Modeling And Simulation Of Spatio-temporal Objects Of Multi-granularity Of Climate Evolution Based On Bayesian Network

In the context of global warming,the impact of climate change is not only a scientific issue,but also a major issue related to sustainable social and economic development,national security,and environmental diplomacy.In order to be able to accurately and comprehensively understand the laws of climate change,scientists present the complex climatic processes and interactions between circles in the climate system through mathematical expressions,namely climate models.Humans use climate models and supercomputers to perform operations on the complex evolutionary processes involved in order to grasp the evolutionary process of the climate system.Therefore,the development of climate models represents mankind's understanding and knowledge of the climate system.The traditional climate models based on numerical simulations,such as atmospheric circulation models,are difficult to support high-precision model simulations on small computers due to their high complexity.Relatively simple climate models such as the A Globally Resolved Energy Balance(GREB)model greatly simplify the physical process of climate evolution,resulting in insufficient accuracy and efficiency in climate simulation.Although the statistical models constructed by machine learning methods show good results in the simulation of climate evolution,the model effects depend on the input of the model,and the model has poor interpretability in the physical laws of the climate process.How to construct a climate evolution model from the perspective of dynamic coupling based on the theoretical basis of the GREB model,organic integration and nested statistical models,so that it can effectively improve the efficiency and accuracy of the simulation,and at the same time have relative stability to the data,and take into account the physical laws of climate evolution.This paper has conducted in-depth research on the expression and simulation methods of the attribute states of the climate evolution process.The main research contents and results are as follows:(1)Based on GREB model theory and full-spatial information modeling theory,model spatio-temporal objects of multi-granularity of climate evolution in GREB model.The GREB model climate evolution modeling ideas are analyzed,and the GREB model climate process is abstracted and modeled by objects.(2)Based on Bayesian network theory,the relationship between spatio-temporal objects of multi-granularity is expressed through the graph model of Bayesian network.The conditional probability table is used to describe the relationship strength of spatiotemporal objects of multi-granularity.Construct an expression method of climate evolution based on Bayes network.(3)Case analysis.Introduced experimental data processing and configuration,designed experimental procedures and experimental methods,including node selection,state grading,relationship construction,data training and state simulation.Through the experimental method in this paper,the state simulation of the average surface temperature,the average temperature of the atmosphere,and the average humidity of the atmosphere is realized,and the simulation results of the GREB model are compared with the experiment.The experimental results verify the applicability and accuracy of the method in this paper,and it has a good effect on the simulation of the average surface temperature,the average atmospheric temperature and the attribute characteristic state of the average atmospheric humidity.It provides strong support for the numerical simulation of climate evolution variables in the future.