A Dynamic Correction Method for Urban CO2 Concentration Based on Ground-Based Observations

Abstract

The application discloses a kind of urban CO2 concentration dynamic correction method based on ground observation, belong to greenhouse gas monitoring and environmental simulation technical field.The method is based on the simulation result of urban CO2 concentration generated by chemical transport model, introduces ground photoacoustic spectroscopy continuous observation data, by constructing dynamic correction factor field that changes with time and space, the output of model is adaptively modified every hour.The method includes model input refinement, photoacoustic observation data acquisition and quality control, dynamic correction factor construction, time smoothing constraint, spatial interpolation generates continuous correction field and simulation result dynamic correction and the like steps.The method can effectively correct systematic deviation generated by model under the condition of complex urban underlying surface and rapid boundary layer evolution without changing the physical structure of mode.

Description

Technical Field

[0001] This invention belongs to the field of greenhouse gas monitoring and environmental simulation technology, specifically involving a dynamic correction method for urban CO2 concentration based on ground-based observation, which is applicable to the real-time optimization and deviation correction of urban carbon emission simulation results. Background Technology

[0002] Carbon dioxide (CO2) is one of the main greenhouse gases contributing to global warming. Although urban areas account for only about 2% of the Earth's land area, they contribute more than 70% of fossil fuel emissions. Current urban CO2 emission monitoring and assessment mainly rely on two technical approaches: (1) numerical simulation methods based on emission inventories and meteorological drivers, such as the WRF-Chem model; and (2) inversion and assimilation methods based on satellite or ground-based observations.

[0003] Currently, urban-scale CO2 emission simulations mainly rely on chemical transport models such as WRF-Chem, which calculate concentration distribution by inputting emission inventories, meteorological fields, and surface parameters. However, due to the limited resolution of emission inventories, the complexity of the underlying surface, and the significant influence of urban heat islands on the boundary layer structure, model outputs often deviate from actual observations. Traditional static linear regression or empirical correction methods cannot effectively capture short-term dynamic changes in urban emissions, resulting in insufficient accuracy in simulation results.

[0004] Ground-based photoacoustic spectroscopy (PAS) offers advantages such as high sensitivity, high temporal resolution, and no need for preprocessing, enabling continuous automatic monitoring of CO2. If photoacoustic observation data can be coupled with numerical model results, and observation-driven dynamic correction algorithms can be used to adjust simulation errors in real time, the model's ability to characterize the spatiotemporal variations of urban CO2 can be significantly improved. However, currently, there is a lack of a comprehensive system of dynamic correction methods for urban CO2 simulations based on photoacoustic observations, making it difficult to achieve a precise and integrated coupling between the model and observations. Summary of the Invention

[0005] To address the aforementioned technical challenges, this invention provides a dynamic correction method for urban CO2 concentration based on ground-based observations. Addressing the complex distribution of urban emission sources, frequent concentration changes, and strong time-varying model biases, it proposes an integrated "observation-driven-model correction" approach to achieve hourly dynamic optimization and accuracy improvement of CO2 simulation results. This method combines the urban CO2 concentration field simulated by a chemical transport model with ground-based photoacoustic spectral observation data. Through steps such as input optimization, bias calculation, dynamic correction factor construction, and spatial interpolation, a temporally continuous and spatially smooth correction field is formed, thereby dynamically correcting the model output and achieving high-precision spatiotemporal simulation of urban CO2 concentration.

[0006] The technical solution of the present invention is as follows:

[0007] A method for dynamic correction of urban CO2 concentration based on ground-based observations includes the following steps:

[0008] S1: Input the urban CO2 emission inventory and surface parameters into the chemical transport model for processing to obtain the background simulation field of CO2 concentration;

[0009] S2: Acquire continuous CO2 observation data from the ground-based photoacoustic spectroscopy observation system, and calibrate and perform quality control on the observation data;

[0010] S3: Calculate the hourly dynamic correction factor based on the observed data and the CO2 concentration background simulation field, and apply time smoothing constraints to the correction factor;

[0011] S4: Spatial interpolation is performed on the time-smoothed correction factor to generate a continuous spatiotemporal correction factor field;

[0012] S5: Apply the correction factor field to the model simulation results to dynamically correct the urban CO2 concentration hourly.

[0013] In the above technical solution, the specific steps of step S1 are as follows:

[0014] Step S11: Integrate energy consumption, traffic activity, point of interest distribution and nighttime light remote sensing data to construct an urban CO2 emission inventory with hourly and kilometer-level resolution;

[0015] Step S12: Replace the model's default land type parameters with high-resolution land cover data, and update albedo, surface roughness, and vegetation cover.

[0016] Step S13: Input the optimized emission inventory and surface parameters into the chemical transport model to obtain the simulated background field of urban CO2 concentration.

[0017] In the above technical solution, the specific steps of step S2 are as follows:

[0018] Step S21: Obtain minute-by-minute concentration observation data using a ground-based CO2 monitoring system based on photoacoustic spectroscopy.

[0019] Step S22: The monitoring system is calibrated using CO2 standard gas with multiple concentration gradients. The measurement accuracy of the equipment is evaluated based on the calibration results. If there is a deviation in the equipment measurement, the correction coefficient obtained from the standard gas calibration is used to quantitatively correct the actual measurement results of the monitoring system to ensure the accuracy of the monitoring data.

[0020] Step S23: Perform outlier removal and time averaging on the observed data to obtain a stable high temporal resolution CO2 concentration sequence.

[0021] The specific steps of step S3 are as follows:

[0022] Step S31: Calculate the simulation deviation for each observation point at time t. ;

[0023] Step S32: Define the dynamic correction factor ;

[0024] Step S33: Use a sliding time window to smooth the correction factor over time to suppress short-term observation noise;

[0025] Step S34: Generate a time series of dynamic correction factors that are updated hourly.

[0026] In the above technical solution, the specific steps of step S4 are as follows:

[0027] Step S41: Generate a preliminary spatial correction factor distribution using the inverse distance weighted interpolation method;

[0028] Step S42: Spatial smoothing of the initial distribution is performed using the Kriging interpolation method based on the spatial semivariance model to obtain the continuous correction factor field. .

[0029] In the above technical solution, the specific steps of step S5 are as follows:

[0030] Step S51: According to the formula Dynamically correct the simulation results of the model;

[0031] Step S52: Apply spatial continuity and temporal constraints to the corrected results to prevent the spread of local anomalies;

[0032] Step S53: Output the hourly updated dynamic correction distribution of urban CO2 concentration.

[0033] The present invention has the following beneficial effects:

[0034] 1. It can significantly reduce the systematic bias in urban CO2 simulations without changing the model structure;

[0035] 2. The correction factor is dynamically updated over time and space, which can effectively reflect emission activities and boundary layer changes;

[0036] 3. Take into account both local characteristics and spatial continuity to avoid spatial irrationality caused by single-point correction;

[0037] 4. The method is highly versatile and can be extended to various greenhouse gas simulation application scenarios. Attached Figure Description

[0038] Figure 1This is a flowchart illustrating the overall technical process of the method of the present invention.

[0039] Figure 2 A diagram showing the laboratory calibration results of the ground-based monitoring equipment;

[0040] Figure 3 This is a graph showing measurement data from ground-based monitoring equipment at urban monitoring points;

[0041] Figure 4 The calculated spatial correction factor field map of urban CO2 concentration;

[0042] Figure 5 This is a graph used to verify the accuracy of the calibration results. Detailed Implementation

[0043] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that this embodiment is only used to illustrate the present invention and is not intended to limit the scope of protection of the present invention. Equivalent modifications or substitutions made to this embodiment by those skilled in the art without departing from the spirit of the present invention should all fall within the scope of protection of the present invention.

[0044] Example:

[0045] This embodiment provides a method for dynamic correction of urban CO2 concentration based on ground-based observations, such as... Figure 1 As shown, this method is applicable to real-time optimization and bias correction of greenhouse gas emission simulation results at the urban scale. It combines numerical simulation results from a chemical transport model with ground-based photoacoustic spectral observation data. Through steps such as input data optimization, dynamic correction factor construction, spatial interpolation, and model output correction, it achieves hourly dynamic correction of the urban CO2 concentration field. The chemical transport model adopted is the WRF-Chem model, and its specific implementation steps are as follows:

[0046] S1: Input the urban CO2 emission inventory and surface parameters into the chemical transport model for processing to obtain the simulated background CO2 concentration field. Specific steps are as follows:

[0047] Step S11: Integrate energy consumption, traffic activity, point of interest distribution and nighttime light remote sensing data to construct an urban CO2 emission inventory with hourly and kilometer-level resolution;

[0048] Step S12: Replace the default land type parameters of the WRF-CHEM model with high-resolution land cover data, and update the albedo, surface roughness and vegetation cover.

[0049] Step S13: Input the optimized emission inventory and surface parameters into the WRF-CHEM model for simulation calculation to obtain the simulated background field of urban CO2 concentration.

[0050] First, the input data of the numerical simulation model was refined and optimized to improve the reliability of the basic simulation results. Specifically, multi-source information, including energy consumption statistics, traffic flow information, point-of-interest (POI) distribution data, and nighttime light remote sensing data of the study area, was acquired. Statistical regression and spatial weighting methods were used to spatially allocate different types of emission sources, constructing an urban CO2 emission inventory with hourly temporal resolution and 1km spatial resolution. Simultaneously, high-resolution land cover data was used to replace the model's default land surface classification scheme, and key parameters such as albedo, surface roughness, and vegetation cover were recalculated, thereby improving the model's ability to characterize the urban underlying surface and boundary layer structure. After updating the emission inventory and surface parameters, the optimized input data was imported into the WRF-Chem model to obtain the initial simulated CO2 concentration field of the urban area.

[0051] S2: Acquire continuous CO2 observation data from the ground-based photoacoustic spectroscopy observation system, and calibrate and quality control the observation data. Specific steps are as follows:

[0052] Step S21: Obtain minute-by-minute concentration observation data using a ground-based CO2 monitoring system based on photoacoustic spectroscopy.

[0053] Step S22: The monitoring system is calibrated using CO2 standard gas with multiple concentration gradients. The measurement accuracy of the equipment is evaluated based on the calibration results. If there is a deviation in the equipment measurement, the actual measurement results of the monitoring system are quantitatively corrected using the correction coefficient obtained from the standard gas calibration to ensure the accuracy of the monitoring data.

[0054] Step S23: Perform outlier removal and time averaging on the observed data to obtain a stable high temporal resolution CO2 concentration sequence.

[0055] While obtaining model simulation results, a ground-based CO2 monitoring system based on photoacoustic spectroscopy was used to continuously monitor the study area, acquiring minute-by-minute CO2 concentration data. This monitoring system features high sensitivity, fast response speed, and long-term operational stability, providing a high-precision concentration benchmark for model calibration. To ensure the accuracy of the observation data, the system was calibrated under laboratory conditions using standard gases of known concentrations to verify its range linearity and long-term stability. The calibration results are as follows: Figure 2 As shown in the figure. Subsequently, the raw observation data were time-averaged and outliers were removed to form a continuous, smooth, and representative CO2 concentration time series. Monitoring data are shown below. Figure 3 As shown.

[0056] S3: Calculate the hourly dynamic correction factor based on the observed data and the simulated CO2 concentration background, and apply time-smoothing constraints to the correction factor. The specific steps are as follows:

[0057] Step S31: Calculate the simulation deviation for each observation point at time t. ;in It is the first The observed CO2 concentration of a sample at time t. It is the first The simulated CO2 concentration of a sample at time t. It is the first The difference between observed and simulated CO2 concentrations for a sample at time t.

[0058] Step S32: Define the dynamic correction factor ;in It is the first CO2 concentration correction factor for each sample at time t.

[0059] Step S33: Use a sliding time window to smooth the correction factor over time to suppress short-term observation noise.

[0060] Step S34: Generate a time series of dynamic correction factors that are updated hourly.

[0061] By comparing and analyzing ground-based photoacoustic observation results with numerical model simulation results, a dynamically changing correction factor is constructed. For each observation station, the deviation between the model simulation value and the observed value is calculated, and the correction factor is defined as the ratio of the observed value to the simulated value to characterize the relative change of the model deviation. To suppress abnormal fluctuations caused by observation noise or local disturbances on short time scales, a sliding time window is introduced into the correction factor for smoothing, thereby obtaining a continuously changing dynamic correction factor sequence over time. This correction factor can reflect the evolution characteristics of model deviation caused by urban emissions activities and changes in meteorological conditions.

[0062] S4: Spatial interpolation is performed on the time-smoothed correction factor to generate a continuous spatiotemporal correction factor field. The specific steps are as follows:

[0063] Step S41: Generate a preliminary spatial correction factor distribution using the inverse distance weighted interpolation method.

[0064] Specifically, based on the spatial correction factor sample values ​​at known locations, a weighted calculation is performed according to the spatial distance between the location to be estimated and each known sample point. Sample points that are closer to each other have a greater influence weight on the location to be estimated. The mathematical expression is: ,in, Indicates the location to be estimated Spatial correction factor value at, Indicates the first Spatial correction factor values ​​at known sample points The spatial distance between the location to be estimated and the sample point. The distance decay index is used to characterize the characteristic that the influence of a sample point gradually weakens as the distance increases.

[0065] Step S42: Spatial smoothing of the initial distribution is performed using the Kriging interpolation method based on the spatial semivariance model to obtain the continuous correction factor field. .

[0066] Specifically, based on the preliminary spatial correction factor distribution obtained in step S41, a spatial random field model is constructed, and the correlation between the correction factors at different spatial locations is characterized by a spatial semivariance function, which is defined as: ;in, It is a spatial displacement vector. Indicates having a displacement vector Number of sample point pairs Indicates position The initial spatial correction factor value at the location.

[0067] Based on this, using a pre-defined or fitted spatial semivariance model, the correction factor value at the estimated location is calculated by solving the Kriging interpolation equations. Its expression is as follows: ;in, To be estimated location The relevant kriging weights are determined by the linear unbiased optimal estimation condition under the constraints of the spatial semivariance model. Through the above kriging interpolation calculations, the preliminary distribution results are spatially smoothed to obtain a continuous spatial correction factor field. .

[0068] The dynamic correction factors obtained from discrete observation stations are extended into a spatially continuous correction factor field. First, an inverse distance weighted interpolation method is used to spatially extend the correction factors to preserve the influence of local emission characteristics on the correction results. Then, ordinary kriging interpolation based on a semi-variance model is introduced to spatially smooth the correction results, obtaining a continuously distributed correction factor field across the entire study area. This correction factor field exhibits good spatial continuity and statistical stability and can be directly used to correct numerical simulation results. The spatial correction factor field is shown below. Figure 4 As shown.

[0069] S5: Apply the correction factor field to the model simulation results to dynamically correct the urban CO2 concentration hourly. The specific steps are as follows:

[0070] Step S51: According to the formula The simulation results of the model are dynamically corrected, whereby... Indicates the model at position and time The original simulation results are output below. This represents the model simulation results after dynamic correction by the spatial correction factor. This is the correction factor field that varies with spatial location and time.

[0071] The above correction process is implemented in the model output stage or post-processing stage. According to the preset time step and spatial resolution, the model simulation results are adaptively adjusted time-by-time and grid-by-grid point, so that the correction coefficients can be dynamically updated as the spatiotemporal conditions change, thereby realizing the dynamic correction of the model simulation deviation.

[0072] Step S52: Apply spatial continuity and temporal constraints to the corrected results to prevent the spread of local anomalies. Specifically, in the spatial dimension, the gradient changes between adjacent spatial grid points are restricted to ensure continuous changes in the corrected results at adjacent locations. In the temporal dimension, the magnitude of changes in the corrected results within adjacent time steps is constrained to ensure smooth evolution of the corrected results over time. By jointly applying the above spatial continuity and temporal constraints, stable control of the corrected results is achieved, preventing the unconstrained propagation of local anomalies in the spatial or temporal dimensions, thereby improving the overall consistency and reliability of the corrected results.

[0073] Step S53: Output the hourly updated dynamic correction distribution of urban CO2 concentration.

[0074] The constructed dynamic correction factor field is applied to the CO2 concentration field output by the WRF-Chem model, and the simulation results are dynamically corrected by updating hourly to obtain the corrected CO2 concentration distribution. To avoid the propagation of local outliers in space or time, spatial smoothing and temporal constraints are applied to the correction results to ensure the continuity and physical rationality of the corrected concentration field. The dynamically corrected CO2 simulation results can more accurately reflect the spatiotemporal variation characteristics of CO2 concentration in urban areas and have higher consistency with ground-based observation data. Validation results are as follows: Figure 5 As shown.

Claims

1. A method for dynamic correction of urban CO2 concentration based on ground-based observation, characterized in that, Includes the following steps: S1: Input the urban CO2 emission inventory and surface parameters into the chemical transport model for processing to obtain the background simulation field of CO2 concentration; S2: Acquire continuous CO2 observation data from the ground-based photoacoustic spectroscopy observation system, and calibrate and perform quality control on the observation data; S3: Calculate the hourly dynamic correction factor based on the observed data and the CO2 concentration background simulation field, and apply time smoothing constraints to the correction factor; S4: Spatial interpolation is performed on the time-smoothed correction factor to generate a continuous spatiotemporal correction factor field; S5: Apply the correction factor field to the model simulation results to dynamically correct the urban CO2 concentration hourly. Step S1 includes: S11: Integrate energy consumption, traffic activity, point of interest distribution and nighttime light remote sensing data to construct an urban CO2 emission inventory with hourly and kilometer-level resolutions; S12: Replace the model's default land type parameters with high-resolution land cover data, and update albedo, land roughness, and vegetation cover; S13: Input the optimized emission inventory and surface parameters into the chemical transport model to obtain the simulated background field of urban CO2 concentration; Step S2 includes: S21: Obtain minute-by-minute concentration observation data using a ground-based CO2 monitoring system based on photoacoustic spectroscopy. S22: The monitoring system is calibrated using CO2 standard concentration gas with multiple concentration gradients to verify its measurement accuracy and stability; S23: Perform outlier removal and time averaging on the observed data to obtain the CO2 concentration sequence; Step S3 includes: S31: Calculate the simulation deviation at time t for each observation point. ;in It is the first The observed CO2 concentration of a sample at time t. It is the first The simulated CO2 concentration of a sample at time t. It is the first The difference between observed and simulated CO2 concentrations for a sample at time t; S32: Define the dynamic correction factor ;in It is the first CO2 concentration correction factor for each sample at time t; S33: The correction factor is smoothed over time using a sliding time window to suppress short-term observation noise; S34: Generate a dynamic correction factor time series that is updated hourly; Step S4 includes: S41: Generate the preliminary spatial correction factor distribution using the inverse distance weighted interpolation method; S42: The Kriging interpolation method based on the spatial semivariance model is used to spatially smooth the initial distribution to obtain a continuous correction factor field. .

2. The method according to claim 1, characterized in that, Step S5 includes: S51: According to the formula Dynamically correct the simulation results of the model; Indicates the model at position and time The original simulation results are output below. This represents the model simulation results after dynamic correction by the spatial correction factor. This is a correction factor field that varies with spatial location and time; S52: Apply spatial continuity and temporal constraints to the corrected results to prevent the spread of local anomalies; S53: Outputs the hourly updated dynamic correction distribution of urban CO2 concentration.

3. The method according to claim 1, characterized in that, In step S33, the width of the sliding time window is 30-90 minutes.

4. The method according to claim 1, characterized in that, The semivariance model described in step S42 adopts a Gaussian model or an exponential model.

5. The method according to claim 1, characterized in that, This method is applicable to the real-time optimization of greenhouse gas simulation results at the urban scale.

6. The method according to claim 5, characterized in that, It can also be used for dynamic correction of CH4 and N2O gases.

Images