Rainfall observation method and device, computer system and medium
By integrating satellite and base station signals using artificial intelligence technology, a rainfall reconstruction and calibration module was constructed, which solved the problems of high equipment cost and low resolution for rainfall observation in remote areas, and achieved real-time, high-precision rainfall observation.
Patent Information
- Application Number
- CN202511648327.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-12
- Publication Date
- 2025-12-12
- Estimated Expiration
- 2045-11-12
AI Technical Summary
Existing technologies for rainfall observation in remote areas suffer from high equipment costs, maintenance difficulties, low spatial resolution, and data gaps, making it difficult to achieve high-precision, real-time regional rainfall observation.
Artificial intelligence technology is used to integrate satellite remote sensing, reanalysis datasets and mobile communication base station signals to construct a rainfall reconstruction model and calibration module. By fusing multi-source data, a rainfall attenuation and intensity model is optimized to achieve real-time and high-precision rainfall observation.
It provides a real-time, dynamic, and high-precision rainfall observation mode, overcoming the shortcomings of satellite remote sensing and microwave link observation, and improving the accuracy and reliability of regional rainfall observation.
Smart Images

Figure CN121118009A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of rainfall observation technology, and more specifically to a rainfall observation method and apparatus based on artificial intelligence coupling satellite and ground base station signals. Background Technology
[0002] Rainfall, as a key meteorological element in agriculture, forestry, water conservancy, hydrology, and mining, has a significant impact on industrial and agricultural production, water resource regulation, flood control and disaster reduction, and engineering safety. High-precision rainfall observation data is not only fundamental to urban planning, water resource development, and disaster early warning, but also a crucial support for maintaining ecological balance and ensuring sustainable economic and social development. Currently, there is an urgent need for real-time and accurate rainfall monitoring data in fields such as water conservancy, meteorology, agriculture, and forestry, especially in remote areas where the severe scarcity of basic rainfall data has become a significant factor restricting the rational use of water resources and ecological environmental protection.
[0003] Currently, rainfall data is mainly obtained through three technologies: satellite remote sensing, weather radar, and ground rain gauges. However, each has significant limitations. While satellite remote sensing has the advantage of wide coverage, its spatial resolution is low, making it difficult to meet the needs of refined monitoring. Weather radar, although providing rainfall information with high spatiotemporal resolution, is susceptible to terrain obstruction and beam lifting, and its high maintenance costs hinder large-scale deployment. Ground rain gauges, while providing the most accurate single-point rainfall data, suffer from high equipment costs, short lifespans, and maintenance difficulties, resulting in insufficient deployment density in remote areas. This data gap problem in a multi-source observation system urgently requires technological innovation to overcome the bottlenecks of existing monitoring methods and build a more accurate, efficient, and economical rainfall observation network.
[0004] Microwave signals are affected by the atmospheric medium during propagation in the near-surface atmosphere, resulting in absorption, attenuation, scattering, polarization, and path bending. Rain and snow particles have a particularly significant impact on microwaves. In recent years, with the rapid development of mobile communication technology, mobile communication base stations offer advantages such as high density and low cost. Some researchers have proposed methods to invert rainfall intensity using the attenuation characteristics of microwave signals. For example, patent CN109697323B discloses a rainfall observation method that integrates satellite remote sensing and mobile communication base station signals. This method leverages the advantages of satellite remote sensing rainfall observation technology and microwave link real-time rainfall intensity observation technology, providing a real-time, dynamic, and high-precision rainfall observation mode.
[0005] Satellite remote sensing and mobile communication base stations provide massive amounts of data, but existing literature rarely makes full use of artificial intelligence technology to mine this massive information, resulting in significant uncertainty and arbitrariness in rainfall observation. On the other hand, existing literature fails to fully integrate the advantages of meteorological satellites, reanalysis datasets, and mobile communication base stations, and cannot solve the problem of high-reliability observation of regional rainfall. Summary of the Invention
[0006] This invention proposes a rainfall observation method, device, computer system, and medium based on artificial intelligence coupling satellite and ground base station signals, which solves the problem of high-reliability observation of regional rainfall.
[0007] Firstly, a rainfall observation method includes: acquiring rainfall data retrieved from a satellite system, a reanalysis dataset, meteorological and hydrological data output from an Earth system model, and microwave signal related data from regional mobile communication base stations; loading a rainfall reconstruction model constructed by an artificial neural network, using the reanalysis dataset and meteorological and hydrological data output from the Earth system model as input, to reconstruct a long series of gridded rainfall data for historical periods in the region; for each microwave link in the region, calculating the corresponding link rainfall intensity using different rainfall attenuation and intensity models, and further calculating the regional rainfall under different rainfall attenuation and intensity models; using the reconstructed long series of gridded rainfall data for historical periods in the region as the observation ground truth, and using the historical periods corresponding to different rainfall attenuation and intensity models... Regional rainfall is used as a simulated value, and the Nash efficiency coefficient, root mean square error, and / or Pearson correlation coefficient are used as evaluation indicators to select the optimal rainfall attenuation and intensity model. Rainfall attenuation data is derived based on microwave signal correlation data. Real-time dynamic rainfall attenuation data from each mobile communication base station in the region are used as the real-time microwave link rainfall observation model with the optimal rainfall attenuation and intensity model. Real-time rainfall intensity information of each microwave link is calculated, and real-time regional rainfall data is further generated. The real-time regional rainfall data, the rainfall data inverted from the satellite system at the previous moment, the real-time microwave signal correlation data observed by each mobile communication base station, and the microwave signal correlation data observed at the previous moment are input into the calibration module constructed by the artificial neural network, and the corrected real-time regional rainfall is output.
[0008] Secondly, a rainfall observation device includes: a multi-source data acquisition module configured to acquire rainfall data retrieved from a satellite system, a reanalysis dataset, meteorological and hydrological data output from an Earth system model, and microwave signal related data from regional mobile communication base stations; a rainfall reconstruction model constructed by an artificial neural network configured to reconstruct a long-series gridded rainfall data from a region's historical period using the reanalysis dataset and meteorological and hydrological data output from an Earth system model as input; and a rainfall attenuation and intensity model selection module configured to: for each microwave link within the region, use different rainfall attenuation and intensity models to estimate the corresponding link rainfall intensity, and further estimate the regional rainfall under different rainfall attenuation and intensity models, using the reconstructed long-series gridded rainfall data from a region's historical period as the observed true value, and selecting the corresponding rainfall attenuation and intensity models. Historical regional rainfall is used as the simulated value, and the Nash efficiency coefficient, root mean square error, and / or Pearson correlation coefficient are used as evaluation indicators to select the optimal rainfall attenuation and intensity model. The real-time rainfall observation module is configured to: derive rainfall attenuation data based on microwave signal correlation data; use real-time dynamic rainfall attenuation data from each mobile communication base station in the region; use the optimal rainfall attenuation and intensity model as the real-time microwave link rainfall observation model; derive real-time rainfall intensity information for each microwave link; and further generate real-time regional rainfall data. A calibration module constructed from an artificial neural network is configured to: take real-time regional rainfall data, rainfall data inverted from the satellite system at the previous moment, real-time microwave signal correlation data observed by each mobile communication base station, and microwave signal correlation data observed at the previous moment as inputs; and output the corrected real-time regional rainfall.
[0009] Thirdly, a computer system includes a processor and a memory, the memory storing one or more computer program modules configured to be executed by the processor to implement the rainfall observation method.
[0010] Fourthly, a non-volatile computer-readable storage medium stores a computer program that, when executed by a computer, enables the aforementioned rainfall observation method.
[0011] This invention is scientifically sound and closely aligned with practical engineering applications. It fully leverages the advantages of satellite remote sensing rainfall observation technology, reanalysis datasets, and microwave link real-time rainfall intensity observation technology, overcoming the shortcomings of low spatial resolution in satellite observations and the difficulty in verifying microwave link observation data. By fusing satellite remote sensing and mobile communication base station signals, it provides a real-time, dynamic, and high-precision rainfall observation mode.
[0012] This invention provides an important and highly operable reference for flood control, disaster reduction, and watershed planning. It utilizes reanalysis data and Earth system model output data to construct a deep learning model of satellite-retrieved rainfall, thereby reconstructing a gridded long-series rainfall dataset. Then, using historical rainfall attenuation information as input, it obtains corresponding historical rainfall simulation series through different rainfall attenuation and intensity models. The weighted link method is then used to deduce regional rainfall under different rainfall attenuation and intensity models. The optimal rainfall attenuation and intensity model is selected from the long-series gridded rainfall data. Furthermore, a deep learning model is constructed using satellite-retrieved data and rainfall data retrieved from the rainfall attenuation and intensity model to achieve real-time correction of regional rainfall. Finally, using the selected rainfall attenuation and intensity model and real-time mobile signals as input, real-time rainfall information for small grids and regions is derived, and the real-time corrected regional average rainfall is obtained, making it more practical. Attached Figure Description
[0013] Figure 1 This is a flowchart of a rainfall observation method based on artificial intelligence coupling satellite and ground base station signals in one embodiment of the present invention.
[0014] Figure 2 This is a schematic diagram of the spatial distribution of mobile communication base stations.
[0015] Figure 3 This is a schematic diagram of the observation area divided into small grids and microwave links in one embodiment of the present invention.
[0016] Figure 4 This is a schematic diagram of a microwave link within a small grid in another embodiment of the present invention.
[0017] Figure 5 This is a schematic diagram of a small grid that does not contain microwave links in another embodiment of the present invention. Detailed Implementation
[0018] This invention employs reanalysis data and Earth system model output data to construct a first deep learning model for satellite-retrieved rainfall, thereby reconstructing a gridded long-series rainfall dataset. Then, using historical rainfall attenuation information as input, different rainfall attenuation and intensity models are used to obtain corresponding historical rainfall simulation series. The weighted link method is then used to deduce regional rainfall under different rainfall attenuation and intensity models. The optimal rainfall attenuation and intensity model is selected from the long-series gridded rainfall data from historical periods. Furthermore, a second deep learning model is constructed using satellite-retrieved data and rainfall data retrieved from the rainfall attenuation and intensity model to achieve real-time correction of regional rainfall. Finally, the optimized rainfall attenuation and intensity model is used with real-time mobile signals as input to deduce real-time rainfall information for small grids and regions, and to obtain the real-time corrected regional average rainfall. This provides a rainfall observation method and device based on artificial intelligence coupled with satellite and ground base station signals. The specific process is detailed in [link to detailed process description]. Figure 1 .
[0019] Step 1: Obtain rainfall data retrieved from meteorological satellites, meteorological and hydrological data from reanalysis datasets, meteorological and hydrological data output from Earth system models, and relevant data on microwave signals from mobile communication base stations in the study area.
[0020] For each grid point in the study area, hourly meteorological and hydrological observation data from 1950 to 2024 were collected from the European Centre for Medium-Range Weather Forecasts' Generation 5 Atmospheric Reanalysis Dataset (ERA5-Land), specifically including air temperature, air pressure, dew point temperature, snowfall, precipitation, runoff depth, shortwave radiation, and longwave radiation.
[0021] Obtain hourly-scale data on temperature, specific humidity, relative humidity, snowfall, precipitation, runoff depth, shortwave radiation, and longwave radiation for the historical period (1950-2014) under the Earth System Model; for the period 2015-2024, obtain hourly-scale data on temperature, specific humidity, relative humidity, snowfall, precipitation, runoff depth, shortwave radiation, and longwave radiation under the SSP370 shared socio-economic path under the Earth System Model; for example, the selected Earth System Model is the CAS-ESM2.0 model, which is a highly complex comprehensive numerical simulation tool based on the physical, chemical, and biological processes in the Earth's spheres and the laws governing the exchange of matter and energy between the spheres.
[0022] Obtain rainfall data retrieved from the Global Precipitation Measurement Mission (GPM) satellite system, such as the GPM IMERG-Final version of the satellite product, which provides hourly rainfall datasets with a spatial resolution of 0.1 degrees covering the study area from 2000 to 2024.
[0023] The study collected the latitude and longitude coordinates of mobile communication base stations within the research area, along with related microwave signals. The microwave signal data comprised all data collected since the station's establishment, generally from 1950 onwards. The microwave signal data included real-time data reflecting signal attenuation and quality, such as dynamically changing received signal level (RSL), bit error rate (BER), and signal-to-noise ratio (SNR), as well as fixed link characteristic parameters (length, frequency, polarization, location coordinates, elevation angle, etc.). Precise timestamps were also included for all data points.
[0024] Step 2: Load the rainfall reconstruction model (first deep learning model) to reconstruct the long series of gridded rainfall data for the study area using the data from Step 1.
[0025] The Clausius-Clapeyron thermodynamic equation is defined as follows: (1) in, Let K be the first integration constant, taken as 273.16 K; The second integration constant is set to 611 Pa. Let be the latent heat of vaporization constant, and take . ; Let be the water vapor gas constant, and take . ; T Temperature is the input variable to the Clausius-Clapeyron thermodynamic equation.
[0026] By substituting the air temperature and dew point temperature from the ERA5-Land dataset for each grid point in the study area into the Clausius-Clapeyron thermodynamic equation as input variables, the relative humidity (RH) for each grid point was calculated. ,in T dew Dew point temperature, T 2m The temperature is 2m.
[0027] wet q The ratio of water vapor mass to total air mass mass is given by ERA5-Land pressure. p The formula for calculating the dew point temperature is as follows: (2) The bilinear interpolation method was used to interpolate the meteorological and hydrological variables from the ERA5-Land dataset and the Earth System Model output to a spatial resolution of 0.1 degrees.
[0028] An artificial neural network was used to construct a rainfall reconstruction model, whose driving factors included two categories: 1) variables obtained or derived from ERA5-Land, including air temperature, relative humidity, snowfall, precipitation, runoff depth, shortwave radiation and longwave radiation; 2) Earth system model output data, including air temperature, relative humidity, snowfall, precipitation, runoff depth, shortwave radiation and longwave radiation.
[0029] Based on the driving factors and rainfall data retrieved from the GPM satellite system, an artificial neural network is constructed, considering a lag of several hours. Specifically, the GPM-retrieved rainfall at each grid point within the study area, the driving factors at the same time, and the driving factors at several lag times (e.g., 3 hours) are all used as calibration data for the artificial neural network to calibrate an artificial intelligence model (rainfall reconstruction model) covering the study area. When calibrating the rainfall reconstruction model, physical constraints on the rainfall variables are further added, as follows: (3) In the formula: These represent the simulations of the rainfall reconstruction model. k The grid point at the th grid pointt +1 and t Hourly rainfall at any given time, in mm; express , The smaller value in the model is used; specifically, the gradient descent method is employed to calibrate the rainfall reconstruction model and obtain the optimal parameters.
[0030] Using the ERA5-Land dataset and Earth system model output data since 1950, a calibrated rainfall reconstruction model was used to reconstruct daily rainfall datasets (e.g., hourly rainfall) for each grid point (e.g., 0.1 degrees) within the study area from 1950 to 2024, as shown below: (4) in, express t Time of the first k The reconstructed rainfall at each grid point express t Time of the first k Input variables for each grid point, express t- 1st moment k Input variables for each grid point, express t- 2nd moment k Input variables for each grid point, express t- 3rd moment k Input variables for each grid point; F This represents the rainfall reconstruction model.
[0031] Step 3: For each microwave link in the study area, different rain attenuation and rainfall intensity models are used to establish the relationship between rain attenuation and rainfall intensity, and the corresponding link rainfall intensity is calculated.
[0032] like Figure 2 As shown, base stations are generally distributed in a honeycomb pattern, with distances between them typically within 10km. Since the spatial distribution of the central base station and surrounding base stations is relatively uniform, microwave links are generally established between the central base station and surrounding base stations. Microwave links can also be established between adjacent surrounding base stations as needed. The average rain attenuation rate of the link is calculated based on the received and transmitted power of adjacent base stations. , length is L Received power on the link With transmission power With rain attenuation rate The relationship is: (5) in, Represents the integral over length.
[0033] From the above formula, the average rain attenuation rate of the link can be derived. for: (6) Furthermore, a rainfall attenuation and rainfall intensity model was established, and the link-average rainfall intensity was derived. R。
[0034] In existing technologies, different rain attenuation and rainfall intensity models are applicable to different frequency bands and different systems of ground-to-ground and ground-to-air links. It is difficult to determine a suitable rain attenuation and rainfall intensity model before evaluation using large-scale satellite remote sensing data. Therefore, different models proposed at home and abroad are used to establish the relationship between rain attenuation and rainfall intensity, that is, the link average rain attenuation rate. and average rainfall intensity R The relationship between them was determined, and the average rainfall intensity of each link was calculated. R Models include, but are not limited to, Crane Global model, Assis Einloft Improved model, SAM model, DAH model, ITU-R P.618 model, China model, ITU-R P.838 model, Moupfouma model, EXCELL model, and Garcia-Lopez model.
[0035] For example, the relationship between rainfall attenuation and rainfall intensity can be established using the ITU-R P.618 model, as shown in the following equation: (7) In the formula, k and The parameter can be estimated using existing methods recommended in ITU reports. According to the above formula, the link average rain attenuation rate... The average rainfall intensity of the link was obtained by reverse calculation. R This means that the average rainfall intensity of the link can be derived using the ITU-R P.618 model.
[0036] In practice, the relationship between rainfall attenuation and rainfall intensity is established, and can be implemented according to the specific model selected. This invention will not elaborate on this point.
[0037] Step 4: Based on the long series of gridded rainfall data reconstructed in Step 2, select the rainfall attenuation and rainfall intensity model suitable for the study area.
[0038] like Figure 3 As shown, considering the geographical conditions of the area, the density of base station deployment, and other availability information, the observation area is reasonably divided into grids of appropriate scale. In this embodiment, it is assumed that the area is divided into G identical grids. Figure 3 Taking a microwave link as an example, point F is the signal transmitting base station and point Z is the signal receiving base station, which constitutes a single microwave link.
[0039] For each rainfall attenuation and intensity model established in step 3, the rainfall amount corresponding to each grid under different models can be obtained: such as Figure 4 As shown, for each grid, if it contains microwave links, under different rain attenuation and rainfall intensity models, the historical microwave link information (microwave signals) included in the grid can be extracted based on the average rainfall intensity of the corresponding links; for example, a certain grid Gr contains M There are 10 microwave links, and the length of each microwave link in the grid is denoted as . L j ( j =1, ... , M The average rainfall intensity of each link, calculated in step 2, is expressed as follows: R j ( j =1, ... , M The weighted link method was used to obtain the rainfall for this grid. Prep for: (8) In the formula, sumL The length of each microwave link in the grid within that grid. L j sum.
[0040] like Figure 5 As shown, some grids in the region do not contain microwave links. For such grids, a reasonable radius threshold is selected, and microwave links within the radius threshold are searched. Then, the reciprocal distance weighting method is used to calculate the rainfall for that grid. For example, M microwave links are found within the radius threshold of a certain grid, and the vertical distance from the grid center point to each link is... S j ( j =1, ... , M If the rainfall in the grid is such that the reciprocal distance weighting method is used, then the rainfall in that grid can be obtained. Prep for: (9) In the formula, sumS For the grid sum.
[0041] Figure 5 In a certain grid G Three microwave links were obtained by searching within a radius threshold of 0. The perpendicular distance from the grid center point to each link is... S 1. S 2. S 3. Rainfall in this grid .
[0042] For each rainfall attenuation and intensity model established in step 3, the corresponding historical regional rainfall under different models can be obtained.
[0043] The number of rain attenuation and rainfall intensity models is N The number of grids is G Let the first g The grid passes through the first i ( i =1,..., N The grid rainfall obtained from the rainfall attenuation and rainfall intensity models is: Then the first one can be obtained. i Historical regional rainfall corresponding to the rainfall attenuation and intensity model. AR i for: (10) Furthermore, the hourly rainfall at each grid point in the study area since 1950, which was reconstructed in step 2, was used as the observed true value. The regional rainfall corresponding to different rainfall attenuation and rainfall intensity models derived from rainfall attenuation data in formula (10) was used as the simulated value. The combined index of Nash (Nash efficiency coefficient), RMSE (root mean square error), and Pearson correlation coefficient was used as the evaluation index to select the best rainfall attenuation and rainfall intensity model.
[0044] (11) In the formula: Index is an evaluation index used to select the optimal rain attenuation and rainfall intensity model. The higher the Index, the better the model performance.
[0045] Step 5: Based on the selected rainfall attenuation and intensity model, a correction module for the rainfall attenuation and intensity model is constructed by further considering the rainfall data retrieved from meteorological satellites.
[0046] Using the optimized rain attenuation and intensity model obtained in step 4, and inputting the microwave link information of the study area, the average regional rainfall over a historical period (in this embodiment, the historical period refers to the period from the establishment of the mobile base station to 2024) can be obtained, denoted as . AAR t ,in t For a moment.
[0047] Furthermore, based on the hourly rainfall dataset from 2000 to 2024 provided by the GPM IMERG-Final version collected in step 1, the average hourly rainfall series of the study area is derived. Furthermore, using the average hourly rainfall series of the study basin retrieved by the meteorological satellite as a benchmark, the microwave signals and AAR observed by each mobile communication station in the study area are used as inputs, and considering the time lag effect (such as 1 hour), a calibration module for the rainfall attenuation and intensity model is established through a calibration convolutional neural network model (second deep learning model).
[0048] Based on the data obtained from the above steps, a convolutional neural network model is constructed. Specifically, this method utilizes the average rainfall in the study area derived from an optimized rainfall attenuation and intensity model and microwave signals observed by various mobile communication stations as explanatory variables, with the average hourly rainfall series of the study basin retrieved from meteorological satellites as the target variable. The method first preprocesses the rainfall data and microwave signals, including data partitioning, normalization, and reshaping, to adapt to the input format of the convolutional neural network. Then, a three-layer neural network architecture containing convolutional layers, pooling layers, and fully connected layers is constructed. The convolutional layers extract spatial features from the data, the pooling layers reduce data dimensionality, and the fully connected layers achieve feature integration and rainfall simulation. Finally, a temperature scaling method is used to calibrate the model's output probability to improve the accuracy of the simulated probability. The convolutional neural network constructed in this implementation is used to correct the errors of the rainfall attenuation and intensity model and improve the accuracy of capturing the average rainfall in the study area.
[0049] The convolutional neural network model constructed using the method described above can be represented as follows: (12) In the formula: Indicates the period from 2000 to 2024 t The hourly average rainfall in the study area after time correction. These represent the periods from 2000 to 2024. t The regional average rainfall at time t-1 was calculated using the rain attenuation and rainfall intensity model optimized in step 4. During the period from 2000 to 2024 respectively t The microwave signals observed by all mobile communication stations in the study area at time t-1; F DL The diagram shows the calibrated convolutional neural network model. To optimize the model's parameters, the minimum batch gradient descent method, a conventional technique in this field, is used to train the convolutional neural network model.
[0050] Step 6: Based on the selected rain attenuation and intensity model and the calibrated correction module, observe real-time rain intensity information with high spatial resolution based on real-time dynamic rain attenuation data from mobile communication base stations.
[0051] Using real-time dynamic rain attenuation data from each mobile communication base station in the study area, and taking the rain attenuation and rainfall intensity model selected in step 4 as the real-time microwave link rainfall observation model, the real-time rainfall intensity information of each microwave link can be derived based on the optimized relationship between rain attenuation and rainfall intensity. At the same time, the small grid and regional rainfall data can be derived in the same way as in step 4.
[0052] Furthermore, the derived real-time regional rainfall data, the rainfall data retrieved from the previous hour, the real-time microwave signals observed from each mobile communication base station, and the observation signals from the previous hour are input into the convolutional neural network model calibrated in step 5, thereby obtaining the corrected real-time regional rainfall using the following formula.
[0053] (13) In the formula: Indicates the corrected study area in the 1st... t Average hourly rainfall over time. express , The smaller value in the formula is used; this formula takes into account the possible estimation bias of the correction model in terms of the probability of rainfall events, and fully integrates the rainfall inversion information before and after correction.
[0054] The invention in this embodiment is as follows: Using reanalysis data and Earth system model output data, a first deep learning model is constructed based on satellite-retrieved rainfall data to reconstruct a gridded long-series rainfall dataset. Then, using historical rainfall attenuation information as input, a corresponding historical rainfall simulation series is obtained through different rainfall attenuation and intensity models. The regional rainfall under different rainfall attenuation and intensity models is then calculated using the weighted link method. The optimal rainfall attenuation and intensity model is selected from the long-series gridded rainfall data from historical periods. Furthermore, a second deep learning model is constructed using satellite-retrieved data and rainfall data retrieved from the rainfall attenuation and intensity model to achieve real-time correction of regional rainfall. Finally, using the selected rainfall attenuation and intensity model with real-time mobile signals as input, real-time rainfall information for small grids and regions is calculated, and the real-time corrected regional average rainfall is obtained.
[0055] The present invention also proposes a rainfall observation device, which includes a multi-source data acquisition module, a rainfall reconstruction model, a rainfall attenuation and rainfall intensity model screening module, a real-time rainfall observation module, and a calibration module.
[0056] The multi-source data acquisition module is configured to acquire rainfall data retrieved from the satellite system, reanalysis datasets, meteorological and hydrological data output from the Earth system model, and microwave signals from regional mobile communication base stations.
[0057] The rainfall reconstruction model is constructed using an artificial neural network, which is configured to reconstruct long-series gridded rainfall data from regional historical periods by taking meteorological and hydrological data from the reanalysis dataset and Earth system model outputs as inputs.
[0058] The rain attenuation and intensity model selection module is configured as follows: for each microwave link in the region, different rain attenuation and intensity models are used to estimate the corresponding link rainfall intensity, and further, the regional rainfall under different rain attenuation and intensity models is estimated. The long series of gridded rainfall data from the reconstructed regional historical period is used as the observed true value, and the regional rainfall corresponding to the historical period of different rain attenuation and intensity models is used as the simulated value. The Nash efficiency coefficient, root mean square error, and / or Pearson correlation coefficient are used as evaluation indicators to select the best rain attenuation and intensity model.
[0059] The real-time rainfall observation module is configured as follows: it uses real-time dynamic rainfall attenuation data from each mobile communication base station in the area, uses the best rainfall attenuation and rainfall intensity model as the real-time microwave link rainfall observation model, calculates the real-time rainfall intensity information of each microwave link, and further generates real-time regional rainfall data.
[0060] The calibration module is constructed using an artificial neural network. Its configuration is as follows: it takes real-time regional rainfall data, rainfall data retrieved from the satellite system at the previous moment, real-time microwave signals observed by each mobile communication base station, and microwave signals observed at the previous moment as inputs, and outputs the corrected real-time regional rainfall.
[0061] This rainfall observation device is a program product coupled with the rainfall observation method. The detailed algorithm implementations of the core functional modules within this device are described in the above-described method embodiment.
[0062] The present invention also provides a computer system including a processor and a memory. The memory and the processor can be interconnected via a bus system and / or other forms of connection. The processor may include any one or more processors such as a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP). The memory may include volatile memory, such as random access memory (RAM). The memory may also include non-volatile memory. volatile memory), such as read-only memory (read... The memory can be ROM, flash memory, hard disk drive (HDD), or solid state drive (SSD). The memory stores executable program code, which the processor executes to implement the wind field identification method for the parachute system. In other words, the memory contains instructions for executing the rainfall observation method.
[0063] The present invention also provides a computer-readable storage medium. For example, this computer-readable storage medium is a non-transitory computer-readable storage medium, such as a read-only memory (ROM). Read-only memory (ROM), random access memory (RAM), and compact disc (CD-ROM) are also mentioned. Only memory, CD ROM, magnetic tape, floppy disk, and optical data storage devices, etc. This computer-readable storage medium is used to store non-transitory computer-readable instructions, which, when executed by a computer, can perform one or more steps of the above-described rainfall observation method.
Claims
1. A method for observing rainfall, characterized in that, include: Acquire rainfall data retrieved from satellite systems, reanalysis datasets, meteorological and hydrological data output from Earth system models, and microwave signal data from regional mobile communication base stations; A rainfall reconstruction model constructed by an artificial neural network is loaded, and meteorological and hydrological data output from the reanalysis dataset and the Earth system model are used as inputs to reconstruct long-series gridded rainfall data of the region's historical periods. For each microwave link in the region, different rain attenuation and intensity models are used to calculate the corresponding link rainfall intensity, and the regional rainfall under different rain attenuation and intensity models is further calculated. Using the long series of gridded rainfall data from the reconstructed regional historical period as the observed true values, and the regional rainfall corresponding to different rainfall attenuation and intensity models as the simulated values, the Nash efficiency coefficient, root mean square error and / or Pearson correlation coefficient are used as evaluation indicators to select the best rainfall attenuation and intensity model. Rainfall attenuation data is derived based on microwave signal correlation data. Real-time dynamic rainfall attenuation data from each mobile communication base station in the region are used. The optimal rainfall attenuation and rainfall intensity model is used as the real-time microwave link rainfall observation model. Real-time rainfall intensity information of each microwave link is calculated, and real-time regional rainfall data is further generated. The calibration module, constructed using an artificial neural network, inputs real-time regional rainfall data, rainfall data retrieved from the satellite system at the previous moment, real-time microwave signal data observed by each mobile communication base station, and microwave signal data observed at the previous moment. The module then outputs the corrected real-time regional rainfall.
2. The method according to claim 1, characterized in that, The driving factors of the rainfall reconstruction model include two categories: 1) variables obtained or derived from the reanalysis dataset, including air temperature, relative humidity, snowfall, precipitation, runoff depth, shortwave radiation, and longwave radiation; 2) data output by the Earth system model, including air temperature, relative humidity, snowfall, precipitation, runoff depth, shortwave radiation, and longwave radiation. The rainfall data retrieved from satellite systems at each grid point within the region, the driving factors at the same time, and the driving factors at several hourly lag times are all used as calibration data for the rainfall reconstruction model. When calibrating the rainfall reconstruction model, the following formula is used as a constraint: In the formula: These represent the simulations of the rainfall reconstruction model at the first and second stages. k The grid point at the th grid point t +1 and t Hourly rainfall at any given time, in mm; express , The smaller value in the range.
3. The method according to claim 2, characterized in that, Using historical reanalysis datasets and meteorological and hydrological data output from Earth system models, the calibrated rainfall reconstruction model is driven to reconstruct long-series gridded rainfall data for the region's historical periods.
4. The method according to claim 1, characterized in that, Using the regional average rainfall and microwave signals observed by each mobile communication station as explanatory variables, and the regional average hourly rainfall series retrieved by the satellite system as the target variable, the calibration module is calibrated. The calibration module is represented as follows: In the formula: Representing historical periods t The hourly average rainfall in the region after time correction. Representing historical periods t and t The regional average rainfall at time -1 was estimated using the optimal rain attenuation and rainfall intensity model. Historical periods t and t Microwave signals observed by all mobile communication stations in the region at time -1; F DL The calibration module is indicated after calibration.
5. The method according to claim 4, characterized in that, The corrected real-time regional rainfall is described by the following mathematical formula: In the formula: Indicates the area after correction at the 1st t Average hourly rainfall over time. express , The smaller value in the range.
6. The method according to claim 1, characterized in that, The region is divided into several grids, and the rainfall of the grid containing microwave links is obtained by weighted link method.
7. The method according to claim 6, characterized in that, For grids that do not contain microwave links, a radius threshold is selected, microwave links within the radius threshold are searched, and then the rainfall of the grid is obtained by using the inverse distance weighting method.
8. A rainfall observation device, characterized in that, include: The multi-source data acquisition module is configured to acquire rainfall data retrieved from satellite systems, reanalysis datasets, meteorological and hydrological data output from Earth system models, and microwave signal data from regional mobile communication base stations. The rainfall reconstruction model, constructed using artificial neural networks, is configured to reconstruct long-series gridded rainfall data from historical periods in a region, taking reanalysis datasets and meteorological and hydrological data output from Earth system models as inputs. The rain attenuation and intensity model selection module is configured as follows: for each microwave link in the region, different rain attenuation and intensity models are used to estimate the corresponding link rainfall intensity, and further, the regional rainfall under different rain attenuation and intensity models is estimated. The long series of gridded rainfall data of the reconstructed regional historical period is used as the observed true value, and the regional rainfall corresponding to the historical period of different rain attenuation and intensity models is used as the simulated value. The Nash efficiency coefficient, root mean square error and / or Pearson correlation coefficient are used as evaluation indicators to select the best rain attenuation and intensity model. The real-time rainfall observation module is configured as follows: based on microwave signal correlation data, rainfall attenuation data is derived; real-time dynamic rainfall attenuation data of each mobile communication base station in the region is used; the optimal rainfall attenuation and rainfall intensity model is used as the real-time microwave link rainfall observation model; real-time rainfall intensity information of each microwave link is derived; and real-time regional rainfall data is further generated. as well as The calibration module, constructed using an artificial neural network, is configured to take real-time regional rainfall data, rainfall data retrieved from the satellite system at the previous moment, real-time microwave signals observed by each mobile communication base station, and microwave signals observed at the previous moment as inputs, and output the corrected real-time regional rainfall.
9. A computer system, characterized in that, It comprises: a processor and a memory, the memory storing one or more computer program modules configured to be executed by the processor to implement the rainfall observation method according to any one of claims 1-7.
10. A non-volatile computer-readable storage medium, characterized in that, It contains a computer program that, when executed by a computer, enables the implementation of the rainfall observation method according to any one of claims 1-7.
Citation Information
Patent Citations
A method for rainfall observation that integrates satellite remote sensing and mobile communication base station signals
CN109697323B
A rainfall observation method integrating satellite remote sensing and mobile communication base station signals
CN109697323A
Rainfall estimation method based on microwave rain attenuation and rainfall monitoring system
CN111666656A
Rain intensity self-adaptive estimation system and method based on clustering rain attenuation relation
CN113177360A
Rainfall monitoring method based on mobile phone signal
CN116893451A
Cited By
Satellite and ground base station signal coupled drought and flood sudden turning early warning method and system
CN121640646A
Signal interference strategy generation method and device based on deep learning, and medium
CN122151009A
