Approximation of planetary waves from combined satellite and ground-based observations using an adaptive metaheuristic algorithm
V.I. Sivtseva1, A.V. Savvin1, V.V. Grigoriev1, I.I. Koltovskoi2
1Federal State Autonomous Educational Institution of Higher Education "M.K. Ammosov North-Eastern Federal University", Yakutsk, Russia
2Yu.G. Shafer Institute of Cosmophysical Research and Aeronomy of the Siberian Branch of the RAS, Yakutsk, Russia
Keywords: planetary waves, Rossby waves, Artificial Bee Colony (ABC), satellite data, Aura (MLS), rotational temperature, hydroxyl
Abstract
The study of large-scale atmospheric processes, such as planetary waves, plays a crucial role in understanding the coupling between the lower and upper layers of the atmosphere. However, accurate modeling of these waves is challenging due to the heterogeneity and sparsity of observational data (both satellite and ground-based), as well as the high dimensionality of the parameter space when describing wave structures. In this paper, an approach to approximation of planetary waves with the use of the Two Strategy adaptive Artificial Bee Colony (TSaABC) algorithm based heterogeneous satellite and ground-based data is suggested. The TSaABC algorithm is used to optimize the parameters of a nonlinear spatiotemporal model representing atmospheric temperature data obtained from the Aura satellite (MLS) and three ground-based stations measuring hydroxyl OH(3, 1) emission bands. The temperature data are approximated using the sum of planetary wave harmonics with unknown parameters including amplitudes and wavenumbers, which are selected from a dictionary of harmonics. By solving the inverse problem of minimizing the data divergence and the L1-norm of harmonic amplitudes, the method achieves approximation accuracy and sparsity in a large dictionary of harmonics. To solve the L1-minimization problem, a hard thresholding strategy was developed within the TSaABC algorithm. The use of a hard threshold value allows us to reduce the dimensionality of the solution search, thus inereasing computational efficiency. The results demonstrate the potential of the algorithm for assimilating heterogeneous data and improving the modeling of atmospheric processes.
|
