Hi
not knowing what region you are simulating and what the timeseries is taken from (a single grid cell in an urban area, regional average,...) it is not possible to see whether something is wrong or not. Assuming your near-zero values (or are they exactly zero?) are at nighttime, it might be worthwhile checking the NOx values. Possible you have very strong titration happening? And also check what your boundary layer does at night.
One other thing to check is to include emissions in your output files to ensure they are all read in.
Gabi