We measured the components of the 31-m-long vector between the two very-long-baseline interferometry (VLBI) antennas at the Kokee Park Geophysical Observatory (KPGO), Hawaii, with approximately 1 mm precision using phase delay observables from dedicated VLBI observations in 2016 and 2018. The two KPGO antennas are the 20 m legacy VLBI antenna and the 12 m VLBI Global Observing System (VGOS) antenna. Independent estimates of the vector between the two antennas were obtained by the National Geodetic Survey (NGS) using standard optical surveys in 2015 and 2018. The uncertainties of the latter survey were 0.3 and 0.7 mm in the horizontal and vertical components of the baseline, respectively. We applied corrections to the measured positions for the varying thermal deformation of the antennas on the different days of the VLBI and survey measurements, which can amount to 1 mm, bringing all results to a common reference temperature. The difference between the VLBI and survey results are 0.2 ± 0.4 mm, −1.3 ± 0.4 mm, and 0.8 ± 0.8 mm in the East, North, and Up topocentric components, respectively. We also estimate that the Up component of the baseline may suffer from systematic errors due to gravitational deformation and uncalibrated instrumental delay variations at the 20 m antenna that may reach ± 10 and −2 mm, respectively, resulting in an accuracy uncertainty on the order of 10 mm for the relative heights of the antennas. Furthermore, possible tilting of the 12 m antenna increases the uncertainties in the differences in the horizontal components to 1.0 mm. These results bring into focus the importance of (1) correcting to a common reference temperature the measurements of the reference points of all geodetic instruments within a site, (2) obtaining measurements of the gravitational deformation of all antennas, and (3) monitoring local motions of the geodetic instruments. These results have significant implications for the accuracy of global reference frames that require accurate local ties between geodetic instruments, such as the International Terrestrial Reference Frame (ITRF).

我们测量了夏威夷科基公园地球物理观测站 (KPGO) 的两个超长基线干涉测量 (VLBI) 天线之间的 31 米长矢量的分量，使用来自专用 VLBI 观测的相位延迟观测值精度约为 1 毫米在 2016 年和 2018 年。两个 KPGO 天线是 20 m 传统 VLBI 天线和 12 m VLBI 全球观测系统 (VGOS) 天线。国家大地测量局 (NGS) 在 2015 年和 2018 年使用标准光学调查获得了两个天线之间矢量的独立估计。后一项调查的不确定性在基线的水平和垂直分量中分别为 0.3 和 0.7 mm . 我们在 VLBI 和测量测量的不同日期对天线的不同热变形的测量位置进行了校正，可以达到 1 毫米，使所有结果达到一个共同的参考温度。在东、北和上地心分量中，VLBI 和测量结果之间的差异分别为 0.2 ± 0.4 mm、-1.3 ± 0.4 mm 和 0.8 ± 0.8 mm。我们还估计，由于重力变形和 20 m 天线处未校准的仪器延迟变化可能分别达到 ± 10 和 -2 mm，基线的 Up 分量可能会遭受系统误差，导致精度不确定性约为天线的相对高度为 10 mm。此外，12 m 天线的可能倾斜将水平分量差异的不确定性增加到 1.0 mm。这些结果突出了以下重要性：（1）将场地内所有大地测量仪器的参考点的测量值校正到一个共同的参考温度，（2）获得所有天线的重力变形测量值，以及（3）监测当地大地测量仪器的运动。这些结果对全球参考框架的准确性具有重要意义，全球参考框架需要在大地测量仪器之间建立准确的局部联系，例如国际陆地参考框架 (ITRF)。