Rapid distinguishing method for predicting nitrogen source produced by flooded soil through combination of in-situ observation and model

Abstract

The invention discloses a rapid distinguishing method for predicting a nitrogen source produced by flooded soil through combination of in-situ observation and a model, and relates to the technical field of agro-ecology. Field micro-areas are divided in a target area, an intelligent sensing system is deployed, and soil environment parameters are monitored in real time; the method comprises the following steps: sampling soil in a target area, measuring soil bulk density, root layer depth, total nitrogen content and 15N abundance parameters of a soil sample, and obtaining soil basic data; and calculating the optimal 15N fertilizer abundance according to the recommended nitrogen fertilizer dosage recommended by the target area in combination with the soil nitrogen library content, the fertilizer application amount and the soil background 15N abundance. According to the method, dynamic changes of soil environment parameters are captured through combination of in-situ observation and an intelligent sensing system, the generation rate of nitrogen isotopes is monitored synchronously, the limitation that a traditional method depends on a single parameter is avoided, the distinguishing precision of nitrogen loss sources is remarkably improved, and the contribution rate of the in-situ observation and the intelligent sensing system to nitrogen loss can be rapidly calculated by combining model analysis.

Description

Technical Field

[0001] The present invention relates to the field of agricultural ecological technology, and in particular to a method for quickly distinguishing nitrogen production sources in flooded soil by combining in-situ observation and models. Background Art

[0002] In agricultural ecosystems, rice fields are important food production bases, and their nitrogen cycle and loss mechanisms have a profound impact on environmental quality and agricultural sustainability. Gaseous nitrogen loss is one of the main pathways of nitrogen loss in rice fields, accounting for 40% to 60% of the total nitrogen application. Denitrification and anaerobic ammonia oxidation are key microbial processes that lead to nitrogen loss in the form of nitrogen gas. These processes are particularly active under flooding conditions, causing nitrogen losses of up to 20% to 50% of the applied nitrogen, seriously affecting the nitrogen utilization efficiency and crop yields in rice fields. However, nitrogen transformation in rice field soil is a dynamic process, affected by multiple factors such as temperature, humidity, and microbial activity. In situ field observations can continuously record these changes and capture the critical moments and rates of nitrogen loss.

[0003] Nitrogen in paddy soil mainly comes from the soil nitrogen pool and nitrogen applied to the soil with fertilizers in the current season. The soil nitrogen pool is mostly organic nitrogen, which changes slowly, while the nitrogen applied to the soil with fertilizers is mainly inorganic nitrogen, which decomposes quickly. It is generally believed that nitrogen lost as nitrogen gas mainly comes from fertilizer nitrogen, and the contribution of the soil nitrogen pool is relatively small. However, due to the lack of effective differentiation methods, there is currently a lack of empirical evidence on the contribution of the soil nitrogen pool to nitrogen gas loss. Since it is impossible to effectively distinguish the sources of nitrogen gas produced in paddy soil, there is a lack of accurate targets in finding corresponding methods and technologies for reducing nitrogen gas loss. Therefore, how to combine intelligent sensing systems and in-situ field observations, and conduct model analysis to distinguish the sources of nitrogen gas produced in flooded soils, and improve the efficiency of distinguishing the sources of nitrogen loss, is the problem to be solved by the present invention. To this end, a method for quickly distinguishing the sources of nitrogen gas produced in flooded soils by combining in-situ observations and models is proposed. Summary of the Invention

[0004] The present invention aims to provide a method for quickly distinguishing the sources of nitrogen production in flooded soil by combining in-situ observation and model prediction, so as to solve the problems raised in the above-mentioned background technology.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A method for rapidly distinguishing nitrogen sources in flooded soils by combining in-situ observations and models, comprising the following steps:

[0007] S1. Divide the field into micro-areas in the target area and deploy an intelligent sensing system to monitor soil environmental parameters in real time;

[0008] S2. Take soil samples in the target area and measure the soil bulk density, root layer depth, total nitrogen content and 15 N abundance parameters, to obtain basic soil data;

[0009] S3. Recommended nitrogen fertilizer dosage for the target area, combined with soil nitrogen pool content, fertilizer application rate and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance;

[0010] S4. Soak the soil and apply 15 N fertilizers and use the deployed smart sensing system to monitor changes in soil environmental parameters, including temperature, moisture, and pH;

[0011] S5. Use a sampler to continuously collect water samples in situ in the field, and use a mass spectrometer to determine the water samples, including 28 N2, 29 N2 and 30 Nitrogen isotope content of N2;

[0012] S6. Combine field observation data and intelligent sensing system data to build a nitrogen production model and optimize model parameters to improve prediction accuracy;

[0013] S7. Based on the output of the nitrogen production model, calculate the contribution of soil nitrogen pools and fertilizer nitrogen to nitrogen loss, predict the source of nitrogen production in flooded soils, and quickly identify the source of nitrogen loss.

[0014] A further improvement of the technical solution of the present invention is that: S1 includes:

[0015] Select an observation area within the target rice paddy field and divide it into several micro-areas of at least 0.15 square meters based on the terrain and soil uniformity. Use stainless steel or plastic barriers to isolate the micro-areas, implanted at a depth of at least 0.4 meters in the soil to ensure isolation from soil and water, and to ensure the independence and stability of the micro-areas.

[0016] An integrated intelligent sensing system is deployed in each micro-zone, including temperature, humidity, pH, and redox potential sensors, ensuring that the sensor probes are evenly distributed within the root layer depth range and transmitting data in real time to the cloud platform via a wireless transmission module;

[0017] After the deployment of the intelligent sensor system is completed, a 12-hour continuous pre-operation test is carried out to verify the stability of data transmission and the accuracy of parameters. After confirming that the system is normal, all-weather real-time monitoring is started, and the soil environment dynamic data is recorded every 10 minutes and stored synchronously in the database of the cloud platform.

[0018] A further improvement of the technical solution of the present invention is that: S2 includes:

[0019] Based on the topography, soil type and planting layout of the target area, sampling points are evenly planned to ensure coverage of different areas. Based on the distribution of crop roots, the root layer depth range is determined and the sampling layers are located.

[0020] Use standard sampling tools to collect soil at planned locations and depths, remove surface debris, and place soil samples in clean, sealed containers to avoid contamination and composition changes, ensuring that the samples truly reflect the soil conditions.

[0021] The collected soil samples were sent to the laboratory to measure soil bulk density, root layer depth, total nitrogen content and 15 N abundance, obtain soil basic data, organize the measurement results, and establish a soil basic database.

[0022] A further improvement of the technical solution of the present invention is that: S3 includes:

[0023] Check the recommended nitrogen fertilizer application rate for the target area, measure the soil nitrogen pool content (total nitrogen content) and soil background 15 N abundance data;

[0024] Based on the obtained recommended nitrogen fertilizer application rate, total soil nitrogen content, soil bulk density, root layer depth and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance to ensure the scientific and effective application of fertilizers;

[0025] The calculation results were preliminarily verified and the optimal 15 N fertilizer abundance and recommended nitrogen fertilizer application rates should be adjusted.

[0026] A further improvement of the technical solution of the present invention is that: the optimal 15 The calculation process of N fertilizer abundance is:

[0027] Combined with the total soil nitrogen content, soil bulk density and root depth, the nitrogen reserves of the soil nitrogen pool within the target depth range were calculated. 15 N abundance, analyzing the isotopic characteristics of nitrogen in the soil nitrogen pool and clarifying the contribution of the soil nitrogen pool;

[0028] Based on the recommended nitrogen fertilizer application rate and the contribution of the soil nitrogen pool, determine the actual amount of nitrogen that needs to be supplemented through fertilizer;

[0029] Combined with the recommended nitrogen fertilizer application rate, total soil nitrogen content, soil bulk density, root depth and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance, where too high 15 Although N abundance can improve the tracking accuracy of fertilizer nitrogen, it will increase the cost of fertilizer. Too low an abundance will cause the signal of fertilizer nitrogen in the soil to be masked by the background nitrogen in the soil, affecting the tracking effect.

[0030] A further improvement of the technical solution of the present invention is that: S4 includes:

[0031] In the micro-area of ​​the target area, the top 0-20 cm of soil is plowed, and enough water is added to cover the soil surface by 3-5 cm. Soak for 5-10 days to simulate flooding conditions and ensure that the soil reaches a flooded state;

[0032] According to the calculated optimal 15 N fertilizer abundance and recommended nitrogen fertilizer application rate, 24 hours in advance 15 Apply nitrogen fertilizer evenly to the soil surface and mix thoroughly to ensure that the labeled nitrogen is in full contact with the soil and to initiate the nitrogen conversion process;

[0033] Start the deployed intelligent sensing system to monitor the changes of soil environmental parameters in real time, including temperature, humidity and pH value. The sensor collects data every 10 minutes through the wireless module and transmits it to the cloud platform in real time, which then automatically integrates the sensor data with the 15 N fertilizer application time node, establish a time series database.

[0034] A further improvement of the technical solution of the present invention is that: S5 includes:

[0035] Use a sampler to continuously collect water samples in the field, ensuring that the water samples do not come into contact with air during the collection process to avoid external interference. Each collection should be done every 2 hours, and the samples should be collected continuously for at least 3 times. The water samples should be sealed and stored and sent to the laboratory;

[0036] The collected water samples are transferred to a sealed gas extraction device, and the dissolved N2 gas is separated by low-temperature distillation. The purified gas is injected into an isotope mass spectrometer to quantitatively determine the water sample, including 28 N2, 29 N2 and 30 The nitrogen isotope content of N2 is measured, and based on the measurement results and the sampling time interval, the production rate of each nitrogen isotope per unit time is calculated.

[0037] A further improvement of the technical solution of the present invention is that: S6 includes:

[0038] The nitrogen isotope production rate observed in the field and the soil temperature, humidity, and pH values ​​monitored in real time by the intelligent sensing system were integrated and simultaneously imported into the database. The data were then cleaned (outliers removed), standardized (Z-score normalization), and time-aligned to ensure consistent spatiotemporal resolution across parameters, thus constructing a basic data set.

[0039] Based on the theory of microbial-driven nitrogen cycle, a nitrogen production model including denitrification and anaerobic ammonium oxidation was established. The soil environmental parameters were used as driving variables and the nitrogen isotope production rate was used as the response variable to construct a nonlinear relationship equation.

[0040] The Bayesian optimization algorithm is used, with the root mean square error between the predicted value and the measured value as the objective function, to dynamically adjust the model parameters. The model generalization ability is evaluated through cross-validation, and finally a high-precision nitrogen production model and the optimal parameter solution are output.

[0041] A further improvement of the technical solution of the present invention is that: S7 includes:

[0042] The nitrogen isotope production rate data were obtained from the nitrogen production model to calculate the total N2 production rate of paddy soil and analyze the 15 The N2 production rate of N fertilizers, and then based on the total N2 production rate of paddy soil, 15 N2 production rate of N fertilizers and fertilizers 15 N abundance, to obtain the N2 contribution from fertilizer nitrogen;

[0043] The N2 production rate of each sampling during the rice growth period was counted, and the total nitrogen loss and fertilizer nitrogen contribution rate during the entire growth period were calculated;

[0044] Based on the calculated contribution rate and total nitrogen loss, the nitrogen production sources of flooded soils are predicted, and the contribution ratios of soil nitrogen pools and fertilizer nitrogen to nitrogen loss are quickly distinguished, providing a scientific basis for precise fertilization and nitrogen management.

[0045] A further improvement of the technical solution of the present invention is that the calculation process of the total nitrogen loss and the fertilizer nitrogen contribution rate during the entire growth period is:

[0046] According to the characteristics of the rice growth period, a sampling plan was formulated to collect water samples every 1-2 days after fertilization for 3-6 consecutive times. In the early stage of non-fertilization, water samples were collected every 5-10 days until the rice was harvested. No sampling was done during the rice drying period.

[0047] Each time sampling is performed, the sampling point number, sampling time, soil environmental parameters, and nitrogen isotope production rate are recorded in detail, and all sampling data are organized into a unified database to ensure data integrity and consistency;

[0048] Within the sampling times n during the entire growth period, calculate the total N2 production rate of the nth sampling, multiply the total N2 production rate of the nth sampling by the time interval between adjacent sampling times, and sum them up to obtain the total nitrogen loss during the entire growth period. The time interval between adjacent sampling times is the time interval between the nth sampling and the n-1th sampling. Within the sampling times n during the entire growth period, calculate the total N2 production rate of the nth sampling during the entire growth period. 15 N2 production rate of N fertilizers;

[0049] Based on the total nitrogen loss during the entire growth period, 15 N2 production rate of N fertilizers and fertilizers15 N abundance, calculate the fertilizer nitrogen contribution rate during the entire growth period.

[0050] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art:

[0051] 1. The present invention provides a method for quickly distinguishing the sources of nitrogen production in flooded soil by combining in-situ observations and models. By combining in-situ observations with intelligent sensing systems, the dynamic changes of soil environmental parameters can be captured in real time, and the production rate of nitrogen isotopes can be simultaneously monitored. This avoids the limitations of traditional methods that rely on a single parameter, significantly improves the accuracy of distinguishing the sources of nitrogen loss, and combined with model analysis, the contribution rate of the two to nitrogen loss can be quickly calculated.

[0052] 2. The present invention provides a method for quickly distinguishing the sources of nitrogen production in flooded soil by combining in-situ observation and model. 15 N-labeled fertilizers and optimal application rate calculations can accurately track the transformation pathway of fertilizer nitrogen in the soil. Combined with nitrogen production models, they can quantify the proportion of fertilizer nitrogen lost due to denitrification or anaerobic ammonia oxidation, helping to adjust the amount and timing of fertilizer application, avoiding excessive fertilization, and reducing the risk of agricultural non-point source pollution while saving production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0054] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0055] Figure 2 Schematic diagram of the method of the present invention. DETAILED DESCRIPTION

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0057] Example 1, as Figure 1 、 Figure 2 As shown, the present invention provides a method for quickly distinguishing the sources of nitrogen production in flooded soil by combining in-situ observation and model prediction, comprising the following steps:

[0058] S1. Divide the target area into field micro-zones and deploy an intelligent sensing system to monitor soil environmental parameters in real time. Select an observation range within the target rice field area and divide the area into several micro-zones of at least 0.15 square meters based on the terrain and soil uniformity. Use stainless steel or plastic barriers to isolate the micro-zones, implanted at a depth of at least 0.4 meters to ensure isolation between soil and water, and to ensure the independence and stability of the micro-zones. Deploy an integrated intelligent sensing system in each micro-zone, including temperature, humidity, pH, and redox potential sensors, ensuring that the sensor probes are evenly distributed within the root layer depth range. Data is transmitted back to the cloud platform in real time via a wireless transmission module. After the intelligent sensor system is deployed, conduct a 12-hour continuous pre-operation test to verify data transmission stability and parameter accuracy. Once the system is confirmed to be normal, initiate 24 / 7 real-time monitoring, set to record soil environmental dynamic data every 10 minutes, and synchronize the data to the cloud platform's database.

[0059] The specific process of S1 is as follows:

[0060] According to the topographic characteristics, soil type distribution and irrigation system layout of the target rice field area, a representative observation range is comprehensively selected to ensure that the observation results can reflect the nitrogen cycle characteristics of the entire rice field. According to the topographic undulation and soil uniformity, the observation range is divided into several micro-areas with an area of ​​not less than 0.15 square meters. Stainless steel or plastic enclosure materials are used to ensure that the enclosure is embedded in the soil to a depth of not less than 0.4 meters, effectively isolating the soil and water exchange between each micro-area to ensure the independence and stability of the micro-area; in each micro-area, an integrated intelligent sensing system is arranged, including temperature sensor, humidity sensor, pH sensor and redox potential sensor to ensure the sensing The sensor probes are evenly distributed within the root layer depth range to accurately reflect the rhizosphere soil environment. The intelligent sensing system is connected to the cloud platform via a wireless transmission module to achieve real-time data transmission. After the deployment of the intelligent sensor system, a 12-hour continuous pre-operation test is carried out. During this period, the stability of data transmission and the accuracy of parameter measurement are focused on verifying the reliability of the sensors in the complex field environment. After confirming the normal operation of the system, the all-weather real-time monitoring mode is activated, and the dynamic data of the soil environment is recorded every 10 minutes, including key parameters such as temperature, humidity, pH value and redox potential. All data are synchronously stored in the database of the cloud platform.

[0061] S2. Take soil samples in the target area and measure the soil bulk density, root layer depth, total nitrogen content and 15N abundance parameters, obtain basic soil data, plan sampling points evenly based on the topography, soil type and planting layout of the target area to ensure coverage of different areas, determine the root depth range based on the distribution of crop roots, locate the sampling layer, use standard sampling tools, collect soil at the planned points and depths, remove surface debris, and place soil samples in clean and sealed containers to avoid contamination and composition changes, ensuring that the samples truly reflect the soil conditions, and send the collected soil samples to the laboratory to measure soil bulk density, root depth, total nitrogen content and 15 N abundance, obtain basic soil data, organize the measurement results, and establish a basic soil database;

[0062] The specific process of S2 is:

[0063] Collect topographic maps, soil type distribution maps and planting layout maps of the target rice field area, identify the main topographic units, soil types and crop planting patterns in the target rice field area, determine the root depth range (0-15cm, 15-30cm, etc.) according to the distribution characteristics of crop roots (rice roots are mainly distributed in the 0-20cm soil layer), and divide the sampling layers. Use the grid method to evenly arrange sampling points in the target area to ensure coverage of different topography, soil types and planting areas. The number of sampling points is determined according to the area of ​​the area, and the geographical location (latitude and longitude), topographic characteristics, soil type and crop growth status of each sampling point are recorded; use a standard soil sampler (stainless steel ring knife) to ensure that the tool is clean and pollution-free, and equip it with auxiliary tools such as brushes, shovels, label paper, sealed bags or clean glass bottles to arrive at the sampling site. After sampling, remove surface debris (dead branches, leaves, stones, etc.) to avoid contamination. According to the determined root depth range, carry out 0-15cm and 15-30cm layered sampling. When sampling each layer, insert the sampler vertically into the soil layer and take out the complete soil core. When using a ring knife for sampling, ensure that the soil in the ring knife is dense and has no gaps. Put the soil sample into a clean and sealed container to avoid direct sunlight and high temperature to prevent composition changes. Label each sample with the sampling point number, depth, date and sampler information; bring the collected soil samples back to the laboratory for measurement. If they need to be preserved, they can be placed in a 4℃ refrigerator for short-term storage. Before measurement, sieve the soil sample with a 2mm sieve to remove stones and plant residues, mix them evenly and then divide them into groups for use. Then measure the basic soil parameters, including soil bulk density, root depth, total nitrogen content and 15 The soil bulk density was determined by the ring knife method to calculate the dry weight of soil per unit volume. The root layer depth was combined with on-site sampling records and laboratory observations to confirm the actual distribution depth of the root system. The total nitrogen content of the soil was determined by an element analyzer, and the soil nitrogen content was determined by an isotope mass spectrometer. 15N abundance; enter the measurement results into a spreadsheet and store them by sampling point number, depth, and parameter type. Establish a basic soil database, associate it with geographical location, topography, soil type, and other information, and perform quality control on the data, check for outliers, and retest to ensure data accuracy;

[0064] S3. Recommended nitrogen fertilizer dosage for the target area, combined with soil nitrogen pool content, fertilizer application rate and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance, check the recommended nitrogen fertilizer application rate for the target area recommended by local agricultural production, and measure the soil nitrogen pool content (total nitrogen content) and soil background 15 N abundance data, based on the recommended nitrogen fertilizer application rate, total soil nitrogen content, soil bulk density, root depth, and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance, ensure the scientific and effective application of fertilizers, conduct preliminary verification of the calculation results, and optimize 15 N fertilizer abundance and recommended nitrogen fertilizer application rates are adjusted;

[0065] In addition, the best 15 The calculation process of N fertilizer abundance is:

[0066] Combined with the total soil nitrogen content, soil bulk density and root depth, the nitrogen reserves of the soil nitrogen pool within the target depth range are calculated. The soil nitrogen pool is an important source of nitrogen supply for crops, and its reserve size directly affects the fertilizer application strategy. By analyzing the reserves of the soil nitrogen pool, we can preliminarily judge the degree to which the soil itself meets the nitrogen demand of crops, thereby providing a basis for determining the amount of fertilizer nitrogen supplementation. At the same time, combined with the soil background 15 N abundance, analyze the isotopic characteristics of nitrogen in the soil nitrogen pool, clarify the contribution of the soil nitrogen pool, and determine the actual amount of nitrogen that needs to be supplemented through fertilizer based on the recommended application rate of nitrogen fertilizer and the contribution of the soil nitrogen pool. Among them, the recommended application rate of nitrogen fertilizer is a recommended value based on the comprehensive consideration of crop growth needs and soil fertility conditions. When actually fertilizing, the actual contribution of the soil nitrogen pool must be considered to avoid excessive fertilization. Combined with the recommended application rate of nitrogen fertilizer, total soil nitrogen content, soil bulk density, root layer depth and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance, where too high 15 Although nitrogen abundance can improve the tracking accuracy of fertilizer nitrogen, it will increase fertilizer costs. However, too low an abundance will cause the signal of fertilizer nitrogen in the soil to be masked by soil background nitrogen, affecting the tracking effect.

[0067] 15 The calculation formula for N fertilizer abundance is as follows:

[0068]

[0069] Where X is 15 N fertilizer abundance, p is soil bulk density, h is root layer depth, a is total soil nitrogen, b is soil background 15 N abundance, c is the recommended nitrogen fertilizer application rate;

[0070] S4. Soak the soil and apply 15 N fertilizer, and use the deployed intelligent sensor system to monitor changes in soil environmental parameters, including temperature, humidity, and pH. In the micro-area of ​​the target area, the top 0-20cm of soil is plowed, and enough water is added to cover the soil surface by 3-5cm. Soil is soaked for 5-10 days to simulate flooding conditions, ensuring that the soil reaches a flooded state. According to the calculated optimal 15 N fertilizer abundance and recommended nitrogen fertilizer application rate, 24 hours in advance 15 N fertilizer is evenly applied to the soil surface and mixed to ensure that the labeled nitrogen is fully in contact with the soil, starting the nitrogen conversion process and activating the deployed intelligent sensing system to monitor changes in soil environmental parameters in real time, including temperature, humidity and pH value. The sensor collects data every 10 minutes through the wireless module and transmits it to the cloud platform in real time, thereby automatically integrating the sensor data with the soil. 15 N fertilizer application time node, establish a time series database;

[0071] The specific process of S4 is as follows:

[0072] In the micro-area of ​​the target area, the surface 0-20cm soil is plowed to ensure that the soil is loose, and enough water is added to make the water cover the soil surface 3-5cm to simulate flooding conditions. The soaking time is 5-10 days to ensure that the soil reaches a flooded state and create suitable environmental conditions for the nitrogen conversion process; according to the previously calculated optimal 15 N fertilizer abundance and recommended nitrogen fertilizer application rate, prepare the corresponding 15 N fertilizer, 24 hours in advance 15 N fertilizer should be evenly applied to the soil surface. It can be applied manually or mechanically to ensure that the fertilizer is evenly distributed. After applying the fertilizer, use tools to fully mix the fertilizer with the surface soil to make the marked nitrogen fully contact with the soil and start the nitrogen conversion process. 15 Before applying fertilizer, check the operating status of the intelligent sensing system to ensure that the sensor is working properly and the data transmission module is fault-free. 15 At the same time as applying N fertilizer, the real-time monitoring function of the intelligent sensor system is activated to ensure that changes in soil environmental parameters are recorded from the moment of fertilization. The sensor collects data every 10 minutes through the wireless module, including key parameters such as soil temperature, humidity and pH value. The collected data is transmitted to the cloud platform in real time to ensure the timeliness and integrity of the data. On the cloud platform, the sensor data is automatically integrated with the 15N fertilizer application time node, establish a time series database, and clearly record the dynamic changes of soil environmental parameters after fertilization;

[0073] S5. Use a sampler to continuously collect water samples in situ in the field, and use a mass spectrometer to determine the water samples, including 28 N2, 29 N2 and 30 The nitrogen isotope content of N2 is determined by continuously collecting water samples in situ in the field using a sampler. Ensure that the water samples do not come into contact with the air during the collection process to avoid external interference. Each collection is conducted every 2 hours and the collection is conducted continuously for at least 3 times. The water samples are sealed and stored and sent to the laboratory. The collected water samples are transferred to a sealed gas phase extraction device, and the dissolved N2 gas is separated by low-temperature distillation. The purified gas is injected into the isotope mass spectrometer to quantitatively determine the nitrogen isotope content in the water samples, including 28 N2, 29 N2 and 30 The nitrogen isotope content of N2, and based on the measurement results and the sampling time interval, calculate the production rate of each nitrogen isotope per unit time;

[0074] The specific process of S5 is as follows:

[0075] Use a sampler based on patent number 201911416804.9, ensure that it is clean and free of any contamination, check the sealing and function of the sampler, ensure that it can work normally underwater and is leak-free, select a sampling point in the micro-area of ​​the target rice field, and insert the sampler into the water layer of the rice field, ensure that the sampler is completely immersed in water, and the water inlet of the sampler is located in the middle of the water layer to avoid contact with the bottom sediment, start the sampler, and extract the water sample from the rice field into the sealed container of the sampler, ensure that the water sample does not come into contact with the air during the sampling process, and avoid external interference. Each collection interval is 2 hours, and collection is carried out continuously for at least 3 times. After the collection is completed, transfer the water sample from the sampler to a sealed glass bottle. Ensure that the water sample does not come into contact with the air during the transfer process, and place the sealed water sample in an insulated box to avoid direct sunlight and high temperature to prevent changes in the gas composition of the water sample, and label the water sample container with the sampling point number, sampling time, Depth and other information; After the collected water samples are brought back from the field to the laboratory, they should be analyzed as soon as possible. If they cannot be analyzed immediately, the water samples should be stored in a 4°C refrigerator, but the storage time should not be too long to avoid the impact of microbial activity in the water samples on the nitrogen isotope content. The water samples are transferred to a sealed gas extraction device. The gas extraction device has a low-temperature distillation function and can effectively separate the dissolved N2 gas. In the gas extraction device, the dissolved N2 gas in the water sample is separated by low-temperature distillation. The distillation temperature and pressure are controlled to ensure that the N2 gas can be completely separated and not interfered with by other gases. The separated N2 gas is passed through a gas purification device to remove the existing impurity gases (O2, CO2, etc.) to ensure that the gas injected into the isotope mass spectrometer is high-purity N2; the purified N2 gas is injected into the isotope mass spectrometer. During the injection process, the gas flow and pressure must be ensured to be stable. The isotope mass spectrometer is used to quantitatively determine the nitrogen isotope content in the water sample. 28 N2, 29 N2 and 30 The nitrogen isotope content of N2, record the abundance value of each nitrogen isotope; based on the measured nitrogen isotope content and the sampling time interval (2 hours), calculate the production rate of each nitrogen isotope per unit time, and record the calculated production rate of each nitrogen isotope in the data table;

[0076] S6. Combine field observation data and intelligent sensing system data to build a nitrogen production model and optimize model parameters to improve prediction accuracy;

[0077] S7. Based on the output of the nitrogen production model, calculate the contribution of soil nitrogen pools and fertilizer nitrogen to nitrogen loss, predict the source of nitrogen production in flooded soils, and quickly identify the source of nitrogen loss.

[0078] Example 2, as Figure 1 、 Figure 2 As shown, based on Example 1, the present invention provides a technical solution: preferably, S6 includes:

[0079] The nitrogen isotope production rate observed in the field and the soil temperature, humidity, and pH values ​​monitored in real time by the intelligent sensing system were integrated and simultaneously imported into the database. The data were cleaned (outliers were removed), standardized (Z-score normalization), and time-aligned to ensure consistent spatiotemporal resolution of different parameters. A basic data set was constructed. Based on the theory of the microbial-driven nitrogen cycle, a nitrogen production model that includes denitrification and anaerobic ammonia oxidation was established. Soil environmental parameters were used as driving variables and nitrogen isotope production rate as the response variable. A nonlinear relationship equation was constructed. A Bayesian optimization algorithm was used, with the root mean square error between the predicted and measured values ​​as the objective function. The model parameters were dynamically adjusted, and the model generalization ability was evaluated through cross-validation. Finally, a high-precision nitrogen production model and optimal parameter solution were output.

[0080] The specific process of S6 is as follows:

[0081] The nitrogen isotope production rate observed in the field and the soil temperature, humidity and pH value monitored by the intelligent sensing system were uniformly imported into the database. The nitrogen isotope rate data were time-aligned with the soil environmental parameters through timestamp matching. Missing values ​​were filled by linear interpolation to ensure that the time resolution of all parameters was unified to 10 minutes / time. Spatially, the coordinates of the nitrogen isotope sampling points were mapped to the same grid (resolution 1m×1m) using the inverse distance weighted method based on the position of the intelligent sensing node to eliminate spatial scale differences. Based on the 3σ principle, data points in the nitrogen isotope rate that exceeded ±3 times the standard deviation of the mean were removed. For soil temperature, humidity and pH value, the box plot method was used to identify and remove outliers. The cleaned data were Z-score normalized according to the parameter type to eliminate the dimension effect. The preprocessed data were then divided into training set, validation set and test set according to the time series to ensure that each data set covered the complete experimental period and environmental conditions, generating a dataset containing input variables (soil temperature, humidity, pH value) and output variables ( 28 N2, 29 N2 and 30The standardized data set of N2 production rate was used as the basis for model training; based on the microbial-driven nitrogen cycle theory, a dual-path nitrogen production model including denitrification and anaerobic ammonium oxidation was established to determine the dual-path nitrogen production rate, and a nonlinear relationship equation including the denitrification rate equation and the anaerobic ammonium oxidation rate equation was constructed. Among them, the denitrification rate equation was a Monod model modified by the Arrhenius equation with soil temperature, humidity and pH as driving variables, and the anaerobic ammonium oxidation rate equation was an enzyme activity model combined with pH dependence. Then, according to literature analysis and preliminary experimental results, the model parameter range was initialized; the Bayeux method was used to calculate the nitrogen production rate of the two pathways. The Gaussian optimization algorithm is used for parameter tuning. The root mean square error between the predicted value and the measured value is used as the objective function. The optimal solution is dynamically searched in the parameter space. The number of iterations is set to 100. Each iteration updates the parameter probability distribution based on Gaussian process regression. K-fold cross-validation is used to randomly divide the training set into 5 folds. 4 folds are used for training and 1 fold for validation each time. After 5 cycles, the average root mean square error is taken as the model performance indicator. The optimized model is run on the test set to calculate the mean absolute error. At the same time, the preset mean absolute error threshold is used to ensure the model's prediction accuracy for unknown data, thereby outputting a parameter-optimal and high-precision nitrogen production model.

[0082] The expression of the dual-path nitrogen generation model is as follows:

[0083] R total =R Den +R Ann ;

[0084] Where R total is the dual-path nitrogen generation rate, R Den and R Ann are the nitrogen production rates of denitrification and anaerobic ammonium oxidation, respectively;

[0085] The expression of the denitrification rate equation is as follows:

[0086]

[0087] Where k Den is the maximum reaction rate constant of denitrification, E a is the activation energy, P is the gas constant, K w and K P is the half-saturation constant of humidity and pH value, T is soil temperature, W is humidity, and P is pH value;

[0088] The expression of the anaerobic ammonium oxidation rate equation is as follows:

[0089]

[0090] Where k Annis the maximum reaction rate constant of anaerobic ammonium oxidation, K P1 and K P2 is the pH inhibition constant, T min and T opt is the minimum and optimum temperature;

[0091] The expression of the root mean square error is as follows:

[0092]

[0093] Where RMSE is the root mean square error, m is the number of samples, which means the total number of data points used to calculate the root mean square error, and R pred,j is the model’s prediction result for the jth sample, R obs,j is the actual observation value of the jth sample;

[0094] The expression of mean absolute error is as follows:

[0095]

[0096] Where, MAE is the mean absolute error;

[0097] S7 includes:

[0098] The nitrogen isotope production rate data were obtained from the nitrogen production model to calculate the total N2 production rate of paddy soil and analyze the 15 The N2 production rate of N fertilizers, and then based on the total N2 production rate of paddy soil, 15 N2 production rate of N fertilizers and fertilizers 15 N abundance, obtain the N2 contribution from fertilizer nitrogen, for each sampling, calculate 28 N2, 29 N2 and 30 The sum of the N2 production rates is used to obtain the total N2 production rate in paddy soil and calculate 29 The N2 production rate is divided by 2, and 30 The N2 generation rate is added to obtain the 15 The N2 production rate of N fertilizers is due to 30 N2 is composed of two 15 N atoms, so 30 The production rate of N2 mainly comes from 15 N fertilizer, 29 There is also a 15 N atoms, statistics of N2 production rates at each sampling time during the rice growth period, calculation of total nitrogen loss and fertilizer nitrogen contribution rate throughout the growth period, prediction of nitrogen production sources in flooded soils based on the calculated contribution rate and total nitrogen loss, rapid differentiation of the contribution ratios of soil nitrogen pools and fertilizer nitrogen to nitrogen loss, and providing a scientific basis for precision fertilization and nitrogen management;

[0099] The specific process of S7 is as follows:

[0100] Obtained from the nitrogen generation model 28 N2, 29 N2 and 30 The N2 production rate data was used to calculate the total N2 production rate of paddy soil based on the nitrogen isotope production rate output by the model, and the N2 production rate from the paddy soil was calculated based on the model output. 15 The N2 production rate of N fertilizer is based on the total N2 production rate of paddy soil, 15 N2 production rate of N fertilizers and fertilizers 15 N abundance, calculate the contribution rate of fertilizer nitrogen; according to the field sampling plan, collect water samples several times during the rice growth period, record the N2 production rate of each sampling, generally collect once every 1-2 days after fertilization, and collect 3-6 times in a row; collect once every 5-10 days in the early stage of non-fertilization until rice harvest, and do not sample during the rice drying period, count the N2 production rate of each sampling during the entire growth period, calculate the total nitrogen loss during the entire growth period, and calculate the total nitrogen loss during the entire growth period based on the total nitrogen loss during the entire growth period and the nitrogen loss from 15 The N2 production rate of N fertilizers is used to calculate the fertilizer nitrogen contribution rate throughout the growing season. Based on the calculated contribution rate and total nitrogen loss, a nitrogen production model is used to predict the source of nitrogen production in flooded soils. This allows for rapid differentiation of the contribution of soil nitrogen pools and fertilizer nitrogen to nitrogen loss. By comparing nitrogen production rates from different sources, the primary source of nitrogen loss can be identified.

[0101] In addition, the calculation process of the total nitrogen loss and fertilizer nitrogen contribution rate during the entire growth period is as follows:

[0102] According to the characteristics of the rice growing period, a sampling plan was formulated. Water samples were collected every 1-2 days after fertilization, for 3-6 consecutive times. Water samples were collected every 5-10 days in the early non-fertilization period until the rice was harvested. No sampling was done during the rice drying period. Representative sampling points were selected in different micro-areas of the target rice fields to ensure coverage of different terrains, soil types and planting areas. The sampling point number, sampling time, soil environmental parameters and nitrogen isotope production rate were recorded in detail for each sampling. All sampling data were organized into a unified database to ensure data integrity and consistency. Within the number of sampling times n during the entire growing period, the total N2 production rate of the nth sampling was calculated, and the total N2 production rate of the nth sampling was multiplied by the time interval between adjacent sampling times, and the sum was taken to obtain the total nitrogen loss during the entire growing period. Among them, the time interval between adjacent sampling times was the time interval between the nth sampling and the n-1th sampling. Within the number of sampling times n during the entire growing period, the total N2 production rate of the nth sampling was calculated 15 The N2 production rate of N fertilizers is based on the total nitrogen losses during the entire growing period, the 15N2 production rate of N fertilizers and fertilizers 15 N abundance, calculating the fertilizer nitrogen contribution rate during the entire growth period;

[0103] The calculation formula for the total N2 production rate in paddy soil is as follows:

[0104] V n =V 28，n +V 29，n +V 30，n ;

[0105] Where V n is the total N2 production rate of paddy soil sampled at the nth time, V 28，n 、V 29，n 、V 30，n The nth sampling time 28 N2, 29 N2 and 30 N2 production rate;

[0106] From 15 The formula for calculating the N2 production rate of N fertilizer is as follows:

[0107]

[0108] Where, For those from 15 N2 production rate of N fertilizers;

[0109] The calculation formula for the total nitrogen loss during the rice growth period is as follows:

[0110]

[0111] Where M is the total nitrogen loss during the rice growth period, n is the number of sampling times during the entire growth period, and V sn is the total N2 generation rate at the nth sampling time, Δt is the time interval between the nth sampling time and the n-1th sampling time;

[0112] The formula for calculating the total N2 production rate from 15N fertilizer during the rice growing period is as follows:

[0113]

[0114] Where M - The total growth period of rice comes from 15 N2 production rate of N fertilizer, V - s n is the value from the nth sampling 15 N2 production rate of N fertilizers;

[0115] The calculation formula for the total N2 contribution rate from fertilizer nitrogen during the rice growing period is as follows:

[0116]

[0117] Where D f is the total N2 contribution rate from fertilizer nitrogen during the rice growth period, X is 15 N fertilizer abundance.

[0118] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for rapidly distinguishing nitrogen sources in flooded soils by combining in-situ observation and model prediction, characterized in that: The following steps are involved: S1. Divide the field into micro-areas in the target area and deploy an intelligent sensing system to monitor soil environmental parameters in real time; S2. Take soil samples in the target area and measure the soil bulk density, root layer depth, total nitrogen content and 15 N abundance parameters, to obtain basic soil data; S3. Recommended nitrogen fertilizer dosage for the target area, combined with soil nitrogen pool content, fertilizer application rate and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance; S4. Soak the soil and apply 15 N fertilizers and monitor changes in soil environmental parameters using the deployed smart sensing system; S5. Continuously collect water samples in situ in the field using a sampler and determine the nitrogen isotope content in the water samples using a mass spectrometer; S6. Combining field observation data and intelligent sensing system data to build a nitrogen production model; S7. Based on the output of the nitrogen production model, calculate the contribution of soil nitrogen pools and fertilizer nitrogen to nitrogen loss, predict the source of nitrogen production in flooded soils, and quickly identify the source of nitrogen loss.

2. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and modeling according to claim 1, characterized in that: Said S1 comprises: Select an observation area in the target rice field and divide it into several micro-areas of no less than 0.15 square meters based on the terrain and soil uniformity. Use stainless steel or plastic barriers to isolate the micro-areas and plant them in the soil to a depth of no less than 0.4 meters. An integrated intelligent sensing system is deployed in each micro-area, including temperature, humidity, pH value and redox potential sensors, and the data is transmitted back to the cloud platform in real time via a wireless transmission module; After the deployment of the intelligent sensor system is completed, a 12-hour continuous pre-operation test is carried out. After confirming that the system is normal, all-weather real-time monitoring is started, and the soil environment dynamic data is recorded every 10 minutes and stored synchronously in the database of the cloud platform.

3. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and modeling according to claim 1, characterized in that: The S2 includes: Based on the topography, soil type and planting layout of the target area, the sampling points are evenly planned. According to the distribution of crop roots, the root depth range is determined and the sampling layers are located. Use standard sampling tools to collect soil at planned locations and depths, and place the soil samples into clean, sealed containers; The collected soil samples were sent to the laboratory to measure soil bulk density, root layer depth, total nitrogen content and 15 N abundance, obtain soil basic data, organize the measurement results, and establish a soil basic database.

4. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and modeling according to claim 1, characterized in that: The S3 includes: Check the recommended nitrogen fertilizer application rate for the target area and measure the soil nitrogen pool content and soil background 15 N abundance data; Based on the obtained recommended nitrogen fertilizer application rate, total soil nitrogen content, soil bulk density, root layer depth and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance; The calculation results were preliminarily verified and the optimal 15 N fertilizer abundance and recommended nitrogen fertilizer application rates should be adjusted.

5. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and model prediction according to claim 4, characterized in that: The optimal 15 The calculation process of N fertilizer abundance is: Combined with the total soil nitrogen content, soil bulk density and root depth, the nitrogen reserves of the soil nitrogen pool within the target depth range were calculated. 15 N abundance, analyzing the isotopic characteristics of nitrogen in the soil nitrogen pool and clarifying the contribution of the soil nitrogen pool; Based on the recommended nitrogen fertilizer application rate and the contribution of the soil nitrogen pool, determine the actual amount of nitrogen that needs to be supplemented through fertilizer; Combined with the recommended nitrogen fertilizer application rate, total soil nitrogen content, soil bulk density, root depth and soil background 15 N abundance, calculate optimal 15 N fertilizer abundance.

6. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and modeling according to claim 1, characterized in that: The S4 includes: In the micro-area of ​​the target area, the top 0-20 cm of soil is plowed, water is added to cover the soil surface by 3-5 cm, and soaked for 5-10 days to simulate flooding conditions; According to the calculated optimal 15 N fertilizer abundance and recommended nitrogen fertilizer application rate, 24 hours in advance 15 N fertilizer is applied evenly to the soil surface and mixed thoroughly to start the nitrogen conversion process; Start the deployed intelligent sensing system to monitor the changes of soil environmental parameters in real time, including temperature, humidity and pH value. The sensor collects data every 10 minutes through the wireless module and transmits it to the cloud platform in real time, which then automatically integrates the sensor data with the 15 N fertilizer application time node, establish a time series database.

7. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and modeling according to claim 1, characterized in that: The S5 includes: Water samples were collected continuously in the field using a sampler, with an interval of 2 hours between each collection, for at least 3 times. The water samples were sealed and stored and sent to the laboratory. The collected water samples are transferred to a sealed gas extraction device, and the dissolved N2 gas is separated by low-temperature distillation. The purified gas is injected into an isotope mass spectrometer to quantitatively determine the water sample, including 28 N2, 29 N2 and 30 The nitrogen isotope content of N2 is measured, and based on the measurement results and the sampling time interval, the production rate of each nitrogen isotope per unit time is calculated.

8. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and model prediction according to claim 7, characterized in that: The S6 includes: Integrate the nitrogen isotope production rate observed in the field and the soil temperature, humidity, and pH value monitored in real time by the intelligent sensing system, import them into the database, clean, standardize, and time-align the data to build a basic data set; Based on the theory of microbial-driven nitrogen cycle, a nitrogen production model including denitrification and anaerobic ammonium oxidation was established. The soil environmental parameters were used as driving variables and the nitrogen isotope production rate was used as the response variable to construct a nonlinear relationship equation. The Bayesian optimization algorithm is used, with the root mean square error between the predicted value and the measured value as the objective function, to dynamically adjust the model parameters. The model generalization ability is evaluated through cross-validation, and finally a high-precision nitrogen production model and the optimal parameter solution are output.

9. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and modeling according to claim 8, characterized in that: The S7 includes: The nitrogen isotope production rate data were obtained from the nitrogen production model to calculate the total N2 production rate of paddy soil and analyze the 15 The N2 production rate of N fertilizers, and then based on the total N2 production rate of paddy soil, 15 N2 production rate of N fertilizers and fertilizers 15 N abundance, to obtain the N2 contribution from fertilizer nitrogen; The N2 production rate of each sampling during the rice growth period was counted, and the total nitrogen loss and fertilizer nitrogen contribution rate during the entire growth period were calculated; Based on the calculated contribution rate and total nitrogen loss, the source of nitrogen production in flooded soils is predicted, and the contribution ratio of soil nitrogen pool and fertilizer nitrogen to nitrogen loss is quickly distinguished.

10. The method for rapidly distinguishing nitrogen sources in flooded soil by combining in-situ observation and model prediction according to claim 9, characterized in that: The calculation process of the total nitrogen loss and fertilizer nitrogen contribution rate during the entire growth period is as follows: According to the characteristics of the rice growth period, a sampling plan was formulated to collect water samples every 1-2 days after fertilization for 3-6 consecutive times. In the early stage of non-fertilization, water samples were collected every 5-10 days until the rice was harvested. No sampling was done during the rice drying period. Each time sampling is performed, the sampling point number, sampling time, soil environmental parameters, and nitrogen isotope production rate are recorded, and all sampling data are organized into a unified database; Within the sampling times n during the entire growth period, calculate the total N2 production rate of the nth sampling, multiply the total N2 production rate of the nth sampling by the time interval between adjacent sampling times, and sum them up to obtain the total nitrogen loss during the entire growth period. The time interval between adjacent sampling times is the time interval between the nth sampling and the n-1th sampling. Within the sampling times n during the entire growth period, calculate the total N2 production rate of the nth sampling during the entire growth period. 15 N2 production rate of N fertilizers; Based on the total nitrogen loss during the entire growth period, 15 N2 production rate of N fertilizers and fertilizers 15 N abundance, calculate the fertilizer nitrogen contribution rate during the entire growth period.

Images