A flood forecasting method and device based on artificial intelligence and satellite remote sensing

Abstract

The present invention relates to the technical field of data prediction and processing, and in particular to a flood forecasting method and device based on artificial intelligence and satellite remote sensing, wherein the method comprises: obtaining long-series meteorological observation data according to a statistical downscaling model; establishing a watershed hydrological model, and using the watershed hydrological model to simulate the preliminary runoff of each sub-basin; constructing a long short-term memory neural network model for each sub-basin; inputting the long-series meteorological observation data into the watershed hydrological model and the long short-term memory neural network model respectively; extracting flood events from the reconstructed long-series runoff process of each sub-basin, and establishing a random forest model including spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics based on the flood events, so as to forecast large-scale floods in areas with scarce data. Thus, the problem that the relevant technology fails to make full use of satellite telemetry meteorological information and cannot solve the problem of long-series flood forecasting in areas with scarce data is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data prediction and processing, and in particular to a flood forecasting method based on artificial intelligence and satellite remote sensing. Background Art

[0002] In recent decades, the emission of greenhouse gases such as carbon dioxide in the atmosphere has increased dramatically, the energy balance and material circulation process of the earth's land and atmosphere have changed, the global temperature has continued to rise, coral reefs and tropical rainforests have died out on a large scale, and the imbalance and inequality of the social and economic systems and ecological environment have become increasingly serious. In order to meet the above challenges, it is urgent to study the formation mechanism of extreme precipitation and flood events and carry out disaster prevention and mitigation work by improving the accuracy of forecasts.

[0003] In the related technology, it is possible to obtain hydrological and meteorological data, determine the hydrological and meteorological characteristics of the target area in the pre-divided area, and carry out flood forecasting by establishing a linear or nonlinear relationship between meteorological variables and floods.

[0004] However, the relevant technology has failed to fully utilize satellite remote sensing meteorological information, failed to consider the errors in runoff simulation caused by human activities, and cannot solve the problem of long-series flood forecasting in areas with scarce data, and urgently needs to be improved. Summary of the invention

[0005] The present invention provides a flood forecasting method based on artificial intelligence and satellite remote sensing to solve the problem that related technologies require a long series of meteorological and hydrological data. However, hydrological and meteorological data are extremely scarce in most areas, and some areas only have a small amount of measured hydrological and meteorological data. Satellite remote sensing mainly uses sensors carried by meteorological satellites to measure meteorological data, but satellite remote sensing technology has the problems of large grid scale and low spatial resolution, which makes it difficult to directly meet actual requirements. At the same time, hydrological models are suitable for simulating runoff processes under natural conditions. Engineering measures such as dams, reservoirs, agricultural irrigation, water diversion and inter-basin water transfer often destroy the consistency of the underlying surface, resulting in large errors in the basin hydrological model, which restricts the accuracy of hydrological simulation. Related technologies fail to make full use of satellite telemetry meteorological information, fail to consider the errors caused by human activities interference in runoff simulation, and cannot solve the problem of long-series flood forecasting in areas with scarce data.

[0006] The first aspect of the present invention provides a flood forecasting method based on artificial intelligence and satellite remote sensing, which is applied to the model building stage and includes the following steps: based on the limited observation data of ground meteorological stations in the scarce data area and the large-scale grid data of satellite remote sensing, a statistical downscaling model is established, and a long series of meteorological observation data is obtained according to the statistical downscaling model; based on the short series of runoff observation data and downscaled meteorological data of each sub-basin in the scarce data area, a basin hydrological model is established, and the preliminary runoff of each sub-basin is simulated using the basin hydrological model to obtain the simulated runoff of each sub-basin; The simulated runoff of the basin is corrected to obtain a corrected simulated runoff series, and a long short-term memory neural network model of each sub-basin is constructed based on the corrected simulated runoff series; the long series of meteorological observation data are respectively input into the basin hydrological model and the long short-term memory neural network model to obtain the reconstructed long series of runoff processes of each sub-basin in the scarce data area; flood events are extracted from the reconstructed long series of runoff processes of each sub-basin, and a random forest model including spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics is established based on the flood events to predict large-scale floods in the scarce data area.

[0007] Optionally, in one embodiment of the present invention, the simulated runoff of each sub-basin is corrected to obtain a corrected simulated runoff series, including: based on each sub-basin in the scarce data area, statistically analyzing the simulated daily runoff process and the measured daily runoff process to determine the lag time that affects the daily measured runoff; based on the lag time that affects the daily measured runoff, correcting the simulated daily runoff process to obtain the corrected simulated runoff series.

[0008] Optionally, in one embodiment of the present invention, the calculation formula of the corrected simulated runoff series is:

[0009] ,

[0010] in, The hydrological model is The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, represents the lag determined by the LSTM model, Represents the LSTM model.

[0011] Optionally, in one embodiment of the present invention, the long series of meteorological observation data are respectively input into the watershed hydrological model and the long short-term memory neural network model to obtain the reconstructed long series of runoff processes for each sub-basin in the data scarce area, including: based on the statistical downscaling model, using the large-scale raster meteorological data of satellite telemetry to obtain the long series of meteorological series at the watershed scale; inputting the long series of meteorological series into the watershed hydrological model to obtain the long series of runoff processes for each sub-basin; using the long short-term memory neural network model to correct the simulated runoff to reconstruct the long series of runoff processes for each sub-basin to obtain the reconstructed long series of runoff processes for each sub-basin.

[0012] Optionally, in one embodiment of the present invention, flood events are extracted from the reconstructed long series of runoff processes of each sub-basin, and a random forest model is established based on the flood events, which takes into account spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics, including: based on the reconstructed long series of runoff processes of each sub-basin, taking a preset quantile of daily runoff as a threshold, and taking the daily runoff exceeding the threshold as the flood event; based on the flood event, obtaining the basin average meteorological variables of each sub-basin in the scarce data area, and establishing the random forest model based on the basin average meteorological variables.

[0013] A second aspect of the present invention provides a flood forecasting method based on artificial intelligence and satellite remote sensing, which is applied in the model application stage and includes the following steps: obtaining meteorological forecast information and reanalyzing meteorological data for each sub-basin in the area with scarce data;

[0014] The meteorological forecast information and the reanalyzed meteorological data are input into a preset random forest model to drive the random forest model to generate large-scale flood forecast results for the data-scarce area, wherein the preset random forest model is constructed by spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics.

[0015] The third aspect of the present invention provides a flood forecasting device based on artificial intelligence and satellite remote sensing, which adopts the flood forecasting method based on artificial intelligence and satellite remote sensing as described above and is applied to the model construction stage, including: an establishment module, which is used to establish a statistical downscaling model based on the limited observation data of ground meteorological stations in the scarce data area and the large-scale grid data of satellite remote sensing, and obtain long-series meteorological observation data according to the statistical downscaling model; a simulation module, which is used to establish a basin hydrological model based on the short-series runoff observation data and downscaled meteorological data of each sub-basin in the scarce data area, and use the basin hydrological model to simulate the preliminary runoff of each sub-basin to obtain the simulated runoff of each sub-basin; A correction module, used to correct the simulated runoff of each sub-basin to obtain a corrected simulated runoff series, and construct a long short-term memory neural network model of each sub-basin based on the corrected simulated runoff series; an input module, used to input the long series of meteorological observation data into the basin hydrological model and the long short-term memory neural network model respectively, to obtain the reconstructed long series runoff process of each sub-basin in the scarce data area; a forecasting module, used to extract flood events from the reconstructed long series runoff process of each sub-basin, and establish a random forest model including spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics based on the flood events, so as to forecast large-scale floods in the scarce data area.

[0016] Optionally, in one embodiment of the present invention, the correction module includes: a determination unit, used to perform statistical analysis on the simulated daily runoff process and the measured daily runoff process based on each sub-basin in the scarce data area to determine the lag time that affects the daily measured runoff; a correction unit, used to correct the simulated daily runoff process based on the lag time that affects the daily measured runoff to obtain the corrected simulated runoff series.

[0017] Optionally, in one embodiment of the present invention, the calculation formula of the corrected simulated runoff series is:

[0018] ,

[0019] in, The hydrological model is The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, represents the lag determined by the LSTM model, Represents the LSTM model.

[0020] Optionally, in one embodiment of the present invention, the input module includes: an acquisition unit, used to acquire a long series of meteorological series at the basin scale based on the statistical downscaling model using the large-scale raster meteorological data sensed by the satellite; an input unit, used to input the long series of meteorological series into the basin hydrological model to obtain the long series of runoff processes for each sub-basin; and a reconstruction unit, used to correct the simulated runoff using the long short-term memory neural network model to reconstruct the long series of runoff processes for each sub-basin to obtain the reconstructed long series of runoff processes for each sub-basin.

[0021] Optionally, in one embodiment of the present invention, the forecast module includes: a generation unit, which is used to reconstruct the long series of runoff processes of each sub-basin, take the preset quantile of daily runoff as a threshold, and regard the daily runoff exceeding the threshold as the flood event; an establishment unit, which is used to obtain the basin-average meteorological variables of each sub-basin in the scarce data area based on the flood event, and establish the random forest model according to the basin-average meteorological variables.

[0022] The fourth aspect of the present invention provides a flood forecasting device based on artificial intelligence and satellite remote sensing, which adopts the flood forecasting method based on artificial intelligence and satellite remote sensing as described, and is applied to the model application stage, including: an acquisition module, used to obtain meteorological forecast information and reanalysis meteorological data for each sub-basin in the scarce data area; a driving module, used to input the meteorological forecast information and the reanalysis meteorological data into a preset random forest model to drive the random forest model to generate large-scale flood forecast results for the scarce data area, wherein the preset random forest model is constructed by spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics.

[0023] A fifth aspect of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a flood forecasting method based on artificial intelligence and satellite remote sensing as described in the above embodiment.

[0024] A sixth aspect of the present invention provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned flood forecasting method based on artificial intelligence and satellite remote sensing.

[0025] The embodiment of the present invention provides an important and highly operable reference basis for flood forecasting in areas with scarce data by coupling the physical mechanism of flood formation, artificial intelligence, satellite remote sensing and other methods, and provides engineering reference value for responding to climate disasters, disaster prevention and mitigation, and integrated water resources management. Thus, the problem that the relevant technology fails to make full use of satellite telemetry meteorological information, fails to consider the errors caused by human activities in runoff simulation, and cannot solve the problem of long-series flood forecasting in areas with scarce data is solved.

[0026] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0028] Figure 1 A flow chart of a flood forecasting method based on artificial intelligence and satellite remote sensing provided in an embodiment of the present invention applied to a model building stage;

[0029] Figure 2 A schematic diagram of ground meteorological station observations, satellite remote sensing data and runoff data according to an embodiment of the present invention;

[0030] Figure 3 A schematic diagram of the change in the correlation coefficient between the daily measured runoff and the simulated runoff under different lag times according to an embodiment of the present invention;

[0031] Figure 4 A flow chart of a flood forecasting method based on artificial intelligence and satellite remote sensing provided in an embodiment of the present invention applied to a model application stage;

[0032] Figure 5 A schematic structural diagram of a flood forecasting device based on artificial intelligence and satellite remote sensing provided according to an embodiment of the present invention applied to a model building stage;

[0033] Figure 6 A schematic structural diagram of a flood forecasting device based on artificial intelligence and satellite remote sensing provided according to an embodiment of the present invention applied to a model application stage;

[0034] Figure 7 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0035] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0036] The following describes a flood forecasting method and device based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention with reference to the accompanying drawings. In view of the fact that the related technologies mentioned in the above background technology fail to make full use of satellite telemetry meteorological information, fail to consider the errors caused by human activity interference to runoff simulation, and fail to solve the problem of long-series flood forecasting in areas with scarce data, the present invention provides a flood forecasting method based on artificial intelligence and satellite remote sensing. In this method, by coupling the physical mechanism of flood formation, artificial intelligence and satellite remote sensing, an important and highly operational reference basis is provided for flood forecasting in areas with scarce data, and an engineering reference value is provided for responding to climate disasters, disaster prevention and mitigation, and comprehensive water resources management. Thus, the problem that the related technologies fail to make full use of satellite telemetry meteorological information, fail to consider the errors caused by human activity interference to runoff simulation, and fail to solve the problem of long-series flood forecasting in areas with scarce data is solved.

[0037] Specifically, Figure 1 A schematic flow chart of a flood forecasting method based on artificial intelligence and satellite remote sensing provided in an embodiment of the present invention.

[0038] like Figure 1 As shown, the flood forecasting method based on artificial intelligence and satellite remote sensing is applied to the model building stage and includes the following steps:

[0039] In step S101, a statistical downscaling model is established based on limited observation data from ground meteorological stations in areas with scarce data and large-scale grid data from satellite remote sensing, and a long series of meteorological observation data is obtained according to the statistical downscaling model.

[0040] It is understandable that if Figure 2 As shown, the embodiment of the present invention provides a schematic diagram of ground meteorological station observations, satellite remote sensing data and runoff data. For the watershed of the embodiment, there is a mismatch between ground meteorological data and runoff data, and they cannot be used directly to calibrate the hydrological model. Therefore, it is necessary to adopt relatively complete satellite telemetry data and statistical downscaling models to obtain corrected long series of satellite telemetry data, so as to establish a hydrological model with the scarce runoff observation series, and finally realize the watershed hydrological simulation.

[0041] In the actual implementation process, the embodiment of the present invention may include:

[0042] 1.1 Based on the limited observation data in the areas with scarce data, the empirical distribution function of the large-scale grid data of satellite remote sensing and the daily precipitation (including days without precipitation) and temperature series measured on the ground are calculated monthly:

[0043] , (1)

[0044] , (2)

[0045] in, d Indicates the month; G represents the observation data of satellite remote sensing; ref represents the period when the ground meteorological station has data, which is recorded as "ground observation period" in the embodiment of the present invention; ecdf represents the cumulative distribution function; , They are the empirical cumulative distribution functions of precipitation and temperature in each month (January, February, March, … December) during the ground observation period; and They represent the large-scale grid precipitation and temperature series sensed by satellites, respectively; Represents the first d Cumulative distribution function for a surface observation period of 1 month.

[0046] Furthermore, in order to obtain large-scale rainfall data in the observation area, conventional techniques in the art may be used to extract TRMM precipitation satellite data developed by NASA and the Japan Aerospace Exploration Agency or GPM (Global Precipitation Program) precipitation satellite data.

[0047] 1.2 Calculate the correction factors of daily precipitation (ratio of ground observation sequence to satellite remote sensing grid sequence) and daily temperature (difference between ground observation sequence and satellite remote sensing grid sequence) under each probability of empirical distribution on a monthly basis:

[0048] , (3)

[0049] , (4)

[0050] in, , are the monthly precipitation and temperature correction factors in the historical period, , are the empirical cumulative distribution functions of precipitation and temperature in each month (January, February, March, ... December) during the ground observation period, ecdf represents the cumulative distribution function, d Indicates the month; It represents the first d The inverse of the monthly cumulative distribution function; Represents the first dThe inverse of the cumulative distribution function for a surface observation period of 1 month.

[0051] 1.3 Monthly interpolation and extrapolation of daily precipitation and daily temperature correction factors calculated based on empirical distribution:

[0052] , (5)

[0053] (6)

[0054] in, , are respectively the design periods to be derived after processing ( des ) precipitation and temperature correction factors, , are the precipitation and temperature correction factors for each month in the historical period.

[0055] 1.4 Apply the processed correction factors to simulate the long series of daily precipitation and daily temperature during the design period:

[0056] , (7)

[0057] , (8)

[0058] in, , They are the long series of daily precipitation and temperature after statistical downscaling, is the long series of precipitation measured by satellite telemetry during the design period, is the long series of temperature measured by satellite telemetry during the design period, , are respectively the design periods to be derived after processing ( des ) precipitation and temperature correction factors.

[0059] In step S102, a watershed hydrological model is established based on the short series runoff observation data and downscaled meteorological data of each sub-basin in the scarce data area, and the preliminary runoff of each sub-basin is simulated using the watershed hydrological model to obtain the simulated runoff of each sub-basin.

[0060] It can be understood that, for each sub-basin in the data-scarce area, the embodiment of the present invention uses the short series of runoff observation data of the sub-basin and the meteorological data after statistical downscaling in the same period to construct the Xin'anjiang model and the HBV model (a semi-distributed conceptual hydrological model for simulating the hydrological process of the basin), and calibrate the model parameters respectively.

[0061] Among them, the Xin'anjiang model is a rare hydrological model with global influence. The Xin'anjiang model is a lumped hydrological model (becomes a dispersed hydrological model when divided into sub-basins), which can be used in the wet season in humid and semi-humid areas. When the basin area is small, the Xin'anjiang model adopts a lumped model, and when the area is large, a dispersed model is adopted. It divides the entire basin into many unit basins, calculates the runoff for each unit basin, and obtains the outlet flow process of the unit basin. Then, the river flood calculation below the outlet is performed to obtain the flow process at the basin outlet. By adding up the outflow process of each unit basin, the total outflow process of the basin is obtained.

[0062] The model calculates the evapotranspiration of the basin according to the three-layer evapotranspiration model, calculates the total runoff generated by rainfall according to the concept of full storage and flow generation, and uses the basin water storage curve to consider the impact of uneven underlying surface on the change of flow generation area. In terms of runoff component division, for the three water sources, according to the "hillside hydrology" flow generation theory, a free water reservoir with a finite volume and measuring holes and bottom holes is used to divide the total runoff into saturated surface runoff, soil water runoff and groundwater runoff. In terms of confluence calculation, the confluence of surface runoff per unit area generally adopts the unit line method, while the confluence of soil water runoff and groundwater runoff adopts the linear reservoir method.

[0063] The HBV model is a conceptualized hydrological model that uses conceptual functions to simplify the actual hydrological process. By dividing the watershed into sub-basins, the HBV model can be regarded as a fully distributed or semi-partial hydrological model. The HBV model mainly includes four small models: snowmelt and snow accumulation model, soil moisture and effective precipitation model, evapotranspiration model, and runoff model.

[0064] Furthermore, for each sub-basin in the data-scarce area, the daily runoff process was simulated based on the calibrated Xin'anjiang model and HBV model, and the average series of the two models were used as the preliminary simulated daily runoff.

[0065] Therefore, the embodiment of the present invention can establish a basin hydrological model based on the short series runoff observation data and downscaled meteorological data of each sub-basin in the scarce data area, and use the model averaging method to achieve preliminary runoff simulation of each sub-basin.

[0066] In step S103, the simulated runoff of each sub-basin is corrected to obtain a corrected simulated runoff series, and a long short-term memory neural network model of each sub-basin is constructed based on the corrected simulated runoff series.

[0067] Among them, the embodiment of the present invention uses machine learning technology to correct the simulated runoff for each sub-basin in the data-scarce area, thereby obtaining a corrected simulated runoff series, thereby reducing the simulation error caused by human water use activities and constructing a long-short-term memory model.

[0068] Optionally, in one embodiment of the present invention, the simulated runoff of each sub-basin is corrected to obtain a corrected simulated runoff series, including: based on each sub-basin in the area with scarce data, statistical analysis is performed on the simulated daily runoff process and the measured daily runoff process to determine the lag time that affects the daily measured runoff; based on the lag time that affects the daily measured runoff, the simulated daily runoff process is corrected to obtain a corrected simulated runoff series.

[0069] In the actual implementation process, the embodiment of the present invention may include:

[0070] 3.1 For each sub-basin in the data-scarce area, the hysteresis time that affects the daily measured runoff is determined by statistically analyzing the simulated daily runoff process and the measured daily runoff process.

[0071] like Figure 3 As shown, a schematic diagram of the change in the correlation coefficient between the daily measured runoff and the simulated runoff under different lag times is given; the correlation coefficient between the simulated runoff and the measured runoff generally decreases gradually with the extension of the lag time; further, a suitable correlation threshold is selected to determine the simulated runoff time for establishing a machine learning model with the measured runoff; for example, 0.5 can be taken.

[0072] 3.2 For each sub-basin in the data-scarce area, a long short-term memory neural network (LSTM) model is used to correct the simulated daily runoff process to reduce the hydrological model error caused by human activities.

[0073] The embodiment of the present invention can construct a long short-term memory neural network (LSTM) model with a three-layer neural network architecture to generalize the regulating and storage effects of dams, reservoirs or water diversion projects on watersheds and improve the accuracy of hydrological simulation. This embodiment uses a neural network interval simulation mean method to independently run the neural network model multiple times and take the average value as the final simulation result to reduce uncertainty.

[0074] To solve the problem of nonlinear autoregressive exogenous input model (NARX) dynamic neural network in deep learning process (number of hidden layers) 2 layers), the LSTM long short-term memory neural network introduces storage units, namely input gates, forget gates, internal feedback connections, and output gates in the hidden layers of the NARX neural network to select to remember current information or forget past memory information (such as rainfall-runoff mapping) to enhance the long-term memory capacity of the NARX neural network. In short, the LSTM long short-term memory neural network replaces each hidden layer in the NARX dynamic neural network with a storage unit with memory function, referred to as the LSTM unit, and its input layer and output layer are the same as the NARX dynamic neural network.

[0075] Among them, in one embodiment of the present invention, based on the long short-term memory neural network (LSTM) model, the calculation formula of the corrected simulated runoff series is:

[0076] , (9)

[0077] in, The hydrological model is The simulated runoff at time, The hydrological model is The simulated runoff at time, The hydrological model is The simulated runoff at time, The hydrological model is The simulated runoff at time, represents the lag determined by the LSTM model, F LSTM Represents an LSTM model.

[0078] In step S104, the long series of meteorological observation data are respectively input into the watershed hydrological model and the long short-term memory neural network model to obtain the reconstructed long series runoff process of each sub-basin in the area with scarce data.

[0079] It is understandable that the embodiment of the present invention can use the minimum batch gradient descent method to train the LSTM model.

[0080] In the actual implementation process, the embodiment of the present invention can input the long series of meteorological observation data obtained in step S101 into the hydrological model and long short-term memory neural network model established in steps S102 and S103 to reconstruct the long series of runoff processes in each sub-basin in the area with scarce data.

[0081] Optionally, in one embodiment of the present invention, long series meteorological observation data are respectively input into the watershed hydrological model and the long short-term memory neural network model to obtain the reconstructed long series runoff process of each sub-basin in the data scarce area, including: based on the statistical downscaling model, using large-scale raster meteorological data sensed by satellite telemetry to obtain the long series meteorological series at the watershed scale; inputting the long series meteorological series into the watershed hydrological model to obtain the long series runoff process of each sub-basin; using the long short-term memory neural network model to correct the simulated runoff to reconstruct the long series runoff process of each sub-basin, and obtaining the reconstructed long series runoff process of each sub-basin.

[0082] Specifically, the embodiment of the present invention can use the large-scale raster meteorological data of satellite telemetry based on the statistical downscaling model established in step S101 to obtain a long series of meteorological series at the basin scale, and then input the long series data into the Xin'anjiang model of each sub-basin calibrated above to obtain a long series of runoff process, and then use the LSTM model calibrated above to correct the simulated runoff, so as to reconstruct the simulated long series of runoff process for each sub-basin in the scarce data area.

[0083] In step S105, flood events are extracted from the reconstructed long series of runoff processes in each sub-basin, and a random forest model including spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics is established based on the flood events to predict large-scale floods in areas with scarce data.

[0084] In the actual implementation process, the embodiment of the present invention can extract flood events from the reconstructed long series of runoff processes in each sub-basin, and establish a random forest model that considers spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics, drive the random forest model, and carry out large-scale flood forecasting in areas with scarce data, thereby providing an important and highly operational reference basis for flood forecasting in areas with scarce data, and providing engineering reference value for responding to climate disasters, disaster prevention and mitigation, and integrated water resources management.

[0085] Optionally, in one embodiment of the present invention, flood events are extracted from the reconstructed long series of runoff processes of each sub-basin, and a random forest model is established based on the flood events, taking into account spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics, including: based on the reconstructed long series of runoff processes of each sub-basin, taking the preset quantile of daily runoff as a threshold, and treating the daily runoff exceeding the threshold as a flood event; based on the flood events, obtaining the basin-average meteorological variables of each sub-basin in the area with scarce data, and establishing a random forest model based on the basin-average meteorological variables.

[0086] It can be understood that the preset quantile of the daily runoff in the embodiment of the present invention may be the 90% quantile of the daily runoff.

[0087] Specifically, embodiments of the present invention may include:

[0088] 5.1 Based on the reconstructed long series of runoff processes in each sub-basin, the 90% quantile of daily runoff is used as the threshold, and the daily runoff exceeding the threshold is identified and regarded as a flood event; for daily runoff that continuously exceeds the threshold, the highest runoff value is taken as a flood event.

[0089] 5.2 Based on the fifth-generation atmospheric reanalysis dataset (ERA5-Land) of the European Centre for Medium-Range Weather Forecasts, hourly meteorological and hydrological observation data were collected, including 2m (meter) air temperature, air pressure, dew point temperature, snowfall, precipitation, wind speed, shortwave radiation and longwave radiation. The daily average 2m (meter) air temperature, daily average dew point temperature, daily average snowfall, daily precipitation, daily average air pressure, daily wind speed, daily shortwave radiation and daily longwave radiation were obtained through time scale conversion.

[0090] 5.3 Calculate specific humidity and relative humidity based on the Clausius-Clapeyron thermodynamic equation.

[0091] The Clausius-Clapeyron thermodynamic equation is defined as follows:

[0092] , (10)

[0093] in, T 0 is the first integration constant, which is 273.16 K; is the second integral constant, which is 611 Pa; is the latent heat of vaporization constant, taken as 2.5×10 6 J kg -1 ; is the water vapor gas constant, which is 461 J kg -1 K -1 ; T is the air temperature, which is the input variable of the Clausius-Clapeyron thermodynamic equation.

[0094] The 2-meter air temperature and dew point temperature of the meteorological and hydrological observation samples at each grid point in the study basin are substituted into the Clausius-Clapeyron thermodynamic equation as input variables to calculate the relative humidity (RH) of each grid point: , where Tdew is the dew point temperature, T2m is the air temperature at 2m (meters), is the saturated water vapor pressure.

[0095] Specific humidity q is the ratio of water vapor mass to total air mass mass, using ERA5-Land ground pressure p And dew point temperature, the formula is as follows:

[0096] , (11)

[0097] in, is the saturated water vapor pressure, T dew is the dew point temperature.

[0098] 5.4 The Thiessen polygon method is used to derive the basin-average meteorological variables of each sub-basin in the area with scarce data, and a random forest model of flood events considering spatial attributes is established.

[0099] The flood events of each sub-basin are combined with the meteorological variables of the basin. The meteorological variables include the average daily temperature, specific humidity, relative humidity, snowfall, precipitation, wind speed, shortwave radiation and longwave radiation on the day of the flood event and 1-7 days before. That is, for each flood event in each sub-basin, 8×8=64 meteorological data are involved.

[0100] Meteorological and hydrological data in scarce data areas are limited. In order to achieve large-scale flood forecasting, the embodiment of the present invention uses flood and meteorological information of all sub-basins to calibrate the random forest model, thereby achieving the characterization of the spatial attributes of the flood mechanism. For areas with scarce data, the flood magnitude is used as the dependent variable, the 64 meteorological and hydrological variables corresponding to the flood are used as driving factors, and the flood-meteorological and hydrological data of all sub-basins are used as input to calibrate the random forest model.

[0101] The random forest model can rank the importance of each input variable to the simulation variable. Therefore, based on the above random forest model, the important factors affecting floods are selected. In this embodiment, 25% is set as the threshold, that is, after ranking the 16 variables, the top 16 variables are selected as key factors.

[0102] Furthermore, the 16 selected key factors were used as input to construct a flood forecasting model for areas with scarce data.

[0103] Further, Figure 4 A schematic flow chart of a flood forecasting method based on artificial intelligence and satellite remote sensing provided in an embodiment of the present invention applied to a model application stage.

[0104] like Figure 4 As shown, the flood forecasting method based on artificial intelligence and satellite remote sensing includes the following steps:

[0105] In step S401, the weather forecast information and reanalysis weather data of each sub-basin in the data-scarce region are obtained.

[0106] Among them, the embodiment of the present invention can obtain weather forecast information for areas with scarce data from the European Center for Medium-Term Weather Forecasts or the China Meteorological Administration and other units, specifically including the average daily temperature, specific humidity, relative humidity, snowfall, precipitation, wind speed, shortwave radiation and longwave radiation for the next 1-14 days, and use the Thiessen polygon method to calculate the regional average weather forecast information for areas with scarce data; the reanalysis meteorological data of the past 7 days obtained from ERA5-Land (the fifth generation of atmospheric reanalysis land surface data set) is used, and the Thiessen polygon method is used to obtain the regional average reanalysis meteorological data for areas with scarce data.

[0107] In step S402, the meteorological forecast information and the reanalyzed meteorological data are input into a preset random forest model to drive the random forest model to generate large-scale flood forecast results in areas with scarce data, wherein the preset random forest model is constructed by spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics.

[0108] Among them, the embodiment of the present invention can use the regional average weather forecast information for the next day in the area with scarce data and the regional average reanalysis weather data of the past 1-7 days in the area with scarce data as input to drive the flood forecast model, thereby obtaining the flood forecast value for the next day; using a similar method as mentioned above, the weather forecast information with a longer future forecast period is successively used as input to obtain the flood forecast value for the next 2-7 days.

[0109] According to the flood forecasting method based on artificial intelligence and satellite remote sensing proposed in the embodiment of the present invention, by coupling the physical mechanism of flood formation, artificial intelligence and satellite remote sensing, it provides an important and highly operational reference basis for flood forecasting in areas with scarce data, and provides engineering reference value for responding to climate disasters, disaster prevention and mitigation, and integrated water resources management. Thus, the problem that the relevant technology fails to make full use of satellite telemetry meteorological information, fails to consider the errors caused by human activity interference in runoff simulation, and cannot solve the problem of long-series flood forecasting in areas with scarce data is solved.

[0110] Next, a flood forecasting device based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention will be described with reference to the accompanying drawings.

[0111] Figure 5 It is a structural schematic diagram of a flood forecasting device based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention.

[0112] like Figure 5 As shown, the flood forecasting device 10 based on artificial intelligence and satellite remote sensing is applied to the model building stage and includes: a building module 100, a simulation module 200, a correction module 300, an input module 400 and a forecasting module 500.

[0113] Specifically, a module 100 is established to establish a statistical downscaling model based on limited observation data from ground meteorological stations in areas with scarce data and large-scale grid data from satellite remote sensing, and to obtain long-series meteorological observation data based on the statistical downscaling model.

[0114] The simulation module 200 is used to establish a basin hydrological model based on the short series runoff observation data and downscaled meteorological data of each sub-basin in the scarce data area, and use the basin hydrological model to simulate the preliminary runoff of each sub-basin to obtain the simulated runoff of each sub-basin.

[0115] The correction module 300 is used to correct the simulated runoff of each sub-basin to obtain a corrected simulated runoff series, and to construct a long short-term memory neural network model of each sub-basin based on the corrected simulated runoff series.

[0116] The input module 400 is used to input the long series of meteorological observation data into the watershed hydrological model and the long short-term memory neural network model respectively, so as to obtain the reconstructed long series runoff process of each sub-basin in the area with scarce data.

[0117] The forecast module 500 is used to extract flood events from the reconstructed long series of runoff processes in each sub-basin, and to establish a random forest model including spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics based on flood events to forecast large-scale floods in areas with scarce data.

[0118] Optionally, in one embodiment of the present invention, the correction module 300 includes: a determination unit and a correction unit.

[0119] The determination unit is used to perform statistical analysis on the simulated daily runoff process and the measured daily runoff process for each sub-basin in the area with scarce data, so as to determine the lag time that affects the measured daily runoff.

[0120] The correction unit is used to correct the simulated daily runoff process based on the lag time that affects the daily measured runoff to obtain a corrected simulated runoff series.

[0121] Optionally, in one embodiment of the present invention, the calculation formula of the corrected simulated runoff series is:

[0122] ,

[0123] in, The hydrological model is The simulated runoff at time, The hydrological model is The simulated runoff at time, The hydrological model is The simulated runoff at time, The hydrological model is The simulated runoff at time, represents the lag determined by the LSTM model, F LSTM Represents an LSTM model.

[0124] Optionally, in one embodiment of the present invention, the input module 400 includes: an acquisition unit, an input unit and a reconstruction unit.

[0125] Among them, the acquisition unit is used to obtain a long series of meteorological series at the basin scale based on a statistical downscaling model using large-scale raster meteorological data sensed by satellite telemetry.

[0126] The input unit is used to input the long series of meteorological data into the basin hydrological model to obtain the long series of runoff processes in each sub-basin.

[0127] The reconstruction unit is used to correct the simulated runoff by using the long short-term memory neural network model to reconstruct the long series of runoff processes of each sub-basin and obtain the reconstructed long series of runoff processes of each sub-basin.

[0128] Optionally, in one embodiment of the present invention, the forecast module 500 includes: a generating unit and a establishing unit.

[0129] Among them, the generation unit is used to reconstruct the long series runoff process based on each sub-basin, take the preset quantile of daily runoff as the threshold, and regard the daily runoff exceeding the threshold as a flood event.

[0130] A unit was established to obtain the basin-average meteorological variables of each sub-basin in the data-scarce area based on flood events, and a random forest model was established based on the basin-average meteorological variables.

[0131] Figure 6 It is a structural schematic diagram of a flood forecasting device based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention.

[0132] like Figure 6 As shown, the flood forecasting device 20 based on artificial intelligence and satellite remote sensing is applied in the model application stage and includes: an acquisition module 600 and a driving module 700.

[0133] Specifically, the acquisition module 600 is used to obtain the weather forecast information and reanalysis weather data of each sub-basin in the area with scarce data.

[0134] The driving module 700 is used to input the meteorological forecast information and the reanalyzed meteorological data into the preset random forest model to drive the random forest model to generate large-scale flood forecast results in areas with scarce data, wherein the preset random forest model is constructed by spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics.

[0135] It should be noted that the above explanation of an embodiment of a flood forecasting method based on artificial intelligence and satellite remote sensing is also applicable to a flood forecasting device based on artificial intelligence and satellite remote sensing in this embodiment, which will not be repeated here.

[0136] According to the flood forecasting device based on artificial intelligence and satellite remote sensing proposed in the embodiment of the present invention, by coupling the physical mechanism of flood formation, artificial intelligence and satellite remote sensing and other methods, it provides an important and highly operational reference basis for flood forecasting in areas with scarce data, and provides engineering reference value for responding to climate disasters, disaster prevention and mitigation, and integrated water resources management. Thus, the problem that the relevant technology fails to make full use of satellite telemetry meteorological information, fails to consider the errors caused by human activity interference in runoff simulation, and cannot solve the problem of long-series flood forecasting in areas with scarce data is solved.

[0137] Figure 7 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device may include:

[0138] A memory 701 , a processor 702 , and a computer program stored in the memory 701 and executable on the processor 702 .

[0139] When the processor 702 executes the program, a flood forecasting method based on artificial intelligence and satellite remote sensing provided in the above embodiment is implemented.

[0140] Furthermore, the electronic device further comprises:

[0141] The communication interface 703 is used for communication between the memory 701 and the processor 702 .

[0142] The memory 701 is used to store computer programs that can be executed on the processor 702 .

[0143] The memory 701 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0144] If the memory 701, the processor 702 and the communication interface 703 are implemented independently, the communication interface 703, the memory 701 and the processor 702 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0145] Optionally, in a specific implementation, if the memory 701, the processor 702 and the communication interface 703 are integrated on a chip, the memory 701, the processor 702 and the communication interface 703 can communicate with each other through an internal interface.

[0146] The processor 702 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0147] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned flood forecasting method based on artificial intelligence and satellite remote sensing is implemented.

[0148] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0149] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0150] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0151] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection with one or N wirings (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0152] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0153] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0154] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0155] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A flood forecasting method based on artificial intelligence and satellite remote sensing, characterized in that: Applied to the model building phase, it includes the following steps: Based on the limited observation data of ground meteorological stations in areas with scarce data and the large-scale grid data of satellite remote sensing, a statistical downscaling model is established, and a long series of meteorological observation data is obtained according to the statistical downscaling model; Based on the short series runoff observation data and downscaled meteorological data of each sub-basin in the scarce data area, a basin hydrological model is established, and the preliminary runoff of each sub-basin is simulated by using the basin hydrological model to obtain the simulated runoff of each sub-basin; Correcting the simulated runoff of each sub-basin to obtain a corrected simulated runoff series, and constructing a long short-term memory neural network model of each sub-basin based on the corrected simulated runoff series; Inputting the long series of meteorological observation data into the watershed hydrological model and the long short-term memory neural network model respectively, so as to obtain the reconstructed long series of runoff process of each sub-basin in the scarce data area; Extract flood events from the reconstructed long series of runoff processes of each sub-basin, and establish a random forest model including spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics based on the flood events to predict large-scale floods in the scarce data area, wherein the extract flood events from the reconstructed long series of runoff processes of each sub-basin, and establish a random forest model considering spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics based on the flood events, including: based on the reconstructed long series of runoff processes of each sub-basin, taking the preset quantile of daily runoff as a threshold, and taking the daily runoff exceeding the threshold as the flood event; based on the flood event, obtaining the basin average meteorological variables of each sub-basin in the scarce data area, and establishing the random forest model according to the basin average meteorological variables; Among them, the flood events of each sub-basin are combined with the meteorological variables of the basin. The meteorological variables include the average daily temperature, specific humidity, relative humidity, snowfall, precipitation, wind speed, shortwave radiation and longwave radiation on the day of the flood event and 1-7 days before. For each flood event in each sub-basin, 8×8=64 meteorological data are involved.

2. A flood forecasting method based on artificial intelligence and satellite remote sensing according to claim 1, characterized in that: The simulated runoff of each sub-basin is corrected to obtain a corrected simulated runoff series, including: Based on each sub-basin in the scarce data area, the simulated daily runoff process and the measured daily runoff process are statistically analyzed to determine the lag time that affects the measured daily runoff; Based on the time lag that affects the daily measured runoff, the simulated daily runoff process is corrected to obtain the corrected simulated runoff series.

3. The flood forecasting method based on artificial intelligence and satellite remote sensing according to claim 1 is characterized in that: The calculation formula of the corrected simulated runoff series is: ， in, The hydrological model is The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, Indicates that the hydrological model The simulated runoff at time, represents the lag time determined by the LSTM model, Represents the LSTM model.

4. The flood forecasting method based on artificial intelligence and satellite remote sensing according to claim 1 is characterized in that: The long series of meteorological observation data are respectively input into the watershed hydrological model and the long short-term memory neural network model to obtain the reconstructed long series of runoff processes of each sub-basin in the scarce data area, including: Based on the statistical downscaling model, the large-scale grid meteorological data of satellite telemetry are used to obtain a long series of meteorological series at the basin scale; Inputting the long series of meteorological series into the watershed hydrological model to obtain the long series of runoff processes of each sub-watershed; The simulated runoff is corrected using the long short-term memory neural network model to reconstruct the long series of runoff processes of each sub-basin to obtain the reconstructed long series of runoff processes of each sub-basin.

5. A flood forecasting method based on artificial intelligence and satellite remote sensing, characterized in that: The flood forecasting method based on artificial intelligence and satellite remote sensing as described in any one of claims 1 to 4 is applied to the model application stage, comprising the following steps: Obtain weather forecast information and reanalysis weather data for each sub-basin in data-scarce areas; The meteorological forecast information and the reanalyzed meteorological data are input into a preset random forest model to drive the random forest model to generate large-scale flood forecast results for the data-scarce area, wherein the preset random forest model is constructed by spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics.

6. A flood forecasting device based on artificial intelligence and satellite remote sensing, characterized in that: The flood forecasting method based on artificial intelligence and satellite remote sensing as described in any one of claims 1 to 4 is applied to the model building stage, comprising: Establishing a module for establishing a statistical downscaling model based on limited observation data from ground meteorological stations in areas with scarce data and large-scale grid data from satellite remote sensing, and obtaining long-series meteorological observation data according to the statistical downscaling model; A simulation module, for establishing a watershed hydrological model based on the short series runoff observation data and downscaled meteorological data of each sub-basin in the scarce data area, and simulating the preliminary runoff of each sub-basin using the watershed hydrological model to obtain the simulated runoff of each sub-basin; A correction module, used for correcting the simulated runoff of each sub-basin to obtain a corrected simulated runoff series, and constructing a long short-term memory neural network model of each sub-basin according to the corrected simulated runoff series; An input module, used for inputting the long series of meteorological observation data into the watershed hydrological model and the long short-term memory neural network model respectively, so as to obtain a reconstructed long series of runoff processes for each sub-basin in the scarce data area; A forecasting module, used for extracting flood events from the reconstructed long series of runoff processes of each sub-basin, and establishing a random forest model including spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics based on the flood events, so as to forecast large-scale floods in the scarce data area, wherein the extracting flood events from the reconstructed long series of runoff processes of each sub-basin, and establishing a random forest model considering spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics based on the flood events, comprises: based on the reconstructed long series of runoff processes of each sub-basin, taking a preset quantile of daily runoff as a threshold, and taking the daily runoff exceeding the threshold as the flood event; based on the flood event, obtaining the basin average meteorological variables of each sub-basin in the scarce data area, and establishing the random forest model according to the basin average meteorological variables; Among them, the flood events of each sub-basin are combined with the meteorological variables of the basin. The meteorological variables include the average daily temperature, specific humidity, relative humidity, snowfall, precipitation, wind speed, shortwave radiation and longwave radiation on the day of the flood event and 1-7 days before. For each flood event in each sub-basin, 8×8=64 meteorological data are involved.

7. A flood forecasting device based on artificial intelligence and satellite remote sensing, characterized in that: The flood forecasting method based on artificial intelligence and satellite remote sensing as described in claim 5 above is applied in the model application stage, comprising: Acquisition module, used to obtain weather forecast information and reanalysis weather data for each sub-basin in areas with scarce data; A driving module is used to input the meteorological forecast information and the reanalyzed meteorological data into a preset random forest model to drive the random forest model to generate large-scale flood forecast results in the scarce data area, wherein the preset random forest model is constructed by spatial attributes, meteorological forecast information and meteorological-hydrological time lag characteristics.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement a flood forecasting method based on artificial intelligence and satellite remote sensing as described in any one of claims 1 to 4 or a flood forecasting method based on artificial intelligence and satellite remote sensing as described in claim 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement a flood forecasting method based on artificial intelligence and satellite remote sensing as described in any one of claims 1 to 4 or a flood forecasting method based on artificial intelligence and satellite remote sensing as described in claim 5.

Images