Method for estimating biomass of alpine grassland ecosystem
By using random forest models to combine biomass, altitude and climate data, the problem of lack of biomass estimation methods in alpine grasslands is solved, and accurate estimation of biomass in alpine grasslands is achieved, and more reasonable grassland management and ecosystem improvement are supported.
Patent Information
- Application Number
- CN202510031940.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-09
- Publication Date
- 2025-02-14
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
There is a lack of special estimation methods for biomass of alpine grasslands in the prior art, which leads to the inability to effectively manage grassland resources and affects the ecological balance of grasslands.
By obtaining existing grassland area data and alpine grass sample point area data, a random forest model is used to train with biomass, altitude and climate data, the biomass estimate of alpine grassland is output, and the model performance is verified through linear fitting, and finally the model is achieved. Accurate estimation of biomass in alpine grasslands.
Accurate estimation of biomass in alpine grasslands has been achieved, more reasonable animal husbandry management is supported, grassland ecosystems have been improved, and the problem of the inability to effectively estimate biomass in the existing technology is solved.
Smart Images
Figure CN119443529A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of computer systems based on specific computing models, and in particular to a method for estimating biomass of an alpine grassland ecosystem. Background Art
[0002] The aboveground biomass of grassland is one of the main indicators of national grassland monitoring. Accurate estimation and monitoring of plateau grassland biomass is an important prerequisite for efficient management of grazing grasslands and balance of grass-livestock supply and demand. The existing biomass estimation methods include forest biomass estimation methods, desert biomass estimation methods, lake biomass estimation methods, etc. However, there is no biomass estimation method for alpine grasslands, which makes it impossible to guide herders to effectively and reasonably raise livestock, thereby effectively maintaining the ecological balance of grasslands. Summary of the invention
[0003] The purpose of the present application is to provide a method for estimating the biomass of an alpine grassland ecosystem, which solves the problem that the prior art cannot provide a method for estimating the biomass of an alpine grassland specifically.
[0004] The purpose of the present application is achieved through the following technical solutions, including: obtaining existing grassland area data and using it as a training set, obtaining alpine grassland sample area data and using it as a test set; inputting the biomass data, altitude data and climate data corresponding to each existing grassland area in the training set as a group of temperature change factors into the random forest model for training, and outputting the grassland biomass prediction value; setting the alpine grassland biomass parameter value for the training set based on the grassland biomass prediction value, and the alpine grassland biomass parameter value is used to filter the biomass parameters of the sample area of the test set; inputting the biomass data, altitude data and climate data corresponding to each sample area in the test set as a group of temperature change factors into the random forest model in sequence Forest model, random forest model is trained on the training set based on the initial filtering of the alpine grassland biomass parameter values to obtain the prediction model; the temperature change parameters corresponding to each sample area are obtained, and the temperature change parameters and biomass data are input into the prediction model for training, and the alpine grassland biomass estimation value is output; the prediction model is verified by linear fitting: the alpine grassland biomass estimation value and the grassland biomass prediction value are linearly fitted to generate a linear model; according to the results of the linear model, the prediction model performance is evaluated, and the estimation result is determined according to the evaluation result; the optimal alpine grassland biomass estimation result is defined as the prediction data, and the biomass data in the next year is predicted based on the prediction data to obtain the prediction result.
[0005] Preferably, the training set trains the model, adjusts the model parameters, obtains fitting training data, and optimizes filtering capabilities.
[0006] Preferably, the process of acquiring the climate data is specifically as follows: acquiring the daily ERA5-land climate data of each existing grassland area from one year before the data acquisition date to the data acquisition date; averaging the daily temperatures in the weekly ERA5-land climate data to obtain the weekly mean temperature, and using the weekly mean temperature as the weekly climate data.
[0007] Preferably, the alpine grassland biomass parameter value is used to filter the biomass parameters of the test set sample area: if one or more of the biomass data and climate data corresponding to the existing grassland area is inferior to the biomass data and climate data corresponding to the sample area, the alpine grassland biomass parameter value is 2; if the biomass data and climate data corresponding to the existing grassland area are both better than the biomass data and climate data corresponding to the sample area, the alpine grassland biomass parameter value is 0.
[0008] Preferably, the method of obtaining the temperature change parameters corresponding to each sample area is as follows: the temperature change parameters and biomass data are input into the prediction model for training, the data with the alpine grassland biomass parameter value of 0 is filtered out, and the prediction model outputs a set of alpine grassland biomass estimation values.
[0009] Preferably, the verification of the prediction model by linear fitting is specifically as follows: applying the prediction model to an independent verification set, the verification set including biomass data, altitude data, climate data corresponding to the existing grassland area and biomass data, altitude data, climate data corresponding to the sample area; calculating the difference between the grassland biomass prediction value of the calculation model and the alpine grassland biomass estimation value.
[0010] Preferably, the linear fitting formula is:
[0011]
[0012] in, , Among them, P i is the estimated value of alpine grassland biomass; f(t) is the fitted value of alpine grassland biomass; a and b represent the fitting parameters; R 2 is the goodness of fit, t is time, SST is the total sum of square deviations of alpine grassland biomass, and N is the number of times the regional parameters of the sample point were obtained.
[0013] Preferably, the evaluation result is specifically: based on the result of linear fitting, the performance of the prediction model is evaluated. If the performance of the prediction model does not meet the preset parameters, the prediction model is returned to adjust the temperature change parameters corresponding to the sample area of the model; when the performance of the prediction model meets the preset parameters, the predicted value of the model is the optimal alpine grassland biomass estimation result.
[0014] Preferably, the prediction of biomass data within the next year based on the predicted data is specifically as follows: obtaining atmospheric carbon dioxide content data of the sample area from one year before the data acquisition date to the data acquisition date; averaging the atmospheric carbon dioxide content data of each day in the weekly atmospheric carbon dioxide content data to obtain a weekly average of the atmospheric carbon dioxide content data, and using the weekly average of the atmospheric carbon dioxide content data as the weekly temperature change data; all weekly temperature change data are collectively used as a group of temperature change factors and sequentially input into the random forest model for training, and outputting future grassland biomass prediction values.
[0015] The technical solution of the present application has at least the following advantages and beneficial effects: The present application provides a method for estimating the biomass of an alpine grassland ecosystem. The method optimizes the existing biomass estimation method, and starts with the biomass data, altitude data, climate data and other parameters of existing temperate grasslands. After environmental simulation, the biomass is estimated after multiple data filtering. The method can achieve the estimation of the biomass of the alpine grassland ecosystem, so that the alpine grassland can also effectively improve the greenhouse effect and repair the grassland ecosystem according to the changes in biomass, and solve the problem that the existing technology cannot provide a method for estimating the biomass of the alpine grassland. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 The flowchart of the biomass estimation method of the present application is shown in FIG. DETAILED DESCRIPTION
[0017] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0018] In the existing technology, biomass estimates have been made for forests, lakes, rice fields, wheat fields and other areas, but the estimation of grassland biomass is also very important, especially the biomass of alpine grasslands. However, there is still no method that can more efficiently estimate the biomass of alpine grasslands.
[0019] Please refer to Figure 1 The present application provides a method for estimating the biomass of an alpine grassland ecosystem. The method for estimating the biomass of an alpine grassland ecosystem is mainly used to solve the problem that the prior art cannot provide a method for estimating the biomass of an alpine grassland; it also solves the problem that the prior art cannot effectively predict biomass changes.
[0020] The implementation process of the present application includes the following steps: obtaining existing grassland area data and using it as a training set, obtaining alpine grassland sample area data and using it as a test set; inputting the biomass data, altitude data and climate data corresponding to each existing grassland area in the training set as a set of temperature change factors into the random forest model for training, and outputting the grassland biomass prediction value; setting the alpine grassland biomass parameter value for the training set based on the grassland biomass prediction value, and the alpine grassland biomass parameter value is used to filter the biomass parameters of the sample area of the test set; inputting the biomass data, altitude data and climate data corresponding to each sample area in the test set as a set of temperature change factors into the random forest model in sequence The random forest model is trained on the training set based on the initial filtering of the alpine grassland biomass parameter values to obtain a prediction model; the temperature change parameters corresponding to each sample area are obtained, and the temperature change parameters and biomass data are input into the prediction model for training, and the estimated value of alpine grassland biomass is output; the prediction model is verified by linear fitting: the estimated value of alpine grassland biomass is linearly fitted with the predicted value of grassland biomass to generate a linear model; according to the results of the linear model, the performance of the prediction model is evaluated, and according to the evaluation results, the estimation result is determined; the optimal alpine grassland biomass estimation result is defined as the prediction data, and the biomass data in the next year is predicted based on the prediction data to obtain the prediction result.
[0021] Among them, the existing grassland area data is obtained through a local database or a cloud database, and the data content mainly includes remote sensing images of existing temperate grasslands, biomass data, altitude data, climate data, biological species, soil temperature, etc. This embodiment mainly predicts the sample area data of alpine grasslands through the existing grassland area biomass data, altitude data, and climate data.
[0022] Among them, the acquisition of the sample area data of the alpine grassland is mainly obtained through measurement. Since the area of the alpine grassland is large, when acquiring, the vegetation condition of the entire alpine grassland is first collected by using a drone, and then the vegetation condition is evaluated based on the collected images. The alpine grassland condition is mainly divided into three types: excellent, normal and poor according to the color of the plants. The alpine grassland area is divided into regions. To ensure the high efficiency of the prediction, the entire grassland is divided into three blocks. Three sample points of excellent, normal and poor are determined in each area to collect biomass on the spot. The data content of the sample area data of the alpine grassland mainly includes remote sensing images, biomass data, altitude data, climate data, biological species, soil temperature, etc. of the sample area of the alpine grassland. This embodiment mainly uses the biomass data, altitude data and climate data of the sample area of the alpine grassland to predict the biomass of the entire alpine grassland.
[0023] Among them, the alpine grassland is quite different from the existing grassland. The main reasons for the difference are the different altitudes, climate conditions and biomass, which lead to differences in the biological states of grasslands in different locations. The difference in temperature is a key factor affecting the survival status and number of organisms. Therefore, biomass data, altitude data and climate data are used as temperature variation factors to train the random forest model. The environmental state of the alpine grassland is simulated under the training of the temperature variation factor, and then the grassland biomass is predicted to obtain the grassland biomass prediction value.
[0024] Among them, the alpine grassland biomass parameter value is set, and the biomass parameter of the test set sample area is filtered by the alpine grassland biomass parameter value, so as to improve the accuracy of the biomass parameter of the entire test set sample area, exclude invalid data, and reduce the processing cycle. The difference between the training of the training set by the random model and the training of the test set by the random model is that the training of the test set by the random model is based on the results of the training set training by the random model, and is completed through a progressive relationship.
[0025] Among them, the grassland biomass prediction value obtained by training the training set sets the alpine grassland biomass parameter value. The biomass data, altitude data, climate data and alpine grassland biomass parameter value are all foreseeable variables, and are all predictive variables that affect the change of alpine grassland biomass. Therefore, after the initial filtering of the biomass data and climate data based on the alpine grassland biomass parameter value, the test set is trained in the random model under the common predictive variables of the temperature change factor and the alpine grassland biomass parameter value to obtain the prediction model.
[0026] The temperature change parameters of the sample area are mainly obtained by accessing the Internet database and field measurements. Since temperature is the main factor that causes the difference in biological state between alpine grassland and existing temperate grassland, the temperature change parameters and biomass data corresponding to the sample area are input into the prediction model for training. The biomass of the alpine grassland is estimated by simulating the environmental changes of the existing temperate grassland, and the estimated biomass of the alpine grassland is obtained.
[0027] Among them, the estimated value of alpine grassland biomass is linearly fitted with the predicted value of grassland biomass to ensure the accuracy of the estimated value of alpine grassland biomass.
[0028] Among them, the linear fitting method is used to verify the prediction model. The linear fitting method is used to fit the grassland biomass prediction value corresponding to each alpine grassland biomass estimation value to obtain the relationship between the alpine grassland biomass estimation value and the grassland biomass prediction value, so as to evaluate the prediction performance of the model and obtain the best alpine grassland biomass estimation value.
[0029] Among them, the optimal alpine grassland biomass estimation result is used as the prediction data, and the biomass of the alpine grassland in the next year is predicted based on the prediction data, so as to effectively plan the future livestock management method and ensure that the biomass of the alpine grassland is always in a normal state.
[0030] Based on the above content, in some embodiments, the training set is also used to train the model, adjust the model parameters, obtain fitting training data, and optimize the filtering ability. By adjusting the parameters of the model, the simulation of the alpine grassland environment is optimized to ensure that better training data can be obtained and the filtering ability can be optimized, thereby better realizing the training of the test set.
[0031] In some embodiments, the process of obtaining climate data is specifically as follows: obtaining the daily ERA5-land climate data of each existing grassland area from one year before the data acquisition date to the data acquisition date; averaging the daily temperatures in the weekly ERA5-land climate data to obtain the weekly mean temperature, and using the weekly mean temperature as the weekly climate data.
[0032] It should be noted that the climate data was obtained by visiting the Climate Data Store (CDS) website, selecting temperature as the meteorological parameter, selecting the data set and variables, determining the longitude and latitude of the alpine grassland sample area through the global positioning system, selecting the time range, longitude and latitude, data format, etc., downloading the relevant data, and using the CDS Toolbox for online processing. According to the time range, the weekly average of the temperature was downloaded, and the downloaded weekly average of the temperature was used as the weekly climate data input into the model to train the training set.
[0033] In some embodiments, the alpine grassland biomass parameter value is used to filter the biomass parameters of the test set sample area: if one or more of the biomass data and climate data corresponding to the existing grassland area is inferior to the biomass data and climate data corresponding to the sample area, the alpine grassland biomass parameter value is 2; if the biomass data and climate data corresponding to the existing grassland area are both better than the biomass data and climate data corresponding to the sample area, the alpine grassland biomass parameter value is 0.
[0034] It should be noted that since the altitude data of the alpine grassland is higher than the altitude of the existing grassland area, the altitude data is not used as the basis for the alpine grassland biomass parameter value. Considering that the environment of the alpine grassland is worse than that of the existing grassland area, the biomass data and climate data of the existing grassland area are inferior to the sample area as the basis for the alpine grassland biomass parameter value, so as to ensure that the data selected by the alpine grassland biomass parameter value is more referenceable and more practical.
[0035] In some embodiments, the obtaining of the temperature change parameters corresponding to each sample area is specifically to obtain the temperature change parameters corresponding to each sample area, input the temperature change parameters and biomass data into the prediction model for training, filter out the data with the alpine grassland biomass parameter value of 0, and the prediction model outputs a set of alpine grassland biomass estimation values.
[0036] Among them, the biomass data and climate data are initially filtered based on the biomass parameter values of alpine grassland to obtain the prediction model, and the biomass data and climate data are secondary filtered based on the biomass parameter values of alpine grassland. After filtering out the data with the biomass parameter value of alpine grassland being 0, the biomass of the alpine grassland is estimated. The two filterings at different stages can realize the effective extraction and use of data, and further improve the accuracy of the estimation of alpine grassland biomass.
[0037] In some embodiments, the verification of the prediction model by linear fitting is specifically as follows: the prediction model is applied to an independent verification set, the verification set includes biomass data, altitude data, climate data corresponding to the existing grassland area and biomass data, altitude data, climate data corresponding to the sample area; the difference between the grassland biomass prediction value of the calculation model and the alpine grassland biomass estimation value.
[0038] In some embodiments, the linear fitting is specifically: linear fitting is performed on the estimated alpine grassland biomass and the predicted grassland biomass to generate a linear model; the goal of the linear fitting is to minimize the difference between the estimated alpine grassland biomass and the predicted grassland biomass; the result of the linear fitting can be used to evaluate the bias and variance of the prediction model, as well as to minimize the correlation between the estimated alpine grassland biomass and the predicted grassland biomass.
[0039] A large deviation means that the model cannot correctly capture the relationship in the data, causing the prediction results to deviate from the true value; the larger the variance, the more discrete the data points; the smaller the variance, the more concentrated the data points. A large variance may increase the sensitivity and instability of the model to different data sets, and further adjustments to the alpine grassland climate data are needed.
[0040] Linear fitting was used to verify the prediction model, and the model was improved and optimized in combination with the biomass data, altitude data, and climate data of the sample area to improve the accuracy of the estimation results.
[0041] In some embodiments, the evaluation results are specifically: based on the results of linear fitting, the performance of the prediction model is evaluated. If the performance of the prediction model does not meet the preset parameters, the prediction model is returned to adjust the temperature change parameters corresponding to the sample area of the model; when the performance of the prediction model meets the preset parameters, the predicted value of the model is the optimal alpine grassland biomass estimation result.
[0042] In some embodiments, the linear fitting formula is:
[0043]
[0044] in, , Among them, P i is the estimated value of alpine grassland biomass; f(t) is the fitted value of alpine grassland biomass; a and b represent the fitting parameters; R 2 is the goodness of fit, t is time, SST is the total sum of square deviations of alpine grassland biomass, and N is the number of times the regional parameters of the sample point were obtained.
[0045] In some embodiments, predicting the biomass data within the next year based on the predicted data is specifically as follows: obtaining atmospheric carbon dioxide content data of the sampling point area from one year before the data acquisition date to the data acquisition date; The atmospheric carbon dioxide content data of each day in the weekly atmospheric carbon dioxide content data are averaged to obtain the weekly average value of the atmospheric carbon dioxide content data, and the weekly average value of the atmospheric carbon dioxide content data is used as the weekly temperature change data; All weekly temperature change data are taken as a set of temperature change factors and input into the random forest model for training in turn to output the predicted value of future grassland biomass.
[0046] It should be noted that the weekly average of atmospheric carbon dioxide content data is calculated, and all weekly average values of atmospheric carbon dioxide content data are input into the random forest model in sequence as a set of temperature change factors for training, and then the output is used to predict the future grassland biomass prediction value.
[0047] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or basic features of the present application. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present application is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present application. Any figure mark in the claims should not be regarded as limiting the claims involved.
[0048] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.
Claims
1. A method for estimating biomass of an alpine grassland ecosystem, characterized in that: include: Obtain existing grassland area data and use it as a training set, and obtain sample area data of alpine grassland and use it as a test set; The biomass data, altitude data and climate data corresponding to each existing grassland area in the training set are input into the random forest model as a set of temperature change factors for training, and the predicted value of grassland biomass is output; Based on the grassland biomass prediction value, the alpine grassland biomass parameter value is set for the training set, and the alpine grassland biomass parameter value is used to filter the biomass parameters of the sample point area of the test set; The biomass data, altitude data and climate data corresponding to each sample area in the test set are input into the random forest model as a set of temperature change factors. The random forest model trains the training set based on the initial filtering of the biomass parameter values of the alpine grassland to obtain a prediction model. Obtain the temperature change parameters corresponding to each sample area, input the temperature change parameters and biomass data into the prediction model for training, and output the estimated value of alpine grassland biomass; The prediction model was verified by linear fitting: the estimated biomass of alpine grassland was linearly fitted with the predicted biomass of grassland to generate a linear model; Based on the results of the linear model, the performance of the prediction model is evaluated, and based on the evaluation results, the estimation results are determined; The optimal alpine grassland biomass estimation result is defined as the predicted data, and the biomass data in the next year is predicted based on the predicted data to obtain the predicted result.
2. The estimation method according to claim 1, characterized in that: The training set trains the model, adjusts model parameters, obtains fitting training data, and optimizes filtering capabilities.
3. The estimation method according to claim 1, characterized in that: The process of obtaining climate data is specifically as follows: Obtain daily ERA5-land climate data for each existing grassland area from 1 year before the data acquisition date to the data acquisition date; The daily temperatures in the weekly ERA5-land climate data are averaged to obtain the weekly mean temperature, which is used as the weekly climate data.
4. The estimation method according to claim 1, characterized in that: The alpine grassland biomass parameter value is used to filter the biomass parameter of the test set sample area, specifically: If one or more of the biomass data and climate data corresponding to the existing grassland area is inferior to the biomass data and climate data corresponding to the sampling area, the alpine grassland biomass parameter value is 2; If the biomass data and climate data corresponding to the existing grassland area are better than those corresponding to the sample area, the biomass parameter value of the alpine grassland is 0.
5. The estimation method according to claim 1, characterized in that: The method of obtaining the temperature change parameters corresponding to each sample area is as follows: the temperature change parameters and biomass data are input into the prediction model for training, and the data with the alpine grassland biomass parameter value of 0 is filtered out, and the prediction model outputs a set of alpine grassland biomass estimation values.
6. The estimation method according to claim 1, characterized in that: The verification of the prediction model by linear fitting is specifically as follows: The prediction model was applied to an independent verification set, which contained biomass data, altitude data, climate data corresponding to the existing grassland area and biomass data, altitude data, climate data corresponding to the sample area; the difference between the grassland biomass prediction value of the calculation model and the estimated value of the alpine grassland biomass was calculated.
7. The estimation method according to claim 6, characterized in that: The linear fitting formula is: ； ； in, , Among them, P i is the estimated value of alpine grassland biomass; f(t) is the fitted value of alpine grassland biomass; a and b represent the fitting parameters; R 2 is the goodness of fit, t is the period for obtaining the alpine grassland biomass, SST is the total sum of square deviations of the alpine grassland biomass, and N is the number of times the sampling point area parameters are obtained.
8. The estimation method according to claim 1, characterized in that: The evaluation results are as follows: According to the results of linear fitting, the performance of the prediction model is evaluated. If the performance of the prediction model does not meet the preset parameters, the prediction model is returned to adjust the temperature change parameters corresponding to the sample area of the model; When the performance of the prediction model meets the preset parameters, the predicted value of the model is the optimal alpine grassland biomass estimation result.
9. The estimation method according to claim 1, characterized in that: The prediction of biomass data in the next year based on the prediction data is specifically as follows: Obtain atmospheric carbon dioxide content data in the sampling area from one year before the data acquisition date to the data acquisition date; The atmospheric carbon dioxide content data of each day in the weekly atmospheric carbon dioxide content data are averaged to obtain the weekly average value of the atmospheric carbon dioxide content data, and the weekly average value of the atmospheric carbon dioxide content data is used as the weekly temperature change data; All weekly temperature change data are taken as a set of temperature change factors and input into the random forest model for training in turn to output the predicted value of future grassland biomass.