It is currently Sun, 30-04-17, 14:45 GMT

All times are UTC




Post new topic Reply to topic  [ 28 posts ]  Go to page 1, 2  Next
Author Message
PostPosted: Sun, 10-04-16, 19:50 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4439
Location: Hamburg, Germany
Hi all,

let me report a bit about a current, pretty interesting activity of mine, concerning the precision testing of celestia.Sci orbits in our solar system. A standard and VERY sensitive challenge, concerns precision tests of so-called mutual events among Jupiter's four Galilean moons. The procedure I'll be describing here can also be performed by analogy, using your good old Celestia program (until celestia.Sci will be available in public). ;-)

Here are the steps for celestia.Sci:
Before going through the individual testing procedure for a typical example, let me first make several explanatory comments.
  • First of all, we need a list of high precision predictions for such events. Alternatively, of course, precise oservations. In this link you find an up-to-date listing for the time window 2014 - 2015 from the well-known French Institut de Mecanique Celeste et de Calcul des Ephemerides (IMCCE) : http://www.imcce.fr/langues/en/observat ... hemu15.txt
    In 2014-2015, a series of eclipses and occultations took place among the satellites of Jupiter, thanks to the equinox on this planet occurring in 2015.

    Here you can see why usually, such occulting moon events can be detected pretty well photometrically... (from http://www.imcce.fr/langues/en/observat ... index.html)
    Image

  • The events in question use the following general notation: nOm and nEm,
    with n,m = 1..4, labelling the 4 Galilean moons. Hence nOm means that moon n will occult moon m, while nEm means that moon n will cast a shadow on moon m, giving rise to a respective eclipse. The numbering is standardized as follows: Io =1, Europa=2. Ganymede=3 and Callisto=4.

  • On my 24" screen, I managed to display simultaneously a) the main window with a much elongated shape, b) the part of interest of the mutual events listing displayed in Firefox, and c) the Preferences display (which I coded entirely from scratch a while ago). This is how things look now in celestia.Sci (reduced to a 1366 x 768 screen)

    [click on image + fullscreen browser shortcut (F11 for FF)]
    Note, the slightly ellipsoidal shape of the 2 displayed moons is due to a change of aspect ratio when I reduced the images from 1920x1200 to 1366x768 px. The latter size is more popular on laptops...
    Attachment:
    mutual_eventtest1.jpg
    mutual_eventtest1.jpg [ 190.45 KiB | Viewed 1464 times ]


  • Note that I coded the Preferences dialog as a non-modal window, meaning that one may simultaneously work with both windows. E.g. by entering date/time data in Preferences->Time and trying them out right away in the main window. You may also move the dialog anywhere to a convenient place or dock it into the main window on the left or on the right. The latter option is illustrated here:

    [click on image + fullscreen browser shortcut (F11 for FF)]
    Attachment:
    mutual_eventtest1_1.jpg
    mutual_eventtest1_1.jpg [ 201.02 KiB | Viewed 1464 times ]

    The first line in the (time -) Preferences dialog, refers to an optional account of the LightTravelDelay (LT) along with a calculation and display of its value (hour / min /sec) .
    If LT was activated, the green LT flag appears behind the date/time display in the main window along with a brief notice.
  • To characterize mutual events, Universal Time (UTC) zone is selected as usual. Next, the respective date/time data (<=> begin of event) are entered into the combo boxes (year month day / hour min sec) . The corresponding simulation date/time is set in the system by hitting the bar below the combo boxes. By hitting the last bar (Now), one may optionally return to the precise current date / time

Now we are ready to check a particular event from the IMCCE document:
Code:
Date            begin: h  m  s      end: h  m  s      Type    Duration[minutes]
---------------------------------------------------------------------------------------- 
2014  12  20       05 31 14             5 51 33        2O1        20.3

In this 2O1 event, Europe occulted Io.

  • Open a much elongated main window with Earth centered and selected (startup config)

  • Next select the target Galilean moon (Io=1), center it (hitting C) and then apply a strong zoom (via Shift + mouse drag), until you can comfortably see Io's size. Note: The observer remains Earth-based throughout, and thus the LT correction is needed to push the mutual event time backwards in the Jovian frame (i.e.earlier than the earth-based prediction!). The light travel time amounts to 38-40 minutes from Io to the observer!

  • Next hit the first tool button on the bottom left of the main display, which opens the Preferences panel. Click on its Time section. Place the dialog conveniently, not to hide the date/time display (top right) on the main display.

  • Enter the date/time values from the IMCCE data above into the combo boxes on the panel and then hit the "Set simulation time" bar.

  • As the last setting, click on the LT flag. You can observe simultaneously in the date/time display of the canvas that the begin time for our 2O1 event has moved to earlier times. This is of course due to the light travel time (LT = 38':55.5") that needs to be added in before the event is seen by our Earth-based observer!

  • After centering again (C key), and zooming until you see booth involved moons with convenient size, slowly advance the time display (L, K and J keys) until the predicted mutual event time has been reached in the canvas date/time display. By then Europa and Io should be getting VERY close and eventually Europa will mostly occult Io. Compare with the predicted begin, end and duration times... Differences wrto IMCCE predictions should remain within a few minutes typically (satisfied in celestia.Sci!).

I am doing the same type of precision tests with the Pluto - Charon eclipses, which is particularly important due to residual orbit uncertainties.

Fridger

PS:
In case anyone is interested: celestia.Sci is currently running with the latest Qt version = Qt 5.6.0 which has been declared as the first LTS = Long-Time Support release of the Qt5 series.

_________________
Image


Top
 Profile  
 
PostPosted: Tue, 12-04-16, 10:06 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4439
Location: Hamburg, Germany
Hi all,

this post is in continuation of my previous one.

Below I consider the 2002-2003 season of mutual events among the Galilean moons. There is a nice paper by Dr. J.-E. Arlot (IMCCE), where you can also learn why the mutual events in the Jovian system only take place every six years.

(@Astronomy & Astrophysics journal:)
http://www.aanda.org/articles/aa/pdf/2002/08/aa1762.pdf

The numerical predictions corresponding to the paper are available at CDS Strasbourg:
http://cdsarc.u-strasbg.fr/viz-bin/nph- ... F383%2F719
The meaning of the various columns is explained in Arlot's paper.

This time, I made a little video (MP4) from a typical test (1ECL2, 2003 March 13) which is almost perfect in celestia.Sci within the reading accuracy! The video shows the shadow of Io (1) moving across the surface of Europa (2) (and back ;-) ) at the correct time. The eclipse is partial (P). You'll find the predicted times in the listing above by looking up the line starting as: 2003 / 3 / 13 | 1ECL2 | P ...

***********Click for start of video************
1ECL2
************************************************
My impression is that these calculations are more accurate (and detailed) than the above 2014-2015 results. In any case I get VERY good agreement with celestia.Sci, as you may see yourself in the video. I would be interested hearing about corresponding results within Celestia. I guess the accuracy should be about comparable...

Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Sat, 16-04-16, 19:54 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 749
Location: N 42.38846 W 83.45456
Image
i get this but i am also a bit out of date on the spice kernels

a few more images
zoom through earth
Image Image

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
PostPosted: Sat, 16-04-16, 20:15 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4439
Location: Hamburg, Germany
Ah, many thanks for testing with Celestia, John!

Since your time zone is UTC-5 and the eclipsing event (1ECL2) happens around 17:26 UTC-5, it would correspond to about 22:26 UTC. The prediction of Dr. Arlot was 23:06 UTC on 2003/3/13. Hence your result is around 40 minutes early. On the other hand, I am missing the green LT icon on your screenshot, which would mean another ~38 minutes shift (i.e. delay) due to the light travelling time from Europa to the earthbound observer!

In any case this type of tests represents a huge challenge of orbit accuracy!

Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Sat, 16-04-16, 20:41 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 749
Location: N 42.38846 W 83.45456
i have been noticing that the moons correlate rather well with voyager and Galileo and Cassini images

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
PostPosted: Fri, 22-04-16, 1:28 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 385
Location: Tuscany, Tyrrhenian Sea
x/

_________________
Never at rest.


Top
 Profile  
 
PostPosted: Fri, 22-04-16, 18:52 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4439
Location: Hamburg, Germany
fenerit wrote:
x/

:wall: :lol:

_________________
Image


Top
 Profile  
 
PostPosted: Fri, 22-04-16, 21:31 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 385
Location: Tuscany, Tyrrhenian Sea
Well... that was because I wrote that: where is the delta-T? :wall: When I convert TDB to UTC by the function celestia:tdbtoutc() to find the max. time of 08/21/2017 solar eclipse, seem that I got still TDB, being the eclipse resulting predicted (through Meeus, A.A. algorithms) for the 18:26:28 UTC, when indeed should be 18:25:30 UTC (Nasa http://eclipse.gsfc.nasa.gov/SEsearch/SEdata.php?Ecl=+20170821 ). The accuracy range of Meeus is within seconds, not close to minutes. the delta-T for such a date is 70.3 sec. Maybe you have enhanced the sourcecode, but the table of leap seconds of 1.61 -> 1.70 versions is missing of the latest 36sec. of TAI-UTC at 1 Jan 2015 and this do alters some timing. Maybe a better delta-t algorithms is necessary.

_________________
Never at rest.


Top
 Profile  
 
PostPosted: Fri, 22-04-16, 22:08 GMT 
Offline
User avatar

Joined: Tue, 04-09-07, 21:55 GMT
Posts: 749
Location: N 42.38846 W 83.45456
a few minuets ether way is rather good
considering the code dose not use a huge amount of decimal places

for running at Aprox. 60 FPS, a few min. is rather GOOD

_________________
"I don't pitch Linux to my friends, I let Microsoft do that for me."
Using OpenSUSE 42.1 & Scientific Linux 6.7


Top
 Profile  
 
PostPosted: Sat, 23-04-16, 1:51 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 385
Location: Tuscany, Tyrrhenian Sea
Anyhow, if it can be useful, at least for polynomials...

Code:
-- Espenak & Meeus: "Five Millennium Canon of Solar Eclipses"
-- source NASA: <eclipse.gsfc.nasa.gov/SEcat5/deltatpoly.html>
local CANON_DT = function()
   local TDB = EPHEM_TIME();
   local ut = celestia:tdbtoutc(TDB + local_offset);
   local year = ut.year;
   local month = ut.month;
   
   local y, u, t, c, dt = 0, 0, 0, 0, 0;
   
   -- astronomica year, not the historical one
   -- "the year which the historians call 585 B.C. is actually the year -584" cit.
   -- see Meeus, "Mathematical Astronomy Morsels"
   -- already accounted by Celestia
   
   -- decimal year for the middle of the month
   y = year + (month - 0.5) / 12;
   
   -- Correction for ELP-2000/82 Moon's secular acceleration of -25.858 arcsec/cy^2
   if ((year >= 1955) and (year <= 2005)) then   
      c = 0.0;
   else
      c = -0.000012932 * (y - 1955)^2;
   end   
   
   if (year <= (-500)) then
      u = (year - 1820) / 100;
      dt = -20 + 32 * u * u;
   elseif ((year > (-500)) and (year <= 500)) then
      t = y / 100;
      dt = (10583.6 - 1014.41 * t
                 + 33.78311 * t * t
                 - 5.952053 * t * t * t
                 - 0.1798452 * t * t * t * t
                 + 0.022174192 * t * t * t * t * t
                 + 0.0090316521 * t * t * t * t * t * t);
   elseif ((year > 500) and (year <= 1600)) then
      t = (y - 1000) / 100;
      dt = (1574.2 - 556.01 * t
                + 71.23472 * t * t
                + 0.319781 * t * t * t
                - 0.8503463 * t * t * t * t
                - 0.005050998 * t * t * t * t * t
                + 0.0083572073 * t * t * t * t * t * t);
   elseif ((year > 1600) and (year <= 1700)) then
      t = y - 1600;
      dt = (120 - 0.9808 * t
              - 0.01532 * t * t
              + t * t * t / 7129);
   elseif ((year > 1700) and (year <= 1800)) then   
      t = y - 1700;
      dt = (8.83 + 0.1603 * t
               - 0.0059285 * t * t
               + 0.00013336 * t * t * t
               - t * t * t * t / 1174000);
   elseif ((year > 1800) and (year <= 1860)) then      
      t = y - 1800;
      dt = (13.72 - 0.332447 * t
               + 0.0068612 * t * t
               + 0.0041116 * t * t * t
               - 0.00037436 * t * t * t * t
               + 0.0000121272 * t * t * t * t * t
               - 0.0000001699 * t * t * t * t * t * t
               + 0.000000000875 * t * t * t * t * t * t * t);
   elseif ((year > 1860) and (year <= 1900)) then   
      t = y - 1860;
      dt = (7.62 + 0.5737 * t
               - 0.251754 * t * t
               + 0.01680668 * t * t * t
               - 0.0004473624 * t * t * t * t
               + t * t * t * t * t / 233174);
   elseif ((year > 1900) and (year <= 1920)) then   
      t = y - 1900;
      dt = (-2.79 + 1.494119 * t
               - 0.0598939 * t * t
               + 0.0061966 * t * t * t
               - 0.000197 * t * t * t * t);
   elseif ((year > 1920) and (year <= 1941)) then   
      t = y - 1920;
      dt = (21.20 + 0.84493 * t
               - 0.076100 * t * t
               + 0.0020936 * t * t * t);
   elseif ((year > 1941) and (year <= 1961)) then      
      t = y - 1950;
      dt = (29.07 + 0.407 * t
               - t * t / 233
               + t * t * t / 2547);
   elseif ((year > 1961) and (year <= 1986)) then      
      t = y - 1975;
      dt = (45.45 + 1.067 * t
               - t * t / 260
               - t * t * t / 718);
   elseif ((year > 1986) and (year <= 2005)) then      
      t = y - 2000;
      dt = (63.86 + 0.3345 * t
                - 0.060374 * t * t
               + 0.0017275 * t * t * t
               + 0.000651814 * t * t * t * t
               + 0.00002373599 * t * t * t * t * t);
   elseif ((year > 2005) and (year <= 2050)) then      
      t = y - 2000;   
      dt = 62.92 + 0.32217 * t + 0.005589 * t * t;
   elseif ((year > 2050) and (year <= 2150)) then      
      dt = -20 + 32 * ((y - 1820) / 100)^2 - 0.5628 * (2150 - y);
   elseif (year > 2150) then      
      u = (year - 1820) / 100;
      dt = -20 + 32 * u * u;
   else
      dt = 0.0;
   end
   
   return (dt) + (c);
end

_________________
Never at rest.


Top
 Profile  
 
PostPosted: Sat, 23-04-16, 8:27 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4439
Location: Hamburg, Germany
[quote="fenerit"]Well... that was because I wrote that: where is the delta-T? :wall: /quote]

Ah... :°

Here are a few more links about TT (Terestial Time) and its relation to other time standards.

https://en.wikipedia.org/wiki/Terrestrial_Time
http://www.staff.science.uu.nl/~gent011 ... deltat.htm
http://aa.usno.navy.mil/faq/docs/TT.php
http://stjarnhimlen.se/comp/time.html
https://www.cv.nrao.edu/~rfisher/Ephemerides/times.html

Notably these linear approximations are perhaps useful:
  • TT is ahead of TAI, and can be approximated as TT ≅ TAI + 32.184 seconds
  • TT runs a little ahead of UT1 (a refined measure of mean solar time at Greenwich) by an amount known as ΔT = TT − UT1. ΔT was measured at +67.6439 seconds (TT ahead of UT1) at 0h UTC on 1 January 2015;[
  • TT is in effect a continuation of (but is more precisely uniform than) the former Ephemeris Time (ET)

Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Sat, 23-04-16, 8:29 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4439
Location: Hamburg, Germany
John Van Vliet wrote:
a few minuets ether way is rather good
considering the code dose not use a huge amount of decimal places

for running at Aprox. 60 FPS, a few min. is rather GOOD


I agree.

*********************************
A real challenge for people believing to have precise orbits of Pluto and Charon :clap: is to check the predictions of their mutual eclipses given the new data about the Pluto-charon system!
*********************************

Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Sat, 23-04-16, 11:31 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 385
Location: Tuscany, Tyrrhenian Sea
t00fri wrote:
Notably these linear approximations are perhaps useful:
  • TT is ahead of TAI, and can be approximated as TT ≅ TAI + 32.184 seconds
  • TT runs a little ahead of UT1 (a refined measure of mean solar time at Greenwich) by an amount known as ΔT = TT − UT1. ΔT was measured at +67.6439 seconds (TT ahead of UT1) at 0h UTC on 1 January 2015;[
  • TT is in effect a continuation of (but is more precisely uniform than) the former Ephemeris Time (ET)
Fridger


Is Celestia.sci accounting for these? Say - aside UT1-UTC - 32.184 + TAI-UTC -> 32.184 + 36 = 68.184 s at 1 July 1 2015 (in the previour post I've erroneusly wrote Jan), but within Celestia 1.61 and 1.70 the table ends at TAI-UTC = 35s. For the Espenak-Meeus algo. the delta-T should be 69.3s at mid-Jan 2015.

_________________
Never at rest.


Top
 Profile  
 
PostPosted: Sat, 23-04-16, 12:20 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4439
Location: Hamburg, Germany
fenerit wrote:
t00fri wrote:
Notably these linear approximations are perhaps useful:
  • TT is ahead of TAI, and can be approximated as TT ≅ TAI + 32.184 seconds
  • TT runs a little ahead of UT1 (a refined measure of mean solar time at Greenwich) by an amount known as ΔT = TT − UT1. ΔT was measured at +67.6439 seconds (TT ahead of UT1) at 0h UTC on 1 January 2015;[
  • TT is in effect a continuation of (but is more precisely uniform than) the former Ephemeris Time (ET)
Fridger


Is Celestia.sci

Celestia.sci ==> celestia.Sci :°
Quote:
accounting for these?

Yes (like Celestia 1.7.x for the time being). You can check yourself in the SVN archive for Celestia 1.7.x:
https://sourceforge.net/p/celestia/code ... /astro.cpp
Quote:
Say - aside UT1-UTC - 32.184 + TAI-UTC -> 32.184 + 36 = 68.184 s at 1 July 1 2015 (in the previour post I've erroneusly wrote Jan), but within Celestia 1.61 and 1.70 the table ends at TAI-UTC = 35s. For the Espenak-Meeus algo. the delta-T should be 69.3s at mid-Jan 2015.


Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Sat, 23-04-16, 12:46 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4439
Location: Hamburg, Germany
Note: concerning the mutual event times as calculated by Dr. Arlot, he writes in his paper (cf above link)

The (event) dates are provided in Terrestrial Time (TT) since the
UTC for this period is not yet available. The difference
TT – UTC will, however, be near 65 s in 2003. Note that
observations will be recorded referred to UTC.


Since accuracy better than 1 second is not required in these mutual event tests, you can
use UTC = UT1 (with occasional leap seconds added!) and thus

you may convert Arlot's event predictions in TT approximately to UTC as follows:
*****************************************************
UTC = TT - ΔT = TT - 67.6439sec
*****************************************************
Hence UTC runs about 1 minute behind the given event times in TT!

_________________
Image


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 28 posts ]  Go to page 1, 2  Next

All times are UTC


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group