This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Orbits and
Navigation T o FULLY UNDERSTAND and use satellite data it is necessary to under
stand the orbits in which satellites are constrained to move and the geometry with which they view the Earth. This chapter begins with a review of
basic physical principles which reveal the shape of a satellite orbit and how to
orient the orbital plane in space. This knowledge allows us to calculate the position
of a satellite at any time. Orbit perturbations and their effects on meteorological
satellite orbits are then discussed. Next the geometry of satellite tracking and
Earth location of the measurements made from the satellites are explored. This
leads to a discussion of space—time sampling. The chapter concludes with a brief overview of satellite launch vehicles and orbit insertion options. 2.] NEWTON'S LAWS Isaac Newton1 discovered the basic principles that govern the motions of
satellites and other heavenly bodies. 1 English physicist and mathematician, 1642—1727. 16 Chapter 2 Orbits and Navigation Newton’s Laws of Motion 1. Every body will continue in its state of rest or of uniform motion in a
straight line except insofar as it is compelled to change that state by an impressed force. ‘ .
2. The rate of change of momentum IS proportional to the 1mpressed force and takes place in the line in which the force acts.
3. Action and reaction are equal and opposrte. Since momentum is the product of the mass of a body and its velocity, Newton’s
Second Law is the familiar — — E 2.1
F—ma—mdt: ( ) where F is force, m is mass, at is acceleration, v is velocity, and tis time. In addition,
Newton gave us the functional form of the force that determines satelllte motion: Newton’s Law of Universal Gravitation The force of attraction between two point masses m1 and m2 separated by
. . Q.
a distance r 15 F = ——G":;m2, (2.2) where G is the Newtonian (or universal) gravitation constant (see Appen
dix E). Consider the simple circular orbit shown in Fig. 2.1: Assuming that the Earth
is a sphere, we can treat it as a point mass. The centrlpital force required to keep
the satellite in a circular orbit is mull/r, where v is the orbital velour); of the
satellite. The force of gravity that supplies this centripital force IS Gmem/r , where
me is the mass of the Earth (Appendix E) and m 18 the mass of the satellite. Equating the two forces gives
v2 Gmem. (23) r 1'2 _ . SATELLITE FIGURE 2.1. A circular satellite orbit. 2.2 Keplerian Orbits Division by m eliminates the mass of the satellite from the equation, which means
that the orbit of a satellite is independent of its mass. The period of the satellite is the orbit circumference divided by the velocity: T = Z‘JTf/U. Substituting in Eq.
2.3 gives 4172
T’ = r3. 2.4
Gme ( ) The current NOAA satellites orbit at approximately 850 km above the Earth’s
surface.2 Since the equatorial radius of the Earth is about 6378 km, the orbit
radius is about 7228 km. Substituting in Eq. 2.4 shows that the NOAA satellites
have a period of about 102 min. As a second example, we calculate the radius required for a satellite in geosyn
chronous orbit, that is an orbit in which the satellite has the same angular velocity
as the Earth. The angular velocity of a satellite is Far. T (2.5)
Substituting Eq. 2.5 in Eq. 2.4 gives
1'3 = (2.6) Inserting the angular velocity of the Earth (Appendix E), the required radius for a
geosynchronons orbit is 42,164 km, or about 35,786 km above the Earth’s surface. 2.2 KEPLERIAN ORBITS Satellites, however, do not travel in perfect circles, although a circular orbit is
the goal for most meteorological satellites. It is possible to derive the exact form
of a satellite’s orbit from Newton’s laws of motion and the law of universal gravitation.3 The results of this derivation are neatly summarized in Kepler’s laws
and in Kepler’s equation. 2.2.] Kepler's laws Johannes Kepler4 died 12 years before Newton was born and, therefore, did not
have the advantage of Newton’s work. Kepler formulated his laws by analyzing
a mass of data on the position of the planets. This task was complicated by the
rotation of the Earth and the motion of the Earth about the sun, which make 2 Speciﬁcations call for them to orbit at either 833 or 870 km; 850 km is a representative value.
3 The reader is referred to Escobal (1965) and Goldstein (1950) for two quite different, but equally
lengthy, derivations. " German astronomer, 1571—1630. 18 Chapter 2 Orbits and Navigation planetary motions seem very complex. In modern form, Kepler’s laws may be
stated as follows: Kepler ’5 Laws 1. All planets travel in elliptical paths with the sun at one focus. 2. The radius vector from the sun to a planet sweeps out equal areas in
equal times. 3. The ratio of the square of the period of revolution of a planet to the
cube of its semimaior axis is the same for all planets revolving around the sun. The same laws apply if we substitute satellite for planet and Earth for sun.
Equation 2.4 is a statement of Kepler’s third law for the special case of a circu lar orbit. 2.2.2 Ellipse Geometry The parameters that are used to specify satellite orbits are based in part on
geometric terminology. Figure 2.2 illustrates the geometry of an elliptical orbit.
The point where the satellite most closely approaches the Earth is the perigee, or
more generally, the perifocus. The point where the satellite is furthest from the
Earth is called the apogee or apofocus. The distance from the center of the ellipse
to the perigee (or apogee) is the semimajor axis and will be denoted by the symbol
a. The distance from the center of the ellipse to one focus (to the center of the
Earth) divided by the semimajor axis is the eccentricity and will be denoted by
the symbol a. For an ellipse, the eccentricity is a number between zero and one
(0 s e < 1). A circle is an ellipse with zero eccentricity. The equation for the
ellipse, that is, the path that the satellite follows, is given in polar coordinates . SATELLITE PERIGEE APOGEE
(PERIFOCUSl lAPOFOCUS) FIGURE 2.2. Elliptical orbit geometry. 2.2 Keplerian Orbits with the Earth as origin as _ all — 82) r——n
1+ac050 (2.7) The angle 6 is the true anomaly and is always measured counterclockwise (the
direction of satellite motion) from the perigee. 2.2.3 Kepler's Equation A satellite in a circular orbit undergoes uniform angular velocity. By Kepler’s
Second Law, however, a satellite in an elliptical orbit cannot have uniform angular
velocity; it must travel faster when it is closer to Earth. The position of the satellite
as a function of time can be found by applying Kepler’s equation: M = n(t — tp) = e — e sine, (2.8) where M is the mean anomaly; M increases linearly in time at the rate n, called
the mean motion constant, given by _2_1r_ Gm.
— T — 43  (2.9) 11: By deﬁnition M is zero when the satellite is at perigee; therefore, t1, is the time of perigeal passage. The angle e is the eccentric anomaly. It is geometrically related
to the true anomaly (Fig. 2.3): _ cose — a cost) — ——1 _ 8 cose’ (2.10a)
= c050 + s cose ——1 + a GOSH (2.10b) CIRCUMSCRIBED
CIRCLE ELLIPTICAL
ORBIT FIGURE 2.3. The geometric relationship between true anomaly (6) and eccentric anomaly (e). 20 Chapter 2 Orbits and Navigation 2.2.4 Orientation in Space By calculating r and 8 at time t, we have positioned the satellite in the plane
of its orbit; now we must position the orbital plane in space. To do so requires
the deﬁnition of a coordinate system. This coordinate system must be an inertial
coordinate system, that is, a nonaccelerating system in which Newton’s Laws of
Motion are valid. A coordinate system ﬁxed to the rotating Earth is not such a
system. We will adopt an astronomical coordinate system called the right ascen
sion—declination coordinate system.5 In this system (Fig. 2.4) the z axis is aligned
with the Earth’s spin axis. The x axis is chosen such that it points from the center
of the Earth to the sun at the moment of the vernal equinox, when the sun is
crossing the equatorial plane from the Southern Hemisphere to the Northern
Hemisphere.6 The y axis is chosen to make it a righthanded coordinate system.
In this system, the declination of a point in space is its angular displacement
measured northward from the equatorial plane, and the right ascension is the
angular displacement, measured counterclockwise from the x axis, of the projec
tion of the point in the equatorial plane (Fig. 2.5). Three angles are used to position an elliptical orbit in the right ascension—decli
nation coordinate system (Fig. 2.6): the inclination angle, the right ascension of
ascending node, and the argument of perigee. The inclination angle (i) is the angle between the equatorial plane and the
orbital plane. By convention, the inclination angle is zero if the orbital plane
coincides with the equatorial plane and if the satellite rotates in the same
direction as the Earth. If the two planes coincide but the satellite rotates
opposite to the Earth, the inclination angle is 180°. Prograde orbits are those with inclination angles less than 90°; retrograde orbits are those with i greater
than 90°. The ascending node is the point where the satellite crosses the equatorial plane
going north (ascends). The right ascension of this point is the right ascension of
ascending node (9). It is measured in the equatorial plane from the x axis (vernal
equinox) to the ascending node. In practice, the right ascension of ascending node
has a more general meaning. It is the right ascension of the intersection of the
orbital plane with the equatorial plane; thus it is always deﬁned, not just when
the satellite is at an ascending node. Finally, the argument of perigee (cu) is the angle measured in the orbital plane
between the ascending node (equatorial plane) and the perigee. 5 Because the origin of this coordinate system moves about the sun with the Earth, it is not truly
inertial. However, the sun’s gravity causes the satellite to rotate around the sun as does the Earth.
Therefore, the satellite acts as if the right ascension—«declination coordinate system were inertial. 6 This x axis is also referred to as the First Point ofAries because it used to point at the constellation
Aries. Because of the inﬂuence of the sun and moon on the nonspherical Earth, the Earth’s spin axis
precesses like a top with a period of 25,781 years. This causes the vernal equinox to change. Today,
the x axis points to the constellation Pisces, but it is still referred to as the First Point of Aries. 2.2 Ke lerian Orbits AUTUMNAL
EQUINOX SUMMER SOLSTICE SOLSTI CE VERNAL
EQUINOX FIGURE 2.4. The right ascension—declination coordinate system. X FIGURE 2.5. Coordinates used in the right ascension—declination coordinate system: right ascension
(0,), declination (6), and radius (r). EARTH'S
SPIN
AXIS
z A, .. 4,...——PERIGEE
/
A;
CENTER ORBIT
0F EARTH EQUATORIAL
PLANE ASCENDING
NODE FIGURE 2.6. Angles used to orient an orbit in space. 22 Chapter 2 Orbits and Navigation 2.2.5 Orbital Elements The justdiscussed parameters for location of a satellite in space are collectively
known as the classical orbital elements (Table 2.1) or as Brouwer mean orbital
elements (Brouwer and Clemence, 1961). These parameters may be determined
by optical, radar, or radio ranging observations or by matching features of known
locations on the Earth’s surface (landmarks) with observations made by instru
ments on the satellite (Dubyago, 1961; Escobal, 1965). The orbital elements for
particular satellites are available from the agencies that operate them: NOAA or
NASA in the United States,7 the European Space Agency (BSA) in Europe, etc. A
ﬁnal parameter, included in Table 2.1, is the time when these elements are observed
or are “valid.” This time is called the epoch time (to). Some of the orbital elements
change with time, as we shall see below. A subscript “0” on an orbital element indicates a value at the epoch time.
There is some variation in how the orbital elements are speciﬁed. BSA, for example, substitutes true anomaly for mean anomaly. Also, in less formal descrip
tions of satellite orbits, one frequently sees the height of the satellite above the
Earth’s surface substituted for the semimajor axis. Since the Earth is not round,
the height of a satellite will vary according to its position in the orbit. Specifying
the semimajor axis is a much better way to describe a satellite orbit. Orbits in which the classical orbital elements (except M) are constant are called
Keplerian orbits. Viewed from space, Keplerian orbits are simple. The satellite
moves in an elliptical path with the center of the Earth at one fOCus. The ellipse
maintains a constant size, shape, and orientation with respect to the stars (Fig
2.7a). Perhaps surprisingly, the only effect of the sun’s gravity on the satellite is
to move the focus of the ellipse (the Earth) in an elliptical path around the sun
(the Earth’s orbit). Viewed from the Earth, Keplerian orbits appear complicated because the Earth
rotates on its axis as the satellite orbits the Earth (Fig. 2.8). The rotation of the
Earth beneath a ﬁxed orbit results in two daily passes of the satellite near a point
on the Earth (assuming that the period is substantially less than a day and that
the inclination angle is greater than the latitude of the point). One pass occurs
during the ascending portion of the orbit; the other occurs during the descending
portion of the orbit. This usually means that one pass occurs during daylight and
one during darkness. 2.3 ORBlT PERTURBATIONS Although satellites travel in nearly Keplerian orbits, these orbits are perturbed
by a variety of forces (Table 2.2). Forces arising from the last ﬁve processes are
small and can be viewed as causing essentially random perturbations in the orbital
elements. Operationally they are dealt with simply by periodically (1) observing 7NCAA orbital elements for the polarorbiting satellites are broadcast in the form of “TBUS
bulletins.” Barnes and Smallwood (1982) explain how to interpret these bulletins. TABLE 2.1. Classical Orbital Elements Element Symbol Semimajor axis a
Eccentricity e
Inclination i
Argument of perigee mu
Right ascension of ascending node .00
Mean anomaly MD
Epoch time t0 FIGURE 2.7. The change with season of (a) a Keplerian orbit and (b) a sunsynchronous orbit. 24 Chapter 2 Orbits and Navigation FIGURE 2.8. The orbit of a representative satellite as viewed from a point rotating with the Earth. the orbital elements and (2) adjusting the orbit with onboard thrusters. Forces
due to the nonspherical Earth cause secular (linear with time) changes in the
orbital elements. These forces can be predicted theoretically and indeed are useful. The gravitational potential of Earth is a complicated function of the Earth’s
shape, the distribution of land and ocean, and even the density of crustal material.
As a ﬁrstorder correction to a spherical shape, we may treat the Earth as an
oblate spheroid of revolution. In cross section the Earth is approximately elliptical.
The distance from the center of the Earth to the equator is, on average, 6378.140
km, whereas the distance to the poles is 6356.755 km. One can think of the Earth
as a sphere with a 21kmthick “belt” around the equator. The gravitational
potential of the Earth is approximately given by 2
U = _ GT“ [1 + %]2 (1 — 3 sinza) + , (211) where re, is the equatorial radius of the Earth, 8 is the declination angle, and2 is the coefﬁcient of the quadrupole term (Appendix E). The higherorder TABLE 2.2. Orbit Perturbing Forces Force Source Nonspherical gravitational ﬁeld Nonspherical, nonhomogeneous Earth Gravitational attraction of auxiliary bodies Moon, planets Radiation pressure Sun’s radiation Particle flux Solar wind Lift and drag Residual atmosphere Electromagnetic forces Interaction of electrical currents in the satellite with the Earth’s magnetic ﬁeld 2.3 Orbit Perturbations terms are more than two orders of magnitude smaller than the quadrupole term
and will not be considered here, although they are necessary for very accurate
calculations. How does this belt of extra mass affect a satellite’s orbit? One might expect
it to cause the satellite to orbit at a different speed, and indeed it does. The time
rate of change of the mean anomaly (dM/dt) is given by the mean motion constant
n in the unperturbed orbit and by the anomalistic mean motion constant, E, in
a perturbed orbit. Considering only the quadrupole term, Escobal (1965)
shows that 2
cfj—t=ﬁ=rt [1 +%]2 (1 — 52)'3’2(1—%sin2i)]. (2.12) When the inclination angle is less than 54.7", n is greater than n; the satellite
orbits faster than it would in an unperturbed orbit. However, for larger inclina
tions, the satellite orbits more slewly than it otherwise would. Because the belt exerts an equatorward force, one might also expect that it
would have an effect on the inclination angle. This force, however, affects the
right ascension of the ascending node rather than the inclination angle. Just as
the force of gravity causes a top to precess rather than to fall over, so the attraction
of the belt causes the orbit to precess about the z axis rather than to change its
inclination angle. Escobal (1965) gives the rate of change of the right ascension
of ascending node as 2
(3—? = #5 B], (1 '— az)‘zcosi]. (213) The ﬁnal effect of the belt is to cause the argument of perigee to rotate or
precess. Escobal (1965) gives 2 The other three orbital elements, a, a, and 1', undergo small, oscillatory changes that may be neglected.
If SI Units are used, Eqs. 2.9, 2.12, 2.13, and 2.14 result respectively in values of n, n, dQ/dt, and dw/dt whose units are radians per second.
The anomalistic period of a perturbed orbit is simply (2.15) However, because M is measured from perigee, the anomalistic period is the time
for the satellite to travel from perigee to moving perigee. Of more use is the
synodic or nodal period, T, which is the time for the satellitemto travel from one
ascending node to the next ascending node. An exact value of Tmust be calculated 26 Chapter 2 Orbits and Navigation numerically; however, to very good approximation 1”": —21T—— (2.16) _ do)
(n+3) In summary, then, the ﬁrstorder effects of the nonspherical gravitational poten
tial of the Earth consist of a slow, linear change in two of the classical orbital
elements, the right ascension of ascending node and the argument of perigee, and
a small change in the mean motion constant. 2.4 METEOROLOGICAL SATELLITE ORBITS Nearly all present meteorological satellites are in one of two orbits, sunsynchro
nous or geostationary, but other orbits are also useful. 2.4.1 Sunsynchronous Orbits The nonspherical gravitational perturbation of Earth, far from being a problem,
has a very useful application. As shown in Fig. 2.7a, the angle between the lines
that join the sun and the ascending node to the center of the Earth changes in a
Keplerian orbit because the orbital plane is ﬁxed while the Earth rotates around
the sun. This causes the satellite to pass over an area at different times of the
day. For example, if the satellite passes over near noon and midnight in the spring,
it will pass over near 6:00 am and 6 :00 pm in the winter. Several problems result;
among them are (1) the data do not ﬁt conveniently into operational schedules,
(2) orientation of solar cell panels is difﬁcult, and (3) dawn or dusk visible images
may not be as useful as images made at other times. Fortunately, the perturbations
caused by the nonspherical Earth can be employed to keep the sun—Earth—satellite angle constant. The Earth makes one complete revolution about the sun (21: radians) in one
tropical year (31,556,925.9747 5). Thus the right ascension of the sun changes
at the average rate of 1.991064 X 10‘7 rad s‘1 (0.9856473° day“). If the inclina
tion of the satellite is correctly chosen, the right ascension of its ascending node
can be made to precess at this same rate. An orbit that is so synchronized with
the sun is called a sunsynchronous orbit. For a satellite with a semimajor axis of
7228 km and zero eccentricity, Eq. 2.13 requires an inclination of 988° to be
sunsynchronous. Figure 2.7b shows the change with season of a sunsynchro
nous orbit. Because the sun—Earth—ascending node angle is constant3 in a sunsynchronous
orbit, the satellite is often said to cross the equator at the same local time every 3 Apart from small changes due to the elliptical orbit of the Earth. 2.4 Meteorological Satellite Orbits 27 day. Unfortunately, local time is an ambiguous term. We will use it to mean \I’ 15,. (2.17) where tU is coordinated universal time in hours and ‘I’ is the longitude in degrees
of a particular point.9 Equator crossing time (ECT) is the local time when a satellite crosses the equator: ‘PN
a i 2.18
ECT in + 150 ( )
where ‘PN is the longitude of ascending or descending node. If 111,," = —15°(r,, — 12) (2.19) is the longitude of the sun, and if AllI 5 ‘PN — ‘I’sun, it is easy to show that ECT = 12 + (2.20) If N? is constant, as it is for a sunsynchronous satellite, then ECT is constant. Sunsynchronous satellites are classiﬁed by their ECTs. Noon satellites (or
noon—midnight satellites) ascend (or descend) near noon LT (local time). They
must, therefore, descend (or ascend) near local midnight. Morning satellites ascend
(or descend) between 06 and 12 h LT, and descend (or ascend) between 18 and
24 h LT. Afternoon satellites ascend (or descend) between 12 and 18 h LT, and
descend (or ascend) between 00 and 06 h LT. The highest latitude reached by the subsatellite point (in any orbit) is equal to
the inclination angle (or the supplement of i, in the case of retrograde orbits).
Since sunsynchronous orbits reach high latitudes, they are referred to as near
polar orbits. This is frequently shortened to polar orbits, although they do not
cross directly over the poles. These orbits are also called low Earth orbits (LEOs)
to distinguish them from geostationary orbits (GEOs). Note, however, that polar
orbiter is a general term for a satellite that passes near the poles, and low Earth
orbiter is a general term for a satellite that orbits not far above the Earth’s surface.
While sunsynchronous satellites are of necessity polar orbiters and LEOs, the
converse is not necessarily true. The ground track of a satellite is the path of the point on the Earth’s surface
that is directly between the satellite and the center of the Earth (the subsatellite
point). Figure 2.9 shows the ground track for three orbits of the sunsynchronous NCAA 11 satellite. 9 The other use of local time refers to the time on one’s watch, that is, the time in a particular
time zone. Time zones are deﬁned as areas where time is agreed to be the local time (in our sense)
on a particular meridian. Eastern Standard Time, for example, is the local time on the 75“ west
meridian (‘1' = —75°). 23 Chapter 2 Orbits and Navigation NCAA  I
Three Orbits on 22 March [990 Start time= 0258 UTC End iime= 0804 UTC o =7229.606 km no = 29.3059° l = 9897446" (00 = l67.74754° h m s
6 = 0.00I995B EDOCh time = 22 Mar I990 I I5 52.353 UTC
Mo=l92.2866° Nodal Period = 02.0764 min FIGURE 2.9. The ground track of a typical sunsynchronous satellite. 2.4.2 Geostationary Orbits In Section 2.1 we calculated the radius of a geosynchronous orbit to be 42,164
km. Perturbations due to the nonspherical Earth, however, require a slight adjust
ment in this ﬁgure. The adjustment is small because geosynchronous orbit is about
6.6 Earth radii, and the correction terms are inversely proportional to the square
of this ratio. For an orbit with zero eccentricity and zero inclination, Eqs. 2.12,
2.14, and 2.16 require a semimajor axis of 42,168 km to be geosynchronous. The terms geosyncbronous and geostationary are often used interchangeably.
In fact, they are not the same. Geosynchronous means that the satellite orbits with
the same angular velocity as the Earth. Geostationary orbit is geosynchronous, but
it is also required to have zero inclination angle and zero eccentricity. Geostation
ary satellites, therefore, remain essentially motionless above a point on the equator.
They are classiﬁed by the longitude of their subsatellite point. Second—order perturbations cause a geostationary satellite to drift from the
desired orbit. Periodic maneuvers, performed as frequently as once a week, are
required to correct the orbit. These maneuvers keep operational geostationary
satellites very close to the desired orbit. For example, on 11 March 1990, the
GOES 7 satellite had an inclination angle of 0.05°; therefore, it did not venture
more than 005° latitude from the equator. Figure 2.10 sh0ws the ground track
for a geostationary satellite that is no longer used for imaging and therefore whose
orbit is not so carefully maintained. " mgr...“ .. 2.4 Meteorological Satellite Orbits GOES 6
3m (One Orbit on 6 March l990) 2°N APOGEE 0520
l42l85.l2 km) °N E0 I°S 2‘5 3°S 37°W 36°W 35°W I34‘W 33°W I32°W a=427L798 km 510: 30259“ i = l.973l0' We = 22139!" 9900031??? Epoch time = 6 Mar I990 a." a‘“ 20' uTc
Mo I4? 020 Nodal Period = I436. 297mm FIGURE 2.1!]. The ground track of a geostationary satellite. Note that the satellite’s orbit is not quite
geostationary; it drifts west slightly each day. 2.4.3 Other Orbils Geostationary and sunsynchronous are only two of infinite possible orbits.
Others have been and will become useful for meteorological satellites. The Earth Radiation Budget Satellite (ERBS) was launched from the Space
Shuttle and orbits at an altitude of 600 km with an inclination angle of 57°. It
was placed in this orbit so that it would precess with respect to the sun and
sample all local times (see Section 2.6) over the course of a month. The former Soviet Union placed its Meteor satellites in low Earth orbit with
inclination angles of about 82° (see Appendix A). The former Soviet Union also
used a highly elliptical orbit for Molniya communications satellites. It has been " suggested that this orbit would be useful for meteorological observations of the high latitudes (Kidder and Vonder Haar, 1990). The Molniya orbit has an inclina—
tion angle of 63.4"; at which the argument of perigee is motionless (Eq. 2.14);
thus the apogee, from which measurements are made, stays at a given latitude.
The semimajor axis is chosen such that the satellite makes two orbits while the
Earth turns once with respect to the plane of the orbit. The eccentricity is made
as large as possible so that the satellite will stay near apogee longer. However,
the eccentricity must not be so large that the satellite encounters significant atmo
spheric drag at perigee. A semimajor axis of 26,554 km and an eccentricity of
0.72 result in a perigee of 7378 km (1000 km above the equator), an apogee of
45,730 km (39,352 km above the equator), and a period of 717.8 min. The 30 Chapter 2 Orbits and Navigation attractiveness of this orbit is that it functions as a highlatitude, parttime, nearly
geostationary satellite. For about 8 h centered on apogee, the satellite is synchro
nized with the Earth so that it is nearly stationary in the sky. For a meteorological
satellite in a Molniya orbit, the rapid imaging capability, which is so useful from
geostationary orbit, would be available in the high latitudes. As meteorological satellite instruments become more specialized, more custom
orbits are likely to be used. 2.5 SATELLITE POSITIONING, TRACKING, AND NAVIGATION It is important to be able to calculate the position of a satellite in space, to
track it from Earth, and to know where its instruments are pointing. These topics
are discussed in turn in this section. 2.5.1 Positioning in Space To locate a satellite in a perturbed orbit at time t, one needs current values of
the orbital elements. The three constant elements, a, e, and i, are taken directly
from a recent bulletin.10 The other three, M, 0., and w, are calculated: M=Mo+‘ﬂI—4(t— to), (2.2121)
dt
.0. = o, + g—Qu  to), (2.21b)
dt
_ Q _
ro—ro0+ dt (t to). (2.21c) Then the satellite is positioned by one of several methods. We ﬁnd two methods
useful: the vector rotation method and the spherical geometry method. 2.5.“ The Vector Rotation Method Figure 2.11 illustrates what we call the vector rotation method. It is discussed
in a somewhat different form by Escobal (1965) and others. In the ﬁrst step, the
satellite is located in the plane of its orbit; that is, the true anomaly 9 and the
radius r are calculated. This is done by (1) solving for e using Eq. 2.8, (2) calculating
6 using Eq. 2.10a, and (3) calculating 1' using Eq. 2.7. (For a circular orbit, this
step is simpliﬁed because the mean anomaly, the eccentric anomaly, and the true
anomaly are identical, and r is constant.) In the second step, a vector is formed that points from the center of the Earth
to the satellite in the right ascension—declination coordinate system. The Cartesian 1° Such bulletins are available from a variety of sources. Because these sources change rapidly, we su est that the interested reader contact the a encies listed in Section 4.4 to ﬁnd a convenient source
as I _ s of satellite bulletins. 2.5 Satellite Positioning, Tracking, and Navigation FIGURE 2.11. Rotations used to position a satellite in its orbit: (a) the satellite in the plane of its orbit,
(b) rotation about the z axis through the argument of perigee (to), (c) rotation about the x axis through the inclination angle (i), and (d) rotation about the z axis through the right ascension of ascending
node (.0). coordinates of this vector are x r cos!)
y = r sin0  (2.22)
z 0 At this point, the orbital ellipse is assumed to lie in the 36—)! plane with the perigee
on the positive x axis (Fig 2.11a). In the ﬁnal three steps, the vector is rotated so that the orbital plane is properly
oriented in space. In the third step, the vector is rotated about the z axis through the argument
of perigee (Fig. 2.1 1b). This rotation is conveniently accomplished by multiplying 32 Chapter 2 Orbits and Navigation the vector by a rotation matrix, in this case x' cosw —sinw 0 x x cosw — y since
y' = since cosm 0 y = x since + y cosm . (2.23)
z’ 0 O 1 z z In the fourth step, the vector is rotated about the x axis through the inclination
angle (Fig. 2.1 1c). x" 1 0 0 x’ x’
y” = 0 cosi 'sini y' = y’ cosi — z' sini . (2.24)
z" 0 sini cosi z' y' sini + z' cosi In the ﬁfth and ﬁnal step, the vector is rotated about the z axis through the
right ascension of the ascending node (Fig. 2.11d). 96'" cos!) —sinD. 0 x" x" cosi) — y" sinD.
y’” = sinﬂ cosi). 0 y" = as" sin.(). + y" cosﬂ _ (2.25)
z”! 0 0 1 Z" z” The vector (x"’, y’”, z’") is the location of the satellite in the right ascension—
declination coordinate system at time t. This vector may be converted into the
radius, declination, and right ascension of the satellite by rs = 1/xm2 + ymz + szZ = r, Q=mﬂ(£0, nan) rs
(a). mac) After one has calculated the right ascension, declination, and radius of the
satellite, it is useful to calculate the latitude and longitude of the subsatellite point.
Assuming that the Earth is a sphere, the latitude (known as the geocentric latitude)
is simply equal to the declination. The longitude of the subsatellite point is the
difference between the right ascension of the satellite and the right ascension of
the prime meridian (0° longitude) which passes through Greenwich, England (Fig.
2.12). The right ascension of Greenwich can be calculated knowing its right
ascension at a given time and the rotation rate of the Earth.“ Since the rotation
rate changes very slightly, due to the actions of the wind and ocean currents, very
accurate knowledge of the right ascension of Greenwich requires observations.
Some satellite bulletins give the right ascension of Greenwich in addition to the
satellite orbital elements. 11 If nothing else is available, one can use the following: at 0000 UTC on 1 january 1990 the
right ascension of Greenwich was 10038641", and the rotation rate was 7.292115 922 X 10'5 radians
per second or 3603856507" per day. ' “ *da‘iLa‘a'h ._'. . . 2.5 Satellite Positioning, Trackin , and Nani ation VERNAL
EQUINOX SATELITE FIGURE 2.12. The relationship between Earth longitude and right ascension. The inverse problem of ﬁnding when a satellite passes over (or close to) a
particular point is solved iteratively by ( 1) estimating the time, (2) calculating the
position of the satellite, and (3) correcting the time estimate. Steps 2 and 3 are
repeated until a satisfactory solution is found. 2.5.1.2 The Spherical Geometry Method The spherical geometry method can be derived using spherical geometry (Mad
den and Parsons, 1973), but it is also a distillation of the vector rotation method.
Let I", the argument of latitude, be the angle, measured in the orbital plane, from
the ascending node to the satellite. Numerically, F56+w, can where 6 is the true anomaly and to is the argument of perigee. Working through
the mathematics of the vector rotation method results in12 rs = r, (2.28a)
(9, = 8, = sin—1(sinF sini), (2.28b) = :1 sinI‘ cosi _ _ d9, _ d_.Q. _
‘1’, tan (—COSI, ) + .00 0.600) ( dt dt (t to). (2.28c) Here r is the distance of the satellite calculated with Eq. 2.7, @S and ‘1', are its
latitude and longitude, respectively. 0,,(1‘0) is the right ascension of Greenwich at
the epoch time, and therefore, 0.0 — 9.41:0) is the longitude of ascending node at
the epoch time. The quantity (riﬂe/dt  dQ/dt) is the relative Earth rotation
rate, that is, the rotation rate of the Earth relative to the orbital plane. For a
sunsynchronous satellite, it must be exactly 211' radians per day. 12 Normally the arctangent term would be written tan—1(tanFcosi). The form in Eq. 2.28c is used because it allows the quadrant of the angle to be determined unambiguously. In Fortran, for example,
ATAN2(sinFcosi,cosF) will result in the correct angle. 34 Chapter 2 Orbits and Navigation For a circular orbit, or one which is so nearly circular that no signiﬁcant error
occurs from neglecting its elliptical nature, I‘(t) = To + (H + (t — to), (2.29)
and the spherical geometry method is particularly easy to apply. Polar orbiters
can often be treated with this approximation. When they are so treated, the orbital
parameters may come in a different form. The supplied parameters may be (1)
the longitude of ascending node, (2) the nodal period, (3) the radius (or semimajor
axis), (4) the inclination, (5) the time of ascending node, and (6) the nodal longitude
increment (ALON), which is the difference in longitude13 between successive
ascending nodes: dt dt T, (2.30) ALON = (“me am) " where T is the nodal period. The above "equations still apply, but one must
remember that To = 0, (E + day/tit) = 27r/T, and (dﬂe/dt — dQ/dt) = ALON/ T. 2.5.2 Tracking A list of time versus position of a celestial body is called an ephemeris (plural:
ephemerides). To track a satellite, one must be able to point one’s antenna at it.
The elevation angle, measured from the local horizontal, and the azimuth angle,
measured clockwise from the north, can be calculated from the ephemeris data
as follows. Suppose the subsatellite point is at latitude G), and longitude ‘PS, and that the
satellite is at radius r, from the center of the Earth. Suppose also that the antenna
is located at latitude (9,, longitude We, and radius re (the radius of the Earth).
The Cartesian coordinates of the satellite, then, are sung1n . rm ._'_..____,;. mamum . xs rs cos®s cos‘I’,
$35 = 3;, = r, cos®S sin‘IJ‘S , (2.31)
z, r, sin®s whereas the Cartesian coordinates of the antenna are xe re c059,, cos‘I‘"e
7: = ye = re c0569E sin‘I“e , (2.32)
ze re sin®e The difference vector (FD E F, — ﬂ) points from the antenna to the satellite (Fig.
2.13). Assuming a spherical Earth, the vector 7;, points to the local vertical (Fig. ‘3 That is, the next ascending node occurs ALON west of the current ascending node.
2.5 Satellite Positioning, Tracking, and Nani etion FIGURE 2.13. Satellite tracking geometry. 2.14). The cosine of the satellite’s zenith angle t (the complement of the elevation
angle) is given by (2.33) Finding the azimuth angle is a little more difﬁcult. First, we need to ﬁnd two
vectors 1n the tangent plane at the antenna. The ﬁrst points north: , xN —sin®e cos‘I’e
7N = yN = —sin®E sin‘P‘e , (2.34)
zN cosh), The second is the horizontal projection of FD. If we deﬁne unit vectors in the W
0’ FIGURE 2.14. Deﬁnition of zenith angle (1;) and azimuth angle (up). 36 Chapter 2 Orbits and Navigation directions of if; and ?D as e (2.35a) "7:
ill
TLI at
t. >
it“ (2.35b) the required horizontal vector is
7:11 = ?D _ (fe'ﬁﬂfe = 7"u _ lﬁolcosg 7": = WWI) _ COSE fe)’ (236) The azimuth angle is then given by cos d; = A (2.37) M.
erll?Hl One must be careful when taking the inverse cosine. If the satellite is west of the
antenna, gt; will be greater than 180°. It also must be noted that these equations
assume a spherical Earth. Fortunately, most receiving antennas are 1nsens1tive to the slight errors this assumption causes. 2.5.3 Navigation In addition to knowing where a satellite is in its orbit, it is necessary to know
the Earth coordinates (latitude, longitude) of the particular scene it is Viewmg.
The problem of calculating the Earth coordinates is known as ‘the navigation
problem; fundamentally, it is a complex geometry problem. it requires an accurate
knowledge of where the satellite is in its orbit, the orientation of the satellite (its
attitude), and the scanning geometry of the instrument involved. . In simpliﬁed form, we can proceed as follows. Suppose that at a particular
time a satellite is at position (x,,y5,z,) with respect to the center of the Earth
in the right ascension—declination coordinate system. ‘Suppose further, that the
telescope is pointing in a direction speciﬁed by declmatloanland. right ascensmn
QT. A unit vector in the direction that the telescope IS pomtlng is given by nor cosﬁl cosQT
yT : cosBT sinﬂq . (2.38)
ZT sin3T Figure 2.15 shews that the ray from which the telescope receives radiation (that
is, the line in space through the satellite and in the direction of the telescope) lS given by
st:s + s xT x
y = ys+syr . (239)
Z ZS+SZT where s is the distance from the satellite. 2.5 Satellite Positioning, Tracking, and Nani ation FIGURE 2.15. Navigation geometry. The location at which this ray strikes the spherical Earth is the solution of the equation (xS + s JCT); + (315 + s yT)2 + (zS + 5251)2 = r3. (2.40) This is a quadratic equation in s that has no real roots, if the ray does not intersect
the Earth; or two real roots, if it does. The smaller root is to be chosen; the larger
root represents the location from which the ray reemerges from the opposite side
of the Earth. When the ray is just tangent to the Earth, the two roots are equal.14 After a solution for s has been found, Eq. 2.39 gives the Cartesian coordinates
in the right ascension—declination coordinate system of the point on the Earth’s
surface being viewed. The latitude and longitude can then be found as in Sec
tion 2.5.1. Satellite images are usually the result of a scanning instrument. The data come
in the form of scan lines, each divided into elements or samples known as pixels
or scan spots. Because scanning is very carefully timed, each pixel has a unique
time associated with it. Therefore, calculating the latitude and longitude of a pixel
is accomplished using the equations of Sections 2.5.1 and 2.5.3 in the forward
direction; time yields satellite position and telescope pointing angles, which then
yield latitude and longitude. The opposite problem, ﬁnding the pixel which ob
served a particular point on Earth (latitude and longitude), must be solved in an
iterative manner because the exact time when the point was observed is unknown.
In brief, the time of observation is estimated, the actual point being observed at
that time is calculated, and a correction is made in the time which moves the point of observation closer to the desired point. This procedure is iterated until
satisfactory convergence is achieved. The scheme outlined here for ﬁnding latitude and longitude is simple and very
general. It is applicable to a wide variety of satellite orbits and instruments. For H For a geostationary satellite, this occurs about 81" from the satellite subpoint and explains why
geestationary satellites never observe the poles. 38 Chapter 2 Orbits and Navigation each instrument, the difﬁcult part is to determine the telescope pointing angles ST and (IT. 2.5.3.1 Geostationary Geometry Until recently, all geostationary meteorological satellites were spin stabilized.
They spin on an axis which is maintained nearly parallel to the Earth’s spin axis.
The rotation of the satellite changes the right ascension of the telescope and
provides scanning across the Earth. Scanning in the north—south direction is
accomplished by a tilting mirror (see Chapter 4), which changes the telescope
declination. Thus BT and 01 are natural coordinates for spinstabilized geostation
ary satellites. Unfortunately, the satellite’s spin axis is not exactly parallel to the Earth’s spin
axis. Furthermore, although the radiometer’s telescope is rigidly oriented with
respect to the principal axis of the satellite, the spin axis deviates slightly from
the principal axis, which causes deviations similar to pitch, roll, and yaw in a
low Earth orbiter (see next section). Corrections for these effects and for the non
spherical Earth can be made. The interested reader is referred to Hambrick and
Phﬂhps(1980) The parameters that describe the satellite orbit and attitude must be aCCurately
known to perform accurate navigation. These parameters can be determined by
the use of landmarks. Normally the orbit and attitude parameters are accurate,
as is navigation performed with them. However, for up to 18 h after the thrusters
are ﬁred in an orbit or attitudecorrecting maneuver, navigation parameters are
poorly known, and pixels can be signiﬁcantly misplaced. These errors can be
partially corrected by displaying the data as an image and shifting the image up
or down and right or left until a landmark is properly positioned. Rotation of
the image is sometimes necessary to achieve good navigation, especially if a large
(continentsize) area is being studied. It is interesting to note that the GOES satellites can detect a few stars at the
edges of the image frame. These stars can be used to very accurately determine
the attitude of the satellite. Then landmarks can be used to determine the orbital
elements (Hambrick and Phillips, 1980). 2.5.3.2 low Earth Orbit Geometry Low Earth orbit satellite instruments have many scanning patterns. Navigation
of these data can be achieved using different approachs. We outline an approach,
based on the discussion above, which is general enough for use with many scanning
patterns. The basis of the technique is that if we can determine where a scan spot
is in relation to the satellite, then we can use nearl}r the same rotation matrices
with which we position the satellite to position the scan spot. First we must deﬁne
the angles and a coordinate system used to specify satellite attitude. The instruments on many low Earth orbit satellites are mounted on the under
side of the satellite and scan perpendicular to the velocity vector through the
subsatellite point (see Chapter 4). A convenient coordinate system (Fig. 2.16) is
one in which the z axis points from the satellite toward the center of the Earth, 2.5 Satellite Positioning, Tracking, and Naw‘ ation FIGURE 2.16. Coordinate system used for satellite attitude. the x axis points in the direction of satellite motion, and the y axis is chosen to
complete a righthanded coordinate system. Three angles specify the orientation
of the satellite in this coordinate system. Rotation about the y axis is called pitch,
rotation about the x axis is called roll, and rotation about the z axis is called yaw. A combination of these angles can be used to specify nearly any scan geometry.
Instruments that scan through nadir perpendicular to the satellite motion vector
are described by changing the roll angle. Instruments that scan in a cone can be
described by a constant pitch plus a variable yaw. Instruments that scan through
the subpoint but at an oblique angle with respect to the satellite motion vector
can be described with a roll plus a constant yaw. To calculate the position of a scan spot with respect to the satellite, we proceed
as follows. First, position the satellite at radius rs, declination zero, and right
ascension zero, and let its velocity vector point east. Assume that the telescope
is pointing straight down,” or 8T = 0 and QT = 71'. That is, the telescope pointing
vector is xT = —1, yT = 0, zT = 0. Next, rotate the telescope vector through the
pitch, roll, and yaw angles that describe the position of the telescope at time t. 15 If the satellite is not pointing straight down, its deviation is described by pitch, roll, and yaw
bias errors, which are usually small. For example, horizon sensors and a sun sensor on the current
NOAA satellites maintain pitch, roll and yaw bias errors to within i0.2° of zero. If the bias errors
are known, and if the desired precision of the calculation requires it, the initial telescope pointing
vector can be corrected for bias errors at this point. 40 Chapter 2 Orbits and Navigation __—__,__.__.—_———#—— or for the scan position desired. At the assumed position and orientation of the
satellite, in the right ascension—declination coordinate system, the pitch rotation matrix is
COSQP Sinai]: 0
—sinaP cosap 0 , 0 0 1 where or is the pitch angle; the roll rotation matrix is cosaR 0 —sinaR
0 1 0 , sinaR 0 cosaR
where (IR is the roll angle; and the yaw rotation matrix is 1 0 0
0 cosozY —sinay , 0 sinay cosaY where city is the yaw angle.16
After the telescope pointing vector has been determined, the distance 5 from the satellite to the scan spot is calculated using Eq. 2.40, and the position of the
spot relative to the satellite is calculated using Eq. 2.39. Finally, the scan spot is moved along with the satellite to its actual position
by (1) rotating about the z axis through the argument of latitude, (2) rotating
about the x axis through the inclination angle, and (3) rotating about the z axis
through the right ascension of the ascending node minus the right ascension of Greenwich.
An advantage to this method is that if the orbit is sufﬁciently circular, the vectors to the scan spots can be calculated in advance and simply rotated into position at successive times. Note that limb scanners, which scan the atmosphere above the Earth’s horizon,
can be treated with this procedure except that Eq. 2.40 is not applicable because
the ray does not strike the Earth. Instead, the distance to the tangent point, that
is, the point where the ray most closely approaches the Earth, can be used for 5.
If a is the angle between the initial telescope pointing vector (straight down) and the ﬁnal vector, then s : rscosa. (2.41) Finally, we would like to outline a simple calculation that is frequently useful
in satellite meteorology: how to ﬁnd the distance of a scan spot from the subsatellite 1‘ If the satellite is thought of as an airplane, a positive pitch angle is deﬁned here as the nose
pointing up, a positive roll as the right wing pointing up, and a positive yaw as a counterclockwise
rotation of the plane as viewed from above. Irma... 2.6 Space—Time Sampling FIGURE 2.17. Determining the distance of a scan spot from the subsatellite point. point. Figure 2.17 shows the geometry of this calculation. If a is the scan angle
then the law of‘sines gives the angle B as , sinB = since. (2.42) C The angle measured from the center of the Earth is ¢=wBa, (2.43) and the distance from the subsatellite point to the scan spot is (fit .
The swath wraith is the width of the entire scan of the satellite instrument. If the instrument scans equally on each side of the ground track, then the swath
Width 1s 2951's, where a is the maximum scan angle.17 2.6 SPACE—TIME SAMPLING To select an orbit for a satellite or a scan pattern for a particular instrument
several questions must be answered: What areas will the orbit and scan patterii
allow the instrument to observe? How often will an area be observed? At what
local times will the observations be made? At what viewing zenith and azimuth
angles will the observations be made? These questions are all aspects of what is
called space—time sampling. Geostationary satellites are designed to be nearly stationary over a point on
the equator. They therefore view a ﬁxed area (about 42% of the globe). Any 17 With this deﬁnition, the swath width is the distance between the centers of the extreme scan
Spots (see Fig. 4.10). Sometimes, the halfwidth of the radiometer ﬁeld of view is added to each end
of the swath, so that the swath width describes all that the radiometer senses. 42 Chapter 2 Orbits and Navigation lBO 90W 0 90E [80
l:l 0 W  m 2 or more FIGURE 2.18. One day’s coverage by a hypothetical instrument on a sunsynehronous satellite. The
coverage at the equator is 50%. The orbit is circular with a semimalor axis 'of 7228 'km and an
inclination angle of 98.8”. Note that slightly different orbit parameters can result in quite different pat— terns. point in this area can be observed as frequently as their instruments will allow;
that is, it can be observed at any local time. lIowever, smce each pomt has 3
ﬁxed geometric relationship to the satellite, it 15 Viewed at only one zenith an
one azimuth angle. I I ,
For a satellite in low Earth orbit, these questions depend on the satellites
orbit and the scanning geometry of its instruments. Most meteorological satellite
instruments are designed such that the area viewed on one orbit touches or'overlaps
the area viewed on previous and successive orbits. If the satellite’s inclination angle
is large enough, the instrument views every pomt on Earth twice per day, at leailt.
The poles are observed on every orbit. Usually eachpomt is viewed at a w1 e
range of zenith and azimuth angles. Many meteorological satellites are in sunsyn
chronous orbits, which have constant equator crossing times. These satellites V1631!
each point (except near the poles) only in a small range of local times (t T) centered on the two equator crossing times. I I I ft
For instruments whose scans on successrve orbits do not overlap, it is 0 en best to plot the coverage for a day and to determine visually which areas are
observed and which are not. Figure 2.18, for example, shows the one—dgy coverage
of a hypothetical instrument in a sunsynchronous orb1t which has 50 A: coverage
at the equator.18 Some areas are not observed, some are observed once, and some
are observed twice or more. This pattern will be the same on succeeding days,
except that it will drift in longitude. The drift rate can be calculated as follows.
Divide the length of day by the nodal period and round to the nearest integer, ‘3 That is, the swath width divided by the sine of the inclination angle is 50% of ALON. 2.6 Space—Time Sam ling 43 N. The westward longitude increment in N complete orbits is NALON. If we express the nodal longitude increment ALON (Eq. 2.30) in degrees, then the
change in the pattern per day is Drift = —NALON + 360°. (2.44} Sunsynchronous satellites have a relative Earth rotation rate of exactly 360”
per day. They have the interesting property that if they make an integral number
of orbits in an integral number of days, then they must arrive exactly at the
longitude where they started and repeat the ground track. If such a satellite makes
N + k/m orbits per day, where it and m are integers, then the orbit track repeats
every m days after making mN + k orbits. Furthermore, if k and m have no
common factors, all m of the ground tracks, spaced ALON/m, will he traversed.
Earth remote sensing satellites have utilized this property. Landsats 1, 2, and 3,
for example, were designed to make 1317/13 orbits per day. Thus they had a
nodal period of 103.27 min, which means that ALON was 25 .82°. The distance
between ascending nodes at the equator was about 2874 km. The Multispectral
Scanner (MSS) scanned across the satellite track with a ground swath width of
only 185 km; only a small fraction of the equator was observed on any one day.
The daily longitude drift was —1.43° (*ALON/18), or about 160 km west of an
ascending node on the previous day. Since the swath width was greater than the
westward movement, the swaths on successive days overlapped. In 18 days the
satellites observed every point on the equator and began the cycle anew. The French SPOT satellites (see SPOT User’s Handbook) utilize the same type
of repeat cycle, except that they orbit 145/26 times per day. The swaths on
successive days do not overlap, but in 26 days the entire Earth is imaged. Landsats
4 and 5 have similar orbits. Note that this repeat cycle is very sensitive to the semimajor axis. If the orbital
altitude of Landsats 1, 2, and 3 had been decreased by only 19 km, they would
have made exactly 14 orbits in one day. There would have been no westward
progression of the swaths. Some parts of the Earth would be observed every day;
the rest would never be observed. For studies of diurnal variation, a point must be observed at local times through
out the day. Since sunsynchronous satellites view a point at nearly the same two
local times every day, they are not useful for diurnal variation studies. A satellite
designed speciﬁcally to measure diurnal variation is the Earth Radiation Budget
Satellite (ERBS; see Chapter 10). It is in a 57°~inclination orbit at an altitude of
600 km. The right ascension of ascending node moves west by 3355" per day,
while the mean sun moves east 0.9860 per day (360° in one year). Thus the angle
between the sun and the ascending node changes 494° per day. Because the
satellite makes observations both as it ascends and as it descends, all local times
will be sampled when the sun—Earth—ascending node angle has changed by 180°.
The ERBS, then, samples all local times in about 36 days. Many space—time sampling strategies are possible. The reader is encouraged
to use the equations presented above to investigate some of the possibilities. 2.7 LAUNCH VEHICLES AND PROFILES A discussion of satellite orbits would not be complete without mention of the
launch vehicles used and the strategies available for achieving orbit. 2.7.1 Launch Vehicles U.S. meteorological satellites have been la
recently two have predominated (Fig. 2.19). The Delta 3914 was used to place the 345kg GOES 4—7 satellites into 35 ,790km geostationary orbits. The Atlas
EI'F is used to boost the 1421kg NOAA satellites into 850km sunsynchronous
orbits. Other meteorological satellites have been placed into orbit by the French
Ariane rocket, the Soviet F2 rocket, and the Japanese N—Z rocket, and the U.S.
Space Shuttle (see Appendix A). As satellite launches become more commercial
and more competitive, many different rockets are likely to be used for meteorologi cal satellites. It is interesting that a far larger rocket is used to launch the low Earth orbit
satellites than the geostationary satellites. In part this is because the energy
required to achieve orbit is proportional to the mass of the satellite. However,
Fig. 2.20 shows that the Earth is at the bottom of a deep gravitational potential
well. The ﬁrst step into space is the energy—consuming step. It takes approximately
35 M] kg‘1 to lift a satellite into an 850 km orbit; it takes only about 65% (23 M] kg'l) additional energy to increase that orbit by a factor of 42 to geostation ary altitude. unched by a variety of vehicles, but Delta 3914 Alta: I
Centaur FIGURE 2.19. Rockets used to launch recent U.S. meteorological satellites. [After Chen
by permission of Academic Press] (1985). Reprinted 2.7 Launch Vehicles and Proﬁles 0‘ A
O O ENERGY PER UNIT MASS IMJ kg")
N
O 850 km [NW SATELLITES) I
I
l
I
I
i
I
ml
3:
of
g:
x.’
U 2
"I
8 L,,I
(9
I
 5 O
IOO IDOO IQOOO O0.000
ORBlTAL ALTITUDE (km) FIGURE 2.20. The energy per unit mass re ‘ ' ' '
_ _ qmred to place a satellite in orbit as a function of orb't l
altitude. Note that the consulerable energy required to lift the rocket itself is not included in this ﬁglit: 2.7.2 Launch Proliles Three basic strategies are available for orbit insertion. In poweraIIthewa
ascentsthe rocket burns steadily until the desired orbit is achieved. This launcfi
proﬁle is more costly, but less risky, than the others because rockets do not have
to be restarted in space. This proﬁle is used for manned space ﬂights. I The second type of launch proﬁle is called ballistic ascent because of its similar
ity With artillery. A large ﬁrststage rocket is used in the early part of the ﬂight
:10 prolpel bthe payload to high velocity. It then coasts to the location of the
dig; 2:13p: where a secondstage rocket is ﬁred to adjust the trajectory to the The third type of launch proﬁle is called elliptical ascent. Orbit insertion is
achieved in three steps. First the payload is placed in a low Earth orbit by either
of the above means. This ﬁrst orbit is referred to as a parking orbit. In the next
phase, a'rocket is ﬁred to move the payload into an elliptical transfer orbit whose
perigee is the parking orbit and whose apogee is the desired orbit. When the
[:Jéofad rﬁacheslapogﬁeha roccllcet (sometimes called an “apogee kick motor”) iesteorittote esire Li ' ' ' '
for geostationary satellites. ( sually circular) shape. Elliptical ascent is used Bibliography Barnes, J. C., and M. D. Smallwood 1982 TIR  ' ‘ ' B NOAA, waShington, DC. ( ). OS N Series Direct Readout Serwces Users Gwde. Cﬁuwg, 13., and G. M. Clemente (1961). Methods of Celestial Mechanics. Academic Press, New York.
II, . . (1985). Space Remote Sensing Systems: An Introduction. Academic Press, Orlando. 46 Chapter 2 Orbits and Navigation Dubyago, A. D. (1961). The Determination of Orbits. The Macmillan Company, New York. Escobal, P. R. (1965). Methods of Orbit Determination. John Wiley and Sons, New York. Goldstein, H. (1950). Classical Mechanics. Addison—Wesley, Reading, MA. I _ Hambrick, L. N., and Phillips, D. R. (1980). Earth Locating Image Data ofSpinStabrlized Geosyncbm
nous Satellites. NOAA Tech. Memo. NESS 111, Washington, DC. _ . . Kidder, S. Q., and T. H. Vonder Haar (1990). On the use of satellites in Molniya orbits for meteorologi
cal observation of middle and high latitudes. ]. Atmos. Ocean. Tech, 7, 17—522. ' Madden, R., and Parsons, C. (1973). A technique for real—time, quantitative display of APT Scanning Radiometer data. I. Appl. Meteor., 12, 381—385. ‘ I
Short, N. M., P. D. Lowman,]r., S. C. Freden, and W. A. Finch,]r. (1976). Massmn to Earth: Landsat Views the World. NASA, Washington, DC. . . .
Smith, E. A. (1980). Orbital Mechanics and Analytic Modeling of Meteorological Satellite Orbits. Colo. State Univ. Atmos. Sci. Pap. 321, Fort Collins, CO.
SPOT User’s Handbook. SPOT Image Corp., Reston, VA. Radiative
Transfer A LL OF THE information received by a satellite about the Earth and its
atmosphere comes in the form of electromagnetic radiation. It is
necessary, therefore, to understand the mechanisms by which this radiation is
generated and how it interacts with the atmosphere. Several texts listed in the
Bibliography explore atmospheric radiation in detail. Here we concentrate on
those aspects that are essential for satellite meteorology. 3.] BASIC QUANTITIES Electromagnetic radiation consists of alternating electric and magnetic ﬁelds
(Fig. 3.1). The electric ﬁeld vector is perpendicular to the magnetic ﬁeld vector,
and the direction of propagation is perpendicular to both. Radiation is often
speciﬁed by its wavelength, which is the distance between crests of the electric
or magnetic ﬁeld. Figure 3.2. shows the electromagnetic spectrum. A broad range
of wavelengths from the ultraviolet to the microwave region is useful in satel
lite meteorology. An alternate way to describe radiation is to give its frequency, which is the
rate at which the electric or magnetic ﬁeld oscillates when observed at a point. 47 ...
View
Full Document
 Spring '11
 Haijan
 Celestial mechanics, Astrodynamics, Orbital elements, Kepler orbit

Click to edit the document details