Evaluation method for effect of soil salinization on farmland productivity under multi-dimensional dynamic model

Abstract

The application discloses a kind of multi-dimensional dynamic model under the evaluation method of soil salinization to cultivated land productivity, it is related to the field technical field of agricultural resources and environment, and the multi-source data set of fusing remote sensing, ground monitoring, meteorological and farmer management data is first constructed in the scheme, and the minimum data set of soil physical, chemical, biological properties is screened after fusion preprocessing;Soil quality index (SQI) and cultivated land productivity index (CPI) are calculated again;Subsequently, the multi-dimensional dynamic evaluation model containing key salinization index screening, effect quantification, mechanism model and crop sub-model is constructed;Finally, the productivity influence dynamic update is realized in combination with real-time data of Internet of Things, and the optimal improvement technology mode is simulated and screened by structural equation model simulation management scene. The method is comprehensive in data coverage, accurate in evaluation, dynamic in model and targeted, can reveal the influence mechanism of salinization on cultivated land productivity, provide practical decision support, and help efficient and sustainable use of saline-alkali soil.

Description

Technical Field

[0001] This invention relates to the field of agricultural resources and environmental technology, specifically to a method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model. Background Technology

[0002] Soil salinization is a key factor restricting arable land productivity, and assessing its impact on arable land productivity is a core aspect of agricultural sustainable development research. Current technologies mostly monitor soil salinity indicators such as electrical conductivity and pH value, combining this with crop yield data, and employing single models or simple linear analyses to explore the correlation between the two. Some studies, using traditional statistical models and single-index evaluation methods, have preliminarily revealed the inhibitory effect of salinization on crop growth and yield, providing a basic reference for saline-alkali land management. For example, monitoring changes in soil salt ion content at fixed points can help determine the fluctuation trend of arable land productivity.

[0003] However, existing technologies have significant shortcomings in data integration and coverage. Most assessment methods rely solely on single-source data such as ground-based fixed-point monitoring, failing to effectively integrate remote sensing macro-data, meteorological dynamic data, and farmer management data. This results in an incomplete characterization of the spatiotemporal distribution of salinization, making it difficult to reflect the regional heterogeneity of salinization and its comprehensive impact on productivity. Furthermore, soil quality assessments are limited to physicochemical properties, neglecting biological properties such as microbial abundance and enzyme activity, and lack a scientific minimum dataset selection system, leading to redundant indicators or missing key information, thus affecting the accuracy and efficiency of the analysis.

[0004] These shortcomings significantly reduce the practicality and scientific rigor of the assessment methods. On the one hand, due to the lack of multi-dimensional data fusion and dynamic model support, it is difficult to accurately quantify the complex relationship between salinization degree, soil quality, and arable land productivity. It also fails to reveal the mechanism by which salinization affects productivity through both direct toxicity and indirect soil quality impacts, resulting in insufficient reliability of the assessment results. On the other hand, existing methods are mostly static analyses, failing to incorporate crop growth dynamics and seasonal changes to achieve real-time updates on productivity impacts. They also lack differentiated systems for different salt-tolerant crops and lack scenario simulation and optimization functions for field management measures, making it difficult to provide precise and feasible improvement and productivity enhancement schemes, thus hindering the efficient and sustainable utilization of saline-alkali land resources. To overcome the technical bottlenecks of existing assessment methods and accurately quantify the impact mechanism of soil salinization on arable land productivity, this paper proposes a soil salinization-arable land productivity evaluation method based on a multi-dimensional dynamic model. This method aims to systematically address the aforementioned problems and provide scientific support for the efficient utilization of saline-alkali land and the enhancement of arable land productivity. Summary of the Invention

[0005] 1. The technical problem to be solved by the present invention

[0006] The purpose of this invention is to propose a method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model, in order to solve the following problems existing in the prior art: (1) Insufficient data integration and evaluation dimensions: Existing technologies mostly rely on single-source data and do not effectively integrate multi-dimensional data such as remote sensing, meteorology, and farmer management, resulting in an incomplete depiction of the spatiotemporal distribution of salinization; moreover, soil quality assessment is limited to physical and chemical properties, neglecting biological properties, lacking a scientific minimum dataset screening system, and having problems such as redundant indicators or missing key information, making it difficult to accurately reflect the comprehensive impact of salinization on arable land productivity.

[0007] (2) The evaluation model lacks staticity and practicality: Most existing methods are static analyses, and do not combine crop growth period and seasonal changes to realize dynamic updates of production capacity impact, nor do they establish differentiated evaluation systems for different salt-tolerant crops; at the same time, they lack the scenario simulation and optimization functions for field management measures, and cannot reveal the complex mechanism by which salinization affects production capacity through two pathways: direct toxicity and indirect impact on soil quality, making it difficult to provide accurate and feasible improvement and production capacity enhancement solutions.

[0008] 2. Technical Solution To achieve the above objectives, this invention provides a technical solution for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model: Specific steps include: S1. Construct a multi-source dataset: Collect remote sensing data, ground monitoring data, meteorological data and farmer management data of the study area. The ground monitoring data includes soil physicochemical property data and crop growth status data to establish a long-term spatial distribution dataset. S2. Data Fusion and Preprocessing: Multi-source data were integrated using a multivariate stepwise linear regression satellite-ground hyperspectral fusion method. Principal component analysis and discriminant analysis were used to construct a minimal dataset of soil physical, chemical, and biological properties. S3. Index Calculation: Based on the minimum dataset of soil physical, chemical and biological properties obtained by principal component analysis and discriminant analysis, the soil quality index (SQI) is calculated by combining models such as fuzzy logic, random forest and nonlinear scoring model. At the same time, crop yield is converted into arable land productivity index (CPI) to characterize the potential productivity of arable land through normalization model. S4. Construct a multi-dimensional dynamic evaluation model: Key salinization indicators such as soil pH and electrical conductivity (EC) are screened out using the random forest model. The partial least squares path model is used to quantify the degree of salinization, the various effects between soil quality index (SQI) and arable land productivity index (CPI). Empirical models, interaction term multivariate nonlinear regression models and other mechanistic models are constructed to reveal synergistic effects. Sub-crop evaluation sub-models are also established for different salt-tolerant crops. S5. Dynamic Assessment and Decision Support: Combining real-time monitoring data from the Internet of Things with historical data, the system dynamically updates the impact of different crop growth stages and seasons on production capacity. It also uses structural equation modeling (SEM) to simulate production capacity changes under different management scenarios and selects the optimal improvement technology model.

[0009] Preferably, the remote sensing data mentioned in S1 is multispectral satellite remote sensing imagery and UAV imagery data, used to obtain macroscopic salinity distribution information; meteorological data includes precipitation, evaporation, temperature, and light data; and farmer management data includes irrigation methods, fertilizer application rates, and crop variety selection records.

[0010] Preferably, the soil physicochemical properties data in S2 include soil bulk density, water content, average weight diameter of aggregates (MWD), organic matter, total nitrogen, available phosphorus (AP), available potassium (AK), and salt ion content; and the soil biological properties data include microbial biomass and enzyme activity data.

[0011] As a preferred option, the specific expression for the normalized model in S3 is:

[0012] In the formula, X X represents the normalized value; i X represents the measured yield at the i-th sampling point; max and X min These represent the maximum and minimum values ​​of the measured output, respectively.

[0013] Preferably, the soil principal component analysis described in S3 uses linear transformation to transform the original p relevant soil indices X=(X1,X2,…X…) p Convert the p data into p independent composite indices PC = (PC1, PC2, ..., PC2) p Its transformation model expression is:

[0014] In the formula: Z j It is the j-th standardized raw index; PCk is the k-th principal component; v kj It is the load, i.e., the j-th original index Z. j For the first The contribution weights of each principal component PCk constitute the feature vector. The variance Var(PCk) of each principal component is equal to its corresponding eigenvalue λ. k The contribution rate of the principal components is determined by... The calculation of the Soil Quality Index (SQI) involves using a nonlinear membership function to score the indicators in the minimum dataset, followed by weighted summation. The weight is the ratio of the variance of the principal component corresponding to each indicator to the cumulative variance. The expression for the nonlinear membership function is: S

[0015]

[0016] In the formula, S The score is (0–1). x The measured values ​​are the actual values ​​of the indicators; H and L represent the maximum and minimum values ​​among the measured values ​​of a specific indicator, respectively. The specific expression for calculating SQI using the weighted method is as follows:

[0017] In the formula, n represents the number of evaluation indicators; S i w represents the score of the i-th indicator; i This represents the weight of the i-th indicator, which is the ratio of the variance of the principal component corresponding to indicator i in the MDS to the cumulative variance.

[0018] Preferably, the expression for the empirical model described in S4 is:

[0019] In the formula ， CPI is the arable land productivity index; A, B, and n are model parameters; EC is soil electrical conductivity; the expression for the interaction term multivariate nonlinear regression model is:

[0020] In the formula, SQI is the soil quality index; β 0 to β 3 represents the parameter to be estimated. For residuals; As a preferred embodiment, when constructing the evaluation model in S4, crop-specific evaluation sub-models are established for different salt-tolerant crops such as sunflower, corn, and cotton. The sub-models adjust the model parameters based on the growth period characteristics and salt tolerance thresholds of different crops.

[0021] Preferably, the specific expression of the structural equation model (SEM) described in S5 is as follows:

[0022] Among them, path coefficient , ,and The direct effects of salinization on soil quality, soil quality on arable land productivity, and salinization on arable land productivity were estimated using partial least squares method.

[0023] Preferably, the dynamic update in S5 collects data in real time through soil sensors and crop growth sensors, and combines historical salinity change patterns with crop growth response characteristics to achieve real-time assessment and early warning of the impact on productivity. The management scenarios include improvement measures such as irrigation method optimization, organic material input, and spring irrigation to reduce salt content. By simulating changes in salinization degree and soil quality index under different scenarios, the evaluation results of the improvement effect of arable land productivity are output.

[0024] Compared with existing technologies, the method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model provided by this invention has the following beneficial effects: 1. More comprehensive data coverage and more efficient integration; This solution overcomes the limitations of existing technologies that rely on a single data source. It integrates remote sensing data, ground monitoring data, meteorological data, and farmer management data to construct a long-term spatially distributed dataset, capable of comprehensively capturing multi-dimensional information on soil salinization, crop growth, and environmental management. Simultaneously, it employs a multivariate stepwise linear regression-based satellite-ground hyperspectral fusion method, combined with principal component analysis and discriminant analysis to select the minimum dataset, reducing data redundancy while ensuring no core information is lost, thus improving data utilization efficiency.

[0025] 2. The evaluation index is constructed more scientifically and accurately; Compared to existing technologies that simply assess soil quality and arable land productivity, this method combines multiple models, including nonlinear membership functions and weighting methods, to calculate the Soil Quality Index (SQI). It then uses a normalization model to convert crop yield into the Arable Land Productivity Index (CPI), providing a more objective reflection of soil quality and the potential productivity of arable land. The SQI encompasses multiple soil physical, chemical, and biological indicators, while the CPI eliminates the influence of absolute differences in yield data, making the evaluation results more comparable and valuable.

[0026] 3. The evaluation model is more dynamic and targeted; This proposed solution constructs a multi-dimensional dynamic evaluation model that not only uses a random forest model to screen key salinization indicators but also quantifies various effects through a partial least squares path model, revealing the synergistic impact mechanism of salinization degree, soil quality, and arable land productivity. Simultaneously, it establishes crop-specific sub-models for different salt-tolerant crops such as sunflower, corn, and cotton, and combines real-time IoT monitoring data to dynamically update the productivity impact at different growth stages and seasons, solving the problems of static assessment and lack of crop-specificity in existing technologies.

[0027] 4. Decision support is more practical and instructive; Existing technologies often only address impact assessments. This solution, however, uses structural equation modeling (SEM) to simulate changes in productivity under different management scenarios, such as optimized irrigation methods and organic material inputs, directly identifying the optimal improvement technology. It provides not only assessment results but also actionable farmland improvement solutions, offering farmers and relevant departments precise decision-making support and effectively promoting the efficient utilization and sustainable development of saline-alkali land resources. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the multi-dimensional dynamic model framework of the present invention; Figure 2 This is a schematic diagram of the influence path framework based on structural equation modeling in this invention; Figure 3 This is a schematic diagram of the spatial distribution of secondary soil salinization in the study area in this embodiment of the invention; Figure 4 This is a schematic diagram of the average score of soil indicators, soil quality index, and arable land productivity index in an embodiment of the present invention. Figure 5 This is a schematic diagram illustrating the spatial distribution characteristics of the soil quality index according to an embodiment of the present invention. Figure 6 This is a schematic diagram illustrating the correlation analysis between soil salinization index and soil quality index in an embodiment of the present invention. Figure 7 This is a schematic diagram illustrating the importance of soil secondary salinization index to soil quality index in an embodiment of the present invention. Figure 8 This is a schematic diagram of the spatial distribution of the arable land productivity index according to an embodiment of the present invention; Figure 9 This is a schematic diagram illustrating the correlation between the arable land productivity index and the soil secondary salinization index in an embodiment of the present invention. Figure 10 This is a schematic diagram illustrating the importance of soil secondary salinization index to arable land productivity index in an embodiment of the present invention. Figure 11 This is a schematic diagram illustrating the assessment of the impact of secondary soil salinization properties on soil quality and arable land productivity based on the partial least squares path model of this invention. Figure 12 This is a schematic diagram illustrating the correlation analysis between soil quality index and arable land productivity index in an embodiment of the present invention. Figure 13 This is a schematic diagram illustrating the nonlinear relationship between soil secondary salinization index and arable land productivity index in an embodiment of the present invention. Figure 14 This is a schematic diagram illustrating the multivariate nonlinear relationship between secondary salinization index, soil quality index, and arable land productivity index in an embodiment of the present invention. Figure 15 This is a schematic diagram illustrating the impact of agricultural management measures based on structural equation modeling on the arable land productivity index according to the present invention. Figure 16 This is a schematic diagram of the process for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0030] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0031] Example 1; To address the problems mentioned in the technical solutions, this application provides a method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model. The specific steps are as follows: S1. Construct a multi-source dataset: Collect remote sensing data, ground monitoring data, meteorological data and farmer management data of the study area. The ground monitoring data includes soil physicochemical property data and crop growth status data to establish a long-term spatial distribution dataset. S2. Data Fusion and Preprocessing: Multi-source data were integrated using a multivariate stepwise linear regression satellite-ground hyperspectral fusion method. Principal component analysis and discriminant analysis were used to construct a minimal dataset of soil physical, chemical, and biological properties. S3. Index Calculation: Based on the minimum dataset of soil physical, chemical and biological properties obtained by principal component analysis and discriminant analysis, the soil quality index (SQI) is calculated by combining models such as fuzzy logic, random forest and nonlinear scoring model. At the same time, crop yield is converted into arable land productivity index (CPI) to characterize the potential productivity of arable land through normalization model. Among them, the random forest model was used to quickly diagnose the core secondary salinization indicators affecting arable land productivity, and the arable land productivity index (CPI) and soil quality index (SQI) were constructed.

[0032] 4.2.1 Random Forest Model The system is trained with production capacity output as the dependent variable and various indicators as independent variables. The importance ranking of these variables visually reveals which secondary salinization indicators contribute most to production capacity prediction.

[0033] CPI = f (pH, EC, ENa, ESP) + ε

[0034] In the formula, MSE represents the minimum mean squared error. The larger the %IncMSE value, the more important the feature is for predicting the dependent variable (CPI).

[0035] S4. Construct a multi-dimensional dynamic evaluation model: Key salinization indicators such as soil pH and electrical conductivity (EC) are screened out using the random forest model. The partial least squares path model is used to quantify the degree of salinization, the various effects between soil quality index (SQI) and arable land productivity index (CPI). Empirical models, interaction term multivariate nonlinear regression models and other mechanistic models are constructed to reveal synergistic effects. Sub-crop evaluation sub-models are also established for different salt-tolerant crops. S5. Dynamic Assessment and Decision Support: Combining real-time monitoring data from the Internet of Things with historical data, the system dynamically updates the impact of different crop growth stages and seasons on production capacity. It also uses structural equation modeling (SEM) to simulate production capacity changes under different management scenarios and selects the optimal improvement technology model.

[0036] The remote sensing data mentioned in S1 are multispectral satellite remote sensing images and UAV image data, used to obtain macroscopic salinity distribution information; meteorological data include precipitation, evaporation, temperature, and light data; and farmer management data includes irrigation methods, fertilizer application rates, and crop variety selection records.

[0037] The soil physicochemical properties data described in S2 include soil bulk density, water content, average weight diameter of aggregates (MWD), organic matter, total nitrogen, available phosphorus (AP), available potassium (AK), and salt ion content. The soil biological properties data include microbial biomass and enzyme activity data.

[0038] The specific expression for the normalized model in S3 is:

[0039] In the formula, X X represents the normalized value; i X represents the measured yield at the i-th sampling point; max and X min These represent the maximum and minimum values ​​of the measured output, respectively.

[0040] The soil principal component analysis described in S3 transforms the original p relevant soil indicators X=(X1,X2,…X…) using linear transformation. p Convert the p data into p independent composite indices PC = (PC1, PC2, ..., PC2) p Its transformation model expression is:

[0041] In the formula: Z jIt is the j-th standardized raw index; PCk is the k-th principal component; v kj It is the load, i.e., the j-th original index Z. j For the first The contribution weights of each principal component PCk constitute the feature vector. The variance Var(PCk) of each principal component is equal to its corresponding eigenvalue λ. k The contribution rate of the principal components is determined by... The calculation of the Soil Quality Index (SQI) involves using a nonlinear membership function to score the indicators in the minimum dataset, followed by weighted summation. The weight is the ratio of the variance of the principal component corresponding to each indicator to the cumulative variance. The expression for the nonlinear membership function is: S

[0042]

[0043] In the formula, S The score is (0–1). x The measured values ​​are the actual values ​​of the indicators; H and L represent the maximum and minimum values ​​among the measured values ​​of a specific indicator, respectively. The specific expression for calculating SQI using the weighted method is as follows:

[0044] In the formula, n represents the number of evaluation indicators; S i w represents the score of the i-th indicator; i This represents the weight of the i-th indicator, which is the ratio of the variance of the principal component corresponding to indicator i in the MDS to the cumulative variance.

[0045] The expression for the empirical model described in S4 is:

[0046] In the formula ， CPI is the arable land productivity index; A, B, and n are model parameters; EC is soil electrical conductivity; the expression for the interaction term multivariate nonlinear regression model is:

[0047] In the formula, SQI is the soil quality index; β 0 to β 3 represents the parameter to be estimated. For residuals; When constructing the evaluation model in S4, sub-models for different salt-tolerant crops such as sunflower, corn, and cotton are established. The sub-models adjust the model parameters based on the growth period characteristics and salt tolerance thresholds of different crops.

[0048] The specific expression for the structural equation model (SEM) described in S5 is as follows:

[0049] Among them, path coefficient , ,and The direct effects of salinization on soil quality, soil quality on arable land productivity, and salinization on arable land productivity were estimated using partial least squares method.

[0050] The dynamic update described in S5 uses soil sensors and crop growth sensors to collect data in real time. Combined with historical salinity change patterns and crop growth response characteristics, it enables real-time assessment and early warning of the impact on productivity. The management scenarios include improvement measures such as irrigation method optimization, organic material input, and spring irrigation to reduce salt content. By simulating changes in salinization degree and soil quality index under different scenarios, the system outputs the evaluation results of the effect of improving arable land productivity.

[0051] Example 2: Based on Example 1 but with a difference, the following describes the method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model proposed in this invention, with specific examples and accompanying drawings. The specific content is as follows.

[0052] This embodiment, based on typical saline-alkali land in the Yellow River irrigation area of ​​Inner Mongolia, used a systematic grid method to collect soil samples from the 0-20cm soil layer, measuring 17 soil physical, chemical, and microbiological indicators, and analyzing the spatiotemporal variation characteristics of secondary soil salinization indicators. Principal component analysis combined with the weighted method was used to assess the Soil Quality Index (SQI), and the normalization method was used to calculate the Crop Productivity Index (CPI). A random forest model was used to identify key driving factors affecting the CPI. Linear regression and partial least squares path models were used to reveal the mechanism by which the degree of secondary soil salinization affects the CPI. Univariate and multivariate nonlinear regression models were used to construct an assessment model of the impact of the degree of secondary salinization on the CPI. Finally, structural equation modeling was used to reveal the complex relationship between the degree of secondary salinization, SQI, and CPI under irrigation methods and organic input agricultural management practices.

[0053] The specific steps are as follows: Sample collection and determination of basic indicators; 1. Study Area and Sampling Sites: Typical saline-alkali land in the Yellow River irrigation area of ​​Inner Mongolia was selected as the study object. Sampling points were set up in the study area using a systematic grid method. Soil samples from the 0-20cm soil layer were collected in a targeted manner to ensure that the samples covered different terrains, irrigation conditions and planting areas in the area, thus ensuring the representativeness of the data.

[0054] 2. Indicator Measurement: Seventeen basic soil indicators were systematically measured on the collected soil samples, covering three major categories: physical, chemical, and microbiological, providing data support for subsequent analysis. Physical indicators included bulk density, water content, and mean weight diameter (MWD); chemical indicators included pH, electrical conductivity (EC), exchangeable sodium ions (ENa), sodium adsorption ratio (ESP), available potassium (AK), available phosphorus (AP), organic carbon, and total nitrogen; microbiological indicators included bacterial biomass, fungal biomass, total microbial biomass, and the activity of carbon, nitrogen, and phosphorus metabolic enzymes. Simultaneously, sunflower yield data within the study area were recorded as a basis for assessing arable land productivity.

[0055] Analysis of the spatiotemporal distribution characteristics of secondary soil salinization; Table 1 shows the results of secondary soil salinization indices in the Yellow River irrigation area of ​​Inner Mongolia. Over time, soil pH showed relatively small fluctuations, while electrical conductivity (EC) varied considerably. Compared to the average EC in May, the average EC in July and October increased by 82.6% and 161.6%, respectively. Spatially, soil pH ranged from 7.67 to 8.86, with a low coefficient of spatial variation (2.1%–3.3%); while soil electrical conductivity (EC) ranged from 0.31 to 3.30 mS / cm. -1 The spatial variation coefficient ranged from 29.2% to 61.2%. Furthermore, the average exchangeable sodium ion (ENa) content was 2.55 cmol·kg⁻¹. -1 The coefficient of variation was 28.6%; the average sodium adsorption ratio (ESP) was 4.37%, with a coefficient of variation of 26.8%. Soil electrical conductivity (EC), exchangeable sodium ions (ENa), and sodium adsorption ratio (ESP) showed a higher distribution in the north than in the south, and higher distribution in the east than in the west. Figure 3 As shown in the figure. The average electrical conductivity (EC) in the eastern region is 2.70 mS·cm. -1 The average value in the west is 1.02 mS·cm. -1 The mean values ​​of exchangeable sodium ions (ENa) and sodium adsorption ratio (ESP) in the northern region were 3.08 cmol·kg⁻¹. -1 The pH was 4.96%, significantly higher than in other areas. In summary, soil pH showed little variability in time and space, while electrical conductivity (EC) showed greater variability. This is because the local area experiences heavy irrigation in spring to suppress salt, resulting in low surface salt content in May; however, rising summer temperatures lead to increased soil moisture evaporation, and reduced autumn irrigation causes salt to accumulate in the surface layer. Furthermore, the spatial variability of electrical conductivity (EC) may be strongly influenced by micro-topography and field irrigation. The study area in this example exhibits a topography that slopes from west to east and from south to north, causing salt to accumulate in low-lying areas with irrigation or rainfall, resulting in a salt distribution pattern of "high in the northeast and low in the southwest."

[0056] Table 1 is a descriptive statistical analysis table of soil secondary salinization indicators.

[0057] Construction and assessment of the Soil Quality Index (SQI); Principal component analysis (PCA) yielded three principal components with eigenvalues ​​greater than 1, accounting for 72.4% of the cumulative variance, as shown in Table 2. In PC1, the index with the highest loading was MWD; in PC2, the indices with relatively high loadings were AK and Pacq, as shown in Table 2. Figure 4 As shown, only AK was retained because the correlation coefficient between AK and Pacq was 0.68; similarly, AP was selected in PC3. In summary, the minimum dataset (MDS) indices are MWD, AK, and AP. This indicates that soil physical structure and nutrient availability are key indicators for assessing soil quality. Among them, MWD reflects the stability of soil aggregates, and higher salinity (especially sodium ions) in saline-alkali land reduces MWD by dispersing soil particles. In addition, AK and AP are key indicators reflecting soil nutrient availability. In existing technologies, salt stress inhibits the availability of AK, which may be attributed to ion competition. Secondary soil salinization also reduces phosphatase activity, affects the community structure of phosphorus cycling-related microorganisms, hinders the mineralization of organic phosphorus, and thus affects AP availability. In conclusion, secondary soil salinization reduces soil quality by affecting soil physical structure and nutrient supply capacity.

[0058] Based on the membership function, the average scores of soil physical, chemical, and microbiological indicators are as follows: Figure 4 As shown, the SQI ranges from 0.20 to 0.93, with an average of 0.50. Specifically, 31.3% of respondents had an SQI between 0.20 and 0.40, 45.8% between 0.40 and 0.60, 12.5% ​​between 0.60 and 0.80, and only 10.4% had an SQI greater than 0.80. Spatially, the SQI is higher in the south than in the north, and higher in the west than in the east, as shown in the figure. Figure 5 As shown, this is contrary to the distribution characteristics of secondary soil salinization, reflecting that the soil quality in this study area is significantly affected by secondary salinization stress, and the soil quality is lower in areas with higher degrees of secondary salinization.

[0059] Table 2. Statistical table of principal component analysis of soil quality indicators;

[0060] Construction of the Cultivated Land Productivity Index (CPI) and identification of key driving factors; In this embodiment, the sunflower yield in the study area ranged from 1675.3 to 3608.7 kg·hm². -2 The CPI is obtained through the normalization method, such as... Figure 4As shown, the average CPI was 0.62. Among the cultivated land areas, 47.9% had a CPI > 0.80, 8.3% had a CPI between 0.60 and 0.80, 8.3% had a CPI between 0.40 and 0.60, 18.8% had a CPI between 0.20 and 0.40, and 16.7% had a CPI < 0.20. Spatially, the CPI was higher in the south than in the north, and higher in the west than in the east, as shown in the figure. Figure 8 As shown in the figure, this indicates that the overall arable land productivity in this study area is at a moderate level, possibly due to the fact that sunflowers are a moderately salt-tolerant crop, and soils with low secondary salinization have little impact on sunflower growth. However, 16.7% of the arable land still has extremely low productivity (CPI < 0.20), mainly distributed in the northeastern part of the study area, such as... Figure 8 As shown in the figure, this indicates that changes in CPI are closely related to spatial differences in the degree of secondary salinization.

[0061] Linear regression analysis revealed a significant positive correlation between CPI and October pH (P=0.04). Figure 9 As shown; however, CPI and EC showed a highly significant negative correlation (P < 0.001), with CPI explaining 43%, 50%, and 85% of the variance in May, July, and October, respectively. Similarly, CPI and ENa showed a highly significant negative correlation (R² = 0.48, P < 0.001); SQI and ESP showed a highly significant negative correlation (R² = 0.35, P < 0.001). The importance of soil secondary salinization indicators to CPI was analyzed using a random forest model, such as... Figure 10 As shown in the figure, the results indicate that the secondary salinization index with the greatest impact on CPI is EC in July and October, followed by EC and ENa in May, and pH and ESP in July and May. This result suggests that secondary soil salinization from July to October has a greater impact on arable land productivity, implying that arable land productivity is jointly determined by crop growth stage and soil salinity changes. Existing technology shows that sunflower seeds have a certain degree of salt tolerance, resisting the effects of salt stress on seed germination and seedling growth to some extent. In this example study, although May and June are the seedling stage of sunflowers, a vulnerable period for the crop, the soil salinity level is relatively low at this time, thus having a limited impact on final productivity. However, July to October is the critical growth period for sunflowers (budding, flowering, grain-filling, and maturity stages), which is also the period of rising soil salinity. At this time, higher salinity reduces root water absorption through osmotic stress, reduces the photosynthetic performance and leaf area of ​​sunflower leaves, and thus leads to crop yield reduction. In conclusion, the results of this study indicate that higher levels of secondary soil salinization significantly reduce arable land productivity, especially during critical crop growth periods when the impact of secondary soil salinization on arable land productivity is even greater.

[0062] Salting, the relationship between SQI and CPI, and model construction; This study revealed the impact of secondary soil salinization on arable land productivity through multi-model analysis. Partial least squares path model analysis showed that soil secondary salinization indices and the soil quality index (SQI) could explain 65.8% of the changes in the CPI. Figure 11 As shown, secondary soil salinization indicators directly affect CPI (standard path coefficient = -0.610; P < 0.001) and also influence CPI through a negative effect on SQI (standard path coefficient = -0.694; P < 0.001), with SQI having a direct positive effect on CPI (standard path coefficient = 0.258; P < 0.05). The total effect of secondary soil salinization indicators on CPI is -0.789. This reflects the dual pathway by which secondary soil salinization affects arable land productivity. On the one hand, the direct pathway of secondary soil salinization on CPI reflects the physiological toxicity of salt to crops, especially the accumulation of exchangeable sodium ions in the soil, which damages plant cells and affects crop growth. On the other hand, secondary soil salinization reduces CPI through an indirect pathway by inhibiting SQI. A significant positive correlation exists between SQI and CPI, such as... Figure 12 As shown in Table 2, SQI has a direct positive effect on CPI (standard path coefficient = 0.258). This indicates that soil quality is a key intermediary linking secondary salinization and productivity. Soil physical properties (MWD) and nutrient properties (AP and AK) are the core indicators for assessing SQI. On the one hand, increased secondary soil salinization leads to degradation of soil physical structure (reduced MWD); on the other hand, increased secondary soil salinization affects the availability of basic nutrients (nitrogen, phosphorus, and potassium), thus limiting crop growth and yield. In summary, secondary soil salinization inhibits crop growth and nutrient absorption by damaging soil physical structure and limiting nutrient supply, thereby affecting crop productivity. Therefore, improving arable land productivity requires not only controlling salt levels during critical crop growth periods to reduce the physiological toxicity of salt to crops, but also fundamentally improving soil quality through a series of tillage, irrigation and drainage, and organic improvement measures.

[0063] Based on the above analysis, this study used a univariate nonlinear regression model to analyze the relationship between the soil secondary salinization index (EC) and CPI, obtaining the function CPI = 1.02 - 0.17. EC 1.48 (R) 2 =0.88; for example Figure 13 (As shown). This indicates that arable land productivity exhibits a slow-to-rapid decrease process with increasing secondary soil salinization. When EC < 1.3 ms cm -1 When CPI remains at a high level (>0.8), the rate of decline in CPI gradually accelerates as EC increases; when EC > 3.4 ms cm -1 At that time, the CPI dropped to 0.

[0064] Furthermore, by constructing an interactive multiple nonlinear regression model, the effects of soil secondary salinization index (electrical conductivity EC), SQI, and their interaction on CPI were systematically analyzed, yielding the regression model: CPI = 0.64 - 0.44EC - 0.07SQI + 0.188(EC × SQI) + 0.64(R²) 2 =0.93; for example Figure 14 (As shown in the figure). The model's fit is significantly higher than that of the univariate nonlinear regression model, indicating that secondary soil salinization and the Soil Quality Index (SQI) do not act independently on arable land productivity, but rather have a significant synergistic effect. This result highlights that improving the productivity of saline-alkali land requires a dual approach of "reducing secondary soil salinization" and "improving soil quality," providing a scientific basis for the efficient utilization and productivity enhancement of saline-alkali land.

[0065] Verification of the effects of agricultural management measures; This study used structural equation modeling (SEM) to assess the effects of irrigation methods and organic input agricultural management practices on secondary salinization, soil quality index (SQI), and soil saturation index (CPI). Irrigation methods had a significant negative impact on secondary salinization (path coefficient = -0.48), indicating that optimized irrigation can effectively reduce the risk of soil salt accumulation; organic input also had a negative impact on secondary salinization (path coefficient = -0.10). Both irrigation methods and organic input had significant positive effects on SQI, indicating that they can directly improve soil quality. Secondary salinization had a highly significant negative impact on both SQI and CPI, while SQI had a very strong positive driving effect on CPI (0.82). Analysis of the overall effects of various factors on CPI revealed that irrigation method and organic input had significant positive overall effects on CPI, with the positive effect of organic input slightly greater than that of irrigation method. Secondary salinization had a significant negative effect on CPI, being a key factor restricting arable land productivity. Soil quality (SQI) had a significant positive overall effect on CPI, with a high intensity, further highlighting the supporting role of soil quality in arable land productivity. These results indicate that changing irrigation methods (from flood irrigation to drip irrigation) positively impacts arable land productivity through a dual pathway of directly reducing salinization and directly increasing SQI. This may be related to drip irrigation reducing salt accumulation on the surface and maintaining soil moisture balance.

[22] Organic inputs (adding cow manure, straw, and carbon-based fertilizers) increase soil organic matter, promote the formation of soil aggregates, and enhance soil nutrient retention capacity. Combining drip irrigation with organic material additions can provide technical support for salinization control, soil quality improvement, and increased arable land productivity.

[0066] Integration of research findings; This study, based on typical saline-alkali land in the Yellow River irrigation area of ​​Inner Mongolia, collected soil samples from the 0-20 cm soil layer using a systematic grid method, measured 17 soil physical, chemical, and microbiological indicators, and analyzed the spatiotemporal variation characteristics of secondary soil salinization indicators. Principal component analysis combined with the weighted method was used to assess the Soil Quality Index (SQI), and the normalization method was used to calculate the Crop Productivity Index (CPI). A random forest model was used to identify key driving factors affecting CPI. Linear regression and partial least squares path models were used to reveal the mechanism by which the degree of secondary soil salinization affects CPI. Univariate and multivariate nonlinear regression models were used to construct an assessment model of the impact of secondary salinization on CPI. Finally, structural equation modeling revealed the complex relationship between the degree of secondary salinization, SQI, and CPI under irrigation methods and organic input agricultural management practices. The results show that electrical conductivity (EC) is a sensitive indicator of secondary soil salinization. In this study area, EC exhibits significant spatiotemporal differentiation—gradually increasing from May to October, and spatially distributed with a "high in the northeast and low in the southwest" pattern influenced by micro-topography. The soil quality index (SQI) constructed based on average weight diameter (MWD) of aggregates, available potassium (AK), and available phosphorus (AP) shows that the overall soil quality in this region is moderately low, with soil quality decreasing further in areas with higher levels of secondary salinization. The arable land productivity index (CPI) exhibits significant spatial heterogeneity, with the strongest negative impact from the end-of-season (EC) period (July–October). Secondary soil salinization affects arable land productivity through two pathways: direct salt toxicity and regulation of the SQI, with a total effect of -0.789. A univariate nonlinear model was constructed to determine the impact of secondary soil salinization on arable land productivity: CPI = 1.02–0.17. EC 1.48 (R) 2 =0.88), indicating that arable land productivity decreases slowly at first and then rapidly with the increase of secondary soil salinization. An interactive multiple nonlinear regression model was constructed to obtain the regression model for "secondary soil salinization—SQI—CPI": CPI = 0.64–0.44. EC-0.07 SQI+0.188 (EC SQI)+0.64(R) 2 The coefficient of variation (0.93) demonstrates that secondary soil salinization and the Soil Quality Index (SQI) do not act independently on arable land productivity, but rather exhibit a significant synergistic effect. Structural equation modeling results indicate that drip irrigation and organic material addition improve soil conditions and enhance arable land productivity through a dual pathway of directly reducing salinization and increasing SQI, making them effective and comprehensive measures for achieving sustainable development of saline-alkali land.

[0067] Please refer to the above work process. Figures 1 to 15 .

[0068] It should be noted that the term "comprising" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0069] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model, characterized in that, Includes the following steps: S1. Construct a multi-source dataset: Collect remote sensing data, ground monitoring data, meteorological data and farmer management data of the study area. The ground monitoring data includes soil physicochemical property data and crop growth status data to establish a long-term spatial distribution dataset. S2. Data Fusion and Preprocessing: Multi-source data were integrated using a multivariate stepwise linear regression satellite-ground hyperspectral fusion method. Principal component analysis and discriminant analysis were used to construct a minimal dataset of soil physical, chemical, and biological properties. S3. Index Calculation: Based on the minimum dataset of soil physical, chemical and biological properties obtained by principal component analysis and discriminant analysis, the soil quality index (SQI) is calculated by combining models such as fuzzy logic, random forest and nonlinear scoring model. At the same time, crop yield is converted into arable land productivity index (CPI) to characterize the potential productivity of arable land through normalization model. S4. Construct a multi-dimensional dynamic evaluation model: Key salinization indicators such as soil pH and electrical conductivity (EC) are screened out using a random forest model. Partial least squares path model is used to quantify the degree of salinization, various effects between soil quality index (SQI) and arable land productivity index (CPI). Empirical models, interaction term multivariate nonlinear regression models and other mechanistic models are constructed to reveal synergistic effects. Sub-crop evaluation sub-models are also established for different salt-tolerant crops. S5. Dynamic Assessment and Decision Support: Combining real-time monitoring data from the Internet of Things with historical data, the system dynamically updates the impact of different crop growth stages and seasons on production capacity. It also uses structural equation modeling (SEM) to simulate production capacity changes under different management scenarios and selects the optimal improvement technology model.

2. The method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to claim 1, characterized in that, The remote sensing data mentioned in S1 are multispectral satellite remote sensing images and UAV image data, used to obtain macroscopic salinity distribution information; meteorological data include precipitation, evaporation, temperature, and light data; and farmer management data includes irrigation methods, fertilizer application rates, and crop variety selection records.

3. The method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to claim 1, characterized in that, The soil physicochemical properties data described in S2 include soil bulk density, water content, average weight diameter of aggregates (MWD), organic matter, total nitrogen, available phosphorus (AP), available potassium (AK), and salt ion content. The soil biological properties data include microbial biomass and enzyme activity data.

4. The method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to claim 1, characterized in that, The specific expression for the normalized model in S3 is: In the formula, X X represents the normalized value; i X represents the measured yield at the i-th sampling point; max and X min These represent the maximum and minimum values ​​of the measured output, respectively.

5. The method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to claim 1, characterized in that, The soil principal component analysis described in S3 transforms the original p relevant soil indicators X=(X1,X2,…X…) using linear transformation. p Convert the p data into p independent composite indices PC = (PC1, PC2, ..., PC2) p Its transformation model expression is: In the formula: Z j It is the j-th standardized raw index; PCk is the k-th principal component; v kj It is the load, i.e., the j-th original index Z. j For the first The contribution weights of each principal component PCk constitute the feature vector. The variance Var(PCk) of each principal component is equal to its corresponding eigenvalue λ. k The contribution rate of the principal components is determined by... The calculation of the Soil Quality Index (SQI) involves using a nonlinear membership function to score the indicators in the minimum dataset, followed by weighted summation. The weight is the ratio of the variance of the principal component corresponding to each indicator to the cumulative variance. The expression for the nonlinear membership function is: S In the formula, S The score is (0–1). x The measured values ​​are the actual values ​​of the indicators; H and L represent the maximum and minimum values ​​among the measured values ​​of a specific indicator, respectively. The specific expression for calculating SQI using the weighted method is as follows: In the formula, n represents the number of evaluation indicators; S i w represents the score of the i-th indicator; i This represents the weight of the i-th indicator, which is the ratio of the variance of the principal component corresponding to indicator i in the MDS to the cumulative variance.

6. The method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to claim 1, characterized in that, The expression for the empirical model described in S4 is: In the formula ， CPI is the arable land productivity index; A, B, and n are model parameters; EC is soil electrical conductivity; the expression for the interaction term multivariate nonlinear regression model is: In the formula, SQI is the soil quality index; β 0 to β 3 represents the parameter to be estimated. It represents the residual.

7. The method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to claim 1, characterized in that, When constructing the evaluation model in S4, sub-models for different salt-tolerant crops such as sunflower, corn, and cotton are established. The sub-models adjust the model parameters based on the growth period characteristics and salt tolerance thresholds of different crops.

8. The method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to claim 1, characterized in that, The specific expression for the structural equation model (SEM) described in S5 is as follows: Among them, path coefficient , ,and The direct effects of salinization on soil quality, soil quality on arable land productivity, and salinization on arable land productivity were estimated using partial least squares method.

9. The method for evaluating the impact of soil salinization on arable land productivity under a multi-dimensional dynamic model according to claim 1, characterized in that, The dynamic update described in S5 uses soil sensors and crop growth sensors to collect data in real time. Combined with historical salinity change patterns and crop growth response characteristics, it enables real-time assessment and early warning of the impact on productivity. The management scenarios include improvement measures such as irrigation method optimization, organic material input, and spring irrigation to reduce salt content. By simulating changes in salinization degree and soil quality index under different scenarios, the system outputs the evaluation results of the effect of improving arable land productivity.

Images