KR101463493B1 - Supplementation method for global climate model using stochastic typhoon simulation - Google Patents

Abstract

A method of supplementing a global climate model using a stochastic hurdle simulation according to an embodiment of the present invention includes constructing rainfall data for a storm simulation for a past period; Reviewing rainfall for storm events affecting the watershed; Analyzing the statistical characteristics of typhoon events; Applying a Poisson distribution to simulate the frequency of occurrence of a typhoon in the forecast period; Applying the Campbell distribution to simulate the occurrence of typhoons in the forecast period; And finding the number of occurrences of the typhoon and the amount of generated rainfall, thereby overcoming the inability to reflect the extreme rainfall such as typhoon in the existing global climate model (GCM).

Description

A supplementary method for a global climate model using stochastic typhoon simulations.

The present invention relates to a process for applying downscaling to obtain high resolution climatic or climate change information from hydrological values of large scale atmospheric model grid scales such as Global Climate Model (GCM) And to overcome the disadvantages of not simulating the global climate model using a stochastic typhoon simulation.

The Global Climate Model (GCM) is being used as the basis for many studies for climate change research. Global climate models are based on the assumption that greenhouse gases in natural atmospheres are converted to equivalents for carbon dioxide, and the climate factors (temperature, precipitation, and temperature) are calculated considering the trends of greenhouse gas increase for 100 years from 2001 to 2100, Humidity, wind, etc.) as a numerical simulator of future climate information, which can be used as a basis for preparing countermeasures for minimizing the impact of climate change impacts on how and where to look.

However, hydrological components such as rainfall, soil moisture, evaporation, and basal runoff associated with river runoff may have significant variability on a sub-grid scale than on large-scale atmospheric model lattices such as global climate models (GCM) many. Therefore, in order to perform a reliable hydrological prediction based on the GCM simulation, it is necessary to reproduce the fluctuation component on the sub-grid scale through appropriate downscaling process.

In climate change studies, scale detailing refers to the technique of obtaining high resolution climate or climate change information from the relatively rough GCM grid scale hydrological values, and several preconditions must be made for scale detailing to be meaningful.

First, the impact model should be a situation that requires a much more detailed scale information than the scale of the GCM model. Second, the calculation load of the detailed model should be small so that it is advantageous for iterative calculation. Third, the optimal specification technique should be selected considering the original scale and the target scale.

There are statistical and mechanistic refinement techniques for refinement. The mechanistic refinement technique is a technique to exploit the physical relationship between the simulated values of the GCM and the scale characteristics of the meso- or micro-scale between the observations in the watershed. It is a large grid of GCM (100 to 300 km) And a detailed analysis of the area is carried out. However, computation, large-scale demand data, calibration of models, and so on, are often the parameters used in the global climate model, which are determined by local experiments and are used to carry out global simulations. They are also suspicious of reliability in simulating values. Moreover, due to the lack of understanding of the mechanical process, the need for parametrization of sub-grid units, and the lack of accurate information on initial conditions or boundary conditions, the present physical-based atmospheric model is not sufficient for the statistical properties of the observed rainfall There is a problem in simulating.

The statistical refinement technique has the advantage of analyzing the uncertainty through the low calculation load and the hydrologic scenario ensemble production by simulating the future view by estimating the correlation between the GCM climate variables from the past observation data. Recently, a nonlinear fit model using artificial intelligence techniques such as artificial neural network (ANN) and fuzzy theory has been actively applied, and a hybrid model with a probability model up to ± 25 km, model is being attempted.

As mentioned above, hydrological components such as rainfall, soil moisture, evaporation, and basin runoff associated with river runoff often have large variability in scale at small watershed scale in large-scale atmospheric model lattices such as GCM . Therefore, in order to perform reliable hydrological forecast based on GCM simulations, it is necessary to utilize it through appropriate downscaling process.

However, GCM itself does not simulate extreme rainfall such as typhoon, so it shows limitations in detailing.

The present invention provides a supplementary method of a global climate model applying a stochastic typhoon simulation capable of generating a hurricane thought reflecting the statistical characteristics of typhoon affecting the Korean Peninsula.

According to another aspect of the present invention, there is provided a method of supplementing a global climate model using stochastic hurdle simulation, comprising: constructing rainfall data for a storm simulation for a past period; Reviewing rainfall for storm events affecting the watershed; Analyzing the statistical characteristics of typhoon events; Applying a Poisson distribution to simulate the frequency of occurrence of a typhoon in the forecast period; Applying the Campbell distribution to simulate the occurrence of typhoons in the forecast period; And obtaining the number of times of occurrence of the typhoon and the amount of generated rainfall.

By configuring as described above, it is possible to perform the local refinement process by solving the conventional global climate model (GCM) which does not reflect extreme rainfall such as a typhoon.

In analyzing the statistical characteristics of the typhoon event, the monthly average typhoon frequency, average rainfall or maximum rainfall can be analyzed.

In the application of the Poisson distribution, the number of occurrence of the typhoon in the prediction period is applied to Poisson distribution which describes the number of events occurring within a specific range, and the formula for the Poisson distribution is

Where λ is the average number of hurricanes and β is the scale factor.

In the step of applying the Gumbel distribution, the formula for the < RTI ID = 0.0 >

이고,

ego,

, 위치변수

이며,Here,

, Positional variables

Lt;

는 과거 태풍 발생강우량에 대한 평균, σ는 과거 태풍 발생강우량에 대한 표준편차이다.

Is the mean of past typhoon rainfall, and σ is the standard deviation of past typhoon rainfall.

In the step of obtaining the number of occurrences of the typhoon and the amount of generated rainfall, it is possible to predict or estimate the typhoon rainfall according to past statistical characteristics by multiplying the occurrence frequency of the typhoon with the occurrence rainfall of the typhoon.

In the step of obtaining the number of occurrences of the typhoon and the amount of generated rainfall, the past statistical characteristics may include the monthly average frequency of the typhoon, the maximum typhoon rainfall, the minimum typhoon rainfall amount, or the average typhoon rainfall amount.

The step of calculating the number of occurrences of the typhoon and the amount of generated rainfall can calculate the extreme rainfall event.

The step of applying the Poisson distribution to simulate the number of times of occurrence of the typhoon in the prediction period and the step of applying the < RTI ID = 0.0 > Campbell < / RTI > distribution to simulate the amount of rainfall of the typhoon in the prediction period may be mixed.

The complementary method of the global climate model using the stochastic typhoon simulations according to the present invention can obtain accurate hydrological components even if the scale specification is performed on a small watershed scale because the extreme rainfall such as a typhoon can be simulated.

The complementary method of the global climate model using the stochastic typhoon simulation according to the present invention can generate the typhoon rainfall which exists within the range where the statistical characteristic and the large error do not occur in the past.

1 is a flowchart illustrating a method of supplementing a global climate model using stochastic typhoon simulation according to an embodiment of the present invention.
FIGS. 2 to 4 are diagrams illustrating programmed screens for performing a method of supplementing a global climate model according to an embodiment of the present invention.
FIG. 5 is a diagram showing an example of a result of applying a supplementary method of a global climate model applying stochastic typhoon simulation according to the present invention.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to or limited by the embodiments. Like reference symbols in the drawings denote like elements.

FIG. 1 is a flow chart for explaining a method of supplementing a global climate model applying stochastic hurricane simulation according to an embodiment of the present invention. FIGS. 2 to 4 illustrate a method of supplementing a global climate model according to an embodiment of the present invention FIG. 5 is a diagram showing an example of a result of applying a supplementary method of a global climate model applying stochastic typhoon simulation according to the present invention.

The complementary method of the global climate model applying the stochastic hurdle simulation according to an embodiment of the present invention is about localization of the statistical refinement techniques and it is possible to map or simulate extreme rainfall such as a typhoon on a small watershed scale such as the Korean Peninsula To a complementary method.

Referring to FIG. 1, a supplementary method of a global climate model applying a stochastic hurricane simulation according to an embodiment of the present invention includes a step 1100 of constructing rainfall data for simulation of a hurricane for a past period (1100) A step 1300 of analyzing the statistical characteristics of the typhoon event, a step 1400 of applying the Poisson distribution to simulate the occurrence frequency of the typhoon in the forecast period, A step 1500 of applying the Campbell distribution to simulate the generated rainfall of the typhoon of the typhoon, and a step 1600 of obtaining the number of occurrences of the typhoon and the amount of generated rainfall.

By complementing the global climate model by the above process, we can solve the problem of not reflecting extreme rainfall such as typhoon in the existing global climate model (GCM), and it can suggest the process of local refinement that can be precisely applied to small watershed scale .

Referring to FIG. 2, a supplementary method of global climate model according to an embodiment of the present invention first constructs detailed input data for inputting observation rainfall and removing typhoon rainfall. In addition, we construct detailed input data to perform principal component analysis of GCM climate variable. The artificial neural network structure is determined based on the detailed input data and the artificial neural network (ANN) is used to simulate the forecast period rainfall.

Then, the rainfall of the artificial neural network simulation result is corrected by applying the abnormal or dynamic condition. At the same time, a stochastic typhoon simulation process is performed to select typhoon storms with past statistical characteristics.

Through this simulation process, the final predicted rainfall is selected and grid refinement is performed using a random cascade.

The method of supplementing the global climate model according to the embodiment of the present invention described above is related to stochastic hurdle simulation.

In the step 1100 of constructing rainfall data for the simulation of the hurricane for the past period, observational data such as the frequency of occurrence of typhoons and the occurrence rainfall of typhoon are collected for the past period (Baseline, 1976 to 2000) of the GCM model . In other words, the statistical parameters such as the monthly average frequency of typhoons, the maximum typhoon rainfall, the minimum typhoon rainfall or the average typhoon rainfall are constructed or collected for the past period in which the observed value exists.

FIG. 3 illustrates an exemplary programmed initial screen for performing stochastic typhoon simulation, and FIG. 4 illustrates an exemplary programmed initial screen for estimating final predicted rainfall. Step 1100 can construct or collect rainfall data for past periods through the program shown in FIG.

In step 1200 examining the rainfall for a typhoon event that has affected the watershed, the rainfall intensity of the hurricane history that influenced a small watershed size (for example, the Korean peninsula) based on past hurricane observations constructed at step 1100 .

After analyzing the rainfall amount of the typhoon event, a step 1300 of analyzing the statistical characteristics of the typhoon event is performed. In the step 1300 of analyzing the statistical characteristics, the number of occurrence of the typhoon's monthly average, Statistical parameters such as rainfall, minimum typhoon rainfall or mean typhoon rainfall can be analyzed.

After analyzing the statistical characteristics, the Poisson distribution is applied (1400) and the Campbell distribution is applied (1500). The complementary method of the global climate model applying stochastic hurdle simulation according to an embodiment of the present invention is a method of generating a value by applying a random sample to a given cumulative distribution function (CDF) Inverse Transform Method is applied as well.

The frequency of occurrence of typhoons in the forecast period (future period) uses the Poisson distribution to explain the number of events occurring within a certain range, and for typhoon rainfall, Gumbel) distribution. In this way, the simulation period reflecting the statistical characteristics (monthly average number and maximum / minimum / average rainfall) of typhoon rainfall over the past 30 years through the mixed distribution of Poisson distribution and the Campbell distribution for the occurrence frequency and total rainfall of typhoon The process of generating monthly monthly typhoon simulation values can be constructed. That is, the Poisson distribution may be applied 1400 to simulate the occurrence frequency of the typhoon in the prediction period, and the Sambell distribution 1500 may be applied to simulate the occurrence of the typhoon.

In case of statistical data of the removed typhoon rainfall used as the input data, the user can easily construct the detailed input data (observation rainfall) through the input screen as shown in FIG. 3, and the result of the monthly simulated typhoon rainfall event Can be confirmed.

In the step 1400 of applying the Poisson distribution to simulate the number of occurrences of the typhoon in the forecast period, the number of occurrence of the typhoon in the forecast period is determined by applying the Poisson distribution which describes the number of events occurring within a specific range, The formula for the distribution is as shown in [Equation 2].

Where χ is the number of hurricanes, λ is the average number of hurricanes, and β is the scale factor.

The formula for the Campbell distribution in step 1500 of applying the Gumbel distribution to simulate the occurrence of typhoons in the forecast period is as follows:

, 위치변수

이며,Here,

, Positional variables

Lt;

는 과거 태풍 발생강우량에 대한 평균, σ는 과거 태풍 발생강우량에 대한 표준편차이다. 검벨(Gumbel) 분포의 매개변수인 축척변수(β)와 위치변수(μ)는 모멘트법을 사용하여 추정할 수 있다.

Is the mean of past typhoon rainfall, and σ is the standard deviation of past typhoon rainfall. The scale parameter (β) and the positional parameter (μ), which are parameters of the Gumbel distribution, can be estimated using the moment method.

The complementary method of the global climate model applying the stochastic typhoon simulation according to the present invention is a simulation period reflecting the statistical characteristics of the typhoon rainfall over the past period through the mixed distribution of Poisson distribution and the Campbell distribution for the frequency of typhoon occurrence and total rainfall A monthly storm simulation value can be generated.

In step 1600 of calculating the number of occurrences of the typhoon and the amount of generated rainfall, it is possible to predict or estimate the typhoon rainfall according to the statistical characteristics of the past period by multiplying the occurrence frequency of the typhoon and the occurrence rainfall of the typhoon. That is, the monthly typhoon simulation value reflecting the statistical characteristics (the number of occurrences of the monthly average and the maximum / minimum / average rainfall) of the typhoon rainfall over the past period can be generated through the multiplication of the finally obtained typhoon occurrence frequency and the generated rainfall amount.

Meanwhile, in the step 1600 of obtaining the number of occurrences of the typhoon and the generated rainfall, the past statistical characteristics may include a monthly average occurrence frequency of the typhoon, a maximum typhoon rainfall amount, a minimum typhoon rainfall amount or an average typhoon rainfall amount. The results are summarized in Table 1, which is applied to the Yangpyeong Observatory in order to verify the complementary method of the global climate model using the stochastic typhoon simulations.

In Table 1, * represents the error rate for the observed values in [Table 1].

As shown in [Table 1], the supplementary method of the global climate model applying the stochastic typhoon simulation according to the present invention to the Yangpyeong Observation Site is applied to the statistical characteristics of the past period (the monthly average frequency and the maximum typhoon rainfall / mean typhoon rainfall) Typhoon rainfall with an error rate within the range of ± 15% could be generated. Therefore, it is confirmed that the supplementary method of the global climate model according to the embodiment of the present invention can overcome the disadvantage of GCM which can not simulate typhoon rainfall, and can more accurately perform local refinement.

Meanwhile, the step 1600 of obtaining the number of times of occurrence of the typhoon and the amount of generated rainfall can calculate the extreme rainfall event. As shown in FIG. 5, according to the method of supplementing the global climate model using the stochastic typhoon simulation according to an embodiment of the present invention, extreme rainfall events of 400 mm or more, which was not seen in the conventional GCM, can be generated. Referring to FIG. 5, the GCM indicated by the thick dotted line has no rainfall pattern of more than 400 mm, whereas the supplementary method of the global climate model applying the stochastic typhoon simulation according to the present invention indicated by the fine dotted line ) Can produce rainfall events of over 400mm.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Accordingly, the spirit of the present invention should not be construed as being limited to the embodiments described, and all of the equivalents or equivalents of the claims, as well as the following claims, belong to the scope of the present invention .

Claims (8)

Inputting observation rainfall and removing typhoon rainfall;
Determining an artificial neural network structure based on the constructed detailed input data;
Simulating the rainfall of the prediction period using the determined artificial neural network;
Correcting the rainfall of the artificial neural network simulation result by applying a dynamic decimation;
Stochastic typhoon simulation process to select typhoon storms with past statistical characteristics;
Selecting a final predicted rainfall through a simulation process; And
And performing grid refinement using a probability distribution model,
The step of performing the stochastic typhoon simulation may include:
Establishing rainfall data for typhoon simulation for past periods;
Obtaining rainfall data for typhoon events that have affected the watershed;
Obtaining statistical parameters for typhoon events;
Applying a Poisson distribution to simulate the frequency of occurrence of a typhoon in the forecast period;
Applying the Campbell distribution to simulate the occurrence of typhoons in the forecast period; And
And a step of calculating the number of times of occurrence of the typhoon and the amount of generated rainfall,
A method of complementing global climate models using stochastic hurricane simulations performed by a computer, characterized by using inverse transforms as a method of generating a value by applying a random sample to a cumulative probability distribution.
The method according to claim 1,
Wherein the step of obtaining the statistical parameter for the typhoon event comprises analyzing the monthly average hurricane frequency, the average rainfall amount or the maximum rainfall amount, and a computer-implemented method for supplementing the global climate model using stochastic typhoon simulation.
3. The method of claim 2,
In the application of the Poisson distribution, the number of occurrence of the typhoon in the prediction period is applied to Poisson distribution which describes the number of events occurring within a specific range, and the formula for the Poisson distribution is

Where λ is the average number of occurrences of typhoon, and β is a scale factor. A computerized method of supplementing the global climate model using stochastic typhoon simulations.
The method of claim 3,
In the step of applying the Gumbel distribution, the formula for the < RTI ID = 0.0 >

ego,
Here,

, Positional variables

Lt;

Is an average of past typhoon-generated rainfall, and σ is a standard deviation of past typhoon-generated rainfall. This is a supplementary method for a global climate model using a stochastic typhoon simulation performed by a computer.
5. The method of claim 4,
Wherein the step of generating the typhoon generates the monthly typhoon simulation value reflecting the statistical characteristics of the typhoon rainfall during the past period by multiplying the finally obtained typhoon occurrence frequency and the generated typhoon rainfall, A complementary method of global climate model applying stochastic typhoon simulations.
6. The method of claim 5,
Wherein the statistical characteristic of the typhoon rainfall during the past period includes the occurrence frequency of the typhoon, the maximum typhoon rainfall amount, the minimum typhoon rainfall amount or the average typhoon rainfall amount in the step of obtaining the number of occurrences of the typhoon and the generated rainfall amount. A complementary method of global climate model using stochastic typhoon simulations.
5. The method of claim 4,
Wherein the step of obtaining the number of occurrences of the typhoon and the amount of generated rainfall includes calculating an extreme precipitation event, and a computer-implemented method of supplementing the global climate model using the stochastic typhoon simulation.
8. The method according to any one of claims 1 to 7,
Wherein the step of applying the Poisson distribution to simulate the number of times of the occurrence of the typhoon in the prediction period and the step of applying the Sambel distribution to simulate the amount of the generated typhoon in the prediction period are applied in combination. A complementary method of global climate model applying the typhoon simulation.

Images