scaled to produce the same total production rate. Several intermediate cases were
also tested. The results showed that for cases where only 20% of the surface was
actively emitting, the flow field was essentially indistinguishable from the
insolation-driven case. Only when the total area contributing to activity was reduced
to the minimum (2% of the surface) did the flow field become clearly different.
These tests were performed for a low production case (1.25 kg s
À1 ) and obviously
as the production rate is increased the mean free path is reduced (Fig. 3.19). For 67P
close to perihelion, values in the range 1–5 m would be expected. For these higher
production rates, the reduction in the Knudsen penetration number leads to a more
radial outflow but, even here, the strong variations in production close to the
terminators (where the mean free path rises sharply) and possible deviations from
insolation-driven production at scales larger than the mean free path, will result in
complex flow geometries leading to degeneracy (Marschall et al. 2020).
It is to be noted that several authors have mapped gas emissions using ROSINA/
COPS and ROSINA/DFMS data under the assumption that the activity distribution
is solely a function of the solar insolation so that observations in four dimensional
space and time can be related by simple functions (e.g. Combi et al. 2020). However,
the evidence of transient ice patches on the surface (e.g. Fig. 2.106) and large
transient dust jets (see below) implies that this assumption must break down at
some level. At present, however, that level is not well defined.
3.4.6 Re-Condensation and Surface Reflection of Gas
Molecules
As we have seen, gas molecules can be transported from their source on the dayside
to the nightside. Rubin et al. (2014a) also showed this and concluded that condensation onto the nightside surface could occur producing surface frosts because the
nightside acts as a regional cold trap. This would result in bright frosts being evident
on the surface at the morning terminator for example.
Condensation can also occur in shadowed areas on the dayside. The temperature
of shadowed surfaces will be low because of the low thermal inertia and the absence
of any warming atmosphere. As we have seen in Fig. 2.66, decreases in the surface
temperature upon transition from a surface being illuminated to entering shadow can
be rapid and large. Hence, molecules returning to the surface will have a good
chance of sticking to the surface on contact once the temperature is significantly
below 200 K. Even under fairly rarefied conditions, collisions of molecules in the
source regions are sufficient to generate a backflux of molecules as can be seen from
the velocity distribution functions (Fig. 3.21; Table 3.7). Hence, shadowed areas can
also act as local cold traps for emitted gas and the production of visible surface frosts.
Evidence for condensation has been found in OSIRIS images of 67P. Two
examples are given in Figs. 3.34 and 3.35. Figure 3.34 shows an extensive condensate close to a shadowed area in the Hapi region. Hapi was the most active region on
3.4 Gas Expansion
241
also tested. The results showed that for cases where only 20% of the surface was
actively emitting, the flow field was essentially indistinguishable from the
insolation-driven case. Only when the total area contributing to activity was reduced
to the minimum (2% of the surface) did the flow field become clearly different.
These tests were performed for a low production case (1.25 kg s
À1 ) and obviously
as the production rate is increased the mean free path is reduced (Fig. 3.19). For 67P
close to perihelion, values in the range 1–5 m would be expected. For these higher
production rates, the reduction in the Knudsen penetration number leads to a more
radial outflow but, even here, the strong variations in production close to the
terminators (where the mean free path rises sharply) and possible deviations from
insolation-driven production at scales larger than the mean free path, will result in
complex flow geometries leading to degeneracy (Marschall et al. 2020).
It is to be noted that several authors have mapped gas emissions using ROSINA/
COPS and ROSINA/DFMS data under the assumption that the activity distribution
is solely a function of the solar insolation so that observations in four dimensional
space and time can be related by simple functions (e.g. Combi et al. 2020). However,
the evidence of transient ice patches on the surface (e.g. Fig. 2.106) and large
transient dust jets (see below) implies that this assumption must break down at
some level. At present, however, that level is not well defined.
3.4.6 Re-Condensation and Surface Reflection of Gas
Molecules
As we have seen, gas molecules can be transported from their source on the dayside
to the nightside. Rubin et al. (2014a) also showed this and concluded that condensation onto the nightside surface could occur producing surface frosts because the
nightside acts as a regional cold trap. This would result in bright frosts being evident
on the surface at the morning terminator for example.
Condensation can also occur in shadowed areas on the dayside. The temperature
of shadowed surfaces will be low because of the low thermal inertia and the absence
of any warming atmosphere. As we have seen in Fig. 2.66, decreases in the surface
temperature upon transition from a surface being illuminated to entering shadow can
be rapid and large. Hence, molecules returning to the surface will have a good
chance of sticking to the surface on contact once the temperature is significantly
below 200 K. Even under fairly rarefied conditions, collisions of molecules in the
source regions are sufficient to generate a backflux of molecules as can be seen from
the velocity distribution functions (Fig. 3.21; Table 3.7). Hence, shadowed areas can
also act as local cold traps for emitted gas and the production of visible surface frosts.
Evidence for condensation has been found in OSIRIS images of 67P. Two
examples are given in Figs. 3.34 and 3.35. Figure 3.34 shows an extensive condensate close to a shadowed area in the Hapi region. Hapi was the most active region on
3.4 Gas Expansion
241