Improving Predictions of the Earth

1
Improving Predictions of the Earths Rotation
UsingOceanic Angular Momentum
Richard Gross and Maggi Glasscoe Jet Propulsion
LaboratoryCalifornia Institute of
TechnologyPasadena, CA 911098099, USA ECCO2
Annual Meeting November 910, 2009Pasadena, CA
2
Introduction
Accurate navigation of interplanetary
spacecraft requires accurate knowledge of
Earths orientation Must know Earths
orientation in space to know spacecrafts
position in space from Earth-based tracking
measurements Uncertainty in Earths
orientation can be a major, if not the dominant,
source of error in spacecraft navigation and
tracking (Estefan and Folkner, 1995) Error in
UT1 of 0.1 ms (4.6 cm) produces an error of 7
nrad in spacecraft right ascension,
corresponding to a position error at Mars of 1.6
km Earths orientation in space given by 5
parameters 2 precession-nutation parameters
(Dy, De) Specifies location of spin axis in
celestial reference frame 2 polar motion
parameters (PMX, PMY) Specifies location of
spin axis in terrestrial reference frame 1
spin parameter (UT1) Specifies angle
through which Earth has rotated about spin axis
3
Improved Predictions using OAM?
Earths rotation varies rapidly and
unpredictably Universal Time (UT1) variations
particularly difficult to predict Rapid UT1
variations caused mainly by changes in angular
momentum of winds Predicting UT1 is as
challenging as predicting the weather Atmospheri
c angular momentum (AAM) Exchanged with
angular momentum of solid Earth Axial
component causes length-of-day (UT1-rate)
variations Equatorial components contribute
to causing the Earth to wobble as it
rotates Computed by operational numerical
weather forecast models NCEP, ECMWF, NOGAPS,
AAM forecasts used to improve accuracy of
UT1 predictions Oceanic angular momentum
(OAM) Can including pressure forcing improve
prediction accuracy? Can OAM forecasts improve
prediction accuracy?
4
(Gross et al., 2004)
5
(No Transcript)
6
Approach
Evaluate potential impact of OAM forecasts on
accuracy of UT1 predictions Add OAM forecasts
to AAM forecasts when predicting UT1 But OAM
forecasts are not yet available from ocean
modeling groups Ocean weather forecast
systems currently being developed Use OAM
analyses as proxy OAM forecasts Forecasts are
less accurate than analyses Using analyses
as proxy forecasts provides upper limit to
improvement in accuracy of UT1 predictions that
can be expected to be obtained when OAM
forecasts are used to predict UT1 variations
7
Data Sets
Atmospheric angular momentum US National
Centers for Environmental Prediction (NCEP)
Angular momentum of winds and surface
pressure variations Analyses and forecasts to
7.5 days into future at 6-hour intervals Only
5-day forecasts of wind angular momentum used
here European Centre for Medium-Range Weather
Forecasts Angular momentum of winds and
surface pressure variations Analyses and
forecasts to 10 days into future at 12-hour
intervals Only 5-day forecasts of wind
angular momentum used here Oceanic angular
momentum ECCO/JPL data assimilating ocean
model (kf066b) Global model of the general
circulation of the oceans Assimilates
altimetric sea surface height measurements Ang
ular momentum of ocean current and bottom
pressure variations summed UT1
predictions Generated using JPLs Kalman Earth
orientation filter
8
Kalman Earth Orientation Filter
A Kalman filter is used operationally at JPL
to Combine available Earth rotation
measurements Including NCEP AAM forecasts
(5-day winds only) Predict future changes in
the Earths rotation Out to 78 days in
advance Input data are corrected prior to
combination Data are placed within same
terrestrial reference frame ITRF2005 Stated
uncertainties are adjusted to reflect accuracy,
not precision, of measurement Adjusted
uncertainties are used as weights by Kalman
filter Used here to generate UT1
predictions No AAM AAM only AAM
OAM Predictions generated 73 times during
19 March 2004 to 22 July 2004
9
UT1 Prediction Error, cm
AAM Forecast Series Prediction Interval,
days 0 1 2 3 4 5 6 7 No AAM
forecasts 4.1 4.7 5.8 8.5 12.9 18.7 25.7 33.7 NC
EP 5-day wind forecasts 4.2 4.7 5.5 6.9 9.0 11.7 1
5.0 19.2 ECMWF 5-day wind forecasts 4.1 4.7 5.6 7
.2 9.6 12.4 15.9 20.1 NCEP 5-day wind forecasts
OAM 4.2 4.7 5.5 6.8 8.7 11.1 14.1 17.9 ECMWF
5-day wind forecasts OAM 4.1 4.7 5.6 7.2 9.5 12.
1 15.4 19.4 Tabulated UT1 prediction error is
the rms with respect to an accurate UT1
reference series of the 73 predictions given at
the stated lead time. Prediction day 0 is the
epoch of the last length-of-day measurement. The
epoch of the last UT1 measurement is typically a
few days earlier. A change in rotation
equivalent to a 1 millisecond change in time
corresponds to a 46.3 centimeter displacement of
the Earths surface at the equator.
10
Summary
Potential impact of OAM forecasts on accuracy
of UT1 predictions has been evaluated Add OAM
analyses to AAM forecasts when predicting UT1
Adding OAM to AAM forecasts improves accuracy
of UT1 predictions only slightly 7-day UT1
prediction accuracy improves from 19.2 to 17.9 cm
when OAM is added to NCEP 5-day wind AAM
forecasts UT1 LOD variations predominantly
caused by wind AAM Oceans are minor source of
UT1 LOD variations But what about polar
motion (wobble)? Adding OAM to AAM forecasts
should greatly improve polar motion
predictions Oceans are known to be a major
source of polar motion excitation