Added lower limit to temperature used by RRTMGP.

[dustinswales]

This PR contains a snippet of code to apply a bounds check on the layer temperature provided to RRTMGP.

RRTMGP reports an error and quits when out of range temperatures are provided. This was discovered to occur in the upper atmosphere when using 127 vertical layers.

[yangfanglin]

Dustin,
Could you please explain in more detail the change you made ? Is the temperature being reset to a T_min if it is smaller than T_min (160K) ? If so, this temporary fix needs to be updated at the earliest possible.

if (t_lay(iCol,iLay) .le. lw_gas_props%get_temp_min()) then
t_lay = lw_gas_props%get_temp_min() + epsilon(lw_gas_props%get_temp_min())
endif

[dustinswales]

Dustin,
Could you please explain in more detail the change you made ? Is the temperature being reset to a T_min if it is smaller than T_min (160K) ? If so, this temporary fix needs to be updated at the earliest possible.

if (t_lay(iCol,iLay) .le. lw_gas_props%get_temp_min()) then
t_lay = lw_gas_props%get_temp_min() + epsilon(lw_gas_props%get_temp_min())
endif

@yangfanglin
You're correct.
The temperature is set to the lowest value allowed in RRTMGP, lw_gas_props%get_temp_min(), for cases with temperatures lower.

[climbfuji]

@dustinswales I added it to the commit queue, hope to get this in soon.

[climbfuji]

@dustinswales Can you confirm whether an update of the rrtmgp-rte submodule pointer is required or not?

[dustinswales]

@climbfuji
No update needed to the rte-rrtmgp submodule.