Finding percentiles of daily gridded data
Added by Kyle Lesinger almost 4 years ago
Hello.
I am trying to find the percentile values for each day of the year out of 38 years of daily data with 1 variable (soil moisture). Each grid cell should have it's own daily percentile with the same number of time steps as the original file (e.g., 1792) I am looking at the percentiles for each day (e.g., Jan 1 - Dec. 31) from all 38 years of data. I currently do not see any CDO operators that work with this type of data.
I have investigated ydaypctl, ydrunpctl, and fldpct but these all produce a single year of percentile values at one specific value (e.g., 40th percentile). Additionally I am actually trying to determine the 5th, 10th, ... 95th, 100th percentile values since I have to do investigate changes over time with my future analysis.
Do I have to manually code the actual percentile value in Python or R to see what each day's percentile value is when grouped by month and day?
Any advice about an operator that I have not heard of would be very much appreciated. Currently my python code is very slow for this analysis.
Thanks,
Kyle L.
Replies (1)
RE: Finding percentiles of daily gridded data - Added by Karin Meier-Fleischer almost 4 years ago
Hi Kyle,
what about daypctl (https://code.mpimet.mpg.de/projects/cdo/embedded/cdo.pdf#subsection.2.8.21)?
-Karin