Variance of Newtonian constant from local gravitational acceleration measurements

We use IGETS absolute gravitational acceleration measurement data to study the variance. The relative variance of $\delta g /g$ in 22 years is less than $4\times 10^{-8}$. Since G /G\lessapprox\delta /g $, this implies Newtonian constant $3\times 10^{-9}$ based on an sine-like oscillation hypothesis. This limit at least 4 orders magnitude better existing $G$ measurements. scattered values reported measurements coming from different experiments are most probably systematic errors associated with these and not due intrinsic time variation $G$. also find that $\dot{ G} /G<5.61\times 10^{-10} \text{yr}^{-1}$ a linear best terrestrial result so far.