MORE THAN YOU PROBABLY WANTED TO KNOW ABOUT ECLIPSES


***********************************************************

SEQUENCE OF EVENTS TO OBSERVE IN A SOLAR ECLIPSE

-80 minutes:
watch the partial phase shrinking (USE ONLY SAFE METHODS - see below)

-5 minutes:
shadow bands on ground and walls
strange shadows and spots of sunlight

minus few seconds:
diamond ring, Baily's Beads

minus one second:
chromospheric (pink-red) flash (you can look directly AFTER this)

totality:
see the prominences and sun corona
see the planets and stars
notice the terrestrial environment

mid-eclipse:
corona at best now, block central to see outer part
with binoculars, can you see features on the moon?
notice: wind stopped, cooler, birds quiet

chromospheric flash-out:
bring back filter viewing IMMEDIATELY
watch 5 more minutes for shadow bands on ground and walls

***********************************************************

MORE ON WHAT IS VISIBLE

SHADOW BANDS: These are lighting fluctuations on the ground caused by density fluctuations in the ocean of air above us. They arise from the localized gathering and spreading of the paths of the sun's rays, and appear as parallel bright and dark bands of light in alternation, some 10 to 30 cm in width, moving at a few meters per second, and lasting a few seconds to about a minute.

PHOTOSPHERE: This is the visible disk of the sun, the layer of gas from which the photons of light, emitted by the hot gas, at last escape into space instead of being recaptured again in the gas. About 300 km thick. Never look at it directly except at sunrise, sunset, or through thick fog. We don't see the gas deeper in than the photosphere because the photosphere is opaque.

CHROMOSPHERE: Meaning "colour sphere" because of its pink-red colour, this quite transparent gas lies above the photosphere, about 4000 km thick and with an outer edge made jagged by 10,000 km long "spicules" or little spikes. Never visible except during a solar eclipse.

CORONA: Very rarefied gas one thousand times hotter than the sun's surface but a million times fainter than it. Its shape is molded by magnetic field forces which guide the current of stream of solar particles outward. Near the maximum of the sunspot cycle as now, the corona is stretched out in all directions.

***********************************************************

EYE SAFETY

There is great danger of retinal burning (permanent eye damage) from doing the wrong thing, even if the view looks dim (the infrared rays still get through) and even if you don't feel discomfort. Follow these rules for safety.

(1) Never look directly until the Bailey's Beads and chromosphere have disappeared. You can still see the latter through your filter. Take up your filter again the instant the returning side of the sun starts to brighten up, before the sun disk reappears.

(2) Never leave an optical instrument pointed at the sun which bystanders, e.g. children, may be tempted to look through.

(3) Use only a safe filter to look at the sun's disk, and even then look for some seconds at a time, not continuously with prolonged staring.

(4) Indirect viewing is always to be recommended:
(a) projection through a tiny hole onto a shaded surface
(b) projection through an optical instrument onto a shaded surface
(c) reflection off a small mirror onto a shadowed wall some distance off

(5) Direct viewing is safe through:
(a) commercial mylar solar filters, but CHECK FOR FLAWS before using
(b) number 14 welders glass
(c) certain exposed photographic Ag-based glass plates

(6) Definitely unsafe are:
(a) most exposed films and negatives
(b) smoked glass, sunglasses, photographic filters
(c) solar filters provided with cheap telescopes, UNLESS you stop down the telescope opening
and are on guard for the filter cracking, which can easily happen
(d) REMEMBER: just because the sun looks dim through a filter does not prove it is safe; the infrared light still can be getting through

***********************************************************

PHOTOGRAPHY

IN GENERAL. The solar image in mm is approximately equal to the telephoto focal length, divided by 108. Hence a 500mm telephoto is adequate but not less. Test out your tripod stability and install fresh batteries ahead of time. Don't spend the whole time looking at your equipment, not the eclipse. Professional photographs will be available anyway.

TOTALITY PHOTOGRAPHY. Ahead of time, use the full moon for focus and exposure trials. Don't use Neutral Density filters. Turn off the automatic flash as it interferes with and irritates totality watchers. Bracket the exposure times. Tips: ASA 50-100, not ND, 1/60 second (prominences) to 1/2 second (corona).

PARTIAL PHASE PHOTOGRAPHY. Ahead of time, use the sun for focus and exposure trials. Do the partial phase exposures after totality to avoid interfering. Need to use Neutral Density filters. Bracket the exposure times. Tips: ASA 50-100, ND 5.0, f/8, 1/125 second.

***********************************************************

PARTS OF A SHADOW

Imagine the sun, on the left, as the light source and the moon, on the right, as the light "blocker". It helps to see (or draw) a diagram for this. There will be a cone of shadow to the right of the moon, called the umbra. From anywhere in the umbra, you can't see the sun, because the moon blocks it. If you draw a line from the top of the sun to the top of the moon, and carry on to the right, that is the top edge of the umbra, and a line from the bottom of the sun to the bottom of the moon, carried on to the right, is the bottom edge of the umbra.

Outside the umbra there is a broader region of space called the penumbra. From anywhere in the penumbra, you can see some, but not all, of the sun. (A line from the bottom of the sun to the top of the moon, and carried on to the right: this is the top edge of the penumbra. A line from the top of the sun to the bottom edge of the moon, and carried on to the right: this is the bottom edge of the penumbra). The degree of darkness within the penumbra varies greatly, and smoothly, from nearly no darkening at the outer edge, to nearly as dark as the umbra, near the umbra. If you hold an object, e.g. your hand, near a wall, in direct sunlight (or torch-light in a darkened place), you can see these effects.

The umbra of the moon stretches about about 380 thousand km behind it, just about far enough to reach the earth. The umbra of the earth stretches out 1.4 million km behind it, about 3 1/2 times farther than the distance to the moon. These umbrae are very small, less than one percent, compared with the distance between the earth and the sun, 150 million km.

More technically, by drawing lines from the top, center, and bottom of the sun, through the top, and bottom, of the blocker (which could be the moon or the earth or anything), one can see that the shadow has a number of regions, defined by where these lines cross. Specifically, it has THREE REGIONS WHERE THE BLOCKER LOOKS BIGGER THAN THE SUN:

(1) and where it covers all of the sun (= the UMBRA)
(2) and where it covers the center, but not all, of the sun
(3) and where it covers some of the sun, but not the center

AND FOUR REGIONS WHERE THE BLOCKER LOOKS SMALLER THAN THE SUN:

(4) and it does cover the center of the sun, but doesn't extend over the edge
(5) and it doesn't cover the center of the sun and doesn't extend over the edge
(6) and it does cover the center of the sun and does extend of the edge
(7) and it does not cover the center of the sun and does extend over the edge

The locations of all of these regions can be precisely and very simply calculated in terms of only 4 angles, called the solar parallax, the solar semidiameter, the lunar parallax, and the lunar semidiameter.

***********************************************************

ECLIPSE OCCURRENCE THRESHOLD

If                     crosses the shadow of            you have a
--                     ---------------------            ---------- 
earth (leading edge)       moon (zone 1)            total solar eclipse

earth (leading edge)       moon (zone 4 or 5)       annular solar eclipse

earth (leading edge)       moon (zone 2,3,6 or 7)   partial solar eclipse
  
moon (trailing edge)       earth (zone 1)            total lunar eclipse 

moon (leading edge)        earth (zone 1)           partial lunar eclipse 

moon (leading edge)        earth (zone 2 or 3)     penumbral lunar eclipse

(An annular eclipse is like a total one except the moon appears too small to cover the entire sun's disk, because the moon is too far from the earth on the occasion. Anulus is Latin for "ring" and a ring-like circle of sun is still visible entirely surrounding the moon even at maximum annular eclipse.)

The different kinds of eclipses can be visualized as follows in terms of their appearance at maximum eclipse, the moment of closest approach on the sky as seen from the earth:

total solar eclipse  ...   none of the sun extends beyond the edge of the moon
annular solar eclipse      none of the moon extends beyond the edge of the sun
partial solar eclipse      some of the sun extends beyond the edge of the moon
                       and some of the moon extends beyond the edge of the sun
total lunar eclipse        all of the moon lies within the earth's umbra
partial lunar eclipse      some of the moon lies within the earth's umbra
penumbral lunar eclipse    some of the moon lies within the earth's penumbra

***********************************************************

ECLIPSE LIMITS

How close must the moon center and sun center (as seen from the center of the earth) approach each other to produce each kind of solar eclipse? And how close must the moon center and anti-sun point (i.e. the center of the earth's shadow) on the sky approach each other to produce each kind of lunar eclipse?

Here are some of the technical details, but don't worry about them. For the eclipse on the left, the bodies must approach to within the angular separation (as seen from the center of the earth) given on the right:

eclipse type:    body centers must approach within an angle:
-------------    --------------------------------------------
lunar penumbral: 1.02(p + P + S) + s       (the 1.02 is due to the earth's
lunar partial  : 1.02(p + P - S) + s     atmosphere making the earth "bigger")
lunar total    : 1.02(p + P - S) - s
solar partial  :      s + S + p - P
solar annular  :      S - s + p - P
solar total    :      s - S + p - P

where
P is the sun  parallax,     = 8.8"
S is the sun  semidiameter, = 16.0' (varies between 15.8' and 16.3')
p is the moon parallax,     = 57'   (varies between 54' and 62')
s is the moon semidiameter  = 15.5' (varies between 14.7' and 16.8')

( note: an arcminute ' is a 1/60th of a degree
  and an arcsecond " is a 1/60th of an arcminute)

( P is an angle of about 1 part in 23500, semiamplitude range of about 1.7%
  S is an angle of about 1 part in 215  , same range
  p is an angle of about 1 part in 60   , semiamplitude range of about 6.8%
  s is an angle of about 1 part in 222  , same range)

And so the values turn out to be:
lunar penumbral: 90.1' (86.0',96.8')
lunar partial  : 57.4' (53.3',64.0')
lunar total    : 26.4' (21.8',32.5')

solar partial  : 88.4' (84.4',95.0')
solar annular  : 57.4' (53.9',63.5')
solar total    : 57.0' (53.9',62.9')

This means that, e.g, a total solar eclipse will occur if the moon lines up to within about 57 arcminutes (nearly a degree, 60 arcmin) of the sun, and a partial solar eclipse will occur if it lines up to within about 88 arcmin, or about 1 1/2 degrees, from the sun. Similarly, a total lunar eclipse will occur if the moon lines up to within about half a degree of the center of the earth's shadow, a partial lunar if within about a degree of the center of the earth's shadow, and a penumbral lunar if within about 1 1/2 degrees of the center of the earth's shadow.

(Incidentally, for the geometrically minded, the solar eclipse limits can be found from a diagram with the moon drawn as outwardly tangent (for partial eclipses) or inwardly tangent (for total eclipses) to the outer tangent between the earth and sun, and making use of the law of exterior angles. The lunar eclipse limits can be found from a diagram with the moon drawn as outwardly tangent to the earth's penumbra, or umbra (for penumbral and partial lunar eclipses), or inwardly tangent to the earth's umbra (for total lunar eclipses), again making use of the law of exterior angles.)

***********************************************************

ECLIPSE SEASONS

The sun circles the starry sky taking 365.2422 days to get back to the same reference point (called the vernal equinox). The sun's path is called the ecliptic. This is essentially the plane of the solar system, in which the earth is going around the sun, but to us it looks as though the sun is going around the starry sky. The sun's "celestial longitude" is zero when it crosses the vernal equinox point on the sky, and increases with time until it completes 360 degrees, which is the same 0 degrees and the start of the new cycle.

The moon circles the earth taking 29.53 days (on average) to get back to the same reference point, which is when it overtakes, i.e. has the same celestial longitude as, the sun. This instant is called conjunction, or new moon. The moon's orbit is however tilted by 5.15 degrees compared with the ecliptic, i.e. the path of the moon is somewhat slanted on the sky compared with the path of the sun. Since we learned above that the two bodies must approach to within something like a degree to have an eclipse, we can see that at most conjunctions (new moons) and oppositions (full moons) throughout the year, the moon will pass way above or way below the sun or the earth's shadow, hence no eclipse happens.

There are two places where the moon's orbit crosses the ecliptic (two slanting great circles on the sky have to cross at two places 180 degrees apart), and these points are called the nodes (of the moon's orbit). There are only two short intervals each year when the sun is near one or other of the nodes. (Occasionally it can make it from the first node to the second, and back to the first, all within one calendar year, however.) That is why there are usually two, rarely three, "eclipse seasons" per year, when one hears about eclipses taking place in a burst of activity, typically 2 +- 1 eclipses, and then nothing happening for the 5-6 months or so in between.

How far can the sun, or the earth's shadow, be from the node and still have an eclipse? Remember that the angle between the sun's path and moon's path is quite shallow, so they don't have to meet EXACTLY at the node to eclipse. The geometry implies a "projection factor" about 11.1, so then:

There will be    if the moon approaches      which happens if the sun at
this kind of     within ... arcminutes of    conjunction or opposition is
eclipse:         the sun or earth's umbra    within ... degrees of the node:
---------------  ------------------------    -------------------------------
                                             
lunar penumbral    90.1'(86.0',96.8')          16.7(15.9,17.9) deg
lunar partial      57.4'(53.3',64.0')          10.6( 9.9,11.8)
lunar total        26.4'(21.8',32.5')           4.9( 4.0, 6.0)

solar partial      88.4'(84.4',95.0')          16.4(15.6,17.6)
solar annular      57.4'(53.9',63.5')          10.6(10.0,11.7)
solar total        57.0'(53.9',62.9')          10.5(10.0,11.6)

(The second column is from above. The third column is the second times 11.1 and divided by 60 arcminutes-per-degree. The parenthesis give the range of variation due to the changing distances between the earth, moon, and sun.)

So you see that out of the full possible 360 degrees, the sun has to be within +-10.5 degrees of the node of the lunar orbit in order for a total eclipse to occur, and so forth. Otherwise the moon's umbra will pass above or below the earth at conjunction.

So the eclipses in any one year occur in a "typical" but not rigid pattern. If "S" is a solar eclipse and "L" is a lunar eclipse, a typical year would be:

  S (..1/2 month..) L (..5 1/2 months..) S (..1/2 month..) L (..5 1/2 months..)
or
  L (..1/2 month..) S (..5 1/2 months..) L (..1/2 month..) S (..5 1/2 months..)

with each of the SL or LS pairs making an "eclipse season." One can understand the clumping of eclipses in pairs, and sometimes triples, as follows. In 1/2 month, the sun (and the earth's shadow) has moved 15.3 degrees relative to the lunar nodes, and in a whole month it has moved 30.7 degrees. We saw above that a solar eclipse can occur if the sun is within about a 33-degree-wide band centered on a node, and a lunar eclipse can occur if the earth's shadow (the anti-sun point) is within about a 21-degree-wide band centered on a node (disregarding lunar penumbral eclipses as rather insignificant).

During would-be successive eclipse pairs like S-L or L-S, the sun has moved about 15 degrees, and during triples like S-L-S, the sun has moved about 31 degrees, but the eclipse "window" for the sun's position near the node is wide enough to allow these repeats. Perhaps you can visualize this as the sun's position (at conjunction), or the earth's umbra position (at opposition) passing the lunar node thusly:

|                            LUNAR NODE                            |
|                               here                               |
|                                                                  |
|<---              SOLAR ECLIPSE IF sun anywhere                   |
|                     in this 33 degree window                 --->|
|                                                                  |
            |                                          |
            |<---  LUNAR ECLIPSE IF umbra anywhere     |
            |         in this 21 degree window     --->|

In terms of this picture, you can get a lone solar eclipse if it occurs with the sun near the lunar node. Fifteen degrees either side, the lunar eclipse can't occur since the earth's umbra is too far from the node and "falls outside the window". You can get a solar eclipse and lunar eclipse pair if they occur spaced somewhat to either side of the node, and you can get a S-L-S triple season if the solar eclipses occur just inside the two edges of the window.

The long-term statistics are as follows:

(1) in a typical year,
there are 2 solar eclipses
but there can be between 2 and 5
working out to an average of 2.38 solar eclipses per yr

there are two lunar eclipses (excl. penumbral)
but there can be between 0 and 2
working out to an average of 1.54 lunar eclipses per yr

(2) in a ten-year average, there are

6.6 total solar
1.1 annular-total solar
7.7 annular solar
8.4 partial solar

7.2 total lunar
8.3 partial lunar eclipses

***********************************************************

ECLIPSE CYCLES

The eclipses occurring one after another are very different one from another because the moon has a different closest approach to the sun or the earth's umbra, and the earth-sun-moon are at quite different points in their orbits and thus have several percent differences in their mutual distances. However by waiting long enough, in fact 18 years (called one Saros), there finally does occur an eclipse which is pretty nearly a repeat of the earlier one.

Say that at a particular new moon there is a solar eclipse. At most following new moons there won't be an eclipse because the new moon is too far above or below the sun, but at about 6 month intervals there will be an eclipse because the sun is near the node of the moon's orbit and thus the new moon must block off at least part of the sun. Each such eclipse will be very much different in character (kind of eclipse, duration, track on the earth). But the 223rd new moon (18 years later) since we started keeping track, there will come another eclipse nearly identical to the first, and this will happen again after another 18 years, and another, ... This is the Saros cycle or sequence of returning eclipses.

For the technically minded, I include the following information which is the essential foundation for understanding how the eclipse cycles arise. There are 6 "objects" whose eastward motion on the sky creates the eclipse cycles, due to their relative rates of motion.

The "objects" are: S(sun), M(moon), LP (lunar perigee), FS(fixed stars), VE(vernal equinox), LN(lunar node). Their relative rates of motion are:

M  -  LN :   2 Pi /  27.21222  rad/d       M  -  LP :   2 Pi /  27.55455  rad/d
S  -  LN :   2 Pi / 346.61979  rad/d       S  -  LP :   2 Pi / 411.78423  rad/d
LP -  LN :   2 Pi /   5.99696  rad/y       FS -  LP :  -2 Pi /   8.85062  rad/y
FS -  LN :   2 Pi /  18.59962  rad/y       VE -  LP :  -2 Pi /   8.84758  rad/y
VE -  LN :   2 Pi /  18.61305  rad/y       LN -  LP :  -2 Pi /   5.99696  rad/y
                                                                        
M  -  VE :   2 Pi /  27.32158  rad/d       M  -  S  :   2 Pi /  29.53059  rad/d
S  -  VE :   2 Pi / 365.24220  rad/d       LP -  S  :  -2 Pi / 411.78423  rad/d
LP -  VE :   2 Pi /   8.84758  rad/y       FS -  S  :  -2 Pi / 365.25636  rad/d
FS -  VE :   2 Pi /     25770  rad/y       VE -  S  :  -2 Pi / 365.24220  rad/d
LN -  VE :  -2 Pi /  18.61305  rad/y       LN -  S  :  -2 Pi / 346.61979  rad/d
                                                                        
M  -  FS :   2 Pi /  27.32166  rad/d       S  -  M  :  -2 Pi /  29.53059  rad/d
S  -  FS :   2 Pi / 365.25636  rad/d       LP -  M  :  -2 Pi /  27.55455  rad/d
LP -  FS :   2 Pi /   8.85062  rad/y       FS -  M  :  -2 Pi /  27.32166  rad/d
VE -  FS :  -2 Pi /     25770  rad/y       VE -  M  :  -2 Pi /  27.32158  rad/d
LN -  FS :  -2 Pi /  18.59962  rad/y       LN -  M  :  -2 Pi /  27.21222  rad/d

***********************************************************

SAROS CYCLE

What the above rates of motion are telling us is that:

(1) the moon comes back to same ecliptic longitude shift relative to
    the sun in a "synodic month" of  29.53059 days on average

    223 times later it is  6585.322 days   ( =  1 SAROS = 18.03 yr)

    which is 18 calendar years plus 10 1/3 days if there were 5 leap-yrs
                               plus 11 1/3 days if there were 4 leap-yrs
                        rarely plus 12 1/3 days if there were 3 leap-yrs
    thus moon will have the same PHASE as before

(2) the sun comes back to same ecliptic longitude shift relative to
    the LN in an "eclipse year" of  346.61979 days

    i.e. sun comes back to same ecliptic latitude shift relative to
    the lunar orbit at that ecliptic longitude

    19 times later it is  6585.776 days

    moon will have nearly same DISPLACEMENT above or below the sun as before


WITH (1) AND (2) THE MOON-SHADOW WILL MAKE A VERY SIMILAR PASS BY THE EARTH
as it did 223 months before

(3) the moon comes back to the same distance from the earth in an
    "anomalistic month" of  27.55455 days

    239 times later it is  6585.537 days

    thus the moon shadow will have nearly the same width as before because
    (a) earth-moon distance similar because nearly a whole number of
        anomalistic months
    (b) earth-sun distance similar because nearly a whole number of years

BECAUSE OF (1) AND (2) AND (3) HAPPENING ALL TOGETHER THE MOON-SHADOW WILL:

   -- Make a VERY similar pass by the earth, and
   -- Be of a VERY similar size, thus
   -- Will cause a VERY VERY similar eclipse

In summary, when the moon lines up with (above or below) the sun at new moon, 223 months later it will have a very similar offset above or below it, and the apparent sizes of the sun and the moon on the sky will also be very similar to before. Only a minority of the intervening 222 new moons will produce an eclipse, and none of those will be particularly like the first eclipse, but the 223rd return to new moon will produce the "same" eclipse. The sequence of "same" eclipses coming every 18 years is called the Saros sequence. The same effect occurs for lunar eclipses at full moon but is usually ignored.

***********************************************************

SUCCESSIVE ECLIPSES IN THE SAROS SEQUENCE

ARE SIMILAR BUT NOT IDENTICAL

Just as twins can be closely similar but not identical, the successive eclipses in a single Saros sequence have distinguishing features.

(1) the Saros is not a precise whole number of days

=> the next one will occur shifted about 120 W in longitude since the earth has had an addition 1/3 day to rotate

(2) the Saros is slightly shorter than precisely 19 eclipse years

=> the next eclipsed sun is 28.3 arcmin (.472 deg) further W compared with the node of the lunar orbit

=> the Saros can't go on forever because the successive eclipses will eventually leave the "eclipse window" of positions near the node

(3) the Saros is slightly shorter than precisely 239 anomalistic months (time it takes for the moon to cycle back to the same apparent size)

=> eclipses cycle in moon size with period of 128 Saros steps, so the next eclipsing moon has a 0.78% change of favorability, occurring progressively earlier compared with perigee (the moon being near perigee produces the best eclipses)

(4) the Saros is slightly longer than precisely one sidereal year (time it takes for the sun to cycle back to the same apparent size)

=> eclipses cycle in sun size on the sky with period of 34 Saros steps so the next eclipsing sun has a 2.93% change of favorability, occurring progressively later compared with aphelion (the earth being near aphelion in early July produces the best eclipses)

A "family" of eclipses, a Saros sequence, is born with very weak eclipses, grows to "strong" eclipses as the sun-at-conjunction gradually approaches the lunar node, and then weakens and eventually disappears as the sun-at- conjunction heads away from the node and back out of the "eclipse window." A Saros sequence lasts for about 70 steps, or about 1300 years, which you can deduce from the size of the eclipse window (2 x 16.4 degrees) and the speed at which the Saros steps cross the window (0.472 deg per step).

It starts "weaker" with about 15 partial eclipses, then peaks with about 40 total (or annular) eclipses, then dies off finally with about 15 partials. Every eclipse going on currently is a member of its own Saros family, so there are about 42 Saros families running simultaneously. One Saros family, presumably the oldest, dies out and is replaced by a new one about every 29 years.

A newborn Saros sequence where the sun in approaching the descending node of the lunar orbit will begin with an eclipse near the earth's south pole and subsequent eclipses will occur approximately 200 km further northward, 18 years later. A newborn Saros sequence where the sun in approaching the ascending node of the lunar orbit will begin with an eclipse near the earth's north pole and subsequent eclipses will occur approximately 200 km further southward.

***********************************************************

ECLIPSE EVENT LINES ON THE GLOBE

Eclipse maps shown as lines on the earth can be somewhat confusing. The only fairly obvious part is the line of totality, which is the path of the moon's umbra crossing the surface of the earth, as both of these are moving in time. Some explanatory notes about the other lines is given here without much elaboration, other than to explain that the sunrise and sunset lines are the lines on the surface of the globe which are on the plane which is perpendicular to the direction to the sun.

ECLIPSE STARTS AT SUNRISE LINE =
globe locations where leading edge of penumbra intersects sunrise line;
moon starts covering top of sun the instant the sun breaks above horizon;
3rd furthest west line;br> occurs first in time

MAXIMUM ECLIPSE OCCURS AT SUNRISE LINE =
globe locations where "penumbra midline" intersects sunrise line;
moon is midway across sun as sun breaks above horizon;
2nd furthest west line;
occurs second in time

ECLIPSE ENDS AT SUNRISE LINE =
globe locations where trailing edge of penumbra intersects sunrise line;
moon departs the bottom of the sun as the sun breaks above the horizon;
furthest west line;
occurs third in time

ECLIPSE STARTS AT SUNSET LINE =
globe locations where leading edge of penumbra intersects sunset line;
moon starts covering bottom of sun the instant the top of sun goes beneath the horizon;
furthest east line;
occurs fourth in time

MAXIMUM ECLIPSE OCCURS AT SUNSET LINE =
globe locations where "penumbra midline" intersects sunset line;
moon is midway across sun as top of sun goes beneath the horizon;
2nd furthest east line;
occurs fifth in time

ECLIPSE ENDS AT SUNSET LINE =
globe locations where trailing edge of penumbra intersects sunset line;
moon departs the top of the sun just as the top of the sun goes beneath the horizon;
3rd furthest east line;
occurs sixth in time

The central track (for the umbra) is quite analogous to the preceding except in miniature.

The speed at which the shadow crosses the ground can be estimated. The speed of the moon's shadow in space, relative to the earth's center, is about 57 (with range of 54 to 61) km per minute.

The speed of the ground at the equator relative to the earth's center is about 28 km per minute. So the slowest eastward progress of the moon's shadow would be about 29 km per minute, but it is usually somewhere in between 29 and 57 because
(1) the ground moves more slowly at higher latitude
(2) the ground makes a steeper angle to the moon's shadow as the latter falls non-centrally on the earth
So 45 +- 10 km per minute is a fairly typical ground speed of the shadow.

The typical duration of solar eclipses can be estimated from the shadow's time to cross the earth, and is in the neighborhood of 3-4 hours. Totality at any one place on the earth's surface can last up to a possible 7:40 minutes, but is more typically shorter, half that time or less. The typical duration of lunar eclipses can be estimated from the moon's time to cross the earth's umbra (at a speed relative to the shadow of about 0.5 arcmin per minute) and is again in the neighborhood of 3-4 hours. Total lunar eclipse can last up to about 103 minutes.


John Caldwell   ,   Dave Laney


South African Astronomical Observatory
PO Box 9, Observatory, 7935, South Africa
Phone: (021) 447-0025 / Fax: (021) 447-3639
SAAO home page
http://www.saao.ac.za
Page updated 28 June 1999