Published values for the long‐term, global mean sea level rise determined from tide gauge records exhibit considerable scatter, from about 1 mm to 3 mm/yr. This disparity is not attributable to instrument error; long‐term trends computed at adjacent sites often agree to within a few tenths of a millimeter per year. Instead, the differing estimates of global sea level rise appear to be in large part due to authors' using data from gauges located at convergent tectonic plate boundaries, where changes of land elevation give fictitious sea level trends. In addition, virtually all gauges undergo subsidence or uplift due to postglacial rebound (PGR) from the last deglaciation at a rate comparable to or greater than the secular rise of sea level. Modeling PGR by the ICE‐3G model of Tushingham and Peltier (1991) and avoiding tide gauge records in areas of converging tectonic plates produces a highly consistent set of long sea level records. The value for mean sea level rise obtained from a global set of 21 such stations in nine oceanic regions with an average record length of 76 years during the period 1880–1980 is 1.8 mm/yr±0.1. This result provides confidence that carefully selected long tide gauge records measure the same underlying trend of sea level and that many old tide gauge records are of very high quality.

No abstract

No abstract

Woodworth [1990] has investigated past apparent sea level acceleration and observes that the large number of tide gauges in the Baltic-Northern European region, although subject to linear vertical movements from glacial rebound, provides data suitable for study of the nonlinear component of sea level change. However, in his study he did not exploit the regional spatial correlation of the sea level signal at low frequencies, which has the effect of overemphasizing the importance of the very large number of tide gauges in northern Europe. We shall see below that a compact oceanic region containing records of varying lengths should be modeled by a linear trend for each tide gauge record along with a single acceleration parameter that simultaneously satisfies (in the sense of least squares) all records. This paper is not subject to U.S. copyright. Published in 1992 by the American Geophysical Union.Paper number 92JC01133. Woodworth [1990] also considered existing long records outside of Europe and concluded that there is little or no evidence for a statistically significant apparent acceleration of sea level in them. In addition, he attempted to show by an extrapolation of (meteorologically corrected) historic sea level residual data from the Newlyn, England, site that the acceleration in a region might be detected in about 3 decades, subject to certain assumptions concerning sea level variability. The present paper is less optimistic, taking note of the ubiquity of large chaotic regional sea level events that endure over a decade or more. These events can distort an estimate of regional sea level acceleration over even very long records and are devastating to records of a few decades' duration.Before proceeding further, it is necessary to clarify what is meant in this paper, as well as in others such as Woodworth's [1990], by the term acceleration. This usage refers to the deviation of sea level from a linear trend over the data span in question that is modeled by an algebraic term of the second degree in time. It is necessary to explicitly state this because the true spectrum of sea level is red. This has the consequence that any particular span of sea level data, such as the period 1850 to present considered in this paper, will represent as an acceleration to a lesser or greater extent those low-frequency changes of sea level whose periods are significant in comparison to, or longer than, the data span. Thus the accelerations derived and referred to in this paper are really apparent ones, whose significance lies only in their amplitude relative to the acceleration predicted to accompany global warming.