Add a "ptop" argument to delta_pressure and associated masking

[kafitzgerald]

Describe the feature you'd like added to this project
It'd be good to have "ptop" (top of column pressure) as a separate argument for delta_pressure so we could compute the top pressure level similarly to dpres_plevel in NCL and do any associated masking of upper levels as needed.

Currently ptop is inferred from the pressure_lev array, which restricts how the top level thickness is defined and calculated.

Link to original NCL functionality, if applicable
https://www.ncl.ucar.edu/Document/Functions/Built-in/dpres_plevel.shtml