Known Issues

Extra timesteps in HIRHAM runs

The original files for some HIRHAM variables had extra timesteps, which we have removed.

For the HIRHAM ERA-Interim and EC-Earth historical runs, the files for 2001-2005 and 1961-1965 (respectively) had extra timesteps for a number of variables: clivi, clt, clwvi, prhmax, prsn, prw, sfcWindmax, snm, snw, sund, tasmax, and tasmin. (Not all of these variables are part of the NA-CORDEX archive.)

In these files, there were 10 extra days of data inserted between the correct data from timesteps 760-770 (0-based counting), which corresponds to the first 10 days of February in 2003 (ERA-I) or 1963 (EC-Earth). We believe this is due to a post-processing error at the original modeling center.

The extra timesteps look like weather, but come from a different time of year. They cause the files in question to be 10 days too long, and to have a non-monotonic time coordinate: at the end of the file, it jumps back by 10 days and counts up to the end again.

We have resolved the problem by removing the extra days, but be aware that the remaining values in the affected ten-day period may be suspect.