A new method for directly and precisely determining the threshold of flood events based on percentiles

Abstract

The present invention discloses a new method for directly and accurately determining the threshold of flood events based on percentiles, which relates to the field of flood monitoring in land water storage inversion technology, and solves the problem that the existing methods cannot directly and accurately determine the MFPI threshold. The key points of its technical solution specifically include the following steps: calculating the GRACE terrestrial water storage anomaly data TWSA by using the spherical harmonic coefficients SHCs of the earth's gravity field obtained by the Gravity Recovery and Climate Experiment GRACE satellite; reading precipitation data, and constructing a modified flood potential index MFPI by using precipitation and detrended terrestrial water storage anomaly data TWSA<subgt;detrend< / subgt>. The effect is that the new method proposed in this study directly and accurately determines the MFPI threshold based on the percentile method. This method avoids the influence brought by human factors and step complexity in the traditional method, and effectively improves the accuracy of flood event monitoring.

Description

Technical Field

[0001] The present invention relates to the field of flood monitoring based on terrestrial water storage inversion technology, and specifically to a new method for directly and accurately determining flood event thresholds based on percentiles. Background Art

[0002] With the intensification of global warming, the acceleration of the hydrological cycle has led to frequent flood events, so it has become particularly important to strengthen flood monitoring. The Gravity Recovery and Climate Experiment (GRACE) satellite provides an effective means for large-scale regional surface and groundwater monitoring. The Modified Flood Potential Index (MFPI) calculated based on precipitation and GRACE data has been widely used in flood monitoring because it can comprehensively evaluate the impact of precipitation and terrestrial water saturation on flood disasters. However, accurately determining the MFPI threshold is the key to judging flood events, which directly affects the evaluation results of MFPI in identifying floods. Therefore, it is particularly important to accurately determine this threshold.

[0003] Currently, the threshold of MFPI is usually determined indirectly through runoff data. The specific steps are as follows: First, the flood events and their occurrence times during the study period are determined by statistically analyzing runoff data based on the percentile method; second, the MFPI data corresponding to the time set are extracted; finally, the average value of these MFPI data is calculated to finally determine the MFPI threshold. This method has the following deficiencies: the steps are cumbersome, and it lacks universality in basins lacking runoff data. In addition, the impact of reservoir regulation on runoff data cannot be ignored, and over-reliance on runoff data may reduce the accuracy of MFPI in identifying flood events.

[0004] In summary, in order to improve the accuracy of flood event monitoring, it is particularly important to propose a new method for determining the MFPI threshold. Summary of the Invention

[0005] The purpose of the present invention is to provide a new method for directly and accurately determining flood event thresholds based on percentiles to solve the problems raised in the above background art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A new method for directly and accurately determining flood event thresholds based on percentiles, based on precipitation data and detrended terrestrial water storage anomaly data (TWSA) provided by the GRACE satellite detrend)Construct the MFPI. Directly calculate the MFPI according to the percentile method, and obtain the thresholds of the MFPI under different probabilities. The new threshold determination method can accurately monitor flood events and avoid the influence of the disadvantages of existing methods on flood monitoring. The specific steps are as follows:

[0008] Step a) Calculate T using the GRACE RL06 spherical harmonic coefficients detrend (T detrend represents the variable of detrended terrestrial water storage data), to prepare the variable I for calculating the MFPI (step b);

[0009] Step b) Read precipitation data, and use the T detrend data in a) and precipitation data to calculate the variable A of the Modified Flood Potential Amount (MFPA), and standardize A to obtain I;

[0010] Step c) Based on the I data calculated in step b), according to the newly proposed method of directly and accurately determining the MFPI threshold by percentile, calculate the thresholds of the MFPI under different probabilities;

[0011] Step d) Judge flood events based on the thresholds obtained in step c), and then determine the accuracy of the threshold determination method in identifying flood events.

[0012] As a further technical solution of the present invention: Calculate the Terrestrial Water Storage Anomaly (TWSA) from the GRACE RL06 spherical harmonic coefficients: ① In the technical note document Techinical Note 13, replace the first-order term of the GRACE spherical harmonic coefficients, and use the product SLR RL06 released by satellite laser ranging to replace the C 20 term of the GRACE spherical harmonic coefficients; ② Deduct the average value of the GRACE spherical harmonic coefficients from January 2004 to December 2009; ③ Use the SWENSON decorrelation filter to weaken the north-south strip error caused by the correlation of the spherical harmonic coefficients, and use a 300km Gaussian filter to smooth the spherical harmonic coefficients to weaken the high-frequency error; ④ Remove the Glacial Isostatic Adjustment (GIA) effect; ⑤ Calculate T (T represents the variable of TWSA) using the GRACE spherical harmonic data; ⑥ Correct the leakage error in the T data using the RMS_NSE scale factor method.

[0013] As a further technical solution of the present invention, the calculation process of the MFPI is as follows:

[0014] The theory of using MFPI to monitor flood events is as follows: The GRACE satellite has the ability to monitor the regional land surface and groundwater. Based on this, the maximum monthly precipitation ΔH that can be stored in the monitored area is calculated, and A is obtained by subtracting ΔH from the precipitation data. Then, A is standardized to estimate the variable I of MFPI;

[0015] ① Obtain precipitation data and obtain T by detrending T detrend , to prepare for constructing I data;

[0016] ② Use T detrend data to estimate the maximum monthly precipitation that can be stored in the monitored area, as shown in formula (1);

[0017] ΔH(t) = MAX(T detrend ) - T detrend (t - 1) (1)

[0018] In the formula, ΔH is the maximum monthly precipitation that can be stored in the monitored area, T detrend is the detrended T, T is the variable of TWSA, t represents the t-th month, and MAX represents taking the maximum value.

[0019] ③ Use precipitation data and ΔH to calculate A, as shown in formula (2);

[0020] A(t) = R(t) - ΔH(t) (2)

[0021] In the formula, A is the variable of the modified flood potential, R is the precipitation, ΔH is the maximum monthly precipitation that can be stored in the monitored area, and t represents the t-th month.

[0022] ④ Standardize A to obtain I, as shown in formula (3);

[0023]

[0024] In the formula, I is the variable of the modified flood potential index, A is the variable of the modified flood potential, t represents the t-th month, and MAX represents taking the maximum value. The value of I varies between (-∞, 1), and the closer it is to 1, the greater the probability of a flood event occurring.

[0025] As a further technical solution of the present invention: According to the newly proposed method of directly and accurately determining the MFPI threshold based on percentiles, the thresholds of MFPI under different probabilities are determined. The following is the new threshold determination method proposed by the present invention.

[0026] The new method for directly and accurately determining the MFPI threshold using percentiles is as follows: Set the threshold for directly determining the Modified Flood Potential Index (MFPI) at various probabilities based on the percentile method as P. Divide a set of data into two parts, including r% of the data being less than the threshold P and (100 - r)% of the data being greater than P. When the variable I of the Modified Flood Potential Index is greater than P, it is determined as a flood event. The threshold calculation steps are as follows:

[0027] n = 1 + (T - 1)r% (4)

[0028] Where T is the length of the time series, and r% is the probability of possible flooding, taking 85 th 、90 th and 95 th respectively, and n represents the number of non - flood events at the r% probability;

[0029] P = X [n] + (X [n+1] - X [n] )(n - [n]) (5)

[0030] Where [] is the rounding function, P represents the threshold for directly determining the Modified Flood Potential Index (MFPI) at various probabilities based on the percentile method. Sort the I time series from smallest to largest to get X, and n represents the number of non - flood events at the r% probability.

[0031] As a further technical solution of the present invention, first determine the flood events from the calculated threshold, and then determine the accuracy of the threshold determination method in identifying flood events. The specific calculation process is as follows:

[0032] First, compare the value of I with the threshold one by one, and determine the values greater than the threshold as flood events.

[0033] Then, take the flood events determined from the runoff data as a reference, and introduce evaluation indicators to evaluate the accuracy of the threshold determination method in identifying flood events. The method for constructing the evaluation indicators is as follows.

[0034] ① Establish a confusion matrix for determining the accuracy of flood event identification, and determine the authenticity of each flood event in turn according to Table 1.

[0035] Table 1 Confusion matrix for evaluating the accuracy of flood event identification

[0036]

[0037] ② Evaluate the ability of different threshold determination methods in identifying flood events according to formulas (6), (7), and (8), where the flood events monitored by the runoff data are used as reference values.

[0038] Accuracy Rate (ACC): It represents the ratio of the flood index correctly identifying flood and non-flood events, and the ideal value is 1.

[0039]

[0040] True Positive Rate (TPR): It represents the ratio of the flood events identified by the flood index to all flood events, and the ideal value is 1.

[0041]

[0042] False Positive Rate (FPR): It represents the ratio of non-flood events being misidentified as flood events by the flood index, and the ideal value is 0.

[0043]

[0044] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0045] Existing threshold determination methods have problems such as cumbersome steps, low accuracy, and lack of universality in the absence of runoff data. In addition, this method does not consider the influence of reservoir regulation on the accuracy of estimating the threshold. In view of these deficiencies, this study proposes a more robust threshold determination method to improve the existing method. The percentile method used in the newly proposed threshold determination method is an important statistical tool that can determine the threshold of extreme events at different probability levels, providing a scientific and accurate basis for identifying extreme climate events. The study also found that the design concept of MFPI is that the closer the value is to 1, the greater the potential for flood events to occur, which is consistent with the principle of determining extreme events by the percentile method. Therefore, theoretically, the percentile method can be used to directly estimate the threshold of MFPI. This study proposes a new method for directly and accurately determining the threshold of MFPI based on the percentile method, which avoids the influence caused by human factors and step complexity in the traditional method. The verification results show that this method effectively improves the accuracy of flood event monitoring. Description of the Drawings

[0046] Figure 1 is the technical roadmap.

[0047] Figure 2 are the runoff data and I time series and their corresponding thresholds.

[0048] Figure 3 are the comparison results of flood events determined based on runoff data and flood events obtained using different threshold determination methods. Detailed Implementation Manner

[0049] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0050] The main content of the present invention includes:

[0051] An MFPI threshold determination method that is theoretically more robust is invented, that is, a new method for directly and precisely determining the MFPI threshold based on the percentile method. The threshold determined by this method can significantly improve the monitoring accuracy of flood events.

[0052] The research objective of this project is to propose a new method for directly and precisely determining the MFPI threshold based on the percentile to improve the deficiencies of the existing threshold determination methods: ① Simplify the calculation steps: solve the problem that the steps of the existing method are too cumbersome and improve its applicability in the absence of runoff data; ② Eliminate the influence of reservoir regulation: avoid the influence of reservoir regulation on the threshold estimation accuracy. The new method not only overcomes the defect of being affected by human factors in the traditional threshold determination method, but also has a perfect theoretical support, thereby effectively improving the accuracy and reliability of flood event monitoring.

[0053] Please refer to Figure 1 , the present invention discloses a new method for directly and precisely determining the flood event threshold based on the percentile

[0054] First, calculate T from the GRACE RL06 spherical harmonic coefficients detrend , and read precipitation data;

[0055] Then, taking the Yellow River Basin as an example, calculate I according to T detrend and precipitation data;

[0056] Secondly, adopt the new threshold estimation method to determine the thresholds of I at different probability levels;

[0057] Finally, use the determined thresholds to identify flood events during the study period, and judge the accuracy of the thresholds for flood event identification through the introduced evaluation indicators. As Figure 2 shown, in order to analyze the accuracy of the new threshold determination method in flood event identification, the runoff and I time series and their corresponding thresholds are calculated. Figure 3 Sorted out the I exceeding the threshold in Figure 2 , and showed the flood events detected by each index under different threshold determination methods, and further compared the identification effects of different threshold determination methods. It can be seen from Figure 3 that the new method has a higher consistency with the flood events determined by the runoff data.

[0058] As shown in Table 2, to quantitatively analyze the accuracy of the new method, the evaluation indices of I under different threshold determination methods were calculated. In Table 2, I + represents the new threshold determination method proposed in this study, and I * represents the original threshold determination method. It can be seen from the table that the threshold of I * is lower than that of I + . Therefore, the original threshold determination method determines more events as flood events. Although determining more events as flood events increases the number of flood events identified by the method of I * (higher TPR), it also increases the probability of misidentifying flood events (higher FPR), resulting in a lower overall accuracy rate (ACC).

[0059] Table 2 Evaluation indices of I under different threshold determination methods

[0060]

[0061] Thus, it can be seen that adopting the threshold determination method proposed in this study can effectively reduce the probability of misjudging as flood events and improve the recognition accuracy of flood indices at the same time.

[0062] The following are some explanations of terms in this industry:

[0063] 1. Gravity Recovery and Climate Experiment (GRACE) satellite: A satellite project jointly developed by the National Aeronautics and Space Administration (NASA) of the United States and the German Aerospace Center. By precisely measuring the distance and acceleration changes between two satellites, it reveals the changes in the Earth's gravity field.

[0064] 2. Spherical harmonic coefficients: One of the main data products generated by the GRACE satellite mission, used to describe the changes in the Earth's external gravity field. These coefficients are retrieved from the observational data collected by the GRACE satellite and can reflect in detail the changes in the mass distribution on the Earth's surface.

[0065] 3. Terrestrial water storage: The sum of all forms of water stored on the Earth's land, including components such as ice and snow, soil water, groundwater, surface water, and vegetation water.

[0066] 4. North - south striping error: A common type of error in the inversion of the GRACE satellite gravity field. This error is mainly manifested as strip - shaped noise in the gravity field signal along the north - south direction of the Earth, seriously affecting the precise inversion and application of the gravity field.

[0067] 5. Smoothing filtering: averaging adjacent data points to reduce data volatility, thereby making the data smoother.

[0068] 6. Decorrelation filtering: reducing the striping effect by reducing the correlation between odd and even spherical harmonic coefficients, thereby improving the clarity and quality of the data.

[0069] 7. Leakage error: The signal attenuation and distortion caused by the filtering method are called leakage, and the error caused by leakage is called leakage error.

[0070] 8. Hydrological model: A mathematical model used to simulate and predict hydrological processes, widely applied in fields such as water resources management, flood forecasting, and environmental impact assessment.

[0071] 9. Terrestrial water storage anomaly: The terrestrial water storage data during the study period minus the average terrestrial water storage from 2004 to 2009.

[0072] 10. Scale factor method: A method for correcting leakage error. This method quantifies the leakage error of the signal by performing spherical harmonic expansion on the hydrological model and based on the least squares fitting between the water storage before and after filtering.

[0073] 11. RMS_NSE scale factor method: A leakage error correction method that weights the scale factors obtained for different hydrological models using the Nash-Sutcliffe Efficiency Coefficient (NSE) and the Root Mean Squared Error (RMS) to obtain a new scale factor.

[0074] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be encompassed within the present invention. Any reference signs in the claims should not be regarded as limiting the claims involved.

[0075] In addition, it should be understood that although this specification is described according to embodiments, not every embodiment only contains an independent technical solution. This narrative manner of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A new method for directly and accurately determining the threshold of flood events based on percentiles, characterized in that Specifically, it includes the following steps: a) Calculate the GRACE Terrestrial Water Storage Anomaly data TWSA using the spherical harmonic coefficients SHCs of the Earth's gravity field obtained by the Gravity Recovery and Climate Experiment GRACE satellite; b) Read precipitation data and construct the Modified Flood Potential Index (MFPI) using precipitation and detrended Terrestrial Water Storage Anomaly (TWSA) data detrend ; c) Based on the Modified Flood Potential Index MFPI data constructed in b), according to the newly proposed method for directly and precisely determining the MFPI threshold using percentiles, calculate the thresholds of the Modified Flood Potential Index MFPI at different probabilities; d) Use the thresholds obtained in c) to determine flood events and judge the accuracy of the threshold determination method for identifying flood events; The newly proposed method for directly and precisely determining the MFPI threshold using percentiles is as follows: Set the threshold P for directly determining the Modified Flood Potential Index MFPI at various probabilities based on the percentile method. Divide a set of data into two parts, including r% of the data being less than the threshold P and (100 - r)% of the data being greater than P. When the variable I of the Modified Flood Potential Index is greater than P, it is determined as a flood event. The threshold calculation steps are: n = 1 + (T - 1)r% (4) Where T is the length of the time series, and r% is the probability of possible flood occurrence, taking 85 th , 90 th and 95 th respectively; n represents the number of non-flood events under the probability of r%. P = X [n] +(X [n+1] -X [n] )(n - [n]) (5) In the formula, [] is the rounding function, P represents the threshold for directly determining the Modified Flood Potential Index MFPI at various probabilities based on the percentile method. Sort the I time series from smallest to largest to get X, and n represents the number of non - flood events at the r% probability.

2. A new method for directly and precisely determining the threshold of flood events based on percentiles according to claim 1, characterized in that, Step a includes the following sub - steps: a1. Use the product in the Technical Note 13 to replace the first-order term of the GRACE spherical harmonic coefficients, and use the product SLR RL06 released by satellite laser ranging to replace the C 20 term of the GRACE spherical harmonic coefficients; a2. Deduct the average value of the GRACE spherical harmonic coefficients from January 2004 to December 2009; a3. Use the SWENSON decorrelation filter to weaken the north - south striping error caused by the correlation of spherical harmonic coefficients, and use a 300 - km Gaussian filter to smooth the spherical harmonic coefficients to weaken the high - frequency error; a4. Remove the Glacial Isostatic Adjustment GIA effect; a5. Calculate the variable T of the GRACE Terrestrial Water Storage data using the GRACE RL06 spherical harmonic data; a6. Correct the leakage error of T using the RMS_NSE scale factor method.

3. A new method for directly and precisely determining the threshold of flood events based on percentiles according to claim 2, characterized in that, Step b includes the following sub - steps: b1. Read precipitation data and detrend the T data in a6 to obtain T detrend ; b2. Utilize the T obtained in b1 detrend Calculate the maximum precipitation ΔH that can be stored in the basin each month; b3. Calculate the variable A of the Modified Flood Potential using the precipitation data obtained in b1 and the ΔH data obtained in b2; b4. Standardize the A obtained in b3 to get the variable I of the Modified Flood Potential Index MFPI.

4. A new method for directly and accurately determining the threshold of flood events based on percentiles according to claim 3, characterized in that, Based on the percentile method, directly calculate the threshold of the Modified Flood Potential Index MFPI. Step c includes the following sub - steps: c1. Use the time - series length data T and the probability data r% of possible floods to calculate the number n of non - flood events at different probabilities according to formula (4); c2. Rearrange the I data obtained in b4 in ascending order to get the data X time series; c3. Based on the n calculated in c1 and the X time series obtained in c2, calculate the threshold P at different probabilities using formula (5).

5. A new method for directly and precisely determining the threshold of flood events based on percentiles according to claim 4, characterized in that, Step d includes the following sub - steps: d1. Compare the value of I with the threshold obtained in c3 one by one, and judge those greater than the threshold as flood events; d2. Use the flood events determined based on runoff data as a reference, and introduce evaluation indicators to evaluate the accuracy of the threshold determination method for identifying flood events.

Images