The Earth's orientation
is defined as the rotation from the Earth crust (the terrestrial system)
to a geocentric set of axes tied to the quasars (geocentric celestial
system, to be distinguished from the reference celestial system having
for origin the barycenter of the solar system). This rotation is split
into three components :
 The precessionnutation of the figure axis in space. It is specified through the
conventional
precessionnutation model (presently that recommended by
the IAU 2000 General Assembly) and corrections determined by the
VLBI observations : the celestial pole offsets (dψ,dε)
or (dX, dY) according
to the parameterisation which has been adopted. The precessionnutation
model and its corrections (up to 1 mas / IAU 2000 model) define a fictitious
axis, the Celestial Intermediate Pole or CIP.
N.B.:
The CIP was called before 2003 celestial ephemeris pole or CEP. This
change of terminology corresponds to the nonambiguous frequency splitting
between diurnal/subdiurnal polar motion and nutation for this axis
: by definition the precessionnutation of the CIP does not have
spectral components of which the period is smaller than 2 days.
 The diurnal rotation
around the celestial intermediate pole (this one remains close to the
instantaneous rotation axis in a tolerance of 20 milliarcseconds) :
Ω_{N} UT1. It can be decomposed into
an angle uniformly varying with TAI, Ω_{N}
TAI, (with the nominal Earth rotation rate Ω_{N} ,
see useful constants)
and corrections showing variations of the Earth angular velocity. These
corrections are given by the difference (UT1UTC) or (UT1TAI).
 The polar motion
of the celestial intermediate pole with respect to the terrestrial crust.
The CIP has for terrestrial coordinate (x,y,1). As a
first approximation, for periods greater than 10 days (x,y,1) are also
the coordinates of the instantaneous rotation axis. It is not possible
to confuse both axes at diurnal time scale.
The five Earth orientation parameters , derived from the observations, bring
corrections to the uniform diurnal rotation and modelled precessionnutation :
 The celestial pole offsets (dψ,dε)
or (dX,dY)
with a maximum temporal resolution of 2 days.
 UT1UTC
or UT1TAI (from which we can derive the variations of the
length of the mean solar day, ΔLOD,
with respect to its nominal value of 86400 s TAI)
 The polar motion (x,y)
The Earth orientation
is then obtained by inserting those parameters in the coordinate transformation
between the Celestial Reference Frame and the Terrestrial Reference Frame
:
 for the classical way :
[CRF]=PN(dψ,dε) R_{3}(UT1TAI) W(x,y) [TRF]
 for the "non rotating origin" formalism :
[CRF]=PN(dX,dY,s)R_{3}(UT1TAI)W(s',x,y) [TRF] (s
are s' are practically independent from the EOP)
The description
of this matrix can be found in the IERS
conventions 2003, chapter 5. For more insight on the meaning of the
EOP, read the following page.
We propose two kind of EOP series :
 the individual series (most often updated routinely)
as determined from the observations of a given astrogeodetic technique
VLBI, LLR, SLR, GPS, DORIS) by various organisations all over the world.
 the IERS combined series:
 C04: since 1962, daily
 C01: since 1840, 0.05 year sampling
 C02: since 1830 for LOD & UT1  TAI after 1958 and UT  TE before 1958, 100 day sampling
