CN118297222A

CN118297222A - Remote sensing prediction method and device for vegetation net primary productivity

Info

Publication number: CN118297222A
Application number: CN202410381278.1A
Authority: CN
Inventors: 赵东升
Original assignee: Institute of Geographic Sciences and Natural Resources of CAS
Current assignee: Institute of Geographic Sciences and Natural Resources of CAS
Priority date: 2024-04-01
Filing date: 2024-04-01
Publication date: 2024-07-05
Anticipated expiration: 2044-04-01

Abstract

The invention provides a remote sensing prediction method and a device for vegetation net primary productivity, and relates to the technical field of remote sensing prediction, wherein the method comprises the following steps: acquiring longitude and latitude coordinates corresponding to the above-ground vegetation net primary productivity and the underground vegetation net primary productivity data set respectively, and extracting the global years average vegetation net primary productivity, the years average precipitation and the annual precipitation anomaly rate at the observation point in the global data set; building and training a machine learning model according to the average vegetation net primary productivity for many years, the average precipitation amount for many years and the annual precipitation anomaly rate; according to a machine learning model, global annual average vegetation net primary productivity, annual average precipitation and annual precipitation anomaly rate data sets are input to generate predicted global data sets of overground vegetation net primary productivity and underground vegetation net primary productivity. The invention solves the problem of mismatching of time dimension in the prior art, and greatly improves the model performance.

Description

Remote sensing prediction method and device for vegetation net primary productivity

Technical Field

The invention relates to the technical field of remote sensing prediction, in particular to a remote sensing prediction method and device for vegetation net primary productivity.

Background

Net Primary Productivity (NPP) refers to the fraction of plant carbon fixed by photosynthesis minus its own respiration consumption, also known as net primary productivity. NPP is a fundamental indicator of plant dynamics and is also an important component of the global carbon cycle. Above ground net primary productivity (ANPP) is an energy source for heterotrophs including humans and livestock, and below ground net primary productivity (BNPP) is a critical carbon input for carbon in the soil. Accurate estimation of global ANPP and BNPP is significant for a thorough understanding of global carbon balance, and for a comprehensive realization of sustainable development goals.

Currently, model methods are used to estimate NPP at the global scale, including empirical model methods (e.g., miami models), parametric model methods (e.g., light energy utilization models), process model methods (e.g., BIOME _bgc models), and the like. However, these methods have difficulty achieving above-ground and below-ground partitioning of NPPs, and accurate estimation ANPP and BNPP on a global scale remains a challenge.

In the prior art, a global ANPP and BNPP estimation scheme based on field actual measurement data and a machine learning method is provided. But this solution has the following drawbacks:

(1) The environmental factors used are all variables that do not change over time, such as a mean value over years or the results of a survey, whereas the field ANPP and BNPP are measured at different points in time, and a mismatch in the time dimension results in limited reliability of the estimated results.

Disclosure of Invention

The invention aims to solve the technical problem of providing a remote sensing prediction method and device for vegetation net primary productivity, solving the problem of mismatching of time dimension in the existing scheme and greatly improving model performance.

In order to solve the technical problems, the technical scheme of the invention is as follows:

in a first aspect, a method of remote sensing prediction of net primary productivity of vegetation, the method comprising:

Acquiring a global average vegetation net primary productivity product for many years generated based on a parameter model;

acquiring environment covariates according to the global annual average vegetation net primary productivity product, wherein the environment covariates comprise global annual precipitation data and annual average precipitation data;

calculating annual precipitation abnormality rate according to global annual precipitation data;

acquiring a ground vegetation net primary productivity data set of ground actual measurement and an underground vegetation net primary productivity data set;

Acquiring longitude and latitude coordinates corresponding to the above-ground vegetation net primary productivity and the underground vegetation net primary productivity data set respectively, and extracting the global years average vegetation net primary productivity, the years average precipitation and the annual precipitation anomaly rate at the observation point in the global data set;

Building and training a machine learning model according to the average vegetation net primary productivity for many years, the average precipitation amount for many years and the annual precipitation anomaly rate;

According to a machine learning model, global annual average vegetation net primary productivity, annual average precipitation and annual precipitation anomaly rate data sets are input to generate predicted global data sets of overground vegetation net primary productivity and underground vegetation net primary productivity.

Further, obtaining a global years-average vegetation net primary productivity product generated based on the parametric model, comprising:

determining a data source of a global years of average vegetation net primary productivity product;

determining the relation between vegetation growth and environmental factors according to a data source, and constructing a parameter model;

Acquiring input data and inputting the input data into the parametric model so that the parametric model calculates the net primary productivity of the year-by-year vegetation;

screening the net primary productivity of vegetation for years to obtain screening data;

Aligning and space-time matching the screening data to calculate a mean value over a plurality of years;

Based on the years' average, a years-average vegetation net primary productivity product is generated.

Further, obtaining an environmental covariate from the global year-average vegetation net primary productivity product, the environmental covariate comprising global year precipitation data and year-average precipitation data, comprising:

estimating the annual precipitation magnitude of the missing time point by a linear interpolation method according to the known time point data to obtain alignment data;

Determining geographic units and time units for matching based on the alignment data;

Extracting annual precipitation values corresponding to each selected geographic cell and time cell from the alignment data, and extracting an annual average precipitation value corresponding to each geographic cell;

The annual precipitation magnitude value and the annual average precipitation magnitude value are matched with the global annual average vegetation net primary productivity value of the corresponding geographic unit and the time unit to obtain a matching result, wherein the matching result comprises that each geographic position and each time point uniquely correspond to one annual precipitation magnitude value, one annual average precipitation magnitude value and one global annual average vegetation net primary productivity value.

Further, calculating annual precipitation anomaly rate from global annual precipitation data, comprising:

By passing through Calculating the average annual precipitation over a long periodWhere w _i is the weight of the i-th year, β is the decay factor, P _i is the annual precipitation of the i-th year, and N is the number of years used to calculate the long term average;

According to the average annual precipitation over a long period of time Calculating a standard deviation s _P of the long-term annual precipitation, wherein gamma _i is the robustness weight of the ith year;

According to the standard deviation of annual precipitation And calculating the precipitation abnormality Rate _i, wherein,Is the absolute deviation of the annual precipitation from the weighted long-term average annual precipitation, delta being the adjustment parameter.

Further, obtaining a ground measured net primary productivity of the above-ground vegetation and a ground vegetation data set comprising:

Determining a research area and target vegetation;

According to the actual condition of the research area, the feasibility and the potential risk of the field measurement are estimated to obtain an evaluation result; determining an in-field measurement scheme according to the evaluation result, and determining a time table according to the in-field measurement scheme;

According to the on-site measurement scheme, carrying out on-site vegetation and underground vegetation biomass measurement on each sample party to obtain measurement data;

The measured data is converted to a global average vegetation net primary productivity over many years as needed.

Further, acquiring longitude and latitude coordinates corresponding to the above-ground vegetation net primary productivity and the below-ground vegetation net primary productivity data set respectively, and extracting a global years average vegetation net primary productivity, a years average precipitation amount and an annual precipitation anomaly rate at an observation point in the global data set, wherein the method comprises the following steps:

Acquiring data sets of net primary productivity of overground vegetation and net primary productivity of underground vegetation and longitude and latitude coordinates thereof

Acquiring an observation data set of the net primary productivity of the overground vegetation and the net primary productivity of the underground vegetation;

formatting the observation data set to obtain formatted data;

Extracting and sorting longitude and latitude coordinates of an observation point in each formatted data;

And acquiring a global data set, performing space matching with the global data set according to longitude and latitude coordinates of the observation points, and extracting years of average vegetation net primary productivity, years of average precipitation and years of precipitation anomaly values corresponding to each observation point.

Further, after inputting the global annual average vegetation net primary productivity, annual average precipitation and annual precipitation anomaly rate data sets according to the machine learning model to generate the predicted global data sets of above-ground vegetation net primary productivity and below-ground vegetation net primary productivity, further comprising:

The predicted global data sets of the net primary productivity of the overground vegetation and the net primary productivity of the underground vegetation are respectively compared with the data sets of the net primary productivity of the overground vegetation and the net primary productivity of the underground vegetation which are actually measured on the ground to obtain comparison results;

and according to the comparison result, evaluating the prediction performance of the machine learning model to obtain an evaluation result.

In a second aspect, a remote sensing prediction apparatus for net primary productivity of vegetation, comprising:

The acquisition module is used for acquiring a global average vegetation net primary productivity product for many years generated based on the parameter model; acquiring environment covariates according to the global annual average vegetation net primary productivity product, wherein the environment covariates comprise global annual precipitation data and annual average precipitation data; calculating annual precipitation abnormality rate according to global annual precipitation data; acquiring a ground vegetation net primary productivity data set of ground actual measurement and an underground vegetation net primary productivity data set;

The processing module is used for acquiring longitude and latitude coordinates corresponding to the above-ground vegetation net primary productivity and the underground vegetation net primary productivity data set respectively, and extracting the global years average vegetation net primary productivity, the years average precipitation and the annual precipitation abnormal rate at the observation point in the global data set; building and training a machine learning model according to the average vegetation net primary productivity for many years, the average precipitation amount for many years and the annual precipitation anomaly rate; according to a machine learning model, global annual average vegetation net primary productivity, annual average precipitation and annual precipitation anomaly rate data sets are input to generate predicted global data sets of overground vegetation net primary productivity and underground vegetation net primary productivity.

In a third aspect, a computing device includes:

one or more processors;

and a storage device for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement the above-described methods.

In a fourth aspect, a computer readable storage medium stores a program that when executed by a processor implements the above method.

The scheme of the invention at least comprises the following beneficial effects:

the annual precipitation data and the annual precipitation abnormal rate are calculated, so that the change of the vegetation net primary productivity along with time can be better captured, the problem of mismatching of time dimension caused by using environmental factors which do not change along with time in the prior art is solved, and the reliability of an estimation result is improved.

The invention successfully realizes the division of the above-ground vegetation net primary productivity (ANPP) and the underground vegetation net primary productivity (BNPP) on the global scale by using environment covariates such as the global multi-year average vegetation net primary productivity, the multi-year average precipitation amount, the annual precipitation abnormal rate and the like through a machine learning model.

By combining the global average vegetation net primary productivity product for many years and ground actual measurement data generated by the parametric model, ANPP and BNPP in the global scope can be predicted more accurately. In addition, reliability and stability of the predictions are further enhanced by machine learning models.

Drawings

Fig. 1 is a schematic flow chart of a remote sensing prediction method for vegetation net primary productivity according to an embodiment of the present invention.

FIG. 2 is a schematic diagram of a remote sensing prediction apparatus for net primary productivity of vegetation according to an embodiment of the present invention.

Detailed Description

Exemplary embodiments of the present disclosure will be described more closely below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

As shown in fig. 1, an embodiment of the present invention proposes a remote sensing prediction method of net primary productivity of vegetation, the method comprising:

Step 11, acquiring a global average vegetation net primary productivity product for many years generated based on a parameter model;

Step 12, acquiring environment covariates according to the global average vegetation net primary productivity product for many years, wherein the environment covariates comprise global annual precipitation data and average annual precipitation data;

Step 13, calculating annual precipitation abnormal rate according to global annual precipitation data;

Step 14, acquiring a ground vegetation net primary productivity data set of ground actual measurement and a ground vegetation net primary productivity data set of underground vegetation;

Step 15, acquiring longitude and latitude coordinates corresponding to the above-ground vegetation net primary productivity and the underground vegetation net primary productivity data set respectively, and extracting the global years average vegetation net primary productivity, the years average precipitation and the annual precipitation anomaly rate at the observation point in the global data set;

step 16, constructing and training a machine learning model according to the global average vegetation net primary productivity for many years, the average precipitation amount for many years and the annual precipitation abnormal rate;

Step 17, inputting a global data set of annual average vegetation net primary productivity, annual average precipitation and annual precipitation anomaly rate on the global scale according to a machine learning model to generate a predicted global data set of overground vegetation net primary productivity and underground vegetation net primary productivity.

According to the embodiment of the invention, the change of the vegetation net primary productivity along with time can be better captured through annual precipitation data and annual precipitation abnormal rate calculation, and the problem of mismatching of time dimension caused by using environmental factors which do not change along with time in the prior art is solved, so that the reliability of an estimation result is improved. The invention successfully realizes the division of the above-ground vegetation net primary productivity (ANPP) and the underground vegetation net primary productivity (BNPP) on the global scale by using environment covariates such as the global multi-year average vegetation net primary productivity, the multi-year average precipitation amount, the annual precipitation abnormal rate and the like through a machine learning model. By combining the global average vegetation net primary productivity product for many years and ground actual measurement data generated by the parametric model, ANPP and BNPP in the global scope can be predicted more accurately. In addition, reliability and stability of the predictions are further enhanced by machine learning models.

In a preferred embodiment of the present invention, the step 11 may include:

step 111, determining a data source of a global years of average vegetation net primary productivity product;

Step 112, determining the relation between vegetation growth and environmental factors according to the data sources, and constructing a parameter model;

Step 113, obtaining input data and inputting the input data into the parameter model so that the parameter model calculates the net primary productivity of the year-by-year vegetation;

Step 114, screening the net primary productivity of the vegetation for years to obtain screening data;

Step 115, aligning and space-time matching the screening data to calculate a mean value of years;

Step 116, generating a years-average vegetation net primary productivity product based on the years-average.

In an embodiment of the present invention, the data sources of step 111 include satellite remote sensing images, data from ground observation sites, weather datasets, and other relevant environmental data. The data source is helpful to ensure the accuracy and consistency of the subsequent analysis, and the use of the data source can improve the accuracy of NPP estimation, thereby providing more valuable information for global carbon recycling and ecosystem research. Step 112 will analyze statistical relationships between vegetation growth (represented by NPP) and environmental factors (e.g., temperature, precipitation, radiation, etc.) based on the selected data sources, and based on these relationships, construct a parametric model that can predict vegetation NPP under different environmental conditions. The construction of the parametric model is helpful for understanding the complex relationship between vegetation growth and environmental factors and improving the accuracy and generalization capability of NPP prediction. Step 113 collects the input data required for the model, including year-by-year environmental factor data (e.g., temperature, precipitation, etc.), which is then input into the previously constructed parametric model to calculate year-by-year vegetation NPP. Through year-by-year calculations, vegetation NPP data over time series can be obtained. Step 114 will filter the calculated years vegetation NPP data for the purpose of outlier removal, data noise reduction, or selecting data points that meet certain conditions. Data screening helps to improve data quality and reliability of analysis, and by removing outliers and noise, their impact on subsequent analysis can be reduced, thereby making the results more accurate and robust. Step 115 will align and match the screened data in time and space, ensuring that data of different years and places are compared and analyzed on the same spatio-temporal scale. By eliminating errors due to inconsistent data space-time scales, a more accurate years of average vegetation NPP can be obtained, thereby providing more valuable information for global scale carbon recycling and ecosystem research. The average vegetation NPP product over the world generated in step 116 is significant for assessing the health of the ecosystem, monitoring the impact of climate change, and developing sustainable environmental management strategies.

In another preferred embodiment of the present invention, the step 112 may include:

Step 1121, by Constructing a relation between vegetation growth (NPP) and an environmental factor, wherein β ₀ is an intercept term, B _3j is a coefficient of a j-th basis function corresponding to a 3 rd environmental factor R, B _1j is a coefficient of a j-th basis function corresponding to a1 st environmental factor T, B _2j is a coefficient of a j-th basis function corresponding to a2 nd environmental factor P, B _1j (T) is a j-th basis function corresponding to a1 st environmental factor T, B _2j (P) is a j-th basis function corresponding to a2 nd environmental factor P, B _3j (R) is a j-th basis function corresponding to a 3 rd environmental factor R, k ₁、k₂ and k ₃ are numbers of basis functions for fitting the 1 st, 2 nd and 3 rd environmental factors, respectively, e is an error term, environmental factor T is temperature, environmental factor P is precipitation, and environmental factor R is solar radiation;

In step 1122, for each predicted variable X _i (temperature, precipitation and solar radiation), a model matrix B _i is constructed, provided with n observations and k _i basis functions selected for the predicted variable X _i, the model matrix B _i being a matrix of n×k _i, each element B _ijl representing the value of the first observation corresponding to the j-th basis function, wherein the model matrix B _i is:

wherein X _il is the X _i predicted variable for the first observation;

Step 1123, by The estimation coefficient B _ij and the intercept term β ₀, where y _l is the response variable of the first observation, p is the number of predicted variables (temperature, precipitation and solar radiation), λ _i is the penalty parameter controlling the smoothness of each predicted variable, and B "_ij(x_i) is the second derivative of the basis function.

Step 1124, in order to minimize the objective function G (β ₀, b), by Iteratively updating the parameter estimates to obtain an optimal solution, alpha _m being the step size of the mth step, H ^-1 being the inverse of the hessian matrix,Is the gradient of the objective function, these quantities are recalculated in each iteration until a convergence condition is met, where m represents the current number of iterations, m +1 represents the next number of iterations,Representing an estimate of the intercept term beta ₀ at the mth iteration,The updated estimate representing the intercept term β ₀ after the m+1th iteration, b ^(m) is a vector containing estimates of all coefficients b _ij at the m-th iteration, b ^(m+1) is an updated coefficient vector containing new estimates of all coefficients b _ij after the m+1th iteration, gradient of the objective functionIs a vector composed of partial derivatives of the objective function on each parameter, and the calculation formula is as follows:

Wherein, Is the partial derivative of G (beta ₀, b) with respect to intercept term beta ₀,The partial derivative vector of the coefficient vector b comprises the partial derivative of the objective function to each element in the coefficient vector b; the hessian matrix H is a second-order partial derivative matrix of the objective function, and the calculation formula is as follows:

Wherein, Is the second partial derivative of the objective function with respect to the intercept term beta ₀,Is a mixed second partial derivative vector of the objective function with respect to the intercept term beta ₀ and the coefficient vector b,Is a transposed vector of the mixed second partial derivative of the coefficient vector b and intercept term beta ₀,Regarding the second partial derivative matrix of the coefficient vector b,Representing the partial derivative and T representing the transpose of the matrix.

In an embodiment of the present invention, step 1121 establishes a mathematical relationship between vegetation growth (NPP) and three environmental factors, temperature (T), precipitation (P), and solar radiation (R), by a formula. Wherein NPP is expressed as a sum of an intercept term, a linear combination of basis functions of three environmental factors, each fitted by a basis function (B _ij), and an error term; the invention provides a flexible model framework, which can capture possible nonlinear relations between each environmental factor and NPP, adapt to various complex data modes through the combination of basis functions, and improve the expression capacity of the model.

In the embodiment of the present invention, in step 1121, the calculation formula of B _1j (T) is:

B_1j(T)＝T^j；

where j is the order of the polynomial; the calculation formula of B _2j (P) is as follows:

B_2j(P)＝max(0,P-θ_j)；

Where θ _j is the segmentation point, B _2j(P)＝max(0,P-θ_j) is a segmentation function, also called ReLU (RECTIFIED LINEAR Unit) function, P represents an input value, which may be any real number, and in precipitation, P represents precipitation at a certain point in time, and the behavior of the function may be divided into two parts:

When P < θ _j, the function value is 0, because P- θ _j is a negative number, and max (0, negative number) =0;

when P is equal to or greater than θ _j, the function value is P- θ _j, because P- θ _j is a non-negative number and max (0, non-negative number) =non-negative number.

Solar radiation has significant daily and seasonal variations, which, like temperature, can be represented using polynomial basis functions, e.g., a combination of sinusoidal functions can be used to represent the daily and seasonal variations of solar radiation:

Where t is the number of days in a year (from 0 to 364), j is the frequency factor, controlling the periodicity of the basis function, the annual period of solar radiation can be captured.

Step 1122 builds model matrices for each of the predicted variables, and for the three predicted variables, temperature, precipitation, and solar radiation, respectively, builds model matrices (B _i) comprising values for each observation corresponding to the selected basis function, and matrixing the relationship between the predicted variables and the basis functions for mathematical operations and parameter estimation, thereby facilitating the independent construction of model matrices for each predicted variable, and helping to maintain the modularity and scalability of the model.

In the embodiment of the present invention, step 1123 estimates the coefficient (b _ij) and intercept term β ₀ of the model by minimizing the objective function G (β ₀, b), the objective function is composed of two parts, the data fitting term ensures that the model can fit the observed data well, and the smooth penalty term is used to control the complexity of the model, prevent overfitting, balance the fitting ability and complexity of the model, help to improve the generalization performance of the model, and can directly control the smoothness of the model by introducing the penalty term, thereby increasing the interpretation of the model.

In an embodiment of the present invention, in step 1123, the calculation process of the second derivative b″ _ij(x_i) of the basis function includes:

There is a cubic B-spline basis function B _ij (x) which is a cubic polynomial over a specific interval t _m,t_m+1, where t _m and t _m+1 are adjacent nodes, which cubic polynomial can be expressed as:

B_ij(x)＝a_m(x-t_m)³+b_m(x-t_m)²+c_m(x-t_m)+d_m;

over this interval, where a _m,b_m,c_m,d_m is the coefficient of the polynomial, determined by the B-spline construction and node position; calculating the second derivative of the third polynomial, and obtaining the second derivative of B _ij (x) to obtain:

the above calculation formula is the second derivative over the interval t _m,t_m+1, and this process needs to be repeated over each interval for the entire basis function, since each interval has its own polynomial representation and coefficients.

In embodiments of the present invention, a cubic B-spline basis function allows the model to capture the nonlinear relationship between the predicted and response variables, and by using different polynomial coefficients at each interval, the basis function can adapt to local variations in the data, thereby providing a more accurate fit. Calculation of the second derivative is key to controlling model smoothness, and in the penalty term of GAM, the sum of squares of the second derivatives is used to avoid overfitting, ensuring that the model's predicted curve is smooth, rather than excessively fluctuating, which helps generalize the model over unknown data. The cubic B-spline basis function and its derivatives are continuous at the nodes, which ensures that the prediction of the model is continuous throughout the definition domain, which continuity is necessary for many practical applications, as it ensures the stability and reliability of the prediction. By examining the shape of the basis function and the values of the second derivative, an intuitive understanding can be obtained as to how the predicted variable affects the response variable. For example, the sign and magnitude of the second derivative may provide information about the rate of change of the response variable. Cubic B-spline basis functions can accommodate various types of data distributions and patterns, as they are implemented by fitting different polynomials over each interval.

Step 1124 iteratively updates the parameter estimates to obtain an optimal solution, minimizes the objective function using an iterative optimization algorithm, and iteratively updates the parameter estimates until a convergence condition is met. Each iteration calculates the gradient of the objective function according to the current parameter valueAnd the inverse H ^-1 of the Heisen matrix, then updating parameters by using the information, the iterative optimization algorithm can efficiently find the minimum value of the objective function, can effectively work even under the condition that the parameter space is very complex, and can obtain high-precision parameter estimation under limited computing resources by gradually approaching the optimal solution.

In a preferred embodiment of the present invention, the step 12 may include:

Step 121, estimating the annual precipitation value of the missing time point by a linear interpolation method according to the known time point data to obtain alignment data;

step 122, determining geographic units and time units for matching according to the alignment data;

step 123, extracting annual precipitation values corresponding to each selected geographic unit and time unit from the alignment data, and extracting an annual average precipitation value corresponding to each geographic unit;

Step 124, matching the annual precipitation magnitude, the annual average precipitation magnitude and the global annual average vegetation net primary productivity value of the corresponding geographic unit and time unit to obtain a matching result, wherein the matching result comprises a annual precipitation magnitude, an annual average precipitation magnitude and a global annual average vegetation net primary productivity value uniquely corresponding to each geographic position and time point.

In the embodiment of the present invention, step 121 can estimate the annual precipitation value at the missing time point by using a linear interpolation method, so as to improve the integrity and continuity of the data set, and reduce the deviation caused by the data missing to a certain extent. Step 122, determining the geographic and temporal units facilitates a clear analysis of the spatial and temporal dimensions, making the study more targeted and operable, and the clear geographic and temporal unit partitioning facilitates a more accurate matching of the data of the different data sources. Step 123, extracting annual precipitation magnitude and annual average precipitation magnitude of specific geography and time units, facilitating comparative analysis of different regions and different time periods, and providing a basis for matching the subsequent value with the global annual average vegetation net primary productivity value. Step 124, the matching result includes geographic location, time point, annual precipitation, annual average precipitation and global annual average vegetation net primary productivity value, which provides abundant information for comprehensive analysis; by ensuring that each geographic position and each time point uniquely correspond to one set of data, the accuracy and the reliability of analysis are improved; the matched data structure supports analysis of vegetation net primary productivity from different dimensions (e.g., space, time, precipitation, etc.).

In another preferred embodiment of the present invention, the step 123 may include:

Step 1231, extracting annual precipitation magnitude by P _ij＝α×Lat_i+β×Lon_i+γ×A_k+δ×Season(T_j)+∈_ijk, wherein P _ij represents annual precipitation prediction values under geographic unit (latitude and longitude), time unit T _j (Season) and altitude a _k, lat _i represents latitude value of geographic unit, lon _i represents longitude value of geographic unit, a _k represents altitude value of geographic unit, season (T _j) represents mapping time unit T _j (such as month) onto corresponding seasonal category; alpha, beta, gamma and delta are regression coefficients of latitude, longitude, altitude and season, respectively, and epsilon _ijk represents a random error term;

step 1232, by Calculating an average precipitation magnitude over a plurality of years, wherein,Representing the annual average precipitation magnitude of a geographical unit, P _ij representing the annual precipitation magnitude under the geographical unit and the time unit, w _j representing the weight of the jth time unit, Q _ij representing the data quality control factor, N representing the total number of time units associated with the geographical unit, Q _ij being a number between 0 and 1, wherein 0 represents that the data is completely unreliable and 1 represents that the data is completely reliable.

In the embodiment of the invention, step 1231 considers a plurality of influencing factors such as geographic position (longitude and latitude), altitude, season and the like, and can more comprehensively reflect the variation condition of the precipitation under different geographic and time conditions; regression coefficients of all influence factors are estimated through regression analysis, a more accurate annual precipitation prediction model can be established, and therefore prediction accuracy and reliability are improved; random error terms in the model can take unobserved influence factors into consideration, so that the model has certain adaptability to precipitation prediction of different geographic areas and time scales. Step 1232 introduces a weight of the time unit, which can reflect the importance difference of the precipitation data of different time periods, for example, the data of more recent years may have more reference value; the abnormal value or other data which does not accord with the quality standard is eliminated through the data quality control factor, so that reliable data can be used when the average precipitation amount of a plurality of years is calculated, the accuracy of the result is improved, and the average precipitation amount of a plurality of years can reflect the precipitation characteristics of a geographic unit on a long time scale.

In the embodiment of the invention, the specific mapping process of the Season (T _j) is as follows:

Using standard four seasons divisions, the spring, summer, autumn, winter are divided into corresponding months, respectively, then the Season (T _j) mapping process can be represented by the following steps:

A specific value of time cell T _j is determined. This is a date, month or other time identifier. In this example, assume that T _j is a month value, ranging from 1 to 12.

The determination of which season it belongs to can be achieved by comparing T _j with the seasonal demarcation boundary value based on the value of T _j.

T _j is mapped onto the corresponding seasonal category. This is accomplished by assigning a season tag (e.g. "spring", "summer", "autumn", "winter") or a season code (e.g. using the number 3 for spring, 6 for summer, 9 for autumn, 12 for winter).

In a preferred embodiment of the present invention, the step 13 may include:

Step 131, by Calculating the average annual precipitation over a long periodWhere w _i is the weight of the i-th year, β is the decay factor, P _i is the annual precipitation of the i-th year, and N is the number of years used to calculate the long term average;

step 132, according to the average annual precipitation over a long period of time, by Calculating a standard deviation s _P of the long-term annual precipitation, wherein gamma _i is the robustness weight of the ith year;

step 133, according to standard deviation of annual precipitation for a long period of time, passing And calculating the precipitation abnormality Rate _i, wherein,Is the absolute deviation of the annual precipitation from the weighted long-term average annual precipitation, delta being the adjustment parameter.

In the embodiment of the present invention, step 131, through the attenuation factor β, enables the recent annual precipitation to have a greater weight when calculating the long-term average, and more accords with the characteristics of more important recent data in actual situations; by adjusting the annual weight w _i, the importance of the data of different years can be flexibly considered, for example, for the years with higher or more critical data quality, a larger weight can be given; the average annual precipitation over a long period can be used as an important reference value for evaluating the long-term climate conditions of a certain area, and is helpful for knowing the climate characteristics of the area. Step 132, by calculating the standard deviation of the annual precipitation, the annual variation degree of the precipitation in a certain area can be quantified, and by setting the robustness weight gamma _i, the influence of the abnormal value on the standard deviation calculation can be reduced, and the robustness of the result can be improved. Step 133, the precipitation abnormality rate comprehensively considers the absolute deviation of annual precipitation and weighted long-term average annual precipitation and the proportion of the deviation to the weighted long-term average annual precipitation, provides more comprehensive precipitation abnormality condition evaluation, and can adjust the sensitivity degree to the deviation according to actual needs when calculating the precipitation abnormality rate by adjusting the parameter delta, thereby increasing the flexibility of the method, and the precipitation abnormality rate can be used as an early warning index of extreme climate events such as drought, flood and the like, thereby being beneficial to timely taking countermeasures to reduce losses.

In a preferred embodiment of the present invention, the step 14 may include:

Step 141, determining a research area and target vegetation;

Step 142, estimating the feasibility and the potential risk of the field measurement according to the actual condition of the research area to obtain an estimation result; determining an in-field measurement scheme according to the evaluation result, and determining a time table according to the in-field measurement scheme;

Step 143, according to the in-situ measurement scheme, performing on-ground vegetation and underground vegetation biomass measurement on each sample party to obtain measurement data;

Step 144, converting the measured data into a global average vegetation net primary productivity over many years as needed.

In the embodiment of the invention, step 141 can focus a specific geographic space by determining a research area, so that the research is more targeted and operable, and the determination of the target vegetation is helpful for selecting representative vegetation types, thereby improving the reliability and universality of the research result. In step 142, by predicting the feasibility and the potential risk of the field measurement, the possible problems can be identified and dealt with in advance, so that the risk and the uncertainty in the field measurement process are reduced, and the determination of the field measurement scheme and the time table is helpful for reasonably distributing manpower, material resources and time resources, so as to ensure the smooth performance of the measurement work. Step 143, by measuring the above-ground and underground vegetation biomass simultaneously, the growth status and ecosystem function of the vegetation can be more comprehensively evaluated, and measuring each sample side is helpful to obtain more accurate data, thereby improving the accuracy and reliability of subsequent analysis. Step 144, converting the measured data into the average vegetation net primary productivity for many years worldwide, which is helpful to unify the data of different areas and different times under the same measurement standard, is convenient for the comparison and analysis in the world, and the converted data can be integrated with the data of other researches in the world, thereby expanding the application range and providing powerful support for the researches in the fields of global climate change, ecological system service and the like.

In another preferred embodiment of the present invention, the step 141 may include:

Selecting a proper geographical area, such as a certain river basin or ecological area, and the like according to the research purpose and the actual situation; natural environment conditions of the research area, such as climate, topography, soil and the like, ensure that the area has representativeness and diversity; determining the boundary range of the research area, wherein natural geographic boundaries can be used or the boundaries can be customized according to research requirements; the location and extent of the investigation region are marked on the map and the area of the investigation region is calculated.

According to the vegetation types and research purposes of the research area, determining vegetation types needing to be measured and evaluated in a key way, such as forests, grasslands, farmlands and the like; the vegetation types in the research area are investigated and classified, and the existing vegetation classification system can be referred to or the classification system can be built automatically according to the research requirements; on the basis of vegetation type classification, target vegetation such as conifer, broadleaf forest, shrubs and the like is further subdivided, so that the representativeness and the representativeness of the target vegetation are ensured; and drawing and marking the distribution areas of the target vegetation, and calculating the area and the duty ratio of each target vegetation.

Overlapping and integrating the spatial distribution information of the research area and the target vegetation to generate a distribution map of the target vegetation in the research area, analyzing and evaluating the distribution map, ensuring that the distribution of the target vegetation in the research area is representative and balanced, and preliminarily determining the spot positions and the quantity of the spot measurements according to the distribution conditions of the research area and the target vegetation.

In another preferred embodiment of the present invention, the step 142 may include:

analyzing natural conditions such as terrain, traffic, climate and the like of a research area, and evaluating difficulty and accessibility of field measurement; and combining the factors to obtain the evaluation result of the feasibility of the field measurement, such as high, medium and low grades.

Identifying natural risks, such as extreme weather, wild animals, geological disasters and the like, which may be encountered during the field measurement process, and evaluating the possibility and influence degree of occurrence of the natural risks; identifying technical risks possibly encountered in the field measurement process, such as instrument equipment faults, improper measurement methods, data quality problems and the like, and evaluating the occurrence possibility and influence degree of the technical risks; and combining the factors to obtain the evaluation result of the on-site measurement potential risk, such as high risk, medium risk and low risk level.

Determining an overall scheme of field measurement according to feasibility and risk assessment results, wherein the overall scheme comprises a measurement area, a measurement object, a measurement method, sample point setting, personnel division and the like; according to the measurement scheme, the workload and the required time of the field measurement are estimated, and the start-stop time and milestone nodes of the measurement are preliminarily drawn up by considering factors such as traffic, weather and the like.

And matching the measurement time table with the vegetation growth period, the climatic characteristics and the like, selecting a proper measurement time window, and avoiding carrying out measurement in a period when vegetation growth is inactive or difficult to identify.

And dynamically adjusting and optimizing the time table, and timely correcting and perfecting the measurement plan according to the actual measurement progress and conditions so as to ensure that the measurement task is completed on time.

In another preferred embodiment of the present invention, the step 143 may include:

step 1431, classifying the overground vegetation according to the types to obtain classification results;

step 1432, measuring the coverage area and the average height of each vegetation, selecting a representative soil sample in each sample side to sample the root system, cleaning and processing the root system sample, separating the root system and measuring the total weight of the root system;

Step 1433, estimating root biomass in unit volume according to the soil volume weight and the sampling volume;

Step 1434, for above-ground vegetation, by Calculating biomass for each vegetationBy passing through Calculating to obtain the total biomass on the ground, wherein,Representing the coverage area of the i-th above-ground vegetation,Represents the density (plants per unit area) of the i-th above-ground vegetation,Representing the height correction coefficient of the i-th above-ground vegetation,Represents the unit biomass (weight after removal of water) of the i-th above-ground vegetation dry matter,Representing the growth correction factor of the i-th above-ground vegetation,The water content correction coefficient of the i-th overground vegetation is represented by G, the collection of vegetation functional groups is represented by w _gi, the weight of the i-th vegetation in the G-th functional group is represented by TF _g, the terrain correction factor of the G-th functional group is represented by K _topo, and the terrain correction coefficient of the whole sample party is represented by K _topo;

Step 1435, estimating root biomass of the whole sample party for the underground vegetation according to the sampling result; by passing through The total biomass above and below ground is calculated, where z represents a layered set of soil depths and Bi _be,z represents the below ground biomass at the z-th soil depth.

In the embodiment of the invention, the biomass of each vegetation can be estimated more accurately by classifying the vegetation on the ground according to the types, measuring the coverage area and the average height of each vegetation and selecting a representative soil sample for root sampling. Meanwhile, direct underground biomass data is also provided for root system sampling and measurement of underground vegetation, and estimation accuracy is further improved. In calculating the biomass, this step introduces a plurality of correction factors, such as height correction factors, growth correction factors, and moisture content correction factors, to account for the effects of different heights, growth phases, and moisture content on the biomass. In addition, the terrain correction factors and the terrain correction factors of the whole sample party are also considered, so that the biomass estimated value is adjusted to be more suitable for the actual situation. The biomass of the above-ground vegetation is estimated, the root biomass of the underground vegetation is also estimated through sampling and measurement, and the total biomass is obtained by adding the biomass and the root biomass, so that the biomass distribution and the total biomass of an ecological system can be more comprehensively known. The whole steps follow scientific methods and systematic flows, from vegetation classification, measurement to calculation and analysis, each with a definite purpose and operating specifications. This helps to ensure the reliability and repeatability of the biomass estimation. The biomass estimation result obtained by the step can provide basic data for researches on energy flow, material circulation, carbon storage and the like of the biological system, and is helpful for deep understanding of the structure and function of the ecological system.

In another preferred embodiment of the present invention, the step 144 may include:

Step 1441, selecting a representative plot in the investigation region, and performing a simultaneous measurement of biomass and net primary productivity; measuring above-ground and below-ground biomass and measuring net primary productivity;

step 1442, performing quality control on the measured data, and removing abnormal values and false measured values; calculating the aboveground and underground biomass and the net primary productivity of each plot; biomass and net primary productivity data are classified according to vegetation type, growth stage, and other factors.

In step 1443, a linear regression model is built with biomass as an independent variable and net primary productivity as a dependent variable, and regression models of the above-ground and underground parts are built respectively, such as:

NPP_s＝a×AGB+b；

NPP_r＝c×BGB+d；

wherein NPP _s and NPP _r are above-ground and below-ground net primary productivity, AGB and BGB are above-ground and below-ground biomass, respectively, and a, b, c, d is a regression coefficient.

Step 1444, for the aerial parts, converting the factor into a slope coefficient a of a regression model, representing the net primary productivity corresponding to biomass on the unit ground; for the subsurface portion, the conversion factor is the slope coefficient c of the regression model, representing the net primary productivity per unit subsurface biomass.

Step 1445, converting the above-ground and underground biomass of each sample into a corresponding net primary productivity, respectively, using the established conversion factors; calculating the average value of the net primary productivity of all the sample parties in the research area as the representative value of the net primary productivity of the area; the average value is multiplied by the area of the investigation region to give the total net primary productivity of the investigation region.

Step 1446, obtaining a global vegetation distribution map corresponding to the selected vegetation classification from the global land cover database; determining other areas with the same vegetation category as the research area according to the global vegetation distribution diagram; assuming similar net primary productivity levels for other regions of the same vegetation class as the study region; assigning a net primary productivity representative of the area of investigation to all areas of the same vegetation class worldwide; multiplying each vegetation category by its global area to obtain the global total net primary productivity of that category;

Step 1447, adding the global total net primary productivity of all vegetation categories to obtain global total net primary productivity; collecting global vegetation distribution data within a certain time range (such as 2000-2010), repeating the operation, and calculating global total net primary productivity of each year; averaging the total net primary productivity over the years to obtain an average net primary productivity over the years over the world; dividing the annual average by the global land total area gives the global average net primary productivity density over many years.

In the embodiment of the invention, the accuracy and the representativeness of the data are ensured by synchronously measuring the biomass and the net primary productivity through selecting the representative sample. And the quality control is carried out on the measured data, so that the reliability of the data is further improved. Establishing a relationship between biomass and net primary productivity of the above-ground and below-ground parts, respectively, using a linear regression model; by introducing a conversion factor (slope coefficient of regression model), the biomass data can be conveniently converted into net primary productivity, which is flexible and widely applicable. The net primary productivity representative value of the research area is promoted to all areas of the same vegetation class in the global scope, so that estimation from local to global scale is realized, and an important basis is provided for ecological system estimation in global scale. By collecting global vegetation distribution data over a time frame and repeating the above operations, the global total net primary productivity per year can be calculated, and further the time-varying trend of net primary productivity can be analyzed.

In a preferred embodiment of the present invention, the step 15 may include:

Step 151, acquiring a data set of the net primary productivity of the overground vegetation and the net primary productivity of the underground vegetation and longitude and latitude coordinates of the data set;

Step 152, obtaining an observation data set of the net primary productivity of the above-ground vegetation and the net primary productivity of the below-ground vegetation;

step 153, formatting the observation data set to obtain formatted data;

step 154, extracting and sorting longitude and latitude coordinates of the observation point in each formatted data;

Step 155, acquiring a global data set, performing space matching with the global data set according to longitude and latitude coordinates of the observation points, and extracting years of average vegetation net primary productivity, years of average precipitation and years of precipitation anomaly values corresponding to each observation point.

In the embodiment of the invention, in step 151, the acquisition of the above-ground and underground vegetation net primary productivity dataset provides a necessary data basis for subsequent analysis, and the acquisition of longitude and latitude coordinates ensures the geospatial positioning accuracy of the data, thereby facilitating subsequent spatial matching and analysis. In step 152, the observation data set is used to verify model output or remote sensing data, provide important references for ground real conditions, and combine the model output, remote sensing data and ground observation data, so as to improve the comprehensiveness and accuracy of analysis. In step 153, the formatting process ensures the consistency and comparability of the data, is convenient for subsequent data integration and analysis, and the unified data format can reduce errors and redundancy in the data processing process and improve the data processing efficiency. And 154, extracting and sorting the longitude and latitude coordinates ensures accurate geographic positioning of each observation point, provides an accurate basis for subsequent space matching, and is beneficial to improving data quality and reducing data deviation caused by geographic positioning errors. Step 155, through space matching, the data of the observation points and the global background data can be correlated, more comprehensive information is provided for analysis, and a plurality of indexes such as the average vegetation net primary productivity of many years, the average precipitation of many years, the annual precipitation anomaly value and the like are extracted, so that the comprehensive analysis of the relationship between vegetation growth and climate factors from a plurality of dimensions is facilitated.

In step 155, the longitude and latitude coordinates of the observation point are corresponding to grid cells in the global data set, so as to extract the years average vegetation net primary productivity, years average precipitation and years precipitation anomaly value of the position of the observation point, and the following is a specific method for performing space matching:

In step 1551, the spatial resolution of the global data set is determined, the global data set being provided in a grid format, such as 0.5 ° ×0.5 °, 1km×1km, etc.

In step 1552, the grid size and boundary coordinates of the dataset are determined, e.g., the dataset has a longitude range of-180 ° to 180 °, a latitude range of-90 ° to 90 °, and the grid size is 0.5 ° by 0.5 °.

Step 1553, converting the coordinates of the observation point into a coordinate system of a global dataset, wherein the longitude and latitude coordinates of the observation point are in a geographic coordinate system (e.g. WGS 84), and the global dataset may be in a different projection coordinate system (e.g. equal area projection); the observation point coordinates are converted from a geographic coordinate system to the same projection coordinate system as the global data set using GIS software or programming tools.

Step 1554, determining the grid unit where the observation point is located, and calculating the row number and the column number of the observation point in the global data set grid according to the projection coordinates of the observation point, wherein,

Line number= (Y _max-Y_obs)/size_y;

column number= (X _obs-X_min)/size_x;

Wherein Y _max and X _min are the north-most and west-most coordinates of the dataset, Y _obs and X _obs are the projected coordinates of the observation point, and size _y and size _x are the latitudinal and longitudinal sizes of the grid, respectively.

Step 1555, obtaining the row number and the column number of the grid unit where the observation point is located after rounding, and extracting corresponding years of average vegetation net primary productivity, years of average precipitation and years of precipitation anomaly values from the global data set according to the row number and the column number of the grid unit where the observation point is located; if the observation point is located on the grid boundary, a nearby principle can be selected to extract the data value of the grid cell nearest to the observation point.

Step 1556, processing the data missing or abnormal value, if the data value of the grid unit where the observation point is located is missing or abnormal, estimating by interpolation or adjacent unit averaging and other methods;

Step 1557, combining the longitude and latitude coordinates of each observation point with the extracted years of average vegetation net primary productivity, years of average precipitation and annual precipitation anomaly values into a new data sheet, and storing the matching result.

In a preferred embodiment of the present invention, after the step 17, the method may further include:

Step 18, comparing the predicted global data set of the net primary productivity of the overground vegetation and the net primary productivity of the underground vegetation with the data set of the net primary productivity of the overground vegetation and the net primary productivity of the underground vegetation which are actually measured on the ground respectively to obtain a comparison result;

And step 19, according to the comparison result, evaluating the prediction performance of the machine learning model to obtain an evaluation result.

In the embodiment of the invention, step 18 can directly verify the prediction accuracy of the machine learning model by comparing the prediction data with the ground measured data, so as to ensure the reliability of the prediction result. And step 19, the prediction performance of the machine learning model can be quantitatively evaluated through comparison of the results, the evaluation results can provide guidance for further optimization of the model, and more powerful support can be provided for decisions such as biological system management, climate change adaptation and the like.

As shown in fig. 2, an embodiment of the present invention further provides a remote sensing prediction apparatus 20 for net primary productivity of vegetation, comprising:

An acquisition module 21 for acquiring a global average year-to-year vegetation net primary productivity product generated based on the parametric model; acquiring environment covariates according to the global annual average vegetation net primary productivity product, wherein the environment covariates comprise global annual precipitation data and annual average precipitation data; calculating annual precipitation abnormality rate according to global annual precipitation data; acquiring a ground vegetation net primary productivity data set of ground actual measurement and an underground vegetation net primary productivity data set;

The processing module 22 is configured to obtain longitude and latitude coordinates corresponding to the above-ground vegetation net primary productivity and the below-ground vegetation net primary productivity data set, and extract a global years average vegetation net primary productivity, a years average precipitation amount and a year precipitation abnormal rate at the observation point in the global data set; building and training a machine learning model according to the average vegetation net primary productivity for many years, the average precipitation amount for many years and the annual precipitation anomaly rate; according to a machine learning model, global annual average vegetation net primary productivity, annual average precipitation and annual precipitation anomaly rate data sets are input to generate predicted global data sets of overground vegetation net primary productivity and underground vegetation net primary productivity.

Optionally, obtaining a global years-average vegetation net primary productivity product generated based on the parametric model comprises:

Optionally, obtaining an environmental covariate from the global year-average vegetation net primary productivity product, the environmental covariate including global year precipitation data and year-average precipitation data, comprising:

Optionally, calculating annual precipitation anomaly rate according to global annual precipitation data includes:

Optionally, acquiring a ground measured net primary productivity of the above-ground vegetation and a ground vegetation data set comprising:

Determining a research area and target vegetation;

Optionally, acquiring longitude and latitude coordinates corresponding to the above-ground vegetation net primary productivity and the underground vegetation net primary productivity data set respectively, and extracting a global years average vegetation net primary productivity, a years average precipitation amount and an annual precipitation anomaly rate at an observation point in the global data set, including:

formatting the observation data set to obtain formatted data;

Optionally, after inputting the global annual average vegetation net primary productivity, annual average precipitation and annual precipitation anomaly rate data sets according to the machine learning model to generate the predicted global data sets of overground vegetation net primary productivity and subsurface vegetation net primary productivity, further comprising:

It should be noted that the apparatus is an apparatus corresponding to the above method, and all implementation manners in the above method embodiment are applicable to this embodiment, so that the same technical effects can be achieved.

Embodiments of the present invention also provide a computing device comprising: a processor, a memory storing a computer program which, when executed by the processor, performs the method as described above. All the implementation manners in the method embodiment are applicable to the embodiment, and the same technical effect can be achieved.

Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform a method as described above. All the implementation manners in the method embodiment are applicable to the embodiment, and the same technical effect can be achieved.

Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.

In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.

The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk, etc.

Furthermore, it should be noted that in the apparatus and method of the present invention, it is apparent that the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent aspects of the present invention. Also, the steps of performing the series of processes described above may naturally be performed in chronological order in the order of description, but are not necessarily performed in chronological order, and some steps may be performed in parallel or independently of each other. It will be appreciated by those of ordinary skill in the art that all or any of the steps or components of the methods and apparatus of the present invention may be implemented in hardware, firmware, software, or a combination thereof in any computing device (including processors, storage media, etc.) or network of computing devices, as would be apparent to one of ordinary skill in the art after reading this description of the invention.

The object of the invention can thus also be achieved by running a program or a set of programs on any computing device. The computing device may be a well-known general purpose device. The object of the invention can thus also be achieved by merely providing a program product containing program code for implementing said method or apparatus. That is, such a program product also constitutes the present invention, and a storage medium storing such a program product also constitutes the present invention. It is apparent that the storage medium may be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and method of the present invention, it is apparent that the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent aspects of the present invention. The steps of executing the series of processes may naturally be executed in chronological order in the order described, but are not necessarily executed in chronological order. Some steps may be performed in parallel or independently of each other.

While the foregoing is directed to the preferred embodiments of the present invention, it will be appreciated by those skilled in the art that various modifications and adaptations can be made without departing from the principles of the present invention, and such modifications and adaptations are intended to be comprehended within the scope of the present invention.

Claims

1. A method for remote sensing prediction of net primary productivity of vegetation, the method comprising:

2. The method of claim 1, wherein obtaining a global multi-year average vegetation net primary productivity product generated based on a parametric model comprises:

3. The method of remote sensing prediction of net primary productivity of vegetation according to claim 2, wherein obtaining environmental covariates from the global multi-year average net primary productivity product of vegetation, the environmental covariates comprising global annual precipitation data and multi-year average precipitation data comprises:

4. A method of remotely sensing a net primary productivity of vegetation as claimed in claim 3 wherein calculating annual precipitation anomaly rate from global annual precipitation data comprises:

5. The method of claim 4, wherein obtaining the ground measured ground net primary productivity of vegetation and the ground net primary productivity dataset of vegetation comprises:

Determining a research area and target vegetation;

6. The method of claim 5, wherein obtaining latitude and longitude coordinates corresponding to the above-ground vegetation net primary productivity and the below-ground vegetation net primary productivity data sets, respectively, and extracting the global years-average vegetation net primary productivity, the years-average precipitation amount, and the annual precipitation anomaly rate at the observation point in the global data set comprises:

formatting the observation data set to obtain formatted data;

7. The method of claim 6, wherein after inputting global multi-year average net primary productivity of vegetation, multi-year average precipitation and annual precipitation anomaly rate data sets from a machine learning model to generate predicted global data sets of net primary productivity of overground vegetation and net primary productivity of subsurface vegetation, further comprising:

8. A remote sensing predictive device for net primary productivity of vegetation, comprising:

9. A computing device, comprising:

one or more processors;

Storage means for storing one or more programs which when executed by the one or more processors cause the one or more processors to implement the method of any of claims 1 to 7.

10. A computer readable storage medium, characterized in that the computer readable storage medium has stored therein a program which, when executed by a processor, implements the method according to any of claims 1 to 7.