University of Twente Student Theses
Comparing alternative methods for a priori optimization of the slow-mode closure for the Lorenz 96 system
Groot, Justin de (2022) Comparing alternative methods for a priori optimization of the slow-mode closure for the Lorenz 96 system.
|
PDF
1MB |
| Abstract: | We coarsen a weather forecast model with multiple time scales by approximating the changes of the slow variables caused by the fast variables, which is the true subgrid. We narrow the problem to the Lorenz 96 model, which is a simplified weather forecast model with two time scales. We try to approximate the true subgrid using a parameterized subgrid, by minimizing a loss function which defines the error of the approximation. There are currently two approaches known for defining this loss function, an a priori and an a posteriori approach, where the former used theoretical deduction and the latter uses empirical observation. A previous paper which introduced the a posteriori approach showed that it gave better results. However, the results found for the a priori approach was done using linear regression, while machine learning was used for the a posteriori approach. So, we wish to use machine learning to solve the a priori approach, to determine if the a priori approach can achieve the same results when also being solved by machine learning. Here we show that while using machine learning improves the accuracy of a priori, there is still a big gap between the results found for a posteriori. |
| Item Type: | Essay (Bachelor) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 31 mathematics, 54 computer science |
| Programme: | Computer Science BSc (56964) |
| Link to this item: | https://purl.utwente.nl/essays/94698 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page