A method and device for predicting sudden drought and flood transition based on artificial intelligence and satellite remote sensing

Abstract

The present invention relates to the field of climate prediction technology, and in particular to a method and device for predicting drought-flood sudden transition based on artificial intelligence and satellite remote sensing, wherein the method includes: determining saturated water vapor pressure deficit and specific humidity using the Clausius-Clapeyron thermodynamic equation; calibrating a long-term and short-term memory model suitable for terrestrial water storage simulation; extracting drought-flood sudden transition events from a reconstructed long series of terrestrial water storage inversion data sets; using a box element scaling function to divide the temperature-drought-flood sudden transition intensity combination of each grid point in the predicted basin into multiple boxes with the same sample capacity; using a scaling coefficient to predict the drought-flood sudden transition intensity of each grid point in the future period, so as to obtain the prediction result of the drought-flood sudden transition intensity of each grid point in the target future period. Thus, the problem that the relevant technology is difficult to achieve retrospective risk assessment of drought and flood disasters over a long time span, the evolution law of drought-flood sudden transition events is rarely considered, and the artificial intelligence technology is not fully utilized to carry out drought-flood sudden transition prediction is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of climate prediction, and in particular to a method and device for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing. Background Art

[0002] Global climate change has changed the energy balance and water cycle of the land-atmosphere system. Extreme climate disasters such as heat waves, droughts, and floods have occurred frequently, posing huge challenges to the sustainable development of the social and economic system and the ecological environment. A deep understanding of the socio-economic impact of drought and flood disasters under climate change scenarios is of great significance for future drought and flood risk prediction, disaster prevention and mitigation, and adaptation management.

[0003] Continuous high temperature and drought, soil dehydration and looseness or caking and hardening, large atmospheric energy accumulation, slight disturbances are prone to severe convective weather and concentrated short-term heavy rainfall, and are prone to sudden flash floods and induce extreme disasters with multiple disaster-causing factors superimposed. In recent years, when there is no precipitation or less precipitation in the early stage and drought has occurred, it is easy to be followed by high-intensity precipitation in a short period of time and form floods. This phenomenon is called drought-flood transition. Due to the rapid transition of natural states, short time and high rainfall intensity, it is easy to cause flash floods, sudden rise of river water, river water intrusion and other phenomena. Drought-flood transition poses a great challenge to flood and drought disaster prevention, and early prevention is needed. However, it is still unclear how to monitor and predict future drought-flood transition disasters.

[0004] In March 2002, the Gravity Recovery and Climate Experiment (GRACE) satellite was successfully launched, providing a continuous, high-precision direct observation method for obtaining global large-scale migration of Earth's surface materials. The gravity field model based on the GRACE satellite signal solution can extract the change information of the Earth's lunar gravity field at a spatial scale of 300 km×300 km, deduct the influence of factors such as crustal material movement, atmospheric movement, ocean currents and tides, and can effectively reflect the gravity changes caused by ice and snow, surface water, soil water, groundwater and human factors, and comprehensively monitor the terrestrial water storage anomaly (TWSA) signal. In May 2018, after the GRACE gravity satellite stopped working for a year, the GRACE-FO (GRACE Follow-On) gravity satellite was successfully launched, continuing the scientific observation mission of the GRACE satellite. The GRACE / GRACE-FO gravity satellite effectively solved the problems of shallow ground observation range, uneven spatial distribution, and difficulty in data acquisition, and showed great potential in tracking regional and global water and drought events.

[0005] However, the observation time of satellites used in related technologies for drought and flood disaster monitoring and assessment is relatively short, making it difficult to achieve retrospective risk assessment of drought and flood disasters over a longer time span. Moreover, they usually only focus on changes in a single disaster type such as drought or waterlogging, and rarely consider the evolution laws of sudden transitions from drought to flood. They fail to fully utilize artificial intelligence technology to carry out sudden transition forecasts from drought to flood, which urgently needs to be improved. Summary of the invention

[0006] The present invention provides a method and device for predicting the sudden transition from drought to flood based on artificial intelligence and satellite remote sensing, so as to solve the problems in the related technology that the observation time of satellites used for drought and flood disaster monitoring and assessment is short, and it is difficult to realize the retrospective risk assessment of drought and flood disasters with a longer time span, and the method usually only focuses on the changes of a single disaster type such as drought or waterlogging, and rarely considers the evolution law of sudden transition from drought to flood events, and fails to make full use of artificial intelligence technology to carry out the prediction of sudden transition from drought to flood.

[0007] The first aspect of the present invention provides a method for predicting sudden transitions from drought to flood based on artificial intelligence and satellite remote sensing, comprising the following steps: collecting at least one of the meteorological, hydrological, vegetation, land use data, inversion data of the gravity satellite GRACE, the El Nino ENSO index and climate model output data in the drought-flood sudden transition prediction basin; based on the at least one data, determining the saturated water vapor pressure deficit and specific humidity using the Clausius-Clapeyron thermodynamic equation, and based on the saturated water vapor pressure deficit, the specific humidity and a preset random forest model, obtaining key factors that meet preset preferred conditions affecting terrestrial water storage at each grid point in the drought-flood sudden transition prediction basin to obtain preferred key factors; based on the preferred key factors and the inversion data of the gravity satellite GRACE, calibrating the long short-term memory model suitable for the terrestrial water storage simulation, and using the long short-term memory model to calculate the terrestrial water storage. A series of key factor data sets drive the long-short term memory model to obtain a reconstructed long series of terrestrial water storage inversion data sets; drought-flood transition events are extracted from the reconstructed long series of terrestrial water storage inversion data sets, and the average temperature in the drought-flood transition events forms a temperature-drought-flood transition intensity pairing series; based on the temperature-drought-flood transition intensity pairing series, the temperature-drought-flood transition intensity combination of each grid point in the predicted basin is divided into multiple bins with the same sample capacity using a box scaling function to obtain a division result, and the scaling coefficient of the drought-flood transition intensity in the predicted basin and the near-ground temperature is obtained according to the division result; based on the temperature data of each grid point in the predicted basin output by the global climate model, the target future drought-flood transition intensity of each grid point is predicted using the scaling coefficient to obtain the prediction result of the target future drought-flood transition intensity of each grid point.

[0008] Optionally, in one embodiment of the present invention, the Clausius-Clapeyron thermodynamic equation describes the nonlinear relationship between saturated water vapor pressure and air temperature as follows:

[0009] ,

[0010] in, and are the first and second integral constants, is the latent heat of vaporization constant, is the water vapor gas constant;

[0011] The calculation formula of the specific humidity is:

[0012] ,

[0013] in, is the dew point temperature, is the ERA5 ground pressure, is the saturated water vapor pressure.

[0014] Optionally, in one embodiment of the present invention, the key factors that meet preset preferred conditions affecting terrestrial water storage are obtained in each grid point in the drought-flood rapid transition prediction basin to obtain the preferred key factors, including: inputting at least one driving factor among air temperature, specific humidity, saturated water vapor pressure deficit, precipitation, soil moisture, water vapor flux divergence, cloud coverage, runoff depth, shortwave radiation, leaf area index, impervious area and the El Nino ENSO index into the preset random forest model to obtain variables of the preset random forest model; sorting the importance of the variables of the preset random forest model to obtain sorted variables, and judging whether the sorted variables exceed a preset threshold; if the sorted variables exceed the preset threshold, generating the key factors from the variables exceeding the preset threshold.

[0015] Optionally, in one embodiment of the present invention, the long short-term memory model is The expression for simulating the land water storage at any time is:

[0016]

[0017] in, express The input variables at time, express Input variables at time 1, express Input variables at time 2, express Input variables at time 3, Represents the long short-term memory model.

[0018] Optionally, in one embodiment of the present invention, extracting drought-flood sudden transition events from the reconstructed long series terrestrial water storage inversion dataset includes: obtaining a long series of drought intensities based on the reconstructed long series terrestrial water storage inversion dataset; calculating thresholds corresponding to the first quantile and the second quantile of the long series drought intensities to obtain drought thresholds and flood thresholds, and determining drought months and flood months based on the drought thresholds and the flood thresholds, and generating the drought-flood sudden transition event when the drought month and the flood month exist in two adjacent months at the same time.

[0019] Optionally, in one embodiment of the present invention, the expression of the scaling coefficient of the predicted intensity of drought-flood transition in the watershed and the near-ground temperature is:

[0020] ,

[0021] in, P 1 and P 2 are the average intensity of the drought-flood transition event in adjacent boxes, is the scaling factor, Indicates the amount of change, T For the temperature.

[0022] Optionally, in one embodiment of the present invention, the calculation formula of the target future drought-flood transition intensity of each grid point is:

[0023] ,

[0024] in, represents the average intensity of drought-flood transition in the target future, represents the average value of the intensity of drought-flood transition, is the scaling factor, is the temperature change in the future period relative to the historical period;

[0025] The calculation formula for the prediction result of the target future drought-flood sudden change intensity of each grid point is:

[0026] ,

[0027] in, a and b represents the parameters of the emergence constraint model, Represents the annual average temperature trend obtained from the ERA5 dataset at each grid point.

[0028] The second aspect of the present invention provides a drought-flood sudden transition prediction device based on artificial intelligence and satellite remote sensing, including: a collection module, used to collect at least one of the meteorological, hydrological, vegetation, land use data, gravity satellite GRACE inversion data, El Nino ENSO index and climate model output data in the drought-flood sudden transition prediction basin; a determination module, used to determine the saturated water vapor pressure deficit and specific humidity based on the at least one data using the Clausius-Clapeyron thermodynamic equation, and based on the saturated water vapor pressure deficit, the specific humidity and a preset random forest model, obtain the key factors that meet the preset preferred conditions affecting terrestrial water storage in each grid point in the drought-flood sudden transition prediction basin to obtain the preferred key factors; a calibration module, used to calibrate the long short-term memory model suitable for the terrestrial water storage simulation based on the preferred key factors and the gravity satellite GRACE inversion data, and use the long series The key factor data set drives the long-short term memory model to obtain a reconstructed long series of terrestrial water storage inversion data set; an extraction module is used to extract drought-flood sudden transition events from the reconstructed long series of terrestrial water storage inversion data set, and form a temperature-drought-flood sudden transition intensity pairing series from the average temperature in the drought-flood sudden transition events; a division module is used to divide the temperature-drought-flood sudden transition intensity combination of each grid point in the predicted basin into multiple bins with the same sample capacity based on the temperature-drought-flood sudden transition intensity pairing series using a bin scaling function to obtain a division result, and obtain a scaling coefficient of the drought-flood sudden transition intensity in the predicted basin and the near-ground temperature based on the division result; a prediction module is used to predict the target future drought-flood sudden transition intensity of each grid point based on the temperature data of each grid point in the predicted basin output by the global climate model using the scaling coefficient to obtain a prediction result of the target future drought-flood sudden transition intensity of each grid point.

[0029] Optionally, in one embodiment of the present invention, the Clausius-Clapeyron thermodynamic equation describes the nonlinear relationship between saturated water vapor pressure and air temperature as follows:

[0030] ,

[0031] in, and are the first and second integral constants, is the latent heat of vaporization constant, is the water vapor gas constant;

[0032] The calculation formula of the specific humidity is:

[0033] ,

[0034] in, is the dew point temperature, is the ERA5 ground pressure, is the saturated water vapor pressure.

[0035] Optionally, in one embodiment of the present invention, the determination module includes: an input unit, used to input at least one driving factor of the air temperature, the specific humidity, the saturated vapor pressure deficit, precipitation, soil moisture, water vapor flux divergence, cloud coverage, runoff depth, shortwave radiation, leaf area index, impervious area and the El Niño ENSO index into the preset random forest model to obtain variables of the preset random forest model; a sorting unit, used to sort the importance of the variables of the preset random forest model to obtain sorted variables, and determine whether the sorted variables exceed a preset threshold; a generation unit, used to generate the key factor from the variable exceeding the preset threshold when the sorted variable exceeds the preset threshold.

[0036] Optionally, in one embodiment of the present invention, the long short-term memory model is The expression for simulating the land water storage at any time is:

[0037]

[0038] in, express The input variables at time, express Input variables at time 1, express Input variables at time 2, express Input variables at time 3, Represents the long short-term memory model.

[0039] Optionally, in one embodiment of the present invention, the extraction module includes: an acquisition unit, used to obtain a long series of drought intensities based on the reconstructed long series of terrestrial water storage inversion data set; a calculation unit, used to calculate thresholds corresponding to the first quantile and the second quantile of the long series of drought intensities to obtain drought thresholds and flood thresholds, and determine drought months and flood months based on the drought thresholds and the flood thresholds, and generate the drought-flood sudden transition event when the drought month and the flood month exist in two adjacent months at the same time.

[0040] Optionally, in one embodiment of the present invention, the expression of the scaling coefficient of the predicted intensity of drought-flood transition in the watershed and the near-ground temperature is:

[0041] ,

[0042] in, P 1 and P2 are the average intensity of the drought-flood transition event in adjacent boxes, is the scaling factor, Indicates the amount of change, T For the temperature.

[0043] Optionally, in one embodiment of the present invention, the calculation formula of the target future drought-flood transition intensity of each grid point is:

[0044] ,

[0045] in, represents the average intensity of drought-flood transition in the target future, represents the average value of the intensity of drought-flood transition, is the scaling factor, is the temperature change in the future period relative to the historical period;

[0046] The calculation formula for the prediction result of the target future drought-flood sudden change intensity of each grid point is:

[0047] ,

[0048] in, a and b represents the parameters of the emergence constraint model, Represents the average annual temperature trend obtained from the ERA5 dataset at each grid point.

[0049] A third 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 method for predicting sudden transitions from drought to flood based on artificial intelligence and satellite remote sensing as described in the above embodiment.

[0050] A fourth 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 method for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing.

[0051] The embodiment of the present invention can combine artificial intelligence methods and GRACE gravity satellite data to construct an artificial intelligence model for simulating terrestrial water storage, and use long-series meteorological and hydrological data to drive the artificial intelligence model to achieve long-series terrestrial water storage data reconstruction; further derive the scale coefficient for predicting the intensity of drought-flood sudden transition in the basin, and use the latest global climate model set to predict the intensity of future drought-flood sudden transition, and finally use the emergence constraint technology to reduce uncertainty, which has important scientific significance for future drought-flood sudden transition risk prediction and disaster prevention and mitigation, and provides a certain theoretical reference and technical basis for further evaluating the environmental and disaster effects caused by the evolution of the earth system. Thus, the problem that the observation time of satellites in related technologies used for drought and flood disaster monitoring and assessment is short, and it is difficult to achieve a long time span of drought and flood disaster retrospective risk assessment, and it usually only focuses on the changes of a single disaster type such as drought or waterlogging, and rarely considers the evolution law of drought-flood sudden transition events, and fails to make full use of artificial intelligence technology to carry out drought-flood sudden transition prediction.

[0052] 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

[0053] 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:

[0054] Figure 1 A flowchart of a method for predicting sudden drought-flood transition based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention;

[0055] Figure 2 A schematic diagram of a sliding window method for predicting a sudden transition from drought to flood based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention;

[0056] Figure 3 A schematic diagram of the structure of a drought-flood rapid transition prediction device based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention;

[0057] Figure 4 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0058] 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.

[0059] The following describes a method and device for predicting drought-flood abrupt transition based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention with reference to the accompanying drawings. Aiming at the problem that the observation time of satellites in the related technologies mentioned in the above background technology is short when used for monitoring and evaluating drought and flood disasters, it is difficult to realize the retrospective risk assessment of drought and flood disasters with a long time span, and they usually only focus on the changes of single disasters such as drought or waterlogging, and rarely consider the evolution law of drought-flood abrupt transition events, and fail to make full use of artificial intelligence technology to predict drought-flood abrupt transition. The present invention provides a method for predicting drought-flood abrupt transition based on artificial intelligence and satellite remote sensing, in which an artificial intelligence model for simulating terrestrial water storage can be constructed by combining artificial intelligence methods and GRACE gravity satellite data, and the artificial intelligence model is driven by long series meteorological and hydrological data to realize the reconstruction of long series terrestrial water storage data; further derive the scale coefficient for predicting the intensity of drought-flood abrupt transition in the basin, and use the latest global climate model ensemble to predict the intensity of drought-flood abrupt transition in the future, and finally use the emergence constraint technology to reduce uncertainty, which has important scientific significance for the prediction of future drought-flood abrupt transition risk and disaster prevention and mitigation, and provides a certain theoretical reference and technical basis for further evaluating the environmental and disaster effects caused by the evolution of the earth system. This solves the problem that satellites in related technologies have short observation times when used for drought and flood disaster monitoring and assessment, making it difficult to conduct retrospective risk assessment of drought and flood disasters over a longer time span, and that they usually only focus on changes in a single disaster type such as drought or waterlogging, and rarely consider the evolution of drought-flood sudden transition events, and fail to fully utilize artificial intelligence technology to carry out drought-flood sudden transition forecasts.

[0060] Specifically, Figure 1 A schematic flow chart of a method for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing provided in an embodiment of the present invention.

[0061] like Figure 1 As shown, the drought-flood rapid transition prediction method based on artificial intelligence and satellite remote sensing includes the following steps:

[0062] In step S101, at least one of the following data is collected: meteorological, hydrological, vegetation, land use data, GRACE inversion data, El Nino ENSO index, and climate model output data in the drought-flood rapid transition prediction basin.

[0063] In the actual implementation process, the embodiment of the present invention can first collect meteorological, hydrological, vegetation and land use data sets since 1940, including the monthly meteorological, hydrological and vegetation data of the fifth generation atmospheric reanalysis data set (ERA5) of the European Centre for Medium-Range Weather Forecasts, specifically including 2m air temperature, dew point temperature, air pressure, precipitation, soil moisture, water vapor flux divergence, cloud coverage, runoff depth, shortwave radiation and leaf area index (LAI); the land use data set adopts the land use data set published by international and domestic scholars, and the present invention extracts the monthly scale impervious area (denoted as impervious surface area, ISA) of each grid point through Google Earth Engine technology.

[0064] Satellite gravity data is used to invert terrestrial water storage. Internationally, the main institutions responsible for data interpretation include the Potsdam Geoscience Center in Germany, the Jet Propulsion Laboratory (JPL) of the California Institute of Technology, the Center for Space Research at University of Texas, Austin (CSR) and the Goddard Space Flight Center (GSFC) of NASA, and publish the global gravity field output data on a monthly scale. The present invention uses the latest sixth-generation (RL06) products released by JPL, CSR and GSFC at the same time. These three data sources have different spatial resolutions and all provide mascon monthly-scale equivalent water heights (minus the average field from 2004 to 2009) based on the concentrated mass block gravity field solution, and finally output the long sequence TWSA data set. In order to consider the possible uncertainties caused by different data products, the three GRACE / GRACE-FO gravity satellite datasets were interpolated to a 0.25°×0.25° spatial grid, and the average of each product was taken in each time step, finally obtaining a monthly dataset of terrestrial water storage from 2002 to the present.

[0065] We further obtained the monthly time series of three ENSO (El Nino Southern Oscillation) indices from 1940 to the present, including the NINO 1+2 region sea surface temperature anomaly index, the NINO 3 region sea surface temperature anomaly index, and the NINO 3.4 region sea surface temperature anomaly index.

[0066] In order to predict future climate scenarios, the five global climate models (GFDL-ESM4, IPSL-CM6A-LR, MPI-ESM1-2-HR, MRI-ESM2-0, UKESM1-0-LL) newly released by the International Coupled Model Inter-comparison Project Phase 6 (CMIP6) are used. Compared with the earlier version (CMIP5), CMIP6 uses the matrix framework of Shared Socioeconomic Pathway (SSP) and Representative Concentration Pathway (RCP). The embodiment of the present invention uses the model output data of the highest greenhouse gas emission scenario SSP5-8.5. The output variables of the five global climate models include daily temperature data, and 1985-2014 (30 years in total) is selected as the historical period; in order to predict the meteorological characteristics under the future climate change scenario, the last 30-year window of this century (2071-2100) is set as the future period.

[0067] In step S102, based on at least one data, the Clausius-Clapeyron thermodynamic equation is used to determine the saturated water vapor pressure deficit and specific humidity, and based on the saturated water vapor pressure deficit, specific humidity and a preset random forest model, key factors that meet preset preferred conditions affecting terrestrial water storage are obtained in each grid point in the drought-flood rapid transition prediction basin to obtain the preferred key factors.

[0068] Specifically, the embodiment of the present invention can use the Clausius-Clapeyron thermodynamic equation to determine the saturated water vapor pressure deficit and specific humidity, and based on the saturated water vapor pressure deficit, specific humidity and preset random forest model, combined with the sliding window method, obtain the key factors that meet the preset preferred conditions affecting terrestrial water storage in each grid point in the drought-flood rapid transition prediction basin, so as to obtain the preferred key factors, and provide support for the subsequent calibration of the long-short-term memory model suitable for terrestrial water storage simulation.

[0069] In one embodiment of the present invention, the Clausius-Clapeyron thermodynamic equation describes the nonlinear relationship between saturated water vapor pressure and air temperature as follows:

[0070] , (1)

[0071] in, and are the first and second integral constants, which are 273.16 K and 611 Pa respectively; 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.

[0072] The dew point temperature represents the temperature at which the air is cooled to water vapor saturation under the condition of constant water vapor content and air pressure. Substituting it into the Clausius-Clapeyron equation can measure the actual water vapor pressure. T 2m ) and dew point temperature ( T dew ) are substituted into formula (1) to deduce the near-saturation vapor pressure deficit .

[0073] Specific humidity q is the ratio of water vapor mass to total air mass mass, using ERA5 ground pressure p And dew point temperature, the formula is as follows:

[0074] , (2)

[0075] in, is the dew point temperature, is the ERA5 ground pressure, is the saturated water vapor pressure.

[0076] Optionally, in one embodiment of the present invention, key factors that meet preset preferred conditions affecting terrestrial water storage are obtained in each grid point in the watershed to obtain preferred key factors, including: inputting at least one driving factor among air temperature, specific humidity, saturated vapor pressure deficit, precipitation, soil moisture, water vapor flux divergence, cloud coverage, runoff depth, shortwave radiation, leaf area index, impervious area and El Nino ENSO index into a preset random forest model to obtain variables of the preset random forest model; sorting the importance of the variables of the preset random forest model to obtain sorted variables, and judging whether the sorted variables exceed a preset threshold; if the sorted variables exceed the preset threshold, generating key factors from the variables exceeding the preset threshold.

[0077] It is understandable that the preset threshold in the embodiment of the present invention may be 25%. Figure 2 As shown in the figure, for each grid point, 5×5 is set as the spatial sliding window threshold, and the window is slid sequentially according to the interval range. This method can improve the robustness of the machine learning model.

[0078] In the actual implementation process, the embodiment of the present invention can use the random forest algorithm to construct a relationship model between the driving factors within the above-mentioned spatial range and the GRACE gravity satellite terrestrial water storage, and the driving factors include temperature, specific humidity, saturated water vapor pressure deficit, precipitation, soil moisture, water vapor flux divergence, cloud coverage, runoff depth, shortwave radiation, leaf area index, impervious area and three ENSO indexes (a total of 14 driving factors); the input data of the random forest model are all GRACE inversion data period (April 2002 to present), and considering the time lag effect of each driving factor on terrestrial water storage, for each month, the driving factors of this month and the previous 1-3 months are selected as the input of the random forest model, and the embodiment of the present invention has a total of 14×4=56 variables as model input.

[0079] Furthermore, the random forest model can sort the importance of the simulation variables according to each input variable, and determine whether the sorted variables exceed a certain threshold. If the sorted variables exceed a certain threshold, the key factors are generated by the variables exceeding the preset threshold. Therefore, based on the above random forest model, the important factors affecting terrestrial water storage are selected. The embodiment of the present invention can set 25% as the threshold, that is, after sorting 56 variables, the top 14 variables are selected as key factors.

[0080] It should be noted that the preset threshold can be set by those skilled in the art according to actual conditions and is not specifically limited here.

[0081] In step S103, based on the preferred key factors and the inversion data of the gravity satellite GRACE, the long short-term memory model suitable for terrestrial water storage simulation is calibrated, and the long short-term memory model is driven by the long series of key factor data sets to obtain a reconstructed long series of terrestrial water storage inversion data sets.

[0082] It can be understood that the long short-term memory model in the embodiment of the present invention can be an artificial intelligence model.

[0083] As a possible implementation method, the embodiment of the present invention can construct a long-short-term memory model based on the preferred key factors, using the key factors and GRACE gravity satellite data since April 2002, and based on the key factors selected in step S102, use the long series data set since 1940 to drive the long-short-term memory model established in step S103, so as to reconstruct the long series terrestrial water storage data set since 1940.

[0084] In one embodiment of the present invention, the long short-term memory model constructed above is used to simulate the terrestrial water storage since April 2002. The expression for simulating the land water storage at any time is:

[0085] (3)

[0086] in, express The input variables at time, express Input variables at time 1, express Input variables at time 2, express Input variables at time 3, Represents the long short-term memory model.

[0087] In step S104, drought-flood transition events are extracted from the reconstructed long series of terrestrial water storage inversion data sets, and a temperature-drought-flood transition intensity pairing series is formed based on the average temperature in the drought-flood transition events.

[0088] In the actual implementation process, the embodiment of the present invention can extract drought-flood transition events from the reconstructed long series of terrestrial water storage inversion data sets, and extract the average temperature in the drought-flood transition events for each grid point in the predicted watershed, thereby forming a temperature-drought-flood transition intensity pairing series.

[0089] Optionally, in one embodiment of the present invention, drought-flood sudden transition events are extracted from a reconstructed long series terrestrial water storage inversion dataset, including: obtaining a long series of drought intensity based on the reconstructed long series terrestrial water storage inversion dataset; calculating thresholds corresponding to the first quantile and the second quantile of the long series drought intensity to obtain drought thresholds and flood thresholds, and determining drought months and flood months based on the drought thresholds and flood thresholds, and generating drought-flood sudden transition events when there are both drought months and flood months in two adjacent months.

[0090] It can be understood that, in the embodiment of the present invention, the first quantile may be 20%, and the second quantile may be 80%.

[0091] In the actual implementation process, the embodiment of the present invention can calculate the long series drought intensity according to the reconstructed long series terrestrial water storage inversion data set since 1940, as follows:

[0092] , (4)

[0093] in, Representative j Monthly drought intensity based on terrestrial water storage anomalies, Represents the reconstruction j Months of land water storage data, and Respectively m The mean and standard deviation of monthly TWSA,m =1,2…,4, according to j Determine the calendar month in which the month falls.

[0094] Furthermore, based on the above long series of drought intensities, the thresholds corresponding to the 20% quantile and 80% quantile of DWI were calculated, which were called drought threshold and flood threshold, respectively; months below the drought threshold were called drought months, and months above the flood threshold were called flood months; if there was one drought month and one flood month in two adjacent months, it meant that a drought-flood sudden transition event had occurred.

[0095] For a certain dry month, if multiple consecutive months with this month are flood months, the month with the highest DWI index is taken as the flood month; for a certain flood month, if multiple consecutive months with this month are dry months, the month with the lowest DWI index is taken as the dry month; therefore, the drought-flood rapid transition event defined in the present invention includes two scenarios: from flood to drought and from drought to flood, and the duration of the drought-flood rapid transition event is 2 months.

[0096] The above method is used to extract and predict drought-flood transition events at each grid point in the basin since 1940. Furthermore, the intensity measurement index of drought-flood transition events is proposed, as follows:

[0097] , (5)

[0098] Among them, CDHW represents the intensity of drought-flood transition events; DWI W and DWI D They respectively characterize the DWI intensity in the flood month and the dry month in the drought-flood abrupt transition event; The standard deviation of the DWI index in the basin is used to predict the sudden transition from drought to flood.

[0099] In step S105, based on the temperature-drought-flood transition intensity pairing series, the temperature-drought-flood transition intensity combination of each grid point in the predicted basin is divided into multiple bins with the same sample capacity using a bin scaling function to obtain the division result, and the scaling coefficient of the predicted basin drought-flood transition intensity and the near-ground temperature is obtained based on the division result.

[0100] It is understandable that the multiple bins with the same sample capacity in the embodiment of the present invention can be 4 bins with the same sample capacity. The binning scaling function in the embodiment of the present invention can better characterize the nonlinear relationship between variables and has been widely used in recent years to analyze the climate response characteristics of extreme hydrological events.

[0101] In the actual implementation process, the embodiment of the present invention can use the temperature-drought-flood transition intensity pairing series of each grid point in the basin as the input of the box element scaling method, that is, the average temperature in the drought-flood transition event is used as the sorting criterion, and the temperature-drought-flood transition intensity combination series of the same month of the study period is divided into 4 boxes with the same sample capacity; the median of the daily temperature is taken to represent the box element temperature, and the median of the drought-flood transition intensity in each box element is used to represent the average intensity of the drought-flood transition event, and finally 4 groups of temperature-drought-flood transition series are obtained.

[0102] In one embodiment of the present invention, the following formula is used to measure the intensity of drought-flood transition in the basin and the temperature change amplitude: The response function of the watershed drought-flood transition intensity and the scaling coefficient of the near-surface temperature are expressed as follows:

[0103] , (6)

[0104] in, P 1 and P 2 are the average intensity of drought-flood transition events in adjacent boxes, is the scaling factor (Scaling rate), also known as the response intensity, Indicates the amount of change, T For the temperature.

[0105] The local weighted sliding regression smoothing method is used to fit equation (4). Previous studies have found that the climate response types of extreme hydrological events can be divided into monotonic rising type, monotonic falling type and Hook structure (the inflection point is defined as the peak temperature, T pp ). The embodiment of the present invention first identifies the climate response type and T pp , and then the scaling coefficient is derived by the least squares method.

[0106] It should be noted that in the process of deriving the scaling coefficient, only the temperature below T pp The temperature-drought-flood sudden transition intensity data in the box element.

[0107] In step S106, based on the temperature data of each grid point in the forecast basin output by the global climate model, the target future drought-flood transition intensity of each grid point is predicted using the scaling coefficient to obtain the prediction result of the target future drought-flood transition intensity of each grid point.

[0108] During the actual implementation process, the embodiment of the present invention can use the scaling coefficient to preliminarily predict the target future drought-flood sudden transition intensity of each grid point based on the temperature data of each grid point in the predicted basin output by the global climate model, so as to obtain the prediction results of the target future drought-flood sudden transition intensity of each grid point, and use the emergence constraint technology to reduce the uncertainty of the prediction, which has important scientific significance for the prediction of future drought-flood sudden transition risks and disaster prevention and mitigation, and provides a certain theoretical reference and technical basis for further evaluating the environmental and disaster effects caused by the evolution of the earth system.

[0109] At each grid point in the forecast basin, for each global climate model under the SSP585 path, the temperature change value of the future period relative to the historical period is calculated, denoted as ΔT. Since the atmospheric thermal effect caused by rising temperatures is an important mechanism driving the increase in drought-flood abrupt transition events, ΔT is used. T The intensity of drought-flood transitions in the future is predicted to be as follows:

[0110] In one embodiment of the present invention, the calculation formula of the target future drought-flood transition intensity of each grid point is:

[0111] , (7)

[0112] in, It represents the average intensity of drought-flood transition in the future period of this grid point under a certain global climate model. It represents the average value of drought-flood transition intensity in the historical period obtained based on the reconstructed DWI index at this grid point. is the scaling factor, is the temperature change in the future period relative to the historical period.

[0113] For each grid point in the forecast basin, the trend of the historical average annual temperature under each global climate model (denoted as ). Furthermore, the data of all grid points in the basin predicted by the five global climate models were integrated to obtain the trend of annual average temperature change in the historical period and the trend in the future period. The emergence constraint model constructed by the pairing combination is as follows:

[0114] , (8)

[0115] in, a and b represents the parameters of the emergence constraint model, and the least square method can be used to solve the parameters of the emergence constraint model a and b .

[0116] Finally, based on the reanalysis data of the ERA5 dataset, the changing trend of the average annual temperature at each grid point during 1985-2014 was deduced, and the trend term was substituted into the constructed emergence constraint model, that is, the calculation formula for the prediction result of the target future drought-flood abrupt transition intensity of each grid point is:

[0117] , (9)

[0118] in, a and b represents the parameters of the emergence constraint model, Represents the annual mean temperature trend obtained from the ERA5 dataset at each grid point.

[0119] According to an embodiment of the present invention, a drought-flood transition prediction method based on artificial intelligence and satellite remote sensing can be proposed. An artificial intelligence model for simulating terrestrial water storage can be constructed by combining artificial intelligence methods and GRACE gravity satellite data, and a long series of meteorological and hydrological data is used to drive the artificial intelligence model to achieve long series of terrestrial water storage data reconstruction; further derive the scale coefficient for predicting the intensity of drought-flood transition in the basin, and use the latest global climate model set to predict the intensity of drought-flood transition in the future, and finally use the emergence constraint technology to reduce uncertainty, which has important scientific significance for the prediction of future drought-flood transition risks and disaster prevention and mitigation, and provides a certain theoretical reference and technical basis for further evaluating the environmental and disaster effects caused by the evolution of the earth system. Thus, the problem that the observation time of satellites in the relevant technology is short when used for drought and flood disaster monitoring and assessment, and it is difficult to achieve a long time span of drought and flood disaster retrospective risk assessment, and it usually only focuses on the changes of a single disaster type such as drought or waterlogging, and rarely considers the evolution law of drought-flood transition events, and fails to make full use of artificial intelligence technology to carry out drought-flood transition prediction.

[0120] Next, a drought-flood rapid transition prediction device based on artificial intelligence and satellite remote sensing according to an embodiment of the present invention is described with reference to the accompanying drawings.

[0121] Figure 3 The present invention is a schematic structural diagram of a drought-flood rapid transition prediction device based on artificial intelligence and satellite remote sensing.

[0122] like Figure 3 As shown, the drought-flood rapid transition prediction device 10 based on artificial intelligence and satellite remote sensing includes: a collection module 100, a determination module 200, a calibration module 300, an extraction module 400, a division module 500 and a prediction module 600.

[0123] Specifically, the acquisition module 100 is used to collect at least one of the following data: meteorological, hydrological, vegetation, land use data, inversion data of the gravity satellite GRACE, El Nino ENSO index and climate model output data in the drought-flood rapid transition prediction basin.

[0124] The determination module 200 is used to determine the saturated water vapor pressure deficit and specific humidity based on at least one data using the Clausius-Clapeyron thermodynamic equation, and obtain the key factors that meet the preset optimal conditions affecting the terrestrial water storage in each grid point in the drought-flood rapid transition prediction basin based on the saturated water vapor pressure deficit, specific humidity and a preset random forest model, so as to obtain the optimal key factors.

[0125] The calibration module 300 is used to calibrate the long short-term memory model suitable for terrestrial water storage simulation based on the preferred key factors and the inversion data of the gravity satellite GRACE, and use the long series of key factor data sets to drive the long short-term memory model to obtain a reconstructed long series of terrestrial water storage inversion data sets.

[0126] The extraction module 400 is used to extract drought-flood transition events from the reconstructed long series of terrestrial water storage inversion data sets, and to form a temperature-drought-flood transition intensity pairing series based on the average temperature in the drought-flood transition events.

[0127] The division module 500 is used to divide the temperature-drought-flood sudden transition intensity combination of each grid point in the predicted basin into multiple bins with the same sample capacity based on the temperature-drought-flood sudden transition intensity pairing series and using the bin scaling function to obtain the division result, and obtain the scaling coefficient of the predicted basin drought-flood sudden transition intensity and the near-ground temperature based on the division result.

[0128] The prediction module 600 is used to predict the target future drought-flood transition intensity of each grid point based on the temperature data of each grid point in the prediction basin output by the global climate model, and use the scaling coefficient to obtain the prediction result of the target future drought-flood transition intensity of each grid point.

[0129] Optionally, in one embodiment of the present invention, the Clausius-Clapeyron thermodynamic equation describes the nonlinear relationship between saturated water vapor pressure and air temperature as follows:

[0130] ,

[0131] in, and are the first and second integral constants, is the latent heat of vaporization constant, is the water vapor gas constant;

[0132] The calculation formula for specific humidity is:

[0133] ,

[0134] in, is the dew point temperature, is the ERA5 ground pressure, is the saturated water vapor pressure.

[0135] Optionally, in one embodiment of the present invention, the determination module 200 includes: an input unit, a sorting unit and a generating unit.

[0136] Among them, the input unit is used to input at least one driving factor of air temperature, specific humidity, saturated vapor pressure deficit, precipitation, soil moisture, water vapor flux divergence, cloud cover, runoff depth, shortwave radiation, leaf area index, impervious area and El Nino ENSO index into the preset random forest model to obtain the variables of the preset random forest model.

[0137] The sorting unit is used to sort the importance of the variables of the preset random forest model to obtain sorted variables, and to determine whether the sorted variables exceed a preset threshold.

[0138] The generating unit is used to generate key factors from the variables exceeding the preset threshold when the sorted variables exceed the preset threshold.

[0139] Optionally, in one embodiment of the present invention, the long short-term memory model is The expression for simulating the land water storage at any time is:

[0140]

[0141] in, express The input variables at time, express Input variables at time 1, express Input variables at time 2, express Input variables at time 3, Represents the long short-term memory model.

[0142] Optionally, in one embodiment of the present invention, the extraction module 400 includes: an acquisition unit and a calculation unit.

[0143] The acquisition unit is used to obtain the long series drought intensity according to the reconstructed long series terrestrial water storage inversion data set.

[0144] The calculation unit is used to calculate the thresholds corresponding to the first quantile and the second quantile of the long series drought intensity to obtain the drought threshold and the waterlogging threshold, and determine the drought month and the waterlogging month according to the drought threshold and the waterlogging threshold, and generate a drought-waterlogging sudden transition event when there are drought months and waterlogging months in two adjacent months at the same time.

[0145] Optionally, in one embodiment of the present invention, the expression for predicting the scaling coefficient of the intensity of the sudden change from drought to flood in the basin and the near-ground temperature is:

[0146] ,

[0147] in, P 1 and P 2 are the average intensity of drought-flood transition events in adjacent boxes, is the scaling factor, Indicates the amount of change, T For the temperature.

[0148] Optionally, in one embodiment of the present invention, the calculation formula of the target future drought-flood transition intensity of each grid point is:

[0149] ,

[0150] in, It represents the average intensity of drought-flood transition in the target future. It represents the average intensity of drought-flood transition. is the scaling factor, is the temperature change in the future period relative to the historical period;

[0151] The calculation formula for the prediction results of the target future drought-flood transition intensity of each grid point is:

[0152] ,

[0153] in, a and b represents the parameters of the emergence constraint model, Represents the annual mean temperature trend obtained from the ERA5 dataset at each grid point.

[0154] It should be noted that the above explanation of an embodiment of a method for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing is also applicable to an apparatus for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing in this embodiment, and will not be repeated here.

[0155] According to an embodiment of the present invention, a drought-flood transition prediction device based on artificial intelligence and satellite remote sensing can be constructed by combining artificial intelligence methods and GRACE gravity satellite data to simulate land water storage. The artificial intelligence model is driven by long-series meteorological and hydrological data to achieve long-series land water storage data reconstruction. The scaling coefficient of the intensity of drought-flood transition in the basin is further derived, and the latest global climate model set is used to predict the intensity of drought-flood transition in the future. Finally, the emergence constraint technology is used to reduce uncertainty, which has important scientific significance for the prediction of future drought-flood transition risks and disaster prevention and mitigation, and provides a certain theoretical reference and technical basis for further evaluating the environmental and disaster effects caused by the evolution of the earth system. Therefore, the problem that the observation time of satellites in the relevant technology is short when used for drought and flood disaster monitoring and assessment, and it is difficult to achieve a long time span of drought and flood disaster retrospective risk assessment, and it usually only focuses on the changes of a single disaster such as drought or waterlogging, and rarely considers the evolution law of drought-flood transition events, and fails to make full use of artificial intelligence technology to carry out drought-flood transition prediction.

[0156] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device may include:

[0157] Memory 401 , processor 402 , and a computer program stored in the memory 401 and executable on the processor 402 .

[0158] When the processor 402 executes the program, a method for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing provided in the above embodiment is implemented.

[0159] Furthermore, the electronic device further comprises:

[0160] The communication interface 403 is used for communication between the memory 401 and the processor 402 .

[0161] The memory 401 is used to store computer programs that can be executed on the processor 402 .

[0162] The memory 401 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.

[0163] If the memory 401, the processor 402 and the communication interface 403 are implemented independently, the communication interface 403, the memory 401 and the processor 402 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 4 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.

[0164] Optionally, in a specific implementation, if the memory 401, the processor 402 and the communication interface 403 are integrated on a chip, the memory 401, the processor 402 and the communication interface 403 can communicate with each other through an internal interface.

[0165] The processor 402 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.

[0166] This embodiment also provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the method for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing is implemented as described above.

[0167] 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.

[0168] 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.

[0169] 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.

[0170] 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.

[0171] 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.

[0172] 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.

[0173] 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.

[0174] 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 method for predicting sudden drought-flood transition based on artificial intelligence and satellite remote sensing, characterized in that: The following steps are involved: Collect at least one of the following data in the drought-flood rapid transition prediction basin: meteorological, hydrological, vegetation, land use data, GRACE inversion data, El Nino ENSO index, and climate model output data; Based on the at least one data, the saturated water vapor pressure deficit and the specific humidity are determined using the Clausius-Clapeyron thermodynamic equation, and based on the saturated water vapor pressure deficit, the specific humidity and a preset random forest model, key factors that meet preset optimal conditions affecting terrestrial water storage are obtained at each grid point in the drought-flood abrupt transition prediction basin to obtain optimal key factors; Based on the preferred key factors and the inversion data of the gravity satellite GRACE, a long short-term memory model suitable for the terrestrial water storage simulation is calibrated, and the long short-term memory model is driven by a long series of key factor data sets to obtain a reconstructed long series of terrestrial water storage inversion data sets; Extract drought-flood transition events from the reconstructed long series of terrestrial water storage inversion data sets, and form a temperature-drought-flood transition intensity pairing series based on the average temperature in the drought-flood transition events; Based on the temperature-drought-flood transition intensity pairing series, the temperature-drought-flood transition intensity combination of each grid point in the predicted basin is divided into multiple bins with the same sample capacity using a bin scaling function to obtain a division result, and the scaling coefficient of the predicted basin drought-flood transition intensity and the near-ground temperature is obtained according to the division result; Based on the temperature data of each grid point in the predicted basin output by the global climate model, the target future drought-flood sudden transition intensity of each grid point is predicted using the scaling coefficient to obtain the prediction result of the target future drought-flood sudden transition intensity of each grid point.

2. The method for predicting drought-flood transition based on artificial intelligence and satellite remote sensing according to claim 1, characterized in that: The Clausius-Clapeyron thermodynamic equation describes the nonlinear relationship between saturated water vapor pressure and air temperature as follows: ， in, and are the first and second integral constants, is the latent heat of vaporization constant, is the water vapor gas constant; The calculation formula of the specific humidity is: ， in, is the dew point temperature, is the ERA5 ground pressure, is the saturated water vapor pressure.

3. The method for predicting drought-flood transition based on artificial intelligence and satellite remote sensing according to claim 1, characterized in that: The key factors satisfying the preset optimal conditions affecting terrestrial water storage are obtained at each grid point in the drought-flood rapid transition prediction basin, and the optimal key factors are obtained, including: Inputting at least one driving factor of air temperature, the specific humidity, the saturated water vapor pressure deficit, precipitation, soil moisture, water vapor flux divergence, cloud cover, runoff depth, shortwave radiation, leaf area index, impervious area and the El Nino ENSO index into the preset random forest model to obtain variables of the preset random forest model; Sorting the importance of the variables of the preset random forest model to obtain sorted variables, and determining whether the sorted variables exceed a preset threshold; If the sorted variables exceed a preset threshold, the key factor is generated by the variables exceeding the preset threshold.

4. The method for predicting sudden drought-flood transition based on artificial intelligence and satellite remote sensing according to claim 1, characterized in that: The long short-term memory model is The expression for simulating the land water storage at any time is: in, express The input variables at time, express Input variables at time 1, express Input variables at time 2, express Input variables at time 3, Represents the long short-term memory model.

5. The method for predicting sudden drought-flood transition based on artificial intelligence and satellite remote sensing according to claim 1, characterized in that: Extracting drought-flood abrupt transition events from the reconstructed long series of terrestrial water storage inversion data sets includes: Obtaining a long series of drought intensities according to the reconstructed long series of terrestrial water storage inversion data set; The thresholds corresponding to the first quantile and the second quantile of the long series of drought intensities are calculated to obtain drought thresholds and waterlogging thresholds, and drought months and waterlogging months are determined according to the drought thresholds and the waterlogging thresholds. When the drought months and the waterlogging months exist simultaneously in two adjacent months, the drought-waterlogging sudden transition event is generated.

6. The method for predicting drought-flood transition based on artificial intelligence and satellite remote sensing according to claim 1, characterized in that: The expression of the predicted intensity of drought-flood transition in the basin and the scaling coefficient of the near-ground temperature is: ， in, P 1 and P 2 are the average intensity of the drought-flood transition event in adjacent boxes, is the scaling factor, Indicates the amount of change, T For the temperature.

7. The method for predicting drought-flood transition based on artificial intelligence and satellite remote sensing according to claim 1, characterized in that: The calculation formula for the target future drought-flood transition intensity of each grid point is: ， in, represents the average intensity of drought-flood transition in the target future, represents the average value of the intensity of drought-flood transition, is the scaling factor, is the temperature change in the future period relative to the historical period; The calculation formula for the prediction result of the target future drought-flood sudden change intensity of each grid point is: ， in, a and b represents the parameters of the emergence constraint model, Represents the annual average temperature trend obtained from the ERA5 dataset at each grid point.

8. A drought-flood rapid transition prediction device based on artificial intelligence and satellite remote sensing, characterized in that: include: A collection module is used to collect at least one of the following data: meteorological, hydrological, vegetation, land use data, GRACE inversion data, El Nino ENSO index and climate model output data in the drought-flood rapid transition prediction basin; a determination module, for determining the saturated water vapor pressure deficit and the specific humidity by using the Clausius-Clapeyron thermodynamic equation based on the at least one data, and obtaining the key factors satisfying the preset optimal influencing conditions of the terrestrial water storage at each grid point in the drought-flood abrupt transition prediction basin based on the saturated water vapor pressure deficit, the specific humidity and a preset random forest model, so as to obtain the optimal key factors; A calibration module, for calibrating a long short-term memory model applicable to the terrestrial water storage simulation based on the preferred key factors and the inversion data of the gravity satellite GRACE, and using a long series of key factor data sets to drive the long short-term memory model to obtain a reconstructed long series of terrestrial water storage inversion data sets; An extraction module is used to extract drought-flood transition events from the reconstructed long series of terrestrial water storage inversion data sets, and to form a temperature-drought-flood transition intensity pairing series based on the average temperature in the drought-flood transition events; A partitioning module is used to partition the temperature-drought-flood transition intensity combination of each grid point in the predicted watershed into a plurality of boxes with the same sample capacity based on the temperature-drought-flood transition intensity pairing series using a box scaling function to obtain a partitioning result, and obtain a scaling coefficient of the predicted watershed drought-flood transition intensity and the near-ground temperature according to the partitioning result; The prediction module is used to predict the target future drought-flood transition intensity of each grid point based on the temperature data of each grid point in the prediction basin output by the global climate model, and use the scaling coefficient to obtain the prediction result of the target future drought-flood transition intensity of each grid point.

9. 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 method for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement a method for predicting sudden transition from drought to flood based on artificial intelligence and satellite remote sensing as described in any one of claims 1 to 7.

Images