Thursday, February 27, 2014

Very Short Update

I've been so busy these past couple of days with the data taken on 20 Feb 2014 UTC that I haven't been able to update this blog.  Rest assured, I'll get you fully informed.  That's my goal for Saturday morning (1 March).

Last night I took more data.  Had to wait until about 07:30 UTC for the clouds to completely clear off, and then went at it for about three hours.  This kind of data collection I'm doing is hard hard hard.  I'll tell you why on the next blog post.

For now, here's an image of the heart of M35.  I'm looking at M35 in order to better understand the photometry I'm seeing from the Jupiter data from 19 Feb.

Figure 1 is my data, Figure 2 is a chart of the same region from the very way cool AAVSO chart maker (http://www.aavso.org/vsp).
Figure 1: Center of M35 image from 27 Feb 2014 UTC


Figure 2: AAVSO chart of same region
The orientation of the two aren't exactly the same, but I'm assuming that my readership is smart enough to make the translation.

The AAVSO chart shows stars down to magnitude 14.  My guess is that my image is getting down to 12-13.  Not bad for a 0.2 second exposure, eh?  Focus looks really good and there doesn't appear to be much in the way of clouds.

More Saturday.

One more thing.  I took a couple of pictures of my "observatory" last night.  In the years to come, I don't know if I'll laugh or cry when I see these.  Maybe both.  In any case, this might be a good argument for "anything can be done given the right amount of determination."
Wayaway Observatory


Taking flats and enjoying the sunset

Tuesday, February 25, 2014

X Flare

I'm glad that I now have access to calibrated data from SDO so I can examine features on the sun in ways that I want rather than relying on uncalibrated jpeg crap.

I'm not too keen on the "exciting" events like today's X4.9 blast, but it sure is a pretty sight.  Here's an image taken by SDO with the 1600A filter at 00:45 UTC on 25 Feb 2014:

An X4.9 boomer at 00:45 UTC
I've played with the scaling to bring out that cool-looking loop structure and added false color to show off features that may otherwise not be noticeable.

The width of that outer loop is about 37,400km (three Earth diameters).  Do you feel small yet?

I also like the fact that this event was so bright that it caused internal reflections in the optics.

An hour later, this region is still "glowing":

Same region at 01:45 UTC
Same scale, same false-color.

Sunday, February 23, 2014

Scatter

Looking at all of my 20 Feb 2014 UTC data, the photometry is all over the place.

Why?

There's one hint: the photometry is "better" for targets furthest away from Jupiter.

In this data set, there were four targets: Europa, Ganymede (both near Jupiter -- approx 60 pixels away), Callisto, and a background star (both far away from Jupiter -- approx 185 pixels away)  As you can see below, the two nearby targets have very very very poor photometric continuity, while the two far away targets are more consistent (but still quite terrible):

Figure 1: Callisto Photometry

Figure 2: Star Photometry

Figure 3: Europa Photometry

Figure 4: Ganymede Photometry
Since Europa and Ganymede were so close to each other, their photometry began to mix which is why there's an absense of Europa photometry after about 03:42 UTC and a corresponding "brightenting" in the Ganymede data after that time.

But as you can see from the plots, the scatter with Europa and Ganymede is totally unacceptable and in fact all of this data as it is right now is totally useless.

However, the big hint is the fact that the quality of the photometry changes in relation to the target's distance to Jupiter.  IF this was just due to clouds, I'd expect the shapes of the plots to more or less be the same.  I see that to a certain extent, but there are other things going on that can only be due to the "instrumental effects", i.e., internal reflections (see images below in Figure 5), and the fact that Jupiter is such a huge blob.

Looking at the low-level "muck" in some of calibrated images, I see that yes indeed things are very dynamic:

Figure 5: Calibrated Image Samples
Figure 5 is a sequence of four images taken as the targets scan across the field of view.  Jupiter is the bright spot.  You can also see Callisto (moving in from the bottom) and the background star (on the right).  Europa and Ganymede are hidden in the Jupiter blob.

It's obvious from these images that my photometric measurements are going to change dramatically over time, depending not only on where a single target falls, but also its proximity to Jupiter.

So I have two things to help me move forward and they both have equal priority.

First, model the PSF of Jupiter and subtract it out.  Depending on the quality of the fit, I should see a change (hopefully for the better) in the photometry for all four targets.

To test for consistent photometry, I've decided to take a series of at least 200 images of M35, which is near Jupiter but out of its photometric influence.  What I'm hoping to see is consistent photometry.  I certainly won't be lacking targets -- which will give my detection software a good workout, too.

The clouds have continued to inhabit the sky over my location, so no new data has been collected.

Frustrated, but still very enthusiastic and hopeful.

======= UPDATE =======

I decided to take a look at some of the images I got back on 13 Feb 2014 UTC when there weren't any clouds, and I now realize that I was dealing with more clouds than what I thought I was.

Below are two more figures, both showing the same raw images.  Figure 6 shows a raw image from 13 Feb 2014 UTC (top) and a raw image from 20 Feb 2014 UTC (bottom) with a certain level of contrast/brightness adjustment.  It's clear that the 13 Feb image is much sharper and as you can see Jupiter's light is leaking onto the satellites MUCH less.  Figure 7 shows the same two images, but I've adjusted the contrast/brightness to bring out the low-level muck.  As you can see, the two satellites above Jupiter in the top image (strangely almost the same distance and configuration as the 20 Feb satellites) are still visible while the bottom image shows them completely washed out.

So I'm pretty much worrying over nothing.  20 Feb 2014 UTC showed high clouds the entire time I was taking data and that's the explaination for the poor photometry results.

Having said that, I've located a Gaussian function maker in IRAF that will allow me to model the Jupiter PSF and subtract it out if necessary.

Figure 6: Raw data clear (top) and cloudy (bottom)

Figure 7: Same as Fig 6, brightness/contrast adjusted
So yes indeed there is still much hope for this project.  Now all I need is some clear sky!  Maybe tomorrow night if this new system stays to my south:



Friday, February 21, 2014

Cloudy Photometry First 200 Images

I've calibrated the first 200 images I took on the evening of 20 Feb 2014 UTC.  Unfortunately there were quite a few clouds, so the results aren't all that pretty.  Still, I'm encouraged by the results which I show and describe below.

Much of the time I've spent these past few days has been with the software.  Now that I have lots of data, I can no longer manually determine the approximate locations of the Jovian satellites to tell the photometry software I've written where to look.  So most of the programming I've been doing has been to automatically detect the locations of the satellites in each image.  No small task, but I knew that already.

Anyhow, what I've got now is a pretty robust program that does several things:

1. Locates the position of each Jovian satellite
2. Calculates the pixel distance from the satellite to Jupiter
3. Does square aperture photometry of the satellite
4. Detects cosmic ray events

At the moment, the accounting for all of this is still pretty messy as far as the output goes, so it's only really possible to see the results when plotted.  That will change once I'm able to figure out a way to say "that satellite is Europa and here are all the measurements".  This will come in time as I let the data teach me what I need to learn.

These first 200 images included quite a few clouds in the area, which shows up pretty clearly on the plots -- especially the photometry one.  The night got cleared as time went on, so I'm hopeful that the scatter with even the best of these gets smaller.  I'm still seeing about a 4% variation.

Ok, so here's the photometry plot.  The x-axis is the number of seconds after 0h UTC on 20 Feb 2014.  The y-axis is the count measured at the moment in ADU's (analog-to-digital units).
Figure 1: Photometry of Ganymede, Europa, Callisto
Yes, this is a messy plot.  But I hope you can see that as you scan from left to right there are three distinct groups.  Maybe the annotated plot below will help:

Figure 2: Annotated Figure 1
So the brightest satellite is Ganymede, next brightest is Europa, and the faintest is Callisto.  I hope that helps to make sense of this otherwise messy plot.

So as you can see, the scatter is pretty large and as I got to the end of this first set of 200 images, the clouds came in and totally destroyed the photometry which you can clearly see in the right third of the plot (from time 10700 through 11000).

As I said above, the best photometry I can see is the Europa photometry at time ~10150.  This is about a 4% scatter, which is still pretty terrible but considering the clouds I'm not too surprized.  I look forward to making similar plots of data later in the evening when the clouds moved out.

Here's a plot of the best Europa photometry.  Once again, the x-axis the time and the y-axis is the pixel count.  My impression is "not bad":

Figure 3: Best photometry of this first set of 200 images


The next three plots show the distance from satellite to Jupiter.  The x-axis is once again the relative time and the y-axis is the distance in pixels.

Here's Callisto slowly moving towards Jupiter:

Figure 4: Callisto Distance from Jupiter (in pixels)
Next is Europa moving towards Jupiter faster then Callisto (which is nice to see since Europa is closer to Jupiter and SHOULD have a faster orbital motion):

Figure 5: Europa Distance from Jupiter (in pixels)
And finally Ganymede moving away from Jupiter faster than Callisto but slower than Europa:

Figure 6: Ganymede Distance from Jupiter (in pixels)
So I'm still encouraged by all of these results, although I'm hopeful that the photometry gets better.  I really need as close to 1% scatter as I can get and if I'm not able to get that, I'll need to look and see if I can make any improvements or modifications to my calibration process.  If that fails, then I need to look at the hardware to try to determine why there's such a scatter.

Tomorrow I should be able to get through more of the data and make similar plots.

Thursday, February 20, 2014

Jupiter Moon Photometry 20 Feb 2014 UTC

Just a very quick note and picture about last night's observing session.  The clouds parted (well, mostly) and I got a pretty good couple of hours in.  I grabbed a total of 1357 raw images.  Some of those were flats, soe of those were darks, some of those were biases, some of those were binary stars.  But 1100 of them were Jupiter system images.  Because of the lack of tracking, I'd estimate that about 1/2 to 2/3 were images with the satellites in the FOV.

So the next task is the put the flats, darks, and biases together and start some data calibration.  Lotsa work!

Tonight looks like it's gonna be clear again, so I'll be back at it again for a couple of hours.

I was really pleased with the flat field technique I used.  I'll talk more about that in another post.  As for now, here's a picture from the end of the session.

The Jupiter system from last evening

Wednesday, February 19, 2014

Sun Observation Project

I've finally figured out how to access the raw (well, Level 1.5) image data from the SDO and I'm very excited about this.  I've been wanting to do "something" with the sun and the only way was to have access to the actual data rather than silly jpeg's and aminations.

I still have much to do, but I wanted to post some pictures of some image sequences of a "flash" or what I'm calling a "lightning bolt" that took place last night (local time for me) for a few minutes.  These are the kinds of things I'm interested in studying to get a feel for them and look for patterns of behavior.

The image frames you'll see below are all organized according to time, and they read from left to right, top to bottom.  There are sixteen images per frame three minutes apart, going from 02:00 UTC (upper left) to 02:45 UTC (lower right) on 19 Feb 2014.

There are a total of four frames here, representing four different wavelengths: 94A, 131A, 171A, and 1600A.  There are a total of nine different wavelengths but I limited this quick look to these four.  I want to make a plot of the intensity of this flash as a function of time for each wavelength, but I need to write some additional software to do this.  Today's task was to start understanding the kinds of scripts I need to create (actually, programs to write the scripts for me) in order to access the data and get it in a usable form.

Notice how the flash dims over time, but differently for each wavelength.

In any case, here are the image frames.  I've put some false coloring on them instead of leaving them grayscale to bring out some of the background features -- all of which are very interesting and changing in practically every image.

I will continue to look at all of this data.  There's a lot to look at and it's always being updated every three minutes!

One more thing, a 3 by 3 pixel area in these images corresponds to the size of the North America (24,709,000 km2).  These images are 1820 times that size!!!!!

Enjoy.
94A

131A

171A

1600A
All images courtesy of NASA/SDO and the AIA, EVE, and HMI science teams.

Tuesday, February 18, 2014

Under a Nice Blanket of Clouds

An astronomer's nemasis:
I'm somewhere underneath these clouds
But the great thing is that I'm clear-sky limited not telescope time limited.  The "professionals" have nothing on me!

Clear skies are on the way, however.  Tomorrow night is suppose to be clear and windy.  Ah well, at least it's been an unusually warm winter.

Hopefully at that time I'll be able to try out my flat field technique and get images of some binary stars so I can measure my pixel scale.  If the photometry is good enough (which I expect it will be) I'll also be imaging the Jupiter system.

Here's how it's supposed to look 24h from now:

The Jupiter system on 20 Feb 2014 at 03:26 UTC
So it looks like more Callisto stuff for sure.  Europa and Ganymede might be too close for good photometric separation.  Io is transiting and will therefore be invisible.

Stay tuned!

Sunday, February 16, 2014

Long-term, High Time-resolution Photometry of Jupiter's Galilean Satellites

Long-term, High Time-resolution Photometry of Jupiter's Galilean Satellites

I will be making photometric measurements of the Galilean satellites at least
over the next several months until about one month (I'm not sure how close to
the sun I can get before light from sunset begins the interfere) before
superior conjunction that takes place on 14 July 2014.  At that point I should
have sufficient data to determine if there's enough "interesting stuff" to
continue this project after superior conjunction.

So this means I have from now (16 Feb 2014) until about 14 June 2014 for my
observations.  This is a maximum of 119 nights.  I will collect data for
at least 2 hours each night.

Assuming a 50% clear night return, and a data collection rate of 522 images per
night (based on 87 images over the course of 20 minutes on 12 Feb 2014), I
should have at that time 31059 data points (this is a rate of 261 measurements
per hour, so if I observe more hours per night, this number will increase) for
at least one of the satellites.

Because of the dynamic nature of the OTA (nightly collimation changes & camera
mount angle), I will have to obtain new flats every single night to ensure
that they properly correspond to the system for that particular evening.  I
will monitor this over the course of several nights to see if there is indeed
a measureable difference.  I suspect that there will be.  Hopefully I can get
this sufficiently routine to not spend more than 30 minutes on this per
observing session.  Hardware improvements listed below will alleviate some of
this tediusness.

The one great advantage I have over "professional" observations is that I have
access to my telescope at any time, and setup time is a matter of minutes.
This means I don't have to write observing proposals for obtaining telescope
time and then hope that the nights I request will be photometric (or at least
clear).  I have no competition for observing time.  I can cover long periods of
time continuously with high temporal resolution -- something that has never
been done (as far as I can tell with an ADS and arxiv.org literature search)
when it comes to observing the Jupiter system.

=========== What am I looking for? ============

1. Photometric light curves

    From what I've read in the literature, I should expect to see fluctuations
    in the light curves.  Simonelli (2002AGUSM.P22A..07S) says "Io exhibits
    probably the most diverse set of different photometric behaviors of any
    object in the Solar System".  That catches my attention.

    These curves should be pseudo-sinusoidal in general, but will probably be
    more complex than this based on phase angle (where it is in relation to
    Jupiter), changes in albedo (surface features), solar angle (as Jupiter
    and it's system orbit the sun), and sudden changes in brightness due to
    unknown events (volcanic eruptions, meteor impacts, solar CME events,
    etc.).  All four moons are tidally locked, which means that they always
    face towards Jupiter (just like our moon does with us).  So observing each
    satellite in its orbit will allow me to measure the entire surface (except
    when transiting or being occulted by Jupiter).

    While it'll be interesting and useful to add to the database of
    observations, what I'm most interested in is

        1. detecting those unknown events
        2. detecting novel patterns

    Once I get into this more, my guess is that the data will help me see other
    things to look at more closely.  This is a very organic project.

2. Distance measurements from Jupiter

    I'm fairly certain that the orbital parameters of these satellites are
    very well known and that I won't find anything new here, but nevertheless
    I will be making those measurements.

3. Mutual Occultations

    These are occultations of one satellite by another satellite.  I'm
    not aware of anything interesting I could get from this information, but
    I will observe them and see if the data tells me to look for anything.

4. Stellar Occultations

    These are occultations of stars by a satellite.  The timings of these
    events will allow for measurements of physical diameters.  Again, it's
    likely that there are things I'm ignorant of and I'll just have to wait
    for the data to suggest what to look for.

5. Jupiter Transit and Occultation timings

    Obtain timing measurments of transits and occultations of Jupiter.
    Once again, I'm not sure if there's anything useful in this data, but I'll
    be able to do it so I might as well measure it.


6. Cosmic Ray detection and correlation

    Most astronomers detect cosmic rays just to eliminate them from the

    images they have.  I think there is something interesting about cosmic
    ray events and the fact that they are so easily detected on CCD cameras
    make them another thing I can take a hard look at instead of just
    throwing away as junk.  So for starters I will count the number of cosmic ray
    events and correlate them to solar activity.  Here once again I'm going to 
    let the data tell me how to proceed by looking for patterns.

============ Things To Do For The Future ============

I am very confident that my current system and observing methods will provide
me with very encouraging results that will justify the following future
improvements.

1. Get a telescope that tracks

    At the moment, my telescope doesn't track.  This limits the amount of
    data that I can collect.  At the moment, I'm able to get about 8 images
    before the targets move out of field.  I then have to move the telescope
    slightly west in RA which I usually overshoot.  So then I wait several
    images until the targes are back in view.  Using 2x2 binning, the
    duty cycle is six seconds.  This is barely tolerable, especially when it
    comes to occultation timings and spacial resolution (pixel scale)

    This also limits the exposure time to something very short since longer
    the exposure time, the longer the streak caused my earth's rotation.  This
    is not a problem for this project since very short exposures (<0.1s) is
    adequate to obtain seeing-limited data.  However, if I ever want to
    extend this project to fainter targets (Saturn satellites, asteroids,
    variable stars, etc.) I would require a longer exposure time which would
    produce streaks which makes photometry problematic (streaks tend to
    overlap if they're too long).

    The ideal arrangement will be for me to retire my current system that I've
    proudly had in my posession for 30+ years and replace it with a new,
    similarly sized system with all of the modern accurate tracking and
    optical capabilities.

2. Recoat the primary mirror

    Due to non-use and neglect, my 25.4cm primary mirror is in pretty
    desperate need of a recoating.  This diminishes the sensitivity of the
    system, but as long as my calibrations are good, this shouldn't be a
    problem in overal performance.  There would be no need for this if I'm
    able to replace my current system with a new one.

3. Get a new camera

    The two downsides to the camera that I now have (an SBIG ST-7ME)

        1. the array size is small (765x510 Pixels, 9 micron)
        2. the download time is long (6s for 382x255 -- 2x2 binning)

    The array size is actually ok for now since it's a lot easier to keep
    the target in the large FOV.  That limits pixel scale, though.  But
    in truth 9 micron pixels is pretty good so the only way to decrease
    the pixel scale would be to modify the focal length of the system and
    currently that would mean narrowing the FOV which makes it all that
    more difficult to keep the target within the field.  This will change
    once I'm able to track.

    The download time is the major problem with this system and limits the
    time resolution of the data.  Still, a 6-12 second time resolution should
    be adequate for this "first look" project.  A further analysis of the
    data will tell me if shortening the duty cycle would help in any way.
    However, with a system that can download faster (an ST-402ME claims
    800,000 pixels per second using USB -- so that's a 0.48s download time
    at 1x1 binning) I could set the time between exposures which would be
    nice to be able to control rather than having that limitation.

4. UBVRI filters

    The current system has 'Red', 'Green', and 'Blue' filters which are
    pretty wide band and not necessarily useful for scientfic measurements.
    It would be very nice to collect color information which would add
    another dimesion to the photometry.  I would expect the "color" curves
    to be much more complex than single color photometry.

    So I would like to obtain a UBVRI filter set and create an observing
    program that collects data with all of these filters.  Simultaneous
    multispectral photometry would be an ideal situation.

5. Automation

    At the moment, I need to be with the telescope during all observations.
    I can see a time when I could set up an observing "queue" and allow the
    system to run itself throughout the night.  This would increase the
    amount of data throughput by at least a factor of four since most nights
    are at least 8 hours long (excluding time to take calibration data).

6. Networking

    24 hour coverage of the Jovian system is an ideal, but from a single
    location I will average 6 hours a night over the course of a year.
    The ideal sitation would be to have a number of observing locations
    spread geographically over the planet to provide with 24 hour coverage
    with perhaps an hour of overlap for each location.  This means eight
    stations equally spread over the planet.

    This is unlikely to happen, but even finding one or two other locations
    with willing observers would be great.


Saturday, February 15, 2014

First Look At Callisto Photometry

With a fury of programming over the past couple of days, I've come to a stopping point with the data that I have.  As this was a first look just to collect some data and not worry two much about data quality, the results can't be taken too seriously.  Having said that, the results look pretty good and very encouraging.

So what I did was write code to measure the position and brightness of Callisto over the course of the observing "night" (which was only over the couse of 20 minutes, 24 seconds).

Knowing the position of Callisto and the position of Jupiter, I can also compute the distance between them, which is shown in this first plot:
Separation (in pixels) between Callisto & Jupiter
The x-axis is the time in arbitrary seconds, and the y-axis is the separation in pixels.  As you can see, even over the short 20 minute period, I'm seeing a smaller and smaller separation.  Indeed, Callisto was moving closer and closer to Jupiter in its orbit.  So that's very nice and very encouraging.  I haven't done a numerical analysis of the scatter, but by eye I'm seeing a scatter of less than 1/4 pixel.  That's just fine.

(One thing I'll be doing on the next clear night is imaging a number of known visual binary stars to get a value of the pixel scale and field of view)

The next plot shows the photometry for Callisto over the same set of images and over the same timescale.  I'm using aperture photometry with this and at the moment I'm using a 9x9 box centered on the computed centroid of the target.  My aperture photometry computation will improve as I write better code (I eventually want fractional-pixel circular photometry). Here's the plot:

Dark-subtracted aperture photometry of Callisto
The x-axis is once again the time in arbitrary seconds, and the y-axis is the count.  At this point, the images have just been dark subtracted with one of the darks that I took.  I'll be doing standard calibrations (hopefully) from here on out.  This was just supposed to be a quick look to start figuring out what I need to do to analyze this data.

Again, I haven't done a numerical analysis of the photometric scatter, but it's on the order of 500 counts, which in this case is on the order of 5%.  That's not good enough and I'm wondering why there is such a large scatter.

I'm also curious why the values seem to be changing over time.  I wouldn't think that Callisto varies by that much (10%!!) over such a short period of time.  There were high thin clouds in the area, so that might account for it.  It's really hard to tell what's going on here.  So at the moment I'll leave this and hope that I can better understand what's going on with more data.

Wednesday, February 12, 2014

Two New Astro Projects

The first involves Jupiter's four biggest moons.  Here's an image I took just moments ago through my telescope:

Jupiter and three of it's moons
The upper left is Ganymede; the upper right is Europa; the lower one is Callisto.  Io was at the time transiting Jupiter, which of course is that big overexposed blob in the middle.  This is going to be an interesting project if I'm able to do some decent photometry.  This was just a test run, but I might get some decent results from this.

The second project involves looking at star streaks.  Here's an example image, also from tonight:

5 second exposure star streaks

Stay tuned for more information on these projects.