search

PCMDI / PCMDIobs-cmor-tables

JSON Tables create the PCMDI observational database using CMOR3
Other
0 stars 0 forks source link

Cmorized TRMM_3hourly #28

Open chengzhuzhang opened 4 years ago

chengzhuzhang commented 4 years ago

I was looking at the original TRMM 3hourly files and cmorized version, it appears that the original data was generated at 0Z, 3Z, 6Z while after cmorization, the time was shifted back for 1.5 hour.

To look at the time stamps: Original: ncdump -ct /p/user_pub/pmp/pmp_obs_preparation/orig/data/TRMM_3hr_ccovey/gpfs/fs1/p/cgd/cas/zhangyx/TRMM.3HRLY/3B42.201312.3hr_V7.nc

Cmorized:ncdump -ct /p/user_pub/PCMDIobs/PCMDIobs2/atmos/3hr/pr/TRMM-3B43v-7/gn/v20200402/pr_3hr_TRMM-3B43v-7_BE_gn_v20200402_201312010130-201312312230.nc

gleckler1 commented 4 years ago

@chengzhuzhang Thanks for pointing this out. The bounds were set at t,t+3 rather than t-1.5,t+1.5, and cmor appears to have shifted t. a new version is being run here: /p/user_pub/PCMDIobs/PCMDIobs2/atmos/3hr/pr/TRMM-3B43v-7/gn/v20200430 pls confirm.

chengzhuzhang commented 4 years ago

Thank you @gleckler1 ! The time bounds are fixed. After a closer examination, I think there is also unit conversion problem, and minor issues in the global attribute of the metadata. The values in the cmorized files are too large for precipitation flux with units [km/m2/s]. The original files don't have a units attribute for precipitation, which might cause the problem? I should have checked it earlier...In the meantime, I downloaded the latest version of TRMM 3hourly, which has the unit attributes [mm/hr] and includes year span (2000-2019), I think I will try include it into PCMDIobs2.0

gleckler1 commented 4 years ago

@chengzhuzhang Can you confirm that this is still fixed (v20200707) after all of the changes made?