Houston, J.R., and Dean, R.G. 2013. Effects of sea-level decadal variability on acceleration and trend difference.
Previous research has shown that sea-level acceleration determined from individual tide gauge records has remarkably large scatter as record lengths decrease due to decadal variations in sea level. We extend previous data sets to the present time and find even greater acceleration scatter. Using analytic solutions, sinusoidal oscillations with amplitudes and periods of typical decadal variations are shown to basically account for the relationship between record length and both acceleration and trend difference. Data show that decadal variations will obscure estimates of underlying accelerations if record lengths of individual gauges are not greater than at least 75 years. Although worldwide data are less affected by decadal variations than individual gauge data, decadal variations still significantly affect estimates of underlying accelerations, in particular for record lengths less than about 60 years. We give two examples of recent studies that use record lengths of about 30 to 60 years to determine acceleration or related trend difference. Previous authors dismissed the importance of decadal variations on their results and, as a result, reached invalid conclusions.