Method for calculating river nitric oxide emissions based on land-river-atmosphere simulations

Abstract

The application discloses a river nitric oxide emission calculation method based on a land-river-air simulation, and belongs to the field of environmental engineering. The existing method for measuring river N2O has poor accuracy. A RF regression model is trained by using a nitrogen emission prediction set, a geographical variable prediction set and a climate variable prediction set, and a trained RF regression model is obtained. The nitrogen emission test set, the geographical variable test set and the climate variable test set are input into the trained RF regression model, and the river water concentration of each sub-basin in each region is output. The river hydrological parameters of each sub-basin in each region and the river water concentration of each sub-basin are input into an air-water interface gas exchange model, and the total river nitric oxide emission of each sub-basin in each region is obtained. The application is used for obtaining the total river nitric oxide emission.

Description

TECHNICAL FIELD

[0001] The application relates to river N2O emission calculation and belongs to the cross field of environmental engineering, environmental system simulation and prediction technology and computer technology. BACKGROUND

[0002] Nitrous oxide (N2O) is a long-lived atmospheric trace gas with an atmospheric residence time of more than 100 years and a very strong greenhouse effect. The global warming potential of N2O is 298 times that of CO2, and it is also the most important anthropogenic atmospheric pollutant causing stratospheric ozone depletion. River ecosystems have been identified as an important source of global N2O. Current research shows that the N2O emissions of river ecosystems exceed 10% of the total global anthropogenic N2O emissions, but the formation and emission mechanism is still unclear. Therefore, accurately identifying and calculating river N2O emissions, determining the spatial and temporal distribution of river N2O emissions in large-scale river basins, and proposing pollution control strategies based on local conditions are of great significance for water safety.

[0003] The river N2O emission factor (EF5) proposed by the Intergovernmental Panel on Climate Change (IPCC) has been widely used in the estimation of river N2O emissions worldwide in the past 20 years. On the one hand, the application of constant emission factor will lead to overestimation or underestimation of N2O emissions in the world rivers, depending on the time and location of the emission factor measurement. On the other hand, the emission factor method ignores the function of river as a link between land and atmospheric systems, and cannot determine the nitrogen transformation efficiency according to the hydrological and water quality state of a specific river, thus failing to achieve the coordinated control of the pollution source and process of greenhouse gases. Therefore, the existing methods for estimating river N2O emissions in different places have poor accuracy, and the use of accurate and effective model tools to quantitatively describe the nitrogen transfer process in the "land-river-atmosphere" system has become a key problem in the accurate calculation of river N2O emissions. SUMMARY

[0004] The purpose of the present application is to solve the problem of poor accuracy of existing methods for measuring river N2O emissions, and to propose a method for simulating river nitrous oxide emissions based on land-river-atmosphere.

[0005] The method for simulating river nitrous oxide emissions based on land-river-atmosphere comprises the following steps:

[0006] Step 1: Obtain the nitrogen emissions on the land in each region;

[0007] Step 2, divide the nitrogen emissions on land in each region into a nitrogen emission prediction set and a nitrogen emission test set, divide the geographical variables in each region into a geographical variable prediction set and a geographical variable test set, divide the climate variables in each region into a climate variable prediction set and a climate variable test set, train the RF regression model using the nitrogen emission prediction set, the geographical variable prediction set, and the climate variable prediction set, obtain the trained RF regression model, input the nitrogen emission test set, the geographical variable test set, and the climate variable test set into the trained RF regression model, and output the river water concentration of each sub-basin in each region;

[0008] Step 3, obtain the river hydrological parameters of each sub-basin in each region, and the river hydrological parameters of each sub-basin include river water depth, water flow velocity, water temperature, and water surface area;

[0009] Step 4, input the river hydrological parameters of each sub-basin in each region and the river water concentration of each sub-basin into the air-water interface gas exchange model for concentration conversion, and obtain the total river N2O emission amount of each sub-basin in each region.

[0010] Preferably, the nitrogen emissions on land include urban life human-caused nitrogen emissions, industrial human-caused nitrogen emissions, urban rainwater surface source nitrogen emissions, rural life nitrogen emissions, farmland planting nitrogen emissions, and livestock and poultry breeding nitrogen emissions.

[0011] Preferably, the urban life human-caused nitrogen emissions include:

[0012]

[0013] In the formula, URN discharge represents the nitrogen emissions from urban life to the water environment; URN direct represents the amount directly discharged into the water environment without sewage treatment; URN treat represents the urban life nitrogen emissions released to the water environment after treatment by the urban sewage treatment plant; URPop represents the number of urban population; URCof water represents the water consumption coefficient of urban residents (per capita); URRate direct represents the ratio of direct discharge to total sewage; URConc direct represents the nitrogen emission concentration of direct discharge sewage; URRate treat represents the proportion of sewage treated by the sewage treatment plant to the total sewage; URRate reuse represents the reuse rate of effluent of the sewage treatment plant; URConc treat represents the pollutant emission concentration of the sewage treatment plant;

[0014] Urban rainwater surface source nitrogen emissions:

[0015]

[0016] USRN discharge indicates the nitrogen discharged by urban surface runoff, USRNRate i indicates the nitrogen discharged per unit area of runoff corresponding to functional area i, i = 1, 2, 3, 4, respectively corresponding to residential area, commercial area, industrial area and other areas; UArea i indicates the area of functional area i; NCon i indicates the nitrogen discharge concentration of functional area i; PDen i，j indicates the urban population density parameter of functional area i in j year; SF represents the cleaning frequency of urban community, 1 for cleaning once a day; AP i indicates the annual precipitation (cm) of the city where functional area i is located in j year; NCoef represents the nitrogen discharge correction coefficient; PDen i，j is the correction coefficient of functional area; is the population density of administrative region; UACoef j indicates the correction coefficient of urban area in j year, UACoef 2018 indicates the correction coefficient of urban area in 2018, UArea 2018 indicates the area of functional area in 2018;

[0017] Rural life nitrogen discharge:

[0018]

[0019] RRN discharge indicates the nitrogen discharged by rural life into the water environment, RRN direct indicates the water quantity directly discharged into the water environment without sewage treatment; RRN treat indicates the pollution discharge released into the water environment after treatment by rural sewage treatment facilities; RRRate treat indicates the proportion of sewage treated by sewage treatment facilities to the total sewage in rural areas; RRCoef removal indicates the pollutant removal rate of rural sewage treatment; RRPop indicates the total number of rural residents; RRRate DryT indicates the ratio of dry toilets to total toilets in rural areas; RRCoef DryT indicates the per capita pollutant discharge coefficient of rural residents using dry toilets; RRRate FlushT indicates the proportion of flushing toilets to total toilets in rural areas; RRCoef FlushT indicates the pollutant discharge coefficient of rural residents using flushing toilets;

[0020] Farmland planting nitrogen discharge:

[0021]

[0022] CFN = CFArea * CFCoef discharge represents the pollution emission of crop cultivation discharged into the water environment; CFArea refers to the total sowing area of farmland; CFCoef discharge is the pollution loss coefficient of farmland; CFCoef discharge，2017 is the standard nitrogen loss coefficient of farmland based on the data of the Second National Pollution Source Census in 2017; CFFertilizer i represents the amount of chemical fertilizer applied in the i-th year; CFFertilizer 2017 represents the amount of chemical fertilizer applied in 2017;

[0023] Nitrogen emission of livestock and poultry breeding:

[0024]

[0025] LFN = LFArea * LFNumber * LFCoef discharge represents the nitrogen discharged into the water environment by livestock farming; LFN centralized is the pollution emission of centralized livestock farming; LFN free is the pollution emission of free-range livestock farming; LFNumber i is the number of fattening livestock i; LFRate centralized，i is the ratio of the number of centralized breeding of species i to the total number; LFCoef centralized，i is the nitrogen emission coefficient of centralized crop species i; LFRate free，i is the ratio of the number of free-range of species i to the total number; LFCoef free，i is the nitrogen emission coefficient of free-range livestock species i;

[0026] The process of obtaining the industrial artificial nitrogen emission is as follows:

[0027] The industrial nitrogen emission of each region is obtained, and according to the area proportion of each sub-basin in the region, the industrial nitrogen emission of each region is converted into the nitrogen emission of each sub-basin, which is the nitrogen discharged into the water environment by industry.

[0028] Preferably, in step 2, the specific process of outputting the river water concentration of each sub-basin in each region is as follows:

[0029] The environmental investment data in each region is divided into an environmental investment data prediction set and an environmental investment data test set, and the social statistical data of population and economy in each region is divided into a social statistical prediction set of population and economy, a social statistical test set of population and economy;

[0030] The RF regression model is trained by using the nitrogen emission prediction set, the geographic variable prediction set, the climate variable prediction set, the environmental investment data prediction set, and the population and economic social statistics prediction set, and the trained RF regression model is obtained. The nitrogen emission test set, the geographic variable test set, the climate variable test set, the environmental investment data test set, and the population and economic social statistics test set are input into the trained RF regression model, and the river water concentration of each sub-basin in each region is output.

[0031] Preferably, the environmental investment data includes environmental pollution investment proportion and environmental regulation number;

[0032] The population and economic social statistics data include population density, gross national product, fertilizer application amount, mobile phone household number, and grade highway kilometer number;

[0033] The geographic variables include soil bulk density, soil organic matter, soil conductivity, soil pH value, soil type proportion, land use proportion, maximum patch index, edge density, landscape shape index, Shannon diversity index, and median of landscape perimeter area ratio;

[0034] The climate variables include average temperature, accumulated temperature greater than 10℃, average rainfall, humidity index, and normalized vegetation index.

[0035] Preferably, step 3, obtaining the river hydrological parameters of each sub-basin, the specific process is:

[0036] The climate data of each sub-basin in each region is input into the SWAT model for hydrological parameter simulation, and the river hydrological parameters of each sub-basin are output, wherein the climate data includes rainfall, temperature, wind speed, relative humidity, and solar radiation data.

[0037] Preferably, the total river N2O emission amount of each sub-basin is represented as:

[0038]

[0039] In the formula, is the N2O emission flux of the river to the atmosphere; C w is the dissolved N2O concentration in the river surface water; C eq is the theoretical N2O concentration in the surface water in equilibrium with atmospheric N2O; is the N2O gas transfer velocity; 240 is the unit conversion coefficient; is the total N2O emission amount of the river basin in a given year; SA is the water surface area in the given sub-basin; N is the number of days in a given month; DIN is the dissolved inorganic nitrogen concentration simulated by the RF regression model; T K is the water temperature simulated by SWAT; is the molecular weight of N2O in N; C airis the monthly scale N2O concentration in the air calculated according to the N2O combined data set provided by NOAA Global Monitoring Laboratory; is the Schmidt number, T is the water temperature; n is the index; k 600 is the transfer rate of the gas at 20 DEG C.

[0040] Preferably, n is 2 / 3, W 10 is the wind speed at 10 m high.

[0041] Preferably, n is 1 / 2, k 600 = 1.0 + 1.719 x (V / H) 0.5 )+ 2.58 x W 10 , W 10 is the wind speed at 10 m high, V and H are the flow rate and water depth respectively.

[0042] The beneficial effects of the present application are:

[0043] 1. A cross-media river N2O emission estimation method is proposed, which breaks through the calculation boundary limitation of the traditional method, utilizes the influence of nitrogen emission on land on nitrogen in the river, describes the land-river-atmosphere process of river N2O emission, realizes the accurate simulation of river N2O from the source to the emission process, and provides scientific means and technical tools for the coordinated control of land pollution, river pollution and atmospheric pollution;

[0044] 2. The RF regression model is used to obtain the river water concentration of each sub-basin, and the air-water interface gas exchange model is used to obtain the total amount of river N2O emission of each sub-basin in each region, so that a hybrid modeling method for river N2O emission based on machine learning and mechanism model is proposed, which respectively gives full play to the prediction performance of machine learning and mechanism model, effectively improves the calculation accuracy of the existing method, and realizes the calculation accuracy of river N2O in the large-scale basin range to the monthly scale and the catchment resolution (<500 km 2 ).

[0045] 3. The statistical data and conventional meteorological monitoring data are combined, and the river hydrological water quality process and the gas transfer law are coupled, which avoids the calculation result error caused by the traditional method deviating from the actual law, can accurately identify the river greenhouse gas emission state of each region, and provides reliable data support for pollution risk evaluation, pollution tracing and pollution control strategy;

[0046] 4、The application considers the influence of atmospheric background N2O concentration on river N2O gas emission, effectively reduces the parameter uncertainty in the calculation process of river N2O emission, and at the same time, the parameterized calculation formula and the programmed calculation mode make the application applicable to various basins in China, and can effectively promote the implementation of scientific pollution control work in China. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 Flow chart of the calculation method of river nitric oxide emission based on land-river-atmosphere simulation;

[0048] Figure 2 Figure of human nitrogen emission at sub-basin level in the Yangtze River Basin in 2019;

[0049] Figure 3 Figure of river DIN at sub-basin level in the Yangtze River Basin in 2019;

[0050] Figure 4 Figure of river dissolved N2O concentration at sub-basin level in the Yangtze River Basin in 2019;

[0051] Figure 5 Figure of river water surface area simulation results at sub-basin level in the Yangtze River Basin in 2019;

[0052] Figure 6 Figure of gas transfer rate calculation results at sub-basin level in the Yangtze River Basin in 2019;

[0053] Figure 7 Figure of N2O emission flux calculation results at sub-basin level in the Yangtze River Basin in 2019;

[0054] Figure 8 Figure of N2O emission calculation results at sub-basin level in the Yangtze River Basin in 2019. DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0056] It should be noted that the embodiments in the application and the features in the embodiments can be combined with each other without conflict.

[0057] The application will be further described below in combination with the drawings and specific embodiments, but not as a limitation of the application.

[0058] Specific embodiment one:Figure 1 The present embodiment describes a method for calculating riverine nitric oxide emissions based on land-river-atmosphere simulation, which comprises:

[0059] Step 1, obtaining the nitrogen emissions on land in each region;

[0060] Step 2, dividing the nitrogen emissions on land in each region into a nitrogen emissions prediction set and a nitrogen emissions test set, dividing the geographical variables in each region into a geographical variable prediction set and a geographical variable test set, dividing the climate variables in each region into a climate variable prediction set and a climate variable test set, training an RF regression model using the nitrogen emissions prediction set, the geographical variable prediction set, and the climate variable prediction set, obtaining the trained RF regression model, inputting the nitrogen emissions test set, the geographical variable test set, and the climate variable test set into the trained RF regression model, and outputting the river water concentration of each sub-basin in each region;

[0061] Step 3, obtaining the river hydrological parameters of each sub-basin in each region, including river water depth, flow velocity, water temperature, and water surface area;

[0062] Step 4, inputting the river hydrological parameters of each sub-basin in each region and the river water concentration of each sub-basin into the air-water interface gas exchange model for concentration conversion, and obtaining the total river N2O emissions of each sub-basin in each region.

[0063] In the present embodiment, random forest (RF) as a mature machine learning algorithm plays a key role in surface water quality prediction. Here, the present application uses RF regression to predict the water quality concentration in each sub-basin of the watershed. Long-term monthly average water quality concentration is used as the dependent variable of each sub-basin, and driving variables (geography, hydrology, climate, and human activity) are used as independent variables. The training process and accuracy verification process of each RF model used 80% and the remaining 20% of the randomly and proportionally selected data set, respectively. After model training and adjustment, the prediction performance of each RF model was evaluated using multiple statistical indicators, including the square correlation coefficient (R 2 ), root mean square error (RMSE), and mean absolute error (MAE). Specifically: input 80% of the data into the RF model, train the model to obtain model parameters, and input the model parameters into the model. When inputting 20% of the data into the model, the model outputs the water quality concentration. The concentration is compared with the existing concentration using the square correlation coefficient (R 2 ), root mean square error (RMSE), and mean absolute error (MAE) to verify the accuracy of the result model.

[0064] Figure 1The input data are: digital elevation model in river hydrological state, meteorological data, land use, soil type and soil property, soil property in river water quality response, temperature, human emission, land use, rainfall, environmental input, landscape index, evaporation and social index, atmospheric N2O concentration in river N2O emission, hydrology, wind speed, river N2O concentration, river flow depth and river water area; the data in the dashed box are the data in the public database; human emission, river N2O concentration, river flow depth and river water area represent the data output by the upper sub-model.

[0065] Specific implementation method two: this implementation method is a further limitation of the specific implementation method one, in this implementation method, the nitrogen emission amount on land includes urban life human nitrogen emission amount, industrial human nitrogen emission amount, urban rainwater surface source nitrogen emission amount, rural life nitrogen emission amount, farmland planting nitrogen emission amount and livestock breeding nitrogen emission amount.

[0066] In this implementation method, the human nitrogen emission model of the terrestrial system is a bottom-up model, which quantifies the nitrogen input to the river from the human production and life process, and considers the dispersion and centralized treatment process of sewage. The sources of nitrogen emission include six pollution sources of urban life, industry, urban rainwater surface source, rural life, farmland planting and livestock breeding. The urban and rural area weighting method is used to convert the human nitrogen emission data based on the county-level administrative boundary to the sub-basin level estimation. First, we calculate the county-level emission amount of various pollution sources based on the county-level nitrogen emission calculation framework. Second, combined with the county-level administrative boundary map and the sub-basin boundary map, and the urban and rural areas of each cross unit are determined respectively. Third, we decompose the county-level human emission data into cross units according to the determined urban and rural areas. Finally, we summarize the human nitrogen emission data of all target river basins based on the cross units located in the same river basin.

[0067] Specific implementation method three: this implementation method is a further limitation of the specific implementation method two, in this implementation method, the urban life human nitrogen emission amount is:

[0068]

[0069] In the formula, URN discharge represents the nitrogen emission amount of urban life to the water environment; URN direct represents the amount directly discharged into the water environment without sewage treatment; URN treat represents the urban life nitrogen emission amount released to the water environment after treatment by the urban sewage treatment plant; URPop represents the number of urban population; URCof water represents the water consumption coefficient of urban residents (per capita); URRate direct represents the ratio of direct sewage discharge to total sewage; URConc directURConc treat URRate reuse URRate treat URRate

[0070] URRate URRate

[0071] URRate URRate URRate

[0072] URRate discharge URRate i URRate i URRate i URRate i，j URRate i URRate i，j URRate URRate j URRate 2018 URRate 2018 URRate URRate

[0073] RRN RRN

[0074] RRN RRN RRN

[0075] RRN discharge RRN direct RRN treat RRN treat RRN removal RRN DryT RRNDryT represents the pollutant emission coefficient per capita of rural residents using dry toilets; RRRate FlushT represents the proportion of flush toilets to total toilets in rural areas; RRCoef FlushT represents the pollutant emission coefficient of rural residents using flush toilets;

[0076] Nitrogen emissions from farmland planting:

[0077]

[0078] CFN, i = ∑ (CFArea x CFCoef x CFFertilizer x CFFertilizer discharge represents the crop breeding pollution emissions discharged into the water environment; CFArea refers to the total sowing area of farmland; CFCoef discharge is the pollutant loss coefficient of farmland; CFCoef discharge，2017 is the standard nitrogen loss coefficient of farmland based on the data of the Second National Pollution Source Census in 2017; CFFertilizer i represents the amount of fertilizer applied in the i-th year; CFFertilizer 2017 represents the amount of fertilizer applied in 2017;

[0079] Nitrogen emissions from livestock and poultry breeding:

[0080]

[0081] LFN, i = ∑ (LFNumber x LFRate x LFCoef x LFRate discharge represents the nitrogen discharged into the water environment by animal husbandry; LFN centralized is the centralized livestock pollution emissions; LFN free is the pollution emissions of free-range livestock; LFNumber i is the number of fattening livestock i; LFRate centralized，i is the ratio of the number of centralized breeding of species i to the total number; LFCoef centralized，i is the nitrogen emission coefficient of centralized farmland species i; LFRate free，i is the ratio of the number of free-range of species i to the total number; LFCoef free，i is the nitrogen emission coefficient of free-range livestock species i;

[0082] The process of obtaining industrial artificial nitrogen emissions is as follows:

[0083] The industrial nitrogen emissions of each region are obtained, and the industrial nitrogen emissions of each region are converted into the nitrogen emissions of each sub-basin according to the area proportion of each sub-basin in the region, as the nitrogen discharged into the water environment by industry.

[0084] In this embodiment, for urban life human nitrogen emissions: by combining the per capita domestic water consumption of residents, urban sewage treatment rate and urban sewage reuse rate, the nitrogen balance of urban life pollution emissions in any given year is determined. And according to the sewage treatment rate, the uncollected sewage and the collected sewage are distinguished, and the calculation method is as formula 1; The formula parameters in formula 1 can be obtained from China Urban Construction Statistical Yearbook, China Environmental Statistical Yearbook, provincial statistical yearbook and the second pollution source census manual.

[0085] For rural life nitrogen emissions, the human emissions of rural residents are divided into direct emissions and emissions after sewage treatment, which are calculated as formula 3;

[0086] For farmland planting nitrogen emissions, the main source of human nitrogen emissions in farmland is the rainfall erosion of fertilizers, which is calculated as formula 4, wherein CFCoef discharge is the pollutant loss coefficient of farmland per unit area; CFCoef discharge，2017 is the standard nitrogen loss coefficient per unit area of farmland based on the second China pollution source census data in 2017.

[0087] Specific implementation method four: this embodiment is a further limitation of the specific implementation method two, in this embodiment, in step 2, the specific process of outputting the river water concentration of each sub-basin in each region is:

[0088] Divide the environmental investment data in each region into environmental investment data prediction set and environmental investment data test set, and divide the population and economic social statistics in each region into population and economic social statistics prediction set, population and economic social statistics test set;

[0089] Use the nitrogen emission prediction set, the geographical variable prediction set, the climate variable prediction set, the environmental investment data prediction set, and the population and economic social statistics prediction set to train the RF regression model to obtain the trained RF regression model. Input the nitrogen emission test set, the geographical variable test set, the climate variable test set, the environmental investment data test set, and the population and economic social statistics test set into the trained RF regression model, and output the river water concentration of each sub-basin in each region.

[0090] Specific implementation method five: this embodiment is a further limitation of the specific implementation method four, in this embodiment, the environmental investment data includes environmental pollution investment proportion and environmental regulation number;

[0091] The population and economic social statistics data include population density, gross national product, fertilizer application amount, mobile phone household number and grade highway kilometer number;

[0092] Geographical variables include soil bulk density, soil organic matter, soil electrical conductivity, soil pH, soil type proportion, land use proportion, maximum patch index, edge density, landscape shape index, Shannon diversity index, and median of landscape perimeter-area ratio.

[0093] Climate variables include mean air temperature, accumulated temperature above 10°C, mean precipitation, humidity index, and normalized difference vegetation index.

[0094] In this embodiment, the river water quality model includes 30 geographical, hydro-climatic, and human activity driving variables, which are monthly or annual time series. The selection of driving factors is based on the combination of their influence on DIN concentration in surface water and data availability. Geographical variables including soil type and properties, land use types, and landscape indices are included because they are widely recognized as possible to affect river DIN concentration. Regarding hydro-climatic variables, precipitation can directly affect the source and transport of pollutants by changing river flow, while temperature can indirectly affect water quality by affecting the transformation patterns and biochemical reaction rates of pollutants in water. Therefore, temperature, precipitation, and humidity index are considered as relevant hydro-climatic variables. For human activity variables, in addition to relevant social statistical data such as population and economy widely used in water quality prediction, we also consider anthropogenic nitrogen emissions and environmental investment data.

[0095] Specific embodiment six: this embodiment is a further limitation of the specific embodiment one, in this embodiment, step 3, obtaining the river hydrological parameters of each sub-basin, the specific process is:

[0096] Input the climate data of each sub-basin in each region into the SWAT model for hydrological parameter simulation, and output the river hydrological parameters of each sub-basin, wherein the climate data includes rainfall, air temperature, wind speed, relative humidity and solar radiation data.

[0097] In this embodiment, according to the river length and average width of each sub-basin, the water surface area of the river in the river hydrological parameters of each sub-basin is calculated.

[0098] SWAT (Soil and Water Assessment Tool) was developed by Dr. Jeff Arnold of the Agricultural Research Service of the U.S. Department of Agriculture (USDA) in 1994. The original purpose of the model development was to predict the long-term effects of land management on water, sediment, and chemical yields under complex and variable conditions of soil type, land use, and management practices in large river basins. SWAT is a GIS-based, distributed hydrologic model that has been rapidly developed and applied in recent years. It mainly uses spatial information provided by remote sensing and geographic information systems to simulate a variety of different hydrological physical and chemical processes, such as water quantity, water quality, and the transport and transformation of pesticides.

[0099] Specific embodiment seven: This embodiment is a further limitation of the specific embodiment one, in which the total amount of river N2O emissions of each sub-basin is represented as:

[0100]

[0101] In the formula, is the N2O emission flux of the river to the atmosphere; C w is the dissolved N2O concentration in the surface water of the river; C eq is the theoretical N2O concentration in equilibrium with atmospheric N2O in surface water; is the N2O gas transfer velocity; 240 is the unit conversion coefficient; is the total amount of N2O emissions of the watershed in a given year; SA is the water surface area in the given sub-basin; N is the number of days in the given month; DIN is the dissolved inorganic nitrogen concentration simulated by the RF regression model; T K is the water temperature simulated by SWAT; is the molecular weight of N2O in N; C air is the monthly scale N2O concentration in the air calculated according to the N2O combined data set provided by NOAA Global Monitoring Laboratory; is the Schmidt number, T is the water temperature; n is the index; k 600 is the gas transfer rate at 20°C.

[0102] Specific embodiment eight: This embodiment is a further limitation of the specific embodiment seven, in which, in the case of low wind speed, n is 2 / 3, and at this time W 10 is the wind speed at 10 m high.

[0103] Specific embodiment nine: This embodiment is a further limitation of the specific embodiment seven, in which, in the case of high wind speed, n is 1 / 2, and at this time k 600 = 1.0 + 1.719 × (V / H)0.5 )+ 2.58 x W 10 , W 10 V and H are the flow velocity and water depth, respectively.

[0104] Experimental verification:

[0105] The application has been successfully applied in the calculation of N2O emissions in the Yangtze River Basin. Taking 2019 as an example, the calculation of artificial nitrogen emissions in the Yangtze River Basin, river water quality response and river N2O.

[0106] The specific process is:

[0107] 1. Calculation of land artificial nitrogen emissions

[0108] Using a bottom-up artificial nitrogen emission calculation model, county-level land artificial nitrogen emissions were calculated based on county-level basic statistical data in the Yangtze River Basin in 2019, including urban life, industry, urban rainwater non-point source, rural life, farmland planting and livestock breeding. The calculation results of artificial nitrogen emissions of various pollution sources in the Yangtze River Basin are shown in Table 1,

[0109] Table 1 Calculation results of artificial nitrogen emissions of various pollution sources in the Yangtze River Basin

[0110]

[0111] The urban and rural area weighting method was used to convert the artificial nitrogen emission data based on the county-level administrative boundary to the sub-basin level estimation. Based on the county-level emissions of various pollution sources, combined with the county-level administrative boundary map and the sub-basin boundary map, and the urban and rural areas of each cross unit were determined. According to the determined urban and rural areas, the county-level artificial emission data was decomposed into cross units. Finally, we summarized the artificial nitrogen emission data of the Yangtze River Basin based on the cross units located in the same basin. The artificial nitrogen emissions at the sub-basin level are shown in Table 3. Figure 2

[0112] According to the time emission law of pollution sources, the artificial nitrogen emissions of various pollution sources were divided into monthly scale. The calculation results are shown in Table 2.

[0113] Table 2 Calculation results of monthly artificial nitrogen emissions in the Yangtze River Basin

[0114]

[0115] 2. Simulation of river water quality response process

[0116] ​Thirty driving variables were selected, including 11 geographical variables: soil bulk density, soil organic matter, soil electrical conductivity, soil pH, soil type proportion, land use proportion, largest patch index, edge density, landscape shape index, Shannon diversity index, and median of landscape perimeter-area ratio. Five climatic variables: mean air temperature, accumulated temperature above 10°C, mean precipitation, humidity index, and normalized difference vegetation index. Fifteen human activity variables: urban domestic nitrogen emission, industrial nitrogen emission, urban stormwater nitrogen emission, rural domestic nitrogen emission, farmland planting nitrogen emission, livestock breeding nitrogen emission, population density, gross domestic product, fertilizer application amount, mobile phone number, grade highway length, environmental pollution investment proportion, environmental regulation number, and “three simultaneous” environmental protection investment.

[0117] RF regression was used to predict the water quality concentration in the surface water of each sub-basin in the watershed. The long-term monthly average water quality concentration was used as the dependent variable for each sub-basin, and the driving variables (geographical, hydrological, and human activity) were used as independent variables. The training process and accuracy verification process of the model used 80% and the remaining 20% of the randomly and proportionally selected data set, respectively. The average concentration of river DIN in the Yangtze River Basin in 2019 was 1.31 mg L -1 . The predicted river DIN and dissolved N2O concentrations are shown in Figure 3 and 4 , respectively.

[0118] 3. River hydrological state simulation

[0119] Digital elevation model (DEM) with a resolution of 90 m, soil type map, and slope map of the Yangtze River Basin were collected, and the watershed was discretized into sub-basins with a threshold watershed area of 500 km 2 . These sub-basins were further subdivided into hydrological response units (HRU) based on land use and slope. Long-term daily climate data were used to drive the SWAT simulation, and the model preheater was used for the first 3 years to alleviate the initial conditions and exclude them from the analysis.

[0120] Using the model output values in 2019, the river length, water depth, and flow velocity of each sub-basin were calculated, and the river surface area was calculated based on the river length and average width of each sub-basin. The calculation results of the river surface area are shown in Figure 5 .

[0121] 4. Calculation of river N2O gas emissions

[0122] According to the simulated values of river dissolved N2O concentration, the monitoring values of air temperature and wind speed, combined with the river water temperature output by the SWAT model, the N2O gas transfer rate and emission flux were calculated according to the solubility equation and the air-water interface gas exchange model. Plus the river water surface area simulated by the SWAT model, the river N2O emission of the Yangtze River Basin was calculated. The total river N2O emission of the Yangtze River Basin in 2019 was 2.8 Gg N2O-N yr -1 . The simulation results of gas transfer rate are shown in Figure 6 , and the calculation results of N2O emission flux and total emission are shown in Figure 7 and 8 .

[0123] While the application has been described with reference to particular embodiments thereof, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present application. It will thus be appreciated that numerous modifications can be made to the illustrative embodiments and that other arrangements can be devised without departing from the spirit and scope of the present application as defined by the appended claims. It will be understood that the features described with respect to one embodiment can be used in other embodiments.

Claims

1. A method for calculating river nitrogen monoxide emissions based on land-river-atmosphere simulation, characterized in that, The method includes: Step 1: Obtain nitrogen emissions over inland areas of each region; Step 2: Divide the nitrogen emissions on land in each region into a nitrogen emission prediction set and a nitrogen emission test set; divide the geographic variables in each region into a geographic variable prediction set and a geographic variable test set; divide the climate variables in each region into a climate variable prediction set and a climate variable test set; train the RF regression model using the nitrogen emission prediction set, geographic variable prediction set, and climate variable prediction set to obtain the trained RF regression model; input the nitrogen emission test set, geographic variable test set, and climate variable test set into the trained RF regression model to output the river water quality concentration of each sub-basin in each region; Step 3: Obtain the river hydrological parameters of each sub-basin within each region. The river hydrological parameters of each sub-basin include river depth, flow velocity, water temperature, and water surface area. Step 4: Input the river hydrological parameters and river water quality concentration of each sub-basin in each region into the air-water interface gas exchange model for concentration conversion to obtain the total N2O emissions of each sub-basin in each region. The total N2O emissions from rivers in each sub-basin are expressed as follows: In the formula, It is the flux of N2O emissions from rivers into the atmosphere; C w It is the concentration of dissolved N2O in river surface water; C eq It is the theoretical N2O concentration in surface water that is in equilibrium with atmospheric N2O; This refers to the gas transfer rate of N2O; 240 is a unit conversion factor. is the total N2O emissions for a given year in the watershed; SA is the water surface area within a given sub-watershed; N is the number of days in a given month; DIN is the dissolved inorganic nitrogen concentration simulated by the RF regression model; T K This is the water temperature simulated by SWAT. N₂O is the molecular weight of N, denoteed as N; C air The monthly N2O concentration in the air was calculated based on a combined N2O dataset provided by NOAA's global monitoring laboratories. It is the Schmidt number, where T is the water temperature; n is the exponent; k 600 The gas transfer rate is at 20°C.

2. The method for calculating river nitrogen monoxide emissions based on land-river-atmosphere simulation according to claim 1, characterized in that, Nitrogen emissions on land include anthropogenic nitrogen emissions from urban life, anthropogenic nitrogen emissions from industry, nitrogen emissions from urban stormwater non-point source sources, nitrogen emissions from rural life, nitrogen emissions from farmland cultivation, and nitrogen emissions from livestock and poultry farming.

3. The method for calculating river nitrogen monoxide emissions based on land-river-atmosphere simulation according to claim 2, characterized in that, Anthropogenic nitrogen emissions from urban life: In the formula, URN discharge This refers to nitrogen discharged into the aquatic environment by urban life; URN direct This indicates the amount of wastewater discharged directly into the aquatic environment without treatment; URN treat URPop represents the amount of nitrogen released into the aquatic environment after treatment at urban wastewater treatment plants; URCof represents the urban population. water URRate represents the per capita domestic water consumption coefficient of urban residents. direct This represents the ratio of direct wastewater discharge to total wastewater volume; URConc direct URRate indicates the nitrogen emission concentration of directly discharged wastewater. treat URRate indicates the ratio of wastewater treated by a wastewater treatment plant to the total wastewater volume. reuse Indicates the effluent reuse rate of a wastewater treatment plant; URConc treat This indicates the concentration of pollutants discharged from the wastewater treatment plant; Urban stormwater non-point source nitrogen emissions: In the formula, USRN discharge USRNRate represents nitrogen emissions from urban stormwater non-point source pollution. i This represents the nitrogen emissions per unit area of ​​runoff corresponding to functional zone i, where i = 1, 2, 3, 4, corresponding to residential area, commercial area, industrial area, and other areas, respectively; UArea i NContains the area of ​​functional region i; NCon i Indicates the nitrogen emission concentration in functional zone i; PDe i，j represents the urban population density parameter of functional area i in year j; SF represents the cleaning frequency of the urban community, with 1 representing cleaning once a day; AP i The annual precipitation (cm) of city j in which the functional zone is located is represented; NCoef represents the nitrogen emission correction factor; PDen i，j DP is the functional area correction factor; i 0.54 It refers to the population density of an administrative region; UACoef j UACoef represents the correction factor for the city area in year j. 2018 UArea represents the correction factor for urban area in 2018. 2018 This indicates the area of ​​functional zone 2018; Nitrogen emissions from rural domestic life: In the formula, RRN discharge RRN represents nitrogen discharged into the aquatic environment from rural areas. direct This refers to the volume of water discharged directly into the aquatic environment without sewage treatment; RRN treat This refers to the pollution emissions released into the aquatic environment after treatment by rural sewage treatment facilities; RRRate treat This indicates the proportion of wastewater treated by wastewater treatment facilities to the total wastewater in rural areas; RRCoef removal RRPop represents the pollutant removal rate of rural sewage treatment; RRRate represents the total number of rural residents; RRRate represents the pollutant removal rate of rural sewage treatment. DryT This indicates the ratio of dry toilets to total toilets in rural areas; RRCoef DryT This represents the per capita pollutant emission coefficient for rural residents using dry toilets; RRRate FlushT This indicates the ratio of flush toilets to total toilets in rural areas; RRCoef FlushT This represents the pollutant emission coefficient of rural residents using flush toilets; Nitrogen emissions from farmland cultivation: In the formula, CFN discharge This refers to agricultural and livestock pollution emissions discharged into the aquatic environment; CFArea refers to the total sown area of ​​farmland; CFCoef discharge It is the pollutant loss coefficient of farmland; CFCoef discharge，2017 This is the standard nitrogen loss coefficient for farmland based on data from the Second National Pollution Source Census in 2017; CFFertilizer i This represents the amount of fertilizer applied in year i; CFFertilizer 2017 This indicates the amount of fertilizer applied in 2017. Nitrogen emissions from livestock and poultry farming: In the formula, LFN discharge This refers to nitrogen discharged into the aquatic environment by livestock farming; LFN centralized It refers to the pollution emissions from centralized livestock farming; LFN free It represents the pollution emissions from free-range livestock farming; LFNumber i It refers to the number of fattening livestock; LFRate centralized，i It is the ratio of the number of species i in concentrated breeding to the total number; LFCoef centralized，i It is the nitrogen emission factor of concentrated crop species i; LFRate free，i It is the ratio of the free-range population of species i to the total population; LFCoef free，i It is the nitrogen emission coefficient of free-range livestock species i; The process of obtaining industrial anthropogenic nitrogen emissions is as follows: The industrial nitrogen emissions of each region are obtained. Based on the area ratio of each sub-basin to the region, the industrial nitrogen emissions of each region are converted into the nitrogen emissions of each sub-basin, which are then used as the nitrogen discharged into the water environment by industry.

4. The method for calculating river nitrogen monoxide emissions based on land-river-atmosphere simulation according to claim 1, characterized in that, In step 2, the specific process of outputting the river water quality concentration of each sub-basin within each region is as follows: Environmental investment data in each region is divided into an environmental investment data prediction set and an environmental investment data test set. Social statistics data on population and economy in each region are divided into a social statistics prediction set and a social statistics test set. The RF regression model was trained using the nitrogen emission prediction set, geographic variable prediction set, climate variable prediction set, environmental investment data prediction set, and population and economic social statistics prediction set. The trained RF regression model was then input into the nitrogen emission test set, geographic variable test set, climate variable test set, environmental investment data test set, and population and economic social statistics test set to output the river water quality concentration of each sub-basin in each region.

5. The method for calculating river nitrogen monoxide emissions based on land-river-atmosphere simulation according to claim 4, characterized in that, Environmental investment data includes the proportion of investment in environmental pollution and the number of environmental regulations; Population and economic social statistics include population density, gross national product, fertilizer application, number of mobile phone subscribers, and kilometers of graded highways. Geographic variables include soil bulk density, soil organic matter, soil electrical conductivity, soil pH, soil type proportion, land use proportion, maximum patch index, edge density, landscape shape index, Shannon diversity index, and median landscape perimeter-to-area ratio. Climate variables include average temperature, accumulated temperature above 10°C, average rainfall, humidity index, and normalized vegetation index.

6. The method for calculating river nitrogen monoxide emissions based on land-river-atmosphere simulation according to claim 1, characterized in that, Step 3: Obtain the river hydrological parameters for each sub-basin. The specific process is as follows: Climate data from each sub-basin of each region are input into the SWAT model to simulate hydrological parameters, and the river hydrological parameters of each sub-basin are output. The climate data includes rainfall, temperature, wind speed, relative humidity and solar radiation data.

7. The method for calculating river nitrogen monoxide emissions based on land-river-atmosphere simulation according to claim 1, characterized in that, When n is 2 / 3 W 10 The wind speed at a height of 10m.

8. The method for calculating river nitrogen monoxide emissions based on land-river-atmosphere simulation according to claim 1, characterized in that, When n is 1 / 2, k 600 = 1.0 + 1.719 × (V / H) 0.5 )+2.58×W 10 W 10 The wind speed is at a height of 10m, and V and H are the flow velocity and water depth, respectively.

Images