The streets flood even without storm, rain, wind, even on a calm sunny day. It didn’t used to happen, but now the highest spring tides of the year (around October) bring flood waters, septic systems back up, waste oozes onto streets and lawns, saltwater leaks into groundwater and spoils drinking supplies. It’s not a pretty sight, and it’s not confined to Miami and New Orleans, it’s all along the coast.
It’s an undeniable sign sea level is rising. But one wonders: how fast?
The best estimate of global sea level changes based on data from tide gauge readings is, in my opinion, that from Church & White (which I’ll call “CW”). Alternative choices have been discussed before, but that’s not what this post is about; here’s the data from CW:
The thin black line follows the data, the thick red line is a smoothed estimate (using the lowess smooth). There’s a blue box in the upper right drawing attention to the time period since 1993, when we also have data from satellites.
That’s all well and good, but how fast is the level rising? The smoothed estimate enables me to compute the rate of change, and I also did so by least squares regression on a 10-year window sliding throughout the data. Result here:
The red line is the rate estimated from smoothing, with pink shading showing its uncertainty range (2σ). Black triangles mark the rates estimated from linear least squares on 10-year windows. At this time scale at least, it seems sea level is rising faster now than it has been before.
It’s interesting to compare the overall rate before and after 1993 — when the satellite data kick in. Still using the CW data, the rate before 1993 is estimated at 1.46 mm/yr., but since then it has been going at 3.56 mm/yr. That suggests that yes, sea level might be rising considerably faster now than is has been before.
What about the satellites? I found four satellite-based data sets for global sea level which were easy to get: from the University of Colorado (U.Colo), the Copernicus group (Copernicus), the European Space Agency (ESA), and CSIRO in Australia. And here they are:
In addition, there are two more from CSIRO. They recently refined their “nominal” estimate by comparing the results to tide gauges, which themselves must be corrected for vertical land movement (VLM). They did so in two ways: using a Glacial Isostatic Adjustment model (GIA), and using data from GPS instruments (GPS). That gives three estimates from CSIRO:
Since we have six satellite-based data sets to try (and there are others I haven’t retrieved), what do they say about the rate of change? If we use least squares regression on each data set we’ll estimate the average rate during the time since 1993. Here’s what I got:
Interestingly, the highest estimate comes from the tide-gauged based data from CW, at 3.56 mm/yr, the lowest is from the GPS-refined CSIRO data at 2.91 mm/yr. Any way you look at it, it’s rising faster than it was before 1993.
When you find a pretty strong linear trend in some data (as we have), it’s often revealing to subtract that linear trend and look at the residuals left over. In this case, they’re fascinating:
It sure looks like something changed about 2011, and it’s easy to confirm that statistically. The rate of sea level rise increased. I used a lowess smooth to estimate how the rate has changed for each of the data sets, including the tide-gauge based data from CW (shown as a dashed line):
My analysis confirms that although the rate of sea level rise has averaged about 3 mm/yr during the satellite era, it’s faster now, rising at about 5 mm/yr.
Yes different data sets disagree. Of course — if they didn’t they wouldn’t be different. But the things they agree on are that sea level rise has accelerated, and that its rate now is closer to 5 mm/yr than to 3 mm/yr.
This blog is made possible by readers like you; join others by donating at My Wee Dragon.