A comet ionospheric model assuming the plasma to move radially outward with the same bulk speed as the neutral gas and not being subject to severe reduction through dissociative recombination has previously been tested in a series of case studies associated with the Rosetta mission at comet 67P/Churyumov-Gerasimenko. It has been found that at low activity and within several tens of km from the nucleus such models (which originally were developed for such conditions) generally work well in reproducing observed electron number densities, in particular when plasma production through both photoionization and electron-impact ionization is taken into account. Near perihelion, case studies have, on the contrary, showed that applying similar assumptions overestimates the observed electron number densities at the location of Rosetta. Here we compare ROSINA/COPS driven model results with RPC/MIP derived electron number densities for an extended time period (2015 November through 2016 March) during the post-perihelion phase with southern summer/spring. We observe a gradual transition from a state when the model grossly overestimates (by more than a factor of 10) the observations to being in reasonable agreement during 2016 March.