Precipitation product correction method combining dry and wet season division and weight distribution

Abstract

The invention provides a rainfall product correction method combining dry and wet season division and weight distribution based on climate adjustment, and the method comprises the steps: calculating an adjustment amount and a scaling factor according to the rainfall data of a region over the years, generating a deviation adjustment value through the adjustment amount AA and the scaling factor SF, applying the deviation adjustment value to a corresponding IMERG rainfall estimation value, and carrying out the correction of a rainfall product. The rainfall estimation value after adjustment and correction is obtained, so that the accuracy of rainfall estimation value correction can be improved, error correction is effectively performed on the satellite rainfall data, and the reliability of the satellite rainfall data is improved.

Description

Technical Field

[0001] The invention relates to the technical field of rainfall prediction, and in particular to a correction method for precipitation products combining dry and wet season division and weight distribution. Background Art

[0002] Traditional precipitation data acquisition mainly relies on ground observation stations, such as rain gauges and weather radars. However, ground observation stations have limited coverage, especially in remote or complex geographical areas, where observation data are scarce and incomplete.

[0003] The rapid development of satellite remote sensing technology has provided a new way to obtain precipitation data with large-scale and high spatial and temporal resolution. Through sensors carried on satellites, precipitation conditions around the world can be obtained in real time. For example, the Integrated Multi-Satellite Inversion for Global Precipitation Measurement (IMERG) product proposed by the Global Precipitation Measurement Project (GPM) jointly developed by the National Aeronautics and Space Administration (NASA) and the Japan Aerospace Exploration Agency (JAXA) is recognized to have good performance. The latest version of IMERGV06B has extended its time coverage to 2000, and now has access to 20-year data sets from 2000 to the present, thus providing the possibility of longer-term precipitation feature assessment. However, satellite observation data still have large systematic errors and random errors due to the limitations of observation principles and data processing methods. For example, satellite-observed precipitation data may be affected by factors such as cloud thickness, surface reflectivity, and sensor performance, and there is a difference with the actual ground precipitation conditions.

[0004] The accuracy of IMERG estimated precipitation data limits its application and development. Therefore, obtaining a more accurate precipitation data set through error correction is of great significance to the quality improvement of satellite precipitation products. The currently widely used statistical error correction method for satellite rainfall data is mainly the quantile mapping method. The quantile mapping method effectively reduces the systematic deviation by adjusting the probability distribution of the data to make it consistent with the probability distribution of the reference data. However, the error correction accuracy of these methods for satellite rainfall data is still low, and the overall effect still has a lot of room for improvement. In response to these problems, the present invention proposes a precipitation product correction method that combines dry and wet season division and weight distribution, aiming to further improve the accuracy and reliability of satellite observation precipitation data, so as to better serve the fields of meteorological forecasting, climate research and water resources management. Summary of the invention

[0005] In order to solve the technical problems raised in the above-mentioned background technology, the present invention provides a precipitation product correction method based on climate adjustment combined with dry and wet season division and weight distribution, which can improve the accuracy of correction of precipitation estimation values, effectively correct errors in satellite precipitation data, and improve the reliability of satellite precipitation data.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] A correction method for precipitation products combining dry and wet season division and weight distribution includes the following steps:

[0008] S1, historical data preprocessing: obtain the historical measured precipitation data of ground meteorological stations and the corresponding IMERG historical precipitation estimation data, divide the obtained historical measured precipitation data and the corresponding IMERG historical precipitation estimation data according to the dry and wet seasons, and obtain the dry season historical data set and the wet season historical data set;

[0009] S2, calculation of adjustment amount AA: respectively calculating the difference between each IMERG historical precipitation estimation data and the corresponding historical measured precipitation data in the historical data set of the dry season and the historical data set of the wet season, and calculating the probability density function of the difference, and then using the probability density function as a weight to assign to the corresponding difference, so as to respectively calculate the adjustment amount AA corresponding to the historical data set of the dry season and the adjustment amount AA corresponding to the historical data set of the wet season;

[0010] S3, calculation of scaling factor SF: setting several value intervals to classify the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set, respectively, and calculating the scaling factor SF according to the deviation type of the IMERG historical precipitation estimation data in each value interval;

[0011] S4. Correction of the IMERG precipitation estimate during the correction period: The correction method for the IMERG precipitation estimate in the rainfall area during the correction period is as follows: a deviation adjustment value is generated by using the adjustment amount AA corresponding to the season in which the correction period is located and the scaling factor SF corresponding to the value interval in which the IMERG precipitation estimate is located, and the generated deviation adjustment value is applied to the IMERG precipitation estimate to obtain an adjusted and corrected precipitation estimate.

[0012] Furthermore, the difference is calculated according to the following formula:

[0013] The IMERG historical precipitation estimate data is denoted as x IMERG , the historical measured precipitation data of the ground meteorological station is recorded as x ground , then the difference D is:

[0014] Bias = x IMERG -x ground (1)

[0015] D=|Bias|=|x IMERG -x ground | (2)

[0016] Where Bias is the deviation between the historical precipitation estimated data of IMERG and the historical measured precipitation data.

[0017] Furthermore, the historical data preprocessing in step S1 specifically includes the following steps:

[0018] S11, determining a rainfall area to be studied, and determining a historical time period for performing precipitation estimation correction in the rainfall area;

[0019] S12, obtaining the historical measured precipitation data of the ground meteorological stations in the rainfall area every day in the historical time period and the corresponding IMERG historical precipitation estimation data, and arranging the historical measured precipitation data of the ground meteorological stations in the historical time period and the corresponding IMERG historical precipitation estimation data in chronological order to obtain a total historical data set;

[0020] S13, dividing the historical measured precipitation data in the total historical data set and the corresponding IMERG historical precipitation estimation data according to the dry and wet seasons to obtain a dry season historical data set and a wet season historical data set.

[0021] Furthermore, the calculation of the adjustment amount in step S2 specifically includes the following steps:

[0022] S21, during the historical time period, respectively calculating the difference between the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set and the historical measured precipitation data of the corresponding ground meteorological station, and arranging the calculated differences in order from the smallest to the largest to obtain a difference sequence;

[0023] S22, calculating the probability density function of each difference in the difference sequence;

[0024] S23, assigning a weight to each difference, wherein for each difference, the assigned weight is a corresponding probability density function;

[0025] S24, obtaining the adjustment amount AA by adding the product of each difference and its corresponding weight.

[0026] Furthermore, the weight ω i And the calculation formula of the adjustment amount AA is as follows:

[0027] ω i =PDF i (3)

[0028]

[0029] Where i is a number from 1 to n, n is the total number of differences in the difference sequence, ω irepresents the weight of the i-th difference in the difference sequence, PDF i Represents the probability density function of the i-th difference in the difference sequence.

[0030] Furthermore, the calculation of the scaling factor SF in step S3 specifically includes the following steps:

[0031] S31, according to the size of the IMERG historical precipitation estimation data, setting a number of value intervals to classify the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set respectively;

[0032] S32, for all IMERG historical precipitation estimation data in each value interval, comparing the IMERG historical precipitation estimation data with the corresponding historical measured precipitation data to determine whether the deviation type of the IMERG historical precipitation estimation data is underestimation;

[0033] S33, calculating the percentage of the underestimated deviation type in each value interval according to the number of underestimated deviation types in each value interval;

[0034] S34, calculating a scaling factor SF of a corresponding value interval according to the percentage of the underestimated deviation type.

[0035] Furthermore, the calculation formula of the scaling factor SF is as follows:

[0036] SF=a-(1-a) (5)

[0037] Where a is the percentage of underestimated bias type within a certain value interval; the value range of SF is -1 to 1.

[0038] Furthermore, the percentage of the underestimated bias type is calculated as:

[0039]

[0040] Where n is the number of events of the underestimated bias type within a certain value interval; N is the number of events of all types within the interval.

[0041] Furthermore, when correcting a certain IMERG precipitation estimate, the corresponding adjustment amount AA and the scaling factor SF of the value interval of the IMERG precipitation estimate are associated according to the following formula to obtain the deviation adjustment value BA required for correction:

[0042] BA=AA×ST (7)

[0043] Furthermore, in step S4,

[0044] When the IMERG precipitation estimate is <1 mm, BA is not applied to the IMERG estimate;

[0045] When the IMERG precipitation estimate is ≥1 mm and the deviation adjustment value is less than the IMERG precipitation estimate, adding the calculated corresponding deviation adjustment value to the IMERG precipitation estimate to obtain a corrected precipitation estimate;

[0046] When the IMERG precipitation estimate is ≥1 mm and the bias adjustment value is greater than the IMERG precipitation estimate, the IMERG precipitation estimate is corrected to 0.1 mm, as shown in formula (8):

[0047]

[0048] In the formula, x originalIMERG and x adjustedIMERG denote the raw and adjusted IMERG precipitation estimates during the correction period, respectively.

[0049] Due to the adoption of the above technical solution, the present invention has the following beneficial effects:

[0050] The above-mentioned precipitation product correction method based on climate adjustment combined with dry and wet season division and weight allocation calculates the adjustment amount and scaling factor according to the regional precipitation data over the years, generates a deviation adjustment value through the adjustment amount AA and the scaling factor SF, and applies the deviation adjustment value to the corresponding IMERG precipitation estimate to obtain the adjusted and corrected precipitation estimate, thereby improving the accuracy of the correction of the precipitation estimate, effectively correcting the error of the satellite precipitation data, and improving the reliability of the satellite precipitation data; at the same time, since the satellite precipitation deviation distribution has a certain sensitivity to time and period, the deviation amplitudes in the dry season and the wet season are different, and the occurrence frequency of different deviation values ​​is also very different, the above-mentioned precipitation product correction method based on climate adjustment combined with dry and wet season division and weight allocation divides the obtained measured precipitation data and the corresponding IMERG precipitation estimation data according to the dry and wet seasons, which is conducive to obtaining the deviation distribution characteristics and the accuracy of the deviation correction is higher. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is an overall flow chart of a precipitation product correction method combining dry and wet season division and weight distribution proposed by the present invention;

[0052] Figure 2 is the RB and RMSE box plot between the original IMERG estimates and the adjusted IMERG estimates;

[0053] Figure 3 It is a comparison chart of the effectiveness of the precipitation product correction method combining dry and wet season division and weight distribution proposed in the present invention and the traditional quantile mapping method in reducing RB and RMSE in each climate and terrain partition. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0056] See also Figure 1 A preferred embodiment of the present invention provides a correction method for precipitation products combining dry and wet season division and weight distribution, comprising the following steps:

[0057] S1, historical data preprocessing: obtain the historical measured precipitation data of ground meteorological stations and the corresponding IMERG historical precipitation estimation data, divide the obtained historical measured precipitation data and the corresponding IMERG historical precipitation estimation data according to the dry and wet seasons, and obtain the dry season historical data set and the wet season historical data set.

[0058] In this embodiment, the historical data preprocessing in step S1 specifically includes the following steps:

[0059] S11, determine the rainfall area to be studied, and determine the historical time period for the rainfall area to perform precipitation estimation correction. The rainfall data in the historical time period is used to analyze the deviation distribution characteristics of the measured precipitation data of the ground meteorological station in the rainfall area and the IMERG precipitation estimation data, which should be more than 15 years to ensure a relatively stable deviation distribution.

[0060] S12, obtaining the historical measured precipitation data of the ground meteorological stations in the rainfall area every day during the historical time period and the corresponding IMERG historical precipitation estimation data, and arranging the historical measured precipitation data of the ground meteorological stations in the historical time period and the corresponding IMERG historical precipitation estimation data in chronological order to obtain a total historical data set.

[0061] S13, dividing the historical measured precipitation data in the total historical data set and the corresponding IMERG historical precipitation estimation data according to the dry and wet seasons to obtain a dry season historical data set and a wet season historical data set.

[0062] Specifically, a year can be divided into two seasons, dry and wet, according to the seasonal characteristics of rainfall in the rainfall area. Taking Guangxi Zhuang Autonomous Region as an example, April to September is usually the wet season, and October to March of the following year is the dry season. The data in the total historical data set is divided into dry and wet seasons according to time, and the historical data sets for the dry season and the wet season are obtained and entered into subsequent calculations respectively.

[0063] S2, calculation of adjustment amount AA: respectively calculate the difference between each IMERG historical precipitation estimation data and the corresponding historical measured precipitation data in the historical data set of the dry season and the historical data set of the wet season, and calculate the probability density function of the difference, and then use the probability density function as a weight to assign to the corresponding difference, so as to respectively calculate the adjustment amount AA corresponding to the historical data set of the dry season and the adjustment amount AA corresponding to the historical data set of the wet season.

[0064] In this embodiment, the basic principle of calculating the adjustment amount AA is that the higher the frequency of a certain difference between historical time periods, the higher the weight of the difference in the adjustment. Specifically, the calculation of the adjustment amount AA in step S2 specifically includes the following steps:

[0065] S21, during the historical time period, respectively calculate the difference between the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set and the historical measured precipitation data of the corresponding ground meteorological station, and arrange the calculated differences in order from small to large to obtain a difference sequence. Specifically, the difference is calculated according to the following formula:

[0066] The IMERG historical precipitation estimate data is denoted as x IMERG , the historical measured precipitation data of the ground meteorological station is recorded as x ground , then the difference D is:

[0067] Bias = x IMERG -x ground (1)

[0068] D=|Bias|=|x IMERG -x ground | (2)

[0069] Where Bias is the deviation between the historical precipitation estimated data of IMERG and the historical measured precipitation data.

[0070] S22, calculating the probability density function of each difference in the difference sequence.

[0071] S23, assigning a weight to each difference, and for each difference, the assigned weight is the corresponding probability density function.

[0072] S24, obtaining the adjustment amount AA by adding the product of each difference and its corresponding weight. Specifically, the weight ω i And the calculation formula of the adjustment amount AA is as follows:

[0073] ω i =PDF i (3)

[0074]

[0075] Where i is a number from 1 to n, n is the total number of differences in the difference sequence, ω i represents the weight of the i-th difference in the difference sequence, PDF i represents the probability density function of the ith difference in the difference sequence. The calculation of the probability density function belongs to the prior art and will not be described here for the sake of space omission.

[0076] S3, calculation of scaling factor SF: several value intervals are set to classify the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set respectively, and the scaling factor SF is calculated according to the deviation type of the IMERG historical precipitation estimation data in each value interval. Specifically, the calculation of scaling factor SF in step S3 includes the following steps:

[0077] S31, according to the size of the IMERG historical precipitation estimation data, several value intervals are set to classify the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set respectively.

[0078] Specifically, in this embodiment, five value intervals are proposed according to the size of the IMERG estimated value of the entire mainland of China, and the critical values ​​for dividing the intervals are 5, 10, 25 and 50 mm, respectively, constituting five value intervals of rainfall between 0-5 mm, 6-10 mm, 11-25 mm, 26-50 mm and above 50 mm. Different estimation intervals represent different rainfall intensities estimated by the IMERG product. For example, the meteorological department usually uses 10, 25 and 50 mm rainfall as critical values ​​for dividing light, medium, heavy and torrential rain. Subsequently, the IMERG historical precipitation estimation data in the historical data set of the dry season and the historical data set of the wet season are classified according to these five value intervals. For example, if an IMERG historical precipitation estimation data in the historical data set of several seasons is 4 mm, it is divided into the value interval of 0-5 mm.

[0079] S32, for all IMERG historical precipitation estimation data in each value interval, by comparing the IMERG historical precipitation estimation data with the corresponding historical measured precipitation data, it is determined whether the deviation type of the IMERG historical precipitation estimation data is underestimated. Specifically, if the value of the IMERG historical precipitation estimation data is less than the value of the corresponding historical measured precipitation data, it is determined to be an underestimated deviation type event, otherwise it belongs to other deviation type events.

[0080] S33, calculating the percentage of the underestimated deviation type in each value interval according to the number of underestimated deviation types in each value interval, and the percentage can be regarded as the frequency of the underestimated event.

[0081] S34, calculating a scaling factor SF of a corresponding value interval according to the percentage of the underestimated deviation type. Specifically, the calculation formula of the scaling factor SF is as follows:

[0082] SF=a-(1-a) (5)

[0083] Where a is the percentage of underestimated bias type within a certain value interval; the value range of SF is -1 to 1.

[0084] When SF is equal to -1 and 1, it means that the bias type of all IMERG estimates in the value interval is overestimation or underestimation, respectively. In the rainfall interval, if SF<0, it means that the bias adjustment has reduced the precipitation estimate of IMERG during the correction period, and if SF>0, it means that the bias adjustment has increased the precipitation estimate of IMERG during the correction period. It can be seen that the greater the proportion of overestimation events or underestimated events, the greater the reduction or increase in the precipitation estimate of IMERG.

[0085] Specifically, the percentage of underestimated deviation types is calculated as:

[0086]

[0087] Where n is the number of events of the underestimated bias type within a certain value interval; N is the number of events of all types within the interval, that is, the total number of IMERG historical precipitation estimation data or the total number of historical measured rainfall data.

[0088] S4. Correction of the IMERG precipitation estimate during the correction period: The correction method for the IMERG precipitation estimate in the rainfall area during the correction period is as follows: a deviation adjustment value is generated by using the adjustment amount AA corresponding to the season in which the correction period is located and the scaling factor SF corresponding to the value interval in which the IMERG precipitation estimate is located, and the generated deviation adjustment value is applied to the IMERG precipitation estimate to obtain an adjusted and corrected precipitation estimate.

[0089] Specifically, when correcting a certain IMERG precipitation estimate, first, determine whether the correction period of the IMERG precipitation estimate is the dry season or the wet season. Secondly, obtain the value interval of the IMERG precipitation estimate. Finally, associate the adjustment amount AA of the corresponding season with the scaling factor SF of the value interval of the IMERG precipitation estimate according to the following formula to obtain the deviation adjustment value BA required for correction:

[0090] BA=AA×SD (7)

[0091] After the deviation adjustment value BA is calculated, the deviation adjustment value BA is applied to the corresponding IMERG precipitation estimation value to obtain an adjusted and corrected precipitation estimation value, which is specifically:

[0092] When the IMERG precipitation estimate is <1 mm, the ground conditions usually correspond to very little or no precipitation, therefore BA is not applied to the IMERG estimate when the IMERG precipitation estimate is <1 mm.

[0093] When the IMERG precipitation estimate is ≥1 mm and the deviation adjustment value is less than the IMERG precipitation estimate, adding the calculated corresponding deviation adjustment value to the IMERG precipitation estimate to obtain a corrected precipitation estimate;

[0094] When the IMERG precipitation estimate is ≥1 mm and the bias adjustment value is greater than the IMERG precipitation estimate, the IMERG precipitation estimate is corrected to 0.1 mm, as shown in formula (8):

[0095]

[0096] In the formula, x originalIMERG and x adjustedIMERG denote the raw and adjusted IMERG precipitation estimates during the correction period, respectively.

[0097] If multiple IMERG precipitation estimates within the calibration period need to be calibrated, after the calibration of each IMERG precipitation estimate is completed, the IMERG precipitation estimates adjusted according to the dry and wet seasons are reintegrated and merged in chronological order to obtain the complete IMERG estimate calibration results within the calibration period.

[0098] The following is a specific embodiment to verify the effectiveness and reliability of this method through example tests.

[0099] The Chinese mainland is used as the test area. The Chinese mainland is located in East Asia and the west coast of the Pacific Ocean. The longitude ranges from 73°41′E to 135°02′E, and the latitude ranges from 18°10′N to 53°33′N. According to the climate characteristics, the Chinese mainland is divided into several climate zones. The specific division standards are shown in Table 1. In addition, the Chinese mainland is divided into 7 terrain regions according to the definition of different terrain regions based on altitude. The specific division standards are shown in Table 2.

[0100] During the experiment, the measured precipitation data of the selected ground meteorological stations came from the China National Meteorological Center, and the satellite precipitation data came from three sub-products of the IMERGV06 version: the near-real-time products IMERG-E and IMERG-L, and the post-real-time product IMERG-F. The historical time period was from January 1, 2000 to December 31, 2017, and the verification period (i.e., the correction period) was from January 1, 2018 to December 31, 2019. In order to intuitively demonstrate the effect of the present invention, the quantile mapping method, a mainstream method with a higher degree of recognition, was used as a reference to compare the correction effects of the present invention and the quantile mapping method on the IMERG precipitation data. In the following description, the quantile mapping method is referred to as the QM (Quantile mapping) method, and the present invention is referred to as the DW (Dry-wet season division and weight allocation) method.

[0101] In order to reflect the effectiveness of the present invention in correcting the satellite precipitation estimation value, several typical statistical indicators are selected for measurement: CC (correlation coefficient), RB (relative deviation), and RMSE (root mean square error), as shown in Table 3.

[0102] Figure 2 is a box plot of RB and RMSE of the original estimate and the corresponding adjusted estimate. For the QM method, the RB and RMSE values ​​of the 75th, 50th, and 25th percentiles of the adjusted IMERG estimate are higher than those of the original estimate. For the DW method, the RB and RMSE results are opposite to those of the QM method above. Overall, the QM method performs poorly, while the DW method plays an effective role in reducing RB and RMSE.

[0103] Figure 3The effectiveness of both methods in reducing RB and RMSE in each climate and terrain subregion is described. The effectiveness of the DW method is higher than or equal to that of the QM method in all regions. In terms of reducing RB, the DW method maintains high performance in regions 1, N, and NE, while the QM method maintains high performance in regions 6, 7, and QT. Regions S and N correspond to the climate regions with the worst performance under the DW and QM methods, respectively. For reducing RMSE, the DW method performs better than other regions in regions NW and N, while the QM method performs best in region QT. In all terrain regions, the highest performance of the QM method is generally found in regions 6 and 7.

[0104] In general, the DW method of the present invention is significantly more effective than the QM method used for comparison in increasing CC, reducing RB and RMSE values. The DW method is significantly better than the QM method in almost all indicators, as shown in Table 3.

[0105] Table 1 Regions in mainland China divided according to climate conditions

[0106]

[0107] Table 2 Topographic regions in mainland China divided by elevation

[0108] Terrain division Terrain Type Altitude (E) / m 1 Plains E≤100 2 Hills or plains 100<E≤250 3 hills 250<E≤500 4 Mountains or plateaus 500<E≤1000 5 Mountains or plateaus 1000<E≤2000 6 Mountains or plateaus 2000<E≤3000 7 plateau 3000<E

[0109] Table 3. Effectiveness of two error adjustment methods in increasing CC value and reducing RB and RMSE values

[0110]

[0111] From the above, it can be seen that the precipitation product correction method based on climate adjustment combined with dry and wet season division and weight distribution proposed in the embodiment of the present invention has a significant improvement in correction accuracy compared to the traditional frequency correction method quantile mapping method, and can effectively correct satellite precipitation data and improve the accuracy and applicability of near-real-time satellite precipitation products.

[0112] The above-mentioned precipitation product correction method based on climate adjustment combined with dry and wet season division and weight allocation calculates the probability distribution from a large sample to capture the deviation distribution characteristics, uses the probability density function to confirm the weight of each deviation value in the deviation adjustment, calculates the value of the adjustment amount AA, and then divides the IMERG precipitation estimation data into several value intervals, and calculates the value of the scaling factor SF according to the deviation type of the IMERG precipitation estimation data in each value interval. Finally, the adjustment amount and the scaling factor are applied to the initial IMERG precipitation estimation data to obtain the corrected satellite precipitation data, which can improve the accuracy of the correction of the precipitation estimation value, effectively correct the error of the satellite precipitation data, and improve the reliability of the satellite precipitation data; at the same time, since the satellite precipitation deviation distribution has a certain sensitivity to time and period, the deviation amplitudes of the dry season and the wet season are different, and the occurrence frequency of different deviation values ​​is also very different. The above-mentioned precipitation product correction method based on climate adjustment combined with dry and wet season division and weight allocation divides the obtained measured precipitation data and the corresponding IMERG precipitation estimation data according to the dry and wet seasons, which is conducive to obtaining the deviation distribution characteristics, and the accuracy of the deviation correction is higher.

[0113] The above description is a detailed description of the preferred feasible embodiments of the present invention, but the embodiments are not intended to limit the scope of the patent application of the present invention. All equivalent changes or modified changes completed under the technical spirit suggested by the present invention should fall within the patent scope covered by the present invention.

Claims

1. A correction method for precipitation products combining dry and wet season division and weight distribution, characterized in that: The following steps are involved: S1, historical data preprocessing: obtain the historical measured precipitation data of ground meteorological stations and the corresponding IMERG historical precipitation estimation data, divide the obtained historical measured precipitation data and the corresponding IMERG historical precipitation estimation data according to the dry and wet seasons, and obtain the dry season historical data set and the wet season historical data set; S2, calculation of adjustment amount AA: respectively calculating the difference between each IMERG historical precipitation estimation data and the corresponding historical measured precipitation data in the historical data set of the dry season and the historical data set of the wet season, and calculating the probability density function of the difference, and then using the probability density function as a weight to assign to the corresponding difference, so as to respectively calculate the adjustment amount AA corresponding to the historical data set of the dry season and the adjustment amount AA corresponding to the historical data set of the wet season; S3, calculation of scaling factor SF: setting several value intervals to classify the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set, respectively, and calculating the scaling factor SF according to the deviation type of the IMERG historical precipitation estimation data in each value interval; S4. Correction of the IMERG precipitation estimate during the correction period: The correction method for the IMERG precipitation estimate in the rainfall area during the correction period is as follows: a deviation adjustment value is generated by using the adjustment amount AA corresponding to the season in which the correction period is located and the scaling factor SF corresponding to the value interval in which the IMERG precipitation estimate is located, and the generated deviation adjustment value is applied to the IMERG precipitation estimate to obtain an adjusted and corrected precipitation estimate.

2. The correction method for precipitation products combining dry and wet season division and weight distribution as claimed in claim 1, characterized in that: The difference is calculated according to the following formula: The IMERG historical precipitation estimate data is denoted as X IMERG , the historical measured precipitation data of the ground meteorological station is recorded as x ground , then the difference D is: Bias=x IMERG -x ground (1) D=|Bias|=|x IMERG -x ground | (2) Where Bias is the deviation between the historical precipitation estimated data of IMERG and the historical measured precipitation data.

3. The correction method for precipitation products combining dry and wet season division and weight distribution as claimed in claim 1, characterized in that: The historical data preprocessing in step S1 specifically includes the following steps: S11, determining a rainfall area to be studied, and determining a historical time period for performing precipitation estimation correction in the rainfall area; S12, obtaining the historical measured precipitation data of the ground meteorological stations in the rainfall area every day in the historical time period and the corresponding IMERG historical precipitation estimation data, and arranging the historical measured precipitation data of the ground meteorological stations in the historical time period and the corresponding IMERG historical precipitation estimation data in chronological order to obtain a total historical data set; S13, dividing the historical measured precipitation data in the total historical data set and the corresponding IMERG historical precipitation estimation data according to the dry and wet seasons to obtain a dry season historical data set and a wet season historical data set.

4. The correction method for precipitation products combining dry and wet season division and weight distribution as claimed in claim 3, characterized in that: The calculation of the adjustment amount in step S2 specifically includes the following steps: S21, during the historical time period, respectively calculating the difference between the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set and the historical measured precipitation data of the corresponding ground meteorological station, and arranging the calculated differences in order from the smallest to the largest to obtain a difference sequence; S22, calculating the probability density function of each difference in the difference sequence; S23, assigning a weight to each difference, wherein for each difference, the assigned weight is a corresponding probability density function; S24, obtaining the adjustment amount AA by adding the product of each difference and its corresponding weight.

5. The correction method of precipitation product combining dry and wet season division and weight distribution as claimed in claim 4, characterized in that: The weight ω i And the calculation formula of the adjustment amount AA is as follows: oh i =PDF i (3) Where i is a number from 1 to n, n is the total number of differences in the difference sequence, ω i represents the weight of the i-th difference in the difference sequence, PDF i Represents the probability density function of the i-th difference in the difference sequence.

6. The correction method of precipitation product combining dry and wet season division and weight distribution as claimed in claim 3, characterized in that: The calculation of the scaling factor SF in step S3 specifically includes the following steps: S31, according to the size of the IMERG historical precipitation estimation data, setting a number of value intervals to classify the IMERG historical precipitation estimation data in the dry season historical data set and the wet season historical data set respectively; S32, for all IMERG historical precipitation estimation data in each value interval, comparing the IMERG historical precipitation estimation data with the corresponding historical measured precipitation data to determine whether the deviation type of the IMERG historical precipitation estimation data is underestimation; S33, calculating the percentage of the underestimated deviation type in each value interval according to the number of underestimated deviation types in each value interval; S34, calculating a scaling factor SF of a corresponding value interval according to the percentage of the underestimated deviation type.

7. The correction method of precipitation product combining dry and wet season division and weight distribution as claimed in claim 6, characterized in that: The calculation formula of the scaling factor SF is as follows: SF=a-(1-a) (5) Where a is the percentage of underestimated bias type within a certain value interval; the value range of SF is -1 to 1.

8. The correction method for precipitation products combining dry and wet season division and weight distribution as claimed in claim 7, characterized in that: The percentage of underestimation bias type is calculated as: Where n is the number of events of the underestimated bias type within a certain value interval; N is the number of events of all types within the interval.

9. The correction method for precipitation products combining dry and wet season division and weight distribution as claimed in claim 1, characterized in that: When correcting a certain IMERG precipitation estimate, the corresponding adjustment amount AA and the scaling factor SF of the value interval of the IMERG precipitation estimate are related according to the following formula to obtain the bias adjustment value BA required for correction: BA=AA×SF (7).

10. The correction method of precipitation product combining dry and wet season division and weight distribution as claimed in claim 1, characterized in that: In the step S4, When the IMERG precipitation estimate is <1 mm, BA is not applied to the IMERG estimate; When the IMERG precipitation estimate is ≥1 mm and the deviation adjustment value is less than the IMERG precipitation estimate, adding the calculated corresponding deviation adjustment value to the IMERG precipitation estimate to obtain a corrected precipitation estimate; When the IMERG precipitation estimate is ≥1 mm and the bias adjustment value is greater than the IMERG precipitation estimate, the IMERG precipitation estimate is corrected to 0.1 mm, as shown in formula (8): In the formula, x originalIMERG and x adjustedIMERG denote the raw and adjusted IMERG precipitation estimates during the correction period, respectively.

Images