INDEX

• EXAMPLE
• ABOUT
• HOWTOUSE
• ENVIRONMENT
• INSTALATION

EXAMPLE

ABOUT

• Track cyclone center position.
• This project is created as 爆弾低気圧データベース and pytrack
• output format is csv.

Algorithm

1. Find cyclone formation position and datetime.

1. Find local minima by scipy.ndimage.filters.minimum_filter() assuming a cyclic boundary conditon in longitude.
2. Find the grid of the minimum closest to the initial guess lat lon(seted by stdin). This minimum is candidate of cyclone center.
3. If cyclone center presure is 0.5 hPa less than around area pressure, define it low pressure and continue 1,2 step for 6h ago's data. If cannot find low pressure, define cyclone formation datetime is 6h ago.

2. Track cyclone

1. Find local minima by scipy.ndimage.filters.minimum_filter() assuming a cyclic boundary conditon in longitude.
2. Find the grid of the minimum closest to the initial guess lat lon(seted by stdin). This minimum is candidate of cyclone center.
3. Generate nine point stencil with the minimum on grid at the centre
4. Interpolate for the minimum in-between grid using biquadratic interpolation. If interpolation fails use the minimum on grid

Data

• I used GPV data.

• GPV netcdf files have the diffrent datetime data and pressure's unit is Pa. So, I use cdo to split and convert to Pa to hPa. see sample shell script

---

HOWTOUSE

data set construction.

ls ~/data_ini/prmsl
surface-2020-12-01_00-prmsl_hPa
surface-2020-12-01_06-prmsl_hPa
surface-2020-12-01_12-prmsl_hPa
surface-2020-12-01_18-prmsl_hPa
surface-2020-12-02_00-prmsl_hPa
surface-2020-12-02_06-prmsl_hPa
surface-2020-12-02_12-prmsl_hPa
surface-2020-12-02_18-prmsl_hPa
surface-2020-12-03_00-prmsl_hPa
surface-2020-12-03_06-prmsl_hPa

run

python3 -m cyclonetrack -x 135 -y 37 --dir ~/data_ini/prmsl -t 2021-01-06_12 --filetype GPV
python3 -m cyclonetrack -x 135 -y 37 --dir ~/jra55/anl_surf125/202101 -t 2021-01-06_12 --filetype jra55

---

INSTALATION

install library

pip install -r requirements.txt

build (not necessary)

build from source cord.

git clone https://github.com/RyosukeDTomita/cyclone_track.git
cd cyclone_track
python3 setup.py install

build from tar.gz(sometime flozen file is not up to date.)

wget https://github.com/RyosukeDTomita/cyclone_track/blob/master/dist/cyclone-track-0.0.1.tar.gz
pip install cyclone-track-0.0.1.tar.gz

make cyclone-track-0.0.1.tar.gz

git clone https://github.com/RyosukeDTomita/cyclone_track.git
cd cyclone_track
python3 setup.py sdist
pip install ./dist/cyclone-track-0.0.1.tar.gz

uninstall

python setup.py develop -u
pip uninstall cycloen-track

---

ENVIRONMENT

I tested the following environment.

• Python3.8
• Ubuntu 20.04 LTS about library, see requirement.txt

---

ANARYSINGTOOL

Anarysing tool are saved in analyze_tool/.See README.md

OTHER PROGRAM

Calcureate Local Deeping Rate