It is true that this file has a leap day in 2100, a year which is not a leap year in the Gregorian calendar. See the comment in this file: "Leap days (Feb 29) have been left in the data, and may not match actual leap days for all years after 2008."
This file is only used in compsets that use the NO_LEAP calendar (1955-2005 and RCP compsets), so the presence of February 29 is irrelevant in that context. It will cause the model to crash when used with the SD-WACCM compset, which uses the Gregorian calendar. This compset, however, is set up to use a solar data file based on observations by default.
For those who wish to use the wasolar_1948-2140_daily file with the Gregorian calendar, I have created a new version, which will be committed to our subversion input data directory under inputdata/atm/waccm/solar/wasolar_1948-2140_daily_c120306.nc
Until it is available there, you can create the same file by deleting February 29, 2100 with this NetCDF operators command:
Code:
ncks -d time,0,55187 -d time,55189,70099 wasolar_1948-2140_daily_c100805.nc wasolar_1948-2140_daily_c120306.nc
Please note the comments in this file: "Parameters for 2009-2140 created by repeating the last 4 cycles (cycle 20-23) with values from 1965 to 2008 inclusive mapping to 2009-2052, 2053-2096, 2097-2140." Hence, all data for 2009 and later is not based on actual observed solar parameters.
The current file we are using based solely on observed parameters is wasolar_1947-2011_daily_c120227.nc, which goes through December 22, 2011.
