A method for optimizing supersaturation in fog

Abstract

The present invention relates to a method for optimizing supersaturation in fog, including: using cubic spline interpolation, a preprocessing method, to smooth and refine the logarithmic concentration before reverse integration, which can effectively smooth the data and reduce the error of numerical integration; this method estimates supersaturation through reverse integration and temperature dynamic compensation, and introduces the actual activation rate and the theoretical activation rate for error analysis to determine the optimal integration threshold, so as to reduce the error caused by improper setting of the integration threshold in the traditional method, and finally realize the optimization of the supersaturation estimation method. This method solves the problems of isolated data, error accumulation and insufficient environmental adaptability in the prior art, can reduce the gap between model simulation and measurement results, reduce the uncertainty of aerosol-cloud interaction simulation, and at the same time improve the parameterization process of microphysical processes.

Description

Technical Field

[0001] The invention relates to the field of information technology, and in particular to a method for optimizing supersaturation in fog based on κ-Köhler theory. Background Art

[0002] Accurate estimation of supersaturation in fog is a key problem in studying the formation mechanism of fog and improving fog forecast models. The existing technology has significant deficiencies in estimation accuracy and is difficult to accurately reflect the microscopic physical changes in the growth process of droplets. The technical problem is how to effectively integrate and analyze complex observation data, including droplet spectra, aerosol spectra, and cloud condensation nucleus spectra. The technical difficulties faced in the data processing process include: how to match and interpolate observation data of different time scales to ensure the continuity and consistency of the data; how to select a suitable integration threshold in the reverse integration process to balance the computational efficiency and estimation accuracy; how to accurately evaluate the activation process of cloud condensation nuclei and effectively associate it with the supersaturation estimation results. In addition, since the formation and evolution of fog are affected by many factors, how to consider the dynamic changes of environmental conditions in the estimation process and how to evaluate and reduce the uncertainty of the estimation results are technical problems that need to be solved urgently. The solution to these problems is not only related to the accuracy of supersaturation estimation, but also directly affects our understanding of the microscopic process of fog formation and the improvement of our prediction ability. Summary of the invention

[0003] The present invention provides a method for optimizing supersaturation in fog, which mainly comprises the following steps:

[0004] S101: Obtain observation data of fog droplet spectrum, aerosol spectrum and cloud condensation nucleus spectrum during fog process in the studied area, as well as raindrop spectrum observation data during the same study period;

[0005] S102: determining the fog process according to the fog droplet spectrum and raindrop spectrum data;

[0006] S103: performing cubic spline interpolation on the droplet spectrum data, the interpolated particle sizes are 2.5 μm, 3 μm and 3.5 μm respectively, and negative values ​​are processed in different cases;

[0007] S104: unifying the time of the aerosol spectrum and the droplet spectrum, extracting the common time interval, and accumulating the droplet spectrum data according to the time window;

[0008] S105: converting the original counts into number concentration distribution according to the sampling volume and the particle size interval width, and directly reading the spectrum data of the aerosol spectrum as aerosol spectrum data;

[0009] S106: setting key parameters of the supersaturation calculation model, calculating supersaturation using the reverse integration method and the dynamic threshold correction method, and performing abnormal processing and result optimization;

[0010] S107: Obtain the time series of cloud condensation nuclei number concentration under the estimated supersaturation, calculate the actual activation rate of cloud condensation nuclei under the actual supersaturation, and perform a significance test;

[0011] S108: Find the wet critical diameter corresponding to each supersaturation, calculate the activation critical diameter corresponding to each wet critical diameter, regard the particles with a particle size larger than the activation critical diameter as activated particles, integrate the activated particles, and calculate the theoretical activation rates corresponding to different supersaturations; S109: Calculate the cumulative relative error, comprehensively consider the two actual situations of the smaller cumulative relative error and the increase of interpolation stability with the interpolation particle size, determine the optimal integral threshold, and obtain the supersaturation corresponding to the threshold.

[0012] Furthermore, the criterion for determining a fog process is that the average number concentration of fog droplets is greater than or equal to 10 cm -3 And the average liquid water content is greater than or equal to 0.001 gm -3 ;

[0013] At the same time, the raindrop spectra with a number of less than 10 and a rainfall intensity less than 0.002 mm h were removed. -1 If the precipitation rate after elimination is not 0, it is considered as a precipitation process, and the fog process containing precipitation period is eliminated.

[0014] Furthermore, negative values ​​are handled in different situations, including:

[0015] For the number concentration after treatment, if there is a negative value and it is located at -5cm -3 -0 cm -3 If the interval is 0, the difference between the two numerical concentrations is considered small and can be approximately corrected to zero.

[0016] If the negative value exceeds -5cm -3 -0 cm -3 interval, it is considered that the interpolation function makes opposite judgments on the changing trends of the two-level number concentrations, and the opposite of the negative value is taken as the correction value.

[0017] Furthermore, basic observation data is obtained, including:

[0018] Obtain the dry aerosol particle size distribution data and corresponding timestamp of the aerosol spectrum, with the particle size range of 10nm-10000nm;

[0019] Read the droplet spectrum data and timestamp of the droplet spectrum, with a particle size range of 2μm - 50μm. The droplet spectrum data is multiplied by 1000 to convert to nanometer units, and the aerosol spectrum is matched with the droplet spectrum data through time alignment.

[0020] Preferably, in step S104, the timestamps of the aerosol spectrum and the droplet spectrum are unified to hour precision (time intersection range: from 22:19:01 on January 3, 2024 to 10:19:01 on January 4, 2024, a total of 11 time points), and the common time interval is extracted. The droplet spectrum data is accumulated according to a time window (the previous hour or adjacent time intervals), and the time step is unified.

[0021] Preferably, in step S105, the droplet number concentration is calculated according to the sampling volume and the particle size interval width, and the original count is converted into a number concentration distribution. The calculation steps are as follows:

[0022] First, calculate the sampling volume Volume according to the sampling area of the droplet spectrum and the air flow velocity PAS:

[0023] (1)

[0024] In the formula, Volume is the sampling volume, PAS is the air flow velocity (unit: m / s), t is the unit time (usually taken as 1s), and S a is the sampling area (S a = 0.41 mm 2 )

[0025] Secondly, calculate the droplet number concentration per unit volume (cm -3 ) according to the original statistical number of the droplet spectrum:

[0026] (2)

[0027] In the formula, bin nc is the droplet number concentration per unit volume, bin n is the total droplet number concentration of the droplet spectrum in the sampling volume, and Volume is the sampling volume.

[0028] Preferably, the setting of the key parameters of the model includes: the hygroscopicity parameter κ is calculated from the observed data, the temperature T is obtained in real time through the droplet spectrum data, and the droplet surface tension parameter A is calculated. The reverse integration method mainly finds an interval that makes the integrated total concentration of the activated aerosol closest to the actual measured concentration (droplet spectrum) by reverse integrating the aerosol spectrum, and obtains the activated critical diameter D a when the two are equal by interpolation, and then reversely deduces the wet critical diameter D dc of the droplet. The formula for the supersaturation SS after reverse integration is:

[0029] (3a)

[0030] Among them, A is the droplet surface tension parameter, and its calculation formula is:

[0031] (3b)

[0032] In the formula, is the droplet surface tension coefficient (assumed to be the surface tension of pure water at ambient temperature), is the molar mass of water, R is the gas constant (R is approximately equal to 8.314 J / (mol·K)), T is the atmospheric ambient temperature, is the density of liquid water.

[0033] The dynamic threshold correction method distinguishes activated and non-activated particles by introducing an integral threshold, and adjusts the supersaturation SS value by iteratively matching the droplet number concentration. Combining piecewise linear interpolation, it optimizes the SS calculation accuracy under low supersaturation conditions. The part of anomaly handling and result optimization includes: eliminating invalid values (such as infinite values in the estimation results), limiting the SS range (0.02% - 0.1%), and ensuring that the results conform to the actual atmospheric conditions.

[0034] Preferably, the formula for calculating the actual activation rate of cloud condensation nuclei at the actual supersaturation in step S107 is as follows:

[0035] (4)

[0036] In the formula, AR r is the actual activation rate, the cloud condensation nuclei number concentration N CCN can be fitted based on the time series of N CCN under a certain set of SS to obtain the total concentration at a specific SS. N d is the total aerosol number concentration, which can first calculate the hourly average within the corresponding time range and then integrate over time.

[0037] The cloud condensation nuclei number concentration N CCN The fitting formula is as follows:

[0038] , (5)

[0039] Both C and k in the formula are fitting coefficients.

[0040] The significance test of the exponential regression equation for the total cloud condensation nuclei concentration can use the F test. For the index F, there is:

[0041] (6)

[0042] It follows the degrees of freedom of of distribution. Compare the F value with the in the F distribution table. If , it is considered that at the significance level, the regression equation of the total number concentration of cloud condensation nuclei is significant, and the exponential fitting method can be used to fit the number concentration of cloud condensation nuclei. If the F test is unqualified, it is considered that there are fewer data available for fitting, resulting in a large fitting error, and more measurement data of cloud condensation nuclei under different set supersaturations SS should be considered.

[0043] Preferably, in step S108, the relationship between the supersaturation SS and D dc is:

[0044] (7);

[0045] A in the formula is shown in formula (3b).

[0046] D dc and D a is:

[0047] (8),

[0048] Particles with a particle size greater than D a are regarded as activated particles, and the total number of activated particles N act is obtained by integrating these activated particles. Calculate the activation rate corresponding to different supersaturations SS, and thus the relationship between the activation rate AR and the supersaturation SS can be found. The formula is:

[0049] (9).

[0050] Preferably, the calculation of the relative error in step S109 is carried out on the premise of assuming the theoretical activation rate as the true value.

[0051] The present invention has the following beneficial effects compared with the prior art:

[0052] 1) Based on objective observation data and strict reliability analysis, the present invention directly correlates the aerosol spectrum with the fog droplet number concentration by using the backward integration method, avoiding the cumulative error of traditional forward integration. By dynamically adjusting the integration threshold to distinguish activated / unactivated particles, the calculation accuracy under low supersaturation conditions is significantly improved, and a method based on the κ-Köhler theory is proposed for estimating the supersaturation in fog, which has a solid theoretical basis;

[0053] 2) The present invention introduces the cubic spline interpolation method to refine and smooth the data of the fog droplet number concentration, further improving the accuracy of backward integration, reducing the error caused by the irregularity of the original number concentration data, and improving the accuracy of estimating the supersaturation;

[0054] 3) The present invention performs error analysis through the activation rate, and finally determines the optimal integration threshold, which is closely connected to the actual physical process, improving the accuracy and reliability of the research. At the same time, a significance test is carried out on the fitting of the cloud condensation nucleus concentration index, making the results highly credible;

[0055] 4) According to the estimation method of the present invention, it is easy to operate, can improve the accuracy, and has a positive effect on the research of estimating the supersaturation in fog;

[0056] 5) The present invention realizes the accurate calculation of aerosol supersaturation through multi-source data fusion, reverse integration algorithm and temperature dynamic compensation, solves the problems of data isolation, error accumulation and insufficient environmental adaptability in the prior art, and provides an efficient and reliable tool for atmospheric cloud and fog research. Description of the Drawings

[0057] Figure 1 is the overall technical route of this method;

[0058] Figure 2 is the comparison chart of the dynamic threshold correction effect;

[0059] Figure 3 is the normalized time series of the fog droplet number concentration at particle sizes of 2.5 μm, 3 μm, and 3.5 μm obtained by using the spline interpolation method;

[0060] Figure 4 are the theoretical activation rate and actual activation rate values of the 1st to 7th fog processes under the supersaturation corresponding to five integration thresholds (2 μm, 2.5 μm, 3 μm, 3.5 μm, and 4 μm);

[0061] Figure 5 is the variation of the supersaturation with time during seven fog processes where the supersaturation can be estimated. Detailed Embodiments

[0062] The present invention will be further described below in conjunction with specific embodiments and the corresponding drawings.

[0063] The observation instruments used in this embodiment mainly include a fog droplet spectrometer (Fog Monitor model 120, FM-120) produced by DMT company, a Scanning Mobility Particle Sizer (SMPS), and a Continuous-Flow Streamwise Thermal-Gradient CCN Counter (CCNc). The raindrop spectrum used to determine the precipitation process is measured by a raindrop spectrometer (OTT Particle Size Velocity (Parsivel) disdrometer) produced by OTT Hydromet company in Germany.

[0064] The FM-120 mainly obtains the fog droplet spectrum by capturing the scattered light emitted when particles pass through the laser beam to count the particle size and number. The particle size measurement range of this instrument is 2 μm to 50 μm, divided into 30 bins, and the sampling frequency is 1 Hz. The SMPS is mainly used to obtain the dry aerosol spectrum (PNSD). This instrument consists of an Electrostatic Classifier (EC, model 3082) and a Condensation Particle Counter (CPC, model 3776). The Differential Mobility Analyzer (DMA, TSI company, model 3081) used by the EC has a measurement range of 11.3 to 532 nm. The time resolution of the SMPS is 5 min, that is, it completes a scan of the aerosol spectrum every 5 min. The data of the cloud condensation nuclei number CCN is measured by particle size. The set supersaturation SS values are 0.07%, 0.1%, 0.2%, 0.4%, and 0.8% respectively. The measurement time for the first set supersaturation SS is 20 min (it takes a longer stabilization time for the supersaturation SS to switch from 0.8% to 0.07%), and the measurement times for the other four set supersaturation SS are 10 min, and it cycles once every 1 h.

[0065] According to the measurement time range of the fog droplet spectrum data and the aerosol spectrum data, the data observed at the Shanghai Fudan site from November 27, 2023 to February 29, 2024 are selected for the statistics of relevant physical quantities, with a total of 8 fog processes. According to the temperature record of the fog droplet spectrum, the optimal integration threshold is determined to be 3.5 μm by using this method, and the supersaturation of 7 fog processes is obtained (the supersaturation of one fog process cannot be estimated due to the missing measurement of the aerosol spectrum data).

[0066] In the low supersaturation region, there is a difference in supersaturation between before and after dynamic threshold correction at the initial moment, and they are the same in the remaining regions ( Figure 2 , showing that during the fifth fog process, the supersaturation SS1 without dynamic threshold correction and the corrected supersaturation SS1 2 in the low supersaturation region). Taking the fifth fog process as an example, we standardized the droplet number concentration in different particle size ranges using the cubic spline interpolation method ( Figure 3 , Figure 3 in (a) is the standardized time series of droplet number concentration at a particle size of 2.5 μm obtained by using the spline interpolation method, Figure 3 in (b) is the standardized time series of droplet number concentration at a particle size of 3 μm obtained by using the spline interpolation method, Figure 3 in (c) is the standardized time series of droplet number concentration at a particle size of 3.5 μm obtained by using the spline interpolation method. The orange thick solid line represents the change of the standardized number concentration of droplets corresponding to the interpolation particle size at different times, and the blue solid line and green dashed line represent the change of the standardized number concentration of droplets with particle sizes of 2 and 4 μm respectively). It can be seen that as the interpolation particle size (orange line) increases, the consistency between the change trend of the interpolated number concentration and the original data trend gradually improves, that is, the interpolation effect gradually gets better. When the interpolation particle size is 3.5 μm, the interpolation effect is relatively the best. When calculating the actual activation rate, the exponential fitting equation of the CCN number concentration passed the significance test, and the determination coefficient R 2 is 0.98. By comparing the activation rate errors, from Figure 4 ( Figure 4 in (a) - (g) are the theoretical activation rate and actual activation rate values of the 1st - 7th fog processes under the supersaturations corresponding to five integration thresholds (2 μm, 2.5 μm, 3 μm, 3.5 μm, and 4 μm). The horizontal axis represents the supersaturations at integration thresholds of 2 μm, 4 μm, 2.5 μm, 3 μm, and 3.5 μm from left to right in sequence. AR t represents the theoretical activation rate at a specific supersaturation, and different colors represent different aerosol spectrum integration lower limits. The error bars represent the difference between the theoretical activation rate and the actual activation rate), it can be seen that relatively small errors mainly appear in the intervals corresponding to the latter three new thresholds (2.5 μm, 3 μm, and 3.5 μm). Further calculating the cumulative relative error (CRE), the CRE is the smallest at 10.61 when the integration threshold is 2.5 μm, and the second smallest at 15.09 for 3.5 μm, but the difference between the two is small. The CRE under other integration thresholds is greater than 30, and the CRE under the traditional 2 μm threshold even reaches 210.37. Considering the interpolation stability and error analysis results comprehensively, it is determined that 3.5 μm is the best integration threshold. The SS time series in all cases estimated by this threshold is as shown in Figure 5As shown, the supersaturation values of all processes vary within the low supersaturation range of the natural environment.

[0067] The supersaturation estimation method includes the following specific steps:

[0068] (1) Obtain the observational data of the fog droplet spectrum, aerosol spectrum, and cloud condensation nucleus spectrum.

[0069] (2) Perform cubic spline interpolation on the fog droplet spectrum data, with the interpolation particle sizes being 2.5 μm, 3 μm, and 3.5 μm respectively. Subtract the total number concentration value less than the particle size range from the concentration value interpolated for a certain particle size range. For the processed number concentration, if there is a negative value and it is within the range of -5 cm -3 ~ 0 cm -3 interval, it is considered that the difference in the number concentration between the two ranges is small and can be approximately corrected to 0. If the negative value exceeds this interval, it is considered that the interpolation function judges the change trend of the number concentration between the two ranges to be opposite, and the opposite value of this negative value is taken as the correction value.

[0070] (3) Unify the time of the aerosol spectrum (SMPS) and the fog droplet spectrum (FM), extract the common time interval, and accumulate the FM data according to the time window.

[0071] (4) Convert the original count into the number concentration distribution bin nc according to the sampling volume Volume and the particle size interval width, and directly read the spectral data of the SMPS as the aerosol spectrum data. Calculate the sampling volume according to the sampling area S and the air flow velocity PAS of the FM:

[0072] (1)

[0073] In the formula, Volume is the sampling volume, PAS is the air flow velocity (unit: m / s), t is the unit time (usually taken as 1 s), and S a is the sampling area (S a = 0.41 mm 2 ).

[0074] Secondly, calculate the fog droplet number concentration per unit volume (cm -3 ) according to the original statistical number of the fog droplet spectrum:

[0075] (2)

[0076] In the formula, bin nc is the fog droplet number concentration per unit volume, bin n is the total fog droplet number concentration of the fog droplet spectrum within the sampling volume, and Volume is the sampling volume.

[0077] (5) Set the key parameters of the supersaturation calculation model. Calculate the supersaturation using the backward integration method and the dynamic threshold correction method respectively, and perform anomaly handling and result optimization to ensure that the results conform to the actual atmospheric conditions. The setting of the key parameters of the model includes: the hygroscopic parameter κ is obtained by observation calculation (κ is set to 0.34 in the embodiment), the temperature T is obtained in real time through FM data, and the droplet surface tension parameter A is calculated. The backward integration method mainly finds an interval that makes the total integrated concentration of activated aerosols closest to the actual measured concentration (fog droplet spectrum) by backward integrating the aerosol spectrum, and obtains the activated critical diameter D when the two are equal by interpolation a , and back-calculate the wet critical diameter D of the fog droplet dc . The formula for calculating the supersaturation SS after backward integration is:

[0078] (3a)

[0079] where A is the droplet surface tension parameter, and its calculation formula is:

[0080] (3b)

[0081] In the formula, is the droplet surface tension coefficient (assumed to be the surface tension of pure water at the ambient temperature), is the molar mass of water, R is the gas constant (R is approximately equal to 8.314 J / (mol·K)), T is the atmospheric ambient temperature, is the density of liquid water.

[0082] (6) Calculate the CCN activation rate at the actual supersaturation:

[0083] (4).

[0084] In the formula, AR r is the actual activation rate, the cloud condensation nucleus number concentration N CCN can be fitted on the basis of the time series of N CCN at four set SSs to obtain the total concentration at a specific SS. N d is the total aerosol number concentration, which can first calculate the hourly average value within the corresponding time range and then integrate over time.

[0085] Obtain the time series of the CCN number concentration at the estimated SS through the exponential fitting method. The fitting formula is as follows:

[0086] (5),

[0087] In the formula, SS represents the supersaturation, and both C and k are fitting coefficients.

[0088] The significance test of the exponential regression equation for the total CCN concentration can use the F-test. For the index F, there is:

[0089] (6),

[0090] It follows an distribution with degrees of freedom . Compare the F value with the in the F-distribution table. If , it is considered that at the significance level, the regression equation of the total CCN concentration is significant, and the exponential fitting method can be used to fit the CCN number concentration. If the F-test fails, it is considered that there is less data available for fitting, resulting in a large fitting error, and more CCN measurement data under different set supersaturations SS should be considered.

[0091] (7) Based on the estimated supersaturation SS in step (5), according to the inverse relationship between the wet critical diameter D dc and the supersaturation SS, find the wet critical diameter D dc (unit: μm) at each supersaturation SS. The formula is:

[0092] (7).

[0093] A in the formula is shown in formula (3b).

[0094] (8) Calculate the activation critical diameter D dc corresponding to each wet critical diameter D a (unit: nm). The relationship between D dc and D a is:

[0095] (8).

[0096] Particles with a particle size greater than D a are regarded as activated particles, and the total number of activated particles N act is obtained by integrating these activated particles. Calculate the activation rate corresponding to the supersaturation SS in step (5), and thus the relationship between AR and the supersaturation SS can be found. The formula is:

[0097] (9).

[0098] Among them, the total number of aerosol particles may deviate from the normal value due to changes in the measurement environment, so the aerosol integration interval should be appropriately changed to adapt to the actual activation rate value.

[0099] (9) Calculate the cumulative relative error (CRE), which is the sum of the absolute values of the relative errors under the same integration threshold. Considering the two actual situations where the smaller CRE and the interpolation stability increase with the increase of the interpolation particle size, determine the optimal integration threshold. Then the SS corresponding to this threshold is the final result.

[0100] As described above, only the specific preferred embodiments of the present invention are provided, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes should be covered within the protection scope of the present invention.

Claims

1. A method for optimizing supersaturation in fog, characterized in that: include: Obtain observational data of fog droplet spectrum, aerosol spectrum and cloud condensation nucleus spectrum during fog process in the studied area, as well as observational data of raindrop spectrum in the same study period; Determining the fog process according to the fog drop spectrum and rain drop spectrum data; The droplet spectrum data is interpolated by cubic spline, and the interpolated particle sizes are 2.5 μm, 3 μm and 3.5 μm respectively, and negative values ​​are processed in different cases; Unifying the time of the aerosol spectrum and the droplet spectrum, extracting the common time interval, and accumulating the droplet spectrum data according to the time window; According to the sampling volume and the particle size interval width, the raw counts are converted into number concentration distribution, and the spectrum data of the aerosol spectrum is directly read as aerosol spectrum data; Set the key parameters of the supersaturation calculation model, calculate the supersaturation using the reverse integration method and the dynamic threshold correction method, and perform abnormal processing and result optimization; Obtain the time series of cloud condensation nuclei under estimated supersaturation, calculate the actual activation rate of cloud condensation nuclei under actual supersaturation, and conduct significance test; Find the wet critical diameter corresponding to each supersaturation, calculate the activation critical diameter corresponding to each wet critical diameter, regard the particles with a diameter larger than the activation critical diameter as activated particles, integrate the activated particles, and calculate the theoretical activation rates corresponding to different supersaturations; The cumulative relative error is calculated, and the optimal integral threshold is determined by combining the two actual situations: the smaller cumulative relative error and the interpolation stability increases with the interpolation particle size, and the supersaturation corresponding to the threshold is obtained; the smaller one refers to the smallest and second smallest ones among all the cumulative relative errors calculated under different integral thresholds.

2. The optimization method according to claim 1, characterized in that: The criterion for judging a fog process is that the average number concentration of fog droplets is greater than or equal to 10 cm -3 And the average liquid water content is greater than or equal to 0.001 gm -3 ; At the same time, the raindrop spectra with a number of less than 10 and a rainfall intensity less than 0.002 mm h were removed. -1 If the precipitation rate after elimination is not 0, it is considered as a precipitation process, and the fog process containing precipitation period is eliminated.

3. The optimization method according to claim 1, characterized in that: The processing of negative values ​​according to different situations includes: For the number concentration after treatment, if there is a negative value and it is located at -5cm -3 -0 cm -3 interval, then it is considered that the concentration difference between adjacent particle size bins is less than 5 cm -3 , correct negative values ​​to 0; If the negative value exceeds -5cm -3 -0 cm -3 interval, it is considered that the interpolation function makes opposite judgments on the changing trends of the two-level number concentrations, and the opposite of the negative value is taken as the correction value.

4. The optimization method according to claim 1, characterized in that: Get basic observation data, including: Obtain the dry aerosol particle size distribution data and corresponding timestamp of the aerosol spectrum, with the particle size range of 10nm-10000nm; The droplet spectrum data and timestamp of the droplet spectrum are read, and the particle size range is 2 μm-50 μm. The droplet spectrum data is multiplied by 1000 to convert into nanometer units, and the aerosol spectrum and droplet spectrum data are matched by time alignment.

5. The optimization method according to claim 1, characterized in that: Unify the time of aerosol and droplet spectra, including: Unifying the timestamps of the aerosol spectrum and the droplet spectrum to hourly precision, and extracting a common time interval; The droplet spectrum data are accumulated according to the time window, and the time step is unified.

6. The optimization method according to claim 1, characterized in that: The droplet number concentration calculation converts the original count into a number concentration distribution based on the sampling volume and particle size interval width. The calculation process is as follows: According to the droplet spectrum sampling area and air flow velocity, the sampling volume is calculated: (1) Where Volume is the sampling volume, PAS is the airflow velocity, unit: m / s, t is the unit time, S a is the sampling area; Calculate the droplet concentration bin per cubic centimeter based on the original statistical number of droplet spectra nc : (2) In the formula, bin nc is the droplet number concentration per unit volume, bin n is the total droplet number concentration of the droplet spectrum in the sampling volume, and Volume is the sampling volume.

7. The optimization method according to claim 1, characterized in that: Calculation of the actual activation rate AR of cloud condensation nuclei under actual supersaturation r The formula is as follows: (4) In the formula, is the actual activation rate, the cloud condensation nucleus number concentration N CCN N under four SS settings CCN The total concentration at a specific SS is obtained by fitting based on the time series, N d is the total aerosol concentration; Cloud condensation nucleus number concentration N CCN The exponential fitting formula is as follows: (5) C and k in the formula are fitting coefficients.

8. The optimization method according to claim 1, characterized in that: Supersaturation SS and critical diameter D dc The relationship is: (6) Where A is the droplet surface tension parameter, and SS is the supersaturation.

Images