| Abstract |
Methods to extract the phenomena that have a great impact to society from big analysis or forecast data are important. One of them is extratraopical cyclone tracking. Conventionally it is realised by searching the temporally-neighbouring cyclone centres from a set of adjacent synoptic charts or equivalently isobaric-surface data. This method hardly evaluated the merging as observed around Japan, however. I have developed the new tracking method named NEAT to search the overlap of temporally adjacent cyclone regions. Applying it to reanalysis data, I found that the cyclone merging more frequently happened above the western boundary current. This method can be applied to the tracking of typhoons and that of oceanic meso-scale eddies. Moreover, I am developing another method with persistent homology, one of the computational topology algorithms, in order to achieve a successful tracking of thin-shaped phenomena such as atmospheric front. I will talk the new results with a brief introduction to homology. |
