Dear colleagues is there a solution for using eca_r95p and eca_r99p as indicated in the CDO manual eca_r95p infile1 infile2 outfile note that infile1 has many timesteps since it is contain wet days in mm/days, while infile2 contain few timestep of 95percentile of rainfall. Now when you use eca_r95p infile1 infile2 outfile it complains cdo eca_r95p (Abort): Input streams have different number of time steps!. Please help the solution on how to adjust the timesteps or how to go about to solve the challenges. This question has been asked several times in CDO but there is no common solution. Could you please help to address this challenge I tried to interchange the files and started with file2 then file1 eca_r95p infile2 infile1 outfile, this give the answer and no complains but I dont know if ok to interchange the files
Dear Luhunga, you can compute the number of days with partially calculate the percentile and differences. To compute percentile values, you can use "cdo -ydaypctl,95 -mul -gec,1 INFILE.nc INFILE.nc 95pctl_OUTFILE.nc" then "cdo -yearsum -ltc,0 -sub 95pctl_OUTFILE.nc -mul -gec,1 infile1.nc infile1.nc OUTFILE.nc" gives you yearly sum of number of wet days that are above the 95 percentile of reference period. Let the make clear to understand that script:
--INFILE.nc: reference period daily TOTAL precipitation in MM.
--infile1.nc: The period you want to find number of wet days.
--"-gec,1 INFILE.nc": this part changes all values to 1 above 1 (which is 1mm, default value to define wet days.) to 0 below 1. Let say "INFILE_wet_0-1.nc" to output of this function. This contains just 0's and 1's.
--"-mul INFILE_wet_0-1.nc INFILE.nc" part gives the values of precipitation above 1 mm and gives 0 below 1 mm. Let say that INFILE_wet.nc
--"-ydaypctl,95 INFILE_wet.nc 95pctl_OUTFILE.nc" the final part gives you the 365 values that contains 95 percentiles of everyday.
FOR THE SECOND CODE:
--"-mul -gec,1 infile1.nc infile1.nc" part similarly gives the values for wet days, say infile1_wet.nc
--"-sub 95pctl_OUTFILE.nc infile1_wet.nc" gives the difference from 95 percentiles. note that it calculates 95 percentile minus the values. So we are looking for negative values to obtain days above 95 percentile.
--"-ltc,0" makes 1 all values below 0 and makes 0 all values above 1.
--Now we have 0's and 1's that 1's are the days above 95 percentile of reference period. "-yearsum" function takes the sum of every year, so the result gives you the number of wet days that are above the 95 percentile of reference period. Note that I personally suggest to use "-timpctl,95" instead of "-ydaypctl,95". so you will obtain everyday above 95 percentile of whole timeseries of reference period.
Dear Demiralay, thank you very much. I figured to use the similar approach to what you are suggesting in the last two weeks.
I computed the percentile, let say 99percentile, then I compared the file with wet days (1mm) and the file with 99percentile using the code cdo sub file with wet days file with 99percentile. Here will have the file with negative and positive values. Then for extremely wet days I wrote a code cdo gtc,0 then file of output above then file of results. Then cdo time sum of results ....
This code is exactly similar to the script you have suggested. Once again thank you very much
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression