Flake.v6

[YihuaWu-NOAA]

No description provided.

[HelinWei-NOAA]

In some routines they assume one=oceanfrac + landfrac, so lake is either part of ocean or land. For example, in atmos_model.F90

!$omp parallel do default(shared) private(i,j,nb,ix)
do j=jsc,jec
do i=isc,iec
nb = Atm_block%blkno(i,j)
ix = Atm_block%ixp(i,j)
! use land sea mask: land:1, ocean:0
lsmask(i,j) = floor(one + epsln - GFS_data(nb)%SfcProp%oceanfrac(ix))
enddo
enddo

So we need to double-check if lakefrac+oceanfrac+landfrac=1 in the oro data, otherwise there will be some mismatches for land-sea mask, which will cause the model crash.

[YihuaWu-NOAA]

These are not consistent. We need to fix these. Some people wrote this in atmos_model.F90 without considering lakes, some people wrote a section considering lakes in different logic in GFS_surface_composites.F90, and in FV3GFS_io.F90. Maybe, similar issues are in some other files. *Yihua----------------------------------------------------Yihua Wu, Ph D*I.M. Systems Group NOAA/NWS/NCEP/EMC 5830 University Research Court, Room: 2031 College Park, Maryland 20740, USA Phone: (301)-683-3691 Fax: (301)-683-3703

…

On Mon, Apr 11, 2022 at 9:00 AM HelinWei-NOAA ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In physics/GFS_surface_composites.F90 <#25 (comment)> : > @@ -137,9 +161,11 @@ subroutine GFS_surface_composites_pre_run (im, flag_init, flag_restart, lkm, fra else do i = 1, IM + oceanfrac(i) = one - landfrac(i) - lakefrac(i) In some routines they assume one=oceanfrac + landfrac, so lake is either part of ocean or land. For example, in atmos_model.F90 !$omp parallel do default(shared) private(i,j,nb,ix) do j=jsc,jec do i=isc,iec nb = Atm_block%blkno(i,j) ix = Atm_block%ixp(i,j) ! use land sea mask: land:1, ocean:0 lsmask(i,j) = floor(one + epsln - GFS_data(nb)%SfcProp%oceanfrac(ix)) enddo enddo So we need to double-check if lakefrac+oceanfrac+landfrac=1 in the oro data, otherwise there will be some mismatches for land-sea mask, which will cause the model crash. — Reply to this email directly, view it on GitHub <#25 (review)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ANJBYQGFAA3GJBS47R7ZH7TVEQO5VANCNFSM5SUJ4S2A> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[HelinWei-NOAA]

So oceanfrac+lakefrac+landfrac is set to 1 in the oro data? How about those oro data which don't include lake fraction and lake depth?

[YihuaWu-NOAA]

Lake model should not be on for those oro_data files without lake fraction and lake depth, ie, set LKM =0 *Yihua----------------------------------------------------Yihua Wu, Ph D*I.M. Systems Group NOAA/NWS/NCEP/EMC 5830 University Research Court, Room: 2031 College Park, Maryland 20740, USA Phone: (301)-683-3691 Fax: (301)-683-3703

…

On Mon, Apr 11, 2022 at 4:33 PM HelinWei-NOAA ***@***.***> wrote: So oceanfrac+lakefrac+landfrac is set to 1 in the oro data? How about those oro data which don't include lake fraction and lake depth? — Reply to this email directly, view it on GitHub <#25 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ANJBYQCPPUSBXB3NCTIWEKLVESD75ANCNFSM5SUJ4S2A> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[HelinWei-NOAA]

no matter if flake is on or not. We just want to confirm that in the oro dataset with lakefrac, landfrac+oceanfrac+lakefrac=1, for the oro data without lakefrac, landfrac+oceanfrac=1. @shansun6

[ShanSunNOAA]

In the oro_data, there are 2 mask fields, "landfrac" and "lakefrac". There is no oceanfrac in oro_data to save disk space, since the sum of these three are 100%.

In atmos_model.F90, one=oceanfrac + landfrac is used, simply because 'lsmask' is a "land-sea" mask for the air-sea interaction with the ocean model.

Let me know if I am not clear.

[HelinWei-NOAA]

will have a new PR to have new version of flake