Forest carbon sink climate change-based space-time pattern prediction method

Abstract

The invention discloses a space-time pattern prediction method based on forest carbon sink climate change. The method comprises the following steps: collecting vegetation data, soil data, meteorological data, topographic data and special interference factor data of a target forest area; preprocessing the collected multi-source data, and performing space-time matching on the preprocessed data; quantifying a climate index of the target forest region; taking a Biome-BGC ecological process model as a core, and integrating a graph neural network to construct a carbon sink-climate index-terrain coupling model; setting various climatic change scenes, and predicting a carbon sink spatial and temporal distribution pattern under different scenes based on a carbon sink-climatic index-terrain coupling model. The space-time pattern prediction method based on the forest carbon sink climate change is adopted, and the prediction precision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of spatiotemporal pattern prediction, in particular to a spatiotemporal pattern prediction method based on forest carbon sink climate change. BACKGROUND

[0002] As the core carbon pool of terrestrial ecosystems, the carbon sink function of forests plays an important role in mitigating global climate change. However, the carbon sink process of forests is influenced by multiple factors such as climate change, topographic conditions, soil properties, and human activities, and exhibits high spatiotemporal heterogeneity and dynamics. Therefore, accurately predicting the spatiotemporal evolution pattern of forest carbon sinks under the background of climate change is of great significance for formulating regional ecological management policies.

[0003] As the "Asian Water Tower", the forest ecosystem of the Tibetan Plateau is a core component of the plateau carbon cycle and is sensitive to global climate change. However, the existing spatiotemporal pattern prediction methods for forest carbon sinks have significant shortcomings when applied to the plateau region: first, they do not fully consider the extreme climate characteristics of the plateau (extreme low temperature in winter, widespread permafrost), and the adaptability of climate factors is poor; second, existing data-driven models (such as single neural networks) rely heavily on meteorological and vegetation data, ignoring human disturbance factors such as road construction. SUMMARY

[0004] The purpose of the present application is to provide a spatiotemporal pattern prediction method based on forest carbon sink climate change to solve the problems mentioned in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides a spatiotemporal pattern prediction method based on forest carbon sink climate change, comprising the steps of: S1, collecting vegetation data, soil data, meteorological data, topographic data, and special disturbance factor data of the target forest region; S2, preprocessing the collected multi-source data and spatiotemporally matching the preprocessed data; S3, quantifying the climate index of the target forest region; S4, integrating a Biome-BGC ecological process model as the core and constructing a carbon sink-climate index-topography coupling model using a graph neural network; S5, setting multiple climate change scenarios and predicting the spatiotemporal distribution pattern of carbon sinks under different scenarios based on the carbon sink-climate index-topography coupling model.

[0006] Preferably, the vegetation data in step S1 includes vegetation type, forest age and physi-ecological parameters; the soil data includes soil type and soil texture; the meteorological data includes extreme temperature, daily average temperature, daily maximum temperature, daily minimum temperature, monthly average temperature, annual accumulated temperature, precipitation; the topographic data includes landform type, elevation, slope; the special interference factor data includes human interference factors and road construction land occupation.

[0007] Preferably, the preprocessing in step S2 includes missing value filling, noise filtering and standardization.

[0008] Preferably, the climate index of the target forest region in step S3 includes: an abnormal index of winter extreme low temperature, the formula of which is: ; wherein, represents the average temperature of the corresponding month, represents a low temperature critical value; an abnormal index of strong radiation in the growing season, the formula of which is: ; wherein, is the monthly average UV-B radiation intensity, is a radiation critical value; an abnormal index of frozen soil freezing and thawing, the formula of which is: ; wherein, is the thickness of the active layer in the current year, is the average thickness of the active layer in multiple years, a positive value indicating that the freezing and thawing abnormality is enhanced; an abnormal index of precipitation concentration period in the growing season, the formula of which is: ; wherein, = 1 when the monthly precipitation is greater than 80 mm, = 0; an abnormal index of annual accumulated temperature, the formula of which is: .

[0009] Preferably, step S4 specifically includes: S41, inputting the spatio-temporally matched data into a Biome-BGC model for processing to obtain ecosystem carbon flux parameters; S42, constructing a graph structure based on the carbon flux parameters, climate indexes, special interference factor data and topographic data; S43, constructing a carbon sink-climate index-topography coupling model by taking the graph structure as input.

[0010] Preferably, step S42 specifically includes: The target forest region is divided into spatial grids, with each grid serving as a node in a graph structure. v A weighted approach is used to construct a feature fusion vector for each node based on carbon flux parameters, climate indices, special disturbance factor data, and terrain data; Define the edges connecting nodes and calculate the weight for each edge, specifically: For the grid Establish edge connections with adjacent grids, if the grid non-adjacent mesh Feature cosine similarity , Indicate the threshold and supplement the edge connections; The weight of each edge is calculated using the following formula: ; In the formula, Indicates the distance attenuation coefficient. For nodes With nodes Spatial Euclidean distance.

[0011] Preferably, step S43 specifically includes: constructing a carbon sink-climate index-topography coupling model using a graph convolutional neural network and a fully connected layer, capturing spatial dependencies through the graph convolutional layer, and fitting the nonlinear mapping between carbon sink and multiple factors through the fully connected layer.

[0012] Preferably, the loss function of the carbon sink-climate index-topography coupling model is: ; In the formula, Indicates the first i Measured carbon sequestration data for each node, Spatial regularization coefficient, This represents the spatial smoothing term.

[0013] Preferably, multiple climate change scenarios are set up. For any climate change scenario, the grid carbon sink amount for different future periods is obtained based on the carbon sink-climate index-topography coupling model. The annual carbon sink change rate of each grid under the scenario is calculated, and the carbon sink value of each grid under the scenario is extracted. A carbon sink time pattern distribution map is generated based on the grid carbon sink value at each time step.

[0014] Therefore, the present invention employs the above-mentioned method for predicting the spatiotemporal pattern of forest carbon sink climate change, which has the following beneficial effects: (1) Five types of plateau-specific climate indices, such as extreme low temperatures in winter and freeze-thaw cycles of permafrost, are specifically quantified to address the problem of poor adaptability of climate factors in existing methods and improve the accuracy of predictions. (2) Considering special interference factors, meteorological and topographic factors, etc., the error caused by spatiotemporal heterogeneity is reduced and the prediction accuracy is improved by coupling the Biome-BGC ecological mechanism model with the graph neural network.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] Figure 1 This is a flowchart of a method according to an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will be presented in summary. The technical solutions of the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0018] Example like Figure 1 As shown, this invention provides a method for predicting the spatiotemporal patterns of forest carbon sink climate change, including the following steps: S1. Collect vegetation data, soil data, meteorological data, topographic data, and data on special interference factors in the target forest area.

[0019] Taking the Tibetan Plateau region as an example, vegetation data includes vegetation type, forest age, and physiological and ecological parameters. Vegetation types include Tibetan longleaf pine, longleaf spruce, and fir, while physiological and ecological parameters include maximum photosynthetic rate and leaf area index. Soil data includes soil type and texture, such as loam. Meteorological data includes extreme temperatures, daily average temperature, daily maximum temperature, daily minimum temperature, monthly average temperature, annual accumulated temperature, and precipitation. Topographic data includes landform type, altitude, and slope. Data on special disturbance factors includes human disturbance factors and land occupation for road construction, such as grazing, logging, and railway land occupation.

[0020] S2. Preprocess the collected multi-source data and perform spatiotemporal matching on the preprocessed data. The preprocessing includes missing value imputation, noise filtering, and standardization.

[0021] S3. Quantify the climate index of the target forest area. Specifically: Climate indices for the target forest area include: The Winter Extreme Low Temperature Anomaly Index is used to calculate the cumulative deviation of the average temperatures in December, January, and February from a critical value. The formula is: ; In the formula, This represents the average temperature for the corresponding month. Indicates the critical value for low temperature; The growing season strong radiation anomaly index is used to calculate radiation data from June to August. The formula is: ; In the formula, The monthly average UV-B radiation intensity, This is the critical value for radiation. The frozen soil freeze-thaw anomaly index is calculated using the following formula: ; In the formula, The thickness of the active layer in that year, This represents the multi-year average thickness of the active layer; a positive value indicates an enhanced freeze-thaw anomaly. The abnormal index of concentrated rainfall during the growing season is calculated using the following formula: ; In the formula, =1, when the monthly precipitation is greater than 80mm. =0; The annual accumulated temperature anomaly index is calculated using the following formula: .

[0022] S4. Using the Biome-BGC ecological process model as the core, a carbon sink-climate index-topography coupled model is constructed by integrating graph neural networks. The specific process includes: S41. Input the spatiotemporally matched data into the Biome-BGC model for processing to obtain ecosystem carbon flux parameters. S42. Construct a map structure based on carbon flux parameters, climate indices, data on special disturbance factors, and topographic data.

[0023] Step S42 specifically includes: The target forest area is divided into 1km sections. A 1km spatial grid is used, with each grid cell serving as a node in a graph structure. v A weighted approach is used to construct a feature fusion vector for each node based on carbon flux parameters, climate indices, special disturbance factor data, and terrain data; Define the edges connecting nodes and calculate the weight for each edge, specifically: For the grid Establish edge connections with adjacent grids, if the grid non-adjacent mesh Feature cosine similarity , Indicate the threshold and supplement the edge connections; The weight of each edge is calculated using the following formula: ; In the formula, This represents the distance attenuation coefficient, ranging from 0.1 to 0.3. For nodes With nodes Spatial Euclidean distance.

[0024] S43. Using a graph structure as input, construct a carbon sink-climate index-topography coupling model. Specifically, a graph convolutional neural network and two fully connected layers are used to construct the carbon sink-climate index-topography coupling model. The graph convolutional layers capture spatial dependencies, and the fully connected layers fit the nonlinear mapping between carbon sinks and multiple factors.

[0025] The loss function of the carbon sink-climate index-topography coupling model is: ; In the formula, Indicates the first i Measured carbon sequestration data for each node, Spatial regularization coefficient, This represents the spatial smoothing term.

[0026] S5. Set up multiple climate change scenarios and predict the spatiotemporal distribution pattern of carbon sinks under different scenarios based on the carbon sink-climate index-topography coupling model.

[0027] Various climate change scenarios are shown in Table 1.

[0028] Table 1. Various Climate Change Scenarios

[0029] For any climate change scenario, the carbon sink amount for different future time periods is obtained based on the carbon sink-climate index-topography coupled model. The annual carbon sink change rate for each grid under the scenario is calculated, and the carbon sink value for each grid under the scenario is extracted. A carbon sink temporal pattern distribution map is generated based on the grid carbon sink values ​​at each time step. The carbon sink amount calculation formula is as follows: ; In the formula, For the first time in this scenario The node t The feature fusion vector for each year.

[0030] The formula for calculating the annual rate of change of carbon sink is: ; By calculating the annual rate of change of carbon sequestration, the dynamic trend of carbon sequestration can be reflected. A value greater than 0 indicates enhanced carbon sequestration; when... A value less than 0 indicates a weakening of carbon sequestration.

[0031] It is worth noting that all contents not described in detail in this invention are existing technologies and are well known to those skilled in the art.

[0032] Therefore, this invention adopts the above-mentioned spatiotemporal pattern prediction method based on forest carbon sink climate change, considering multiple sources such as special disturbance factors, meteorology, and topography. By coupling the Biome-BGC ecological mechanism model with a graph neural network, the error caused by spatiotemporal heterogeneity is reduced and the prediction accuracy is improved. At the same time, it solves the problem of poor climate factor adaptability of existing methods.

[0033] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for predicting the spatiotemporal patterns of climate change based on forest carbon sinks, characterized in that, Including the following steps: S1. Collect vegetation data, soil data, meteorological data, topographic data, and data on special interference factors in the target forest area; S2. Preprocess the collected multi-source data and perform spatiotemporal matching on the preprocessed data; S3. Quantify the climate index of the target forest area; S4. Using the Biome-BGC ecological process model as the core, integrate graph neural networks to construct a carbon sink-climate index-topography coupling model; S5. Set up multiple climate change scenarios and predict the spatiotemporal distribution pattern of carbon sinks under different scenarios based on the carbon sink-climate index-topography coupling model.

2. The method for predicting the spatiotemporal pattern of forest carbon sink climate change according to claim 1, characterized in that: The vegetation data in step S1 includes vegetation type, forest age, and physiological and ecological parameters; the soil data includes soil type and soil texture; the meteorological data includes extreme temperature, daily average temperature, daily maximum temperature, daily minimum temperature, monthly average temperature, annual accumulated temperature, and precipitation; the topographic data includes landform type, altitude, and slope; and the special interference factor data includes human interference factors and land occupied by road construction.

3. The method for predicting the spatiotemporal pattern of forest carbon sink climate change according to claim 1, characterized in that: Preprocessing in step S2 includes missing value imputation, noise filtering, and standardization.

4. The method for predicting the spatiotemporal pattern of forest carbon sink climate change according to claim 1, characterized in that: The climate indices for the target forest area in step S3 include: The extreme low temperature anomaly index in winter is calculated using the following formula: ； In the formula, This represents the average temperature for the corresponding month. Indicates the critical value for low temperature; The formula for the strong radiation anomaly index during the growing season is: ； In the formula, The monthly average UV-B radiation intensity, This is the critical value for radiation. The frozen soil freeze-thaw anomaly index is calculated using the following formula: ； In the formula, The thickness of the active layer in that year, This represents the multi-year average thickness of the active layer; a positive value indicates an enhanced freeze-thaw anomaly. The abnormal index of concentrated rainfall during the growing season is calculated using the following formula: ； In the formula, =1, when the monthly precipitation is greater than 80mm. =0; The annual accumulated temperature anomaly index is calculated using the following formula: 。 5. The method for predicting the spatiotemporal pattern of forest carbon sink climate change according to claim 1, characterized in that, Step S4 specifically includes: S41. Input the spatiotemporally matched data into the Biome-BGC model for processing to obtain ecosystem carbon flux parameters. S42. Construct a map structure based on carbon flux parameters, climate indices, data on special disturbance factors, and topographic data; S43. Using the graph structure as input, construct a carbon sink-climate index-topography coupled model.

6. The method for predicting the spatiotemporal pattern of forest carbon sink climate change according to claim 5, characterized in that, Step S42 specifically includes: The target forest region is divided into spatial grids, with each grid serving as a node in a graph structure. v A weighted approach is used to construct a feature fusion vector for each node based on carbon flux parameters, climate indices, special disturbance factor data, and terrain data; Define the edges connecting nodes and calculate the weight for each edge, specifically: For the grid Establish edge connections with adjacent grids, if the grid non-adjacent mesh Feature cosine similarity , Indicate the threshold and supplement the edge connections; The weight of each edge is calculated using the following formula: ； In the formula, Indicates the distance attenuation coefficient. For nodes With nodes Spatial Euclidean distance.

7. The method for predicting the spatiotemporal pattern of forest carbon sink climate change according to claim 6, characterized in that, Step S43 specifically includes: constructing a carbon sink-climate index-topography coupling model using a graph convolutional neural network and a fully connected layer, capturing spatial dependencies through the graph convolutional layer, and fitting the nonlinear mapping between carbon sinks and multiple factors through the fully connected layer.

8. The method for predicting the spatiotemporal pattern of forest carbon sink climate change according to claim 7, characterized in that: The loss function of the carbon sink-climate index-topography coupling model is: ； In the formula, Indicates the first i Measured carbon sequestration data for each node, Spatial regularization coefficient, This represents the spatial smoothing term.

9. A method for predicting the spatiotemporal pattern of forest carbon sink climate change according to claim 8, characterized in that, Step S5 specifically includes: setting up multiple climate change scenarios; for any climate change scenario, obtaining the grid carbon sink amount for different future time periods based on the carbon sink-climate index-topography coupling model; calculating the annual carbon sink change rate of each grid under the scenario; extracting the carbon sink value of each grid under the scenario; and generating a carbon sink time pattern distribution map based on the grid carbon sink values ​​at each time step.

Images