CN-121983165-A - Atmospheric nitrogen sedimentation source analysis method and system coupled with stable isotope tracing and Bayes mixed model

CN121983165ACN 121983165 ACN121983165 ACN 121983165ACN-121983165-A

Abstract

The invention discloses an atmosphere nitrogen sedimentation source analysis method and system of a coupling stable isotope tracing and Bayes mixed model, which belong to the technical field of environmental monitoring and pollution tracing, and comprise the steps of collecting atmosphere dry sedimentation samples and wet sedimentation samples, recording sampling metadata, obtaining nitrogen to-be-detected solutions in various forms through automatic separation and purification, and measuring stable isotope values; the method comprises the steps of integrating multisource data to construct a dynamic source end member isotope characteristic database, inputting an improved Bayesian mixed model, carrying out parameter optimization and polymorphic nitrogen differential analysis, outputting a nitrogen source contribution proportion and an uncertainty interval, coupling an LSTM prediction model, combining emission scenes to realize short-term prediction, medium-term prediction and long-term prediction, and generating emission reduction decision suggestions through space-time dynamic visualization. According to the invention, through the coupling of an automatic process, model optimization and multiple technologies, the analysis precision and efficiency are improved, and technical support is provided for the accurate treatment of the atmospheric nitrogen pollution, the cross-regional cooperative control and the ecological risk prevention and control.

Inventors

ZHANG HANDAN
WEI HUAIDONG
Zhang Hancui
GAI XU
LIAN GANG
LI XIAOGANG
YANG LE
HU JIANYONG

Assignees

浙江水利水电学院

Dates

Publication Date: 20260505
Application Date: 20260127

Claims (7)

1. An atmospheric nitrogen sedimentation source analysis method coupled with a stable isotope tracing and Bayes mixed model is characterized by comprising the following steps of: collecting atmospheric dry and wet sedimentation samples, and synchronously recording sampling metadata; carrying out automatic separation and purification on the sample to obtain a solution to be tested of nitrogen in various forms; Measuring stable isotope values of nitrogen in various forms by using a solution to be measured to obtain measurement data; Acquiring an emission source list, meteorological data and land utilization data of a measurement area, integrating the measurement data, the emission source list, the meteorological data and the land utilization data, and constructing a dynamic updating source end member isotope characteristic database; Inputting the integrated data into an improved Bayesian mixed model, carrying out probabilistic source analysis, calculating the contribution proportion and uncertainty interval of each preset nitrogen source, and optimizing the measured data and the parameters of the Bayesian mixed model based on the output and input data of the Bayesian mixed model; Based on the analysis result, combining with future emission situations, carrying out short-term and medium-long-term prediction of source contribution proportion through a coupled LSTM prediction model to obtain a prediction result; and carrying out space-time dynamic visualization on the analysis result and the prediction result, and automatically generating a corresponding pollution emission reduction decision suggestion.
2. The method for analyzing an atmospheric nitrogen sedimentation source coupled with a stable isotope labeling and bayesian mixed model according to claim 1, wherein collecting dry and wet sedimentation samples of the atmosphere and synchronously recording sampling metadata comprises the steps of: setting a plurality of sampling points in a measurement area by adopting a three-level point distribution strategy; Sampling by adopting a precipitation and dust fall automatic sampler with an automatic opening and closing function; The synchronously recorded sampling metadata comprise sampling point longitude and latitude, altitude, sample type, sampling start-stop time, precipitation amount, precipitation pH value, conductivity, temperature, humidity, wind speed and wind direction.
3. The method for analyzing an atmospheric nitrogen sedimentation source coupled with a stable isotope labeling and bayesian mixed model according to claim 1, wherein the automatic separation and purification of the sample comprises: automatically separating NH 4 + 、NO 3 - in the sample by using a full-automatic ion analyzer through an ion exchange resin column; Calculating DON concentration by difference subtraction; Automatically concentrating and purifying the separated solution by adopting a rotary evaporator; And setting blank control and standard sample for quality control, and automatically triggering reprocessing of unqualified samples.
4. The method for atmospheric nitrogen settlement source resolution coupling stable isotope labeling and bayesian mixed model according to claim 1, wherein said improved process for optimizing bayesian mixed model comprises: setting prior contribution probabilities of all nitrogen sources based on an emission source list; Mapping the precision grade of the measured data into a model input weight; Dynamically adjusting fractionation coefficient parameters by using the sampling metadata; obtaining posterior probability distribution of the contribution proportion of each nitrogen source through iterative calculation by adopting a Markov chain Monte Carlo simulation algorithm; When the width of the postdelay distribution probability interval exceeds a preset value, triggering the diagnosis of the measured data, and correcting model parameters according to the diagnosis result, wherein the diagnosis content comprises the measured data duty ratio and the source end member characteristic overlapping degree of which the data precision is lower than the preset precision.
5. The method for atmospheric nitrogen settlement source analysis coupled with stable isotope labeling and bayesian mixed model according to claim 1, wherein the probabilistic source analysis comprises differential analysis of polymorphic nitrogen: For NH 4 + -N, distinguishing agricultural fertilization, livestock and poultry cultivation and industrial ammonia escape sources based on delta 15 N value; For NO 3 - -N, combining delta 15 N and delta 18 O double indexes to distinguish atmospheric sedimentation, industrial emission, traffic tail gas and biomass combustion sources; For DON, inversion of source contribution is performed based on a preset organic nitrogen source end member delta 15 N range and total nitrogen duty cycle constraint.
6. The method for analyzing the atmospheric nitrogen settlement source by coupling the stable isotope labeling and bayesian mixed model according to claim 1, wherein the predicting by the coupled LSTM prediction model comprises: Historical Bayesian analysis posterior distribution data is used as a training set; Inputting future emission scene parameters including a reference scene, a strong emission reduction scene and a moderate emission reduction scene; Outputting a source contribution predicted value with a confidence interval larger than a preset interval by combining with meteorological trend data; And (5) adopting a rolling training strategy to retrain the model based on a preset period.
7. An atmospheric nitrogen settlement source analysis system coupled with a stable isotope tracing and bayesian mixed model, which is characterized by comprising: the automatic sampling module is used for collecting atmospheric dry and wet sedimentation samples and synchronously recording sampling metadata; the pretreatment module is used for automatically separating and purifying the sample to obtain a solution to be tested of nitrogen in various forms; The isotope measurement module is used for measuring stable isotope values of nitrogen in various forms by using the solution to be measured to obtain measurement data; The data integration module is used for acquiring an emission source list, meteorological data and land utilization data of the measurement area, integrating the measurement data, the emission source list, the meteorological data and the land utilization data and constructing a dynamic updating source end member isotope characteristic database; the analysis module is used for inputting the integrated data into the improved Bayesian mixed model, carrying out probabilistic source analysis, calculating the contribution proportion and the uncertainty interval of each preset nitrogen source, and optimizing the parameters of the measured data and the Bayesian mixed model based on the output and input data of the Bayesian mixed model; and the prediction module is used for carrying out short-term and medium-long-term prediction on the source contribution proportion by combining the future emission scene based on the analysis result and a coupled LSTM prediction model to obtain a prediction result. And the visualization module is used for carrying out space-time dynamic visualization on the analysis result and the prediction result and automatically generating a corresponding pollution emission reduction decision suggestion.

Description

Atmospheric nitrogen sedimentation source analysis method and system coupled with stable isotope tracing and Bayes mixed model Technical Field The invention relates to the technical field of environmental monitoring and pollution tracing, in particular to an atmospheric nitrogen sedimentation source analysis method and an atmospheric nitrogen sedimentation source analysis system for a coupling stable isotope tracking and Bayesian mixed model. Background Atmospheric nitrogen sedimentation is a process that nitrogen-containing compounds in the atmosphere enter the ground surface in a dry-wet sedimentation mode, and excessive nitrogen sedimentation causes ecological problems such as soil acidification, water eutrophication, reduced biological diversity and the like. Accurate traceability is a key for making regional atmosphere pollution control plans and implementing cross-regional cooperative control. The current technology mainly relies on the combined application of stable isotope labeling (EA-IRMS measurement delta 15N、δ18 O) and a Bayesian mixed model (SIAR), and the latter is more advantageous in theory than the traditional model by the uncertainty of source contribution of MCMC simulation quantification. However, the existing isotope-bayesian coupling scheme still has a deep bottleneck in practical engineering application: Although stable isotope tracing (delta 15N、δ18 O measurement) is used for nitrogen source identification, different nitrogen source isotope characteristics are overlapped (for example, delta 15 N of traffic and industrial NO x can be in the range of 5-12 per mill), an organic nitrogen DON isotope database is lost, the total nitrogen analysis deviation reaches 25-35%, and then sampling, pretreatment and measurement are manually completed, the period of a single sample is 3-5 days, and large-area high-density sampling cannot be adapted, so that space-time dynamics are difficult to capture. Further, the traditional methods such as the traditional mathematical model has limited resolving power, the chemical mass balance model, the positive definite matrix factorization model and the like are highly dependent on complete and accurate emission source list data, and have huge errors in areas with imperfect lists. Different forms of nitrogen generated by the same source cannot be distinguished, static analysis can be performed only based on historical data, dynamic prediction capability for future emission situations is lacking, and prospective management decisions cannot be supported. Therefore, how to provide an atmospheric nitrogen sedimentation source analysis method and system capable of deeply fusing stable isotope labeling and bayesian mixed model is a problem to be solved by those skilled in the art. Disclosure of Invention In view of the above, the invention provides an atmospheric nitrogen sedimentation source analysis method and an atmospheric nitrogen sedimentation source analysis system for coupling stable isotope labeling and Bayesian mixed models, which provide a set of atmospheric nitrogen sedimentation source analysis method and an atmospheric nitrogen sedimentation source analysis system integrating high precision, high efficiency, dynamic prediction and decision support into a whole through deep fusion of the stable isotope labeling and the Bayesian mixed models, and provide powerful technical support for the accurate management of atmospheric nitrogen pollution and the ecological environment protection. In order to achieve the above purpose, the present invention adopts the following technical scheme: in one aspect, the invention provides an atmospheric nitrogen sedimentation source analysis method coupled with a stable isotope tracing and Bayes mixed model, comprising the following steps: collecting atmospheric dry and wet sedimentation samples, and synchronously recording sampling metadata; carrying out automatic separation and purification on the sample to obtain a solution to be tested of nitrogen in various forms; Measuring stable isotope values of nitrogen in various forms by using a solution to be measured to obtain measurement data; Acquiring an emission source list, meteorological data and land utilization data of a measurement area, integrating the measurement data, the emission source list, the meteorological data and the land utilization data, and constructing a dynamic updating source end member isotope characteristic database; Inputting the integrated data into an improved Bayesian mixed model, carrying out probabilistic source analysis, calculating the contribution proportion and uncertainty interval of each preset nitrogen source, and optimizing the measured data and the parameters of the Bayesian mixed model based on the output and input data of the Bayesian mixed model; Based on the analysis result, combining with future emission situations, carrying out short-term and medium-long-term prediction of source contribution proportion through a coupled LSTM pr