map of climate sensitivity. This module integrates GIS spatial analysis functions, supports various spatial interpolation algorithms such as Kriging interpolation and IDW interpolation, and automatically generates variogram analysis results and interpolation error estimates. The climate sensitivity analysis function supports importing measured data from meteorological stations or gridded reanalysis data to calculate the correlation coefficient matrix between annual ring parameters and climate factors. The distribution map output function supports generating GeoTiff raster format, Shapefile vector format, and PNG image format to meet different application needs.

[0103] In summary, this invention, by constructing a multi-site batch measurement framework for watersheds, employing a strategy of joint measurement of width and density and cross-dating, and combining it with a three-dimensional quality assessment system and spatial distribution analysis function, has achieved a technological leap from single-site single-parameter measurement to watershed-scale multi-parameter joint measurement, providing a systematic measurement method and data products for watershed paleoclimate reconstruction and ecological environment assessment.

[0104] The above description is only a preferred embodiment of the present invention. Any modifications, equivalent substitutions, and improvements 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 joint determination of tree ring width and density at multiple stations in a watershed, characterized in that, The method includes: The sample collection and preprocessing steps involve obtaining the location data of multiple sampling stations within the watershed, and performing surface polishing and fixation on the tree core samples collected from each sampling station to generate processed sample data. The optical measurement procedure for tree ring width involves using a high-resolution linear array scanner to scan the surface of the processed samples at each site, acquiring sample surface image data, automatically identifying the tree ring boundary positions through an image grayscale gradient analysis algorithm, measuring and outputting tree ring width sequence data for each ring. The steps for determining annual ring density involve using a microfocus X-ray scanning system to perform transmission scanning on the processed samples at each site to obtain X-ray attenuation profile data. Based on the quantitative relationship between the X-ray attenuation coefficient and wood density, the earlywood density, latewood density, and maximum density values ​​are calculated. Simultaneously, the blue light reflectance measurement mode is used to obtain blue light reflectance intensity data, which is converted into a density sequence according to a preset calibration relationship. The annual ring density sequence data is then integrated and output. The multi-parameter cross-dating step utilizes the complementary temporal variation features of the tree ring width sequence data and the tree ring density sequence data for cross-validation. An abnormal tree ring detection algorithm is used to identify missing rings, pseudo-rings, and damaged rings and generate abnormal tree ring marker data. Based on the collaborative matching of the width sequence and the density sequence, the absolute age of each tree ring is determined, and the dating result data is output. The data quality assessment steps involve calculating the signal-to-noise ratio, inter-sample correlation coefficient, and expression population signal intensity values ​​for the dating results data from each site, comprehensively evaluating the statistical quality of the measurement sequence, and outputting a quality assessment report. The spatial distribution analysis step involves spatial interpolation of multi-site measurement data within the watershed to generate continuous distribution field data for the watershed, calculating the climate sensitivity coefficient data of annual ring parameters and climate factors for each station, and outputting a spatial distribution map of climate sensitivity.

2. The method for joint determination of tree ring width and density at multiple stations in a watershed according to claim 1, characterized in that, In the optical measurement step of the growth ring width, the scanning resolution of the high-resolution linear array scanner is not less than 2400 dpi, and the measurement accuracy of the growth ring width sequence data reaches a preset width measurement accuracy threshold, which is 0.01 mm.

3. The method for jointly determining tree ring width and density at multiple stations in a watershed according to claim 1, characterized in that, In the step of determining the annual ring density, the focal size of the microfocus X-ray scanning system is no greater than 20 μm, the scanning step distance is 10 μm, the tube voltage range is 20-60 kV, and the tube current range is 0.1-1.0 mA.

4. The method for joint determination of tree ring width and density at multiple stations in a watershed according to claim 1, characterized in that, In the annual ring density measurement step, the preset calibration relationship is established through regression analysis of blue light reflection intensity and X-ray density measurement results, with a regression determination coefficient of not less than 0.

85.

5. The method for joint determination of tree ring width and density at multiple stations in a watershed according to claim 1, characterized in that, In the data quality assessment step, when the signal-to-noise ratio is greater than the preset quality qualification threshold and the expression population signal strength is greater than 0.85, the dating result data is determined to meet the quality requirements for watershed climate reconstruction applications. The preset quality qualification threshold is 0.

85.

6. The method for joint determination of tree ring width and density at multiple stations in a watershed according to claim 1, characterized in that, In the multi-parameter cross-dating step, the abnormal tree ring detection algorithm includes: When the rate of change of width of adjacent annual rings exceeds the width change threshold but the corresponding rate of change of density does not exceed the density change threshold, it is marked as a suspected pseudo-ring; When the tree ring width sequence data shows consecutive missing values ​​and the tree ring density sequence data has a valid density value at the corresponding position, it is marked as a suspected missing ring. When both the width and density measurements show significant abnormal deviations, the wheel is marked as damaged.

7. The method for joint determination of tree ring width and density at multiple stations in a watershed according to claim 1, characterized in that, In the spatial distribution analysis step, the spatial interpolation process uses the Kriging interpolation algorithm or the inverse distance weighted interpolation algorithm, and the interpolation grid resolution is adaptively set according to the watershed area and the sampling station density.

8. The method for jointly determining tree ring width and density at multiple stations in a watershed according to claim 1, characterized in that, In the spatial distribution analysis step, the climate sensitivity coefficient data is obtained by calculating the Pearson correlation coefficient between the annual ring parameter series of each station and the time series of climate factors. The climate factors include annual precipitation, average temperature during the growing season, and effective accumulated temperature during the growing season.

9. The method for jointly determining tree ring width and density at multiple stations in a watershed according to claim 1, characterized in that, In the sample collection and preprocessing steps, the number of sampling stations in the watershed shall not be less than 3, and the number of tree core samples collected at a single station shall not be less than 15. The surface polishing treatment shall be carried out by polishing with a sandpaper sequence of decreasing grit size until the cell outline is visible.

10. A system for jointly measuring tree ring width and density at multiple stations in a watershed, used to implement the method for jointly measuring tree ring width and density at multiple stations in a watershed as described in any one of claims 1 to 9, characterized in that, The system includes: The sample collection and preprocessing module is used to acquire the location data of multiple sampling stations in the watershed, and to perform surface polishing and fixation processing on the tree core samples collected from each sampling station. The optical measurement module for annual ring width includes a high-resolution linear array scanner and an image processing unit, which is used to perform surface scanning on the processed sample to obtain sample surface image data, identify annual ring boundaries through an image grayscale gradient analysis algorithm, and output annual ring width sequence data. The tree ring density measurement module includes a microfocus X-ray scanning unit and a blue light reflectance measurement unit, which are used to acquire X-ray attenuation profile data and blue light reflectance intensity data, and calculate and output tree ring density sequence data including earlywood density, latewood density and maximum density. A multi-parameter cross-dating module is connected to the optical measurement module for tree ring width and the measurement module for tree ring density. It is used to perform cross-validation, abnormal tree ring detection and absolute age determination based on tree ring width sequence data and tree ring density sequence data, and outputs dating result data and abnormal tree ring marker data. The measurement data quality assessment module is connected to the multi-parameter cross-dating module and is used to calculate the signal-to-noise ratio index, the correlation coefficient between samples and the signal strength of the expressed population, and output the quality assessment report data. The spatial distribution analysis module, connected to the measurement data quality assessment module, is used to perform spatial interpolation on multi-site measurement data to generate continuous distribution field data of the watershed, calculate the climate sensitivity coefficient, and output a spatial distribution map of climate sensitivity.

Images