The Impact on EOP Predictions of AAM Forecasts From the ECMWF and NCEP

In support of the operational tracking and navigation of interplanetary spacecraft, a Kalman filter has been used at the Jet Propulsion Laboratory (JPL) for nearly two decades to combine independent measurements of the Earth's orientation and to predict its future evolution. Changes in the Earth's orientation can be described as a randomly excited stochastic process; consequently, between measurements, the uncertainty in our knowledge of the Earth's orientation grows and rapidly becomes much larger than the uncertainty in the measurements. Thus, measurements of the Earth's orientation must be taken frequently and processed rapidly in order to meet the demanding accuracy requirements of the spacecraft navigation teams. Short-term predictions of the Earth's changing orientation have been shown to be improved by using the dynamical model-based analyses and forecasts of the axial component of atmospheric angular momentum (AAM) as proxy length-of-day measurements and forecasts. Such AAM analyses and forecasts are produced as part of the numerical weather prediction activities at governmental agencies like the National Centers for Environmental Prediction (NCEP) and the European Centre for Medium-Range Weather Forecasts (ECMWF). For example, the accuracy of JPL's predictions of the Earth's rate of rotation can be improved by nearly a factor of 2 when AAM analysis and forecast data from NCEP are used. Here we compare the results we get using NCEP forecasts with those we get using ECMWF forecasts. We find that both NCEP and ECMWF 5-day wind AAM forecasts agree extremely well with LOD during 19 March 2004 to 22 July 2004, with the ECMWF having a slightly higher correlation with the LOD observations but with the NCEP explaining a bit more of the LOD variance. We find that the 7-day UT1 prediction accuracy is slightly better with the NCEP forecasts. Finally we find that adding oceanic angular momentum from an ocean general circulation model (ECCO/JPL kf066b) to the AAM forecasts improves the accuracy of the UT1 prediction only slightly.