I saw some models (ACCES1-0, GFDL-CM3, CCSM4) of CMIP5 for the specific humidity variable have undefined values at the level of 1000hPa on the continents. Why does this happen? Thank you.
I am also facing this, I thought it may be related to the interpolation method they used, like Duncan said the 1000hPa surface intersects with the land surface. However I also find some missing value over west Pacific at some time especially in Boreal winter. showed is the monthly mean specific humidity at 1000 hPa at 1986.January.
It is still almost certainly the 1000 hPa surface intersecting with the land/ocean surface as this is the time when the Australian monsoon is at its peak (boreal winter / austral summer). The region is almost entirely in the Southern Hemisphere too, which is again consistent with the above.