Empirical modeling and impact of transient effects on the mean sea level trend estimates from the global tide gauge data

Abstract

The mean sea level trend estimates from shorter records (less than 50 years) are easily influenced by a number of additional transient effects including atmospheric pressure variations, interannual and decadal changes in the mean sea level or various station dependent disturbances, which are not always accounted for in previous studies because their influences may be negligible for stations longer than 100 years or simply such information may not have been available. In this study, we detected and modeled 8,072 transient changes (a nearly periodic, or non-periodic variation, a shift or an episodic change in the mean sea level that may last several months or longer) from all of the globally distributed 1,862 tide gauge stations with approximately 47,000 years of tide gauge data in the Permanent Service Mean Sea Level repository. It was shown that 1,264 out of 1,862 globally distributed tide gauge station solutions were affected significantly by modeling transient changes in the mean sea level. The solutionsR2 values improved at a 95% confidence level with the inclusion of new empirical model parameters representing transient changes as mean shifts as well as the trend estimates that fall within [-1 to +3] mm/year interval.