The K-band Celestial Reference Frame (K-CRF) is created from a Global least-squares solution of all K-band 24-hour astrometric VLBI observing sessions taken since 2002. In the Global solution, an astrometric/geodetic analysis package is used to make a least-squares solution between the observed baseline delays and modelled theoretical delays. The observed baseline delays are the differences in arrival times of a wavefront at the two antennas of each baseline. Theoretical delays are computed using the latest geophysical and astronomical models in compliance with the latest IERS Conventions. These models account for the effects of Earth rotation, precession/nutation, polar motion (wobble), solid Earth tides, ocean loading, Galacto-centric acceleration, the dry troposphere, and several other effects.  For the K-CRF the ionosphere corrections are computed by interpolating GNSS ionosphere maps. Global VLBI solutions normally solve for source coordinates as well as VLBI station positions and velocities, the five daily EOPs and residual atmosphere delays (mainly due to water vapour).

K-CRF Global solutions are computed and regularly updated by the United States Naval Observatory (USNO) in the US, using the Calc/Solve analysis package, and by the Technische Universität Wien (TU Wien) in Austria, using the VieVS analysis package. To improve our K-CRF astrometric catalogs and move us towards a better next ICRF solution, it is important to understand systematic differences seen between the astrometric catalogs produced at the two analysis centres and by the two independent analysis packages. It is also important to study, understand, and improve the methods and models used in our K-CRF analysis, and to probe and understand the underlying assumptions about the methods and models used. Two major challenges which limit the accuracy of the current K-CRF solutions are (1) external ionospheric corrections based on GNSS data and (2) the non-uniform observing network geometry, especially the lack of observations in the deep south.  A recent K-CRF study was conducted by Krásná et al., 2023, to try to understand the systematic differences seen between the astrometric catalogs produced by different analysis packages, to quantify the effect of modified ionospheric mapping functions on the K-CRF solution, and to explore the effect of geometry bias by computing elevation-dependent weighting to down weight low elevation observations. In addition, a collaborative effort between South Africa (SARAO), Switzerland (ETH), and Austria (TU Wien), funded through Leading House Africa and Swiss TPH, will drive the development of a machine learning-based ionospheric model for K-band. Known as KOSMIC (K-band VLBI Observations with Improved Scheduling and Ionospheric Corrections), this project is set to commence in February 2023.

Ideally, the effects of source structure should also be accounted for in Global VLBI astrometric solutions. While VLBI images of CRF sources show that, in general, they appear more compact at K-band (24 GHz) than at X-band (8.4 GHz), they can still exhibit measurable extended emission at K-band. However, such modelling require a more or less quasi-permanent ongoing imaging effort for all CRF sources. K-band offers that possibility because of the readily available VLBI images from our K-CRF observations and the dedicated campaign to map and monitor the structure of our K-band sources. To this end, we started a project to apply structure corrections directly to the data during the astrometric analysis process, using updated models of the source structure.

Under our K-band CRF Data Analysis and Optimisation VLBI Program we aim to study and understand the systematic differences between our estimated reference frames and address the challenges which limit the accuracy of our astrometric/geodetic solution by improving the methods and models used in our analysis.