wrf_alessandro
Member
Hi everyone,
I am running WRF4.2.2 at urban scale (1km) after having successfully completed the simulations for the parent domains D01 and D02 (15km and 5km).
Unfortunately, in one of my urban domain I get a segmentation fault after 6 hours from the beginning of the simulation. Playing with debugging options I found that the error originates in the TWOSTREAM soubroutine, where I have a divided by 0 at this line:
As in the above equation the only denominator is S1 ( S1 = EXP(-H*VAI) ), I printed the values of the variables used to compute S1 finding that for some reasons the VAI (one-sided leaf+stem area index) becomes too large in one or more points:
As cross check, I prescribed S1 to a fix value when it is null and I have been able to overcome the issue and complete the simulation. Besides, I reviewed the namelist and I found out that the error occurs only when I use the dynamic vegetation option within Noah-MP. Interestingly, the same simulation (same namelist) is running correctly on a different urban domain.
As this issue does not originate into the TWOSTREAM subroutine, but where the VAI variable is computed, I was wondering whether I can prescribe VAI when it is larger than a given value? If so, which is the range of variability of VAI? Anyway, other solutions to find the bug and fix this error are welcome.
Thanks into advance for support,
Alessandro
I am running WRF4.2.2 at urban scale (1km) after having successfully completed the simulations for the parent domains D01 and D02 (15km and 5km).
Unfortunately, in one of my urban domain I get a segmentation fault after 6 hours from the beginning of the simulation. Playing with debugging options I found that the error originates in the TWOSTREAM soubroutine, where I have a divided by 0 at this line:
D1 = P1*TMP2/S1 – P2*TMP3*S1
As in the above equation the only denominator is S1 ( S1 = EXP(-H*VAI) ), I printed the values of the variables used to compute S1 finding that for some reasons the VAI (one-sided leaf+stem area index) becomes too large in one or more points:
H, VAI, S1 = 0.8423577, 108.3101, 0.0000000E+00
As cross check, I prescribed S1 to a fix value when it is null and I have been able to overcome the issue and complete the simulation. Besides, I reviewed the namelist and I found out that the error occurs only when I use the dynamic vegetation option within Noah-MP. Interestingly, the same simulation (same namelist) is running correctly on a different urban domain.
As this issue does not originate into the TWOSTREAM subroutine, but where the VAI variable is computed, I was wondering whether I can prescribe VAI when it is larger than a given value? If so, which is the range of variability of VAI? Anyway, other solutions to find the bug and fix this error are welcome.
Thanks into advance for support,
Alessandro