Analysis of earth rotation solution from Starlette

In this study, a consider covariance analysis of error sources limiting the accuracy of the Earth rotation parameters (ERP), which include the polar motion coordinates x and y and the length of the day (LOD), determined from satellite laser ranging (SLR) data to near‐Earth geodetic satellites, has been performed. The solution for the ERP has been obtained from the analysis of the SLR data to Starlette collected primarily during the 14‐month MERIT Campaign that began September 1, 1983. A preliminary Starlette ERP solution based on a Starlette‐tailored Earth gravity model, PGS 1331, was compared with the solutions derived from SLR to LAGEOS, and the weighted rms about the mean of the difference between these solutions was 13 milliarcseconds (mas), 11 mas, and 0.6 milliseconds (ms) for Δx, Δy, and ΔLOD, respectively. A covariance analysis indicated that the accuracy of the Starlette ERP solution was limited primarily by errors in the Earth's gravity field model. In particular, this analysis showed that the first‐order geopotential coefficients produced a systematic perturbation in the solutions for x and y with a maximum value up to 29 mas and a dominant period of 74 days. The LOD determination by Starlette is limited primarily by errors in the Earth's zonal harmonics and long‐period tides. With the improvement of the Starlette force model by simultaneous estimation of selected geopotential coefficients and ocean tide parameters, the weighted rms of the Starlette solution with respect to LAGEOS was reduced to 9 mas and 6 mas for Δx and Δy, respectively. However, this technique was not able to improve the accuracy of the LOD solution. Using an improved gravity field recently developed at the University of Texas Center for Space Research, the weighted rms differences between Starlette and LAGEOS ERP solutions have been further reduced to 4.4 mas and 3.6 mas for Δx and Δy, respectively. These results show that a significant improvement of the ERP solution from Starlette can be achieved through further refinements in the Earth's gravity field and tide model and that a precision of <5 mas in x and y coordinates of the pole is feasible from Starlette.