search

FlexTRKR / PyFLEXTRKR

PyFlexTRKR
Other
66 stars 21 forks source link

setting up 30-minute output for MCS tracking #48

Closed weiming9115 closed 1 year ago

weiming9115 commented 1 year ago

Hi, I am using GPM-IMERGE and MERGE-IR, 30-minute temporal resolution, for MCS tracking. I set up " datatimeresolution: 0.5" to inform the data temporal resolution (0.5 hour), but it doesn't give me outputs of 30-minute when checking the "base_time" of tracked MCSs, only hourly time stamps.

Is it possible to identify the objects every 30 minutes by setting parameters in config.yml? because I only see the "hourly" option in "config_imerg_mcs_tbpf_example.yml", which probably led me to the hourly outputs:

idclouds_hourly: 1 # 0: No, 1: Yes idclouds_minute: 0 # 0 or 30 min, which minute mark to use the Tb data idclouds_dt_thresh: 5 # [minute], time difference allowed between actual data and idclouds_minute

Thank you, Wei-Ming

feng045 commented 1 year ago

Hi, I think if you set idclouds_hourly: 0, the code should use all available times in each of the input files.

Although you may want to also set clouddatasource to something other than 'gpmirimerg' (as in the provided IMERG config example). This is because for our previous working developing the global MCS tracking dataset (Feng et al. 2021 JGR), we kept the 30 min MERGED-IR (Tb) data but averaged the 30 min IMERG precipitation data to hourly. Hence, the tracking code looks for a specific flag clouddatasource: 'gpmirimerg', if that flag is set, then it only uses IMERG from the first time in the hourly file, which is probably not what you want.

What does your GPM-IMERGE and MERGE-IR look like? Do you have two 30 min frames in an hourly file? If that's the case, what I described above should work.

Best, Zhe

weiming9115 commented 1 year ago

Hi,

Got it. I used idclouds_hourly: 1 because I was wondering if idclouds_hourly: 0 may skip the idclouds process. Maybe a comment in the config file help, such as "using NO for all available timeframes"?

I do have two timeframes in an hourly data for GPM and MERGE-IR, so it should work fine as you described. Thank you so much!

Best, Wei-Ming

feng045 commented 1 year ago

Added a comment to explain this parameter in #51. Closing this issue.