do not correct zqsn in condensation step when ktherm=2

[zhaobin74]

This PR fixed a bug of correcting snow enthalpyzqsn(1) prematurely when ktherm=2.

In 1440x1080 configuration, snow layer sometimes could get above freezing when mushy layer scheme is used. There is no check for such condition in the thermo solver. Instead, thickness_change subroutine handles such case in an energy and mass conserving manner

Icepack/columnphysics/icepack_therm_vertical.F90

Lines 1314 to 1336 in 008f5f6

	! Remove internal snow melt
	!--------------------------------------------------------------
	! more efficient formulation using Ts, dhs > 0 (not BFB)
	! Ts = (Lfresh + zqsn(k)/rhos) / cp_ice
	! if (ktherm == 2 .and. Ts > c0) then
	! dhs = -dzs(k) * cp_ice*Ts/Lfresh ! dhs < 0
	if (ktherm == 2 .and. zqsn(k) > -rhos * Lfresh) then
	dhs = max(-dzs(k), &
	-((zqsn(k) + rhos*Lfresh) / (rhos*Lfresh)) * dzs(k))
	mass = massice(k) + massliq(k)
	massi = c0
	if (dzs(k) > puny) massi = max(c0, c1 + dhs/dzs(k))
	massice(k) = massice(k) * massi
	massliq(k) = mass - massice(k) ! conserve mass
	dzs(k) = dzs(k) + dhs ! dhs < 0
	zqsn(k) = -rhos * Lfresh
	melts = melts - dhs
	! delta E = zqsn(k) + rhos * Lfresh
	endif

In GEOS coupling, when condensation happens over snow, there is a line to correct round-off error of updated top snow layer enthalpy zqsn(1)

Icepack/columnphysics/icepack_therm_vertical.F90

Lines 1347 to 1348 in abbbd62

	! avoid roundoff errors
	zqsn(1) = min(zqsn(1), -rhos*Lfresh)

This correction is supposed to be for round-off, not for the above freezing case as seen in mushy layer scheme. Note: for BL99, snow layer temperature never gets above freezing so the correction is valid here. If mushy layer scheme applies this correction, energy is not conserved and the model will crash in conservation check. So an if condition is added here to only apply correction for BL99 scheme. In mushy layer, the correction is done as above ⬆️.

Zero-diff for AMIP and coupled runs using CICE4 sea ice model

[zhaobin74]

Thanks @mathomp4. Sorry to bother you in your vacation.