PREC_ACC_NC in wrfsfc files not resetting with adaptive time step option

[RCarpenter]

I have a related issue to report. Recently, PREC_ACC_NC in wrfsfc files on nested has not been resetting at hourly intervals (prec_acc_dt = 60, 60, 60), but has continued to accumulate. This behavior began in our real-time runs on 8 May 2021 (which I casually note is the 128th day of the year). The issue does not seem to be present going back at least to July 2020. It does not seem to occur in the wrfout files. The problem exists only if adaptive time stepping is used. It is present in 4.1.2 and 4.2.1. I have not had the chance yet to try V4.3 or the fixes mentioned here.

Even though step_to_output_time and adjust_output_times are set, odd behavior is noted in the output timing. Note that the wrfout_d03 file has a fractional XTIME value but no seconds, while the wrfsfc_d03 file has both fractional XTIME and non-zero seconds.

ncdump -v Times,XTIME wrfout_d03_2021-05-24_01:00:00
 Times = "2021-05-24_01:00:00" ;
 XTIME = 240.0765 ;

ncdump -v Times,XTIME wrfsfc_d03_2021-05-24_01:00:00
 Times = "2021-05-24_01:00:04" ;
 XTIME = 240.0765 ;

Last year I posted a similar issue regarding restart output timing: https://forum.mmm.ucar.edu/phpBB3/viewtopic.php?f=45&t=9286.

 

[kwerner]

This topic has been split from the other, as although it's an issue with adaptive time-stepping, it's different. We want to keep all topics separate. Someone will respond to your inquiry soon.