Hi guys, I have been working on idealized cases using WRF-ARW V4.4.2. I try to average the temperature tendency caused by radiative heating at each level and at each time step (RRTMG scheme is used), but it seems rather difficult in a mpirun since the calculation is divided into many tiles and only the boundary information of a tile will be transported to neighbor tiles. However, I have to sum variable values of all grids in the domain first and then substitute the value of each grid with the horizontally averaged one. Here is the procedure by which I tried to realize my goal:
1) in the module_radiation_driver.F, I found the loop was from 1 to num_tiles (which I thought was the total number of tiles), so I averaged the variable (RTHRATEN) in each tile at every loop;
2) I added up the horizontally averaged values of all tiles and stored it into a local variable;
3) after the loop, I divided the local variable by the num_tiles;
4) I created another loop (from 1 to num_tiles) to substitute each grid with the local variable.
The compilation was successful but the averaging procedure didn't work. The result was exactly the same with the run that was not modified.
I hope you could kindly give me some advice on this issue.
1) in the module_radiation_driver.F, I found the loop was from 1 to num_tiles (which I thought was the total number of tiles), so I averaged the variable (RTHRATEN) in each tile at every loop;
2) I added up the horizontally averaged values of all tiles and stored it into a local variable;
3) after the loop, I divided the local variable by the num_tiles;
4) I created another loop (from 1 to num_tiles) to substitute each grid with the local variable.
The compilation was successful but the averaging procedure didn't work. The result was exactly the same with the run that was not modified.
I hope you could kindly give me some advice on this issue.