Dynamic management and analysis system of microbial residue carbon under global change conditions

Abstract

The present invention discloses a dynamic management and analysis system for microbial residue carbon under global change conditions, which relates to the technical field of microbial residue carbon analysis and prediction, and includes the following steps: regularly collecting the microbial residue carbon content of the soil, and performing data preprocessing to obtain residue carbon content data, and simultaneously obtaining relevant climate data; constructing a residue carbon climate-related model based on a multi-sensor model; obtaining relevant climate data at future moments, and using the residue carbon climate-related model to perform predictive analysis to obtain initial prediction data; performing data correction processing on the initial prediction data to obtain residue carbon prediction data; the present invention is used to solve the problem that the existing microbial residue carbon analysis and prediction technology does not reasonably plan the distribution of sampling points when predicting changes in microbial residue carbon based on climate conditions and using a machine learning model, and cannot dynamically correct errors in the output results of the model according to different seasonal environments.

Description

Technical Field

[0001] The present invention relates to the technical field of microbial residue carbon analysis and prediction, and in particular to a microbial residue carbon dynamic management and analysis system based on global change conditions. Background Art

[0002] Microbial residue carbon analysis and prediction technology refers to the technology that uses a series of specific methods and means to accurately measure the content of microbial residue carbon in the soil, and uses relevant models and data to predict its future content change trend.

[0003] Existing microbial residue carbon analysis and prediction technologies often lack scientific planning in sample collection when predicting changes in microbial residue carbon based on climate conditions and using machine learning models. Traditional sample collection may not fully consider the spatial heterogeneity of the soil, and the sampling points are concentrated in certain specific areas, resulting in the collected samples not being representative of the microbial residue carbon content of the entire land area. In addition, the distribution of sampling points is not reasonably planned, resulting in the sampling points being too close to each other or having spatial correlation, which affects the accuracy of subsequent predictions. In addition, the trained prediction model has its own assumptions and scope of application, and there are certain theoretical errors. The existing technology lacks an effective error correction mechanism when predicting the microbial residue carbon content, and the model's prediction under different seasonal environments is not effective. The measurement errors are different, and the existing technology cannot correct the model output results in time according to the actual situation, thereby reducing the accuracy and reliability of the prediction results, resulting in large errors in the prediction results; for example, the patent application with publication number CN116307177A discloses a soil microbial residue carbon prediction method and training method based on deep learning. This solution does not set a corresponding error correction mechanism for the output results of the model, resulting in insufficient accuracy and reliability of the prediction results; therefore, the existing microbial residue carbon analysis and prediction technology does not reasonably plan the distribution of sampling points when predicting changes in microbial residue carbon based on climatic conditions and using machine learning models, and cannot dynamically correct the errors of the model output results according to different seasonal environments. Summary of the Invention

[0004] The present invention aims to solve one of the technical problems in the prior art to at least a certain extent, by regularly collecting the microbial residue carbon content of the soil and performing data preprocessing to obtain residue carbon content data, and at the same time obtaining relevant climate data; constructing a residue carbon climate-related model based on a multi-sensor model; obtaining relevant climate data at future moments, and performing predictive analysis to obtain initial prediction data; performing data correction processing on the initial prediction data to obtain residue carbon prediction data, so as to solve the problem that the existing microbial residue carbon analysis and prediction technology does not reasonably plan the distribution of sampling points when predicting changes in microbial residue carbon based on climate conditions and using machine learning models, and cannot dynamically correct errors in the output results of the model according to different seasonal environments.

[0005] To achieve the above objectives, in a first aspect, the present application provides a system for managing and analyzing the dynamics of carbon in microbial residues under global change conditions, comprising a data collection module, a model management module, a prediction and analysis module, and a data correction module;

[0006] The data collection module includes a data collection unit and a preprocessing unit. The data collection unit is used to regularly collect the microbial residue carbon content of the soil and obtain relevant climate data at the same time. The preprocessing unit is used to preprocess the data to obtain the residue carbon content data.

[0007] The model management module is based on a multi-sensor model and uses the residual carbon content data and related climate data to construct a residual carbon climate related model;

[0008] The prediction and analysis module is used to obtain relevant climate data at future moments and perform prediction and analysis using a residual carbon climate-related model to obtain initial prediction data;

[0009] The data correction module is used to perform data correction processing on the initial prediction data according to the residue carbon content data to obtain the residue carbon prediction data.

[0010] Furthermore, the data collection unit is configured with a data collection strategy, which includes:

[0011] The land area to be managed is recorded as managed land; the partition rectangle size is set to b1*b2, and the managed land is divided into multiple blocks of land using the partition rectangle. All the blocks of land are numbered to ensure that the numbers of any two blocks of land are different; the total number of blocks of land is obtained, recorded as A0; the total area of ​​the managed land is obtained, recorded as V0;

[0012] Obtaining the microbial residue carbon content of the managed land at a first time interval within a first time period, including: selecting k1 block land areas from all block land areas, and ensuring that there are no other selected block land areas in the 8-neighborhood of any selected block land area, and the area of ​​the minimum enclosing circle of the geometric center of the k1 selected block land areas is less than or equal to k2*V0; recording the selected k1 block land areas as collection block areas; wherein the first time period is T0, the first time interval is t1, k1 is the set number, and k2 is the set proportional coefficient.

[0013] Furthermore, the data collection strategy also includes:

[0014] For any sampling block area, obtain the microbial residue carbon content of the soil at the depths of 0-C1cm, C1-C2cm, and C2-C3cm, and record them in order as the upper microbial residue carbon content, the middle microbial residue carbon content, and the lower microbial residue carbon content; repeatedly collect the microbial residue carbon content of the soil at the corresponding depths in all sampling blocks, and record the collection date, which is recorded as the original soil residue carbon data of the corresponding date;

[0015] Repeatedly obtain the original land residue carbon data of the managed land at the first time interval, and ensure that the collection block areas used for the two consecutive collections of the original land residue carbon data are completely different;

[0016] For each collection of original land residue carbon data, obtain the corresponding first time period of the collection date, the global average temperature, the average temperature of the managed land area, and the rainfall at the managed land area, and record them in order as the global temperature, regional temperature, and rainfall, as the meteorological data for the corresponding first time period;

[0017] The meteorological data of the first time period are arranged in chronological order and recorded as relevant meteorological data.

[0018] Furthermore, the preprocessing unit is configured with a preprocessing strategy, which includes:

[0019] For each collected original soil residue carbon data, the carbon content of the upper microbial residue, the carbon content of the middle microbial residue, and the carbon content of the lower microbial residue in the collected block area are sorted in ascending order and recorded as the original upper carbon content sequence, the original middle carbon content sequence, and the original lower carbon content sequence respectively;

[0020] The smallest [k3*k1] and largest [k3*k1] data of the original upper carbon content sequence, the original middle carbon content sequence, and the original lower carbon content sequence were removed respectively, and the average values ​​of the remaining parts were calculated respectively. They were recorded in order as the upper residue carbon content, middle residue carbon content, and lower residue carbon content of the corresponding collection date, and marked as the initial land residue carbon data of the corresponding collection date, where k3 is the set proportional coefficient.

[0021] Furthermore, the preprocessing strategy also includes:

[0022] The original upper layer residue carbon content, the original middle layer residue carbon content, and the original lower layer residue carbon content of all collection dates in each first time period are sorted in ascending order; and recorded in order as the initial upper layer carbon content sequence, the initial middle layer carbon content sequence, and the initial lower layer carbon content sequence;

[0023] Remove the smallest [k3*A1] and largest [k3*A1] data of the initial upper carbon content sequence, the initial middle carbon content sequence, and the initial lower carbon content sequence, and calculate the average value of the remaining parts, which are recorded in order as the upper period average content SM, the middle period average content ZM, and the lower period average content XM; where A1 is the number of collection dates in the first time period;

[0024] The average microbial residue carbon content of the first time period is calculated according to the average residue carbon formula. The average residue carbon formula is as follows: , where PMN represents the average carbon content of microbial residues in the first time period; q1, q2, and q3 are weight coefficients, q1+q2+q3=1;

[0025] The average microbial residue carbon content of all the first time periods is arranged in chronological order and recorded as residue carbon content data.

[0026] Furthermore, the model management module is configured with a model management strategy, which includes:

[0027] Normalize the residual carbon content data and related meteorological data according to specific data types, and scale all data sizes in the residual carbon content data and related meteorological data to [0, 1]. After completion, obtain the normalized carbon content data and normalized meteorological data in sequence;

[0028] Combine the global temperature, regional temperature, and rainfall of the same first time period in the normalized meteorological data into a related feature vector, denoted as TH={h1, h2, h3}, where TH represents the related feature vector; h1, h2, and h3 represent the global temperature, regional temperature, and rainfall in order, and the meteorological characteristic data is obtained after completion;

[0029] The meteorological characteristic data and the normalized carbon content data are combined and stored according to the corresponding first time period and recorded as carbon content characteristic data.

[0030] Furthermore, the model management strategy also includes:

[0031] Repeatedly obtain carbon content characteristic data corresponding to other land areas in the area where the managed land is located and with the same soil type as the managed land, and merge them with the carbon content characteristic data to record them as carbon content-related data;

[0032] An original correlation model is constructed based on the multi-perceptron model. The original correlation model includes an input layer, a hidden processing layer, and an output layer. The number of neurons in the input layer is set to e1, the number of hidden layers in the hidden processing layer is i, the number of neurons in each hidden layer is f1-fi, and the number of neurons in the output layer is e2. The original correlation model is trained using carbon content-related data, and a residual carbon climate correlation model is obtained after completion.

[0033] Furthermore, the prediction analysis module is configured with a prediction analysis strategy, which includes:

[0034] Obtain relevant meteorological data of the first future time length and with the first time period as the time interval, record them as future meteorological data, normalize the future meteorological data according to the data type, and combine the corresponding data of the corresponding time into relevant feature vectors to obtain future feature data; input the future feature data into the residual carbon climate related model to obtain the residual carbon content data of the first future time length and with the first time period as the time interval, mark them as initial prediction data, where the first time length is R1.

[0035] Furthermore, the data correction module is configured with a data correction strategy, which includes:

[0036] The carbon content-related data were divided according to the season to which the corresponding first time period belonged, and the spring carbon content data, summer carbon content data, autumn carbon content data, and winter carbon content data were obtained. These data were input into the residual carbon climate-related model respectively, and the overall error corresponding to each season was calculated according to the average error formula. The average error formula is as follows: , where WC represents the average error, Yj represents the actual average microbial residue carbon content of the first time period, Xj represents the average microbial residue carbon content of the first time period output by the residue carbon climate correlation model, and n represents the number of samples input into the residue carbon climate correlation model; the overall error in spring, summer, autumn, and winter are obtained respectively.

[0037] Furthermore, the data correction strategy also includes:

[0038] For any residual carbon content data in the initial prediction data, record it as the first predicted content CY, obtain the regional temperature, rainfall and season corresponding to the first predicted content, and record them as the first temperature Y0, the first rainfall U0 and the first season respectively in order;

[0039] Screening data similar to the first predicted content from the carbon content-related data includes: recording data corresponding to the first season in the carbon content-related data as carbon content data for the same season; calculating similarity values ​​between each carbon content data for the same season and the first predicted content in sequence according to a similarity formula, and sorting the data from smallest to largest to record the similarity values ​​as a similarity value sequence; the similarity formula is as follows: , where XS represents the similarity value, Y1 and U1 represent the regional temperature and rainfall in the carbon content data of the same season respectively; the carbon content data of the same season corresponding to the smallest k4 similarity values ​​in the similarity value sequence are recorded as the similarity data of the same season;

[0040] All similar data from the same season are input into the residual carbon climate correlation model, and the average error corresponding to each similar data from the same season is calculated according to the average error formula, which is recorded as the seasonal individual error of the first predicted content;

[0041] According to the seasonal individual error of the first predicted content and the overall error corresponding to the first season, the first predicted content is corrected using the correction formula. The correction formula is as follows: , where ZY represents the first predicted content after error correction, CY is the first predicted content, GT represents the seasonal individual error of the first predicted content, ZT represents the overall error corresponding to the first season, and q4 and q5 are the set weight coefficients;

[0042] Repeat the error correction for all the residual carbon content data in the initial prediction data, and obtain the residual carbon prediction data after completion.

[0043] The beneficial effects of the present invention are as follows: the present invention obtains residue carbon content data by regularly collecting the microbial residue carbon content of the soil and performing data preprocessing, and simultaneously obtains relevant climate data; constructs a residue carbon climate-related model based on a multi-sensor model using the residue carbon content data and relevant climate data; obtains relevant climate data at future moments, and uses the residue carbon climate-related model to perform predictive analysis to obtain initial prediction data; performs data correction processing on the initial prediction data according to the residue carbon content data to obtain residue carbon prediction data; when predicting changes in microbial residue carbon, the accuracy and reliability of the prediction results are ensured by rationally planning the distribution of sampling points and dynamically correcting the output results of the model according to different seasonal environments.

[0044] The present invention divides the managed land into blocks and ensures that the collection block areas are not adjacent and that the distribution area of ​​the collection block areas is greater than a certain proportion, thereby ensuring the representativeness and scientific nature of the sampling, avoiding the problems of excessive concentration or uneven distribution of sampling points and the existence of spatial correlation, and reducing the inaccurate prediction results caused by sampling bias; by calculating the overall error corresponding to each season of the model and the individual error corresponding to each output data, personalized error correction is performed on each data, which can accurately grasp the overall prediction deviation of the model in different seasons, make the error correction more seasonally targeted, avoid error residues caused by general corrections, and at the same time, the calculated individual error can better reflect the error characteristics of the predicted value under specific environmental conditions, making the error correction more accurate, and ensuring the accuracy and reliability of the final result. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a functional block diagram of the system of the present invention;

[0046] Figure 2 is a flow chart of the steps of the method of the present invention;

[0047] Figure 3 Schematic diagram of 8 neighborhoods of the divided land area of ​​the present invention;

[0048] Figure 4 Flowchart of the steps of the data correction strategy of the present invention.

[0049] Figure 5 Schematic diagram of the structure of the electronic device of the present invention. DETAILED DESCRIPTION

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0051] Example 1, please refer to Figure 1 As shown, in the first aspect, the present application provides a microbial residue carbon dynamics management and analysis system based on global change conditions, including a data collection module, a model management module, a prediction analysis module, and a data correction module;

[0052] The data collection module includes a data acquisition unit and a preprocessing unit. The data acquisition unit is used to regularly collect the microbial residue carbon content of the soil and obtain relevant climate data at the same time; the preprocessing unit is used to preprocess the data to obtain the residue carbon content data;

[0053] The data collection unit is configured with a data collection strategy, which includes: recording the land area to be managed as managed land; setting the partition rectangle size to b1*b2, using the partition rectangle to divide the managed land into multiple block land areas, and numbering all the block land areas to ensure that the numbers of any two block land areas are different; obtaining the total number of block land areas, recorded as A0; obtaining the total area of ​​the managed land, recorded as V0; the partition rectangle size b1*b2 can be set according to the actual application scenario. In this embodiment, b1*b2=5m*5m;

[0054] Obtaining a microbial residue carbon content of the managed land at first time intervals during a first time period, comprising: Figure 3 As shown, k1 block land areas are selected from all block land areas, and it is ensured that there are no other selected block land areas in the 8-neighborhood of any selected block land area, and the area of ​​the minimum enclosing circle of the geometric center of the selected k1 block land areas is less than or equal to k2*V0; the selected k1 block land areas are recorded as collection block areas; wherein the first time period is T0, the first time interval is t1, k1 is the set number, and k2 is the set proportional coefficient; in order to facilitate the subsequent acquisition of relevant climate data, T0 is generally 1 month; t1 is generally an integer multiple of days and less than T0; k1 is at least 3; k2 can be set according to the actual application scenario, and the normal value range is [0.25, 0.6]; in this embodiment, T0=1 month, t1=1 day, k1=8, k2=0.4;

[0055] For any collection block area, obtain the microbial residue carbon content of the soil at a depth of 0-C1cm, C1-C2cm, and C2-C3cm, and record them in order as the upper microbial residue carbon content, the middle microbial residue carbon content, and the lower microbial residue carbon content; repeatedly collect the microbial residue carbon content of the soil at the corresponding depths in all collection block areas, and record the collection date, and record it as the original soil residue carbon data C1, C2, and C3 of the corresponding date. The data can be set according to the actual application scenario. Under normal circumstances, C3 is less than or equal to 20cm. In this embodiment, C1=5cm, C2=15cm, and C3=15cm

[0056] Repeatedly obtain the original soil residue carbon data of the managed land at the first time interval, and ensure that the collection block areas used for the original soil residue carbon data collected twice are completely different; this can reduce the deviation caused by the concentration of sampling points, make the collected data more representative, and thus more accurately reflect the changes in the microbial residue carbon content of the entire managed land; if sampling is carried out twice in the same area, excessive attention may be paid to the characteristics of certain specific areas, while ignoring the changes in other areas, resulting in the sampling results not representing the true situation of the entire managed land;

[0057] For the original land residue carbon data collected each time, according to the first time period in which the collection date falls, the global average temperature, the average temperature of the management land location, and the rainfall of the management land location of the corresponding first time period are obtained, and recorded in order as the global temperature, regional temperature, and rainfall, respectively, as the meteorological data of the corresponding first time period; the meteorological data of the first time period are arranged in chronological order and recorded as relevant meteorological data; the global average temperature is a key indicator reflecting global climate change, and global climate change will indirectly affect the content of soil microbial residue carbon by affecting physical properties such as soil moisture and aeration, as well as chemical properties such as soil pH and nutrient content, so that subsequent models can more comprehensively consider the comprehensive effect of temperature on the dynamics of microbial residue carbon, reduce prediction errors, and improve the accuracy and reliability of predictions;

[0058] The preprocessing unit is configured with a preprocessing strategy, which includes: for each collected original land residue carbon data, the carbon content of the upper microbial residue, the carbon content of the middle microbial residue, and the carbon content of the lower microbial residue in the collected collection block area are sorted in ascending order, and recorded in order as the original upper layer carbon content sequence, the original middle layer carbon content sequence, and the original lower layer carbon content sequence; that is, the same layer is sorted together in ascending order;

[0059] The smallest [k3*k1] and largest [k3*k1] data of the original upper carbon content sequence, the original middle carbon content sequence, and the original lower carbon content sequence are removed respectively, and the average values ​​of the remaining parts are respectively calculated and recorded in order as the upper residue carbon content, the middle residue carbon content, and the lower residue carbon content on the corresponding collection date, and marked as the initial land residue carbon data on the corresponding collection date, where k3 is a set proportional coefficient; in this embodiment, k3 is 0.1, and [k3*k1] is a rounded integer and is at least 1;

[0060] The original upper layer residue carbon content, the original middle layer residue carbon content, and the original lower layer residue carbon content of all collection dates in each first time period are sorted in ascending order; and recorded in order as the initial upper layer carbon content sequence, the initial middle layer carbon content sequence, and the initial lower layer carbon content sequence;

[0061] Remove the smallest [k3*A1] and largest [k3*A1] data of the initial upper carbon content sequence, the initial middle carbon content sequence, and the initial lower carbon content sequence, and calculate the average value of the remaining data, which are recorded in order as the upper period average content SM, the middle period average content ZM, and the lower period average content XM; where A1 is the number of collection dates in the first time period; [k3*A1] is also a rounded integer and is at least 1;

[0062] The average microbial residue carbon content of the first time period is calculated according to the average residue carbon formula. The average residue carbon formula is as follows: , where PMN represents the average microbial residue carbon content in the first time period; q1, q2, and q3 are weight coefficients, q1+q2+q3=1; in this embodiment, q1=0.3, q2=0.4, q3=0.3, which can be set according to the actual application scenario; by sorting the collected raw data, removing extreme values, and averaging the average, the impact of abnormal data on the results can be effectively reduced, and the accuracy and stability of the data can be improved; at the same time, calculating the average microbial residue carbon content in different time periods further enhances the regularity and usability of the data.

[0063] Arrange the average microbial residue carbon content of all first time periods in chronological order and record it as residue carbon content data;

[0064] During the specific implementation process, the managed land is divided into blocks and numbered, and constraints are set to select the sampling block areas. This sampling method avoids excessive concentration of sampling points, makes the sampling points more evenly distributed in space, and can more comprehensively represent different areas of the managed land, reducing inaccurate results caused by local sampling deviations.

[0065] The model management module is based on a multi-sensor model and uses residual carbon content data and related climate data to build a residual carbon climate related model;

[0066] The model management module is equipped with a model management strategy, which includes: normalizing the residual carbon content data and related meteorological data according to specific data types, scaling all data sizes within the residual carbon content data and related meteorological data to [0, 1]; normalizing the four data types of global temperature, regional temperature, rainfall, and average microbial residual carbon content, and obtaining normalized carbon content data and normalized meteorological data in sequence.

[0067] Combine the global temperature, regional temperature, and rainfall of the same first time period in the normalized meteorological data into a related feature vector, denoted as TH={h1, h2, h3}, where TH represents the related feature vector; h1, h2, and h3 represent the global temperature, regional temperature, and rainfall in order, and the meteorological characteristic data is obtained after completion;

[0068] Merging and storing the meteorological characteristic data and the normalized carbon content data according to the corresponding first time period, and recording them as carbon content characteristic data;

[0069] Repeatedly obtain carbon content characteristic data corresponding to other land areas with the same soil type as the managed land in the region where the managed land is located, and merge them with the carbon content characteristic data to record them as carbon content-related data; for example, if the managed land is forest land, repeatedly obtain carbon content characteristic data for other forest lands in the region where the managed land is located; the scope of the region generally does not exceed the prefecture-level city, that is, the other land areas selected must be in the same prefecture-level city as the managed land;

[0070] An original correlation model is constructed based on a multi-perceptron model. The original correlation model includes an input layer, a hidden processing layer, and an output layer. The number of neurons in the input layer is set to e1, the number of hidden layers in the hidden processing layer is i, the number of neurons in each hidden layer is f1-fi, and the number of neurons in the output layer is e2. The original correlation model is trained using carbon content-related data, and a residual carbon climate correlation model is obtained after completion. In this embodiment, e1=3, i.e., h1, h2, and h3 are input; i=2, f1=f2=128; and e2=1, i.e., the average microbial residue carbon content is output.

[0071] During the specific implementation process, the carbon content characteristic data corresponding to the same area are repeatedly obtained because a large amount of data is needed to train and verify the model when building it; repeatedly obtaining data from multiple identical areas can provide richer information, enabling the model to better fit the actual situation and improve the accuracy and predictive ability of the model; at the same time, it also ensures that the model has good applicability and reliability.

[0072] The prediction and analysis module is used to obtain relevant climate data at future times and use the residual carbon climate-related model to perform prediction analysis to obtain initial prediction data;

[0073] The prediction analysis module is configured with a prediction analysis strategy, which includes: obtaining relevant meteorological data of a first future time length and a first time period as a time interval, recording as future meteorological data, normalizing the future meteorological data according to data type, and combining corresponding data of corresponding time into relevant feature vectors to obtain future feature data; inputting the future feature data into a residual carbon climate correlation model to obtain residual carbon content data of a first future time length and a first time period as a time interval, marking as initial prediction data, wherein the first time length is R1; in this embodiment, the first time length is 12 months and the first time period is 1 month, that is, obtaining relevant meteorological data for each month of the next 12 months, and predicting residual carbon content data for each month of the next 12 months;

[0074] During the specific implementation process, relevant future meteorological data can be obtained from relevant meteorological websites or platforms, or you can make predictions based on historical meteorological data, but the accuracy of the predictions must be guaranteed.

[0075] The data correction module is used to perform data correction processing on the initial prediction data according to the residue carbon content data to obtain the residue carbon prediction data;

[0076] The data correction module is configured with a data correction strategy, which includes: Figure 4 As shown, the carbon content-related data are divided according to the season to which the corresponding first time period belongs, and the spring carbon content data, summer carbon content data, autumn carbon content data, and winter carbon content data are obtained. They are input into the residual carbon climate-related model respectively, and the overall error corresponding to each season is calculated according to the average error formula. The average error formula is as follows: , where WC represents the average error, Yj represents the actual average microbial residue carbon content of the first time period, Xj represents the average microbial residue carbon content of the first time period output by the residue carbon climate-related model, and n represents the number of samples input into the residue carbon climate-related model; the spring overall error, summer overall error, autumn overall error, and winter overall error are obtained respectively; the climatic conditions in different seasons are different, and the error of the residue carbon climate-related model will change with the change of seasonal climatic conditions; by calculating the seasonal overall error separately, the overall prediction deviation of the model in different seasons can be accurately grasped, making the error correction more seasonally targeted and avoiding the error residual caused by general correction;

[0077] For any residual carbon content data in the initial prediction data, record it as the first predicted content CY, obtain the regional temperature, rainfall, and season corresponding to the first predicted content, and record them in order as the first temperature Y0, the first rainfall U0, and the first season; for example, if the season corresponding to the first predicted content is summer, then the first season is summer;

[0078] Screening data similar to the first predicted content from the carbon content-related data includes: recording data corresponding to the first season in the carbon content-related data as carbon content data for the same season; calculating similarity values ​​between each carbon content data for the same season and the first predicted content in sequence according to a similarity formula, and sorting the data from smallest to largest to record the similarity values ​​as a similarity value sequence; the similarity formula is as follows: , where XS represents the similarity value, Y1 and U1 represent the regional temperature and rainfall in the carbon content data of the same season respectively in sequence; the carbon content data of the same season corresponding to the smallest k4 similarity values ​​in the similarity value sequence are recorded as the similarity data of the same season; in this embodiment, k4=4;

[0079] All similar data from the same season are input into the residual carbon climate correlation model, and the average error corresponding to each similar data from the same season is calculated according to the average error formula, which is recorded as the seasonal individual error of the first predicted content;

[0080] According to the seasonal individual error of the first predicted content and the overall error corresponding to the first season, the first predicted content is corrected using the correction formula. The correction formula is as follows: , where ZY represents the first predicted content after error correction, CY is the first predicted content, GT represents the seasonal individual error of the first predicted content, ZT represents the overall error corresponding to the first season, and q4 and q5 are set weight coefficients; in this embodiment, q4=0.7, q5=0.3; generally, q4>q5, because the seasonal individual error directly reflects the deviation between a single first predicted content and the true value, and it has a more direct impact on the accuracy of a specific predicted value; while the overall error reflects the overall error of the entire seasonal data, it may mask some special changes and differences in individual data;

[0081] Repeat the error correction for all the residual carbon content data in the initial prediction data, and obtain the residual carbon prediction data after completion;

[0082] During the specific implementation process, seasonal individual errors and seasonal overall errors are combined for correction, realizing the integration of the individual error of a specific prediction value and the overall error of the entire season; the seasonal overall error reflects the general deviation of the model in that season, and the seasonal individual error reflects the error characteristics of a specific prediction value under similar environments; the combination of the two takes into account both the overall performance of the model in that season and the special environmental conditions of the prediction value, making the error correction more targeted and more accurate.

[0083] Example 2, please refer to Figure 2 As shown, the present application provides a method for analyzing the dynamic management of carbon in microbial residues under global change conditions, comprising the following steps:

[0084] Step S1, regularly collecting the microbial residue carbon content of the soil and performing data preprocessing to obtain residue carbon content data, while simultaneously obtaining relevant climate data; Step S1 includes the following sub-steps:

[0085] Step S101: Record the land area to be managed as managed land; set the partition rectangle size to b1*b2, use the partition rectangle to divide the managed land into multiple sub-land areas, and number all the sub-land areas to ensure that the numbers of any two sub-land areas are different;

[0086] Step S102: Obtain the total number of divided land areas, recorded as A0; obtain the total area of ​​managed land, recorded as V0;

[0087] Step S103, obtaining the microbial residue carbon content of the managed land at a first time interval within a first time period, including: selecting k1 block land areas from all the block land areas, and ensuring that no other selected block land areas exist in the 8-neighborhood of any selected block land area, and that the area of ​​the minimum enclosing circle of the geometric center of the k1 selected block land areas is less than or equal to k2*V0; recording the k1 selected block land areas as collection block areas; wherein the first time period is T0, the first time interval is t1, k1 is a set number, and k2 is a set proportional coefficient;

[0088] Step S104: For any sampling block area, the microbial residue carbon content of the soil at the depths of 0-C1 cm, C1-C2 cm, and C2-C3 cm is obtained respectively, and recorded in order as the upper layer microbial residue carbon content, the middle layer microbial residue carbon content, and the lower layer microbial residue carbon content; the microbial residue carbon content of the soil at the corresponding depths of all sampling blocks is repeatedly collected, and the collection date is recorded as the original soil residue carbon data of the corresponding date;

[0089] Step S105, repeatedly acquiring the original land residue carbon data of the managed land at a first time interval, and ensuring that the acquisition block areas used for the two consecutive acquisitions of the original land residue carbon data are completely different;

[0090] Step S106: For each collected raw land residue carbon data, the global average temperature, the average temperature at the managed land location, and the rainfall at the managed land location for the corresponding first time period are obtained based on the first time period of the collection date. These are recorded in order as the global temperature, the regional temperature, and the rainfall, respectively, as the meteorological data for the corresponding first time period.

[0091] Step S107, arranging the meteorological data of the first time period in chronological order and recording them as relevant meteorological data;

[0092] Step S108: For each collected original soil residue carbon data, the carbon contents of the upper microbial residue, the carbon contents of the middle microbial residue, and the carbon contents of the lower microbial residue in the collected block area are sorted in ascending order, and recorded in order as the original upper layer carbon content sequence, the original middle layer carbon content sequence, and the original lower layer carbon content sequence;

[0093] Step S109: Remove the smallest [k3*k1] and largest [k3*k1] data of the original upper layer carbon content sequence, the original middle layer carbon content sequence, and the original lower layer carbon content sequence, and calculate the average value of the remaining parts, which are recorded in order as the upper layer residue carbon content, the middle layer residue carbon content, and the lower layer residue carbon content on the corresponding collection date, and marked as the initial land residue carbon data on the corresponding collection date, where k3 is the set proportional coefficient;

[0094] Step S110, sorting the original upper layer residue carbon content, the original middle layer residue carbon content, and the original lower layer residue carbon content of all collection dates within each first time period in ascending order; recording them in order as an initial upper layer carbon content sequence, an initial middle layer carbon content sequence, and an initial lower layer carbon content sequence;

[0095] Step S111: Remove the smallest [k3*A1] data and the largest [k3*A1] data from the initial upper carbon content sequence, the initial middle carbon content sequence, and the initial lower carbon content sequence, and calculate the average values ​​of the remaining data, which are recorded in order as the upper period average content SM, the middle period average content ZM, and the lower period average content XM; where A1 is the number of data collection dates in the first time period;

[0096] Step S112, the average microbial residue carbon content of the corresponding first time period is calculated according to the average residue carbon formula. The average residue carbon formula is as follows: , where PMN represents the average carbon content of microbial residues in the first time period; q1, q2, and q3 are weight coefficients, q1+q2+q3=1;

[0097] Step S113: Arrange the average carbon contents of microbial residues in all first time periods in chronological order and record them as residue carbon content data.

[0098] Step S2, based on the multi-sensor model, and using the residue carbon content data and related climate data, constructs a residue carbon climate correlation model; Step S2 includes the following sub-steps:

[0099] Step S201: normalize the residue carbon content data and related meteorological data according to specific data types, scaling all data sizes in the residue carbon content data and related meteorological data to [0, 1]. After completion, normalized carbon content data and normalized meteorological data are obtained in sequence.

[0100] Step S202: Combine the global temperature, regional temperature, and rainfall for the same first time period in the normalized meteorological data into a related feature vector, denoted as TH={h1, h2, h3}, where TH represents the related feature vector; h1, h2, and h3 represent the global temperature, regional temperature, and rainfall, respectively, in that order. Upon completion, meteorological feature data is obtained.

[0101] Step S203, combining the meteorological characteristic data and the normalized carbon content data according to the corresponding first time period and storing them as carbon content characteristic data;

[0102] Step S204, repeatedly obtaining carbon content characteristic data corresponding to other land areas in the area where the managed land is located and having the same soil type as the managed land, and merging the data with the carbon content characteristic data to record the data as carbon content related data;

[0103] Step S205: construct an original correlation model based on the multi-perceptron model. The original correlation model includes an input layer, a hidden processing layer, and an output layer. The number of neurons in the input layer is set to e1, the number of hidden layers in the hidden processing layer is i, the number of neurons in each hidden layer is f1-fi, and the number of neurons in the output layer is e2. Use carbon content-related data to train the original correlation model, and obtain a residual carbon climate correlation model after completion.

[0104] Step S3, obtaining relevant climate data for the future time, and performing prediction analysis using a residual carbon climate-related model to obtain initial prediction data; Step S3 includes the following sub-steps:

[0105] Step S301, obtaining relevant meteorological data of a first time length in the future and with a first time period as a time interval, and recording it as future meteorological data;

[0106] Step S302 , normalizing the future meteorological data according to the data type, and combining the corresponding data at the corresponding time into a related feature vector to obtain future feature data;

[0107] Step S303 , inputting future characteristic data into the residual carbon climate correlation model, obtaining residual carbon content data of a first future time length and a first time period as a time interval, and marking it as initial prediction data, wherein the first time length is R1.

[0108] Step S4, performing data correction processing on the initial prediction data according to the residue carbon content data to obtain residue carbon prediction data; Step S4 includes the following sub-steps:

[0109] In step S401, the carbon content-related data is divided according to the season to which the corresponding first time period belongs, obtaining spring carbon content data, summer carbon content data, autumn carbon content data, and winter carbon content data. The data are input into the residual carbon climate-related model respectively, and the overall error corresponding to each season is calculated according to the average error formula. The average error formula is as follows: , where WC represents the average error, Yj represents the actual average microbial residue carbon content of the first time period, Xj represents the average microbial residue carbon content of the first time period output by the residue carbon climate correlation model, and n represents the number of samples input into the residue carbon climate correlation model; the overall error in spring, summer, autumn, and winter are obtained respectively;

[0110] Step S402: For any residual carbon content data in the initial prediction data, record it as a first predicted content CY, obtain the regional temperature, rainfall, and season corresponding to the first predicted content, and record them in order as the first temperature Y0, the first rainfall U0, and the first season respectively;

[0111] Step S403, screening data similar to the first predicted content from the carbon content related data, including: recording the data corresponding to the first season in the carbon content related data as the carbon content data of the same season; calculating the similarity value between each carbon content data of the same season and the first predicted content in sequence according to a similarity formula, and sorting them from small to large to record them as a similarity value sequence; the similarity formula is as follows: , where XS represents the similarity value, Y1 and U1 represent the regional temperature and rainfall in the carbon content data of the same season respectively; the carbon content data of the same season corresponding to the smallest k4 similarity values ​​in the similarity value sequence are recorded as the similarity data of the same season;

[0112] Step S404: input all similar data of the same season into the residual carbon climate correlation model, and calculate the average error corresponding to each similar data of the same season according to the average error formula, and record it as the seasonal individual error of the first predicted content;

[0113] Step S405: Based on the seasonal individual error of the first predicted content and the overall error corresponding to the first season, the first predicted content is corrected using a correction formula. The correction formula is as follows: , where ZY represents the first predicted content after error correction, CY is the first predicted content, GT represents the seasonal individual error of the first predicted content, ZT represents the overall error corresponding to the first season, and q4 and q5 are the set weight coefficients;

[0114] Step S406, repeatedly performing error correction on all the residue carbon content data in the initial prediction data, and obtaining the residue carbon prediction data after completion.

[0115] Example 3, please refer to Figure 5 As shown, Figure 5 A schematic diagram of the structure of an electronic device is provided. The electronic device may include: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus. The memory stores computer-readable instructions, and the processor can call the instructions from the memory. When the computer-readable instructions are executed by the processor, the steps of a method for dynamic management and analysis of microbial residue carbon under global change conditions are executed to achieve the following functions: regularly collecting soil microbial residue carbon content and performing data preprocessing to obtain residue carbon content data, while also obtaining relevant climate data; constructing a residue carbon-climate correlation model based on a multi-sensor model and using the residue carbon content data and relevant climate data; obtaining relevant climate data for the future and performing predictive analysis using the residue carbon-climate correlation model to obtain initial predicted data; and performing data correction processing on the initial predicted data based on the residue carbon content data to obtain predicted residue carbon data.

[0116] In addition, the logical instructions in the above-mentioned memory can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application, or the part that contributes to the existing technology, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0117] Example 4. The present application also provides a computer-readable storage medium. The present application provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the dynamic management and analysis method of microbial residue carbon under global change conditions are executed to achieve the following functions: regularly collect the microbial residue carbon content of the soil, and perform data preprocessing to obtain residue carbon content data, and at the same time obtain relevant climate data; based on a multi-sensor model, and using residue carbon content data and relevant climate data to construct a residue carbon climate related model; obtain relevant climate data at future times, and use the residue carbon climate related model to perform predictive analysis to obtain initial predicted data; perform data correction processing on the initial predicted data according to the residue carbon content data to obtain residue carbon predicted data.

[0118] Through the description of the above embodiments, the embodiments of the present invention can be provided as methods, systems, or computer program products. Based on this understanding, the essence of the above technical solutions or the portion that contributes to the prior art can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for enabling a computer device (such as a personal computer, server, or network device) to execute the methods described in various embodiments or certain portions of the embodiments.

[0119] In the embodiments provided in this application, it should be understood that the disclosed system or method can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of systems, modules and units can be electrical, mechanical or other forms.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A dynamic management and analysis system for microbial residue carbon under global change conditions, characterized by: It includes data collection module, model management module, prediction analysis module and data correction module; The data collection module includes a data collection unit and a preprocessing unit. The data collection unit is used to regularly collect the carbon content of microbial residues in the soil and obtain relevant climate data at the same time. The preprocessing unit is used to perform data preprocessing to obtain the residue carbon content data; The model management module is based on a multi-sensor model and uses the residual carbon content data and related climate data to construct a residual carbon climate related model; The prediction and analysis module is used to obtain relevant climate data at future moments and perform prediction and analysis using a residual carbon climate-related model to obtain initial prediction data; The data correction module is used to perform data correction processing on the initial prediction data according to the residue carbon content data to obtain the residue carbon prediction data; The data acquisition unit is configured with a data acquisition strategy, which includes: The land area to be managed is recorded as managed land; the partition rectangle size is set to b1*b2, and the managed land is divided into multiple blocks of land using the partition rectangle. All the blocks of land are numbered to ensure that the numbers of any two blocks of land are different; the total number of blocks of land is obtained, recorded as A0; the total area of ​​the managed land is obtained, recorded as V0; Obtaining the microbial residue carbon content of the managed land at a first time interval within a first time period, including: selecting k1 block land areas from all block land areas, and ensuring that there are no other selected block land areas in the 8-neighborhood of any selected block land area, and the area of ​​the minimum enclosing circle of the geometric center of the k1 selected block land areas is less than or equal to k2*V0; recording the selected k1 block land areas as collection block areas; wherein the first time period is T0, the first time interval is t1, k1 is the set number, and k2 is the set proportional coefficient.

2. The microbial residue carbon dynamics management and analysis system based on global change conditions according to claim 1 is characterized in that: The data collection strategy also includes: For any sampling block area, obtain the microbial residue carbon content of the soil at the depths of 0-C1cm, C1-C2cm, and C2-C3cm, and record them in order as the upper microbial residue carbon content, the middle microbial residue carbon content, and the lower microbial residue carbon content; repeatedly collect the microbial residue carbon content of the soil at the corresponding depths in all sampling blocks, and record the collection date, which is recorded as the original soil residue carbon data of the corresponding date; Repeatedly obtain the original land residue carbon data of the managed land at the first time interval, and ensure that the collection block areas used for the two consecutive collections of the original land residue carbon data are completely different; For each collection of original land residue carbon data, obtain the corresponding first time period of the collection date, the global average temperature, the average temperature of the managed land area, and the rainfall at the managed land area, and record them in order as the global temperature, regional temperature, and rainfall, as the meteorological data for the corresponding first time period; The meteorological data of the first time period are arranged in chronological order and recorded as relevant meteorological data.

3. The microbial residue carbon dynamics management and analysis system based on global change conditions according to claim 2, characterized in that: The preprocessing unit is configured with a preprocessing strategy, which includes: For each collected original soil residue carbon data, the carbon content of the upper microbial residue, the carbon content of the middle microbial residue, and the carbon content of the lower microbial residue in the collected block area are sorted in ascending order and recorded as the original upper carbon content sequence, the original middle carbon content sequence, and the original lower carbon content sequence respectively; The smallest [k3*k1] and largest [k3*k1] data of the original upper carbon content sequence, the original middle carbon content sequence, and the original lower carbon content sequence were removed respectively, and the average values ​​of the remaining parts were calculated respectively. They were recorded in order as the upper residue carbon content, middle residue carbon content, and lower residue carbon content of the corresponding collection date, and marked as the initial land residue carbon data of the corresponding collection date, where k3 is the set proportional coefficient.

4. The system for dynamic management and analysis of microbial residue carbon under global change conditions according to claim 3, characterized in that: Preprocessing strategies also include: The original upper layer residue carbon content, the original middle layer residue carbon content, and the original lower layer residue carbon content of all collection dates in each first time period are sorted in ascending order; and recorded in order as the initial upper layer carbon content sequence, the initial middle layer carbon content sequence, and the initial lower layer carbon content sequence; Remove the smallest [k3*A1] and largest [k3*A1] data of the initial upper carbon content sequence, the initial middle carbon content sequence, and the initial lower carbon content sequence, and calculate the average value of the remaining parts, which are recorded in order as the upper period average content SM, the middle period average content ZM, and the lower period average content XM; where A1 is the number of collection dates in the first time period; The average microbial residue carbon content of the first time period is calculated according to the average residue carbon formula. The average residue carbon formula is as follows: , where PMN represents the average carbon content of microbial residues in the first time period; q1, q2, and q3 are weight coefficients, q1+q2+q3=1; The average microbial residue carbon content of all the first time periods is arranged in chronological order and recorded as residue carbon content data.

5. The system for dynamic management and analysis of microbial residue carbon under global change conditions according to claim 4, characterized in that: The model management module is configured with a model management strategy, which includes: Normalize the residual carbon content data and related meteorological data according to specific data types, and scale all data sizes in the residual carbon content data and related meteorological data to [0, 1]. After completion, obtain the normalized carbon content data and normalized meteorological data in sequence; Combine the global temperature, regional temperature, and rainfall of the same first time period in the normalized meteorological data into a related feature vector, denoted as TH={h1, h2, h3}, where TH represents the related feature vector; h1, h2, and h3 represent the global temperature, regional temperature, and rainfall in order, and the meteorological characteristic data is obtained after completion; The meteorological characteristic data and the normalized carbon content data are combined and stored according to the corresponding first time period and recorded as carbon content characteristic data.

6. The system for dynamic management and analysis of microbial residue carbon under global change conditions according to claim 5, characterized in that: Model management strategies also include: Repeatedly obtain carbon content characteristic data corresponding to other land areas in the area where the managed land is located and with the same soil type as the managed land, and merge them with the carbon content characteristic data to record them as carbon content-related data; An original correlation model is constructed based on the multi-perceptron model. The original correlation model includes an input layer, a hidden processing layer, and an output layer. The number of neurons in the input layer is set to e1, the number of hidden layers in the hidden processing layer is i, the number of neurons in each hidden layer is f1-fi, and the number of neurons in the output layer is e2. The original correlation model is trained using carbon content-related data, and a residual carbon climate correlation model is obtained after completion.

7. The system for dynamic management and analysis of microbial residue carbon under global change conditions according to claim 6, characterized in that: The predictive analysis module is configured with predictive analysis strategies, which include: Obtain relevant meteorological data of the first future time length and with the first time period as the time interval, record them as future meteorological data, normalize the future meteorological data according to the data type, and combine the corresponding data of the corresponding time into relevant feature vectors to obtain future feature data; input the future feature data into the residual carbon climate related model to obtain the residual carbon content data of the first future time length and with the first time period as the time interval, mark them as initial prediction data, where the first time length is R1.

8. The system for dynamic management and analysis of microbial residue carbon under global change conditions according to claim 7, characterized in that: The data correction module is configured with a data correction strategy, which includes: The carbon content-related data were divided according to the season to which the corresponding first time period belonged, and the spring carbon content data, summer carbon content data, autumn carbon content data, and winter carbon content data were obtained. These data were input into the residual carbon climate-related model respectively, and the overall error corresponding to each season was calculated according to the average error formula. The average error formula is as follows: , where WC represents the average error, Yj represents the actual average microbial residue carbon content of the first time period, Xj represents the average microbial residue carbon content of the first time period output by the residue carbon climate correlation model, and n represents the number of samples input into the residue carbon climate correlation model; the overall error in spring, summer, autumn, and winter are obtained respectively.

9. The system for dynamic management and analysis of microbial residue carbon under global change conditions according to claim 8, characterized in that: Data correction strategies also include: For any residual carbon content data in the initial prediction data, record it as the first predicted content CY, obtain the regional temperature, rainfall and season corresponding to the first predicted content, and record them as the first temperature Y0, the first rainfall U0 and the first season respectively in order; Screening data similar to the first predicted content from the carbon content-related data includes: recording data corresponding to the first season in the carbon content-related data as carbon content data for the same season; calculating similarity values ​​between each carbon content data for the same season and the first predicted content in sequence according to a similarity formula, and sorting the data from smallest to largest to record the similarity values ​​as a similarity value sequence; the similarity formula is as follows: , where XS represents the similarity value, Y1 and U1 represent the regional temperature and rainfall in the carbon content data of the same season respectively; the carbon content data of the same season corresponding to the smallest k4 similarity values ​​in the similarity value sequence are recorded as the similarity data of the same season; All similar data from the same season are input into the residual carbon climate correlation model, and the average error corresponding to each similar data from the same season is calculated according to the average error formula, which is recorded as the seasonal individual error of the first predicted content; According to the seasonal individual error of the first predicted content and the overall error corresponding to the first season, the first predicted content is corrected using the correction formula. The correction formula is as follows: , where ZY represents the first predicted content after error correction, CY is the first predicted content, GT represents the seasonal individual error of the first predicted content, ZT represents the overall error corresponding to the first season, and q4 and q5 are the set weight coefficients; Repeat the error correction for all the residual carbon content data in the initial prediction data, and obtain the residual carbon prediction data after completion.

Images