NavList:

I am guessing people will suggest a linear fit. A linear fit for some part of the curve is not too bad.

16.5 - M*0.12  for M(onth) = 1 through 6

(though  this gives a bit too high estimate for Jan  16.38 vice 16.27.  Memorize this "special case" ?)

and for July (M=7) a bit low 15.66 vice 15.73

For month's past July use the value for 14-M.  (this could be tweaked.)

If you want to interpolate for days, subtract 0.12/30=0.004 per day. (Maybe adjust for the nearest week

i.e. use -0.03 per week is more appropriate?)