CN112800677A - Simplified global climate change prediction method - Google Patents

Abstract

The invention discloses a simplified climate change prediction method. The method comprises the steps of firstly, obtaining annual average net radiation compelling according to the heat absorption and heat release of the earth, and carrying out regression analysis according to the change of annual average compelling radiation and the change of average air temperature to obtain a simplified linear model; and then, considering the ocean temperature change, performing data fitting on the existing ocean temperature data, predicting the future global ocean temperature change by adopting an improved neural network model, and finally performing polynomial fitting on the ocean temperature change and the earth surface temperature change to determine the global climate change.

Description

Simplified global climate change prediction method

Technical Field

The invention relates to the field of climatology application, in particular to a simplified climate change prediction method

Background

The main cause of global warming is the large use of fossil fuels by mankind over the last century and the emission of large amounts of greenhouse gases. The existing prediction method is mainly a chaos-artificial neural network, and further predicts the change of climate and analyzes the trend by analyzing the change processes of temperature, precipitation and runoff of a basin in recent years. However, the climate is an average state for a long time, the climate changes little in a short time, the global warming is a global problem on a climate scale, and the global temperature change needs to be observed and accumulated for a long time in a global scope from a climate perspective, rather than the problem that the climate change of a certain area in a certain time period is singly researched and the neural network model has distortion of a long-term result, the existing prediction method has the problem of inaccurate prediction. In addition, the traditional method does not consider factors such as ocean heat absorption, and the influence of the ocean heat absorption on global climate change is great. Meanwhile, the traditional method has complex parameters, and is not beneficial to non-professionals to understand and know the global climate change situation.

Disclosure of Invention

The invention discloses a method for predicting climate change through long-term observation data analysis. The climate change is analyzed and predicted through long-term observation data, and the model is simplified, the method is different from the traditional complex climate model, and factors such as ocean heat absorption are considered by the model, so that the model prediction accuracy is improved. The model is beneficial for non-professionals to understand and know the global climate change situation, the internal characteristics of factors influencing the climate change are proved, the understanding of people on the climate change is enhanced, and a decision maker is urged to rapidly make a policy for coping with the climate change. Firstly, in the research, relevant actually measured data information is searched, collected data are preprocessed by using a spread software, and according to the heat absorption of the earth (including solar radiation forcing, greenhouse gases, fuel consumption and aerosol) and main factors of global temperature change caused by the heat release of the earth, the collected data are converted into the annual average change of radiation forcing by adopting different methods, so that the annual average net radiation forcing is obtained. A simplified linear relation model is obtained by performing regression analysis on the change of the average forced radiation and the change of the average air temperature in 1895-1955. The year average net radiation of 1956-2000 is then forcedly substituted into the model to obtain simulated earth surface average air temperature, which is compared with the actual observed value. And modifying and adjusting the model according to the difference between the simulation value and the observation value to obtain an optimized model. And then, considering the influence factor of the ocean temperature change, performing data visualization processing and data fitting on the acquired ocean temperature data, predicting future global ocean temperature change by adopting an improved neural network model, and finally performing polynomial fitting on the ocean temperature change by combining the earth surface temperature change to determine global climate change. The method is characterized by comprising the following steps of:

s1, calculating the net radiation compelling of the earth except the ocean temperature relative to 1880;

s2, analyzing a linear relation between the earth net radiation compulsive change and the global earth surface temperature change from 1895 to 1955 to obtain a final optimization model;

s3, combining the earth surface temperature change of S1, and predicting future global ocean temperature change by adopting an improved neural network model;

s4, fitting the global earth surface temperature and ocean temperature change predicted values by adopting two-dimensional polynomial data,

the step S1 further includes the steps of:

s1.1, analyzing the forced change of solar radiation;

s1.2, calculating radiation compelling generated by earth heat absorption:

wherein W_{Suction device}Mean radial force W/m for Earth's endotherms²；W_CO2Is CO₂Resulting radiation forced W/m²；W_CH4Is CH₄Resulting radiation forced W/m²；W_NO2Is N₂O generated radiation forced W/m²，W_FuelRadiation directly produced by fuel consumptionForced W/m²；W_{Aerosol and method of making}Forced W/m of radiation generated for aerosol²；W_{Solar radiation}Forcing W/m for solar radiation to earth²；

S1.3, calculating radiation compelling generated by earth heat release:

wherein W_PutMean radial force W/m for exothermic heat generation of the earth²(ii) a Mu is the emissivity of the earth, and is 0.6; t is_kiThe temperature is the global average temperature variation caused by K years of net radiation compelling except the ocean temperature;

s1.4, calculate the net radiation compelling of the earth other than ocean temperature versus 1880:

W_{net i}＝W_{Suction device}-W_Put

The step S2 further includes the steps of:

s2.1, obtaining a linear model: the global mean air temperature and annual net radiation coercion were modeled with 1896-1955 correlation data:

T_ki＝m.W_{net i}+n

Note: m, n are coefficients related to the net radiation intensity;

s2.2, optimizing a linear model: and comparing and analyzing the simulated value and the actual observed value, and modifying and optimizing the model to obtain a final optimized model. According to the difference between the temperature change simulation value and the actual observation value, the model is continuously corrected, and the finally corrected model is as follows:

T_ki＝0.215W_{net i}-0.131

The step S3 further includes the steps of:

s3.1, performing data visualization processing and data fitting on the acquired ocean temperature data by considering the influence factor of ocean temperature change;

s3.2, predicting future global ocean temperature change by adopting an improved neural network model in combination with the earth surface temperature change of S1;

the step S4 further includes the steps of:

s4.1, predicting surface temperature change in the future 25 years of the world;

s4.2, drawing a data scatter diagram through known data, and determining the degree n of fitting a polynomial;

s4.3, calculating

And

s4.4, establishing a polynomial coefficient equation set, and solving a polynomial coefficient a_k；

S4.5, obtaining a fitting polynomial

S4.6, fitting by using a two-dimensional polynomial data to obtain an equation:

T_k＝-0.09779T_ki ²+0.2302T_ki-0.3355T_kj ²+0.6605T_kj+0.5999T_kiT_kj-0.01053

and S4.7, predicting the average temperature of 25 years in the future in the world.

Drawings

FIG. 11881- -graph of solar radiation forced change in year 2000 (relative to 1880);

FIG. 21881- -2000 Global CO₂A graph of concentration change;

FIG. 3 Global CO₂The intensity of the radiation generated;

FIG. 41881- -2000 year global CH₄A graph of concentration change;

FIG. 5 Global CH₄The resulting radiation forcing map;

FIG. 61881- -Global NO of 2000₂A graph of concentration change;

FIG. 7 Global NO₂The resulting radiation forcing map;

FIG. 8 direct generation of radiation compels by global fuel consumption;

FIG. 9 is a graph of the radiation forcing variation produced by the main global aerosol;

FIG. 10 is a graph of total radiation variation for global major aerosol production;

FIG. 11 is a graph of variation of earth long wave radiation;

FIG. 12 is a graph of net radiation forcing versus global mean air temperature change;

FIG. 13 is a graph showing the relationship between the simulated value and the actual observed value of the global earth surface average temperature variation;

FIG. 14 is a graph comparing ocean temperature variation to surface temperature variation;

FIG. 15 predicts ocean temperature related results using neural networks;

FIG. 16 is a binary quadratic regression graph;

FIG. 17 is a global predicted trend graph of the average air temperature change in the future 25 years.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. The method comprises the following steps, and the overall flow chart is shown in figure 1:

s1, calculating the net radiation compelling of the earth except the ocean temperature relative to 1880;

s2, analyzing a linear relation between the earth net radiation compulsive change and the global earth surface temperature change from 1895 to 1955 to obtain a final optimization model;

s3, combining the earth surface temperature change of S1, and predicting future global ocean temperature change by adopting an improved neural network model;

and S4, fitting the predicted values of the global earth surface temperature and the ocean temperature by adopting two-dimensional polynomial data.

S1.1 analyzing the solar radiation forced change. We import the correlation data and plot it using Matlab as a graph, as shown in FIG. 1.

From the figure we can conclude that the solar radiation is forced to vary periodically with timeAnd each period is about a year. The difference between the maximum and minimum solar radiation forced in each cycle is typically 0.2W/m²Within, and with a small variation between the maximum or minimum values of the respective periods, the maximum difference between 1881 and 2000 was about 0.25W/m²Left and right. Solar radiation forces an overall enhancement change from 1981 to 1950, and then substantially stabilizes.

S1.2, calculating the radiation compelling generated by the earth heat absorption. The research selects CO which has great influence on global warming₂、CH₄And N₂O for analysis, data were all from NASA.

(1)CO₂

We import the relevant data and draw a curve to get FIG. 2. From the figure we can see that since 1880 we have been CO₂The concentration basically showed a growing trend except for a small decrease in concentration around 1940, especially after 1950₂The concentration increases very rapidly. CO provided according to IPCC₂Radiation forced calculation formula:

wherein alpha is 3.35,

g(C)＝ln(1+1.2C+0.005C²+1.4×10^-6C³)

calculated to obtain CO from 1881-2000 years₂The resulting radiation intensity (relative to 1880) is plotted as shown in figure 3 using Matlab in the attached table.

(2)CH4

We import the relevant data and plot it to get FIG. 4.

As can be seen from the figure, the global CH has been in 1881₄The concentration is in a growing trend, wherein the growth rate is particularly high in the last 60 to 80 years; the growth rate has a slow trend since the 90 s.

CH provided according to IPCC₄Radiation forced calculation formula:

in the formula: alpha is 0.036

f(M,N)＝0.47ln[1+2.01×10^-5(MN)^0.75+5.31×10^-15M(MN)^1.52]

Calculating CH from 1881 to 2000₄The resulting radiation intensity (relative to 1880) is plotted in the attached table using Matlab, as shown in fig. 5. As can be seen from the figure, since 1881 CH₄The resulting radiation is forced to change in a trend consistent with its concentration change.

(3)N₂O

We import the relevant data and plot it into a curve, which is shown in FIG. 6. As can be seen from the figure, NO in the global atmosphere since 1981₂The concentration has a tendency to increase, with a particularly rapid increase since 1970.

CH provided according to IPCC₄Radiation forced calculation formula:

wherein: alpha is 0.036

f(M,N)＝0.47ln[1+2.01×10^-5(MN)^0.75+5.31×10^-15M(MN)^1.52]

Calculated NO in 1881-2000₂The resulting radiation intensity (relative to 1880) is shown in the attached table, plotted using Matlab, as shown in fig. 7.

And analyzing the relation change of fuel combustion. Since the energy generated by burning the fuel is partly diffused directly into the environment by heat and partly into other forms of energy, such as electrical energy, mechanical energy, etc., by which the human being can engage in many activities. According to the principle of conservation of energy, all of this energy will eventually be dissipated to the environment as thermal energy. Oil per ton equals 4.187X 10¹⁰J, earth surface area of about 5.105X 10¹⁴m²Based on these dataThe radiation compelling directly resulting from fossil fuel consumption was calculated and plotted using Matlab as shown in fig. 8.

Analysis of changes in the main aerosol. The data of the change of the major aerosol in the world from 1881 to 2000 was imported to draw a curve, as shown in fig. 9. The radiation compulsions generated by the various primary aerosols were summed to give a total radiation compulsion, plotted using Matlab as shown in figure 10. And analyzing factors influencing the heat release of the earth and the temperature change. The earth surface absorbs solar radiation and releases part of the long-wave radiation, which is absorbed by the carbon dioxide, which is the water vapor in the atmosphere, and the atmosphere also releases heat, i.e., atmospheric radiation, most of which goes downward. We consider the effect of the earth's long wave radiation on temperature changes.

Calculating the radiation compelling generated by the earth heat absorption:

s1.3, calculating the radiation compelling generated by the heat release of the earth. The radiation of an object is known to be proportional to its 4 th power of absolute temperature according to the Stefan-Boltzman law. According to the average temperature of the earth in 1881-2000, the formula is utilized:

the earth's long wave radiation data was calculated and plotted using Matlab as shown in fig. 11. As can be seen from fig. 11, the earth's long wave radiation has a significantly increasing trend since 1960.

S1.4, calculating the global net radiation compulsive change. Adding all the radiation compences to obtain the global average net radiation compelling in 1881-2000, namely:

W_{net i}＝W_{Suction device}-W_Put

The resulting earnest radiation forcing was adjusted based on the mean net radiation forcing of 1951-1980, and the net radiation forcing referred to later, unless otherwise specified, refers to the adjusted net radiation forcing.

The step S2 further includes the steps of:

s2.1, obtaining a linear model. The global mean air temperature and annual net radiation coercion were modeled with 1896-1955 correlation data:

T_ki＝m.W_{net i}+n

Note: m, n are coefficients related to the net radiation intensity;

and S2.2, optimizing the model. Regression analysis was performed on 1896-1955-related data, and the results are shown in FIG. 13. And substituting the relevant data of one year into the model to simulate the global average temperature change value, and comparing the global average temperature change value with the actual observed value.

According to the relation formula of the global earth surface average temperature and the radiation compelling obtained from the figure 12, 1W/m²The net radiation forcing of (a) may cause a global mean air temperature rise of about 0.228 ℃, i.e. the linear model is:

T_ki＝0.228W_{net i}-0.131

Wherein, T_kiTo simulate the global earth surface mean temperature, W_{Net i}For year-averaged net radiation forcing, k represents the year.

We simulated a year of global climate change as shown in fig. 13. From the above figure we can see that the simulated value of the global earth surface average temperature variation is larger than the actual observed value. Therefore, the model is continuously corrected according to the difference between the temperature change simulation value and the actual observed value, and the finally corrected model is as follows:

T_ki＝0.215W_{net i}-0.131

The step S3 further includes the steps of:

and S3.1, performing data visualization processing and data fitting on the acquired ocean temperature data by considering the influence factor of the ocean temperature change. According to the proposed model formula: t is_k＝T_ki+T_kjAnd performing visualization processing on the existing ocean temperature change data to obtain a graph 14.

And S3.2, combining the earth surface temperature change of S1, and predicting the future global ocean temperature change by adopting an improved neural network model. And carrying out rationality prediction on the ocean temperature by applying a neural network algorithm. The BP neural network is a model of an error back-propagation neural network proposed by the PDP group in california, usa in the 80 s of the 20 th century. The neural network model has strong learning, self-adaptability and fault tolerance, and is one of neural network models which are widely applied in the field of prediction in recent years. The key to prediction using the BP neural network is the prediction accuracy. The BP algorithm is a supervised learning algorithm, and the main idea is as follows: inputting a learning sample, repeatedly adjusting and training extreme values and deviations of the network by using a back propagation algorithm to enable an output vector to be as close to an expected vector as possible, finishing training when the sum of squares of errors of a network output layer is smaller than a specified error, and storing the weight and the deviations of the network. The neural network toolbox integrates various learning algorithms, is internally provided with functions with rich functions, can be directly applied even by beginners without the essence of a solution, and most importantly, can save a large amount of programming time. Network models involved in current neural network toolkits: perceptrons, linear networks, BP networks, radial basis function networks, self-organizing map networks, feedback networks, and the like. Therefore, the problem is to train the ocean temperature variation model directly using the neural network toolbox in Matlab.

For solving the distortion problem of long-term prediction of the BP neural network model, a cyclic method can be adopted for prediction, the prediction by using the cycle means that after the predicted value of the first year is calculated, the weight of the original neural network is eliminated, the predicted value of the year is added into an input vector P or a target T, namely the predicted value of the year is taken as a true value to continue a new round of training, on the basis, the prediction simulation of the next year is carried out, and the like.

In the Matlab neural network toolbox, the sample characteristics are considered comprehensively, and the related data results are obtained, as shown in fig. 15. As can be seen from fig. 15, the ocean temperature changes in the next 25 years generally show an ascending trend, but fluctuation of different amplitudes occurs in some years. From the error analysis chart, we can see that the prediction error is in a reasonable range, and the prediction result is shown in table 1 below.

TABLE 1 predicted value of ocean temperature change in the next 25 years

And S4, performing polynomial fitting, and establishing a simplified model to determine the future global average temperature change value. The earth surface temperature and ocean temperature change of the world in the future 25 years are comprehensively considered, two-dimensional polynomial data fitting is adopted for the earth surface temperature and the ocean temperature change, and the climate change of the future 25 years is predicted.

S4.1 global future 25 years surface temperature change prediction. We use the model in S3 to perform data fitting to approximate the future trend presented by the data, and we will push the forecast year forward by 18 years because of the limited data, and the predicted values obtained from 2001 are shown in table 2.

TABLE 2 Global surface temperature Change prediction

S4.2, drawing a data scatter diagram through known data, and determining the degree n of fitting a polynomial;

s4.3, calculating

And

s4.4, establishing a polynomial coefficient equation set, and solving a polynomial coefficient a_k；

S4.5, obtaining a fitting polynomial

S4.6, the obtained equation is as follows:

T_k＝-0.09779T_ki ²+0.2302T_ki-0.3355T_kj ²+0.6605T_kj+0.5999T_kiT_kj-0.01053

and S4.7, predicting the average temperature of 25 years in the future in the world. According to the calculation result of the formula, a three-dimensional curve chart is drawn, as shown in fig. 16, and the calculated predicted value of the global future 25-year average temperature change is drawn into a chart, as shown in fig. 17 and table 3.

TABLE 3 predicted value of average temperature of 25 years in the world

From table 3 and fig. 17, it can be seen that the global temperature will increase by about 0.389 ℃ in 2020 to 2045 years, during which the predicted value of the global average temperature will generally increase continuously, but the trend of the increase gradually slows down.

And (4) evaluating the method. Because a large area of oceans exists on the earth, the temperature of the earth is not changed immediately after being forced to radiate, but a buffer process is provided, so that the simulated global average temperature is larger than the annual change of an actually observed value. The result in predicting future global air temperature changes is only a prediction of trends and not the absolute average air temperature for a particular year. Because most of the data selected by the model are relative values, and the data may be greatly different according to the absolute values of different sources, the predicted absolute values are meaningless.

Claims (1)

1. A simplified method of predicting climate change, comprising the steps of:

s1, calculating the net radiation compelling of the earth except the ocean temperature relative to 1880;

s2, analyzing a linear relation between the earth net radiation compulsive change and the global earth surface temperature change from 1895 to 1955 to obtain a final optimization model;

s3, combining the earth surface temperature change of S1, and predicting future global ocean temperature change by adopting an improved neural network model;

s4, fitting the global earth surface temperature and ocean temperature change predicted values by adopting two-dimensional polynomial data,

the step S1 further includes the steps of:

s1.1, analyzing the forced change of solar radiation;

s1.2, calculating radiation compelling generated by earth heat absorption:

wherein W_{Suction device}Mean radial forcing for the Earth's endotherm in W/m²；

Is CO₂Resulting radiation forcing in units of W/m²；

Is CH₄Resulting radiation forcing in units of W/m²；

Is N₂O generated radiation forcing in W/m²，W_FuelIs the radiation compelling directly generated by fuel consumption and has the unit of W/m²；W_{Aerosol and method of making}Radiation forcing for aerosol generation in W/m²；W_{Solar radiation}The radiation force generated by solar radiation on the earth is measured in W/m²(ii) a S1.3, calculating radiation compelling generated by earth heat release:

wherein W_PutMean radial force W/m for exothermic heat generation of the earth²(ii) a Mu is the emissivity of the earth, and is 0.6; t is_kiThe unit of global average temperature variation except ocean temperature caused by K years of net radiation compelling is;

s1.4, calculate the net radiation compelling of the earth other than ocean temperature versus 1880:

W_{net i}＝W_{Suction device}-W_Put

The step S2 further includes the steps of:

s2.1, obtaining a linear model: the global mean air temperature and annual net radiation coercion were modeled with 1896-1955 correlation data:

T_ki＝m.W_{net i}+n

Wherein m, n are coefficients related to the net radiation intensity;

s2.2, optimizing a linear model: and comparing and analyzing the simulated value and the actual observed value, and modifying and optimizing the model to obtain a final optimized model. According to the difference between the temperature change simulation value and the actual observation value, the model is continuously corrected, and the finally corrected model is as follows:

T_ki＝0.215W_{net i}-0.131

The step S3 further includes the steps of:

s3.1, performing data visualization processing and data fitting on the acquired ocean temperature data by considering the influence factor of ocean temperature change;

s3.2, predicting future global ocean temperature change by adopting an improved neural network model in combination with the earth surface temperature change of S1;

the step S4 further includes the steps of:

s4.1, predicting surface temperature change in the future 25 years of the world;

s4.2, drawing a data scatter diagram through known data, and determining the degree n of fitting a polynomial;

s4.3, calculating

And

s4.4, establishing a polynomial coefficient equation set, and solving a polynomial coefficient a_k；

S4.5, obtaining a fitting polynomial

S4.6, fitting by using a two-dimensional polynomial data to obtain an equation:

T_k＝-0.09779T_ki ²+0.2302T_ki-0.3355T_kj ²+0.6605T_kj+0.5999T_kiT_kj-0.01053

s4.7, predicting the average temperature of 25 years in the future of the world through the step S4.6.

Images