Does burning streams introduce error

Discussion created by kear1384 on Feb 5, 2014
Latest reply on Jan 5, 2017 by Lady_Jane
I am trying to model a watershed at a finer resolution than my states NHD shows.  I understand the processes of Hydro but was wondering why it is acceptable to burn the streams into the DEM when doing so not only alters the stream section of the dem by 10M but also smooths the surrounding cells.  It seems that in doing so would introduce error from the start. 

After modeling with and without burning the NHD into the DEM there are clearly differences in the resulting stream layers.  Observation points match closer to the burned results.  Is this just the trade off?  That the total error is lessened by introducing slight error at the start?