4. Discussion

1. From Fig. 2, the correlation at interannual time scales is quite evident between the results of the two different techniques. The similarities of astrometric and gravimetric curves shown in Fig. 3 further enhance the reality of the correlationship. There is a scale difference between the two derived results, which will be explained below.
2. In gravimetry, the Vening-Meinesz formula is used to estimate the plumb line deflection at an observatory (Barlik & Rogowski 1989):
   
   in which, A denotes the azimuth from the observatory to the point with a given gravity anomaly , the angular distance, and the Vening-Meinesz function. In the case at Jozefoslaw observatory, the anomaly integration range has been limited to a very small angular distance radian). The estimate obtained in such a case, different from the real one , and its corresponding correction are the following:
   
   

   
   
   

   From a simulation under the assumption that the anomalies, within a 75 km range, are correlated as , where denotes the anomaly at the observatory and K is a constant, the
   

   scale of the correction will be four times greater than that of obtained. Thus, the scale difference shown in Fig. 3 is possible in practice.

3. The only common component existing in both the latitude residuals and the PLV from the gravimetric technique is the PLV itself at Jozefoslaw observatory. From the correlationship, including the level of significance and zero phase lag between the two time series derived from different techniques, one can conclude that the result of the interannual PLV measured by the astrometric technique is reliable.