Assessment of Local Vertical Variation Aimed at the Integrity of Tide Gauge Time Series

Abstract

Tide gauge time series are fundamental to integrating height and bathymetric reference systems. Therefore, it is essential to separate signals of non-oceanic origin from such observations. This paper aims to evaluate the local vertical motion variation of regions that host tide gauge stations by considering aspects of the quality of the leveling network and temporal analysis. The leveling network of Imbituba's Tide Gauge Station was evaluated according to quality control widely used in Geodesy. Thus, the following analyses were performed: residuals, internal reliability, redundancy number, and overall model test. For the evaluation of the vertical variation of the geodetic stations, a preliminary classification was made concerning the construction type, the magnitude of the temporal variation rates, outlier detection, and confidence level. The quality control routines applied indicated that the network behaved homogeneously, and the observations had better than expected accuracies. Most residuals were less than 0.5 mm, and partial redundancies were classified as "sufficient" (0.1 ≤ r i ≤ 0.3) and "good" " (0.3 ≤ r i ≤ 1). The overall model tests showed that the observations are better than the ones established in the pre-analysis. The vertical variation of the geodetic stations classified as A and B showed an average vertical variation rate of +0.02 mm/year and no outlier. In contrast, the class C stations, after the identification and separation of outliers, 53% of the class sample, showed an average vertical variation rate of -0.17 mm/year, indicating a characteristic of settlement. It was evident that temporal monitoring is indispensable to the corrections for determining integral tide gauge series, free of influences of non-oceanic origin.