I'll dice the data in two ways; both show an acceleration:
- fitting it to a 2nd-order polynomial
- calculating the change in the slope of the most recent 20 years.
The second-order polynomial fit shows an acceleration of 0.005 ± 0.009 mm/yr2, where the uncertainty is calculated assuming no autocorrelation.
That's still small, but it's been increasing steadily since the 2010-11 La Nina that took a lot of water from the ocean and dumped it on land -- which Josh Willis called correctly and Bjorn Lomborg called wrong:
This fit finds that the current rate of sea-level rise is 3.83 mm/yr.
The second method, calculating the change in the linear slope of the last 20-year's worth of data (20 being picked arbitrarily) finds an acceleration of 0.021 mm/yr2. (Sorry, I haven't calculated its uncertainty.)
Even a little bit of acceleration sends the projection up fast: here is the change in the projection to the year 2100, relative to 1993:
In other words, in late 2014 the projection for the year 2100, relative to 1993, was 30 cm; now, just 2/3rds of a year later, it's up to 38 cm.
(The AVISO data record now spans 22.5 years.)
One last number, since it gives me the opportunity to use my favorite unit, the Sverdrup (Sv), which equals one million cubic meters per second: the change in the volume of the ocean from land ice melting + water thermal expansion is 0.04 Sv, or almost 21 times the flow of water over Niagara Falls.
That's more than twice the average discharge of the Mississippi River into the Gulf of Mexico (= 0.017 Sv).