RealClimate logo

Thames Barrier raised again

Filed under: — gavin @ 14 January 2014

Thames_Barrier_JWolfeBack in 2007 I wrote a post looking at the closures of the Thames Barrier since construction finished in 1983. Since then there has been another 7 years of data* and given that there was a spate of closures last week due to both river and tidal flooding, it seems a good time to revisit the topic.

The Barrier is in the Thames Estuary in the South-East UK and is raised whenever there is a risk of flooding in London proper. The Thames below Teddington weir is tidal with quite a wide range and so (the very real) risks are greatest at high tide. Risks are elevated for one of two reasons (and occasionally both) – either the river flow is elevated, and so a normal high tide would cause flooding (a ‘fluvial’ risk) or the tide itself is elevated (usually associated with a storm surge) (a ‘tidal’ risk). Both river floods and storm surges have a potential climate change component, and so the frequency of closure and the reasons why might be useful in assessing whether risks have changed in recent years. Countering that though, there have been non-climatic changes to how flooding risks are managed, improved hydrological modelling which gives better predictability, and local subsidence issues which will increase flooding risks but is not climatic in origin. Connecting flooding (or flood risks) to climate change is always complicated – as this recent Carbon Brief posting reminds us.

So how many closures have their been?


Red denotes the closures that are tidal in origin, grey is the total number of closures. Obviously tidal closures are dominant, but closures due to the river flooding in 2001, 2003 and this season are very notable. However, for the reasons outlined above and discussed previously, the raw number of closures – despite having increased over the years, is not a clean indicator of climate changes. We can look a little deeper at the data though.

For each closure, there is a record of the high water level at Southend (at the mouth of the Estuary), the river flow at Teddington and how much of the high water level was due to a storm surge. These are plotted below for each of the closures. Note that closures come in clusters, all are during the winter months (Sep-Apr) and in cases where the river flow is high, closures can happen for multiple high tides in a row.


Looking at the changes in these causes is interesting. The highest storm surges were in 2007 and 1993 but there is no apparent trend. The river flow data are more enigmatic, with the peak flows post-2000 standing out, but since there is much better data on the history of the Thames flow, that data should be looked at instead to determine trends. There is also a summary description of the first 100 years of data at Teddington available.

The high water record is perhaps the most interesting and there seems to be an upward trend in the peak levels. However, sea level rise at Southend to 1983 was about 1.2 mm/year and in the 30 years since then, one would expect 3.5 to 5.0 cm more, which is much less than what one would deduce from the peaks in the figure. So, again, a deeper look into the more direct data is probably warranted.

I therefore conclude, as I did in 2007, that:

Thames Barrier closings tell a complicated story which mix climate information with management issues and are probably too erratic to be particularly meaningful – if you want to say something about global sea level, then look at the integrated picture from satellites and tide gauges. But it is a good illustration of adaptive measures that are, and will increasingly be, needed to deal with ongoing climate change.

* Updated data thanks to the UK Environment Agency (via @AlanBarrierEA @EnvAgency) with a helping hand from Richard Betts.

40 Responses to “Thames Barrier raised again”

  1. 1
    wili says:

    Can anyone shed any light on the changes (mentioned in the second paragraph) in “how flooding risks are managed.” Obviously, the closures are human decisions and are based on all sorts of shifts in policy, politics, understanding of risks, economics…If there have been obvious major shifts along any of these lines that someone posting here knows of, it would be helpful to know them–and it would be nice to be able to point to some such specific policy shift to counter any denialist claims that the lack of trend in closure graphs show that sea level rise is a ‘hoax,’ or some such nonsense.

    Meanwhile, I have a question about what happens to the surrounding areas when there is a closure. Isn’t their necessarily more flooding in the immediately adjacent low-lying locations? Is there push back from these localities about sacrificing them for the safety of Londoners?

    I’m also not clear on how such a barrier can help if the water is coming from up-stream.

  2. 2
    Alex says:

    Meanwhile, I have a question about what happens to the surrounding areas when there is a closure. Isn’t their necessarily more flooding in the immediately adjacent low-lying locations?…I’m also not clear on how such a barrier can help if the water is coming from up-stream

    The barrier is closed at the low tide preceding the forecast flooding event. Therefore, the whole tidal range is available to hold water coming from inland. When the alarm is over, the barrier is opened, again at low tide, releasing the water out to sea. Elegant!

  3. 3
    mike says:

    “I’m also not clear on how such a barrier can help if the water is coming from up-stream”. Wili

    The barrier is not intended to deal with water coming from upstream. It is unlikely that so great a precipitation event in itself would be problematic for London – the central cities of (old Roman/Medieaval London and Westminster) are built on river terraces well above any conceivable flood level, and there is no evidence that the frequency of extreme precipitation events in the Thames catchment has increased. Much of the catchment is chalk or limestone which allow infiltration to groundwater.

    The barrier is built to deal with extreme tidal and storm surges like that that devastated the east coast of England much of Holland in1952. A combination of high tides and low pressure cyclonic systems will raise sea levels sufficient to transgress sea defences and the banks of the lower Thames (and Rhine). The funnel shape of the North Sea and the Thames estuary itself are sufficient to magnify the rise in sea level which in the open oceans will be more moderate. It is when there are these extreme storm surges combined with high tides that the barrier is operated to prevent a surge moving upstream to flood London.

    [Response: Actually heavy river flow and normal high tides can cause flooding (these are the ‘fluvial’ cases), and the barrier is raised in those circumstances to prevent the high tide from moving upstream. – gavin]

  4. 4
    tom mallard says:

    Current global estimates of sea-level rise are above 3mm/year so the 1.2mm/year represents tectonic rebound still ongoing and that’s not likely to help by 2100 because global sea-level doubled quickly in the most recent past, fast enough to know that the time of doubling is faster as time passes mainly from small glaciers but the ice-sheets are showing rapid change and adding more than assumed by most models, those being updated by recent work in Antarctica and Greenland.

    Another consideration moving forward is that fracking contributes to global warming, just sayin’, if you want something to burn use algae feeding on wastewater in photo-bioreactors at the sewage treatment plant and you’ll have tons of it to burn from a renewable source, just work on catalytic converters to reduce soot and particulates, runs in any IC-engine on the planet with little refinement tweaks versus cracking crude oil.

    Then recent studies by Caltech show vertical-axis windmills grouped extract 10% more power than propeller models in the same footprint and don’t kill birds.

  5. 5
    Jim Larsen says:

    1 wili said, “I’m also not clear on how such a barrier can help if the water is coming from up-stream.”

    That puzzles me as well. I suppose that if the increase in height of the tide was larger/faster than the backup of the river, then closing the gates makes sense.

  6. 6
    GlenFergus says:

    #3 Mike:

    Effectively, tidal energy is being used to “pump” the upstream flood water seawards. One can achieve much the same effect with a fixed barrier and mechanical pumping (as is done at many town levee systems), but the scale rapidly becomes prohibitive on large rivers.

    This illustrates an important point about the cost of adaption to sea level rise. There will typically be a large discontinuity in cost and practicality of barrier systems once the sea level rise exceeds the local tidal range.

  7. 7
    Dave says:

    … Just to be absolutely clear, the Thames barrier is raised to deal with tidal issues and has nothing to do with river flow due to rainfall.

    [Response: Not so. River flooding reduces the threshold at which high tides become a problem, so when high river flows (such as in 2001 and 2003). – gavin]

  8. 8
    Fergus Brown says:

    To understand flood risk for London and surrounding areas relative to the Thames Barrier and the Thames water catchment, you need to know that the river levels above Teddington are controlled in part by the sluice gates which release river water into the tidal Thames, and are opened or closed according to various parameters of risk. It is normal for the Locks and sluices at Teddington to be closed around high water to prevent tidal insurge, then opened with the falling tide to rebalance river levels. It is also usual, at times of high rain levels, to keep the sluice gates open for longer, to prevent upstream flooding.

    In 1968, large areas of the Thames Valley above Teddington suffered extensive flooding. In response, a complex and expensive series of flood protection schemes were put in place, for example on the Rover Mole, a major tributary. A few years back, parts of Berkshire, further upstream, suffered similar flooding, and work has gone on since to prevent repetition.

    This is the first Winter I know of in recent years where the River Mole flood defences failed, and some of the most expensive and richest ‘suburbs’ of Greater London, such as Thames Ditton, have had their share of flooding from the river. Places like Guildford, Woking and Hampton Court are amongst the best protected and managed systems on the Thames, but they have not escaped damaging floods.

    It might be more revealing to look at the records of the barrier along with the Teddington flow rates/sluice gate records, to see whether there is any pattern in the degree of system pressure/risk. Observing at first hand, it would seem that the frequency of high-risk events, either from extreme rain events or tidal surges, has steadily increased in the past twenty years, and actual flooding upstream of London has increased, probably in part as a consequence of needing to keep the river levels higher at risk events, to prevent more damaging flooding downstream.

  9. 9
    Max Beran says:

    I might be able to help a little with Wili’s request for “management” information as I was involved in a 10-year review of the operation of the barrier which included a statistical assessment of joint tidal and fluvial flooding.

    This study (between Halcrow and the Institute of hydrology) was principally aimed at event probabilities though some atention to operational matters was unavoidable. Principal among the management issues that did impinge at that early time was the financial restrictions imposed by the Port of London Authority under which the barrier authority had to pay a hefty sum to PLA for each closure. Unlike today London was still qite an important port. Subsequently an easement to close the waterway was purchased so there is nowadays no marginal cost per closure.

    Another factor was the closure sequence which required a slow and steady progression of gate raising starting not long after the previous low tide. This put a large burden on the reliability of the surge forecast. The reason was to avoid the reflected wave which endangered sea defences seaward of the barrier. Experience has shown that a judicious pattern of closure can avoid or minimise this reflected wave so now the operators can “go to the wire” closing when the need becomes clearer (and of course forecast errors narrowed).

    Prior to the proof of the North Sea “ocean atmosphere” model the tidal forecast was based on a regression of Southend surge on Immingham and other northerly locations. This had quite a large error of estimate – order of 60 cm as I recall. It’s “simultaneity” with high astronimic tide was also quite uncertain. The operating rule – a matrix between forecast Southend level and Teddington river flow – also didn’t allow for a fraction of surges originating not from the north but coming up the Channel.

    Despite what is often written, there was no recognition of AGW-driven sea level rise. Nowhere in Horner’s (the GLC design engineer published in Proc ICE) papers or calcs is this considered. The safety margin was based on standard freeboard practise for high value sea defence works and a linear extrapolation of past high water levels (at Tower bridge I think). My memory is inclined to play tricks but I fancy the sum total of the two was 80 cm above the 1953 level for a century-ish design life. Storage upstream of the barrier means that the fluvial contribution is almost below the level of attention. This is not the case as one moves up the tidal reach towards Teddington. We reviewed the trend at the time of the study and one could equally have described a quadratic with reducing high water levels into the near future. I note that there is little net movement once the land level change is factored in even now. It should be noted that the surge wave is inversely related to water depth so this will offset sea level rise to some degree (possibly small). Positive fluvial flood trends have generally not been found either in terms of magnitude, frequency or volume despite the vast number of assertions to the contrary.

    I hope these recollections are some help.

  10. 10
    Max Beran says:

    Dave (response No 7) is more nearly correct than either Gavin’s reaction to it or Fergus Brown’s input (which lacks hydrological understanding I’m afraid).

    The calculation of flood risk in an estuary requires one to picture a bivariate distribution of tide+surge in the open sea on one axis and fluvial flow entering the estuary on the other axis. On this same piece of graph paper, and for any cross section down the length of the estuary, one can describe lines of equal water level as a function of those two variables. This is termed the “structure function” and can be thought of as contour map of the two variables that combine to yield a given water level.

    The probability of exceeding that given water level then involves integrating under the bivariate probability mound beyond the structure function contour for that required water level.

    For locations close to the tidal limit the structure function is heavily weighted to the river flow – in other words water level is a function of river flow with only a small adjustment for tide+surge. For points low down the estuary, eg Southend, the structure function is entirely a function of the tide + surge – Southend water level is entirely a function of tide+surge with no effect of river flow. For mid-points both terms enter into the function with appropriate weights.

    At the barrier site, down at Woolwich, the function is very much dominated by tide which is why I said that the fluvial component can more or less be ignored. It’s a lot more like Southend than it is like Teddington. A back of envelope calculation could show this comparing a typical flood hydrograph peaking at, say 2,000 cu m/sec with the available storage above low tide level in the tidal reach (3000 m long, 200 m wide, 2 m vertically). Alternatively it is apparent from direction of flow – down at Woolwich one would never not see water moving upstream on a rising tide no matter how much water was entering the tidal reach from Teddington or the other tributaries entering the tidal Thames.

    Fergus appears to be neglecting Conservation of Mass in his notion of how weirs and sluice gates would influence the flux into the tidal reach. Thames weirs and sluices are there to maintain navigable depth at times of low and medium flow; their impact on the flood regime can be mostly forgotten about. The storage upstream is negligible in relation to the amount of water flowing through so it’s not like a flood detention reservoir built to moderate peak discharge. Nor does Teddington Weir have a serious role as a tidal control structure as it is close to the river’s natural tidal limit under most conditions.

  11. 11
    Barry Wells says:

    Dear Sir
    Re your article on the recent water levels in the Thames river and the sequence OF operating the Thames Barrier in recent years. Firstly please see below information as to the relationship of the South England land mass and the sea level due to rebound from the last ice age copied below,

    The University of Durham looked at levels of land uplift and subsidence in the British isles since the Ice Age. As the ice retreated 20,000 years ago the release of the enormous weight meant the north slowly tilted up while the south sank down. Scotland is still experiencing this “springboard” effect while southern Ireland, Wales and England continue to sink.

    The new study shows that land levels could rise by up to 10cm in some areas of Scotland over the next century, offsetting the effects of sea level rise caused by global warming. But in parts of England, where the land is set to sink by up to 5cm over the next century, it could add between 10 to 33 per cent on sea level rises.
    The map is the most accurate projection of land subsidence in the UK ever compiled.

    The Durham team not only looked a “geophyscial” simulations, which predict what will happen to the earth’s crust over time, but studied sediments in the soil at 80 sites around the country to see how the land has been changing in the past.

    Prof ian Shennan, who led the study, said soil sediments showed that sites in the north of the country are still rising.
    “The action of the Ice Age on our landmass has been like squeezing a sponge which eventually regains its shape. The earth’s crust has reacted over thousands of years and is continuing to react,” he said.

    Added to this is the fact that the M4 motorway corridor that runs parallel to the Thames valley has produced vast amounts of both residential and industrial building in the catchment area of Thames vastly increasing the run off.

    Then there are the changes in farming practices, use of heavy machinery compacting the land, the removal of thousands of miles of hedges and field ditches removing run off storage, change of crops increases in oilseed and linseed rape crops for bio fuels these do not hold back the rain water.

    If you took the trouble to read the old chronicles and country books of years gone by, Issac Walton the complete angle is good example you could not overlook the fact that the rivers then took 7 days to rise into flood and 7 days to fall now they rise into flood in hours not days and they fall just as fast, typical of faster rain water runoff rather than higher volumes of rainfall.

    It’s not co2 nor climate change it’s just more and more and more people and their required infrastructure coupled to some bad farming measurers, you need look no farther than this.

    Barry Wells

  12. 12
    wili says:

    Thanks, all, for insights. I’m humbled and awed by the range of experience and expertise of posters here.

    There do seem to be quite a number of issues that affect how frequently these gates are used beside sea level, so tracking that frequency would seem to be a pretty poor proxy for determining even local SLR.

  13. 13

    Gavin – thanks for the update.

    Small typo:

    local subsidence issues which will increase flooding risks but is not climatic in origin

    should be

    issues … are not climatic in origin

  14. 14

    Brisbane is another city where a combination of flooding and tides can be a problem, compounded by the fact that the city is built on a flood plain. The Wivenhoe Dam was built upstream of the city to contain floods on the principle that the top 50% of the dam capacity is for flood mitigation and ideally, the excess can be let out at low tide.

    In January 2011, despite the dam, Brisbane experienced one of the worst floods in a century, only a bit over a metre less than the last pre-dam flood in 1974.

    I am not going to claim a climate change effect for a once-off event though any shift to bigger floods or bigger storm surges and worse, a combination of the two, puts a city like Brisbane at risk. Fortunately, the dominant political trend in Australia is climate change denial, so that solves that one neatly.

  15. 15
    wili says:

    Here’s something on the general issue of flood prevention, working with nature upstream. This is in Yorkshire. Is any of this happening in the Thames watershed?

  16. 16
  17. 17
    Fergus Brown says:

    Max Beran at ~10 you are quite correct to say that I lack your level of hydrological understanding. However, I think we are talking at cross-purposes, inasmuch as I was focusing on the impact upstream of Teddington, rather than the storm surges and primary risks which are addressed by the Barrier. This is no doubt a consequence of my poor expression, for which…

    The point I was trying to get across is that there is (please correct if this is wrong) a link between the tidal state at Teddington and the use of the weirs/sluices to alleviate upstream flow. There may or may not be a correlation between the number of incidents of upstream (non-tidal) flooding in the Thames Valley, and the number of occasions when, during high-rain events, the weirs/sluices are nonetheless closed. There might have been a measurable connection between this and the closure of the barrier, but I accept your explanation and recognise that there cannot be, since the tidal Thames has the capacity to manage fluvial input without increasing flood risk in London itself.

    What remains is that, whilst London has to date remained reasonably dry, Berkshire, Middlesex and Surrey have experienced an increase in the incidence of flooding, cuased by excess rainfall, and possibly affected by the weir/sluice management process.
    This is a bit off-subject, but at the time of the original comment, I thought that some people would find it useful to consider other weather/climate related factors along with the Thames Barrier.

  18. 18
    Steve Fish says:

    This may be a trivial request regarding this topic, but could someone knowledgeable please elaborate on what is shown in the striking photograph that accompanies Gavin’s original topic post. For example, the white structures appear to be much more artful (expensive) than what might be required for a dynamic water barrier.


  19. 19
    Phil. says:

    That’s the barrage, here’s an annotated photo.

  20. 20
    Steve Fish says:

    Re- Comment by Phil. — 17 Jan 2014 @ 8:36 PM

    Thank you for the picture. The only experience I have with flood walls is via comments of a friend living along the Ohio River. Some time ago there was a serious flood warning and nobody knew how to set up the flood wall. They contacted a retired county employee who knew where the large slabs were stored that had to be placed into slots in adjacent towers with a forklift. This was obviously an old system.


  21. 21
    ozajh says:

    Phillip @ 14,

    I can see a serious future problem with that philosophy for the Wivenhoe dam, unless there are features of the underlying geography which allow a LARGE outflow without harm. If the population of Greater Brisbane continues to increase, there will eventually be political pressure to use the entire capacity of the dam for water capture (especially if the drought/flood cycle intensifies).

    Then you get a big rainfall event when the dam is already full . . .

  22. 22
    Max Beran says:

    Fergus at 17: I am confused by your words: “alleviate upstream flow”. If you refer to the discharge from the Thames basin during a moderate to severe flood then the presence of an minor obstruction in the watercourse like a Thames Weir, even one as long as Teddington, is neither here nor there. Outflow has to equal inflow as the change in storage term is minor in relation to these flows. So the discharge (measured in cu m/sec) is unaffected by whether Teddington Weir is open or closed.

    You may be referring to water level rather than discharge. There is a complicated relationship between water level and discharge and the former adjusts itself to the latter mostly by increasing the cross sectional area – increasing the water level and overflowing the flood plain. Gradient also has a secondary effect which in the case of a structure like weirs and sluices equates to the head difference between up and downstream water level. Opening all the sluice gates at Teddington has the effect of increasing the cross sectional area at that point so the upstream water level would be lower than if the sluices remained shut. However all gates would be open during a river flood so this is not a variable term in thinking about what happens during a river flood.

    Turning now to the effect of a high tide, as I said Teddington is the normal tidal limit (ignoring Richmond Lock which is for low flow control only). Having said that I have seen hydrographs (graph of water level versus time) which exhibit a minor tidal pulse towards Molesey – the next weir upstream, about 5 km above Teddington. This is not salt water moving upstream across the weir but is there because the continuity equation of river flow means that the river will try to maintain the head difference across the weir (ie the gradient effect I mentioned above). This raised level above Teddington Weir transmits itself upstream towards Molesey via the “backwater curve”. However this will be a minor effect during river floods as cross-sectional area dominates.

    Much of the above is related to the small scale of the Thames – catchment area 10,000 sq km, floods of 2,000 cumecs etc. The Brisbane River is a much larger estuary and also Wivenhoe Dam has a storage capacity that is meaningful in terms of flood volumes (though this capacity is shared with water supply and recreation) and there is a substantial catchment area below the dam whose runoff the estuary still has to cope with. The storage volume of reaches of the Thames between weirs is trivial in comparison to its floods.

  23. 23
    Max Beran says:

    I feel Barry (at No 11) overstates the effect of urbanisation on Thames floods with his “vast” and “vastly”. An easterly tour via Google Earth to Oxford and beyond would show the Thames to be still essentially a rural catchment. The view from transport arteries close to London might give a different impression but statistically the Thames at Teddington would not fall in the category where one would need to adjust, say, the 100 year return period flood to account for urbanisation.

    A principal hydrological effect of urbanisation is on the speed of response to rainfall which can of course even be helpful if the contribution of those tributaries is over and done with by the time the main flood wave from the bulk of the catchment arrives.

    The impact of other land use changes on flood magnitude and frequency is not as clear cut either as implied.

    I do not refer to field ditches or small brooks at the outfall of housing developments, but catchment areas of the size flowing into an estuary. I am also not at all aware of small trout streams of the sort that Isaak Walton fished like the Dove having characteristic times to rise of 7 days and to fall another 7. Perhaps he was talking about the Trent, or days is a misprint for hours. Something very weird must have taken place with the geology for such a change in response characteristics.

  24. 24
    Max Beran says:

    oops sorry – “westerly” tour

  25. 25
    John Benton says:

    Excellent comment Barry Wells @ 11.

    As a property developer in that area I can confirm your assessment is exactly correct, however you omitted one of the most important factors which has led to increased flooding, not only in this area, but in many others too.

    Over the last 20 – 30 years the river management practices have changed radically, most particularly since the present Environment Agency was formed. Previously many rivers were dredged to ensure free flow in times of heavy rainfall but much of that work has now stopped despite the entirely predictable consequences we are now seeing.

    In fact the Environment Agency is the cause of much of the river flooding in England much to the ire of landowners and householders. Their incompetence is now so well known there is even a website called Inside the Environment Agency where staff, and former staff, chronicle their experiences.

  26. 26
    Fergus Brown says:

    Max at #22 ; You are correct, I should have referred to level rather than flow.
    I believe our correspondence here has illustrated well the difference between a ‘lay’ understanding, partially based on experience and some research, and an ‘expert’ understanding. The former is comprehensively trumped by the latter – an object lesson for those wishing to participate in climate discussions.
    Thank you for your detailed and eloquent responses – I believe I am now better informed than before and trust this is also true for other readers.

    At the risk of diverting the discussion further from the original subject, you would appear to be well-placed to comment on the incidence of flooding in the Thames Valley; is there evidence of an increase in frequency, and if so, is there evidence of climatic changes which would explain this?

  27. 27
    Barry Wells says:

    Re Max at 23 please see below some excerpts from documents that will explain some parts of my post the first comes from a World Wildlife Fund report on the Thames Basin.

    WWF report on the Thames Basin Vulnerability Report Technical Summary

    Nearly a quarter of the population of England and Wales (around 14 million people) lives and works in this region, producing more than a quarter of the Gross National Product.

    The region has some areas that are particularly heavily urbanised and densely populated including London, Oxford, Swindon, Reading, Luton, Stevenage, Guildford and Crawley.

    Trends suggest that the economy in the Thames region will continue to increase in importance both within and outside its boundaries. The number of people living within the region will increase, putting even more pressure on resources, particularly land and water.
    It is estimated that around 56,000 new homes will be built in the Thames River Basin District every year for the next 15 years and that population may increase by around 2 million by 2026 (Environment Agency, 2007). A large percentage of the new development is planned in the lower Lee valley and in the tidal floodplain of the River Thames.

    The second excerpt comes from a consultant report carried for three London councils:-

    JMP Consultants Limited – Flood Analysis Report Final on the 2000 – 2003 flooding

    JMP Consultants Ltd have been commissioned by Spelthorne, Runnymede and
    Elmbridge Borough Councils to undertake an independent investigation into
    the New Year 2003 flooding. The aim of this report is to identify why the
    Borough’s have been affected by flooding twice in 3 years (autumn 2000 and
    January 2003), the extent and source of the recent flooding and measures
    which can be taken to limit the damage caused by future floods.

    2.36 The magnitude of the November 2000 and January 2003 floods is similar at
    Kingston. However, there are some differences at Windsor and Staines. It is not
    valid to precisely compare the flows from the two floods because they were
    caused by different rainfall conditions, which will result in a different river
    response. As previously mentioned, the flow at Kingston is likely to be
    influenced by tributaries which may assist in making the peak flows from
    December 2000 and January 2003 similar.

    2.37 As shown by the last column in Table 2.3 the average rate of rise of the River
    Thames at Windsor, Staines and Kingston is significantly higher in January 2003
    than in December 2000. This mirrors residents’ descriptions of the rate of rise
    being much faster than they experienced in December 2000. At Windsor, the
    rate of increase increased by approximately 30%. At Staines, the rate of
    increase approximately doubled.

    2.38 JMP Consultants consider that the Jubilee River is the major contributing factor responsible for this increase. Prior to construction of the Jubilee River, flooding upstream would have provided attenuation of the flood flows and slowed the rise of the River Thames. The new channel limits flooding upstream and effectively passes the flood flows downstream un-attenuated, so causing the river to rise more quickly. Further studies, which are outside the scope of this report, would be necessary to accurately assess the scale of this effect.

    And finally to show that floods within the UK are nothing new and are nothing as compared to the event described below:-

    River Parrett Somerset
    The Levels and Moors are a largely flat area in which there are some slightly raised parts, called “burtles”[156] as well as higher ridges and hills. It is an agricultural region typically with open fields of permanent grass, surrounded by ditches lined with willow trees. Access to the Levels and Moors is by “droves”, i.e. green lanes. The Levels are a coastal sand and clay barrier about 20 feet (6 m) above mean sea level (roughly west of the M5 motorway) whereas the inland Moors can be 20 feet (6 m) below peak tides and have large areas of peat. The geology of the area is that of two basins mainly surrounded by hills, the runoff from which forms rivers that originally meandered across the plain but have now been controlled by embanking and clyces. The area is prone to winter floods of fresh water and occasional salt water inundations which have occurred, the worst of which in recorded history was the Bristol Channel floods of 1607, which resulted in the drowning of an estimated 2,000 or more people, with houses and villages swept away, an estimated 200 square miles (520 km2) of farmland inundated and livestock destroyed.[157] A further severe flood occurred in 1872–1873 when over 107 square miles (277 km2) were under water from October to March.[158]


    Barry Wells

  28. 28
    Max Beran says:

    Barry: You seem to have lost track that this thread is concerned with what determines the frequency of closure of the Thames Barrier whose operating rule in turn depends on the fluvial inflow to the tidal reach at Teddington and the tide+surge at Southend. I can really only repeat my suggestion that you take a Google Earth tour of the catchment area upstream of Teddington. Imagine you were parachuted at a random point in this area – what do you think is the chance that you would land in a town? The ones you mention of Reading, Oxford and Swindon are the only ones whose populations top 100,000. I happen to live close to the centroid of these three and can assure you that the view from my home is entirely rural and villagey.

    I am interested in your remark about an apparent speeding up of rise time at a lower Thames point but would need to see the way this was evaluated. A river’s temporal response to an individual rainfall event is not purely a property of the land cover of the catchment area as it varies with the temporal and spatial variation of rainfall input. Rain moving down the catchment is a recipe for faster time of rise as would the balance netween rain falling on the more impermeable soil types compared with that over the chalk and limestone areas. So one would have to see the rainfall “hyetographs” for the major subcatchments and the lagtime between their centroids and the tributary flood hydrographs to get a real appreciation of how two events differ.

    I am very familiar with the Somerset Levels but do not understand why you mention them in this context. If you are looking for historic floods then a search of the British Hydrological Society’s “Chronology of Events” could be rewarding.

  29. 29
    Max Beran says:

    Fergus asks (at 26), “is there evidence of an increase in frequency, and if so, is there evidence of climatic changes which would explain this?”

    I am long-since retired but my impression is that the answer is “no” to the first. The person to ask is Terry Marsh of the “Centre for Ecology and Hydrology” at Wallingford. See the title of the first publication of the list here: . His experience is shared by the vast majority of hydrological researchers looking for evidence of climatically driven change in flood time series. This of course is in accord with IPCC’s circumspection on the extreme events issue in general but entirely at odds with almost every lobbyist and media reporter.

  30. 30
    Max Beran says:

    John (at 25): The clue is in the name – Environment Agency. You probably should not be seeking symapthy here for your view; after all saving the planet and saving the riverine environment would tend to appeal to the same constituency.

    In olden times it was a job given to the village idiot to slub out ditches and road and trackside gulleys as “jobfare” and keep them off the poor rate. Ditto a river lengthman who, still at the parish level, would remove the worst obstacles and improve access to the local watercourse.

    Large scale dragline excavation would never have been more than at known pinch points or river reaches where sediment was known to be deposited. This was always going to be a “tail chasing” exercise as increasing the cross section would lower velocity and increase fresh sediment build-up. A natural river’s sediment load, cross section and spectrum of flows would be in balance. The problems arose where that balance – which would in general involve meanders, pools, riffles, sandbanks – was inconvenient from the point of view of users of the river such as fishing, navigation, or water power.

    Schemes such as the Jubilee River mentioned by Barry, though perhaps billed as part of a holistic enterprise, would be financially justified on the basis of known local flooding issues in built-up areas like Maidenhead. Their purpose is to increase the “conveyance” which implies lowering water levels for any given discharge. This again would be quite local and may be to protect existing properties or allow new development to take place. Their effect on the river regime would be apparent at moderate floods but would largely disappear as you get to the real biggies.

  31. 31
    Hasis says:

    “This of course is in accord with IPCC’s circumspection on the extreme events issue in general but entirely at odds with almost every lobbyist and media reporter.”

    And just to further drive that point home:

  32. 32
    Max Beran says:

    Thanks Hasis (at 31) for that reminder of the precise wording. On reflection I would change circumspect to circumlocutory. A commonsense summarising of the weight of evidence would be more along the lines of high confidence of no link, rather than low confidence of a link. Saying low confidence is still an incitement to someone of the risk-averse-raised-to-the-power-of-risk-averse frame of mind to indulge in misdirected responses.

    I realise of course that such a revision is out of synch with the founding IPCC terms of reference which was to assess the evidence for (man-made) climate change. Under that paradigm (pace Prosecutor Fallacy) it makes no sense to express a probability conditional on no change.

  33. 33
    Max Beran says:

    I thought I’d put in my two penn’orth on the factors diagram of the main article. The deficiencies of the data for trend analysis are clearly stated in some places but perhaps not considered in other interventions.

    So to hammer the point home, all three factors – surge residual, astronomic tide and Teddington flow – are time series but the manner in which the subsets of the individual factors that are depicted on the diagram have been extracted from the basic time series are very far from independent and so fail the key requirement for drawing inferences about time trend.

    The points that are plotted on the diagram ultimately arise from occasions when someone in the Thames Barrier control office decided that the event looked like it would satisfy the operating rule for the gates. This of course is not entirely random – in my day it comprised a graph of tide+surge on one axis and Teddington flow on the other. If the event fell in the upper right above a marked threshold curve then the gates were raised. But as stated above (at 9) the decision was metaphorically muddied by (a) operating and financial considerations and (b) the ocean component of the decision being based on an imperfect forecast well ahead of the actual event. There are no “tidal events” of “fluvial events” as such as it is (or was) a continuous curve partitioning the decision space.

    I would judge that the tidal data are “cleaner” as it is this component that dominates down at Woolwich. But one would need more information on the numbers; for example is the surge the maximum residual or the residual that corresponded to the high tide or that which combined to give the maximum tide+surge water level.

    A more relevant source of flow data than the one linked to in the article is: . Click on POT CSV files and download the Okay for Pooling set. Scroll down to 39001 which is the ultrasonic station at Kingston, but the pre 70’s data are for Teddington Weir. There is a small catchment area difference but the main cause of non-stationarity is that the ultrasonic gauging station measures instantaneous discharge so potentially picks up the true peak. The older data are for daily mean discharge, hence a little lower. Metadata are available on the website and as comments in the time series.

  34. 34
    Max Beran says:

    Tom Mallard (at No 4) refers to sea level trends. To the extent that a part is thermosteric, does not this mean that that part can be forgotten about at a coastal or estuary location. The coefficient of linear expansion of sea water has a lot of zero’s after the decimal point which multiply out to almost nothing when applied to single digit metres of water depth.

    It’s not as if there is a mass of water at greater elevation waiting to sweep ashore because presumably density, hence weight adjusts, and hence a lateral gravitationally driven force is not present. I of course realise that there are many other lateral forces and spatial variations that operate in the ocean but why would they interact with the expansion effect to smoothen it horizontally? Arguing from a volume expansion coefficient does not help as one would still expect the local expansion to relate to available linear dimensions within the volume.

    Can anyone advise? Does this mean that a coastal tide gauge also lacks any meaningful thermal influence? Have there been reported trend differences between ocean buoys and coastal locations?

  35. 35
    Dave Peters says:

    Yesterday, out of the blue, I flashed on the concept of pumping seawater over some low coastal range in Antarctica, and posed a query to the open thread (# 626), asking if anyone had heard of this geoengineering remedy to global flooding. I had a figure of 64 cubic kilometers floating in my head from eight years ago, as the Greenland annual melt taken from that sweet trick of the mass-measuring Mutt & Jeff twin satellites, and a companion notion that perhaps 30% of ice loss was from West Antarctica. That would make for about 100 km^3.

    Hoover dam (wiki) flows an average of 637 m^3/s, or ~20 km^3/yr (60 x 60 x 24 x 365 x 637 = 20.09 billion m^3/yr). It also generates an average of 4.2 TWh/yr., from a hydraulic head of about 500 feet. That figure, divided by 8,760 hours per annum, yields a required capacity of 479 MWe, or ~1/2 gigawatt, through a frictionless pipe network. You would only need a single twin 1.3 GWe modern nuclear plant for my original notion, or terra-forming on the cheap. But, the CU GMSL rate is 3.2 mm/yr., and wiki’s ocean area of 361 million km^2 times that yields an order of magnitude greater figure of 1155 cubic kilometers. That roughs out at 11 twin modern nuclear generating stations.

    Not advocating any such thing, mind you, as it would enable both an oceanic hecatomb and turning the land surface hotter than Furnace Creek. Still, it might be cheaper than moving Miami Beach or building higher Thames barriers.

  36. 36
    jamie taylor says:

    I just want to know why doesn’t the government just admit that the Thames barrier is in the wrong place and that a new barrier need’s to be closer to the mouth of the river to stop the tidal surge from coming into the Thames in the first place

  37. 37
  38. 38
  39. 39
    Alistair Connor says:

    Postscript, for the information of Max Beran, John Benton and others who may be sceptical of a global-warming influence on the current flood season. Concensus is indeed that it is too early to allege firm attribution. My guess is that by the time such an unprecedented series of events repeats (say, within the next five years…) then attribution to global warming will be unequivocal.

    A Global Perspective on the Recent Storms and Floods in the UK – Met Office

    This series of winter storms has been exceptional in its duration, and has led to the wettest December to January period in the UK since records began. Heavy rains combined with strong winds and high waves led to widespread flooding and coastal damage, causing significant disruption to individuals, businesses and infrastructure.

    The severe weather in the UK coincided with exceptionally cold weather in Canada and the USA. These extreme weather events on both sides of the Atlantic were linked to a persistent pattern of perturbations to the jet stream, over the Pacific Ocean and North America.

    The major changes in the Pacific jet stream were driven by a persistent pattern of enhanced rainfall over Indonesia and the tropical West Pacific associated with higher than normal ocean temperatures in that region. The North Atlantic jet stream has also been unusually strong; this can be linked to exceptional wind patterns in the stratosphere with a very intense polar vortex.

    As yet, there is no definitive answer on the possible contribution of climate change to the recent storminess, rainfall amounts and the consequent flooding. This is in part due to the highly variable nature of UK weather and climate.

    Nevertheless, recent studies have suggested an increase in the intensity of Atlantic storms that take a more southerly track, typical of this winter’s extreme weather. There is also an increasing body of evidence that shows that extreme daily rainfall rates are becoming more intense, and that the rate of increase is consistent with what is expected from the fundamental physics of a warming world.

    More research is urgently needed to deliver robust detection of changes in storminess and daily/hourly rain rates and this is an area of active research in the Met Office. The attribution of these changes to anthropogenic global warming requires climate models of sufficient resolution to capture storms and their associated rainfall. Such models are now becoming available and should be deployed as soon as possible to provide a solid evidence base for future investments in flood and coastal defences.

  40. 40

    #39–Persistence seems to be the defining characteristic of this winter in many places. Persistent storminess in the UK, persistent cold over much of North America (or, in the case of my region, the Southeast, persistent alternations between unusual cold and seasonal-to-warmer than seasonal conditions–just two weeks to the day after ‘Snowmaggedon’ shutting down Atlanta, everything is closed once again due to anticipated icing), persistent drought in California, persistent warmth in eastern Australia. Oh, and persistent warmth in the Arctic. How could I forget that one?