Land rainfall prediction method and system in changing environment

Abstract

The invention provides a land rainfall prediction method and system in a changing environment, and belongs to the technical field of artificial intelligence big data model prediction.The system comprises a data acquisition module, a training module and a prediction module, long-term multi-source data is acquired through ERA5 data and a WAM-2layers model, and the prediction of land rainfall in the changing environment is realized by combining training and data processing of a long and short-term memory network model. According to the method, global and regional analysis is carried out, historical rainfall data, water vapor source data of different regions and climate indexes are combined during prediction, and correlation among the three kinds of data is fully utilized, so that rapid and high-precision prediction of land rainfall is realized.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence big data model prediction technology, and specifically relates to a method and system for predicting land precipitation under a changing environment. Background Art

[0002] Land precipitation is key to the Earth's water cycle and is essential for maintaining ecosystems, agriculture, and human society. In the wake of ongoing climate change, it is becoming increasingly important to understand the sources and dynamics of future precipitation, and changes in precipitation patterns are of great significance to the study of water resources, food security, and natural habitats. Changes in the sources of land precipitation are affected by a combination of factors, including rising temperatures caused by global warming, increased evaporation, and long-term shifts in atmospheric circulation patterns. Rising sea surface temperatures enhance ocean evaporation, which in turn affects the ocean's contribution to land precipitation. At the same time, deforestation and changes in land use also inhibit evapotranspiration from land, thereby affecting the replenishment of precipitation from land moisture sources. Existing technologies for the analysis of land precipitation often lack the long-term and comprehensive methods needed to capture broader patterns and trends. Most analyses focus on data from specific regions or limited in time, ignoring the impact of long periods and global factors. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for predicting land precipitation under a changing environment, so as to solve the problems of space, time and data limitations in precipitation prediction in the background technology and achieve faster and more accurate land precipitation prediction.

[0004] The present invention provides a method for predicting land precipitation under a changing environment, the method comprising: The land precipitation time series data were obtained based on the reanalysis of daily land precipitation data of the ERA5 dataset; Based on the WAM-2layers water vapor tracking model, the land water vapor contribution time series and the ocean water vapor contribution time series are obtained; based on the land precipitation time series data, the land water vapor contribution time series and the ocean water vapor contribution time series are cross-validated; If the cross-validation is consistent, the water vapor contribution time series of all Köppen climate zones are calculated based on the verified land water vapor contribution time series and the WAM-2layers water vapor tracking model; the water vapor contribution time series of the four oceans are calculated based on the verified ocean water vapor contribution time series and the WAM-2layers water vapor tracking model; Extract the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans respectively; perform correlation analysis between the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans and the main climate index data to obtain key climate index data; Establishing a model based on a long short-term memory network, using the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, the water vapor contribution time series of the four oceans and key climate index data as training sets and test sets, and training and testing the model; Future land precipitation is predicted based on the trained and tested models.

[0005] Furthermore, the cross-validation based on the land precipitation time series data and the land water vapor contribution time series and the ocean water vapor contribution time series includes: verifying whether the sum of the land water vapor contribution time series and the ocean water vapor contribution time series is consistent with the land precipitation time series data.

[0006] Furthermore, the reanalysis of daily-scale land precipitation data based on the ERA5 dataset obtains land precipitation time series data, including: Define the test statistic for land precipitation time series data; Obtaining a standardized test statistic based on the test statistic; Obtaining a qualitative change trend of land precipitation based on the standardized test statistic; The Sen's Slope test is performed on the land precipitation time series data to calculate the trend slope and obtain the quantitative change trend of land precipitation.

[0007] Furthermore, the obtaining of land precipitation time series data based on the reanalysis daily-scale land precipitation data of the ERA5 dataset also includes: dividing the reanalysis daily-scale land precipitation data of the ERA5 dataset from 1941 to 2023 into four stages to calculate the land precipitation time series data, and the four stages include: P1: 1941-1960, P2: 1961-1980, P3: 1981-2000, P4: 2001-2023.

[0008] Furthermore, the WAM-2layers water vapor tracking model is used to obtain the land water vapor contribution time series and the ocean water vapor contribution time series, including: tracking and calculating water vapor based on the following formula: in, Indicates the moisture content of the bottom or top layer of a pixel. and are the lateral and longitudinal wind speeds, and Represents the water vapor balance content in the horizontal and vertical directions, is the evaporation rate, is the amount of precipitation, is the residual, It is the vertical transport of water between the upper and lower layers.

[0009] Furthermore, based on the verified land water vapor contribution time series and the WAM-2layers water vapor tracking model, the water vapor contribution time series of all Köppen climate zones are calculated respectively; based on the verified ocean water vapor contribution time series and the WAM-2layers water vapor tracking model, the water vapor contribution time series of the four oceans are calculated respectively, including: Based on the Köppen climate zone classification vector data, ocean boundary vector data and WAM-2layers water vapor tracking model, the evaporation of all Köppen climate zones and four oceans tracked to land precipitation is tracked and calculated respectively; Based on the evaporation traced to land precipitation and the total land precipitation in all the Köppen climate zones and the four oceans, the water vapor contribution ratios of the corresponding Köppen climate zones and oceans are calculated respectively; Based on the water vapor contribution ratios of the corresponding Köppen climate zones and oceans and the land water vapor contribution time series and the ocean water vapor contribution time series, the water vapor contribution time series of the corresponding Köppen climate zones and oceans are calculated respectively according to the four stages.

[0010] Furthermore, the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans are extracted respectively, including: Based on the water vapor contribution time series of all the Köppen climate zones and the four oceans, the first three principal components with cumulative variance contribution rates greater than 60% were extracted.

[0011] Furthermore, the main climate index data include ENSO index data, IDO index data and NAO index data; The principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans were correlated with the main climate index data, and the calculation formula for key climate index data was obtained as follows: In the formula, and represent water vapor contribution data and climate index data respectively, and Represent the average values ​​of water vapor contribution data and climate index data, respectively.

[0012] Further, interpolation operation is performed on the key climate index data or the main climate index data; The time scale of the key climate index data or the main climate index data is made the same as the time scale of the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, and the water vapor contribution time series of the four oceans.

[0013] Based on the above-mentioned land precipitation prediction method under a changing environment, the present invention also provides a land precipitation prediction system under a changing environment, the system comprising: The data acquisition module is used to obtain the land precipitation time series data based on the reanalysis daily-scale land precipitation data of the ERA5 data set; obtain the land water vapor contribution time series and the ocean water vapor contribution time series based on the WAM-2layers water vapor tracking model; perform cross-validation based on the land precipitation time series data and the land water vapor contribution time series and the ocean water vapor contribution time series; if the cross-validation is consistent, calculate the water vapor contribution time series of all Köppen climate zones based on the verified land water vapor contribution time series and the WAM-2layers water vapor tracking model; calculate the water vapor contribution time series of the four oceans based on the verified ocean water vapor contribution time series and the WAM-2layers water vapor tracking model; extract the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans respectively; perform correlation analysis on the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans with the main climate index data to obtain key climate index data; A training module, for establishing a model based on a long short-term memory network, using the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, the water vapor contribution time series of the four oceans and key climate index data as training sets and test sets to train and test the model; The prediction module is used to predict future land precipitation based on the trained and tested model.

[0014] The above one or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects: The land precipitation prediction method and system provided by the embodiment of the present invention obtains long-term multi-source data through ERA5 data and WAM-2layers model, combines the long short-term memory network model, and provides a high-precision precipitation prediction method through global and regional level analysis, combined with historical precipitation data, water vapor source data in different regions and climate index. The comprehensive use of ERA5 reanalysis data and WAM-2layers water vapor tracking model can more accurately track and analyze the source and change of land precipitation. The ERA5 data set covers a wide range of meteorological elements and provides a detailed data basis for water vapor transmission and precipitation process. At the same time, the WAM-2layers model accurately calculates the contribution ratio of water vapor from different sources through the layered atmospheric structure, thereby realizing a high-precision dynamic evaluation of the source of precipitation. The multi-source data fusion method overcomes the limitations of the previous single data source or simple model. Since the historical precipitation data, water vapor source data in different regions and climate index are combined in the prediction, the correlation between the three data is fully utilized, so the prediction of land precipitation is fast and high-precision, which can more comprehensively reveal the complex mechanism of precipitation source. The above method and system can achieve end-to-end prediction without human intervention. Compared with the traditional prediction method that relies on human experience or simple statistical models, it can respond to changes in precipitation and climate indices more quickly, provide more timely and accurate information support for decision makers, help improve the efficiency and effectiveness of water resources management, reduce the impact of human factors on prediction results, and ensure the objectivity and accuracy of the prediction.

[0015] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flow chart of the prediction method in Example 1 of the present invention; Figure 2 This is an architecture diagram of the prediction system in Example 2 of the present invention. DETAILED DESCRIPTION

[0017] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0018] Embodiment 1: Reference Figure 1This embodiment provides a method for predicting land precipitation under a changing environment, the method comprising the following steps: S1: Time series data of land precipitation are obtained based on the reanalysis of daily land precipitation data of the ERA5 dataset; S2: Based on the WAM-2layers water vapor tracking model, the land water vapor contribution time series and the ocean water vapor contribution time series are obtained; based on the land precipitation time series data, the land water vapor contribution time series and the ocean water vapor contribution time series are cross-validated; S3: If the cross-validation is consistent, the water vapor contribution time series of all Köppen climate zones are calculated based on the land water vapor contribution time series and the WAM-2layers water vapor tracking model; the water vapor contribution time series of the four oceans are calculated based on the ocean water vapor contribution time series and the WAM-2layers water vapor tracking model; S4: Extract the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans respectively; perform correlation analysis between the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans and the main climate index data to obtain key climate index data; S5: Building a model based on a long short-term memory network, using the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, the water vapor contribution time series of the four oceans and key climate index data as training sets and test sets, to train and test the model; S6: Predict future land precipitation based on the trained and tested model.

[0019] For step S1, as global temperatures rise, evaporation increases, and the water vapor content in the atmosphere also increases, which leads to an increase in extreme precipitation events. Some areas may experience more frequent and severe rainstorms and floods, while other areas may experience long-term droughts. Uneven precipitation distribution not only affects agricultural production and water supply, but may also lead to environmental problems such as soil erosion, river diversion, and groundwater level decline. Land precipitation is a key component of the Earth's water cycle. The center of the water cycle is the complex interaction between land and ocean moisture sources. In order to understand the trend of land precipitation and provide a basis for water vapor tracking, a long-term trend analysis of total land precipitation is first performed. The present invention utilizes the ERA5 dataset, which has the advantages of improvements in model physics, dynamic kernel, data assimilation, and increased resolution, which plays an important role in improving the quality of weather and climate data.

[0020] This example is based on the ERA5 reanalysis of daily land precipitation data, and calculates the annual land precipitation data from 1941 to 2023 to obtain the land precipitation time series data. ,in Represents the precipitation data for a certain year. This time series data is independent, identically distributed samples of random variables.

[0021] Specifically, the land precipitation time series data are obtained based on the reanalysis daily-scale land precipitation data of the ERA5 dataset, including: dividing the reanalysis daily-scale land precipitation data of the ERA5 dataset from 1941 to 2023 into four stages to calculate the land precipitation time series data, and the four stages include: P1: 1941-1960, P2: 1961-1980, P3: 1981-2000, P4: 2001-2023.

[0022] In this embodiment, the reanalysis of daily-scale land precipitation data based on the ERA5 data set obtains land precipitation time series data. After obtaining the land precipitation time series data, a long-term trend analysis of total land precipitation can be performed based on the obtained land precipitation time series data to analyze the changing trend of land precipitation. The analysis methods include qualitative and quantitative analysis.

[0023] The specific steps of qualitative analysis are: Define the test statistic for the land precipitation time series data. Specifically, use the Mann-Kendall trend test method to determine the trend significance and define the test statistic as follows: in, , Respectively , The observations corresponding to the time series, and , is a sign function.

[0024] Water vapor contribution time series number n When it is greater than 8, the statistic It roughly obeys the normal distribution. Without considering the existence of equal-valued data points in the sequence, the variance calculation formula is as follows: A standardized test statistic is obtained based on the test statistic, and the standardized test statistic is The calculation is as follows: Based on the standardized test statistic The qualitative change trend of the land precipitation is obtained, wherein: A positive value indicates that the land precipitation is increasing, and a negative value indicates that the land precipitation is decreasing. When the absolute value is greater than or equal to 1.645, 1.96, and 2.576, it means that the significance test with confidence level of 90%, 95%, and 99% has been passed respectively.

[0025] The quantitative analysis method is: perform Sen's Slope test on the land precipitation time series data to calculate the trend slope, and obtain the quantitative change trend of the land precipitation. The calculation formula is: in, Represents the change per unit time in the time series. When it is a positive number, it means that the time series shows an upward trend. When it is negative, the time series shows a downward trend. To find the median function.

[0026] The above qualitative and quantitative trend analysis can provide a qualitative and quantitative understanding of the overall changing trend of land precipitation from a macro perspective, thereby providing data support for the subsequent analysis of the causes and sources of precipitation changes.

[0027] For step S2, the change in the source of land precipitation is affected by a combination of factors, including rising temperatures caused by global warming, increased evaporation, and long-term changes in atmospheric circulation patterns. The rise in sea surface temperature enhances ocean evaporation, which in turn affects the contribution of the ocean to land precipitation. At the same time, deforestation and changes in land use also inhibit evaporation from land, thereby affecting the replenishment of precipitation by land moisture sources.

[0028] This embodiment uses the WAM-2layers water vapor tracking model. WAM-2layers (Water Accounting Model-2layers) is an atmospheric moisture tracking model based on the Euler framework. It is mainly used to analyze the water vapor source and transmission path of regional precipitation. Its core features and applications are as follows: Tracking mode: The ability to determine where precipitation initially evaporates (reverse tracking) or where evaporated water eventually reaches (forward tracking) is important for understanding the water cycle.

[0029] Backtracking: Identifying the sources of evaporation (i.e., “moisture sources”) of precipitation in a specific area.

[0030] Forward tracking: predicting where the evaporated water will end up in the atmosphere (e.g. as precipitation or recycling processes).

[0031] Layered modeling: The atmosphere is vertically divided into two layers (boundary layer and free atmosphere), and the contribution of water vapor transport at different heights is quantified by combining humidity field and wind field data.

[0032] The inherent capabilities of the WAM-2layers model enable it to more effectively identify the different water vapor contributions from different surface evaporation sources.

[0033] The WAM-2layers water vapor tracking model is used to obtain the land water vapor contribution time series and the ocean water vapor contribution time series, including: tracking and calculating water vapor based on the following formula: (6) The above formula is based on the principle of atmospheric water balance. It divides the atmosphere into two levels, the bottom layer and the top layer, to track and calculate water, so as to better understand the transfer of water between different regions. The bottom layer mainly focuses on surface water evaporation and precipitation, while the top layer focuses on water transport and redistribution in the atmosphere. Indicates the moisture content of the bottom or top layer of a pixel. and are the lateral and longitudinal wind speeds, and Represents the water vapor balance content in the horizontal and vertical directions, is the evaporation rate, is the amount of precipitation, is the residual, It is the vertical transport of water between the upper and lower layers.

[0034] Based on the above WAM-2layers water vapor tracking model, the water vapor transport source of land precipitation can be obtained, and the contribution of water vapor sources in different regions (such as the four oceans: Indian Ocean, Arctic Ocean, Atlantic Ocean, Pacific Ocean; different Köppen climate zones: continental climate, polar climate, tropical climate, arid climate, temperate climate) to land precipitation can be obtained. The water vapor contribution calculation formula is as follows: In the formula, Indicates the contribution ratio of specific source areas (four oceans, different Köppen climate zones) to land precipitation, It is the amount of water vapor that is traced from a specific region (the four oceans and different Köppen climate zones) to the land. It is the total precipitation over land, obtained by statistically analyzing precipitation by region.

[0035] Based on the above principle, the water vapor contribution ratios of land and ocean areas in four periods (P1: 1941-1960, P2: 1961-1980, P3: 1981-2000, P4: 2001-2023) can be calculated respectively. The specific formula is as follows: In the formula, is the total precipitation over land, is the water vapor contribution of land to land precipitation, The ocean's contribution to land precipitation is The proportion of land water vapor contribution is To track evaporation from land to land precipitation, The contribution ratio of ocean water vapor is To track evaporation from the ocean to precipitation on land.

[0036] The principle of cross-validation based on the land precipitation time series data and the land water vapor contribution time series and the ocean water vapor contribution time series is that theoretically, the precipitation in a specific area is equal to the water vapor transport amount. Specifically for land precipitation, the cross-validation method used in this embodiment is to verify whether the sum of the land water vapor contribution time series and the ocean water vapor contribution time series is consistent with the land precipitation time series data. The land water vapor contribution time series and the ocean water vapor contribution time series are the sum of the land water vapor contribution time series and the ocean water vapor contribution time series in the above formula (8). and Long-term time series data. Through the above verification, it can be proved that in the case of long time scales, the water vapor contribution data obtained by tracking global water vapor using the WAM-2layers water vapor tracking model is roughly consistent with the precipitation data. Therefore, the water vapor contribution data obtained by the above theoretical calculations and the reanalysis daily-scale land precipitation data of the ERA5 dataset have an inherently consistent relationship. The reanalysis daily-scale land precipitation data of the ERA5 dataset during this period can be used for land precipitation prediction. If the above verification is inconsistent, it means that there are certain errors in the precipitation data during this period and it cannot be used for subsequent model training and testing. The above verification is particularly suitable for the verification of older data, which can effectively remove erroneous data and shield missing data.

[0037] For step S3, the Köppen climate classification is a climate classification system proposed by German climatologist Willard Köppen, which divides the land into different climate types based on temperature and precipitation. This system divides the global land climate into five major climate zones (continental climate, polar climate, tropical climate, arid climate, and temperate climate), and different climate zones have obvious differences in their contribution to land precipitation.

[0038] In this embodiment, based on the verified land water vapor contribution time series and the WAM-2layers water vapor tracking model, the water vapor contribution time series of all Köppen climate zones are calculated respectively; based on the verified ocean water vapor contribution time series and the WAM-2layers water vapor tracking model, the water vapor contribution time series of the four oceans are calculated respectively, including: Based on the Köppen climate zone classification vector data, ocean boundary vector data and WAM-2layers water vapor tracking model, the evaporation of all Köppen climate zones and four oceans tracked to land precipitation is tracked and calculated respectively; Based on the evaporation traced to land precipitation and the total land precipitation of all the Köppen climate zones and the four oceans, respectively, the water vapor contribution ratios of the corresponding Köppen climate zones and oceans are calculated; The method for calculating the water vapor contribution ratio of the corresponding Köppen climate zone is as follows: obtain the classification vector data of the Köppen climate zone, extract the global evaporation data of the corresponding source area respectively, and calculate the first Köppen climate zones are tracked to evaporation of precipitation over land , calculate the water vapor contribution ratio, the specific formula is as follows: In the formula, Indicates The water vapor contribution of the Köppen climate zones to land precipitation.

[0039] The ocean not only has an important impact on the global climate, but is also a key component of the global water cycle and marine ecosystem. The ocean participates in the global water cycle through evaporation and precipitation, and has a significant impact on precipitation distribution and climate patterns.

[0040] In order to analyze the impact of different oceans on land precipitation, the boundary vector data of the Pacific Ocean, Indian Ocean, Atlantic Ocean and Arctic Ocean were obtained respectively, and the global evaporation data of the corresponding source areas were extracted respectively. Based on the WAM-2layers model, the first The evaporation of oceans is tracked to land precipitation , based on the above formula (7), the water vapor contribution ratio is calculated as follows: In the formula, Indicates The water vapor contribution of ocean regions to land precipitation.

[0041] Similar to the method of step S2, based on the water vapor contribution ratio of the corresponding Köppen climate zone and the ocean and the land water vapor contribution time series and the ocean water vapor contribution time series, the water vapor contribution time series of the corresponding Köppen climate zone and the ocean are calculated respectively according to the four stages.

[0042] For step S4, climate change is the main driving factor for changes in the structure of water vapor contribution to land precipitation. Climate indices such as ENSO, IOD and NAO play an important role in various climate zones. There are many types of climate indices, so it is necessary to find climate indices that have a greater impact on changes in the structure of water vapor contribution to land precipitation. In order to find climate indices related to changes in the structure of water vapor contribution to land precipitation, firstly, the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans are extracted respectively. The specific method is to extract the first three principal components with a cumulative variance contribution rate greater than 60% based on the water vapor contribution time series of all Köppen climate zones and the four oceans.

[0043] The specific formula is as follows: For the Köppen climate zone or ocean water vapor contribution time series obtained in step S3 , the time series contains A dataset of samples, where each sample Is a multidimensional vector, calculate the mean vector of the sample : Perform centralization, that is, subtract the corresponding mean from each feature to obtain standardized data: Calculate the covariance matrix of the dataset : in, is the sample size, is a The covariance matrix of .

[0044] Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalue and the corresponding eigenvector , arrange the eigenvalues ​​in descending order and record the corresponding eigenvectors.

[0045] Based on the ranking of the eigenvalues, the variance contribution of each eigenvalue is calculated, and the top eigenvalues ​​that can explain most of the variance are selected. The calculation formula for the cumulative variance contribution rate is: In this embodiment, the first three eigenvalues ​​with cumulative contribution rates greater than 60% are selected as principal components PC1, PC2, and PC3 for subsequent screening of climate index data. Since only three principal components are selected for data calculation, the amount of calculation can be greatly reduced while ensuring the accuracy of the calculation.

[0046] The existing main climate index data include ENSO index data, IDO index data, and NAO index data. Different climate indices reflect the state of complex climate systems such as atmospheric circulation and ocean-atmosphere interaction from multiple dimensions, and these states will have a direct or indirect impact on the formation, distribution and change of precipitation. Through a detailed analysis of the correlation between climate indices and water vapor contributions, key climate indices that have a significant indicative effect on precipitation in specific areas can be accurately screened out. In the subsequent prediction of future precipitation, the prediction model is constructed based on these screened climate indices, which can effectively improve the model's ability to capture precipitation trends and prediction accuracy. The principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans are correlated with the main climate index data, and the calculation formula for obtaining the key climate index data is: (17) In the formula, and represent water vapor contribution data and climate index data respectively, and Represent the average values ​​of water vapor contribution data and climate index data respectively. The correlation between the first three principal components of water vapor contribution in the Köppen climate zone and the climate index is obtained. The water vapor contribution data adopts the three principal components obtained in the above steps, thus obtaining the correlation between the first three principal components PC1, PC2, PC3 of water vapor contribution in all Köppen climate zones and the four oceans and the climate index.

[0047] For step S5, a model is established based on a long short-term memory network, and the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, the water vapor contribution time series of the four oceans and the key climate index data are used as training sets and test sets to train and test the model.

[0048] Among them, the land precipitation time series data is obtained by step S1, and the water vapor contribution time series of the Köppen climate zone and the water vapor contribution time series of the four oceans are obtained by step S3. The key climate index data are selected by screening, such as the ENSO index publicly provided by the Physical Science Laboratory (NOAA / PSL) of the National Oceanic and Atmospheric Administration of the United States, the IOD index is obtained from the NOAA / OOPC website, and the NAO index is from the NOAA Climate Prediction Center (CPC) website. The long short-term memory network model (LSTM) takes into account the characteristics of long-term dependencies and multi-step outputs. The above data are used as the input of the model, and the key climate index data or the main climate index data are interpolated by calculating the correlation coefficient and performing weighted summation; the time scale of the key climate index data or the main climate index data is made the same as the time scale of the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, and the water vapor contribution time series of the four oceans, thereby realizing the effective fusion of the climate index data with the other two sets of time series data. As mentioned above, the time series data are divided into four periods: P1 (1941-1960), P2 (1961-1980), P3 (1981-2000) and P4 (2001-2023), and further divided into training and test sets based on the principle of equal division of time. These data sets are trained and tested, and the prediction output of the model is designed to capture the time series of future land water vapor contribution based on the Köppen climate zone and ocean water vapor contribution based on the four oceans.

[0049] Specifically, taking daily-scale data as input data as an example, it includes: daily-scale time series of global land precipitation ,in Indicates Daily precipitation; sources of water vapor in the four oceans ,in Indicates Water vapor contribution vectors from the four oceans (such as the Pacific Ocean, Atlantic Ocean, Indian Ocean, and Arctic Ocean); Water vapor sources in different Köppen climate zones on land ,in Indicates The water vapor contribution vector from different Köppen climate zones (such as tropical, temperate, and frigid zones) is obtained on the day. The water vapor source is mainly affected by the climate index, namely ENSO, NAO, and IOD mentioned in the above steps. This example obtains monthly climate index data from websites such as NOAA. ,in Indicates The climate index of the day (such as ENSO, NAO, IOD, etc.) is interpolated to unify the time scale of the climate index data to the daily scale to keep it consistent with other data. The above data is used as the input data of the LSTM model.

[0050] Preprocess the data and standardize the above input data. The reference formula is as follows: (18) In the formula, is the original data, is the mean, is the standard deviation.

[0051] Defining the time step , construct the standardized input data into a time series sample, the reference formula is as follows, In the formula, For the The input features at each time step.

[0052] Secondly, the LSTM model is constructed. LSTM is divided into a unit structure and a network structure. The LSTM unit controls the flow of information through a gating mechanism. Its mathematical description is as follows: Forget Gate: (20) Input Gate: (twenty one) (twenty two) Cell status update: (twenty three) Output Gate: (twenty four) (25) In the above formula, is the input of the current time step, is the hidden state of the previous time step, is the cell state at the previous time step, is the weight matrix, is the bias term, is the sigmoid function.

[0053] The network structure of the long short-term memory network LSTM model constructed in this embodiment includes: Input layer: accepts time steps of Input features ; LSTM layer: contains multiple layers of LSTM units, used to extract time series features; Fully connected layer: maps the output of the LSTM layer to the predicted value; Output layer: Output predicted precipitation .

[0054] After the LSTM model is built, it is trained. In order to predict future precipitation conditions on a daily basis, this embodiment uses daily-scale data from 1941 to 2023 to train and test the model, and uses the Adam optimization algorithm to update the model parameters during training to minimize the loss function. The specific formula of the loss function is as follows: (26) In the formula, is the actual precipitation, To predict precipitation, is the sample size.

[0055] After the model is trained, step S6 is executed to predict future land precipitation based on the trained and tested model. Input into the trained LSTM model, the model will output the predicted precipitation in the future time step The above embodiment is based on long-term historical data, and uses the LSTM model to summarize historical laws. The land precipitation sources in the training and input data are further refined into land and ocean water vapor contribution sources, and the correlation and influence of climate index and water vapor contribution sources are considered. It can also reject the addition of future data with uncertainty, which can effectively improve the accuracy of future precipitation prediction.

[0056] Embodiment 2: Reference Figure 2 Based on the land precipitation prediction method under a changing environment described in Example 1, a land precipitation prediction system under a changing environment is also provided, the system comprising: The data acquisition module is used to obtain the land precipitation time series data based on the reanalysis daily-scale land precipitation data of the ERA5 data set; obtain the land water vapor contribution time series and the ocean water vapor contribution time series based on the WAM-2layers water vapor tracking model; perform cross-validation based on the land precipitation time series data and the land water vapor contribution time series and the ocean water vapor contribution time series; if the cross-validation is consistent, calculate the water vapor contribution time series of all Köppen climate zones based on the verified land water vapor contribution time series and the WAM-2layers water vapor tracking model; calculate the water vapor contribution time series of the four oceans based on the verified ocean water vapor contribution time series and the WAM-2layers water vapor tracking model; extract the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans respectively; perform correlation analysis on the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans with the main climate index data to obtain key climate index data; A training module, for establishing a model based on a long short-term memory network, using the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, the water vapor contribution time series of the four oceans and key climate index data as training sets and test sets to train and test the model; The prediction module is used to predict future land precipitation based on the trained and tested model.

[0057] The specific implementation method of this embodiment is the same as that of Embodiment 1, which will not be repeated here. Please refer to the description of Embodiment 1 for details.

[0058] The land precipitation prediction method and system under the changing environment of the above-mentioned embodiment are adopted, and the LSTM model is used to combine historical precipitation data, water vapor source data in different regions and climate index to provide a high-precision precipitation prediction method. The LSTM model can capture long-term dependencies in time series, effectively handle nonlinear relationships and dynamic changes in time series, thereby significantly improving the accuracy and reliability of precipitation prediction. By comprehensively using ERA5 reanalysis data and WAM-2layers water vapor tracking model, the source and changes of land precipitation can be tracked and analyzed more accurately. The ERA5 data set covers a wide range of meteorological elements and provides a detailed data basis for water vapor transmission and precipitation processes. At the same time, the WAM-2layers model accurately calculates the proportion of water vapor contribution from different sources through the layered atmospheric structure, thereby realizing a high-precision dynamic evaluation of the source of precipitation. This method of multi-source data fusion overcomes the limitations of previous single data sources or simple models, fully considers the correlation and influence of climate indices and water vapor contribution sources, and can more comprehensively reveal the complex mechanism of precipitation sources. The above-mentioned prediction method and system are highly flexible and adaptable, and can adapt to different data formats and time scales. Whether it is precipitation prediction on a daily, monthly or annual scale, it can be achieved by adjusting the parameters and input data of the LSTM model. Compared with the traditional prediction method that relies on manual experience or simple statistical models, the present invention has a higher degree of automation and can respond more quickly to changes in precipitation and climate indices, providing decision makers with more timely and accurate information support, helping to improve the efficiency and effectiveness of water resources management, reduce the impact of human factors on prediction results, and ensure the objectivity and accuracy of the prediction.

[0059] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0060] Obviously, those skilled in the art can make various changes and modifications to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Thus, if these modifications and variations of the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for predicting land precipitation under a changing environment, characterized in that: The method comprises: The land precipitation time series data were obtained based on the reanalysis of daily land precipitation data of the ERA5 dataset; Based on the WAM-2layers water vapor tracking model, the land water vapor contribution time series and the ocean water vapor contribution time series are obtained; based on the land precipitation time series data, the land water vapor contribution time series and the ocean water vapor contribution time series are cross-validated; If the cross-validation is consistent, the water vapor contribution time series of all Köppen climate zones are calculated based on the verified land water vapor contribution time series and the WAM-2layers water vapor tracking model; the water vapor contribution time series of the four oceans are calculated based on the verified ocean water vapor contribution time series and the WAM-2layers water vapor tracking model; Extract the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans respectively; perform correlation analysis between the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans and the main climate index data to obtain key climate index data; Establishing a model based on a long short-term memory network, using the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, the water vapor contribution time series of the four oceans and key climate index data as training sets and test sets, and training and testing the model; Future land precipitation is predicted based on the trained and tested models.

2. The method for predicting land precipitation under a changing environment according to claim 1, characterized in that: The cross-validation based on the land precipitation time series data and the land water vapor contribution time series and the ocean water vapor contribution time series includes: verifying whether the sum of the land water vapor contribution time series and the ocean water vapor contribution time series is consistent with the land precipitation time series data.

3. The method for predicting land precipitation under a changing environment according to claim 1, characterized in that: The reanalysis of daily land precipitation data based on the ERA5 dataset obtains land precipitation time series data, including: Define the test statistic for land precipitation time series data; Obtaining a standardized test statistic based on the test statistic; Obtaining a qualitative change trend of land precipitation based on the standardized test statistic; The Sen's Slope test is performed on the land precipitation time series data to calculate the trend slope and obtain the quantitative change trend of land precipitation.

4. The method for predicting land precipitation under a changing environment according to claim 1, characterized in that: The method of obtaining the land precipitation time series data based on the reanalysis daily-scale land precipitation data of the ERA5 data set also includes: dividing the reanalysis daily-scale land precipitation data of the ERA5 data set from 1941 to 2023 into four stages to calculate the land precipitation time series data, and the four stages include: P1: 1941-1960, P2: 1961-1980, P3: 1981-2000, P4: 2001-2023.

5. The method for predicting land precipitation under a changing environment according to claim 4, characterized in that: The WAM-2layers water vapor tracking model is used to obtain the land water vapor contribution time series and the ocean water vapor contribution time series, including: tracking and calculating water vapor based on the following formula: in, Indicates the moisture content of the bottom or top layer of a pixel. and are the lateral and longitudinal wind speeds, and Represents the water vapor balance content in the horizontal and vertical directions, is the evaporation rate, is the amount of precipitation, is the residual, It is the vertical transport of water between the upper and lower layers.

6. The method for predicting land precipitation under a changing environment according to claim 5, characterized in that: Based on the verified land water vapor contribution time series and the WAM-2layers water vapor tracking model, the water vapor contribution time series of all Köppen climate zones are calculated respectively; based on the verified ocean water vapor contribution time series and the WAM-2layers water vapor tracking model, the water vapor contribution time series of the four oceans are calculated respectively, including: Based on the Köppen climate zone classification vector data, ocean boundary vector data and WAM-2layers water vapor tracking model, the evaporation of all Köppen climate zones and four oceans tracked to land precipitation is tracked and calculated respectively; Based on the evaporation traced to land precipitation and the total land precipitation in all the Köppen climate zones and the four oceans, the water vapor contribution ratios of the corresponding Köppen climate zones and oceans are calculated respectively; Based on the water vapor contribution ratios of the corresponding Köppen climate zones and oceans and the land water vapor contribution time series and the ocean water vapor contribution time series, the water vapor contribution time series of the corresponding Köppen climate zones and oceans are calculated respectively according to the four stages.

7. The method for predicting land precipitation under a changing environment according to claim 1, characterized in that: The principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans are extracted separately, including: Based on the water vapor contribution time series of all the Köppen climate zones and the four oceans, the first three principal components with cumulative variance contribution rates greater than 60% were extracted.

8. The method for predicting land precipitation under a changing environment according to claim 1, characterized in that: The main climate index data include ENSO index data, IDO index data and NAO index data; The principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans were correlated with the main climate index data, and the calculation formula for key climate index data was obtained as follows: In the formula, and represent water vapor contribution data and climate index data respectively, and They represent the average values ​​of the principal component data of the water vapor contribution time series and the climate index data, respectively.

9. The method for predicting land precipitation under a changing environment according to claim 1, characterized in that: Performing an interpolation operation on the key climate index data or the main climate index data; The time scale of the key climate index data or the main climate index data is made the same as the time scale of the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, and the water vapor contribution time series of the four oceans.

10. A land precipitation prediction system under a changing environment, characterized in that: The system comprises: The data acquisition module is used to obtain the land precipitation time series data based on the reanalysis daily-scale land precipitation data of the ERA5 data set; obtain the land water vapor contribution time series and the ocean water vapor contribution time series based on the WAM-2layers water vapor tracking model; perform cross-validation based on the land precipitation time series data and the land water vapor contribution time series and the ocean water vapor contribution time series; if the cross-validation is consistent, calculate the water vapor contribution time series of all Köppen climate zones based on the verified land water vapor contribution time series and the WAM-2layers water vapor tracking model; calculate the water vapor contribution time series of the four oceans based on the verified ocean water vapor contribution time series and the WAM-2layers water vapor tracking model; extract the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans respectively; perform correlation analysis on the principal components of the water vapor contribution time series of all Köppen climate zones and the four oceans with the main climate index data to obtain key climate index data; A training module, for establishing a model based on a long short-term memory network, using the land precipitation time series data, the water vapor contribution time series of all Köppen climate zones, the water vapor contribution time series of the four oceans and key climate index data as training sets and test sets to train and test the model; The prediction module is used to predict future land precipitation based on the trained and tested model.

Images