Fine Expression Method for Temporal Variation of Saline-alkali Land Vegetation Status at Regional Scale

Abstract

A refined expression method for the temporal variation of saline-alkali land vegetation status at the regional scale, which involves an expression method for the temporal variation of saline-alkali wasteland vegetation status at the regional scale. The present invention aims to solve the problem that traditional methods cannot simultaneously depict the volatility caused by random factors in the temporal variation trend of regional saline-alkali wasteland vegetation. The method includes: constructing a mathematical model for the change trend of saline-alkali land vegetation status at the regional scale; constructing a fluctuation model for the change of saline-alkali land vegetation status at the regional scale; constructing and validating the accuracy of the composite refined model. The present invention proposes a composite model that can simultaneously and finely describe the volatility and trend of the temporal variation of regional saline-alkali land vegetation status, making the characterization of the temporal variation characteristics of regional salinized grassland more accurate and scientific. The present invention belongs to the field of expressing the temporal variation of saline-alkali land vegetation status.

Description

Technical Field

[0001] The present invention relates to a method for expressing the temporal variation of the vegetation status of saline-alkali wasteland on a regional scale. Background Art

[0002] Most of the characterizations of the changes in the vegetation status of saline-alkali land on a regional scale use vegetation indices to construct mathematical models to describe its overall change trend. In fact, the change in the vegetation status of saline-alkali land is a comprehensive change under the superposition and interference of different factors, and at the same time shows volatility and trend. Existing models are difficult to fit and describe the fluctuation phenomenon in its change process. Summary of the Invention

[0003] The purpose of the present invention is to solve the technical problem that the traditional time series analysis method cannot simultaneously depict the volatility caused by random factors in the temporal change trend of the vegetation of regional saline-alkali wasteland, and provides a refined expression method for the temporal change of the vegetation status of saline-alkali land on a regional scale.

[0004] The refined expression method for the temporal change of the vegetation status of saline-alkali land on a regional scale is carried out according to the following steps:

[0005] 1. Construct a mathematical model for the change trend of the vegetation status of saline-alkali land on a regional scale

[0006] Obtain saline-alkali land patches with unchanged category attributes of cross-section saline-alkali land at the same time in different years, and at the same time obtain the NDVI index and meteorological data representing the vegetation status of saline-alkali land in the same month of each year with a time frequency of 1 year within the range of the patch, and the spatial resolution is 1 km;

[0007] Taking the NDVI representing the vegetation status of saline-alkali land as the independent variable factor, establish a trend change model for the vegetation of saline-alkali land in the study area respectively, and establish the optimal trend mathematical model with variance and mean square error;

[0008] 2. Construct a fluctuation model for the change of the vegetation status of saline-alkali land on a regional scale

[0009] Calculate the simulated value of NDVI representing the vegetation status of saline-alkali land within the corresponding time with the optimal trend mathematical model established in step 1, and calculate the difference between the simulated value and the actual value, which represents the fluctuation value of the vegetation status on a regional scale;

[0010] Taking the difference of the vegetation status of each pixel as the independent variable, establish a regression relationship with factors such as precipitation and evaporation, so as to establish a fluctuation model for the change of the vegetation status of saline-alkali land;

[0011] 3. Construction and accuracy verification of a composite refined model

[0012] Superimpose and couple the trend mathematical model constructed in Step 1 and the fluctuation model constructed in Step 2 to construct a time-refined model at the regional scale, and conduct accuracy verification to obtain the accuracy characterization quantity of the time-varying model of saline-alkali land vegetation in the study area, and determine the advantages, disadvantages and practicability of the established model.

[0013] When obtaining the saline-alkali land map patches with unchanged category attributes of the cross-section in Step 1, the spatial resolution is 30m.

[0014] In Step 1, the NDVI index and meteorological data characterizing the vegetation status of saline-alkali land are obtained for each year from May to September with a time frequency of 1 year within the range of the map patch.

[0015] The trend mathematical model in Step 1 includes a linear function trend model, a power function trend model, a logarithmic function trend model, and a quadratic polynomial function trend model.

[0016] In the present invention, when constructing a time-varying trend model for the vegetation status change of regional saline-alkali land, the difference between the trend simulation value and the true value is proposed to characterize the fluctuation, and further a fluctuation model of the difference fluctuation quantity and its random interference factors of precipitation and evaporation is constructed. The two are coupled to construct a refined expression model of the change of saline-alkali land vegetation, realizing the simultaneous simulation of the trend process and the fluctuation process of the change of saline-alkali land vegetation, so as to achieve the effect of improving the refined expression of the change process of saline-alkali land vegetation.

[0017] The present invention proposes a composite model that can simultaneously and finely describe the volatility and trend of the temporal change of the vegetation status of regional saline-alkali land, and further characterize the temporal change characteristics of the salinization degree, making the characterization of the temporal change characteristics of regional salinized grassland more accurate and scientific.

[0018] By constructing the vegetation status trend and fluctuation models of saline-alkali land at the regional scale, the present invention improves the time change accuracy of the vegetation status within the saline-alkali land map patches from the time scale, reflecting the scientific nature of using temporally discontinuous data to express the change degree of saline-alkali land vegetation. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is the quadratic polynomial function trend model in Step 1 of the experiment;

[0020] Figure 2 It is the logarithmic function trend model in Step 1 of the experiment;

[0021] Figure 3 It is the power function trend model in Step 1 of the experiment;

[0022] Figure 4 It is the linear function trend model in Step 1 of the experiment;

[0023] Figure 5It is the trend chart of the vegetation status of saline-alkali land in the study area in Experiment 1;

[0024] Figure 6 It is the fluctuation chart of the vegetation status of saline-alkali land in the study area in Experiment 1;

[0025] Figure 7 It is the relationship chart between NDVI and precipitation of saline-alkali land in the study area in Experiment 1;

[0026] Figure 8 It is the relationship chart between NDVI and evaporation of saline-alkali land in the study area in Experiment 1;

[0027] Figure 9 It is the comparison chart of the actual and simulated numerical trends of NDVI of saline-alkali land in the study area in Experiment 1. Specific implementation manners

[0028] The technical solution of the present invention is not limited to the following specific implementation manners, and also includes any combination between the specific implementation manners.

[0029] Specific implementation manner 1: The method for refined expression of the temporal variation of the vegetation status of saline-alkali land at the regional scale in this implementation manner is carried out according to the following steps:

[0030] 1. Construct a mathematical model for the trend of the vegetation status of saline-alkali land at the regional scale

[0031] Obtain saline-alkali land patches with unchanged category attributes of the cross-section saline-alkali land at the same time in different years, and at the same time obtain the NDVI index and meteorological data representing the vegetation status of saline-alkali land in the same month of each year with a time frequency of 1 year within the range of the patch, and the spatial resolution is 1 km;

[0032] Taking the NDVI representing the vegetation status of saline-alkali land as the independent variable factor, establish a trend change model for the vegetation of saline-alkali land in the study area respectively, and establish the optimal trend mathematical model with variance and mean square error;

[0033] 2. Construct a fluctuation model for the change of the vegetation status of saline-alkali land at the regional scale

[0034] Calculate the simulated value of NDVI representing the vegetation status of saline-alkali land within the corresponding time with the optimal trend mathematical model established in step 1, and calculate the difference between the simulated value and the actual value, which represents the fluctuation value of the vegetation status at the regional scale;

[0035] Taking the difference of the vegetation status of each pixel as the independent variable, establish a regression relationship with factors such as precipitation and evaporation, so as to establish a fluctuation model for the change of the vegetation status of saline-alkali land;

[0036] 3. Construction and accuracy verification of the composite refined model

[0037] Superimpose and couple the trend mathematical model constructed in Step 1 and the fluctuation model constructed in Step 2 to construct a time-refined model at the regional scale, and conduct accuracy verification to obtain the accuracy characterization quantity of the time-varying model of saline-alkali land vegetation in the study area, and determine the advantages, disadvantages and practicability of the established model.

[0038] Specific Embodiment 2: The difference between this embodiment and Specific Embodiment 1 is that when obtaining the saline-alkali land map patches with unchanged category attributes of the cross-section in Step 1, the spatial resolution is 30m. Others are the same as Specific Embodiment 1.

[0039] Specific Embodiment 3: The difference between this embodiment and Specific Embodiment 1 or 2 is that in Step 1, the NDVI index and meteorological data characterizing the saline-alkali land vegetation status in May-September of each year with a time frequency of 1 year within the range of the map patch are obtained simultaneously. Others are the same as Specific Embodiment 1 or 2.

[0040] Specific Embodiment 4: The difference between this embodiment and any one of Specific Embodiments 1 to 3 is that the trend mathematical model in Step 1 includes a linear function trend model, a power function trend model, a logarithmic function trend model, and a quadratic polynomial function trend model. Others are the same as any one of Specific Embodiments 1 to 3.

[0041] The following experiment is used to verify the effect of the present invention:

[0042] Experiment 1:

[0043] The method for refined expression of the temporal variation of the saline-alkali land vegetation status at the regional scale is carried out according to the following steps:

[0044] I. Construct a mathematical model of the change trend of the saline-alkali land vegetation status at the regional scale

[0045] First, extract the saline-alkali land map patches in the land use types at 4 time cross-sections (2000, 2005, 2010, 2015, and 2020) during the period from 2000 to 2020. The data accuracy (spatial resolution) is the pixel size of 30m, and select the pixels with unchanged saline-alkali land map patch types during the period to determine the spatial distribution range of the saline-alkali land.

[0046] Second, extract the NDVI index data characterizing the growth status of the saline-alkali land vegetation in May-September of each year with a time frequency of one year. The data accuracy (spatial resolution) is 1km, which is used to characterize the degree of salinization.

[0047] Third, taking the NDVI characterizing the saline-alkali land vegetation status as the independent variable factor, establish a mathematical model of the change trend of the saline-alkali land vegetation status at the regional scale from 2000 to 2020, including a linear function trend model, a power function trend model, a logarithmic function trend model, and a quadratic polynomial function trend model, as shown in Figures 1-4, each model parameter is shown in Table 1.

[0048] Compare the constructed trend change models of saline-alkali land vegetation status in each time series, and compare the accuracy fitting of the simulated NDVI values output by each trend change model with the actual values, as shown in Table 1. Among them, the linear function model R 2 is 0.67 and the RMSE is 0.021; the power function model R 2 is 0.48 and the RMSE is 0.025; the logarithmic function model R 2 is 0.45 and the RMSE is 0.026; the quadratic polynomial function model R 2 is 0.68 and the RMSE is 0.021. It is thus determined that the quadratic polynomial function model has the best fitting effect on the change trend of the saline-alkali land vegetation status in the study area, and the quadratic polynomial function model is established as the optimal trend mathematical model.

[0049] Table 1 Mathematical models of the change trend of the vegetation status of saline-alkali land pixels from 2000 to 2020

[0050]

[0051] II. Construct a fluctuation model for the change of saline-alkali land vegetation status at the regional scale

[0052] According to the quadratic polynomial function trend model, the fitting trend change of the saline-alkali land vegetation status in the study area Figure 1 , by comparing the simulated values and the actual values of the NDVI time series mathematical trends of the vegetation status of each saline-alkali land in the test area, it is found that although the quadratic polynomial function model has a high fitting accuracy for the change trend of the NDVI of the saline-alkali land vegetation status in the study area, the simulated NDVI values output by the model still have a large gap from the actual values, and the fluctuation phenomenon of the change trend of the saline-alkali land vegetation status in the study area cannot be reflected, as shown in Figure 5 .

[0053] Through further data analysis of the saline-alkali land vegetation status, precipitation, and evaporation, it is found that it has a significant positive correlation with precipitation and a significant negative correlation with evaporation. The large evaporation and small precipitation make the climate in the plain arid, the soil humidity low, and the soil prone to salt accumulation, which in turn affects the change trend of the saline-alkali land vegetation status. The relationship between the NDVI of the saline-alkali land in the study area and precipitation and evaporation is as shown in Figure 6 .

[0054] Based on this, a binary linear regression model related to precipitation and evaporation is constructed for the difference between the simulated NDVI value and the actual value output by the quadratic polynomial function trend model. The specific formula of the model is as follows:

[0055] Δ = a·P + b·ET + c…

[0056] In the formula, Δ is the difference between the NDVI simulated value and the actual value output by the quadratic polynomial function trend model, P is the average precipitation from May to September in each year, ET is the average evaporation from May to September in each year, a is the precipitation coefficient, b is the evaporation coefficient, and c is a constant.

[0057] Taking the difference between the NDVI simulated value and the actual value of each year output by the quadratic polynomial function trend model as the independent variable, and the average precipitation and average evaporation from May to September in each year as the dependent variable, the preprocessed precipitation and evaporation data are brought into the constructed fluctuation model, and the coefficient a is obtained as 0.00047, the coefficient b is -0.00036, and the constant c is 0.0079. The specific formula of the fluctuation model is:

[0058] Y = 0.00047·P - 0.00036·ET + 0.267

[0059] III. Construction and accuracy verification of the composite refinement model

[0060] Combining the trend mathematical model and the fluctuation model to obtain a refined model of the temporal variation of the vegetation status of saline-alkali land in the Songnen Plain. The formula is:

[0061] Y = 0.0001x 2 + 0.0017x + 0.00047·P - 0.00036·ET + 0.267

[0062] Perform accuracy verification on the refined model of the temporal variation of the vegetation status of saline-alkali land that has been constructed. R 2 is 0.72, and RMSE is 0.021. The simulation results of the temporal refinement model for the change trend of the vegetation status of saline-alkali land in the study area are as Figure 9 shown. From the accuracy verification results, the error is significantly reduced, and the correlation coefficient is significantly improved. It can be seen from the trend comparison chart of the simulated values that the fitting performance is improved, indicating that the accuracy of the composite model constructed by the present invention is significantly improved.

Claims

1. A refined expression method for the temporal variation of saline-alkali land vegetation status at the regional scale, characterized in that The refined expression method for the temporal variation of saline-alkali land vegetation status at the regional scale is carried out according to the following steps:

1. Construct a mathematical model for the change trend of saline-alkali land vegetation status at the regional scale Obtain saline-alkali land patches with unchanged category attributes of cross-section saline-alkali land at the same time in different years. At the same time, obtain the NDVI index and meteorological data representing the vegetation status of saline-alkali land in the same month of each year with a time frequency of 1 year within the range of the patch, and the spatial resolution is 1 km; Taking the NDVI representing the vegetation status of saline-alkali land as the independent variable factor, establish a trend change model for the vegetation of saline-alkali land in the study area respectively, and determine the optimal trend mathematical model with variance and mean square error; 2. Construct a fluctuation model for the change of saline-alkali land vegetation status at the regional scale Calculate the simulated value of NDVI representing the vegetation status of saline-alkali land within the corresponding time by using the optimal trend mathematical model established in step 1, and calculate the difference between the simulated value and the actual value, which represents the fluctuation value of the vegetation status at the regional scale; Taking the difference of the vegetation status of each pixel as the independent variable, establish a regression relationship with precipitation and evaporation factors, so as to establish a fluctuation model for the change of saline-alkali land vegetation status; 3. Construction and accuracy verification of the composite refined model Superpose and couple the trend mathematical model constructed in step 1 and the fluctuation model constructed in step 2 to construct a time-refined model at the regional scale, and conduct accuracy verification to obtain the accuracy characterization quantity of the time change model of saline-alkali land vegetation in the study area, and determine the advantages, disadvantages and practicability of the established model.

2. The refined expression method for the temporal variation of saline-alkali land vegetation conditions at the regional scale according to claim 1, characterized in that When obtaining the saline-alkali land patches with unchanged category attributes of the cross-section saline-alkali land in step 1, the spatial resolution is 30 m.

3. The refined expression method for the temporal variation of saline-alkali land vegetation conditions at the regional scale according to claim 1, characterized in that In step 1, the NDVI index and meteorological data representing the vegetation status of saline-alkali land in May-September of each year with a time frequency of 1 year are obtained within the range of the patch at the same time.

4. The refined expression method for the temporal variation of saline-alkali land vegetation status at the regional scale according to claim 1, characterized in that The trend mathematical model in step 1 includes a linear function trend model, a power function trend model, a logarithmic function trend model, and a quadratic polynomial function trend model.

Images