Introduction

I have blogged about sea level rise, comparing data from PSMSL with the projections made by DEFRA in 2006 and 2009. These projections are used by Environment Agency and other organizations to plan how to defend the UK coastline. Unfortunately most of these plans seem to feature knocking holes in sea walls and flooding parts of UK.

www.psmsl.org is the global data bank for long term sea level change information from tide gauges and bottom pressure recorders. PSMSL hold records for just over 2087 stations around the world, 55 from UK.

The main purpose of this page is to list the rates of sea level rise for the UK stations which PSMSL do not show. For each station there is a link to a graph showing

 

Meaning of columns in table below

ID

PSMSL's station number. Clicking on an ID links to data held by PSMSL for this station.

It is well worth looking at the notes that PSMSL holds for a station before accepting the numbers. For example the notes for Cromer say

Complete station flagged as suspect

Station Name
Clicking on a station name links to a page which shows a graph with
  • monthly observations
  • trend line
  • DEFRA predictions
And a separate graph showing unadjusted observations (explained below).
Completeness

Most stations have some months with missing observations. In the case of Sheerness there are gaps of 50 years!. Completeness indicates the percentage of months which have observations between Start Year and End Year.

Each monthly observation is actually the average value of daily observations for that month. Sometimes daily observations are missing. This is NOT indicated by Completeness.

Data was downloaded from PSMSL on 16 Nov 2011

Id
Station Name
SLR
mm/year
Start Year
End Year
Years
Completeness
-0.11
1957
2010
53
 91%
1.22
1965
2010
45
 94%
-3.03
1959
1971
12
 99%
5.12
1971
1982
11
 99%
0.96
1931
2010
79
 92%
0.59
1862
1965
103
100%
1.86
1964
1993
29
 99%
4.03
1956
1971
15
 81%
2.25
1989
2010
21
 89%
0.45
1913
1950
37
100%
1.94
1895
2010
115
 93%
7.76
1981
2010
29
 87%
0.95
1959
2010
51
 97%
11.21
1988
2010
22
 93%
2.47
1955
2010
55
 95%
0.05
1980
2010
30
 86%
1.17
1960
1974
14
 93%
5.15
1968
1978
10
 87%
1.25
1933
1983
50
 96%
1.69
1961
1983
22
 96%
1.67
1832
2009
177
 55%
2.37
1961
2010
49
 94%
3.83
1991
2010
19
 93%
1.51
1961
2010
49
 95%
2.98
1996
2010
14
 99%
1.25
1991
2010
19
 95%
2.19
1961
2010
49
 95%
1.79
1915
2010
95
 99%
3.69
1994
2010
16
 92%
2.08
1983
2010
27
 81%
4.02
1990
2010
20
 96%
6.99
1986
2000
14
 87%
8.16
1993
2010
17
 92%
0.22
1960
1985
25
 44%
-2.24
1963
1983
20
 74%
7.15
1987
2010
23
 90%
-4.22
1974
1986
12
 97%
5.37
1988
2010
22
 96%
6.07
1991
2010
19
 83%
2.80
1938
2010
72
 82%
1.90
1955
1972
17
 97%
1.08
1858
1983
125
 70%
9.86
1991
2010
19
 90%
2.66
1960
2010
50
 84%
-7.24
1992
2010
18
 96%
0.35
1938
1977
39
 86%
2.32
1968
2010
42
 97%
1.23
1968
2010
42
 77%
8.26
1991
2010
19
100%
1.76
1989
2010
21
 85%
1.04
1981
2010
29
 93%
2.23
1977
2010
33
 89%
2.82
1991
2010
19
 84%
-0.30
1957
1979
22
 66%
-0.17
1917
1963
46
100%

 

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

Whitby

Notes at PSMSL state
Several problems in 90's. Treat as suspect from 97 onwards.

Cromer

Notes at PSMSL state
Complete station flagged as suspect

Milton Haven

Hakin and Newton Noyes have very different rates of sea level rise (7.15 and -2.34) despite being about a mile apart.

Fishguard

Fishguard 2 and 1 also have very different rates (5.37 and -4.22) despite being even closer than the 2 Milton Haven stations.

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.

Station Name
SLR
mm/year
Land Movement
mm/year
Net SLR
mm/year
Aberdeen 1
0.96
0.11
1.07
Aberdeen 2
0.59
0.11
0.70
North Shields
1.94
-0.48
1.46
Lowestoft
2.47
-1.47
1.00
Sheerness
1.67
-1.09
0.58
Portsmouth
1.51
-0.85
0.66
Newlyn
1.79
0.46
1.33
Liverpool Georges And Princes Piers
1.08
0.25
1.33

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

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.

Station
Woodworth SLR
mm/year
JeremyShiers.com
SLR mm/year
Absolute Difference
mm/year
Note
Lerwick
-0.68
-0.11
0.57
Wick
1.55
1.22
0.33
Aberdeen Composite
0.87
0.96
0.09
1
Rosyth
1.99
1.86
0.13
Dunbar
0.47
0.45
0.02
North Shields
1.92
1.94
0.02
Immingham
0.54
0.95
0.41
Lowestoft
2.57
2.47
0.1
Southend
1.22
1.25
0.03
Tilbury
1.58
1.69
0.11
2
Sheerness
2.23
1.67
0.56
2
Dover
2.18
2.37
0.19
Portsmouth
1.58
1.51
0.07
Devonport
2.55
2.19
0.36
Newlyn
1.70
1.79
0.09
Holyhead
2.31
2.80
0.49
Liverpool Composite
1.60
1.08
0.52
3
Heysham
2.73
2.66
0.07
Douglas
0.26
0.35
0.09
Portpatrick
1.95
2.32
0.37
Millport
1.20
1.23
0.03
Ullapool
2.12
1.04
1.08
Stornoway
2.22
2.23
0.01
Belfast 2
-0.25
-0.17
0.08

Note
Comment
1
Woodworth combines both Aberdeen stations. JeremyShiers.com uses Aberdeen2
2
Woodworth finds quite different values for Sheerness and Tilbury even though these are close.
3
Woodworth combines both Liverpool stations. JeremyShiers.com uses only Georges And Princes Piers. There is only 19 years of data from Gladstone dock and there seems to be a surprising jump of about 100mm after 2000.

 

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.

raw data for sea level observations at sheerness

and here is the adjusted data

adjusted data for sea level observations at sheerness

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.

graph of nasa satellite sea level meaurements

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.

UK Mean Sea Level Change 20 Century

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

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:

EA
Environment Agency

Purpose to protect or enhance the environment, taken as a whole so as to promote the objective of achieving sustainable development.

One part of EA's responsibility is flood and coastal risk management. So in a practical sense they are concerned with sea level rise or fall.

NE
Natural England

Responsible for natural environment including land, soil, flora and fauna, marine and freshwater environment. NE is required to see these are protected and improved.

In addition NE has a responsibility to help people enjoy, understand and access the natural environment. This includes improving public access to the countryside.

NE is responsible for saltmarsh and intertidal habitats which they feel are being eroded by rising sea levels. So they are concerned with sea level rise too.

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

Sea Level Rise - DEFRA 2006 Predictions

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

 
London, East England
South West and Wales
Scotland and North England
Period
mm/year
Period Total
mm/year
Period Total
mm/year
Period Total
1990-2025
4.0
140.0
3.5
122.5
2.5
87.5
2025-2055
8.5
255.0
8.0
240.0
7.0
210.0
2055-2085
12.0
360.0
11.5
345.0
10.0
300.0
2085-2115
15.0
450.0
14.5
435.0
13.0
390.0
Total Sea Level Rise
1990 to 2115
1205.0
1142.5
987.5

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.

 
London
Cardiff
Edinburgh
Belfast
High
Med
Low
High
Med
Low
High
Med
Low
High
Med
Low
2000
35
30
25
35
29
25
22
16
12
23
17
13
2010
73
62
53
73
62
53
47
35
26
49
38
28
2020
115
97
82
115
97
82
75
57
43
78
60
46
2030
160
135
114
159
134
114
107
82
61
111
86
66
2040
208
175
148
208
175
148
142
109
82
147
114
87
2050
258
218
184
259
218
184
180
139
105
186
145
111
2060
314
263
222
314
263
222
221
171
130
229
178
137
2070
372
312
263
371
311
263
266
206
157
274
214
165
2080
433
363
305
433
362
305
314
244
186
323
253
196
2090
497
416
350
497
416
350
365
284
218
376
294
228
2095
531
444
373
531
444
373
392
305
234
403
316
245

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. comparision of sea level rise measured by tide gauge and satellite

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.

 

Station Name
16-Nov-2011
31-May-2011
Difference
-0.11
-0.08
-0.03
1.22
1.39
-0.17
0.96
0.99
-0.03
2.25
2.52
-0.27
1.94
1.95
-0.01
7.76
7.54
0.22
0.95
0.88
0.07
11.21
11.78
-0.57
2.47
2.43
0.04
0.05
-0.21
0.26
2.37
2.36
0.01
3.83
4.05
-0.22
1.51
1.49
0.02
2.98
3.00
-0.02
1.25
2.20
-0.95
2.19
2.20
-0.01
1.79
1.77
0.02
3.69
3.54
0.15
2.08
1.91
0.17
4.02
4.27
-0.25
8.16
7.80
0.36
7.15
7.01
0.14
5.37
6.19
-0.82
6.07
7.23
-1.16
2.80
2.77
0.03
9.86
10.39
-0.53
2.66
2.82
-0.16
-7.24
7.85
-15.09
2.32
2.40
-0.08
1.23
1.31
-0.08
8.26
8.51
-0.25
1.76
2.19
-0.43
1.04
1.39
-0.35
2.23
2.48
-0.25
2.82
3.74
-0.92

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