The target of the NASA DIXI mission, comet 103P/Hartley 2 is in a non-principal-axis (NPA) rotational state that was observed to be varying with time. However, there is still no consensus on the specific NPA rotational state during the 103P/Hartley 2 encounter by DIXI. We analyze the continuum features observed in ground-based images bracketing the time of the DIXI encounter taken from a number of observatories located at different time zones to provide a dense temporal coverage. The continuum features are present at nearly all rotational phases (for the published periodicities of 18-hours and 55-hours). Therefore, we suggest the following possibilities for the origin of the continuum features and the implications. (a) If the features are originating from fixed source regions on the nucleus, short-axis-mode NPA states are more likely than the long-axis-mode NPA states, as the position angles of the observed features are restricted to a small range ( 65 degrees) in the sunlit side. (b) If the features are due to the aggregate effect of the sunward emission from the nucleus, then the continuum features cannot be used to exclude a particular mode of NPA rotation. However, the correct rotational model together with the activity model should be able to reproduce the observed coma features as a function of time. Dynamically, the moments of inertia of the near-prolate nucleus based on the shape model place severe constraints on the range of short-axis-mode states that are feasible. We will explore how non-uniform density distributions would affect the plausible NPA states and how they contrast with the case of uniform density. We will discuss both our analysis of the multi-observatory continuum images as well as the implications of non-uniform density distributions on the rotational state.