Basic doubts

[met-sree]

Dear Sir,
I successfully installed PyFLEXTRKR on my system, and the demo run was successful with the data you provided. However, I am now trying to run it for another region for April 2023. I have downloaded the precipitation and brightness temperature (Tb) data for April 2023. I merged the entire month's precipitation files, which are half-hourly, into a single file, and did the same for Tb in another file. My question is:

1. Does each hour's Tb and precipitation need to be in a single file, or is it possible to use individual merged files (one for Tb and one for precipitation for the entire month)?
2. In the demo data, you provided half-hourly data, but you used date-time-resolution: 1.0 # [hour]. Does this mean you considered the data as hourly?
3. 4.The land and sea mask data (e.g., IMERG_landmask_saag.nc) typically uses a standard masking convention, where values are zero for the ocean and one for land. However, when I plotted your file, it showed a value of 100 over the ocean and different values over the land. Could you clarify this discrepancy?

Thank you.

[feng045]

Here's my answers to your questions:
1. Does each hour's Tb and precipitation need to be in a single file, or is it possible to use individual merged files (one for Tb and one for precipitation for the entire month)?

No, all input variables must be in the same file. There are example codes to preprocess Tb and precipitation data for IMERG. Specifically, this script combines the regridded Tb (matching the IMERG grid) with the IMERG data into single files.

2. In the demo data, you provided half-hourly data, but you used date-time-resolution: 1.0 # [hour]. Does this mean you considered the data as hourly?

The half hourly data is for Tb. The 30min IMERG data has been averaged to hourly in the script above. If you don’t want to do that, you can modify the script and not average the IMERG precipitation and retain the half hourly data. In developing our global MCS dataset (Section 2.1), we used one of the 30 min Tb snapshots for identifying cloud systems (set in the example config_imerg_mcs_tbpf_example.yml) but use the hourly mean IMERG precipitation data to mitigate issues with IMERG temporal discontinuities (Rajagopal et al. 2023).
idclouds_minute: 30

3. The land and sea mask data (e.g., IMERG_landmask_saag.nc) typically uses a standard masking convention, where values are zero for the ocean and one for land. However, when I plotted your file, it showed a value of 100 over the ocean and different values over the land. Could you clarify this discrepancy?

The land mask file is from the IMERG dataset, where they use 100% as water and 0% as land, and a fraction in between for partial water/land.

[met-sree]

Dear Sir,
I have some questions regarding the tracking of Mesoscale Convective Systems (MCS) using GPM rainfall and brightness temperature data (with default config file in your test run):

1. Based on my observations, it seems that the inner core of cyclone clouds is not being tracked, with only some outer rainbands being captured. Could you please clarify why this is happening? Just for the understanding only.
2. Additionally, I would like to know if it is possible to set a time limit to track the life cycle of an MCS from 1 to 4 hours.

Thank you,
Sree

[feng045]

Hi Sree,
If you are referring to tropical cyclones (TCs), I would argue that TCs are not MCSs. Because of the simplicity in the current MCS tracking criteria (only requiring Tb and precipitation data), TCs are often incorrectly identified as MCSs in most of the Tb/precipitation based tracking algorithms because they easily meet both the size and duration requirements of MCSs. In our global MCS climatology study (Feng et al. 2021), we used the IBTrACS database to exclude tracked MCSs in the vicinity of TCs to handle that problem. This is not something built in to PyFLEXTRKR.

If you want to adjust the MCS duration thresholds, you can change these two parameters in the config file (e.g., config_imerg_mcs_tbpf_example.yml):

• mcs_tb_duration_thresh: 4 # [hour] Tb minimum length of a mcs
• mcs_pf_durationthresh: 4 # [hour] PF minimum length of mcs