A rapid prediction method for marine organic carbon content in coastal wetlands based on soil particle size distribution.

Abstract

This invention belongs to the field of environmental monitoring technology, specifically relating to a method for predicting soil marine organic carbon. (1) The content of each particle size, organic carbon and total nitrogen content, and isotope δ in soil samples. 13 (2) The C value was determined based on the ratio of organic carbon to total nitrogen content and the isotope δ. 13 C value, calculate the total organic carbon content of marine organic carbon; (3) fit the content of each particle size in the soil sample; (4) perform ridge regression fitting between the soil particle size data and the marine organic carbon content, determine the optimal K value according to the ridge trace map, and construct the marine organic carbon ridge regression prediction model accordingly, so as to realize the prediction of marine organic carbon content. This invention performs regression analysis on each particle size component of soil based on the ridge regression model, avoids the influence of data collinearity on the model, and increases the accuracy and versatility of the prediction; during the measurement, it is not necessary to perform a large number of element and isotope measurements on the sample, which greatly reduces the workload.

Description

Technical Field

[0001] This invention belongs to the field of environmental monitoring technology, specifically relating to a method for predicting soil and marine organic carbon. Background Technology

[0002] Coastal wetlands, located in the transition zone between land and sea, are unique areas primarily formed by the alluvial and depositional processes of upstream rivers. They possess high carbon deposition rates and carbon burial capacity, making them a crucial component of the "blue carbon" ecosystem. The strong carbon sequestration capacity of coastal wetlands benefits from unique tidal phenomena. As tides scour the intertidal zone, they also deposit large amounts of marine organic carbon particles into the soil. These sediments constitute a significant portion of the organic carbon in coastal wetland soils, a key characteristic distinguishing them from inland wetlands. Simultaneously, particulate organic carbon carried by upstream rivers gradually settles in the estuary area as the terrain flattens, forming the terrestrial organic carbon component of the soil. Furthermore, the accumulated litter from locally grown plants contributes to the plant-based organic carbon component of the coastal wetland soils. Due to the unique hydrological conditions of the confluence of rivers and the sea in coastal salt marshes, the composition of organic carbon in the soil exhibits a complex and diverse nature.

[0003] Coastal wetlands are flat and expansive, with the hydrodynamic intensity gradually decreasing as upstream rivers and tides arrive. Under hydrodynamic sorting, marine organic carbon particles of different sizes are deposited sequentially, resulting in a predictable distribution of soil particle size from sea to land. Furthermore, different particle size components exhibit varying adsorption capacities for particulate organic carbon due to their different specific surface areas; finer particles, with their larger surface area, have a stronger adsorption capacity than coarser particles. Under the combined influence of hydrodynamic sorting and the inherent properties of sediment particle size, the marine organic carbon content in the intertidal soils of coastal wetlands exhibits a regular gradient distribution characteristic from sea to land along with sediment particle size.

[0004] It is worth noting that there is a strong correlation among different particle size components in the intertidal soils of coastal wetlands. This leads to strong collinearity among the independent variables when fitting the particle size components and marine organic carbon using ordinary least squares, resulting in unstable coefficient estimates and excessive variance in the prediction model, thus affecting the model's generalizability. Ridge regression models, by regularizing the coefficients, can effectively address the inaccuracy caused by collinearity among particle size components in the intertidal soils of coastal wetlands, mitigating overfitting and improving the model's generalization ability. Therefore, constructing a ridge regression model for particle size components and marine organic carbon in the intertidal soils of coastal wetlands allows for a faster and more accurate prediction of marine organic carbon in the soil.

[0005] Currently, several methods exist for tracing the source of organic carbon in wetlands. Chinese patent document CN202210193492.5 discloses a method for determining the organic carbon content in marine sediments, which uses a modified potassium dichromate oxidation-spectrophotometric method to determine the organic carbon content in the ocean. However, this method is only suitable for determining the organic carbon content in samples from a single marine source and is not applicable to tracing the source of organic carbon in coastal wetlands from multiple sources. Chinese patent document CN202310646963.8 discloses a method for comprehensive tracing of the source of organic carbon in wetlands. This method determines the organic carbon from plant and microbial sources in wetland soils by measuring two biomarkers: n-alkanes and amino sugars. Chinese patent document CN202010196327.6 discloses a method, apparatus, equipment, and system for predicting the content of terrestrial organic carbon, which constructs a predictive model for deep-water terrestrial-marine source rocks based on hydrodynamic conditions and source distance.

[0006] Currently, the source tracing of organic carbon in coastal wetland soils mainly focuses on terrestrial, plant, and microbial organic carbon. There is still a lack of methods for predicting the content of marine organic carbon in the intertidal zone, so there is an urgent need for a method that can fill this gap. Summary of the Invention

[0007] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method for predicting marine organic carbon in intertidal soils of coastal wetlands under complex sources, thereby filling the gap in existing methods for tracing the source of soil organic carbon and providing direct guidance for exploring the distribution pattern of soil organic carbon in coastal wetlands.

[0008] To achieve the above objectives, the technical solution adopted by this invention is: a rapid prediction method for marine organic carbon content in coastal wetlands based on soil particle size distribution, comprising the following steps:

[0009] (1) Content of various particle sizes, organic carbon and total nitrogen content, and isotope δ in soil samples 13 The C value was measured.

[0010] (2) Based on the ratio of organic carbon to total nitrogen content and the isotope δ¹⁸O of the soil sample 13 C-value, used to calculate the proportion and specific content of organic carbon in total organic carbon from marine sources;

[0011] (3) Constructing a ridge regression prediction model for marine organic carbon: The soil particle size data measured in the laboratory were fitted with the marine organic carbon content using ridge regression. The optimal K value was determined based on the ridge trace map, and regression modeling was performed accordingly. The prediction model is as follows:

[0012] ;

[0013] in, Expressed as the organic carbon content of the seawater , All are constants. ; Expressed as the content of each particle size component;

[0014] (4) Fitting the content of each particle size in the soil sample:

[0015] ;

[0016] in, All are constants. ; Expressed as the content of clay components in sediments, Expressed as the content of fine silt components in sediments, Expressed as the content of medium-silt components in sediments, Expressed as the content of coarse silt components in sediments, Expressed as the content of sand components in sediments;

[0017] (5) By substituting the fitted data of the particle size content of each level in the soil sample into the prediction model, the prediction of marine organic carbon content in coastal wetlands based on soil particle size composition can be realized.

[0018] Preferably, in step (1), before determining the content of each particle size in the soil sample, the soil sample is pretreated as follows: H2O2 and hydrochloric acid are added to the air-dried soil sample in sequence to remove soil organic matter and inorganic salts; after the reaction, the sample is washed repeatedly with deionized water and allowed to stand; finally, sodium hexametaphosphate solution is added and shaken well.

[0019] Preferably, in step (1), before determining the organic carbon and total nitrogen content in the soil sample, the soil sample is pretreated as follows: hydrochloric acid is added to the air-dried soil sample; then the sample is repeatedly washed with deionized water until the supernatant is transparent and close to neutral, and then dried for later use.

[0020] Preferably, in step (2), a Bayesian three-terminal organic carbon mixture model of terrestrial, marine, and plant sources is constructed using the MixSIAR package in R language. The formula of this model is:

[0021] ;

[0022] in, and These are the measured values ​​of the samples, and These are the land-source end-member values ​​for each indicator. and These are the ocean endmember values ​​for each indicator. and These are the plant endmember values ​​for each indicator. The values ​​represent the percentage contributions of terrestrial, marine, and plant-derived sources to soil organic carbon in the samples, respectively.

[0023] The advantages of this invention compared to existing technologies are: it provides a rapid prediction method for marine organic carbon in coastal wetland soils; this method uses a ridge regression model to perform regression analysis on various particle size components of the soil, avoiding the influence of data collinearity on the model and increasing the accuracy and versatility of the prediction; this method does not require extensive elemental and isotopic analysis of samples when determining marine organic carbon in coastal wetland soils, greatly reducing the workload; and it fills the gap in prediction methods for marine organic carbon in coastal wetlands. Attached Figure Description

[0024] Figure 1 Linear fitting diagram of the conversion between soil clay content, medium silt content, coarse silt content, sand content and fine silt content;

[0025] Figure 2 This is a graph showing the correlation between measured and predicted values ​​of soil marine organic carbon and soil particle size distribution in a ridge regression model according to an embodiment of the present invention. Detailed Implementation

[0026] The present invention will be further described below with reference to specific embodiments and accompanying drawings. More details are set forth in the following description in order to provide a full understanding of the present invention. However, the present invention can obviously be implemented in many other ways different from those described herein. For those skilled in the art, any alternative improvements or modifications made to the embodiments of the present invention are within the protection scope of the present invention, and the protection scope of the present invention should not be limited by the content of this specific embodiment.

[0027] This embodiment demonstrates how to utilize the rapid prediction method for marine organic carbon content in coastal wetlands based on soil particle size composition according to the present invention. Soil sediment samples were collected from the Yellow River Delta Coastal Wetland Nature Reserve in Dongying City, Shandong Province, from river to sea. These samples were then brought back to the laboratory for analysis of soil particle size composition, soil organic carbon content, soil nitrogen content, and isotope δ¹⁸O. 13 C was determined by soil C / N ratio and isotope δ¹⁴. 13 Using the C-value combined with a Bayesian endmember mixture model, the marine organic carbon content in soil was calculated. Regression analysis was then performed between soil particle size distribution and marine organic carbon content, enabling rapid prediction of marine organic carbon content in coastal wetland soils. The specific scheme is as follows:

[0028] (1) Study Area: The Yellow River Delta Coastal Wetland Nature Reserve (37°38′N, 119°15′E) is located at the mouth of the Yellow River in Dongying City, Shandong Province. It is a highly representative estuarine wetland ecosystem in the world. It is a fan-shaped alluvial plain formed by the siltation brought by the Yellow River into the Bohai Sea. It is mainly composed of coastal tidal soil and saline soil, with wide and flat tidal flats and densely dendritic tidal channels. Tidal action is a unique hydrological condition of coastal wetlands. The combined action of the Yellow River runoff and the Bohai Sea water leads to silt deposition. This makes the Yellow River Delta estuary system constantly changing. The source of organic carbon is no longer a simple mixture of terrestrial and marine sources. Its distribution and migration patterns are also complex and diverse.

[0029] (2) Sampling method: Sampling points were selected in the area on the north bank of the Yellow River Delta Nature Reserve, which is simultaneously affected by the Yellow River and the Bohai Sea. A transect from the Yellow River beach to the coastline was set up. Three 1×1m quadrats were set up at each sampling point, with each quadrat spaced 20m apart horizontally. Information such as the latitude and longitude of the sampling points and the dominant vegetation species were recorded. Soil samples were collected from the quadrats using a soil auger according to the three-point sampling principle and placed in numbered self-sealing bags.

[0030] (3) Pretreatment for sample particle size determination: Soil samples were air-dried in a cool place, and obvious plant tissues and stones and other impurities were removed. 10.0 g of the air-dried soil sample was weighed and placed in a 50 ml numbered centrifuge tube. 10 mL of 20% H2O2 was added to the centrifuge tube until no more bubbles were produced to remove organic matter from the soil sample. Then, 10 mL of 10% hydrochloric acid was added to remove residual H2O2 and carbonates from the soil sample. After the reaction, the sample was washed three times with deionized water until the solution was close to or reached neutral, and then allowed to stand for 12 hours. Finally, 10 mL of 0.05 mol / L sodium hexametaphosphate solution was added and shaken well, and the soil particle size composition was ready for determination.

[0031] (4) Determination and classification of soil particle size distribution: After soil sample pretreatment, particle size analysis was performed using a Mastersizer 2000 laser particle size analyzer, with a measurement range of 0.01~3500μm. Each sample was tested three times, and the average value was taken as the result. Soil particle size classification adopted the American Soil Particle Size Scale, which divided soil particle components into clay (0.01~2μm), fine silt (2~5μm), medium silt (5~10μm), coarse silt (10~50μm), and sand (>50μm).

[0032] (5) Organic carbon (SOC), total nitrogen (TN) and isotope δ 13Soil sample pretreatment for C-value determination: Weigh 5g of air-dried soil sample into a 25ml numbered centrifuge tube, and add 1mol L⁻¹ hydrochloric acid in multiple batches until no more bubbles emerge, to remove inorganic carbon from the soil sample. Then, centrifuge the tube at 2000 rpm for five minutes, discard the supernatant, and add distilled water to wash away any residual hydrochloric acid. Repeat the above steps until the supernatant in the centrifuge tube is clear and nearly neutral. Finally, dry the acidified sample in a 65℃ oven to constant weight.

[0033] (6) Determination of Soil Organic Carbon (SOC) and C / N Ratio: An acidification combustion method was used. Approximately 10 mg of soil sample was weighed and placed in a tin boat, then sent to an elemental analyzer (B104 Elementar Unicube). Organic matter containing C and N elements in the sample was burned under high temperature and high-purity oxygen conditions, and then reduced in a reduction tube to a stable mixed gas containing CO2 and N2. After capture and separation by an adsorption column, combined with the original weight of the sample, the soil organic carbon (SOC) and total nitrogen (TN) content in the sample could be calculated. Furthermore, the C / N ratio of the sample could be calculated using the following formula:

[0034] .

[0035] (7) Isotope δ 13 C-value determination: Stable isotope combustion mass spectrometry was used. Approximately 40 mg of soil sample was weighed into a tin container, then folded into a small sphere and sent to a stable isotope mass spectrometer (253Plus, Thermo Fisher Scientific). The pretreatment method for this index was the same as that for the determination of SOC and TN content, coupled with an elemental analyzer (B104 Elementar Unicube). The sample was combusted by the elemental analyzer, and after elemental separation, it was sent to the mass spectrometer for analysis. The signal intensity of the target element was obtained by comparing it with the reference gas. Repeated tests of the sample showed δ 13 The analytical precision of Corg is higher than 0.05‰, and the calculation formula is as follows:

[0036] .

[0037] (8) Calculation of marine organic carbon content: The marine organic carbon content was calculated using the Bayesian mixture model (MixSIAR). This model uses the MixSIAR package in R language to combine prior knowledge and observed data, utilizing the C / N value and isotope δ of the soil sample. 13 The C-value is used to calculate the contribution ratio of organic carbon from different sources in the soil to the total organic carbon. This model is based on the following formula:

[0038] ;

[0039] in, and These are the measured values ​​of the samples, and These are the land-source end-member values ​​for each indicator. and These are the ocean endmember values ​​for each indicator. and These are the plant endmember values ​​for each indicator. The values ​​represent the percentage contributions of terrestrial, marine, and plant-derived sources to soil organic carbon in the samples, respectively.

[0040] Referring to the three end-members of terrestrial, marine, and plant origin Values ​​and isotopes The contribution ratio and specific content of marine organic carbon in the soil of each sampling point were calculated.

[0041] (9) Ridge regression fitting was performed on the soil particle size data and the marine organic carbon content. The optimal K value was determined based on the ridge trace map, and regression modeling was performed accordingly. The prediction model is as follows:

[0042] ;

[0043] in, This is expressed as the organic carbon content of the seawater. Expressed as the content of clay components (< 2 μm) in sediments. Expressed as the content of fine silt particles (2~5 μm) in sediments. Expressed as the content of medium-grained silt (5~10 μm) in sediments. Expressed as the content of coarse silt particles (10~50 μm) in sediments. Expressed as the content of sand components (> 50 μm) in sediments. Expressed as the content of each particle size component, , All are constants ( ).

[0044] In this embodiment, the ridge regression prediction model based on marine organic carbon and soil particle size composition in coastal wetland soil is as follows:

[0045] .

[0046] (10) According to the principle of hydrodynamic sorting, as the hydrodynamic intensity changes, there are regular changes in the content of particle size components at different orders of magnitude. Specifically, when seawater washes over tidal flats, the flow velocity slows down, the hydrodynamic intensity of the tide gradually weakens, and the coarse marine sediment particles attached to the seawater settle before the fine sediment particles, resulting in an increase in the proportion of fine sediment particles in the total particulate sediment in the soil, which constitutes the mutual conversion between the contents of particle size components at different orders of magnitude. The particle size component model is as follows:

[0047] ;

[0048] in, All are constants. .

[0049] In this embodiment, based on hydrodynamic sorting conditions, after fitting actual data, it was found that the contents of soil clay, medium silt, coarse silt, and sand were significantly correlated with the contents of fine silt, as shown in the figure. Figure 1 As shown, to simplify the problem, the particle size distribution of soil at each level is converted to fine silt particle size distribution, and the formula is solved as follows:

[0050] .

[0051] (11) By converting the soil particle size components and fine silt components and then inputting them into the marine organic carbon and soil particle size ridge regression model, the prediction of marine organic carbon in the intertidal zone of coastal wetlands based on soil particle size components can be achieved:

[0052] .

[0053] In this embodiment, the correlation diagram between the measured and predicted values ​​of soil marine organic carbon and soil particle size distribution using the ridge regression model shows a significant correlation between the two (p < 0.001). Figure 2 As shown.

[0054] As can be seen from the above description, the embodiments of the present invention, through a large amount of field measurement data, based on the ridge regression model, find the parameters of the correlation between marine organic carbon in coastal wetland soil and soil particle size components at all levels, reduce the deviation between model coefficients and actual values ​​caused by data collinearity of different soil particle size components, and can achieve rapid and accurate prediction of marine organic carbon content in intertidal soil of coastal wetlands, greatly reducing the workload of index measurement.

Claims

1. A rapid prediction method for marine organic carbon content in coastal wetlands based on soil particle size distribution, characterized in that, Includes the following steps: (1) Content of various particle sizes, organic carbon and total nitrogen content, and isotope δ in soil samples. 13 The C value was measured. (2) Based on the ratio of organic carbon to total nitrogen content and the isotope δ¹⁸O of the soil sample 13 C-value is used to calculate the proportion and specific content of organic carbon in the total organic carbon from marine sources. (3) Constructing a ridge regression prediction model for marine organic carbon: The soil particle size data measured in the laboratory were fitted with the marine organic carbon content using ridge regression. The optimal K value was determined based on the ridge trace map, and regression modeling was performed accordingly. The prediction model is as follows: in, This is expressed as the organic carbon content of the marine source; , All are constants. ; Expressed as the content of each particle size component; (4) Fitting the content of each particle size in the soil sample: ； in, All are constants. ; Expressed as the content of clay components in sediments, Expressed as the content of fine silt components in sediments, Expressed as the content of medium-silt components in sediments, Expressed as the content of coarse silt components in sediments, Expressed as the sand component content in sediments; (5) By substituting the fitted data of the particle size content of each level in the soil sample into the prediction model, the prediction of marine organic carbon content in coastal wetlands based on soil particle size composition can be realized.

2. The rapid prediction method for marine organic carbon content in coastal wetlands based on soil particle size distribution according to claim 1, characterized in that, In step (1), before determining the content of each particle size in the soil sample, the soil sample is pretreated as follows: H2O2 and hydrochloric acid are added to the air-dried soil sample in sequence to remove soil organic matter and inorganic salts; after the reaction, the sample is washed repeatedly with deionized water and allowed to stand; finally, sodium hexametaphosphate solution is added and shaken well.

3. The rapid prediction method for marine organic carbon content in coastal wetlands based on soil particle size distribution according to claim 1, characterized in that, In step (1), before determining the organic carbon and total nitrogen content in the soil sample, the soil sample is pretreated as follows: hydrochloric acid is added to the air-dried soil sample; then the sample is repeatedly washed with deionized water until the supernatant is transparent and close to neutral, and then dried for later use.

4. The rapid prediction method for marine organic carbon content in coastal wetlands based on soil particle size distribution according to claim 1, characterized in that, In step (2), a Bayesian three-terminal organic carbon mixture model of terrestrial, marine, and plant sources is constructed using the MixSIAR package in R language. The formula of this model is: ； in, and These are the measured values ​​of the samples, and The values ​​are: (1) the land-source end-member values ​​for each indicator; (2) the land-source end-member values ​​for each indicator. and These are the ocean endmember values ​​for each indicator. and These are the plant endmember values ​​for each indicator. The values ​​represent the percentage contributions of terrestrial, marine, and plant-derived sources to soil organic carbon in the samples, respectively.

Images