Previous Abstract Return to Session C4 Next Abstract

Session C4: Military PNT User Equipment: Other

Curve-Fitting Enhancement of Angular Momentum Forecasts for Earth Orientation Parameter Estimation
Nick Stamatakos, Jessica Page, Dennis McCarthy, Mark Psiaki, David Salstein, US Naval Observatory
Location: Room 1-3
Date/Time: Monday, Jun. 3, 5:05 p.m.

Robust curve-fitting tools are developed to enhance atmospheric and oceanic angular momentum forecasts in support of optimal estimation of Earth Orientation Parameters (EOP), and the resulting techniques are used in a retrospective analysis of past leap second insertions. These techniques are being developed and evaluated in support of ongoing improvements to data assimilation algorithms that provide estimates of past, present, and forecast EOP. Non-dimensional effective atmospheric and oceanic angular momentum function analysis values (i.e., past values) and prediction values along all three Earth-fixed Cartesian axes are provided by atmospheric and oceanic fluid/thermal modeling groups. Their raw outputs may have biases, ramping offsets, and other types of errors. The high-frequency signal content from the best modelers, however, tends to match reality to a significant degree, as evidenced by comparisons with “truth” values as derived from smoothed estimates of past EOP and their time rates of change. (Models of systematic errors in the angular momentum data were developed, and a range of multiple regression models were considered to determine the optimal weighting and length of the time series of the input data.) These retrospective comparisons between modelers’ raw angular momentum values and “truth” values are used to guide the development of pre-processing algorithms for the modeler’s raw inputs. Previous investigations have shown the potential of improving the accuracy of estimates of the direction of the rotational pole and of the angular velocity of the Earth by using improved pre-processing of atmospheric and/or ocean angular momentum data along with improved optimal estimation algorithms that assimilate these data and fuse them with other data to form EOP estimates. Pre-processed atmospheric and oceanic angular momentum data are useful to make short-term predictions of the EOPs by considering the conservation of the angular momentum of the Earth/atmosphere/ocean system. Investigations have shown that predictions of UT1 – UTC estimates can be improved by 45% for 1-day predictions and 30% for 7-day predictions. This work is a continuation of previous efforts to investigate procedures to handle outliers in EOP input data using improved robust curve-fitting tools and improved optimal estimation tools.
In addition, recent models of potential leap second changes are investigated using various standard correlation and ARIMA techniques. In this study, we investigate whether robust curve-fitting tools can improve the ability to forecast previous leap second changes.

Previous Abstract Return to Session C4 Next Abstract