Method and apparatus, system, storage medium for photosynthetically active radiation data interpolation

Abstract

The application discloses a photosynthetically active radiation data interpolation method and device, system and storage medium, and comprises the following steps: acquiring a photosynthetically active radiation time sequence sample set observed by a target year of a to-be-interpolated station; acquiring historical data of photosynthetically active radiation and meteorological driving factors of the to-be-interpolated station; screening key meteorological driving factors according to the historical data of photosynthetically active radiation and meteorological driving factors of the to-be-interpolated station; constructing a normal period regression decision tree model and an extreme period regression decision tree model according to the historical annual sequence variation characteristics of the key meteorological driving factors and the photosynthetically active radiation of the target year of the to-be-interpolated station; inputting the photosynthetically active radiation time sequence sample set into the normal period regression decision tree model and the extreme period regression decision tree model respectively for intelligent interpolation, and obtaining a long-time complete photosynthetically active radiation time sequence. The technical scheme of the application realizes high-frequency photosynthetically active radiation long-time sequence high-precision intelligent interpolation based on ground observation.

Description

Technical Field

[0001] The present invention belongs to the field of photosynthetically active radiation technology, and in particular relates to a photosynthetically active radiation data interpolation method and device, a system, and a storage medium. Background Art

[0002] There are two main methods for measuring photosynthetically active radiation (PAR): indirect and direct. Estimation based on meteorological data and remote sensing are currently the primary indirect methods for measuring PAR. Meteorological data-based estimation uses meteorological data such as sunshine duration and solar radiation to estimate PAR. This method can suffer from significant errors under unusual weather conditions. Remote sensing estimation uses satellite data to invert PAR. For example, the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite estimates PAR using a three-dimensional radiative transfer model (3D RT) or empirical models (such as those based on the Normalized Difference Vegetation Index (NDVI). This method provides continuous data over a large scale, but suffers from low temporal resolution and poor spatial and temporal consistency. Ground-based observations based on stations are the primary direct method for measuring PAR. PAR is measured directly using specialized photon sensors (such as quantum meters). This method offers high accuracy but is subject to data loss.

[0003] Currently, the following methods are commonly used to interpolate photosynthetically active radiation based on ground-based observations: 1. Mean Diurnal Variation (MDV): This method is simple and uses the average of observations from adjacent time periods to fill in data gaps. 2. Marginal Distribution Sampling (MDS): This method uses a dynamic averaging window and a sliding lookup table to fill in missing data. 3. Nonlinear Regression (NLR): This method uses the ecosystem respiration equation and light response curves to estimate photosynthetically active radiation.

[0004] Existing technical solutions suffer from three deficiencies, resulting in significant errors in interpolation results. The empirical relationship employed is suitable for short-term data loss (no more than 15 days), but accuracy decreases significantly when long-term data loss (e.g., more than 30 days) is present. The nonlinear effects and interactions of environmental factors are ignored, leading to significant deviations in interpolation results under extreme weather conditions (such as extreme heat, drought, combined dry-heat conditions, strong winds, stationary heat waves, and extreme total radiation). The limited interpolation data sample makes it difficult to capture the long-term variations in photosynthetically active radiation, limiting interpolation accuracy and stability. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a photosynthetic active radiation data interpolation method and device, system and storage medium.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A photosynthetically active radiation data interpolation method, comprising:

[0008] Step S1, obtaining a sample set of time series of photosynthetically active radiation and meteorological driving factors observed in the target year at the station to be interpolated;

[0009] Step S2: Obtain historical data of photosynthetically active radiation and meteorological driving factors for the site to be interpolated; wherein, based on the historical data of meteorological driving factors, obtain an extreme climate identification sample set; based on the historical data of photosynthetically active radiation, obtain a historical annual sequence change feature set of photosynthetically active radiation;

[0010] Step S3: Based on the synchronous observation data of photosynthetically active radiation and meteorological driving factors for the target year of the interpolation station, the importance scores of meteorological driving factors are calculated using random forest, and the key meteorological driving factors that are most effective in predicting the periodic changes of photosynthetically active radiation are selected;

[0011] Step S4: obtaining an extreme weather state discrimination result of the station to be interpolated based on the extreme climate identification sample set; dividing the photosynthetically active radiation time series sample set observed in the target year of the station to be interpolated into normal period samples and extreme period samples based on the extreme weather state discrimination result; training a random forest model with the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the normal period samples to construct a normal period regression decision tree model that reflects the variation law of photosynthetically active radiation under normal weather conditions; training a random forest model with the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the extreme period samples to construct an extreme period regression decision tree model that identifies the nonlinear response of photosynthetically active radiation to meteorological driving factors under extreme weather scenarios;

[0012] Step S5: Input the target year's photosynthetically active radiation time series sample set into the normal period regression decision tree model and the extreme period regression decision tree model for intelligent interpolation according to whether the extreme climate conditions are met, so as to obtain a long-term and complete photosynthetically active radiation time series.

[0013] The present invention also provides a photosynthetically active radiation data interpolation device, comprising:

[0014] The first processing module is used to obtain a sample set of photosynthetically active radiation time series observed in the target year at the site to be interpolated;

[0015] The second processing module is used to obtain historical data of photosynthetically active radiation and meteorological driving factors for the site to be interpolated; wherein, based on the historical data of meteorological driving factors, an extreme climate identification sample set is obtained; based on the historical data of photosynthetically active radiation, a historical annual sequence change feature set of photosynthetically active radiation is obtained;

[0016] The third processing module is used to calculate the importance scores of meteorological driving factors using random forests based on the historical data of photosynthetically active radiation and meteorological driving factors of the stations to be interpolated, and to screen the key meteorological driving factors that are most effective in predicting the periodic changes of photosynthetically active radiation;

[0017] The fourth processing module is used to obtain the extreme weather state discrimination result of the station to be interpolated based on the extreme climate identification sample set; based on the extreme weather state discrimination result, the photosynthetically active radiation time series sample set observed in the target year of the station to be interpolated is divided into normal period samples and extreme period samples; the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the normal period samples are used to train a random forest model to construct a normal period regression decision tree model that reflects the law of photosynthetically active radiation variation under normal weather conditions; the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the extreme period samples are used to train a random forest model to construct an extreme period regression decision tree model that identifies the nonlinear response of photosynthetically active radiation to meteorological driving factors under extreme weather scenarios;

[0018] The fifth processing module is used to input the photosynthetically active radiation time series sample set into the normal period regression decision tree model and the extreme period regression decision tree model for intelligent interpolation to obtain a long-term and complete photosynthetically active radiation time series.

[0019] The present invention also provides a photosynthetically active radiation data interpolation system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a photosynthetically active radiation data interpolation method when executed by the processor.

[0020] The present invention also provides a storage medium, on which a computer program is stored. When the computer program is run, the method for interpolating photosynthetically active radiation data is executed.

[0021] The present invention simultaneously considers the influence of long-term temporal dynamic change trends and extreme weather conditions, and performs intelligent interpolation by combining the nonlinear response relationship between photosynthetically active radiation and meteorological driving factors to achieve high-precision reconstruction of long-term time series data of high-frequency photosynthetically active radiation. Compared with traditional interpolation methods, the present invention provides a robust interpolation strategy that can effectively utilize historical observation data and fully consider the changing laws of meteorological variables. The present invention is not only suitable for filling in short-term data gaps, but can also cope with long-term data gaps, ensuring the reliability and physical consistency of the interpolation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0023] Figure 1 This is a flow chart of a photosynthetically active radiation data interpolation method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0025] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] Example 1:

[0027] like Figure 1 As shown, an embodiment of the present invention provides a photosynthetically active radiation data interpolation method, comprising:

[0028] Step S1, obtaining a sample set of photosynthetically active radiation time series observed in the target year at the site to be interpolated;

[0029] Step S2: Obtain historical data of photosynthetically active radiation and meteorological driving factors for the site to be interpolated; wherein, based on the historical data of meteorological driving factors, obtain an extreme climate identification sample set; based on the historical data of photosynthetically active radiation, obtain a historical annual sequence change feature set of photosynthetically active radiation;

[0030] Step S3: Based on the synchronous observation data of photosynthetically active radiation and meteorological driving factors for the target year of the interpolation station, the importance scores of meteorological driving factors are calculated using random forest, and the key meteorological driving factors that are most effective in predicting the periodic changes of photosynthetically active radiation are selected;

[0031] Step S4: obtaining an extreme weather state discrimination result of the station to be interpolated based on the extreme climate identification sample set; dividing the photosynthetically active radiation time series sample set observed in the target year of the station to be interpolated into normal period samples and extreme period samples based on the extreme weather state discrimination result; training a random forest model with the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the normal period samples to construct a normal period regression decision tree model that reflects the variation law of photosynthetically active radiation under normal weather conditions; training a random forest model with the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the extreme period samples to construct an extreme period regression decision tree model that identifies the nonlinear response of photosynthetically active radiation to meteorological driving factors under extreme weather scenarios;

[0032] Step S5: Input the target year's photosynthetically active radiation time series sample set into the normal period regression decision tree model and the extreme period regression decision tree model for intelligent interpolation according to whether the extreme climate conditions are met, so as to obtain a long-term and complete photosynthetically active radiation time series.

[0033] As an implementation method of an embodiment of the present invention, in step S1, the photosynthetically active radiation time series sample set includes: photosynthetically active radiation data and meteorological driving factors; wherein the meteorological driving factors include synchronous air temperature, relative humidity, wind speed, wind direction, air pressure and total radiation data observed at the same time and location as the photosynthetically active radiation data.

[0034] As an implementation method of the embodiment of the present invention, in step S2, according to the six types of meteorological driving factors (temperature , relative humidity , wind speed , air pressure , total radiation ), identifying typical extreme meteorological event types within the region. By constructing long-term statistical distributions of each variable, event thresholds were extracted and the following six typical events with significant impacts on photosynthetically active radiation observations were identified: extreme high temperatures, combined dry-heat events, strong winds, static heat waves, and extreme solar radiation events.

[0035] An extreme high temperature event is defined as a situation where the daily maximum temperature exceeds the 90th percentile of its long-term value, that is:

[0036] ;

[0037] Where, is the 90th percentile of the long-term observation series of daily maximum temperature, For the Maximum daily temperature.

[0038] Identify drought events by calculating vapor pressure deficit from air temperature and relative humidity:

[0039] ;

[0040] Where, For the The average daily temperature, For the Daily relative humidity, The calculated Daily saturated vapor pressure deficit, is the 90th percentile of the long-term observation series of relative humidity. A drought event is considered to have occurred.

[0041] For dry-hot composite events, the apparent temperature is used to integrate the effects of air temperature and relative humidity. The calculation formula is as follows:

[0042] ;

[0043] ;

[0044] Where, For the The average daily temperature, For the Daily relative humidity, Indicates time The actual water vapor pressure, Indicates the The average wind speed of the day is finally obtained For time If the perceived temperature exceeds the 90th percentile of the long-term observation series for two consecutive days, that is, , which is defined as a dry-heat composite event.

[0045] In terms of wind disturbances, the criteria for determining strong wind events are:

[0046] ;

[0047] Where, Indicates the The average wind speed of the day, is the 95th percentile of the long-term wind speed observation series.

[0048] For static heat wave events reflecting static heat accumulation, a composite judgment is made by combining the three criteria of high temperature, high pressure and low wind speed:

[0049] ;

[0050] Where, For the The maximum daily temperature, Indicates the The average daily air pressure, Indicates the The average wind speed of the day, 、 and They are the 90th, 95th and 10th percentiles of the long-term observation series of daily maximum temperature, daily average air pressure and daily average wind speed, respectively.

[0051] When all of the above conditions are met simultaneously, it is considered a stationary heat wave event. This type of event is usually accompanied by clear skies with few clouds and strong radiation accumulation, which has a significant driving effect on the volatility and extreme value risk of photosynthetically active radiation levels.

[0052] To identify extreme solar radiation events, the cumulative total radiation on that day exceeding the 95th percentile is considered an extreme value:

[0053] ;

[0054] in, For the Total daily radiation, is the 95th percentile of the long-term wind speed observation series.

[0055] Constructing the discrimination rules for the above six types of extreme events provides a systematic basis for characterizing the observational characteristics of photosynthetically active radiation under different extreme backgrounds, and provides information for optimizing interpolation models under extreme environments.

[0056] In addition, step S2 uses the SSA method to extract the historical annual sequence change feature set of photosynthetically active radiation. The historical data of photosynthetically active radiation are converted into Mapped into a high-dimensional state space and decomposed by singular value decomposition, the historical annual series variation feature set of photosynthetically active radiation is obtained. Specifically, the time series is first embedded into a delay matrix, defined as:

[0057] ;

[0058] in, is the embedding dimension, is the delay matrix composed of time series data. Next, perform singular value decomposition:

[0059] ;

[0060] in, and are the left and right singular vector matrices, respectively, Is a diagonal matrix of singular values. Extract the right singular vector matrix 中前5个主分量的列向量 ,每一个年份 构建一个特征矩阵 ,其中第 行表示一年内第 次观测的5 维投影系数,即:

[0061] ;

[0062] The characteristic matrices of all historical years constitute the historical annual series change feature set of photosynthetically active radiation, and the main components represent the periodicity and trend characteristics in the historical time series.

[0063] As an implementation of the present invention, in step S3, the importance of meteorological driving factors is measured using the mean square error reduction (ΔMSE) of random forest. Specifically, when a meteorological driving factor When used to partition a non-leaf node of a decision tree, the reduction of the prediction error by the partition is regarded as a reflection of its importance. Define a weight function ,表示节点 是否由 进行划分,其形式如下:

[0064] ;

[0065] 因此,每一项气象驱动因子 的整体重要性评分可以表示为:

[0066] ;

[0067] in, 表示森林中树的总数, Indicates the The set of all non-leaf nodes in the tree, using the meteorological driving factor 进行划分的结点 上,通过分裂条件 将光合有效辐射时间序列样本集 Divided into two subsets, the error reduction is defined as:

[0068] ;

[0069] in, and 分别表示依据气象驱动因子 的分裂阈值 The above error reduction indicates that the prediction accuracy of the meteorological driving factor in the current node split is improved. If the meteorological driving factor Frequently appearing in high-weight split nodes in different trees will have higher importance scores. Through the above process, the key meteorological driving factors that have a significant effect on the photosynthetic active radiation fitting effect under the environmental conditions of the interpolation site are identified.

[0070] Step S4: Based on the extreme weather state discrimination results, the photosynthetically active radiation time series sample set of the target year observation of the interpolation station is divided into normal period samples and extreme period samples. The training set and test set are divided into 80% and 20% respectively using the ten-point method in each sample. The model training input consists of the key meteorological driving factor vector , and the historical annual series variation feature set of photosynthetically active radiation extracted by singular spectrum decomposition method , as the model input features, and the measured photosynthetically active radiation value at the target time point 为监督输出,构建样本对:

[0071] ;

[0072] During training, four scenarios were constructed with missing photosynthetically active radiation observations, with missing percentages of 10%, 30%, 50%, and 70%. In each missing percentage, the corresponding percentage of observations was randomly masked from the training set, while the input features remained unchanged and participated in model training.

[0073] To achieve the optimal configuration of hyperparameters, a grid search method is used to systematically traverse the number of decision trees, maximum tree depth, minimum number of samples at internal nodes, and minimum number of samples at leaf nodes of the random forest model. Under each set of parameter combinations, the model performance is evaluated through the test set, and the model efficiency factor is calculated. 、均方根误差 与决定系数 ,作为模型超参数选取标准:

[0074] ;

[0075] in, 表示模型预测值, 为观测值, 为观测均值, is the number of samples. In the normal period regression decision tree model, the final number of decision trees was determined to be 140, with a maximum tree depth of 20, a minimum number of internal node samples of 8, and a minimum number of leaf node samples of 2. In the extreme period regression decision tree model, the final number of decision trees was determined to be 320, with a maximum tree depth of 30, a minimum number of internal node samples of 5, and a minimum number of leaf node samples of 2. Based on the above hyperparameter configuration, the random forest model and the random forest model were trained separately, and the model structure and parameter weights were saved for subsequent interpolation.

[0076] Within each sample set, a ten-point split was used for training and testing, with 80% of the samples used for training and 20% for testing. A random forest model was trained on samples from the normal period to construct a regression prediction structure for photosynthetically active radiation under normal meteorological conditions (a normal period regression decision tree model). Samples from the extreme period were used to train a random forest model to construct a nonlinear prediction structure for the response of photosynthetically active radiation to key meteorological factors during extreme meteorological events (an extreme period regression decision tree model). During the model training phase, to verify the model's stability and generalization under varying levels of missing data, scenarios with 10%, 30%, 50%, and 70% missing photosynthetically active radiation were constructed in the training set. Observations were randomly deleted from both the normal and extreme period training sets, retaining the input features unchanged and using the complete observations as labels for training. This step was used to evaluate how model performance changed under varying levels of missing data.

[0077] As an implementation method of an embodiment of the present invention, in step S5, the photosynthetic active radiation time series sample sets are respectively input into the normal period regression decision tree model to obtain the conventional model output; the photosynthetic active radiation time series sample sets are respectively input into the extreme period regression decision tree model to obtain the extreme scenario model output; the conventional model output and the extreme scenario model output are merged to realize the interpolation of the photosynthetic active radiation time series sample set and obtain a long-term complete photosynthetic active radiation time series.

[0078] Example 2:

[0079] An embodiment of the present invention further provides a photosynthetically active radiation data interpolation device, comprising:

[0080] The first processing module is used to obtain a sample set of photosynthetically active radiation time series observed in the target year at the site to be interpolated;

[0081] The second processing module is used to obtain historical data of photosynthetically active radiation and meteorological driving factors for the site to be interpolated; wherein, based on the historical data of meteorological driving factors, an extreme climate identification sample set is obtained; based on the historical data of photosynthetically active radiation, a historical annual sequence change feature set of photosynthetically active radiation is obtained;

[0082] The third processing module is used to calculate the importance scores of meteorological driving factors using random forests based on the historical data of photosynthetically active radiation and meteorological driving factors of the stations to be interpolated, and to screen the key meteorological driving factors that are most effective in predicting the periodic changes of photosynthetically active radiation;

[0083] The fourth processing module is used to obtain the extreme weather state discrimination result of the station to be interpolated based on the extreme climate identification sample set; based on the extreme weather state discrimination result, the photosynthetically active radiation time series sample set observed in the target year of the station to be interpolated is divided into normal period samples and extreme period samples; the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the normal period samples are used to train a random forest model to construct a normal period regression decision tree model that reflects the law of photosynthetically active radiation variation under normal weather conditions; the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the extreme period samples are used to train a random forest model to construct an extreme period regression decision tree model that identifies the nonlinear response of photosynthetically active radiation to meteorological driving factors under extreme weather scenarios;

[0084] The fifth processing module is used to input the photosynthetically active radiation time series sample set into the normal period regression decision tree model and the extreme period regression decision tree model for intelligent interpolation to obtain a long-term and complete photosynthetically active radiation time series.

[0085] Example 3:

[0086] An embodiment of the present invention also provides a photosynthetically active radiation data interpolation system, characterized in that it includes: a memory and a processor, wherein the memory stores a computer program run by the processor, and the computer program executes a photosynthetically active radiation data interpolation method when run by the processor.

[0087] Example 4:

[0088] An embodiment of the present invention further provides a storage medium having a computer program stored thereon, and the computer program executes a photosynthetic active radiation data interpolation method when running.

[0089] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A photosynthetically active radiation data interpolation method, characterized in that: include: Step S1, obtaining a sample set of photosynthetically active radiation time series observed in the target year at the site to be interpolated; Step S2: Obtain historical data of photosynthetically active radiation and meteorological driving factors for the site to be interpolated; wherein, based on the historical data of meteorological driving factors, obtain an extreme climate identification sample set; based on the historical data of photosynthetically active radiation, obtain a historical annual sequence change feature set of photosynthetically active radiation; Step S3: Based on the synchronous observation data of photosynthetically active radiation and meteorological driving factors for the target year of the interpolation station, the importance scores of the meteorological driving factors are calculated using random forest, and the key meteorological driving factors for the target year of the interpolation station that are most effective in predicting the periodic changes of photosynthetically active radiation are selected; Step S4: obtaining an extreme weather state discrimination result for the station to be interpolated based on the extreme climate identification sample set; dividing the photosynthetically active radiation time series sample set observed in the target year of the station to be interpolated into normal period samples and extreme period samples based on the extreme weather state discrimination result; training a random forest model with the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the normal period samples to construct a normal period regression decision tree model that reflects the variation law of photosynthetically active radiation under normal weather conditions; training a random forest model with the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the extreme period samples to construct an extreme period regression decision tree model that identifies the nonlinear response of photosynthetically active radiation to meteorological driving factors under extreme weather scenarios; Step S5, inputting the photosynthetically active radiation time series sample set of the target year into the normal period regression decision tree model and the extreme period regression decision tree model respectively according to whether the extreme climate conditions are met, and performing intelligent interpolation to obtain a long-term complete photosynthetically active radiation time series; wherein, the photosynthetically active radiation time series sample set is input into the normal period regression decision tree model respectively to obtain a conventional model output; the photosynthetically active radiation time series sample set is input into the extreme period regression decision tree model respectively to obtain an extreme scenario model output; the conventional model output and the extreme scenario model output are merged to realize the interpolation of the photosynthetically active radiation time series sample set, and obtain a long-term complete photosynthetically active radiation time series.

2. A photosynthetically active radiation data interpolation device, characterized in that: include: The first processing module is used to obtain a sample set of photosynthetically active radiation time series observed in the target year at the site to be interpolated; The second processing module is used to obtain historical data of photosynthetically active radiation and meteorological driving factors for the site to be interpolated; wherein, based on the historical data of meteorological driving factors, an extreme climate identification sample set is obtained; based on the historical data of photosynthetically active radiation, a historical annual sequence change feature set of photosynthetically active radiation is obtained; The third processing module is used to calculate the importance scores of meteorological driving factors using random forests based on the historical data of photosynthetically active radiation and meteorological driving factors of the stations to be interpolated, and to screen the key meteorological driving factors that are most effective in predicting the periodic changes of photosynthetically active radiation; The fourth processing module is used to obtain the extreme weather state discrimination result of the station to be interpolated based on the extreme climate identification sample set; based on the extreme weather state discrimination result, the photosynthetically active radiation time series sample set observed in the target year of the station to be interpolated is divided into normal period samples and extreme period samples; the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the normal period samples are used to train a random forest model to construct a normal period regression decision tree model that reflects the law of photosynthetically active radiation variation under normal weather conditions; the key meteorological driving factors of the target year of the station to be interpolated and the historical annual sequence change feature set of photosynthetically active radiation in the extreme period samples are used to train a random forest model to construct an extreme period regression decision tree model that identifies the nonlinear response of photosynthetically active radiation to meteorological driving factors under extreme weather scenarios; The fifth processing module is used to input the photosynthetically active radiation time series sample set into the normal period regression decision tree model and the extreme period regression decision tree model for intelligent interpolation to obtain a long-term complete photosynthetically active radiation time series; wherein, the photosynthetically active radiation time series sample set is input into the normal period regression decision tree model to obtain the conventional model output; the photosynthetically active radiation time series sample set is input into the extreme period regression decision tree model to obtain the extreme scenario model output; the conventional model output and the extreme scenario model output are merged to realize the interpolation of the photosynthetically active radiation time series sample set to obtain a long-term complete photosynthetically active radiation time series.

3. A photosynthetically active radiation data interpolation system, characterized in that: include: A memory and a processor, wherein the memory stores a computer program executed by the processor, and when the computer program is executed by the processor, the photosynthetic active radiation data interpolation method according to claim 1 is executed.

4. A storage medium, characterized in that The storage medium stores a computer program, which executes the photosynthetically active radiation data interpolation method according to claim 1 when running.

Images