Samantha Arundel at USGS NGTOC (Rolla, MO office) has been a lead on this effort.  Inserted is a "nutshell" summary of her most recent approach to doing this ("Sam_z-values-reqs" doc):

 

 

Pete Steeves (USGS New England Water Science Center) has been corresponding with Sam and others off and on for several years on this topic.  Inserted is an old approach he took, which utilized dynamic segmentation in aml.  Much of Pete's focus was maintaining down gradient slope at confluences: