Absolute frequency measurement with uncertainty below 1×10-15\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$1\times 10^{-15}$$\end{document} using International Atomic Time

The absolute frequency of the 87Sr\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^{87}\mathrm{Sr}$$\end{document} clock transition measured in 2015 (Jpn J Appl Phys 54:112401, 2015) was reevaluated using an improved frequency link to the SI second. The scale interval of International Atomic Time (TAI) that we used as the reference was calibrated for an evaluation interval of 5 days instead of the conventional interval of 1 month which is regularly employed in Circular T. The calibration on a 5-day basis removed the uncertainty in assimilating the TAI scale of the 5-day mean to that of the 1-month mean. The reevaluation resulted in the total uncertainty of 10-16\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$10^{-16}$$\end{document} level for the first time without local cesium fountains. Since there are presumably no correlations among systematic shifts of cesium fountains worldwide, the measurement is not limited by the systematic uncertainty of a specific primary frequency standard.