Juno Perijove 58, February 3, 2024 Options

[mcaplinger]

(Started a new thread to avoid cluttering up the PJ57 thread with PJ58 discussion.)

[mcaplinger]

maybe they meant the area covered in Jupiter-shine? Because it would be fairly complimentary to the PJ57 coverage, getting more of the southern hemisphere coverage compared to the northern hemisphere coverage of PJ57. I'm looking forward to seeing East Kanehekili (-18/24) at visible wavelengths, as an example.

I'll do what I can. There are a whole bunch of constraints. Io goes through the FOV rapidly around 17:50 and we can only command a maximum of four images at fixed cadence before incurring a 30s delay. I can't improve the nightside coverage at the cost of degrading the dayside, the best of which happens around 17:51:30.

I'm surprised that people are getting a lot out of the TDI 2 nightside images. I guess they're better than nothing but I was expecting at least TDI 6 would be required if not more, which means the dayside would be blown out for sure.

[StargazeInWonder]

I guess to put this imagery to a rubber-meets-the-road standard, the question of how much Io is changing over the cadence of 1979-2024 (and into the future at a similar timescale) is a key parameter. If at some level of regional imagery we see a change occur and this is our one opportunity to do pin that down chronologically, that's distinctive value. If at some level of resolution Io is probably pretty much the same in 2045 as in 1979, then any image is as good as any other, and then maybe the jupitershine imagery adds little. And I'm making it sound like the change is a known quantity, but it's much more complex than that, with stochastic nature and unknowns pertaining to Io, to future missions, to its activity, all challenging and exciting. However it shakes out, hats off to those of you playing a part. People who haven't been born yet will study these images and potentially (this is partly up to Io itself) advance the science with them.

FWIW, the massive ground based telescopes now in the works will provide the potential for regular monitoring of Io's visible changes, subject to geometry, and observational time. Just a pinch of observation time, though, would easily surpass the scarcity of spacecraft missions we have had so far.

[Bjorn Jonsson]

From the PJ57 thread:

Mike, I am definitely coming around to the idea that if you can get well exposed Jupiter-shine images in the green filter, being able to stack them to improve SNR and getting more of them would be fine. While RED has the best SNR, GREEN has the best balance of SNR and albedo information.

A possible drawback is if the PJ56 problem returns at PJ58. If that happens, red images will be much better. So if possible, it might make sense to also take at least one red-filtered Jupitershine image just in case.

[Brian Swift]

From the PJ57 thread:


A possible drawback is if the PJ56 problem returns at PJ58. If that happens, red images will be much better. So if possible, it might make sense to also take at least one red-filtered Jupitershine image just in case.

NOOOOOO. PJ56 problem is not coming back before PJ112.

[mcaplinger]

A possible drawback is if the PJ56 problem returns at PJ58.

Compare the progression of black offset and noise during PJ56 (a lot) with that in PJ57 (practically none). Even if PJ58 proceeds like PJ56 did (which we hope is unlikely) most of the change would happen after the Io encounter.

We hope.

[kymani76]

Map for PJ58 Io flyby from SPK. Ground track covers the distance of +100.000/-100.000 km.

EDIT: fixed night shadow.

[MarcF]

I expect to see some beautiful images of Ra patera, my favorite volcano. Located near the terminator, we should finally have an idea of the topography of this shield volcano. And then, we'll have some incredible views of Loki. I wonder about the topography of this lava lake and especially its surroundings, which looked quite flat in one of Galileo's last images. Fingers crossed.

[mcaplinger]

My initial cut at PJ58 image timing for Io is:

2024-034T17:32:00, two images, high and low TDI, RGB
2024-034T17:39:00, two images, high and low TDI, RGB
2024-034T17:50:00, four images, first high TDI, three low TDI, RGB
2024-034T17:56:00, two images, high and low TDI, RGB

high TDI will be 6, low TDI will be 2.

All image spacing is 60 second. All compression is lossless. I thought about taking one-band images but due to various constraints I couldn't make it work without compromises.

Feel free to provide feedback. But keep in mind that the FOV of Junocam rapidly slides across Io near closest approach, so if I did something that seems odd, I might have done it intentionally.

[kymani76]

I updated the flyby map with new information above.

[mcaplinger]

I updated the flyby map with new information above.

Thanks. But without seeing where the Junocam FOV is along the ground track, it may be hard to understand why I picked the times I did. Visualizing this on a flat map is tough, as I well know. These times were a product of going back and forth in Cosmographia repeatedly, looking at the evolving FOV coverage.

[volcanopele]

A pretty decent set of observations. I might replace the second set with an extra image at the front of the third set (so roughly 2024-034T17:49:00), but I'm not sure how that works for you for bandwidth and data volume. The second set doesn't really add much to the coverage or resolution, and wouldn't be useful for something like the Masubi plume. With high TDI at 2024-034T17:49:00, should provide highest resolution coverage of Io. Covers Masubi, Kanehekili, Shamshu, and Janus. or just take the High TDI image in the second set? The main "Kodak moment" is in the first set anyway.

[mcaplinger]

A pretty decent set of observations. I might replace the second set with an extra image at the front of the third set (so roughly 2024-034T17:49:00), but I'm not sure how that works for you for bandwidth and data volume.

I'll look at it. Keep in mind that all of these times are +/- 15s and I can only take four images in a set before incurring an additional delay. I could start the third set earlier but only at the cost of dropping the last image it currently has, which I was reluctant to do. I tend to not value the nightside images that highly, but maybe that's the wrong bias to have.

[kymani76]

I made another visualisation of the flyby, including Jason's suggestion for 17:49:00.
No attempt was made to account for Juno's distance from Io, attitude or FOV.

[mcaplinger]

No attempt was made to account for Juno's distance from Io, attitude or FOV.

Pretty, but the FOV is rather important. 17:49 just doesn't have much of Io in it.

Attached thumbnail(s)

 

[volcanopele]

for 17:49 this is what I get:



This does add coverage to the west of what 17:50 can get, particularly Masubi, and at decent resolution, 1.1 km/pixel at center of disk. 17:53 doesn't seem to add much and some of that can be recovered by starting the last set a minute earlier (17:55)

I will note that the leading hemisphere is the other big gap in our map of Io where the best resolution data from Galileo/Voyager/New Horizons was ~7–8 km/pixel and this would definitely help fill that gap.

[mcaplinger]

17:53 doesn't seem to add much and some of that can be recovered by starting the last set a minute earlier (17:55)

17:53 is a backup against timing errors or possible unrecoverable loss of 17:52.

There are constraints on how close the sets can be and I don't think the last set can be moved in.

But I hear you and I'll see what I can do.

[volcanopele]

I presumed that if the third block of four gets shifted forward a minute, so starting at 17:49 instead of 17:50, the fourth block could also be moved up a minute.




Here is a set of preview images starting with the 17:49 opportunity and including 17:53. The last image is the 17:56 opportunity. There is no meaning to the difference in brightness in the Jupiter-shine areas, just me being fast and quick in Photoshop and not paying attention to what exposure setting I used.

[mcaplinger]

for 17:49 this is what I get... This does add coverage to the west of what 17:50 can get, particularly Masubi, and at decent resolution, 1.1 km/pixel at center of disk.

Except Junocam isn't seeing the center of the disk at that point. IMHO 17:49 is too oblique.

I considered bumping everything back 30 seconds so starting at 17:49:30 but that's not a clear winner to me either, thoughts?

And you have to consider that due to spin phase uncertainty the position is uncertain by a spin period.

Whatever we do, I have a soft deadline of EOD today to decide.

[volcanopele]

17:49:30 still has Masubi in the FOV but it is much closer to the limb. Still a better view than 17:50:00. and it has the benefit of having a higher resolution view of East Kanehekili (second highest priority "new volcano" target IMHO at Io from JIRAM data after Tonatiuh) that at least with the current predicts we would miss at 17:49. We would lose the best full-color dayside opportunity.... but there would still be two opportunities. I still think that 17:50 is preferable to 17:49:30 as we get full-disk color at highest resolution (given current predicts of the rotation).

Looking at 17:48:30, so looking at the worst case scenario I am not actually suggesting starting that early, Io is still in the JunoCam FOV. We lose Masubi as it isn't quite in the FOV yet, but Shamshu and Janus would be there, Janus would be a bit more marginal there. My point is that even in the worst-case scenario for the rotation phase, 17:49 would still be useful for gap filling. For added context, during IVO planning we had an entire flyby just to get this coverage.

For context, Janus is a lava lake much like Pele. In the IR, Janus looks almost identical to Pele, but has fewer volatiles (IOW, there's no giant red ring around it)

[mcaplinger]

Have you compared the amount of Jupitershine in this encounter with the previous one? It looks similar, but I was surprised reviewing the PJ51 imaging that the nightside was much dimmer than it was in PJ57.

Probably this is just a Jupiter phase and Io orbit position thing, but the tools I have to look at this are clumsy.

[volcanopele]

comparing 55C00035 with 57C00030 and 57C00022, I get roughly similar pixel values for places like around Fjorgynn Fluctus. Color still looks better for the PJ57 images. No apparent signal in BLUE in the orbit 55 images, it looks to be almost entirely noise. for PJ57, the darkest features like Guawa Patera and Fjorgynn Fluctus and the brightest terrain shows up (bright material near Gauwa and Fjorgynn, and around Acala Fluctus). I presume that this is a factor of noise, reducing effective resolution. So PJ57 data, particularly 57C00022 looks great, but the effective resolution is still reduced by a factor of 2-3 compared to the dayside. but that drop of resolution still makes it better than Galileo/Voyager data.

[mcaplinger]

Have you compared the amount of Jupitershine in this encounter with the previous one? It looks similar, but I was surprised reviewing the PJ51 imaging that the nightside was much dimmer than it was in PJ57.

I think the difference (wait for it) is that Jupiter doesn't shine on the part of Io away from Jupiter, only on the part facing Jupiter. Duh.

So one shouldn't expect Jupitershine on inbound-to-Io images, only on outbound-from-Io images.

[kymani76]

I'm not totally sure, but I believe it is possible to present above discussion in a map form.

The view above shows Io as seen from Jupiter at the time of the flyby. In other words, the center of map projection is at Io's sub-Jupiter point.
This means we are looking at the hemisphere illuminated by Jupiter-shine.

I've used this article to estimate how much of Io's disk Juno can see at selected times.
(17:49 in red, 17:49:30 in green; 17:50 in blue). I used values of 0.25, 0,265 and 0.28 respectively for f to account for the rising altitude, giving me range circles for parts of Io visible to Juno at these times.
You can really see how much the coverage changes in very short time.

I still have to account for Junocam FOV as I suspect not the whole visible disk fits into the camera view, but I haven't figured out how to account for that yet.

And yes, there is no Jupiter-shine on the inbound track.

[mcaplinger]

I still have to account for Junocam FOV as I suspect not the whole visible disk fits into the camera view, but I haven't figured out how to account for that yet.

It's not easy. To do it completely correctly, you need to know the pointing of the spacecraft as a function of time. For many of these flybys, including this one, the spacecraft spin axis (Z) is pointed at Earth, which makes it a little easier. So the Junocam FOV is a locus roughly +/- 30 degrees from the XY plane and spun about the Z axis. If a point on Io is in that locus we can see it, if it's not then we can't. I could imagine sampling a full range of lat/lon, checking to see if that point was in the locus, and drawing a little X there if it was, or something similar.

BTW, after much analysis I decided I couldn't take an image at 17:49, but one did fit at 17:48:30 so I took one there with TDI 12, then three images at 17:50 (TDI 12), 17:51 (TDI 2), and 17:52 (TDI 2). All RGB.

2024-034T17:32:00 RGB 6, 2
2024-034T17:39:00 RGB 2, 12
2024-034T17:48:30 RGB 12
2024-034T17:50:00 RGB 12, 2, 2
2024-034T17:53:30 RGB 12, 6

You have until Monday to change my mind

[volcanopele]

Even for my simulations, I account for the JunoCam FOV by looking at the predicts in Cosmographia and trimming off the excess in Photoshop from the maps I reproject in ISIS.

as far as the timings go, 17:48:30 is even more marginal than 17:50. It does do some gap fill east of Shamshu Patera but it misses Masubi and barely gets Janus. If that's just how the timings work, I mean, that's still really nice resolution at Shamshu and points east. 17:53:30 looks pretty good though maybe reverse the TDIs? 17:53:30 is good for a nice full global shot in sunlight.

Regardless, I'm sure these will be amazing. I'm so thrilled that we are getting such great sub-Jovian/leading-hemisphere coverage!

[mcaplinger]

For posterity, Jason's last message arrived too late to affect command generation, so what I said on 1/12 reflects what we commanded.

I'm never sure how timestamps work on this forum, I'm sending this at 12:46 PST or 20:46 UTC on 1/16.

Now, we wait.

[kymani76]

I made some progress querying SPICE. Here is what I think might be a plot of Junocam's boresight projected on Io's surface for the period from 17:32 to 17:56.
The plot looks a little confusing and I'm still studying how this information might convert to image footprints for specific times.

[volcanopele]

With the timings and TDI settings Mike provided above, here are preview images (combining USGS basemap with PJ57 data, greyscale is USGS basemap and a simulation of Jupiter-shine):

[john_s]

Thanks! How will the resolution on Loki compare to PJ57?

John

[volcanopele]

The resolution will roughly match JNCE_2023364_57C00023_V01, so the second of the images from the last encounter, and at a similar emission angle, but at higher phase angles (110° vs 90°). JNCE_2023364_57C00022_V01 had better pixel scale at Loki, but at a higher emission angle.

[mcaplinger]

Apparently there is going to be a live event with Q&A starting at 9:30 PST tomorrow (2/3) at https://twitch.tv/nasa if that's of interest.

I don't expect the data to be on missionjuno until sometime Monday 2/5, though I have been wrong about this before. There may be some pre-C-kernel images out earlier on social media.

[volcanopele]

Thanks for the heads up Mike. I did let me D&D group know that I might be busy Sunday evening just in case... My legally distinct "Han Solo" character can sit out a session.

[volcanopele]

Full images aren't up yet, but the green filter of one of the crescent sunlit images is now available:



Reflective Loki! the Xihe plume! Jupiter Shine!

[StargazeInWonder]

Here's an effort to align the Loki imagery in green from PJ57 and that first green frame PJ58, to emphasize the specular effects.

This brings back memories of the specular glint seen off Titan's lakes.

Attached image(s)

 

[volcanopele]

BTW, not the first time we’ve seen specular reflection at Loki:

[Bjorn Jonsson]

Wow! This is a beautiful image of Loki, far more so than the I32 Galileo image. And I think there are even some topographic shadows visible at Loki, the best example I can remember of this in images showing Loki.

And the great thing is also that this image implies that the Io flyby was a success (at least for JunoCam). No 'PJ56-like' problems.

[john_s]

New Horizons also got a nice view of strong specular reflection from Loki (red arrow in the attached) at 150 degree phase, albeit at low resolution. The blue arrow points out that the bright deposits around Ra Patera are also very forward scattering.

John

[mcaplinger]

Io images are up on missionjuno now. https://www.missionjuno.swri.edu/junocam/processing

[volcanopele]

First pair of images complete:

[Brian Swift]

PJ58_24 closest approach image from 1499.3 km altitude.
Nightside of Io illuminated by Jupiter-shine.
Needs some cleanup, but not too bad.

[StargazeInWonder]

Wow, the specular effects at Loki vary tremendously between those two images. This flyby is going to provide some unique, perhaps definitive, information about the reflectance of Io's surface.

The Jupitershine imagery is also great.

The only thing mitigating the pure surprise is that PJ57 was already so good, but this adds tremendously.

[volcanopele]

First attempt at the second Jupiter shine image. Need to work on the control network a bit more on this one so this is a first draft.

[Brian Swift]

PJ58_25, altitude 2493.8 km, Jupiter-shine illuminated.

[Brian Swift]

PJ58_25 again, stretched, to bring out hint of plume in upper left.

[volcanopele]

Looks like Masubi lives!

[Bjorn Jonsson]

These images are awesome. They must be by far the best images ever obtained of specular reflections on Io. This is image 26 enlarged by a factor of 2. North is up. I haven't been this excited by new planetary images for a long time (but they appear at a terrible time for me - it's now roughly 3 am where I live!).



A plume is visible at lower right. This part of the image is heavily processed and has been brightened significantly relative to other parts of the image to show the plume more clearly.

[volcanopele]

Improved versions using a global control network

[Daniele_bianchin...]

sorry if the question has already been posted,
What are those that look like lakes filled with?
and does Loki contain magma or a liquid?
Daniel

[StargazeInWonder]

According to previous temperature measurements, the greater part of the surfaces around Loki, while definitely warmed by internal heat, were far too cool to be liquid.

In fact, coincidentally, they are roughly comfortable for human presence!

https://www.sci.news/astronomy/map-loki-pat...n-io-04854.html

So that would be compatible with a solid surface that is glassy.

On the other hand, Io brings uncertainty across time and fine scales of space, so maybe what we're seeing now is different.

[john_s]

Yes, wonderful images. It's interesting that the northernmost patera of Mazda Catena, the site of a strong specular reflection here, was very likely also the location of a specular reflection seen by Voyager 1 45 years ago here. The temporary brightening was interpreted at the time as a blueish cloud emitted by the patera, but it's more likely that it was a specular glint, that was strongest at the time when the blue filter image was taken.

The most striking thing about the Loki specular reflection is how uniform it is- that's telling us something interesting about the resurfacing process.

John

[ngunn]

Wonderful images of a wonderful world. Thanks to all for posting. Jason, I notice that the horizon/limb topography seems to be missing or at least muted in your versions compared with, for example Brian Swift's post 41 which has more the feeling of being in a real landscape. (The horizontal orientation helps too.) This encounter calls to mind the excitement of New Horizons at Pluto even though it's not our first look at Io.

[mcaplinger]

Jason, I notice that the horizon/limb topography seems to be missing or at least muted in your versions compared with, for example Brian Swift's post 41 which has more the feeling of being in a real landscape.

If these are map-projected products without an underlying DTM, then the relief at the limb will get mostly lost -- that's an inevitable consequence of this type of processing that Brian's method avoids.

[volcanopele]

Just two different methods of processing the data. I use ISIS which has the issue of it trimming the images off at the limb when you map project them. But on the other hand they are very useful products for dropping them into ArcGIS.

[Bjorn Jonsson]

A map-projected comparison of images PJ57_24 and PJ58_26:

[antipode]

Has anyone managed to actually spot an impact crater on any of these new images?

As far as I know, no one ever has, but they must be there even if resurfacing is common.

Also, Bjorn, are those specular reflections off peaks at the bottom of your right hand (new Juno) image?

P

[Phil Stooke]

We just recently had a first suggestion of an impact crater in old data:

https://eos.org/articles/amateur-astronomer...le-crater-on-io


Juno would not be able to see craters like that one.

Phil

[Brian Swift]

PJ58_23 with plume. Jupiter-shine illuminated.

[StargazeInWonder]

"Twenty-kilometer diameter craters are made by kilometer-size impactors; such events occur on a Galilean satellite about once in a million years."

https://pubmed.ncbi.nlm.nih.gov/11878353/

Io's surface is "no more than a couple of mission years old."

I don't know what a fair extrapolation is for the rate of cratering for that could be seen in these images, but it seems like it would be very optimistic to hope to see one, and zero is the likeliest number. The ambiguity in appearance adds to the difficulty. (Even on Earth, the cause of origin of weathered, extant craters can be ambiguous, and we're not observing them from orbital distances!)

[Bjorn Jonsson]

This is an updated plot of the values I use for correcting the color in the JunoCam images. It shows the now well known reddening of the images as the mission progresses. Interestingly, the reddening trend became more irregular once the first JunoCam anomaly occurred at PJ48. In particular, the images temporarily became less red after PJ48 but are now again getting redder - there was a very large change (reddening) from PJ57 to PJ58.



There was a very rapid reddening from PJ47 to PJ48 (i.e. immediately before the PJ48 anomaly) and again from PJ53 to PJ54. The third large change is PJ57 to PJ58. I suspect these all correspond to a lot of radiation. Smaller changes might simply be a coincidence or they might be associated with lower levels of radiation or (in some cases) mitigation measures like heating the camera between perijoves as was successfully done before PJ57 in response to the PJ56 anomaly.

[StargazeInWonder]

I keep mulling this over and had the following speculation…

Ionization caused solely by electromagnetic radiation normally has a distinct threshold effect. In the way we encounter this most painfully, we can get a sunburn from the UV in sunlight, but interestingly, we cannot get a sunburn from longer or more intense exposure to visible light. (We can get a thermal burn, of course, but that's another phenomenon.) There is an energy of ionization for a given molecule / bond and a photon either reaches that threshold or does not. For the bonds in most organic compounds, it requires UV to reach that threshold. Intense blue light will not give you a sunburn.

What we see with this reddening is directionally aligned with sunburn: It is as though operations during the mission are harming the shortest wavelength filter a significant degree; the next-shortest, less so. This doesn't display the thresholding that one expects with radiative ionization.

Is it possible that the charged particle radiation from Jupiter and the visible light radiation are somehow additive in damaging camera elements, such that the high end of the stochastic distribution of charged particles is enough to cause damage when added to the visible light of different wavelengths, making blue capable of causing somewhat of a sunburn, green less so, and red less still? If so, then is the imagery performed by JunoCam (which is overwhelmingly, of course, of Jupiter) a causative part of the damage? Put another way, might the reddening have been avoided for arbitrary spans of time if the camera had not been used to image Jupiter?

Heaps of speculation there, and too little too late in any case, but the mystery remains intriguing.

[tau]

A further processing of volcanopele's image of Io from post #48.
Colors enhanced using principal components analysis, image enlarged with Upscayl. North is to the left.

[mcaplinger]

It is as though operations during the mission are harming the shortest wavelength filter a significant degree; the next-shortest, less so.

If there has been wavelength-dependent degradation (I don't think the case is 100% made for this, but it would not be unexpected) it is almost certainly due to radiation darkening, which is caused by exposure to the jovian charged-particle environment. https://sci.esa.int/documents/34530/36042/1...rials_Doyle.pdf is a quick introduction to this and other effects in the context of Jupiter missions.

Junocam was designed with rad-hard glass* types in the front half of the lens and normal glasses farther back, shielded by the front half and also by a large cylinder of titanium. Of course, we were only required to last for 8 orbits, so for radiation effects to still be so subtle as to be debatable this late into the mission speaks to the conservatism of our approach. One could have expected the blue channel to be completely gone by this time otherwise.

It's possible that the annealing recovered some glass transmission that might have been lost, since the heater is attached to the optics and the optics thus warmer than the electronics we were trying to fix.

*I often refer to rad-hard glasses, only somewhat humorously, as "pre-darkened" -- they typically have poorer blue transmission on day 1 than their normal equivalents.

[kymani76]

Here's my version of performing upscale and PCA on Jason's image.


Jupitershine image result.

[tau]

Brian Swift's image from post #44 of Jupiter-shine illuminated Io after further processing
- image artifacts removed
- improved alignement of RGB color channels
- color enhancement by principal components analysis (see here and here for a short explanation)
- image noise reduced

[Holder of the Tw...]

Wen JIRAM images?

[mcaplinger]

Wen JIRAM images?

PJ53 JIRAM data was released on 31 Jan 2024, but the cadence seems somewhat variable so it's hard to be certain. Months, certainly, unless there is some result that the team thinks is worthy of a press release.

https://atmos.nmsu.edu/PDS/data/PDS4/juno_j...undle/data_raw/

[volcanopele]

The PDS release calendar is here: https://pds.nasa.gov/datasearch/subscriptio...ndar-2024.shtml

As Mike mentioned, PJ53 (and 51) data was released to the PDS a couple of weeks ago (I'll post my processed versions of that JIRAM data on those pages in a minute). PJ57/58 PDS data (and thus JIRAM data) should be publicly available on the week of September 24. Mike noted that the schedule is a bit variable so that date is approximate though teams seem to be releasing data a few days to a week early. Also they did change up the release cadence a bit... Last year it was 3 orbits every 4 months, and I was prepared for the PDS release at the end of this month (even asked for time off from HiRISE for it...).

[Bjorn Jonsson]

This is an animation showing a simulation of Io's specular highlights:

https://vimeo.com/914591010

The illumination geometry is constant. The flight path is fictional. It is at uniform speed along a straight line from Juno's location in Io-centered coordinates when image PJ57_24 was obtained to Juno's location when image PJ58_26 was obtained and then back to the starting location.

The map of Io used for rendering the animation is a map-projected mosaic of data from images PJ57_24 and PJ58_27. The mosaic does not contain specular reflections. The specular reflections appearing in the animation were added using a preliminary illumination model with the specular highlights based on the Phong illumination model. This gives a relatively good visual match to the specular highlights in the PJ58_26 and PJ58_27 images.

Interestingly, with the current combination of illumination model parameters it was necessary to assume that Mazda is different from the other sources of specular highlights (e.g. Loki), i.e. more reflective and/or shiny. Without this assumption Mazda was too dark in this simulation. It is probably possible to use a different combination of model parameters where this assumption is not necessary or to use a more sophisticated specular reflection model (the Phong specular reflection model is a very simple empirical model). Interestingly, Mazda is *not* brighter than Loki Patera in the Galileo I32 image showing specular highlights from Loki etc.