Modeling W44 as a supernova remnant in a density gradient, with a partially formed dense shell and thermal conduction in the hot interior. II. The hydrodynamic models

In a previous paper, we presented the analytical background for a new model for W44; in this paper, we report hydrocode experiments verifying many of the details. Our model remnant is evolving ina moderately dense (similar to 6 cm(-3)), ambient medium having a substantial density gradient. At the ob served age (similar to 20,000 yr), the shock is radiative over much of the surface, with expansion speeds of only similar to 130-200 km s(-1) for the dense and rare ends, respectively. With these speeds, the remnant has a coo l periphery and does not produce a limb-brightened X-ray image. It has ther mal conduction within its hot interior, resulting in a nonnegligible densit y there, and its center is X-ray luminous. The combined effect creates a "c enter-filled" X-ray image. A thin, very dense cool shell has formed over th e denser half of the remnant's surface, and its radio synchrotron emission derives from the highly compressed cosmic rays and swept up magnetic field, producing the usual "shell-type" image associated with radiative remnants. This combination of emission patterns results in the remnant being charact erized as having a "thermal composite morphology." Our previous paper demon strated that the intensities and qualitative distributions of the anticipat ed optical, IR line, X-ray, radio synchrotron, and gamma-ray emissions from the model are comparable to those actually observed in W44. In this paper we first use a two-dimensional hydrocode to follow the remnant evolution in a density gradient, verifying that the spatial and velocity structure of t he H I shell are a good match to the observations, without the complication s suggested by Koo & Heiles, and demonstrating that the remnant's asymmetry does not substantially affect the distribution of X-ray emitting material. We also calculate the distribution of radio-continuum emission expected fr om the compression of the ambient magnetic held and cosmic rays into the de nse shell (the van der Laan mechanism) and examine the role of surface crin kling in generating filamentation. The results are in good agreement with o bservations, though the density of ambient cosmic-ray electrons required to match the observed flux is about 4 times greater than that estimated for t he solar neighborhood. A one-dimensional hydrocode model was then used to e xplore the effects of nonequilibrium ionization on the X-ray spectrum and i ntensity. Our model is similar to but more comprehensive than the recent on e by Harrus et at, and, because their model lacked thermal conduction, ours is more successful in providing the thermal X-rays from the hot interior, including a better match to the spectrum. Neither provides the sharpness of the central peaking of the X-ray distribution without further complication s. (For that matter, however, the earlier models with evaporating clouds di d no better.) The paper closes with a discussion of the OH masers associate d with W44, proposing a specific site for their origin, in the dense traili ng edge of the cooling region following the radiative shock.