Cassini UVIS Observations:
Structure and History of
Saturn’s Rings
Larry Esposito
Significant Events from UVIS
Occultations Show Aggregaates
• 13 events found in 44 cuts of F ring
• Widths: 27m to 9km
• Pywacket must be elongated: alpha Sco B
offset by 600m from alpha Sco A
• At least one, and perhaps several may be
opaque
• Consistent with predictions by Cuzzi and
Burns and by Barbara and Esposito
Butterball
Fluffy
Broad and narrow
features in gamma Arae
7
Feature 9
8
Are ancient rings possible?
Regolith model for pollution:
Consider an infinite slab of depth, D
The regolith depth at time t: h(t)
For a moonlet or ring particle, D corresponds to the diameter.
Realistic case for Saturn
• Use Cuzzi and Estrada (1998) impactor size
distribution
• Compare to Quaide and Oberbeck (1975)
lunar regolith model
• Our Markov chain model result gives depth
within a factor of 2 of their values for 105 <
t < 109 years
This implies young rings?
The fractional pollution of the regolith, fp, is
given by
fp =
FG mÝt / 
h(t)
For ring particles, fp is 0.01 in 108 years, a
rough upper limit from ring observations at
microwave

Estimating ring age from the
volume pollution rate
For a ring system with surface mass density,
, we have
fp(vol) =
8
2
10 g /cm / year t

So, fp(vol) = 0.01 and  = 100g/cm2 also
gives t=108 years, consistent with CE98

Ring age
• Ring particles should develop reflectance spectra
showing 1% pollution in 108 years: this is a
problem even if the rings formed a few hundred
million years ago!
• An alternate conclusion is that the ring mass is
underestimated: then, continuing recycling could
renew their composition, consistent with upper
limits on non-icy material
• Clumpy rings mean larger B ring mass, consistent
with less pollution
Numerical simulations show collisions and self-gravity
effects will create transient elongated trailing structures.
Pollution Conclusions
• Our stochastic regolith model shows the
surfaces of ring particles could be 1%
meteoritic after only 108 years
• If rings recycle ejecta, the volume pollution
rate is the same: 1% meteoritic in 108 years
for standard values of ring mass and impact
flux
• For larger mass or lower flux, the ring age
could be proportionately greater: 109 years
if B ring  = 1000 g/cm2
Numerical simulations show a spider-web structure.
The ring opacity underestimates mass of the B ring!
UVIS occs inverted
2D autocorrelation => FFT => image
Reinterpretation of P-11 results:
Colwell’s ‘Granola Bar’ Model
• Cooper etal 1983 assumed a uniform ring to
calculate secondary fluxes from GCR flux
• But, secondary fluxes are double-valued!
• Self-gravity wakes say B ring density is also…
• Instead, assume surface density ring C = ring A =
60g/cm2. For ring B: 80% with 500g/cm2, 20%
with 60g/cm2. These are reasonable values, no free
parameters varied!
• This yields predicted fluxes (Cooper Fig 5):
– Protons 40 (measured 50 +/- 11)
– Gammas 118 (measured 180 +/- 45) in ( m2-sec-ster)-1
UVIS spectrum as a function of age and grain size
Comparison of numerical models with different time steps
Regolith depth simulation data
CE98 Distribution / not extended
Attribute
Diameter
amax
Squarings
10x deltaT
years
Value
91.98269
2.3585305
54
23.64365713
1.35E+10
Units
cm
cm
#
sec
years
Non Zero Probabilities (first row)
Location
Probability
0
1
1
4.67E-13
10
4.11E-14
61
4.77E-15
A plausible ring history
• Interactions between ring particles create
temporary aggregations: wakes, clumps, moonlets
• Some grow through fortunate random events that
compress, melt or rearrange their elements.
Stronger, more compact objects would survive
• Growth rates require only doubling in 105 years
• Ongoing recycling resets clocks and reconciles
youthful features (size, color, embedded moons)
with ancient rings: rings will be around a long time
Backup Slides
Markov chain simulation matches
analytic test case
For an impactor size distribution that is a
power law of index 3, we can solve the
differential equation for h(t), assuming all
material is excavated:
h(t) = Hmax[1 - exp(-t/T0)]
Hmax = H1 amax
T0 =
H max
Ý/ 
FGYm
UVIS HSP 2D autocorrelation