Data was downloaded from PSMSL on 16 Nov 2011

Despite being in the middle of USA Colordao University is a centre of sea level research. This page on their web site lists various estimates of the rate of sea level rise over the last 100 years or so. They are all in range 1-3mm/year. So it would be reasonable to expect values of sea level rise found from PSMSL data to be in this range, or at least close to it.

Looking down the table, in order of decreasing length of station observation record, from Sheerness to Roysth the rate of sea level rise is more or less within the expected range. So having 29 or 30 years of observations seem to be a guide to useable rate of sea level rise.

Looking down to 20 years the following stations seem problematic

So a rough rule of thumb would be to take stations with more than 25 years of observations, and exclude Whitby.

Even so the rate of sea level rise varies from -0.3 to 2.8 mm/year, a range of 3.1mm/year with average and standard deviation of 1.37 ± 0.88 mm/year.

Just considering mainland Britain (leaving out Lerwick, Douglas, Holyhead and Belfast) the range is 0.05 to 2.66 mm/year, a range of 2.61 mm/year with average and standard deviation 1.48 ± 0.75 mm/year.

Part of the reason for this variation is that land moves and moves by different amount in different locations. There are two effects:

1. Across Britain as a whole Scotland and the North of England are supposed to be rising, whilst the South of England sinks, like a big see-saw.
2. In addition land may be rising or falling locally.

There are various estimates of land moving (isostatic), for example page 104 of Absolute Fixing of Tide Gauge Benchmarks and Land Levels

Note if you don't like these estimates there are plenty of others you can choose from.

Taking these estimates of land movement into account the net rate of sea level rise at the seven locations is.

Taken as whole the tide gauges in UK show a range of sea level rise from -7.24 to 11.21 mm/year, a range of 18.45 mm/year! By

• Only taking Stations with at least 25 years of observations
• Excluding Whitby, which is flagged as faulty.
• Excluding stations not on mainland Britain.
• Taking only stations with an estimate for how much land has risen or fallen.

The result is a range 1.67 mm/year with average and standard deviation 1.13 ± 0.55 mm/year.

Update 20 January 2012 - According to Prof Philip Woodworth MBE

Bill Donovan of EA has sent me a paper by Prof Philip Woodworth and others which discusses "Trends in UK Mean Sea Level Revisited" and is dated 4 August 2008.

Philip Woodworth received his MBE in recognition of his work as an international sea-level scientist. He was the Director of the Permanent Service for Mean Sea Level from 1987 to 2007.

You can download the paper from here.

The summary states

The tide gauge MSL trends for 1901 onwards are estimated to be 1.4 ± 0.2 mm/year larger than those inferred from geology or geodetic methods, suggesting a regional sea level rise of climate change origin several 1/10s mm/year lower than global estimates for the 20th century.

1.4 ± 0.2 mm/year is close to value I obtained of 1.13 ± 0.55 mm/year, especially bearing in mind different time periods and some different stations have been used.

There is good agreement for most stations comparing values from table 1 of Woodworth paper with values listed above.

Remember The Data Has Been Adjusted

The data used to generate plots of sea levels that are commonly displayed is adjusted to account for tide gauges being replaced or moved to different locations. To illustrate how dramatic these adjustments can be here is the raw data for Sheerness.

and here is the adjusted data

Clearly there is the possibility of introducing an artificial sea level rise or fall when making these adjustments.

The Trend Is Tiny Compared To Tides

Around UK tides rise and fall about 3 to 4 meters twice a day.

The trend for sea level rise is 1 to 2 millimeters per year, roughly a million times smaller than the tidal movement.

What Are The Measurement Errors

We don't know as PSMSL does not record measurement errors. There is always some error in any measurement and it is normal to give an estimate of this error.

At a guess it would be of the order of ± 1mm.

Sea Level Rise, Climate Change and IPCC

It is often said that sea levels are rising due to ice melting at the poles.

This is not the main reason. The expansion of sea water with rising temperature is responsible for most of this rise. IPCC in AR4 state that thermal expansion is responsible for 70 to 75% of sea level rise. I.e. predicted sea level rise largely depends on predicted rise in temperatures.

The corollary is that if temperatures do not rise as much as predicted, or even fall, then there will be no large rise in sea levels.

NASA Buoys Satellites and Sea Level Rise

This graph which appears on JPL/NASA website clearly shows that sea levels fell by 6mm in 2010. In other words in one year when the trend is for sea levels to rise 3.2mm they fell 6mm, a difference of 9mm. This is nearly 3 years rise at the trend rate.

Does this mean temperatures went down too?

This graph of sea level rise around UK since 1900, by Prof Phil Woodworth, shows that there was an increase in the rate of sea level rise between 1993 and 2000, but this followed a very sharp fall around 1990/1.

Between about 1993 and 1998 sea levels did rise faster than the 1.4mm/year trend.

• Mount Pinatubo erupted in 1991 and lowered temperatures and sea levels for a few years.
• There was a very strong El Nino in 1997/8 which raised temperatures and sea levels
• There was another, weaker, El Nino in 2002/3

And looking over the whole period since 1900 nothing out of the ordinary happened, this pattern of sharp rises and falls has happened several times and the long term trend is 1.4mm/year.

Sea Level Rise, IPCC and DEFRA

DEFRA, the Department for Environment, Food and Rural Affairs, is the UK government agency with overall responsibility for, well, environment food and rural affairs. DEFRA devolves responsibility for specific areas to other agencies which are known as "Delivery Partners" in jargon. For example:

Being somewhat suspicious I sometimes wonder if this proliferation of agencies isn't a scheme worthy of Sir Humphrey.

• EA and NE can claim they are merely following DEFRA's guidance as they are required to do.
• DEFRA can claim they are not responsible as EA and/or NE are the agency implementing a particular project.
• EA and NE can also point the finger at each other.

Sea Level Rise - DEFRA 2006 Predictions

DEFRA 2006 predictions were based on IPCC Third Assessment Report and are as follows.

These predictions feature an accelerating rate of sea level rise. This is strange considering all those graphs with a linear trend line.

As the predictions start from 1990 there is over 20 years of observational data to compare them with. The predictions are clearly far too high. Even the 1990 to 2025 are at least twice what has been observed.

Sea Level Rise - DEFRA 2009 Predictions

DEFRA 2009 predictions were based on IPCC Fourth Assessment Report. This time there are three main versions, high medium and low. The following table gives sea level rise in mm from 1990.

These predictions are roughly half the 2006 predictions. However they still feature an accelerating rate of sea level rise and are still too high as almost all the observations are lower than the "Low" version of predictions.

Sea Level Rise - DEFRA 2009 H++

The previous section stated there are three main versions of the DEFRA 2009 predictions. There is also a worst case scenario, called H++, which is extremely unlikely but, according to DEFRA, physically possible.

DEFRA do not attempt to assign a probability to this H++ scenario and give a wide range of possible sea level rise from 930 to 1900mm by 2100.

The background to this H++ scenario is melting ice at poles. For more details see here.

It is worth stressing that although probability type language is used (e.g. 5th percentile 95th percentile) this is NOT related to the probability of an outcome occurring in the real world. Several computer models were used to make these projections and the percentile refer to the fraction of model runs when a particular result was at or below the quoted figure.

In other words there is absolutely no information about the actual probability of any event occurring in the real world. So the probability of H++ occurring might be 10%, 1% or 0.0000001%.

What do you think?

If you live in the UK it is worth bearing in mind that EA are planning to spend £billions on Managed Realigment based on these projections. NE is planning to spend a £billion or so on compensation to farmers whose land is flooded.

Shoreline Management Plan - Practical Implications Of DEFRA Predictions

EA has been preparing the latest version of it's Shoreline Management Plan (SMP) which outlines in great detail how EA plan to defend the UK coast for next 1000 years.

DEFRA requires EA to use their projections for SMP. A serious consequence of this is that EA decides that it is too costly to defend a stretch of coast and then either abandons it or actively makes a hole in the sea wall (this is called "Managed Realignment").

People have pointed out to EA that observed sea level rise was significantly less than predicted by either DEFRA 2006 or DEFRA 2009. EA have replied that they are required to use DEFRA projections.

The H++ scenario has proved convenient as EA can now say that they are aware the projections are greater than observations but they are required to plan for a worst case scenario, even though it is very unlikely.

It is fine to use a worst case scenario for planning.

Except

When you decided on an extremely unlikely worst case scenario that it is not possible to maintain coastal defenses and you have to knock them down right now.

NASA Buoys Satellites and Sea Level Rise - Again

You may have notice that the trend shown on the NASA Satellite graph was 3.2mm/year.

But the trend from most of the tide gauges, after accounting for land movement, is closer to 1 mm/year.

There is much lively discussion about this for example here and www.appinsys.com/globalwarming/GW_4CE_SeaLevel.html

www.appinsys.com/globalwarming/GW_4CE_SeaLevel.htm has this graph which compares about a century of tide gauge data with 10 years of satellite data. It is clear that the trend that is reported from the satellite data is higher than that from tide gauges. IPCC claim that satellite data shows that sea level rise has accelerated since 1993. This claim is repeated by DEFRA and EA.

Strange sea level rise started accelerating just after the satellites went up.

It is worth remembering that data from satellites is also adjusted before being presented. Maybe this adjustment is responsible for some or all of the difference with tide gauge data.

Also tide gauges measure sea levels close to land at a specific location. Satellites measure levels of entire oceans, but only between latitudes 66 ° N and 66 ° S.

Comparision With Blog Graphs

The data for this page was downloaded from PSMSL on 16 Nov 2011. The data used for posts on my blog during 2011 used data downloaded on 31 May 2011. The following table shows the stations where there was a difference in the rate of sea level rise.

The differences are mostly small or very small apart from Workington which is huge. The notes at PSMSL show the data for this station was adjusted on 2 July 2011!

You can see the table of sea level rise and the graphs based on 31 May 2011 data here

Summary

Here I have provided graphs from stations around UK with linear trendlines.

I hope that I have also given an indication that something that might at first seem simple and well understood is not so simple and there is uncertainty and error.

It is not clear why there are so many stations with short or incomplete records.

If you would like to comment or think I have something wrong please email me