Identification For Physical Parameters Of Sea Ice And Simulation Of Thermodynamic Processes

Sea ice has a very important effect on the global ocean circulation, the atmosphere circulation and the climate change, and its thermodynamic processes directly describe its growth and decay. On the background of the sea ice physical parameters and the thermodynamic processes, based on the sea ice in-situ investigation data at Zhongshan Station obtained by the 22nd Chinese National Antarctic Research Expedition, the control theory for the distributed parameter systems and the numerical computation methods for partial differential equations are used to investigate the identification for the sea ice physical parameters and the simulation of the thermodynamic processes. The main contributions are as follows:1. For the actual heat transfer processes of the snow, the sea ice and the ocean, the one-dimensional and three-dimensional thermodynamic systems are established, the properties of the two non-smooth systems are obtained, L~2 theory for parabolic systems are used to prove the existence and uniqueness of weak solutions of the two systems; then the density, the specific heat, the thermal conductivity and the exchange coefficients of the snow, sea ice and ocean are taken as the identified parameters, the temperature deviation is defined as the performance criterion, and the parameter identification model is put forward; the existence of the identified parameters is discussed, and the first-order necessary conditions for optimality are derived. Therefore, the parameter identification theories of the non-smooth distributed parameter system are applied to the actual sea ice problems, and the mathematical foundation for the numerical computation of the parameter identification problems of the sea ice thermodynamic system is provided.2. At present, the determination method of the sea ice salinity is to estimate it according to the salinity data and the measurement method of sea ice salinity is by hand, but the disadvantages of this method are that the salinity data are very few and the hairlike process can't be obtained. For these reasons, a parameter identification method of determining salinity by the temperature data measured automatically and the few salinity data is put forward. The thermodynamic model during the Polar Night Time is operated, and the Eicken's salinity model is used. First, the existence and uniqueness of solution of the nonlinear thermodynamic system are proved; the coefficients describing the sea ice salinity are taken as the identified parameters, the deviation sum of the sea ice temperature and the salinity is defined as the performance criterion, and a parameter identification model is constructed; the existence of the identified parameters is proved. A new optimization algorithm named Hybrid Accelerating Genetic Algorithm combining the Genetic Algorithm and Hooke-Jeeves Research is constructed to estimate the salinity, and the sea ice data at Nella of Zhongshan Station, Antarctic from June 21 to July 5, 2006 measured by the 22nd Chinese National Antarctic Research Expedition are used. Then another simulation for the sea ice temperature from May 27 to June 20, 2006 is operated. The two results are compared with those computed by Eicken's parameters. Results show that better simulations of the temperature distribution and the salinity distribution are possible with estimated parameters than Eicken's parameters. Thus, the parameter identification method is effective, the obtained salinity function of Nella at Zhongshan Station is applied, and the method can help in interpreting field data and can be used to overcome data gaps.3. Thermodynamic models of sea ice must be accurate, the key of their simulation capacity is not only themselves but also the parameter selection. Thus, it is very crucial to establish and perfect the parameter schemes of thermodynamic models based on field data. In this thesis, the thermodynamic model not during the Polar Night Time is operated, the above obtained salinity function is used, and the identification for three key parameters describing the source term in the nonlinear non-smooth thermodynamic system is operated. First, the domain decomposition method and Galerkin method are used to prove the existence and uniqueness of weak solution of the system. The key parameters are taken as the identified parameters, the temperature deviation is defined as the performance criterion, and the parameter identification model is established. The existence of the identified parameters is proved, the sea ice temperature data at Nella of Zhongshan Station, Antarctic from August 1 to August 31, 2006 measured by the 22nd Chinese National Antarctic Research Expedition are used, and the identified parameters are obtained by the Hybrid Accelerating Genetic Algorithm. Then another simulation for the sea ice temperature from September 1 to September 30, 2006 is operated. The two results are compared with those computed by Zillman's parameters and Shine's parameters. Results show that a better simulation of the temperature distribution is possible with estimated parameters than Zillman's and Shine's. Therefore, the parameter schemes for the source term of thermodynamic models based on field data are perfected.