Ken Stewart, December 2011
Sea level rise is in the news again, with Nils-Axel Morner’s article in the Spectator and news in the Daily Telegraph that papers arguing against official claims of sea level rise have been stopped from being published.
It’s two years since the release of the CSIRO and Climate Change Department’s publication, Climate Change Risks to Australia’s Coast, (Dept of Climate Change and Energy Efficiency ,2009) which predicted sea level rise of up to 1.1 metres above 1990 levels by 2100. Scary stuff. This led to many local governments and State governments suggesting restrictions on coastal developments.
Since then we’ve had Climategate 1 and 2, some leadership changes, an election, floods and cyclones, and two cool years. We’ve also seen the Climate Commissioners broadcasting their doom and gloom, in their May report and their recent Climate Change and Health report. I’ll let others demolish the nonsense in this most recent one. Instead I’ll investigate the honesty and reliability of the Climate Commission, the CSIRO, and the Department of Climate Change, in their continued predictions of dangerous sea level rise.
In Chapter 2 of their May report, the Climate Commissioners state:
A plausible estimate of the amount of sea-level rise by 2100 compared to 2000 is 0.5 to 1.0 m. There is significant uncertainty around this estimate, the largest of which is related to the dynamics of large polar ice sheets
And show this diagram:
This is in line with Climate Change Risks to Australia’s Coast. The authors of both reports base the prediction of 1.0m this century largely on IPCC AR4, 2007, and claim this is supported by high-quality satellite observations which showed a rate of 3.1 mm per year between 1993 and 2003, and National Tidal Centre gauges at 14 locations around Australia which showed a mean rise similar to the global mean between the early 1990s and 2008.
0.5m per century is 5mm per year. 1m per century is 10mm per year. Therefore I will look at evidence for this claim from existing Australian data.
The claim of accelerated sea level rise from the early 1990s to 2008 should be examined in historical context. Here’s what Climate Change Risks to Australia’s Coast says:
For the past 6,000 to 7,000 years, the sea level has oscillated within a narrow band of plus or minus 2 metres. This period of relative stability is sometimes referred to as the ‘stillstand’ period (Box 1.1). The coast that we recognise today stabilised in this period.2
…Global mean sea level has risen about 20 centimetres since pre-industrial times (Figure 2.6), at an average rate of 1.7 millimetres per year during the 20th century.16…
(Sea level has risen) 1.8 millimetres per year from 1961 to 2003.
And satellites show a rise of 3.1mm/year from 1993-2003, which agrees with NTC data. So where do they get claims for acceleration?
These rates of increase are an order of magnitude greater than the average rate of sea-level rise over the previous several thousand years. … implying acceleration, for 2 metres in 6000 years is 0.3mm per year. However, sea level oscillated in a +/- 2 metre band- so the time frame for each rise or fall in this period was much shorter- if it rose or fell 2m in 2000 years, that would be 1mm per year. Is there evidence for this? Well, there is for sea level fall. From an article in the Journal of Coastal Research, here’s a diagram from a study of sand ridges at Beachmere on Moreton Bay in SE Queeensland.
Note on Ridge 7, pebbly sand couldn’t be wind blown, and sits above B Horizon, and B Horizon is at about 1.7m above current mean sea level. Ridge 7 is 1700 +/- 130 years old. 1700mm in 1700 years is 1mm per year. Ridge 6 is 1140+/- 80 years old, and B Horizon is about 1.5m above current MSL, so fall was rapid until 600 yrs ago.
Again from Climate Change Risks to Australia’s Coast
Sea level around Australia rose by about 17 centimetres between 1842 and 2002 – a rise in relative sea level of about 1.2 millimetres per year.17 The rise in sea level has been very variable from decade to decade. The rate of increase was low between the 1970s and early 1990s due to more frequent and severe El Niño events.18
The authors then explain how CSIRO has developed three simple scenarios for sea-level rise (relative to 1990), based on AR4.
Scenario 3 (High end) considers the possible high-end risk identified in AR4 and includes some new evidence on icesheet dynamics published since 2006 and after AR4.
The National Tidal Centre’s sites are at Cape Ferguson, Rosslyn Bay, Port Kembla, Burnie, Spring Bay, Portland, Port Stanvac, Thevenard, Esperance, Hillarys, Broome, Darwin, Milner Bay and Cocos (not used in this analysis). These were set up in the early 1990s as part of an international effort to accurately measure sea level rise. This is how Climate Change Risks shows the data:
The 2009-10 Annual Report of the National Tidal Centre says:
Although the length of record from Baseline stations is relatively short in climate terms there are a number of clear results emerging. The sea level records for all stations, when corrected for local land movement and changes in atmospheric pressure, demonstrate a regional pattern of sea level trends that is consistent with sea level changes detected by satellite-based altimeters.
While the BOM in its annual report 2009-2010 on the National Tidal Centres admits that the record is not long enough to derive any trends, it seems it was quite long enough for the CSIRO and the Climate Change Department.
So what is most likely for the next 89 years? 3.1mm/year, 5mm/year, or 10mm/year? Let’s look at the data from a selection of the NTC sites, and compare with nearby tidal gauges. Note that the Bureau of Meteorology has longer term records from a number of these sites, using earlier equipment, and one hopes the new gauges were calibrated with these records.
I downloaded sealevel data from BOM’s website for NTC sites and for selected nearby sites to get a longer record for comparison. This Table shows NTC sea level data with comparisons:
The grey highlighted lines are for non-NTC sites used for comparison. Column H shows the figures given by Climate Change Risks, which says only that the data is from “the early 1990s” to 2008, so cannot be replicated. Column D shows long term trend, column G shows the short term trend, column K is data for stations with more than 20 years data, and Column L is for those with more than 40 years data. Longer term sites show a lower trend.
All except two of the sites (Cape Ferguson and Milner Bay) have short term trends that are greater than their own long term trends or that of nearby sites. Sometimes they are only a little greater but mostly several times greater. Is this acceleration or natural variability?
Some sample plots:
To avoid the large early gap, here’s the plot from 1984
3.07mm/yr? and finally 1991
Townsville from 1959
1.35mm/yr ; and from 1991
Esperance from 1965
and Esperance from 1991
and Fremantle- from 1897!
1.6mm/yr. Fremantle from 1991:
Port Kembla from 1983 (earlier data spurious with gaps):
0.19mm/yr!! and from 1991
2.6mm/yr. Not too far away Sydney- NOAA has data from 1885
but NTC only from 1914
0.89mm/yr but from 1991
Differences may be due to uneven rise or fall of land and corrections for this.
Here’s a spaghetti plot of anomalies from 1991-2010 for all 19 records.
Apart from Milner Bay (at Groote Eylandt, in the Gulf of Carpentaria),which has a much greater variability, there is little inconsistency. Broome, Hobart and Darwin have some up and down spikes, and Burnie has a period in the 1950s that may be spurious. However, in general they are remarkably consistent- so much so that I am confident in calculating mean and median series.
1.4mm/year. The 6 year running mean clearly shows the interdecadal variability. There was a steady rise from 1920 to the late 1950s, particularly from the mid 1940s, followed by a flat/declining trend to the late 1980s. Note also the step ups every 20-30 years.
It is important to remember however that the early data is based on few stations:
Therefore, while we can say quite confidently that for most of the 20th Century sea level rise could not have been above 1.7mm/year, as stated in Climate Change Risks to Australia’s Coast, we cannot rely on data from only one or two stations. I have therefore plotted graphs of the mean from 1959 (when 4 sites became available) and 1966 (10 sites), to show trends over 52 and 45 years respectively.
1.15 or 1.2mm/year?
Compare this with the 1991-2010 graph for all 19 sites:
But what happens if we start from a different year? Try 1988:
Alright, that’s cherry picking, I started from a high point. Try 1987 instead:
2.63mm/yr – which is still a far cry from 3.8mm/year!
A lot depends on the start year! For another example, here’s another 20 year period of rapid sea level rise: 1938-1957
4.5mm per year! That’s much more than 1991-2010.
Another thing to remember is that the mean tells you nothing about local conditions, and sea level is all about local conditions. High tide is far more significant.
Let’s compare this with Ocean Heat Content (OHC):
Note the rises in 2003-04 and 2007-08- matches the mean graph, but not trending up.
And finally, the 1966-2010 graph of MSL and scaled SOI (monthly values/ 100).
The relationship is pretty obvious.
Sea level has been rising gradually since records began.
Interdecadal variability is highly visible. SOI is a major influence on sea level.
The rapid rise of the late 1990s is not unusual- it was more rapid 1938-1957. Any projections based on this short period are therefore spurious. There is no recent acceleration.
The longer the record, the lower the trend.
3.1mm/year rise has been achieved- in the 20 years since 1991. 5mm/year was almost reached 1938-57, but is almost double the trend since 1987.
10mm per year is more than 2.6 times the trend since 1991, and 8 times the trend of the last 50 years! That’s a lot of glaciers and ice sheets to melt, and a lot of ocean heat content!
There is nothing to indicate that the long term average sea level rise of 1.2 – 1.7mm per year (with periodic increase and decrease with the SOI) should not continue for the foreseeable future. Anything more is speculation. As BOM says in its NTC Report:
Sea level and climate can fluctuate about a long-term climatological mean from one decade to the next. The project to date is only just beginning to span two complete decades, so it is important to recognise that the sea level change observed over this time is largely a measure of decadal variability. Continued monitoring is needed to quantify the longer-term trend that is associated with climate change.
We hope that Dr Jones and his team don’t hear about this- they’ll correct, adjust, and homogenise for sure!