-
Notifications
You must be signed in to change notification settings - Fork 0
jhan_JaC_Redux_Albedo226
Ultimately for JULES-CABLE v CABLE-CABLE testing, I have been testing HAC-226 v CABLE-226. Several differences have been noted and resolved as described in #226.
There is a phase difference due to the different calcs of zenith angle between the models that appears to be unavoidable with out significant work, however otherwise mmagnitudes should be the same. They are every close. I was looking a 4 timesteps 26 days in and noted a difference in the magnitude of albedo quantities at dawn relative to JULES. I then found this followed though to canopy transmitance, reflectance etc. I moved back to a time-series of the ful year, looked at quantities at the start etc. A curiosity was a spike traceable to snow depth> 0 in Jan. This is very odd as precip forcing shows NO snow.
The problem was traced to snow accum. This has to be investigated in further detail. The proble arose @ timestp 462 due to "dq" becoming negative. As a quick fix I restricted this to +ve values, however the issue of snow depth being > 0 returns @ dt=563.
In further detail this is traceable to precip making through to the ground being possible and +ve finite when the air temp is < 0, and then freezing and being counted as adding to the snow depth.
I don't think is particularly right anyway, but nonetheless suggested fix does work. Remains that scales are different though, requiring