Possible bug in get_ijk_from_grid

[eca_hesc]

The subroutine get_ijk_from_grid defined in frame/module_domain.F is a generic interface to get_ijk_from_grid{1,2}. Both the grid1 and grid2 subroutines contain a case statement and the one from grid1 is replicated below. I think the cases for DATA_ORDER_ZXY and DATA_ORDER_YZX are reversed in both subroutines.

data_ordering : SELECT CASE ( model_data_order )
CASE ( DATA_ORDER_XYZ )
imsx = grid%sm31x ; imex = grid%em31x ; jmsx = grid%sm32x ; jmex = grid%em32x ; kmsx = grid%sm33x ; kmex = grid%em33x ;
ipsx = grid%sp31x ; ipex = grid%ep31x ; jpsx = grid%sp32x ; jpex = grid%ep32x ; kpsx = grid%sp33x ; kpex = grid%ep33x ;
imsy = grid%sm31y ; imey = grid%em31y ; jmsy = grid%sm32y ; jmey = grid%em32y ; kmsy = grid%sm33y ; kmey = grid%em33y ;
ipsy = grid%sp31y ; ipey = grid%ep31y ; jpsy = grid%sp32y ; jpey = grid%ep32y ; kpsy = grid%sp33y ; kpey = grid%ep33y ;
CASE ( DATA_ORDER_YXZ )
imsx = grid%sm32x ; imex = grid%em32x ; jmsx = grid%sm31x ; jmex = grid%em31x ; kmsx = grid%sm33x ; kmex = grid%em33x ;
ipsx = grid%sp32x ; ipex = grid%ep32x ; jpsx = grid%sp31x ; jpex = grid%ep31x ; kpsx = grid%sp33x ; kpex = grid%ep33x ;
imsy = grid%sm32y ; imey = grid%em32y ; jmsy = grid%sm31y ; jmey = grid%em31y ; kmsy = grid%sm33y ; kmey = grid%em33y ;
ipsy = grid%sp32y ; ipey = grid%ep32y ; jpsy = grid%sp31y ; jpey = grid%ep31y ; kpsy = grid%sp33y ; kpey = grid%ep33y ;
CASE ( DATA_ORDER_ZXY )
imsx = grid%sm32x ; imex = grid%em32x ; jmsx = grid%sm33x ; jmex = grid%em33x ; kmsx = grid%sm31x ; kmex = grid%em31x ;
ipsx = grid%sp32x ; ipex = grid%ep32x ; jpsx = grid%sp33x ; jpex = grid%ep33x ; kpsx = grid%sp31x ; kpex = grid%ep31x ;
imsy = grid%sm32y ; imey = grid%em32y ; jmsy = grid%sm33y ; jmey = grid%em33y ; kmsy = grid%sm31y ; kmey = grid%em31y ;
ipsy = grid%sp32y ; ipey = grid%ep32y ; jpsy = grid%sp33y ; jpey = grid%ep33y ; kpsy = grid%sp31y ; kpey = grid%ep31y ;
CASE ( DATA_ORDER_ZYX )
imsx = grid%sm33x ; imex = grid%em33x ; jmsx = grid%sm32x ; jmex = grid%em32x ; kmsx = grid%sm31x ; kmex = grid%em31x ;
ipsx = grid%sp33x ; ipex = grid%ep33x ; jpsx = grid%sp32x ; jpex = grid%ep32x ; kpsx = grid%sp31x ; kpex = grid%ep31x ;
imsy = grid%sm33y ; imey = grid%em33y ; jmsy = grid%sm32y ; jmey = grid%em32y ; kmsy = grid%sm31y ; kmey = grid%em31y ;
ipsy = grid%sp33y ; ipey = grid%ep33y ; jpsy = grid%sp32y ; jpey = grid%ep32y ; kpsy = grid%sp31y ; kpey = grid%ep31y ;
CASE ( DATA_ORDER_XZY )
imsx = grid%sm31x ; imex = grid%em31x ; jmsx = grid%sm33x ; jmex = grid%em33x ; kmsx = grid%sm32x ; kmex = grid%em32x ;
ipsx = grid%sp31x ; ipex = grid%ep31x ; jpsx = grid%sp33x ; jpex = grid%ep33x ; kpsx = grid%sp32x ; kpex = grid%ep32x ;
imsy = grid%sm31y ; imey = grid%em31y ; jmsy = grid%sm33y ; jmey = grid%em33y ; kmsy = grid%sm32y ; kmey = grid%em32y ;
ipsy = grid%sp31y ; ipey = grid%ep31y ; jpsy = grid%sp33y ; jpey = grid%ep33y ; kpsy = grid%sp32y ; kpey = grid%ep32y ;
CASE ( DATA_ORDER_YZX )
imsx = grid%sm33x ; imex = grid%em33x ; jmsx = grid%sm31x ; jmex = grid%em31x ; kmsx = grid%sm32x ; kmex = grid%em32x ;
ipsx = grid%sp33x ; ipex = grid%ep33x ; jpsx = grid%sp31x ; jpex = grid%ep31x ; kpsx = grid%sp32x ; kpex = grid%ep32x ;
imsy = grid%sm33y ; imey = grid%em33y ; jmsy = grid%sm31y ; jmey = grid%em31y ; kmsy = grid%sm32y ; kmey = grid%em32y ;
ipsy = grid%sp33y ; ipey = grid%ep33y ; jpsy = grid%sp31y ; jpey = grid%ep31y ; kpsy = grid%sp32y ; kpey = grid%ep32y ;
END SELECT data_ordering

 

[Ming Chen]

They look fine to me. what makes you think they are wrong?

 

[eca_hesc]

XYZ gets 123, YXZ gets 213, ZYX gets 321, and XZY gets 132, so I thought ZXY should get 312 and YZX should get 231. But the comments in this module say Z is 2 and Y is 3, which would make all 6 of these cases wrong. It's also possible I don't have any idea what I'm talking about.

 

[eca_hesc]

I still think the ZXY and YZX cases are reversed in frame/module_domain, and it has not caused a problem because model_data_order is set to DATA_ORDER_XZY in frame/module_driver_constants, so these cases are never used. Is there any situation in which the model_data_order will be something other than XYZ or XZY? If not, why have these other cases in the code?

 

[Ming Chen]

Thank you for raising this issue. I will talk to our expert and get back to you ....