Early warning method for soil drying risk

Abstract

The invention relates to a soil drying risk early warning method. The method comprises the following steps: acquiring a data sequence of rainfall and soil moisture in a historical time period of a target monitoring area; then, the correlation between rainfall and soil moisture under different lag time is analyzed through the mutual information coefficient, and the causal response lag time is determined. Then, based on the rainfall and soil moisture data sequence under the causal response lag time, training a time sequence prediction model to obtain a soil moisture prediction model; and finally, inputting current rainfall data into the model, predicting a future soil moisture condition, and evaluating a soil drying risk according to a prediction result. By means of the method, the future soil moisture condition can be predicted in time, and scientific auxiliary decision-making support is provided for soil moisture adjusting measures such as irrigation. Meanwhile, by evaluating the soil drying risk, the potential soil drying problem can be found and warned in time, and a powerful guarantee is provided for ecological protection and agricultural production.

Description

Technical Field

[0001] The present application relates to the field of geographical science and technology, and in particular to an early warning method for soil desiccation risk. Background Art

[0002] The hydrological cycle of the Earth's surface system is a key factor in maintaining the balance of the ecosystem and the growth of terrestrial vegetation. As the core link in this cycle, the dynamic changes of soil moisture directly affect the health and survival of vegetation. As one of the main sources of soil moisture, the temporal and spatial distribution of precipitation is closely related to the changes in soil moisture.

[0003] At present, the regulation of soil moisture mainly relies on human intervention measures such as irrigation, but there is a lack of scientific early warning mechanisms to guide the implementation of these measures. Traditional soil moisture monitoring methods are often based on single-point observations, which are difficult to fully reflect the soil moisture conditions of the entire monitoring area, and the understanding of the causal relationship between precipitation and soil moisture is not in-depth enough, resulting in a lack of scientific basis for decision-making on regulation measures such as irrigation, which easily increases the risk of soil drying. Summary of the invention

[0004] Based on this, the purpose of this application is to provide an early warning method for soil drying risk, which can accurately capture the causal relationship between precipitation and soil moisture, and predict future soil moisture conditions based on this relationship, thereby effectively warning of soil drying risk.

[0005] The soil desiccation risk early warning method described in the embodiment of the present application comprises the following steps:

[0006] Acquire the first precipitation data sequence and the first soil moisture data sequence in the historical time period of the target monitoring area;

[0007] Acquire a lag time interval, and select a plurality of lag times from the lag time interval; extract a second precipitation data sequence and a second soil moisture data sequence from the first precipitation data sequence and the first soil moisture data sequence respectively according to each of the lag times;

[0008] Obtaining a mutual information coefficient between precipitation and soil moisture in the target monitoring area according to the second precipitation time series and the second soil moisture time series;

[0009] If the mutual information coefficient between precipitation and soil moisture satisfies the mutual information condition, the corresponding lag time is determined as the causal response lag time; the corresponding second precipitation data sequence and the second soil moisture data sequence are input into the time series prediction model for training to obtain a soil moisture prediction model;

[0010] Acquire current precipitation data, input the current precipitation data into the soil moisture prediction model, and obtain soil moisture data after the causal response lag time;

[0011] According to the soil moisture data, soil drying risk information in the target monitoring area is obtained.

[0012] The embodiment of the present application first obtains the data sequence of precipitation and soil moisture in the historical time period of the target monitoring area. Then, the mutual information coefficient is used to analyze the correlation between precipitation and soil moisture at different lag times to determine the causal response lag time. Next, based on the precipitation and soil moisture data sequence under the causal response lag time, the time series prediction model is trained to obtain the soil moisture prediction model. Finally, the current precipitation data is input into the model to predict the future soil moisture conditions, and the soil desiccation risk is evaluated based on the prediction results. Through the method of the embodiment of the present application, it is possible to accurately capture the causal relationship between precipitation and soil moisture in the target monitoring area, and establish a soil moisture prediction model based on this relationship. The model can predict the future soil moisture conditions based on the current precipitation data, and provide scientific auxiliary decision-making support for soil moisture regulation measures such as irrigation. At the same time, by evaluating the risk of soil desiccation, potential soil desiccation problems can be discovered and warned in a timely manner, providing strong guarantees for ecological protection and agricultural production.

[0013] For better understanding and implementation, the present application is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A schematic diagram of the process of the soil desiccation risk early warning method according to an embodiment of the present application;

[0015] Figure 2 This is a schematic diagram of the steps of obtaining a first precipitation data sequence in an embodiment of the present application;

[0016] Figure 3 This is a schematic diagram of the steps of obtaining a first soil moisture data sequence in an embodiment of the present application;

[0017] Figure 4 Schematic diagram of the steps for determining soil desiccation risk information in an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present application clearer, the embodiments of the present application will be further described in detail below in conjunction with the accompanying drawings. Wherein, when the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements.

[0019] It should be clear that the implementation methods described in the embodiments described below do not represent all implementation methods consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0020] The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application are also intended to include plural forms, unless the context clearly indicates other meanings. In addition, in the description of this application, unless otherwise specified, "multiple" refers to two or more. It should also be understood that the term "and / or" used in this article refers to and includes any or all possible combinations of one or more associated listed items, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone; the character " / " generally indicates that the objects associated before and after are in an "or" relationship.

[0021] It should be understood that, although the terms first, second, third, etc. may be used in the present application to describe various information, such information should not be limited to these terms, and these terms are only used to distinguish similar objects, and do not have to be used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances. Depending on the context, the words "if" / "if" used in the present application can be interpreted as "at the time of" or "when" or "in response to determination".

[0022] The embodiments of the present application relate to the field of geographical science and technology, and specifically to soil desiccation monitoring technology. At present, the regulation of soil moisture mainly relies on human intervention measures such as irrigation, but there is a lack of scientific early warning mechanisms to guide the implementation of these measures. Traditional soil moisture monitoring methods are often based on single-point observations, which are difficult to fully reflect the soil moisture conditions of the entire monitoring area, and the understanding of the causal relationship between precipitation and soil moisture is not in-depth enough, resulting in a lack of scientific basis for decision-making on regulation measures such as irrigation, which is likely to increase the risk of soil desiccation.

[0023] In this regard, an embodiment of the present application provides an early warning method for soil drying risk, which can accurately capture the causal relationship between precipitation and soil moisture, and predict future soil moisture conditions based on this relationship, thereby effectively warning of soil drying risk.

[0024] Please refer to Figure 1 The soil desiccation risk early warning method of the embodiment of the present application comprises the following steps:

[0025] S101: Acquire a first precipitation data sequence and a first soil moisture data sequence in a historical time period of a target monitoring area;

[0026] S102: Obtain a lag time interval, and select a plurality of lag times from the lag time interval; and extract a second precipitation data sequence and a second soil moisture data sequence from the first precipitation data sequence and the first soil moisture data sequence respectively according to each of the lag times;

[0027] S103: Obtaining a mutual information coefficient between precipitation and soil moisture in the target monitoring area according to the second precipitation time series and the second soil moisture time series;

[0028] S104: if the mutual information coefficient between precipitation and soil moisture satisfies the mutual information condition, the corresponding lag time is determined as the causal response lag time; the corresponding second precipitation data sequence and the second soil moisture data sequence are input into the time series prediction model for training to obtain a soil moisture prediction model;

[0029] S105: Acquire current precipitation data, input the current precipitation data into the soil moisture prediction model, and obtain soil moisture data after the causal response lag time;

[0030] S106: Obtain soil drying risk information in the target monitoring area based on the soil moisture data.

[0031] The embodiment of the present application first obtains the data sequence of precipitation and soil moisture in the historical time period of the target monitoring area. Then, the mutual information coefficient is used to analyze the correlation between precipitation and soil moisture at different lag times to determine the causal response lag time. Next, based on the precipitation and soil moisture data sequence under the causal response lag time, the time series prediction model is trained to obtain the soil moisture prediction model. Finally, the current precipitation data is input into the model to predict the future soil moisture conditions, and the soil desiccation risk is evaluated based on the prediction results. Through the method of the embodiment of the present application, it is possible to accurately capture the causal relationship between precipitation and soil moisture in the target monitoring area, and establish a soil moisture prediction model based on this relationship. The model can predict the future soil moisture conditions based on the current precipitation data, and provide scientific auxiliary decision-making support for soil moisture regulation measures such as irrigation. At the same time, by evaluating the risk of soil desiccation, potential soil desiccation problems can be discovered and warned in a timely manner, providing strong guarantees for ecological protection and agricultural production.

[0032] The soil desiccation risk early warning method of the embodiment of the present application is executed by a computer device. Each step is described in detail below.

[0033] For step S101, a first precipitation data sequence and a first soil moisture data sequence within a historical time period of a target monitoring area are obtained.

[0034] In this step, the target monitoring area is clearly defined, such as a specific farmland area or ecological protection area. The precipitation data and soil moisture data of the area in the past period of time (such as the past 5 years) are collected through data sources such as meteorological stations, satellite remote sensing or ground sensors. Precipitation data usually records daily or hourly precipitation in millimeters (mm), while soil moisture data can be expressed as volumetric water content or weight water content, recording the moisture content at different soil depths. These data are organized into time series to form the first precipitation data series and the first soil moisture data series, respectively.

[0035] Please refer to Figure 2 In one embodiment, the step of obtaining the first precipitation data sequence of the target monitoring area in the historical time period in step S101 includes:

[0036] Step S1011, obtaining precipitation change data of the target monitoring area in a historical time period;

[0037] Step S1012, obtaining a specified time interval; starting from the specified historical moment, sequentially extracting precipitation data at a plurality of moments from the precipitation change data; wherein any adjacent moments in the plurality of times are separated by the time interval;

[0038] Step S1013: obtaining the first precipitation data sequence according to the precipitation data at the plurality of moments.

[0039] Among them, step S1011 can collect precipitation change data of the target monitoring area in the past specific time period (such as the past 5 years) from a data source (such as a meteorological station, satellite remote sensing data or a ground observation station). These data are presented in the form of a time series, recording the precipitation at each time point. In step S1012, first determine a suitable time interval, which determines the resolution of the precipitation data sequence. For example, you can choose hourly, daily or weekly as the time interval. Then, starting from a specified historical moment (such as the starting point of the time series), according to this time interval, extract precipitation data at several moments from the precipitation change data. These extracted data points will be used to construct a precipitation data sequence. Finally, step S1013 arranges the precipitation data at several moments extracted in step S1012 in chronological order to form a first precipitation data sequence. This sequence will serve as the basis for subsequent analysis to evaluate the dynamic relationship between precipitation and soil moisture.

[0040] In summary, this embodiment ensures the accuracy and completeness of the precipitation data series and provides a reliable basis for subsequent analysis.

[0041] In one embodiment, the step of obtaining the first precipitation data sequence according to the precipitation data at the plurality of moments in step S1013 includes:

[0042] Step S10131, obtaining a specified first sliding time window; determining an average value of precipitation data within the first sliding time window corresponding to each moment; and obtaining the first precipitation data sequence according to the average value of precipitation data corresponding to each moment.

[0043] This embodiment aims to more smoothly represent the changing trend of precipitation data and reduce the influence of noise. Step S1013 is further refined into the following sub-steps.

[0044] In step S10131, first determine a suitable first sliding time window size, which determines the number of precipitation data points used to calculate the average value. For example, the past 24 hours, the past 7 days, or the past 30 days can be selected as the sliding time window. Then, for each moment, calculate the average value of all precipitation data points in the first sliding time window corresponding to the moment. This average value reflects the average condition of precipitation in a given time window, which helps to smooth the data and reduce fluctuations caused by individual extreme values. After obtaining the average value of precipitation data in the first sliding time window corresponding to each moment, arrange these average values ​​in chronological order to obtain the smoothed first precipitation data sequence. This sequence can better reflect the long-term change trend of precipitation than the original precipitation data, and provides more valuable information for subsequent analysis.

[0045] In summary, this embodiment improves the accuracy and stability of the prediction by smoothing the precipitation data. In particular, by introducing a sliding time window and calculating the average value of the precipitation data, the noise and fluctuation in the original data are effectively reduced, so that the precipitation data series can better reflect the actual precipitation conditions. In addition, this smoothing process also helps to improve the performance of the time series prediction model, because the model is more likely to learn the long-term change trend of the data during the training process, rather than being disturbed by short-term fluctuations. Therefore, it shows higher accuracy and practicality in predicting future soil moisture changes and assessing soil drying risks.

[0046] In one embodiment, the first sliding time window is determined according to periodic fluctuation characteristics of precipitation in the target monitoring area.

[0047] In this embodiment, in order to more accurately reflect the periodic fluctuation characteristics of precipitation in the target monitoring area and further optimize the smoothing of the precipitation data sequence, the size of the first sliding time window is determined according to the periodic fluctuation characteristics of precipitation in the area. The periodic fluctuation characteristics of precipitation generally include seasonal changes, interannual changes, and possible periodic drought or wet events. These characteristics can be identified by analyzing historical precipitation data. For example, methods such as spectrum analysis, autocorrelation function or periodogram can be used to detect the periodic components in the precipitation data and determine its main cycle. After determining the periodic fluctuation characteristics of precipitation in the target monitoring area, a first sliding time window size that matches these characteristics can be selected. For example, if there are obvious seasonal changes in the precipitation data and the main cycle is one year, a time window covering the entire season or slightly longer (such as 3 months or 6 months) can be selected to calculate the average precipitation data. If the precipitation changes are more frequent or irregular, it may be necessary to select a shorter time window (such as one week or two weeks) to better capture these changes. In this way, the size of the first sliding time window can more accurately reflect the periodic fluctuation characteristics of precipitation in the target monitoring area, thereby ensuring that the smoothed precipitation data series can more realistically reflect the long-term change trend of precipitation. This helps to improve the accuracy and reliability of subsequent analyses (such as lag time analysis, mutual information coefficient evaluation, and time series prediction model training).

[0048] Please refer to Figure 3 In one embodiment, the step of obtaining the first soil moisture data sequence of the target monitoring area within the historical time period in step S101 includes:

[0049] Step S1014, obtaining soil moisture change data of the target monitoring area within a historical time period;

[0050] Step S1015, starting from the designated historical moment, sequentially extracting soil moisture data at a plurality of moments from the soil moisture change data; wherein any adjacent moments in the plurality of times are separated by the time interval;

[0051] Step S1016: obtaining the first soil moisture data sequence according to the soil moisture data at the plurality of moments.

[0052] In step S1014, the target monitoring area and historical time period are first determined. Then, soil moisture change data of the area in the historical time period are collected by consulting relevant records, using soil moisture sensors for field measurements, or obtaining public data from meteorological, agricultural and other departments. These data usually include information such as soil moisture content, soil moisture index or soil moisture saturation at different time points.

[0053] After obtaining the complete soil moisture change data, it is necessary to extract key data points for analysis. In step S1015, a specified historical moment is first determined as the starting point, and then the soil moisture data of the corresponding moment is extracted from the soil moisture change data in sequence according to a preset time interval (such as daily, weekly or monthly, etc.). These extracted data points will be used to construct the first soil moisture data sequence.

[0054] In step S1016, after extracting soil moisture data at several moments, these data are arranged in chronological order to obtain a first soil moisture data sequence. This sequence reflects the dynamic changes in soil moisture in the target monitoring area during the historical time period, providing basic data for subsequent analysis and prediction.

[0055] In summary, this embodiment ensures the accuracy and integrity of the data by obtaining complete soil moisture change data, extracting key data points at preset time intervals, and constructing a first soil moisture data sequence. In addition, this method also takes into account the complexity and diversity of soil moisture changes and can more flexibly adapt to areas with different climatic conditions and soil types. By constructing the first soil moisture data sequence, reliable basic data is provided for subsequent analysis and prediction, which helps to more accurately assess the risk of soil drying and formulate corresponding countermeasures.

[0056] In one embodiment, the step of obtaining the first soil moisture data sequence according to the soil moisture data at the plurality of moments in step S1016 includes:

[0057] Step S10161, obtaining a specified second sliding time window; determining the average value of soil moisture data in the second sliding time window corresponding to each moment; and obtaining the first soil moisture data sequence according to the average value of soil moisture data corresponding to each moment.

[0058] This embodiment aims to more smoothly represent the changing trend of soil moisture data and reduce the influence of noise. Step S1016 is further refined into the following sub-steps.

[0059] In step S10161, a suitable second sliding time window size is first determined, which determines the number of soil moisture data points used to calculate the average value. The selection of the window size should be determined based on factors such as the research purpose, data characteristics, and the periodicity of soil moisture changes. For example, if the soil moisture changes frequently, a shorter time window (such as one or two weeks) can be selected; if the soil moisture changes are relatively stable, a longer time window (such as one month or longer) can be selected.

[0060] Then, for each moment, the average value of all soil moisture data points in the second sliding time window corresponding to that moment is calculated. This average value reflects the average condition of soil moisture in a given time window, which helps to smooth the data and reduce fluctuations caused by individual extreme values. After obtaining the average value of soil moisture data in the second sliding time window corresponding to each moment, these average values ​​are arranged in chronological order to obtain the smoothed first soil moisture data sequence. This sequence can better reflect the long-term change trend of soil moisture than the original soil moisture data, and provides more valuable information for subsequent analysis.

[0061] In summary, this embodiment effectively reduces the noise and fluctuation in the original data by introducing the second sliding time window and calculating the average value of the soil moisture data, so that the soil moisture data series can better reflect the actual soil moisture conditions. In addition, this smoothing process also helps to improve the performance of the time series prediction model, because the model is more likely to learn the long-term change trend of the data during the training process, rather than being disturbed by short-term fluctuations. Therefore, the method of this embodiment shows higher accuracy and practicality in predicting future soil moisture changes and assessing soil drying risks.

[0062] In one embodiment, the second sliding time window is determined according to periodic fluctuation characteristics of soil moisture in the target monitoring area.

[0063] In this embodiment, in order to more accurately reflect the periodic fluctuation characteristics of soil moisture in the target monitoring area and further optimize the smoothing of the soil moisture data sequence, the size of the second sliding time window is determined according to the periodic fluctuation characteristics of soil moisture in the area. The periodic fluctuation characteristics of soil moisture are usually related to various factors such as climate conditions, soil type, vegetation coverage, and topography. These characteristics can be identified by analyzing historical soil moisture data. For example, methods such as spectrum analysis, autocorrelation function or periodogram can be used to detect the periodic components in soil moisture data and determine its main period.

[0064] After determining the periodic fluctuation characteristics of soil moisture in the target monitoring area, a second sliding time window size can be selected to match these characteristics. For example, if there are obvious seasonal changes in the soil moisture data and the main cycle is one year, a time window covering the entire season or a slightly longer time window (such as 3 months or 6 months) can be selected to calculate the average soil moisture data. If soil moisture changes are more frequent or irregular, you may need to select a shorter time window (such as one week or two weeks) to better capture these changes.

[0065] In this way, the size of the second sliding time window can more accurately reflect the periodic fluctuation characteristics of soil moisture in the target monitoring area, thereby ensuring that the smoothed soil moisture data series can more realistically reflect the long-term change trend of soil moisture. This helps to improve the accuracy and reliability of subsequent analyses (such as lag time analysis, mutual information coefficient evaluation, and time series prediction model training).

[0066] For step S102, a lag time interval is obtained, and several lag times are selected from the lag time interval; according to each of the lag times, a second precipitation data sequence and a second soil moisture data sequence are respectively extracted from the first precipitation data sequence and the first soil moisture data sequence.

[0067] Considering that there may be a hysteresis effect in the impact of precipitation on soil moisture, this step sets a reasonable hysteresis time interval, such as 0 to 10 days, in days. Within this interval, multiple specific hysteresis times are selected, such as 1 day, 3 days, 5 days, 7 days and 10 days. For each selected hysteresis time, the precipitation data of the corresponding time period (for example, the period from the current time minus the hysteresis time to the current time) is extracted from the first precipitation data sequence to form a second precipitation data sequence. Similarly, the soil moisture data corresponding to the time window of the second precipitation data sequence is extracted from the first soil moisture data sequence to form a second soil moisture data sequence.

[0068] For step S103, the mutual information coefficient between precipitation and soil moisture in the target monitoring area is obtained according to the second precipitation time series and the second soil moisture time series.

[0069] The mutual information coefficient is a statistic that measures the dependency between two random variables and can capture the nonlinear relationship between variables. In this step, the mutual information coefficient between the second precipitation data sequence and the second soil moisture data sequence corresponding to each lag time is calculated using a mutual information calculation method (such as the Kruskal-Schmidt algorithm or a method based on kernel density estimation). This coefficient reflects the predictive power of precipitation data for soil moisture data. The higher the coefficient, the stronger the correlation between the two.

[0070] In one embodiment, the step of extracting the second precipitation data sequence and the second soil moisture data sequence from the first precipitation data sequence and the first soil moisture data sequence respectively according to each of the lag times in step S102 includes:

[0071] Step S1021, extracting a plurality of second precipitation data sequences from the first precipitation data sequence according to the lag time, and extracting a second soil moisture data sequence corresponding to each second precipitation data sequence from the first soil moisture data sequence; wherein the plurality of second precipitation data sequences respectively start at different historical moments;

[0072] The step of obtaining the mutual information coefficient of precipitation and soil moisture in the target monitoring area according to the second precipitation time series and the second soil moisture time series in step S103 includes:

[0073] Step S1031, obtaining a mutual information coefficient between precipitation and soil moisture corresponding to each second precipitation data sequence according to each second precipitation data sequence and the corresponding second soil moisture data sequence;

[0074] Step S1032: Calculate a mean mutual information coefficient based on the mutual information coefficient between precipitation and soil moisture corresponding to each of the second precipitation data sequences; and determine the mean mutual information coefficient as the mutual information coefficient between precipitation and soil moisture in the target monitoring area.

[0075] In step S1021, for each predetermined lag time (e.g., from 0 days to a certain maximum lag number of days, such as 30 days, 60 days, etc.), a number of second precipitation data sequences are extracted from the first precipitation data sequence. The starting time of these second precipitation data sequences is different, but the length of each sequence is equal to the lag time plus a fixed time range (the size of this time range can be determined as needed to ensure that there are enough data points for correlation analysis). For example, if the lag time is 10 days and the time range size is 1 month (30 days), the second precipitation data sequence with the 1st day to the (N-40th)th day (N is the total number of days in the first precipitation data sequence) as the starting time can be extracted from the first precipitation data sequence.

[0076] At the same time, for each lag time and the corresponding second precipitation data sequence, the second soil moisture data sequence of the corresponding time period (i.e., the time period after the lag time and the same length as the second precipitation data sequence) is extracted from the first soil moisture data sequence. These second soil moisture data sequences correspond to the second precipitation data sequences one by one in time, so as to facilitate the subsequent correlation analysis.

[0077] It is worth noting that since each lag time can correspond to a second precipitation data sequence with multiple starting times (depending on the total number of days N and the time window size), there will be a set of second precipitation data sequences and corresponding second soil moisture data sequences at each lag time.

[0078] Step S1031 calculates the mutual information coefficient between precipitation and soil moisture corresponding to each second precipitation data sequence according to each second precipitation data sequence and the corresponding second soil moisture data sequence.

[0079] Step S1032 calculates the mean of the mutual information coefficients of all the second precipitation data sequences corresponding to each lag time as the mutual information coefficient between precipitation and soil moisture at the lag time. Then, the variation trend of these mutual information coefficients with the lag time can be further analyzed to evaluate the lag effect of precipitation on soil moisture.

[0080] In summary, the soil desiccation risk warning method proposed in the present invention not only realizes the in-depth mining of the dynamic relationship between precipitation and soil moisture in the target monitoring area through fine step design and parameter optimization, but also comprehensively evaluates the correlation between them by considering the hysteresis effect of precipitation influence and calculating the mutual information coefficient. In particular, by extracting multiple groups of second precipitation data sequences and second soil moisture data sequences from the first precipitation data sequence and the first soil moisture data sequence for each lag time, and calculating the mutual information coefficient and mean corresponding to each group of sequences, the method of this embodiment can more accurately capture the complex dependency between precipitation and soil moisture, and evaluate their correlation strength at different lag times. In addition, the method of this embodiment also takes into account the complexity and diversity of soil moisture changes and the lag of precipitation influence, and can more flexibly adapt to areas with different climatic conditions and soil types. By comprehensively analyzing the mutual information coefficients at different lag times, this method provides strong technical support for assessing soil desiccation risks and helps to formulate more accurate countermeasures.

[0081] In one embodiment, the step of calculating the mean mutual information coefficient according to the mutual information coefficient between precipitation and soil moisture corresponding to each second precipitation data sequence in step S1032 includes:

[0082] Step S10321, performing a significance test on each of the second precipitation data sequences and the corresponding second soil moisture data sequences, determining the second precipitation data sequences and the corresponding second soil moisture data sequences whose significance test results are greater than a preset threshold, and determining the corresponding mutual information coefficients of precipitation and soil moisture;

[0083] Step S10322, based on the mutual information coefficients of precipitation and soil moisture corresponding to the significance check result being greater than a preset threshold, the mean mutual information coefficient is calculated.

[0084] This embodiment aims to ensure that the calculated mutual information coefficient can truly reflect the correlation between precipitation and soil moisture and reduce misjudgment caused by random errors or data fluctuations. Step S1032 is further refined into the following sub-steps.

[0085] In step S10321, a significance test is performed on each second precipitation data sequence and the corresponding second soil moisture data sequence using an appropriate statistical test method (such as a permutation test, a bootstrap test, or a method based on hypothesis testing, etc.). The purpose of the significance test is to determine whether the calculated mutual information coefficient is significantly greater than zero, that is, to determine whether there is a significant correlation between precipitation and soil moisture.

[0086] In the process of significance test, a preset threshold value (such as 0.05, 0.01, etc.) is set to determine whether the significance test result is significant. If the significance test result is greater than the preset threshold value (usually expressed as a p value less than the preset threshold value, i.e., p<0.05 or p<0.01, etc.), it is considered that there is a significant correlation between the second precipitation data sequence and the corresponding second soil moisture data sequence, and the corresponding mutual information coefficient is retained.

[0087] Step S10322, after determining the second precipitation data sequence and the corresponding second soil moisture data sequence whose significance check result is greater than the preset threshold, calculates the mean of the mutual information coefficients corresponding to these sequences. This mean reflects the average level of correlation between precipitation and soil moisture under the condition of significant correlation.

[0088] It is worth noting that, since there may be multiple sets of second precipitation data series and second soil moisture data series that pass the significance test at different lag times, the corresponding mutual information coefficient mean can be calculated for each lag time. In this way, a mutual information coefficient mean sequence that changes with lag time can be obtained, which is used to further analyze the lag effect of precipitation on soil moisture.

[0089] In summary, by introducing the significance test step, the soil desiccation risk warning method proposed in this embodiment can more accurately evaluate the correlation between precipitation and soil moisture. The significance test not only reduces the misjudgment caused by random errors or data fluctuations, but also improves the reliability of the mutual information coefficient as a correlation metric. In addition, by calculating the mean of the mutual information coefficients of precipitation and soil moisture corresponding to the significance check result being greater than the preset threshold, the method of this embodiment can provide a more robust correlation evaluation result. This mean reflects the average level of correlation between precipitation and soil moisture under the condition of significant correlation, which is helpful for subsequent risk assessment and formulation of countermeasures.

[0090] For step S104, if the mutual information coefficient between precipitation and soil moisture satisfies the mutual information condition, the corresponding lag time is determined as the causal response lag time; the corresponding second precipitation data sequence and the second soil moisture data sequence are input into the time series prediction model for training to obtain a soil moisture prediction model.

[0091] This step sets the mutual information condition and selects a set of precipitation data sequences and soil moisture sequences that meet the mutual information condition, and determines the corresponding lag time as the causal response lag time. For example, the lag time with the highest mutual information coefficient (or the lag time with a mutual information coefficient higher than a preset threshold) is selected as the causal response lag time. This means that after this lag time, the effect of precipitation on soil moisture is most significant. Then, the second precipitation data sequence and the second soil moisture data sequence corresponding to the causal response lag time are used as training data and input into a time series prediction model (such as an ARIMA model, an LSTM neural network, or a GRU neural network, etc.). Through model training, a soil moisture prediction model that can predict future soil moisture changes is obtained.

[0092] In one embodiment, before the step of determining the corresponding lag time as the causal response lag time if the mutual information coefficient between precipitation and soil moisture satisfies the mutual information condition in step S104, the method includes the following steps:

[0093] Step S1041, compare the mutual information coefficient of the precipitation and soil moisture with a preset mutual information threshold. If the mutual information coefficient of the precipitation and soil moisture is greater than the mutual information threshold, determine that the mutual information coefficient of the precipitation and soil moisture meets the mutual information condition; wherein the mutual information threshold is determined based on the mutual information coefficient corresponding to the lag time of 0.

[0094] This embodiment first calculates the mutual information coefficient between precipitation and soil moisture when the lag time is 0. This coefficient reflects the instantaneous correlation between precipitation and soil moisture without the lag effect. Then, a mutual information threshold is set according to this instantaneous correlation coefficient. The selection of this threshold should be high enough to ensure that the mutual information condition is considered to be met only when there is a significant correlation between precipitation and soil moisture. Specifically, in one embodiment, the mutual information threshold is set as: no lag (i.e., the lag time is 0) mutual information coefficient + (1-no lag mutual information coefficient) * 10%.

[0095] Next, the mutual information coefficient of precipitation and soil moisture corresponding to each lag time is compared with the preset mutual information threshold. If the mutual information coefficient at a certain lag time is greater than the mutual information threshold, it is considered that the correlation between precipitation and soil moisture at this lag time meets the mutual information condition, that is, there is a significant causal relationship or correlation between them.

[0096] In summary, by comparing the mutual information coefficient of precipitation and soil moisture with the preset mutual information threshold, this embodiment can ensure that the corresponding lag time is determined as the causal response lag time only when there is a significant correlation between them. This not only improves the accuracy and reliability of the method, but also provides a more solid foundation for subsequent analysis of soil desiccation risk. Therefore, the method of this embodiment has broad application prospects and important practical value in soil desiccation risk warning and causal analysis.

[0097] For step S105, current precipitation data is obtained, and the current precipitation data is input into the soil moisture prediction model to obtain soil moisture data after the causal response lag time.

[0098] This step collects the current precipitation data of the target monitoring area in real time or regularly. These data are input into the trained soil moisture prediction model, and the model will output the predicted future soil moisture data corresponding to the time period after the causal response lag time.

[0099] For step S106, soil drying risk information in the target monitoring area is obtained based on the soil moisture data.

[0100] This step determines the soil desiccation risk information based on the predicted soil moisture data. Specifically, risk judgment can be made by setting one or more soil desiccation risk thresholds. For example, when the predicted soil moisture content is lower than a certain critical value, it is considered that there is a soil desiccation risk. By comparing the predicted soil moisture data with these thresholds, the soil desiccation risk information in the target monitoring area is generated. This information may include risk level, risk area distribution, risk duration, etc., to provide decision makers with timely warnings and intervention recommendations.

[0101] Please refer to Figure 4 In one embodiment, the step of obtaining soil drying risk information in the target monitoring area according to the soil moisture data in step S106 includes:

[0102] Step S1061, obtaining soil property information of the target monitoring area; obtaining the soil drying coefficient in the target monitoring area according to the soil moisture data and the soil property information;

[0103] Step S1062: determining soil desiccation risk information in the target monitoring area according to the soil desiccation coefficient.

[0104] In step S1061, soil property information of the target monitoring area is first obtained, including but not limited to soil type, soil texture (such as sand, loam, clay, etc.), soil structure, soil bulk density, soil water content, etc. Such soil property information is crucial for understanding the water retention capacity and drying sensitivity of the soil.

[0105] Next, the soil desiccation coefficient in the target monitoring area is calculated based on the acquired soil moisture data and soil property information. The soil desiccation coefficient is a comprehensive indicator used to quantify the degree of reduction in soil moisture and the degree of soil drying. It can be calculated based on the comparison of soil moisture data and soil property information (such as soil field water holding capacity, soil wilting coefficient, etc.). For example, the soil desiccation coefficient can be calculated using the following formula: Soil desiccation coefficient = (soil field water holding capacity - current soil moisture content) / (soil field water holding capacity - soil wilting coefficient). Among them, soil field water holding capacity is the maximum moisture content that the soil can maintain under field conditions, and the soil wilting coefficient is the moisture content when the soil moisture is reduced to the point where the plant can no longer absorb moisture from it.

[0106] After obtaining the soil desiccation coefficient, step S1062 determines the soil desiccation risk information in the target monitoring area according to the preset soil desiccation risk level classification standard (such as low risk, medium risk, high risk, etc.). This standard can be formulated based on local soil characteristics, climate conditions, crop types and growth stages, etc.

[0107] Specifically, the soil desiccation coefficient can be compared with a preset risk level threshold to determine the soil desiccation risk level in the target monitoring area. For example, if the soil desiccation coefficient is less than a certain threshold, it is judged as low risk; if the soil desiccation coefficient is within a certain range, it is judged as medium risk; if the soil desiccation coefficient is greater than another threshold, it is judged as high risk.

[0108] In summary, by introducing soil property information and the calculation of the soil drying coefficient, this embodiment can more accurately assess the soil drying risk in the target monitoring area. Soil property information provides a basis for understanding the water retention capacity and drying sensitivity of the soil, while the soil drying coefficient quantifies the degree of reduction in soil moisture and the degree of soil drying. In addition, through the preset soil drying risk level classification standard, the method of this embodiment can convert the soil drying coefficient into intuitive soil drying risk information, providing decision makers with clearer and more practical risk assessment results. This not only contributes to scientific research and practical applications in the fields of agriculture, ecology, and environment, but also provides a strong guarantee for formulating more accurate soil drying risk warnings and response measures.

[0109] In one embodiment, after the step of obtaining the soil drying risk information in the target monitoring area according to the soil moisture data in step S1062, the method includes the following steps:

[0110] Step S1063: If it is determined according to the soil desiccation risk information that the soil desiccation risk in the target monitoring area is greater than a preset risk threshold, the soil irrigation equipment in the target monitoring area is controlled to perform irrigation.

[0111] In step S1063, firstly, according to the soil desiccation risk information obtained in step S1062, it is determined whether the soil desiccation risk in the target monitoring area is greater than a preset risk threshold. This risk threshold can be set according to local soil characteristics, climate conditions, crop types and growth stages, etc., to ensure that timely measures are taken when the soil desiccation risk reaches a certain level.

[0112] If the judgment result is yes, that is, the soil drying risk is greater than the preset risk threshold, the soil irrigation equipment in the target monitoring area will be automatically controlled for irrigation. Irrigation equipment can include different types of irrigation systems such as sprinkler irrigation, drip irrigation, and infiltration irrigation. The specific choice depends on local soil conditions, crop needs, and water resources. When controlling irrigation, parameters such as irrigation volume, irrigation frequency, and irrigation time can also be adjusted according to the specific soil drying degree and crop needs in the target monitoring area to ensure the best irrigation effect.

[0113] In summary, when the soil desiccation risk is greater than the preset risk threshold, the present embodiment can automatically control the irrigation equipment to irrigate, thereby effectively alleviating the soil desiccation problem and ensuring the safety of crop growth and ecological environment. In addition, by adjusting the irrigation parameters according to the specific soil desiccation degree and crop demand in the target monitoring area, the method of the present embodiment can also achieve precision irrigation and improve the utilization efficiency of water resources. This not only contributes to the sustainable development of agricultural production, but also provides strong support for ecological environmental protection and resource conservation.

[0114] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present application, and the present application is also intended to include these modifications and modifications.

Claims

1. A soil desiccation risk early warning method, characterized in that: The following steps are involved: Acquire the first precipitation data sequence and the first soil moisture data sequence in the historical time period of the target monitoring area; Acquire a lag time interval, and select a plurality of lag times from the lag time interval; extract a second precipitation data sequence and a second soil moisture data sequence from the first precipitation data sequence and the first soil moisture data sequence respectively according to each of the lag times; Obtaining a mutual information coefficient between precipitation and soil moisture in the target monitoring area according to the second precipitation time series and the second soil moisture time series; If the mutual information coefficient between precipitation and soil moisture satisfies the mutual information condition, the corresponding lag time is determined as the causal response lag time; Inputting the corresponding second precipitation data sequence and the second soil moisture data sequence into a time series prediction model for training to obtain a soil moisture prediction model; Acquire current precipitation data, input the current precipitation data into the soil moisture prediction model, and obtain soil moisture data after the causal response lag time; According to the soil moisture data, soil drying risk information in the target monitoring area is obtained.

2. The soil desiccation risk early warning method according to claim 1, characterized in that: The step of obtaining a first precipitation data sequence of the target monitoring area within a historical time period includes: Obtain precipitation change data in the target monitoring area within the historical time period; Obtaining a specified time interval; starting from a specified historical moment, sequentially extracting precipitation data at a plurality of moments from the precipitation change data; wherein any adjacent moments in the plurality of times are separated by the time interval; The first precipitation data sequence is obtained according to the precipitation data at the plurality of moments.

3. The soil desiccation risk early warning method according to claim 2, characterized in that: The step of obtaining the first precipitation data sequence according to the precipitation data at the plurality of moments comprises: Obtain a specified first sliding time window; determine an average value of precipitation data within the first sliding time window corresponding to each moment; and obtain the first precipitation data sequence according to the average value of precipitation data corresponding to each moment.

4. The soil desiccation risk early warning method according to claim 2, characterized in that: The step of obtaining a first soil moisture data sequence of the target monitoring area within a historical time period includes: Obtain soil moisture change data in the target monitoring area within a historical period; Starting from a specified historical moment, sequentially extracting soil moisture data at a plurality of moments from the soil moisture change data; wherein any adjacent moments in the plurality of times are separated by the time interval; The first soil moisture data sequence is obtained according to the soil moisture data at the plurality of moments.

5. The soil desiccation risk early warning method according to claim 4, characterized in that: The step of obtaining the first soil moisture data sequence according to the soil moisture data at the plurality of moments comprises: Obtain a specified second sliding time window; determine the average value of soil moisture data in the second sliding time window corresponding to each moment; and obtain the first soil moisture data sequence according to the average value of soil moisture data corresponding to each moment.

6. The soil desiccation risk early warning method according to claim 5, characterized in that: The first sliding time window is determined according to the periodic fluctuation characteristics of precipitation in the target monitoring area; the second sliding time window is determined according to the periodic fluctuation characteristics of soil moisture in the target monitoring area.

7. The soil desiccation risk early warning method according to claim 1, characterized in that: The step of extracting the second precipitation data sequence and the second soil moisture data sequence from the first precipitation data sequence and the first soil moisture data sequence respectively according to each of the lag times comprises: According to the lag time, extracting a plurality of second precipitation data sequences from the first precipitation data sequence, and extracting a second soil moisture data sequence corresponding to each second precipitation data sequence from the first soil moisture data sequence; wherein the plurality of second precipitation data sequences respectively start at different historical moments; The step of obtaining the mutual information coefficient between precipitation and soil moisture in the target monitoring area according to the second precipitation time series and the second soil moisture time series comprises: According to each of the second precipitation data sequences and the corresponding second soil moisture data sequences, obtaining a mutual information coefficient between precipitation and soil moisture corresponding to each of the second precipitation data sequences; According to the mutual information coefficient between precipitation and soil moisture corresponding to each of the second precipitation data sequences, a mean mutual information coefficient is calculated; and the mean mutual information coefficient is determined as the mutual information coefficient between precipitation and soil moisture in the target monitoring area.

8. The soil desiccation risk early warning method according to claim 7, characterized in that: The step of calculating the mutual information coefficient mean value according to the mutual information coefficient between precipitation and soil moisture corresponding to each of the second precipitation data sequences comprises: Performing a significance test on each of the second precipitation data sequences and the corresponding second soil moisture data sequences, determining the second precipitation data sequences and the corresponding second soil moisture data sequences whose significance test results are greater than a preset threshold, and determining the corresponding mutual information coefficients of precipitation and soil moisture; According to the mutual information coefficient between precipitation and soil moisture corresponding to the significance check result being greater than the preset threshold, the mean mutual information coefficient is calculated.

9. The soil desiccation risk early warning method according to claim 1, characterized in that: Before the step of determining the corresponding lag time as the causal response lag time if the mutual information coefficient between precipitation and soil moisture satisfies the mutual information condition, the method comprises the following steps: The mutual information coefficient of the precipitation and soil moisture is compared with a preset mutual information threshold. If the mutual information coefficient of the precipitation and soil moisture is greater than the mutual information threshold, it is determined that the mutual information coefficient of the precipitation and soil moisture meets the mutual information condition; wherein the mutual information threshold is determined according to the mutual information coefficient corresponding to the lag time of 0.

10. The soil desiccation risk early warning method according to claim 1, characterized in that: The step of obtaining soil drying risk information in the target monitoring area according to the soil moisture data includes: Acquire soil property information of the target monitoring area; obtain soil drying coefficient in the target monitoring area according to the soil moisture data and the soil property information; According to the soil desiccation coefficient, soil desiccation risk information in the target monitoring area is determined.

Images