Particle Sizes and Clumps from Stellar Occultations

Transcription

1 Particle Sizes and Clumps from Stellar Occultations Josh Colwell, Richard Jerousek, and James Cooney (UCF) Larry Esposito and Miodrag Sremcevic (CU) UVIS Team Meeting June 4-6, 2013, St. George Utah

2 Variance in Occultation Data Related to Particle Sizes For Poisson-distributed data, the variance equals the mean: σ 2 = I Finite size of ring particles compared to the UVIS HSP field of view results in higher variance. Get particle autocorrelation length from excess variance. Technique pioneered by Showalter and Nicholson (1990) with Voyager PPS data. Cassini UVIS data offer higher signal and spatial resolution. UVIS Team Meeting June 4-6, 2013, St. George Utah 2

3 If the rings were a translucent screen like this Then the variance = mean (Poisson statistics) UVIS Team Meeting June 4-6, 2013, St. George Utah 3

4 ...if they look like this, the variance is small. UVIS Team Meeting June 4-6, 2013, St. George Utah 4

5 This simulated data from the previous particle model shows variance equal to the mean, as expected for Poisson counting statistics. (The discretization of the simulated data reflects the simulation of the Cassini UVIS compression algorithm.) DPS October 19, 2012, Reno NV 5

6 If the rings look like this, the variance is large. UVIS Team Meeting June 4-6, 2013, St. George Utah 6

7 This simulated data from the previous particle model shows variance in excess of Poisson counting statistics. DPS October 19, 2012, Reno NV 7

8 Particle Autocorrelation Length R eff = AµΔσ 2 π I 0 2 e τ /µ (1 e τ /µ ) (Esposito and Sremcevic) A=area of region in rings sampled by a single measurement I0=unocculted stellar signal τ=measured optical depth Δσ 2 =excess variance Δσ 2 = π I 2 e τ /µ (1 e τ /µ ) 0 Aµ R eff UVIS Team Meeting June 4-6, 2013, St. George Utah 8

9 Particle Autocorrelation Length R eff = AµΔσ 2 µ π I 2 e 2τ /µ τ 0 (Showalter and Nicholson (1990), assuming large particles) Gives the following relation for excess variance as a function of optical depth for a given Reff: Δσ 2 = π I 2 e 2τ /µ 0 Aµ τ µ R eff UVIS Team Meeting June 4-6, 2013, St. George Utah 9

10 Excess Variance Dependence on Optical Depth UVIS Team Meeting June 4-6, 2013, St. George Utah 10

11 Excess variance is larger for larger characteristic particle sizes. DPS October 19, 2012, Reno NV 11

12 C ring BetCen77_1 The data show multiple distinct populations: points do not follow a single-valued curve. A ring B ring

13 DPS October 19, 2012, Reno NV 13

14 DPS October 19, 2012, Reno NV 14

15 Halo Signature in Particle Autocorrelation Length DPS October 19, 2012, Reno NV 15

16 ! DPS October 19, 2012, Reno NV 16

17 DPS October 19, 2012, Reno NV 17

18 Summary Significant changes in particle autocorrelation length seen at some major ring boundaries as well as at locations with no obvious structural change. Evidence for anti-correlation of autocorrelation length with optical depth in strong density waves. Decreased autocorrelation length in density wave halo regions. DPS October 19, 2012, Reno NV 18

19 VIMS/UVIS Self Gravity Wake Results

20 Self-Gravity Wakes Q c rκ Gσ 1 λ crit = 4π 2 Gσ /κ m UVIS Team Meeting June 4-6, 2013, St. George Utah 20

21 Self-Gravity Wake Properties UVIS Team Meeting June 4-6, 2013, St. George Utah 21

22 Granola Bar Model (Colwell et al.) Pasta Model (Hedman et al.) UVIS Team Meeting June 4-6, 2013, St. George Utah 22

23 S- G Wake Status Previously found discrepant results fi?ng VIMS- only or UVIS- only data to S- G wake model, regardless of model. Differences between VIMS and UVIS likely due to absolute differences in ophcal depths due to diffrachon in VIMS observahons. 23

24 VIMS Occultations Star is placed in one pixel 0.25 by 0.5 mrad Light curves based on 2.9 micron channel (low background) At that wavelength, some light is diffracted out of the pixel by particles smaller than 1.2 cm in diameter. If small particles are present, VIMS will lose more light than UVIS does and will report a higher optical depth, skewing S-G wake parameters. UVIS Team Meeting June 4-6, 2013, St. George Utah 24

25 New S-G Wake Model Unless occultations are at identical geometries between the two instruments, cannot extract particle size information from direct comparison because S-G wakes affect observed optical depth. Use differences in VIMS and UVIS wavelengths to extract information on population of small particles via S-G wake model. Fractional optical depth due to small particles in gaps between S-G wakes is new parameter for S-G wake model. UVIS Team Meeting June 4-6, 2013, St. George Utah 25

26 UVIS Team Meeting June 4-6, 2013, St. George Utah 26

27 UVIS Team Meeting June 4-6, 2013, St. George Utah 27

28 UVIS Team Meeting June 4-6, 2013, St. George Utah 28

29 UVIS Team Meeting June 4-6, 2013, St. George Utah 29

30 UVIS Team Meeting June 4-6, 2013, St. George Utah 30

31 UVIS Team Meeting June 4-6, 2013, St. George Utah 31

32 Derived "small particle" optical depth from UVIS/VIMS joint fit to S-G wakes UVIS Team Meeting June 4-6, 2013, St. George Utah 32

33 Summary Get direct measure of sub-cm population with handful of Vega occultations. Get additional information on small particles in S-G wakes from new S-G wake model: preliminary results show increase abundance of sub-cm particles in outer half of A ring. See also next talk by Becker. Occultation statistics show decrease in particle autocorrelation length in perturbed regions, including haloes. UVIS Team Meeting June 4-6, 2013, St. George Utah 33