Cloud optical thickness inversion method based on earth surface short wave irradiance

Abstract

The invention discloses a cloud optical thickness inversion method based on earth surface short wave irradiance, which comprises the following steps: 1, data preparation and pretreatment: observed radiation data need to contain two parameters of total irradiance and downlink direct irradiance, and the atmosphere is relatively clean; and 2, obtaining the solar altitude angle: directly obtaining observation time and latitude and longitude information of a station from observation or calculating the solar altitude angle by utilizing the observation time and the latitude and longitude information of the station. And 3, calculating the irradiance of the clear sky: calculating the short-wave irradiance received by the earth surface under the cloudless condition of the clear sky by adopting a McClear model, namely, the irradiance of the clear sky. And 4, calculating the cloud optical thickness: calculating and simplifying the key parameters based on the theoretical relationship between the cloud optical thickness and the short-wave irradiance of the earth surface, and giving a relational expression between the cloud optical thickness and the short-wave irradiance of the earth surface by using polynomial fitting. The method has the advantage of providing a new way for inverting the cloud optical thickness only by using the short-wave irradiance and the component of the short-wave irradiance observed on the earth surface without relying on a radar or a satellite.

Description

Technical Field

[0001] The present invention belongs to the technical field of meteorological element inversion, and in particular relates to a cloud optical thickness inversion method based on surface shortwave irradiance. Background Art

[0002] Clouds are a significant factor influencing the radiation budget of the Earth-atmosphere system. However, due to the spatiotemporal heterogeneity of cloud macroscopic properties and the complexity of their radiative effects, clouds are the greatest source of uncertainty in climate predictions. Cloud optical depth is a key parameter describing the macroscopic physical characteristics of clouds, and its changes induce a radiative forcing comparable to a doubling of carbon dioxide. Therefore, obtaining long-term, localized, and high-frequency time series of cloud optical depth can more accurately quantify the modulation of shortwave radiation by clouds, deepening our understanding of radiative transfer processes and cloud-radiation interactions. This is crucial for understanding the feedback mechanisms of clouds on climate and reducing the uncertainty of clouds in climate predictions.

[0003] Long-term observational data are fundamental to a deeper understanding of cloud radiative effects, but methods for cost-effectively retrieving long-term, high-frequency, and local cloud optical thickness are currently lacking. Among the current cloud macro-property detection systems, satellite observations are more suitable for global or regional scale studies, but their description of local high-frequency variations in cloud optical thickness is limited. Airborne experiments are only suitable for case studies. Specialized instruments such as ground-based millimeter-wave radars, microwave radiometers, or all-sky imagers can provide a detailed picture of local cloud optical thickness, but ground-based observatories equipped with these instruments are much more sparse than ground-based radiation observation networks. Currently, the world boasts extensive ground-based radiation observation networks, such as the Atmospheric Radiation Measurement (ARM) program and the Baseline Ground-Based Radiation Network (BSRN), established in the 1990s, which provide minute-level observational data. Successfully utilizing ground-based solar radiation parameters to cost-effectively retrieve long-term, local, and high-temporal-resolution cloud optical thickness time series would be a valuable complement to the existing cloud macro-parameter detection system.

[0004] The presence of clouds causes the solar radiation received by the ground and its components (global radiation, direct radiation, and diffuse radiation) to exhibit different radiative properties. Clouds typically weaken direct radiation and enhance diffuse radiation. The combined effect of these two factors causes the response of global radiation (the sum of downwelling direct and diffuse radiation) to clouds to differ from the other two components. This difference also extends to the cloud radiative forcing component (the difference between the radiation under cloudy and clear-sky conditions) corresponding to the radiation parameters. Therefore, by leveraging these differences and combining them with the characteristics of irradiance and radiative forcing, it is theoretically possible to infer macroscopic cloud parameters, such as cloud cover, based on shortwave radiation parameters at the surface.

[0005] Research has shown that cloud relative radiative forcing is a function of cloud amount and cloud albedo, which in turn are functionally related to cloud optical thickness. However, current schemes for retrieving cloud macroscopic parameters based on shortwave radiation at the surface are computationally complex. For example, the expressions for cloud albedo and cloud amount contain five complex sub-terms, and some parameters require model estimation, limiting their universality. Therefore, improving and optimizing existing schemes and constructing a cloud optical thickness retrieval scheme based on shortwave irradiance at the surface is a highly challenging scientific problem. Retrieving long-term, local, and high-frequency cloud optical thickness time series not only provides a new way to obtain cloud optical property information for meteorological, climate, and environmental monitoring fields, but also effectively supplements the existing cloud observation system. It also provides data support for studying the nonlinear response between cloud optical thickness and cloud radiative forcing, helping to improve our understanding of local climate change mechanisms. Summary of the Invention

[0006] The present invention aims to provide a cloud optical thickness inversion method based on surface shortwave irradiance, comprising the following steps:

[0007] 1. Data Preparation and Preprocessing

[0008] 1) Obtain observation data. The observed radiation data must include two parameters: total irradiance and downlink direct irradiance.

[0009] 2) Use linear interpolation method to interpolate the missing values ​​of radiation data.

[0010] 3) Eliminate total irradiance less than 10W / m 2 The following observation data are used to avoid the influence of sunrise and sunset on the inversion accuracy. The radiation data under relatively clean atmospheric conditions are selected, that is, PM 2.5 Concentration <75 μg / m 3 Or the radiation data for the period when visibility > 10km.

[0011] 2. Get the sun's altitude angle

[0012] If there is observation data of solar altitude angle or solar zenith angle in the observation data, it can be used directly. Solar altitude angle = 90° - solar zenith angle;

[0013] If there is a lack of observation data on the solar altitude angle or solar zenith angle, this technology uses the Solar Position Algorithm (SPA) in the pvlib open source package based on the Python language to calculate the solar altitude angle. This algorithm can be used to calculate the solar position (such as the solar altitude angle, solar azimuth angle) and astronomical parameters related to solar radiation (such as declination angle, hour angle, etc.). It is provided by the National Renewable Energy Laboratory (NREL) of the United States. Based on standard formulas and some correction functions in astronomy, it can calculate the solar altitude angle from 2000 BC to 6000 AD with an error of less than 0.0005°. This algorithm is widely used in solar system design, geopositioning, drone navigation and other fields. When there is a lack of observation data on the solar altitude angle, it is necessary to obtain the latitude and longitude of the site and the observation time, and then use the SPA algorithm to calculate the solar altitude angle θ a , its core formula is:

[0014] sin(θ a )=sin(φ)sin(δ)+cos(φ)cos(δ)cos(H) (1)

[0015] θ a =arccos(sin(θ a )) (2)

[0016] Where φ is the latitude, δ and H are the declination angles and hour angles, respectively, both of which are related to local time and longitude λ.

[0017] 3. Calculating Clear Sky Irradiance

[0018] The invention uses the McClear model to calculate clear-sky irradiance. This model is based entirely on physical principles and utilizes the latest findings from the Monitoring Atmospheric Composition and Climate (MACC) project on aerosol properties, total water vapor content, and ozone content. Combined with the abaci (lookup table) method and interpolation functions, it can calculate shortwave irradiance at the surface under cloudless sky conditions, including both clear-sky total irradiance and clear-sky downward direct irradiance. The calculation speed is approximately 10 times faster than the radiative transfer model libRadtran. 5 The model provides an online version (https: / / www.oie.minesparis.psl.eu / Valorisation / Outils / McClear / ), which can calculate the clear sky irradiance by inputting the latitude and longitude of the station and the observation time.

[0019] 4. Calculating Cloud Optical Depth

[0020] 4.1 Theoretical Relationship between Cloud Optical Depth and Irradiance

[0021] According to the single-layer cloud radiation transfer equation and Beer-Lambert law, the total irradiance at the surface is The following theoretical relationships exist between cloud cover, cloud albedo, cloud absorptivity, and diffuse transmittance:

[0022]

[0023] Where F represents the shortwave irradiance, and the superscripts dn and up represent the vertical downward and upward directions; is the total clear sky irradiance, which can be calculated using the McClear model in step 3; is the upward shortwave irradiance reflected by the surface; α on the right side of the equation r and α a are cloud albedo and cloud absorptivity respectively; f is cloud amount; T is diffuse transmittance.

[0024] For the downward direct irradiance at the surface It has the following relationship with cloud cover, cloud albedo and solar altitude angle:

[0025]

[0026] Wherein, the subscript d represents direct irradiance; is the clear sky direct irradiance, which can be calculated using the McClear model in step 3; τ is the cloud optical thickness; and μ0 is the sine of the solar altitude angle.

[0027] The common parameter in equations (3) and (4) is f. By shifting the terms in the two formulas, we can get:

[0028]

[0029] According to existing research, the cloud absorption rate α a and cloud albedo α r There is an approximate relationship as follows:

[0030] α a =(0.0537+0.0788μ0)α r (7)

[0031] Substituting formula (7) into formula (5) and combining it with formula (6) yields

[0032]

[0033] α=1.0537+0.0788μ0(8c)

[0034]

[0035] Thus, the right side of Equation (8a) contains the radiation parameters, while the left side contains the cloud parameters (cloud optical thickness and cloud albedo) and the solar altitude angle. For a given time and location, the sine of the solar altitude angle, μ0, can be calculated using Step 2, and μ0 can be considered a known quantity.

[0036] According to the second-rate approximation principle, the cloud albedo α r There is the following relationship between it and the cloud optical thickness τ:

[0037]

[0038] Where g is the asymmetry factor of cloud droplets, which is a constant term. According to previous research results, the asymmetry factor of water clouds is 0.86771, while the typical asymmetry factor of ice clouds is 0.84229. Considering the global thermodynamic phase distribution of clouds, in the shortwave band, this paper uses 0.86 as the typical value for calculation.

[0039] Substituting formula (9) into formula (8a) we can get

[0040]

[0041] From this formula, we can see that the cloud optical thickness τ is essentially But this equation cannot be explicitly expressed as τ with respect to Therefore, the expression needs to be further simplified.

[0042] 4.2 Simplification and calculation of key parameters

[0043] 1) Solar altitude angle θ a The solar altitude angle can be obtained according to step 2. The sine of the solar altitude angle μ0 = sin(θ a ).

[0044] 2) According to the representation of F1 and F2, It can be calculated from observations and the downlink direct irradiance output by the McClear model, and in the denominator One, T is the transmittance. Under typical thick cloud or overcast conditions, the transmittance is less than 0.3; under thin or scattered cloud, the transmittance can reach about 0.6. Considering that the global average surface albedo is 0.3, This item accounts for About 1% (0.3*0.6*0.6), so it is ignored in the calculation This one, that is, but The simplified calculation formula is:

[0045]

[0046] The coefficient α in the denominator of F1 is 1.0537+0.0788μ0 (Formula 8c).

[0047] 4.3 Polynomial fitting of cloud optical thickness and surface shortwave irradiance.

[0048] Since formula (10) cannot explicitly give the relationship between the shortwave irradiance at the surface and the cloud optical thickness, As a whole, it is a positive number, so it starts from 0 and increases in steps of 0.01. According to formula (10), we can get a The value of . Fitting and The curve can be obtained by The value of The curve is divided into 4 segments to ensure the most accurate fitting curve. Through polynomial fitting, the following relationship is obtained:

[0049]

[0050] R of the above segmented fitting formula 2 All above 0.99, indicating that it can reflect and The relationship between them.

[0051] When the actual observed irradiance and the calculated clear sky irradiance are calculated After the value of , the corresponding After calculating the sine of the solar altitude angle μ0 through step 2, the cloud optical thickness τ can be inversely calculated.

[0052] The advantages of the present invention are as follows: cloud optical thickness usually requires the use of radar or satellite detection technology. However, for sites that are not equipped with radar or are difficult to match with satellite observations, cloud optical thickness cannot be effectively obtained. The shortwave radiation component at the surface essentially contains cloud optical thickness information. Based on the radiation transfer equation and the simplification of some parameters, an explicit relationship between cloud optical thickness and shortwave irradiance at the surface is established. Therefore, cloud optical thickness can be inverted through observed and calculated surface shortwave irradiance. The promotion of this technology provides a new way to invert local cloud optical thickness information for sites or regions that only have radiation observation equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 For step 4 in the present invention and relationship.

[0054] The picture shows Follow The black dotted line is the collection of calculated values ​​obtained using formula (10), different colored curves represent fitting curves in different intervals, and the black dotted line represents the critical value of the interval.

[0055] Figure 2 A comparison chart of cloud optical thickness detected by radar and effective data. DETAILED DESCRIPTION

[0056] The following is combined with Figure 1-2 The present invention is further described in detail with reference to the accompanying drawings and embodiments.

[0057] Example 1

[0058] Step 1: Data preparation and pre-processing

[0059] Here, we use data from the U.S. Atmospheric Measurement Program (ARM) Southern Great Plains (SGP) Central Observatory as an example. The data was selected from 2014, and the cloud cover was selected based on the cloud cover observed at the station. With this large sample of observations, the accuracy of the inversion algorithm can be verified using ground-based cloud optical thickness.

[0060] 1) Select the shortwave irradiance at the surface from the observation data, including the total irradiance and the downward direct irradiance.

[0061] 2) Perform linear interpolation on the missing irradiance data within the day.

[0062] 3) For total irradiance higher than 10W / m 2 The data of the sunrise and sunset periods are deleted. According to the observation data, the radiation data of the period with visibility above 10km are filtered.

[0063] Step 2: Get the sun's altitude angle

[0064] The following explanation is based on the calculation at 12:00 noon on March 30, 2014 (local time). The calculation steps for other times are the same and will not be repeated here.

[0065] Due to a lack of observational data for the solar altitude angle, the solar altitude angle was calculated using the Solar Position Algorithm (SPA) in the Python open-source package pvlib. The longitude and latitude of the ARM-SGP central observation station are 97.487643°W, 36.607322°N, and the observation time corresponds to 2014-03-30 06:00 UTC (the time difference between the station and UTC is 6 hours). These parameters were input into the SPA algorithm, and the solar altitude angle and its sine were calculated to be 56.401178° and μ0 = 0.832933, respectively.

[0066] Step 3: Calculate clear sky irradiance

[0067] The longitude and latitude of the ARM-SGP central observation station, 36.607322°N, 97.487643°W, and the corresponding universal time 2014-03-30 06:00 were input into the McClear model, and the clear sky total irradiance and clear sky direct irradiance at that moment were obtained to be 918.5 W / m 2 and 807.6W / m 2 .

[0068] Step 4: Calculate cloud optical depth

[0069] The cloud optical thickness is solved according to formula (12).

[0070]

[0071] According to this formula, we can know that to calculate the cloud optical thickness, we must first calculate According to the simplified steps in 4.2 in step 4, The simplified formula is:

[0072]

[0073] Where α=1.0537+0.0788μ0=1.1193, and They are the clear sky total irradiance and the clear sky downward direct irradiance, which are 918.5W / m according to the calculation results in step 3. 2 and 807.6W / m 2 , and They are the total irradiance at the surface and the downward direct irradiance, which can be obtained from the observation results and are 852.9W / m 2 and 573.8W / m 2 , Substituting the above data into formula (11), we can get

[0074] Will Substituting into the first sub-term in formula (12), we can get According to step 2, μ0 = 0.832933, so the final cloud optical thickness τ is 3.4905.

[0075] All valid data of the SGP station in 2014 were used to calculate the cloud optical thickness for the whole year according to the above steps. The comparison between the cloud optical thickness and the cloud optical thickness detected by radar is as follows: Figure 2As shown in the figure, the two have good consistency, with a correlation coefficient of 0.75, passing the significance test of α=0.01. This result shows that the shortwave irradiance at the surface can be used to better invert cloud optical thickness.

[0076] This specific embodiment is merely an explanation of the present invention and is not intended to limit the present invention. After reading this specification, those skilled in the art may make non-creative modifications to this embodiment as needed. However, as long as such modifications are within the scope of the claims of the present invention, they are protected by patent law.

Claims

1. A cloud optical thickness inversion method based on surface shortwave irradiance, characterized in that: The following steps are involved:

1. Data Preparation and Preprocessing 1) This invention inverts cloud optical thickness based on shortwave irradiance observed on the surface. The observed radiation data must include two parameters: total irradiance and downlink direct irradiance. 2) Use linear interpolation to interpolate the missing values ​​of radiation data; 2. Get the sun's altitude angle When there is observation data of solar altitude angle or solar zenith angle in the observation data, use it directly, solar altitude angle = 90° - solar zenith angle; When there is a lack of observation data on the solar altitude angle or solar zenith angle, the latitude and longitude information of the station and the observation time are obtained, and the solar altitude angle is calculated using the Solar Position Algorithm (SPA) in the Python open source package pvlib; 3. Calculating Clear Sky Irradiance The McClear model is used to calculate clear-sky irradiance. This model utilizes the latest findings on aerosol properties, total water vapor content, and ozone content from the Atmospheric Composition and Climate Monitoring Project. Combined with the ABACI lookup table method and interpolation functions, it calculates shortwave irradiance at the surface under clear-sky conditions, including both clear-sky total irradiance and clear-sky downlink direct irradiance.

4. Calculating Cloud Optical Depth 4.1 Theoretical Relationship between Cloud Optical Depth and Shortwave Irradiance at the Earth’s Surface According to the single-layer cloud radiation transfer equation and Beer-Lambert law, the total irradiance at the surface is The theoretical expressions between cloud amount, cloud albedo, cloud absorptivity and diffuse transmittance are: Where F represents the shortwave irradiance, and the superscripts dn and up represent the vertical downward and upward directions, respectively. is the total clear sky irradiance, which can be calculated using the McClear model in step 3; is the upward shortwave irradiance reflected from the surface; α r and α a are cloud albedo and cloud absorptivity respectively; f is cloud cover; T is diffuse transmittance; For the downward direct irradiance at the surface Its theoretical relationship with cloud amount, cloud optical thickness, and solar altitude angle is: Wherein, the subscript d represents the direct irradiance, is the direct downward irradiance of the clear sky atmosphere, which can be calculated by the McClear model in step 3; τ is the cloud optical thickness; μ0 is the sine of the solar altitude angle; it is known that α a =(0.0537+0.0788μ0)α r , by combining equations (3) and (4) and shifting the terms, we can get: α=1.0537+0.0788μ0(8c) The right side of equation (8a) is a radiation-related parameter. The left side of the equation is the cloud parameter including cloud optical thickness and cloud albedo and the solar altitude angle. For a given time and place, the sine of the solar altitude angle μ0 can be calculated by step 2. According to the two-stream approximation principle, the cloud albedo α r There is an approximate relationship between it and the cloud optical thickness τ as follows: Where g is the asymmetry factor of cloud droplets, which is calculated using the typical value of 0.86 in the shortwave band. Substituting formula (9) into formula (8a) yields: From this formula, we can see that the cloud optical thickness τ is essentially function of , but the equation cannot be explicitly expressed as τ with respect to Therefore, after simplifying the key parameters, the polynomial fitting method is used to express the relationship between the two; 4.2 Calculation and simplification of key parameters 1) Solar altitude angle θ a , the solar altitude angle can be calculated according to step 2, then the sine of the solar altitude angle μ0=sin(θ a ); 2) According to the simplification of F1 and F2, It can be calculated from observations and the downlink direct irradiance output by the McClear model; and F1 = in the denominator One term, T is the transmittance, because Ignore here This one, Simplified The calculation formula is where the coefficient α in the denominator of F1 = 1.0537 + 0.0788 μ0 (Formula 8c); 4.3 Polynomial Fitting of Cloud Optical Depth and Surface Shortwave Irradiance Will As a whole, it is a positive number, starting from 0 and increasing in steps of 0.

01. According to formula (10), we can get a The numerical value of and The curve is divided into 4 segments to ensure the most accurate fitting curve. Through polynomial fitting, the following relationship is obtained: Using this relationship, when the actual observed irradiance and the calculated clear sky irradiance are calculated After the value of , the corresponding After calculating the sine of the solar altitude angle μ0 through step 2, the cloud optical thickness τ can be inversely calculated.

2. The cloud optical thickness inversion method based on surface shortwave irradiance according to claim 1 is characterized in that: In step 1, the total irradiance is less than 10W / m 2 Observation data.

3. The cloud optical thickness inversion method based on surface shortwave irradiance according to claim 1 is characterized in that: Determination of cleaning conditions in step 1: visibility>10km or PM 2.5 <75 μg / m 3 .

4. The cloud optical thickness inversion method based on surface shortwave irradiance according to claim 1 is characterized in that: The calculation of the solar altitude angle in step 2 is mainly based on the pvlib open source package of Python language, and the longitude λ, latitude φ and time of the station are input to calculate the solar altitude angle.

Images