[Paper Review] Interaction between celestial and terrestrial reference frames and some considerations for the next VLBI-based ICRF

Zinovy Malkin, H. Schuh+2 more|arXiv (Cornell University)|Jun 10, 2014

GNSS positioning and interference3 citations

TL;DR

This paper proposes a systematic approach to improving the next-generation VLBI-based ICRF by addressing systematic errors through enhanced core source distribution, improved station modeling, and standardized ties to the ITRF. It advocates for a shift from defining to 'core' sources, expanded southern hemisphere coverage, and advanced non-linear station motion modeling to achieve near-uniform sky distribution and sub-milliarcsecond accuracy in the ICRF3 and ICRF4 realizations.

ABSTRACT

In this paper we outline several problems related to the realization of the international celestial and terrestrial reference frames ICRF and ITRF at the millimeter level of accuracy, with emphasis on ICRF issues. The main topics considered are: analysis of the current status of the ICRF, mutual impact of ICRF and ITRF, and some considerations for future ICRF realizations.

Motivation & Objective

• Address persistent systematic errors in the VLBI-based ICRF, particularly those arising from uneven source distribution and ITRF interaction.
• Improve the consistency and accuracy of the celestial and terrestrial reference frames (ICRF and ITRF) at the millimeter and microarcsecond level.
• Develop a standardized, uniform method for modeling non-linear station motions to reduce errors in ICRF solutions.
• Enhance the sky distribution of core sources—especially in the southern hemisphere—to reduce positional errors and improve orientation stability.
• Establish a forward-looking strategy for future ICRF realizations (ICRF3 and ICRF4) based on improved observation planning and multi-baseline sessions.

Proposed method

• Use a global VLBI solution with a single, unified solution for all ICRF2-like source positions, tied to ICRF only via orientation (NNR constraint), to avoid legacy biases.
• Introduce a new concept of 'core' sources instead of 'defining' sources, with a preliminary list established in parallel with each ICRF realization.
• Implement advanced station position modeling using the SOPAC model (linear drift + jumps + seasonal + exponential relaxation) to better represent non-linear motions.
• Conduct dedicated CRF-dedicated observing sessions with global coverage to improve source distribution and data quality.
• Standardize the set of core stations used to tie VLBI solutions to the ITRF, ensuring consistency across analysis centers.
• Leverage regular VLBI2010 operations to systematically increase data volume (50% more than ICRF2) and improve source density to >4100 sources with >410 core sources.

Experimental results

Research questions

• RQ1 How can systematic errors in the current ICRF be reduced through improved source distribution and solution methodology?
• RQ2 To what extent does the choice of reference stations and ITRF model affect the orientation and stability of the ICRF?
• RQ3 What impact do non-linear station motions (e.g., post-seismic relaxation, seasonal signals) have on ICRF solution accuracy, and how can they be better modeled?
• RQ4 Can a shift from 'defining' to 'core' sources in future ICRF realizations lead to more uniform error distribution and improved long-term stability?
• RQ5 What observational and processing strategies are required to achieve near-uniform sky coverage and sub-milliarcsecond accuracy in ICRF3 and ICRF4?

Key findings

• ICRF2 reduced the error floor from 250 μas to 40 μas and improved axis stability to ~10 μas, but systematic errors of up to 200–250 μas remain due to uneven source distribution and ITRF interaction.
• The distribution of ICRF source positions and errors remains highly uneven, with strong bias toward northern sources due to the concentration of northern VLBI stations.
• Non-linear station motions—especially post-seismic relaxation and seasonal height variations—can cause modeling errors exceeding 1 cm when using only linear ITRF models, significantly degrading EOP and ICRF orientation.
• The SOPAC model (linear drift + jumps + seasonal + exponential relaxation) provides a significantly better fit to actual station motion than the standard ITRF model, though dissemination and extrapolation remain challenges.
• ICRF3 is expected to increase the total number of sources to over 4,100 and core sources to over 410, with a focus on improving southern hemisphere coverage and reducing positional errors.
• ICRF4 aims for near-uniform distribution of both sources and position errors across the sky, achievable through regular VLBI2010 operations and advanced observation planning, enabling use in Gaia catalog orientation.