1. Enceladus: UVIS Constraints and Modeling
C. J. Hansen, L. Esposito, D. Shemansky, I. Stewart, A. Hendrix
19 July 2010

2. All plume models must ultimately be consistent with data
Outline
All plume models must ultimately be consistent with data
UVIS Observations
What we observe: Occultations
How we observe: Instrument
Plume Results
Composition
Mass flux
Temporal variability
Gas Jets
Structure
Mach number
Plume
Jets

3. UVIS Observations of Enceladus’ Plume
UVIS observes occultations of stars and the sun to probe Enceladus’ plume
Composition, mass flux, and plume and jet structure
Three stellar and one solar occultation observed to-date
Feb lambda Sco
No detection (equatorial)
July gamma Orionis
Composition, mass flux
Oct zeta Orionis
Gas jets
May Sun
Composition, jets
What do we observe

4. The Occultation Collection
gamma Orionis Occultation
The Occultation Collection
zeta Orionis Occultation
Solar Occultation

5. UVIS Characteristics UVIS has 4 separate channels
For stellar occultations we use:
Far UltraViolet (FUV)
1115 to 1915 Å
2D detector: spectral x 64 one-mrad spatial pixels
Binned to 512 spectral elements
5 sec integration time
High Speed Photometer (HSP)
2 or 8 msec time resolution
Sensitive to 1140 to 1915 Å
Hydrogen-Deuterium Absorption Cell (HDAC) not used
For the solar occultation we used:
Extreme UltraViolet (EUV) solar port
550 to 1100 Å
2D detector: spectral x 64 one-mrad spatial pixels
No spatial information because signal from sun is spread across the detector (deliberately)
Spatial rows binned to two windows of 27 rows each
1 sec integration

6. Outline UVIS Observations Plume Results Gas Jets Occultations
Instrument
Plume Results
Composition
Mass flux
Temporal variability
Gas Jets
Structure
Mach number

7. 2007 - Zeta Orionis Occultation (Alnitak)
gamma Orionis Occultation
(Bellatrix)
Key results:
Dominant composition = water vapor
Plume column density = 1.6 x 1016 /cm2
Water vapor flux ~ 180 kg/sec
Results documented in Science, 2006
Vertical cut through plume
Zeta Orionis Occultation (Alnitak)
Key results:
Average column density = 1.5 x 1016 cm-2
Max column density = 3.0 x 1016 cm-2
Gas jets detected correspond to dust jets
Results documented in Nature, 2008
Horizontal density profile
18 May Solar Occultation
Two science objectives enabled by solar (rather than stellar) occultation:
1. Composition of the plume
New wavelength range: EUV
2. Structure of the jets and plume
Higher time resolution 7

8. Plume: FUV Results from Stellar Occ’s
Absorption is best fit by water vapor
2007 best fit average column density = 1.5 x 1016 cm-2
Error bar: +/- 1.4 x 1015 cm-2
Comparison to 2005 at 15 km altitude
2007 peak column density
= 3.0 x 1016 cm-2
2005 = 1.6 x 1016 cm-2
No detection of CO
formal 2-σ upper limit is 3.6 x 1014 cm-2
corresponds to mixing ratio with H2O of 3.0
Our nondetection of CO excludes 3% CO in the plume at the 2 sigma level
I/I0
Used Mota (2005) cross-sections

9. New EUV Spectrum from Solar Occultation
Navy is unocculted solar spectrum, with typical solar emissions
Red is solar spectrum attenuated by Enceladus’ plume

10. Composition
H2O and N2 have diagnostic absorption features at EUV wavelengths
The primary goal was to look for N2, on basis of INMS detecting a species with amu=28
No N2 (upper limit = 3 x 1013, so < 0.3%)
AMU = 28 detected by INMS is not N2
It is not CO in the plume (or UVIS would have seen it in our stellar occs) or it is < 3%
Maybe very little signal at amu=28 to worry about after all per latest INMS results (personal communication, Waite 2010)
N2 < 0.9%
CO < 0.9%
C2H4 < 0.16%

11. Nitrogen feature at 97.2 nm not detected
Actual
No dip is seen at all at 97.2 nm
Upper limit < 0.3%
Consequences of no N2 for models of the interior
High temperature liquid not required for dissociation of NH3 (if there is NH3 in the plume)
Percolation of H2O and NH3 through hot rock is not required
A catalyst for decomposition is not required at lower temperatures
Clathrate decomposition is not substantiated for N2 as the plume propellant
Predict
N2 feature at 97.2 nm fortuitously coincides with strong lyman gamma emission so lots of signal available
Very sensitive test!

12. Solar Occ results: Water in the Plume
H20 fit to absorption spectrum
Column density of H2O = 0.9 x 1016 cm-2

13. Water Vapor Abundance
To calculate water vapor abundance in the plume the spectra are summed during the center 60 sec of the occultation, then divided by a 650 sec average unocculted sum to compute I / I0
I0 computed at two different times, results were the same
The extinction spectrum is well-matched by a water vapor spectrum with column density = 0.9 +/- 0.1 x 1016 cm-2
Overall amount of water vapor is comparable to previous two (stellar) occultations
2005: 1.6 x 1016 cm-2
2007: 1.5 x 1016 cm-2 (maximum value of 3.0 x 1016 cm-2 at center)
Lower value in 2010 may be at least partially attributable to the viewing geometry

14. Ground Tracks Blue ground track is from zeta Ori occ on Rev 51
Orange is solar occ track, ~orthogonal
Plume is elongated.
If total flux is same then column density will be less by ~ 2/3
Preliminary result:
0.9 x1016 cm-2, ~2/3 of value in 2005
 Ingress
 Egress

15. Water Vapor and other Gases
Some evidence that the mixing ratio of water may change depending on where you are in the plume

16. Solar Occultation Characteristics
FWHM
Total duration of Solar Occ:
2min 19sec
Duration for full-width half max:
56 sec
Line of sight velocity: km/sec
Width of plume at FWHM:
56 sec * 2.85 = 160 km
Compare to zeta Orionis Occ
Zeta Orionis occultation lasted just 10 sec
Line of sight velocity = 22.5 km/sec
Width of plume at FWHM = 110 km
HSP data summed to 200 msec so 50 samples
Zeta Orionis occultation

17. Estimate of Water Flux from Enceladus = 200 kg/sec
S = flux
= N * x * y * vth
= (n/x) * x * y * vth
= n * y * vth
Where
N = number density / cm3
x * y = area
y = vlos * t => FWHM
vth = thermal velocity = 45,000 cm/sec
for T = 170K
n = column density measured by UVIS
note that escape velocity = 23,000 cm/sec x v
Year n
(cm-2) y
(x 105 cm)
vth
(cm / sec)
Flux: Molecules / sec
Flux:
Kg/sec
2005
1.6 x 1016
80 (est.)
45000
5.8 x 1027
170
2007
1.5 x 1016
110
7.4 x 1027
220
2010
0.9 x 1016
160
6.5 x 1027 17

18. One more comparison to tidal energy model
Position of Enceladus in its orbit at times of stellar occultations, and solar occultation
Hurford et al 2007 model predicts tidally-controlled differences in eruption activity as a function of where Enceladus is in its eccentric orbit
Expect fissures to open and close
Substantial changes are not seen in the occultation data, although they would be predicted, based on this model
2010 column density very similar to 2005
Taken from Hurford et al,
Nature 447:292 (2007)
True Anomaly
(deg)
Fraction of orbit from Periapsis
Position in Orbit
Stress
105 Pa
0.0
Periapsis
0.3
0.186
May 18, 2010 90
0.25
One quarter
-0.8
97.76
0.27
July 14, 2005
-0.77
180
0.5
Apoapsis
-0.4
254.13
0.7
October 24, 2007
0.4
270
0.75
Three quarter
0.6

19. Outline UVIS Observations Plume Results Gas Jets Occultations
Instrument
Plume Results
Composition
Mass flux
Temporal variability
Gas Jets
Structure
Mach number

20. Detection of Gas Jets
Perfect geometry to get a horizontal cut through the plume and detect density variations indicative of gas jets 20

21. Zeta Orionis Occultation
Density in jets is twice the background plume
Gas jet typical width = 10 km at 15 km altitude
Egress
Ingress
Because the dimension of the two prominents jets is about 1/8 the width of the largest plume, their gas density needs to be twice that of the surrounding plume at 15 km to yield the 12% opacity increase seen in the jets.
a. Cairo (V)
d. Damascus (II)
c. Baghdad (VI)
b. Alexandria (IV)
Closest point 21

22. 2007 - Plume Structure and Jets
Summary of 2007 results
Significant events are likely gas jets
UVIS-observed gas jets correlate with dust jets in images
Characterize jet widths, opacity, density
Density in jets ~2x density in background plume
Ratio of vertical velocity to bulk velocity = 1.5, supersonic
Supersonic gas jets are consistent with Schmidt et al. model of nozzle-accelerated gas coming from liquid water reservoir

23. Solar Occ Jet Identificationsb c d e f
Minimum altitude
Window 0 and 1 matching features => jets
Repetition of features in window 0 and window 1 shows they are not due to shot noise, therefore likely to be real

24. Jets vs. Tiger Stripes As before, gas jets correlate to dust jets
Higher time resolution because sun’s passage behind the plume was slower
We see gas between jets, although no enhancement as a function of depth in the plume (time at which sun was closest to limb is not a deep attenuation)
Spacecraft viewed sun from this side
Ingress 
Minimum Altitude
Egress

25. Jet Identities Feature Altitude (km) Dust Jet a 20 19.7 b 21 c 27 d 30
Alexandria IV
Closest approach
19.7 b 21
Cairo V and/or VIII c 27
Baghdad I d 30
Baghdad VII e 38
Damascus III f 46
Damascus II

26. Gas Velocity w/h = tan ß = vthermal / vvert = 7 / 27
In the case of jet c (Baghdad I) the source of the jet is close enough to the occultation ground track to calculate the height and width
Half-width of jet c = 2.5 sec * 2.85 km/sec = ~7 km (w) at ~27 km (h) altitude w h
w/h = tan ß = vthermal / vvert = 7 / 27
Mach number = vvert / vthermal = 3.9
Previously estimated mach number (from 2007 occultation) was 1.5
Jets more collimated than previously estimated
New estimate for vertical velocity: if vthermal = 450 m/sec (for ~170 K) then vvert = 1755 m/sec

27. Our maturing understanding…
Composition
Upper limit on N2 of 3 x 1013 cm-2
H20 column density = 0.9 x 1016 cm-2
In family with previous occultations
Precisely corresponds to aspect ratio applied to 2005, which was at a similar true anomaly
Suggests that Enceladus has been steadily erupting for past 5 years
Plume / jet structure
Flux of water from 3 occultations is ~200 kg/sec
Jets are more collimated than estimated from 2007 occultation -> higher mach number of ~3.9

28. Backup Info

29. Occultation is clearly visible
Window 0 has higher counts, but overall shape is the same
Position of sun was slightly offset from center, but not an issue
Observation start time: T05:51:44.45
Observation end time: T06:10:36.45
Ingress: T06:00:40.45
Egress: T06:02:59.45
Velocity of sun across plane of sky ~ 2.75 km/sec
Data shown is summed over wavelength

30. Altitude of Sun above Limbc d a b
Perpendicular distance of ray connecting UVIS and the sun from the limb of Enceladus
Attenuation at a and b bracket closest approach

31. Ethylene at 3% H2O Column Density compared to H2O only
Rev 11 gamma Orionis occultation
Ethylene plus water compared to water only
C2H4 column density = 4.8 x 1014 cm-2
H2O column density = 1.6 x 1016 cm-2
Water only is still best fit to occulted spectrum although there are some interesting matches to small dips with ethylene added in

32. Groundtrack of Occultation
Enhanced HSP absorption features a, b, c, and d can be mapped to dust jets (roman numerals) located by Spitale and Porco (2007) along the tiger stripes
Take time to explain this!
Blue line is groundtrack

33. Gas Absorption Features, Compared to Dust Jet Locations
Plotted here are:
Altitude above Enceladus' limb of the line-of-sight from Cassini to the star
Attenuation of the HSP signal, scaled by a factor of 300
Projections of the 8 jets seen by the ISS into the plane of the figure
Jets assigned a length of 50 km (for purposes of illustration)
C/A marks the closest approach of the line-of-sight to Enceladus.
The times and positions at which the line-of-sight intersected the centerlines of the jets are marked by squares.
The slant of the jets at Baghdad (VII) and Damascus (III) contribute to the overall width of the plume 33

34. Gas Jet Model Key Result: Vvert / Vthermal = 1.5 Flow is supersonic
Gas Jets are idealized as sources along the line of sight with thermal and vertical velocity components
Source strength is varied to match the absorption profile.
The ratio of thermal velocity (vt) to vertical velocity (vb) is optimal at vt / vb = 0.65.
Higher thermal velocities would cause the streams to smear together and the HSP would not distinguish the two deepest absorptions as separate events.
The two deepest absorptions indicate jets are collimated, 10 km wide at 15 km altitude
In the plot the gas streams are shown as dashed vertical lines.
Key Result:
Vvert / Vthermal = 1.5
Flow is supersonic

35. Groundtrack of Ray
2005
2007

36. 2005 Jets Jets mapped to increases in opacity
In this occ we do not see B7 (star is occulted by limb before crossing B7)
Is it OK to compare 2005 and 2007?
IF individual jets are only source of plume then no
If gas from entire tiger stripe probably ok 36

37. Compare 2007 to 2005 - HSP 2005 attenuation <6% at 15 km
2007 attenuation at same altitude ~10%
Overall attenuation clearly higher in 2007 compared to 2005
The ratio of the opacity from 16 to 22 km between 2007 and 2005 is 1.4 +/- 0.4

38. Summary of Results JETS:
HSP data shows 4 features with m < 0.1 (probability of chance occurrence). Typical half-width: 10 km at z = 15 km.
Gas jets can be correlated with dust jets mapped in images on Cairo, Alexandria, Damascus and Baghdad tiger stripes
Jet opacity corresponds to vapor density doubled within jets
Alternate explanation: no excess gas, with all increase due to dust. Then, dust opacity peaks at 0.05 in the jets. This would give 50x more mass in dust compared to vapor within the jet.
Ratio of vertical velocity to thermal velocity in jet = 1.5
Gas is supersonic
Eight or more jets required to reproduce width and shape of absorption
Jet source is approximately 300 m x 300 m
Example Calculations
T surface = 140 K
V thermal = 359 m/sec
V vertical = 552 m/sec
For Tsurface = 180 K (from CIRS)
V thermal = 406 m/sec
V vertical = 624 m/sec

39. Summary of Results PLUME: H2O column density in 2007 = 1.5 x 1016 cm-2
Density at 15 km altitude 2x higher
H2O column density in 2007 ~ 3.0 x 1016 cm-2
Attenuation in HSP data ~10% in 2007, ~6% in 2005
Difference contradicts Hurford et al model of fissures opening and closing
Plume column density goes as ~ z (z is minimum rayheight)
Water vapor flux ~200 kg/sec
No detection of CO, ethylene not definitive

40. High Speed Photometer (HSP) Data
HSP is sensitive to 1140 to 1900Å
Statistical analysis applied to find features that are probably real
Assumes signal is Poisson distribution
Calculate running mean
Six different bin sizes employed, absorptions compared, persistence of feature is part of test
m is the number of such events one would expect to occur by chance in the data set
m<<1 are likely to be real events
Possible real features:
1 (a) m = 0.032
2 (b) m =
3 (c) m =
6 (d) m = 0.026
Statistical analysis is applied to determine which features are likely to be real. The analysis assumes the signal is a Poisson distribution. A running mean is calculated over 1001 integration periods or binned equivalent time [6]. The mean is the baseline for point i. Six different bin sizes were employed. C is the binned stellar signal at a particular bin. The probability that the signal would be ≤C at that bin is given by the sum of the distribution over values ≤C. This step is performed for each bin, i, in the data set to find Pi = P(mi, ≤Ci). Pi is multiplied by the number of bins in the data set, N. This product gives us a value m, where m=NPi [7], or the number of such events one would expect to occur by chance in the data set. 40

41. Water Column Density: FUV comparison to HSP 2007
FUV integrations are 5 sec duration
FUV spectrum shows gas absorption in Rev 51 time records 89 and 90
Higher time resolution of HSP data shows that the peak column density is about 2x higher than the 5 sec average calculated from the FUV data
FUV time record 89
FUV time record 90

42. Plume Composition and Column Density
UVIS Ultraviolet Spectra provide constraints on:
Composition, from absorption features
Column density
Mass Flux
Plume and jet structure
Terminology:
Plume - large body of gas and particles
Jets - individual collimated streams of gas and particles
Plume
Jets
What do we learn

43. Zeta Orionis Occultation 2007 (Alnitak)
gamma Orionis Occultation
(Bellatrix)
Vertical cut through plume
True anomaly = 98
Key results:
Dominant composition = water vapor
Plume column density = 1.6 x 1016 /cm2
Water vapor flux ~ 180 kg/sec
Documented in Science, Hansen et al, 2006
Zeta Orionis Occultation 2007 (Alnitak)
Key results:
Average plume column density = 1.5 x 1016 cm-2
Maximum plume column density = 3.0 x 1016 cm-2
Gas jets detected correspond to dust jets
Horizontal density profile
True anomaly = 254
Results documented in Nature, 2008 43

44. Rev 11 Gamma Orionis Occultation, FUV data
UVIS spectra, occulted and unocculted
Plot I/I0 to see absorption features
Compare I/I0 to water absorption spectrum
Water vapor: uses Mota cross-sections
Best fit column density = 1.6 x 1016 cm-2