This next graph shows the live situation at the barrier for today, and the 48 hours before today. If the Upstream side is higher then the drop is shown by the vertical thickness of the Green. Otherwise if the Downstream side is higher then the drop is shown in Red. Closures are shown by the thickening of the graph.
Where there is neither red nor green this indicates data not available.

And this graph shows the situation at Southend and some of the facts on which barrier closures are decided.

So given the expected height of the next High Water at Southend (see top right of top graphic above) what level will trigger a barrier closure? That depends on the river flow. The chart below shows all the flood defence closures before 2009 plotted against the river flow. The low Teddington flow (exceeded 95% of the time) is about 7 cumecs. The average is 66 cumecs and the high flow, exceeded 10% of the time is 161 cumecs. The average river flow during a closure is 248 cumecs - so the majority of closures occur during high flow conditions. The scatter graph reveals what one might expect that on the left, in times of low river flow, the expected height at Southend which triggers Barrier Closure is higher than on the right. I am not clear whether this could be reduced to a straight line or a curve. If a straight line it seems we might start with 3.70m on the left, going down to 2.80m on the right. Notice that this implies that at a river flow of 280 cumecs closure would be triggered at about the maximum spring tide level without any surge. At the maximum recorded river flow (in the 800s?) most tides would require barrier closure.

Anway plot the next high water on this chart and if it falls amongst, or above, the previous points then a closure is likely. I can't at the moment give you the Kingston flow - but am working on that. In normal times (and barrier closures tend to come at abnormal times) the Maidenhead flow below will not be widely out. Uncertainities in the surge forecast will tend to make closure more likely. So this graph is a "Beta" version and the vertical (river flow) line is not reliable -

If the graphic shows errors (and it has been known!) check the PLA site (selecting a gauge). The PLA gauges sometimes appear to stick or show 0 (Chart Datum) or go completely haywire! An occassional problem in the past has been over zealous error correction which stopped a measurement because the closure of the barrier had created a level out of the normal run of things.

How to read the top online tidal graphic

1. Look at the margin figure below Southend: Red, positive would be a forecast of barrier closure, maybe.
2. Look at the margin figure below Tower: Red, positive would be a forecast of barrier closure, maybe.
3. Look at the red continuous line above Southend. That is the next predicted high tide. (Time in the Tide#1 or Tide#2 row below).
4. If the red continuous line is near the SPRING HW mark then this is a SPRING TIDE (large)
5. If the red continuous line is near the NEAP HW mark then this is a NEAP TIDE (small)
6. The yellow line is the predicted height at the moment.
7. The blue of the measured height may be above or below the predicted height. This shows the SURGE now.
8. The green line shows the height of the next Low Water

In the top graphic you are looking from the south at a notional cross section through the water of the Thames Estuary from Teddington on the left, to Walton on the Estuary on the right. The heights are all in ODN (Ordnance Datum Newlyn) to make them comparable (the CD Chart Datum heights are given in the first row below the graphic.)
The schema of the barrier is shown as if from above, with predicted short or long term closures and full test closures, and in addition if the drop across the barrier indicates closure you will not miss the scarlet colour!
ALL INDICATIONS ABOUT THE BARRIER STATE ARE DEDUCTIONS FROM THE HEIGHTS PROVIDED BY THE PLA COMBINED WITH PLA DATA ABOUT TEST CLOSURES. I have no other way to know.

The figures below are worth coming to terms with:

The Proudman Oceanographic Laboratory allow access to graphed current surge forecasts. You will need to register to see them.

On 28th February 2010 started a run of five closures on successive tides -

Five closures on successive tides

The record of consecutive (or near consecutive) closures is:
3 on 19th & 20th February 1996
5 on 25th - 27th December 1999
7 on 8th - 12th February 2001
4 on 10th - 14th March 2001
14 on 1st - 8th January 2003
5 on 21st - 23rd January 2003
3 on 18th - 22nd January 2007
4 on 18th - 20th March 2007
5 on 28th Feb - 2nd March 2010

Here is a snapshot of the flood closure on 10th February 2009. A deep depression had come up the channel and resulted in a backdoor surge: -

The barrier shut with a drop of 3m.

This is what a full closure would look like, based on the test closure on 4th October 2009. There was a considerable surge, but the river flow was very low. The upstream rise during the closure was estimated at 27cm per hour. After closure there is still a large body of water which has been displaced upstream by the tide and now needs to find its equilibrium, so it will "ebb" - ie move downstream, piling up against the barrier even though its lifeline - the tide - has been shut off. The actual river flow would not affect this very much. The graph comprises vertical lines from the higher level to the lower level at any moment (in green when upstream is higher and red when downstream):

Here is the small test closure on 7th September 2009. It was started a few minutes before low tide at the barrier and so was a 'reverse' closure with the level upstream higher than downstream. This is done so that the barrier machinery can be exercised without too great a disruption to the river. The opening needs to be at a moment when the sides are level to avoid the erosion which might happen as a result of "underspilling" to equalize the levels.

It is fascinating to watch the tide measurements as the high [or low] tide point reaches first Walton and then slowly spreads west up the estuary. Of course tidal water in an estuary flows uphill! It does so because it is moving and its momentum carries it up (not because it is driven from behind!).

In the project pack are explanations of all the closures since the barrier was built.
The barrier closure also depends on river flow - but disappointingly there is no online measure of that. Clues can be gathered from the Teddington Lock site and Thames flood watch
Thames Barrier, (Environment Agency leaflet).
Here are the best pressure forecast maps that I can find (up to 132 hours ahead). The nightmare scenario as I understand it is a low going north west up the Atlantic, then north of Scotland and then down the North Sea with a strong north wind.
 
The disastrous 1953 flood was caused by a surge of 2.59 metres.
The highest recorded surge was 3.66 metres.
But surges tend to peak four hours before high water, though this is not always the case.
 
Animated graph of the North Sea tides today (Week starting last night).

THAMES ESTUARY 2100 CONSULTATION (TE2100)

The Thames Barrier closures are shown here.

PLA Notices to Mariners will have the Planned closures [generally once a month] These are shown in the graphic above.

To find the time differences for places where tidal information is not given above see here

The North Sea is not big enough to have tides!
I am aware that this statement may come as something of a shock to those who live beside it. But the fact remains that taken on its own it would have almost no tide generated astronomically.
Of course it does have tides because it is coupled, both to the north and south, with the much bigger Atlantic which does have astronomic tides. This coupling excites resonances within the North Sea. They centre around three points where the tide is essentially zero, and they swirl around these points in a counterclockwise rotation.

The tide comes round the north of Scotland from the Atlantic and sweeps down the east coast so that the tide at Felixstowe is more or less in the same state as the tide at Wick. The tide in the northern part of the north sea rotates around a point on the southernmost tip of Norway. In the centre of the North sea it rotates around a point half way across about level with Immingham. And in the southern part of the North Sea it rotates around a point about half way across level with Lowestoft. In the Southern rotation as the high water sweeps past the Thames Estuary and on across to the Belgian coast it is reinforced by the high water coming up the channel from the west.

The distance between high and low water at any point is roughly proportional to the distance from the amphydromic point around which it is rotating. As a rule of thumb the spring tide range at Wick is around 3 metres. It increases as one comes south down the east coast to a maximum of about 6 metres at Immingham. It then reduces to a minimum of around 2 metres at Lowestoft and increases again to 5 metres at Sheerness (and 6.5 metres at Richmond).

One of the striking points to be noticed is that in the northern rotation of the North Sea the tide acts as if it does actual rotate even though one half of the rotation is on land. The same thing is to be observed in the centre of the south coast where there is an amphydromic point snugly situated in the middle of the New Forest!

Thames Tides - a description from 'The Geography of London River' by LL Rodwell Jones (1931) -

Co-tidal Tide and Range Lines in the Southern North Sea, Jones, 1931

The tides of north-west Europe are of the semi-diurnal type with nearly equal semi-diurnal amplitude.
[The figure above] taken from the North Sea Tide Chart shows co-tidal time and range lines. It indicates an amphidromic point (point at which the lines meet) in the southern North Sea about the latitude of Lowestoft and nearly midway between the English and Dutch coasts.
The co-tidal time lines show that the tidal pulse, involving a considerable tidal current, moves from north to south on the English side, and that the heaped up water to the south of Dover then passes north eastwards along the Belgian and Holland coasts.
The pulse completes the circle about the amphidromic point each twelve hours.
There is little tidal movement at that point, and the range of tide increases in all directions as shown by the almost circular range lines from that point.
An average tidal range line of 12 ft. runs from North Foreland north-westward, but this range is, of course, magnified in the estuary.
While still calling attention to this diagram we should notice for future reference that the time lines suggest that in the outer estuary High Water occurs in the northern portion (off Essex) some minutes before its occurence in the southern portion (off Kent).

Thus the tidal stream flows from north to south along the Essex coast and propagates a branch up the Thames estuary. The kinetic energy of this vast westward moving column of tidal water is dissipated in overcoming friction at the hydraulic surface, in holding up the upland waters, and, most importantly, in raising the water itself. As the estuary narrows a portion of the energy becomes concentrated upon a gradually decreasing cross-section, and the water is raised to higher and higher levels as one proceeds upstream.

[The figure below] represents the tides of a lunation as registered during a period of little meteorological interference at the Tilbury Pier Head gauge.
Since High Waters are later each day it is possible to record the fortnight's tides on a 24-hour clock drum.
The diagram is self-explanatory. The change in amplitude from springs to neaps is well shown, and the symmetry of the semi-diurnal type.

A Tidal Gauge Automatic Record for a Lunation
Observations taken at Tilbury Pier Head for the period 8 to 23 August 1928

1758: A Description of the River Thames &c -

The Tide ebbs and flows above 70 Miles up this River within the main Land, which is done twice in every 24 Hours; by which means all her channels are filled as often, to the great advantage of Trade and Navigation. Concerning which it is necessary to observe, that, as the Tide is influenced by the Increase and Decrease of the Moon, so the Tides differ in their Times, each one coming 24 Minutes later than the former, which wants but 12 minutes of a whole hour in 24. And therefore, they who have any Dependance on the Ebbing and Flowing of the River Thames, are regulated by such a Table as follows:

In the above table there is at least a mistake at 4 days after the New or Full Moon: instead of '5.52' it reads '6.52'
The 1758 account continues:

N.B. But after all, this Table only serves when the Tide is regular, and not interrupted by any Accidents; for if the Wind proves rough at West or South West, it is known to stop the flowing in of the Tide to its usual Height; and the boisterous North-East Wind has the contrary Effect. Another Accident is the overflowing of the banks of the Thames occassioned by great Rains, which being stopt in their Course to the Sea by the flowing Tide, must consequently make some Alteration in the height of the Water, of which there are several extraordinary Examples recorded, both in ancient and modern Histories *
[* See Maitland's History of London pp.49, 135, 145]

As to the Shifting or preternatural Tides, as some call them, they have either been of that little Consequence, as to deserve no Remark, or may be properly accounted for by what has already been observed concerning the Influence of the North-West Wind encountering a flow Ebb at the Thames's Mouth; which at least, for a certain space, must cause a return of the Tide.

But the most general Rule to know the Time of Tide at London Bridge is, that when the Moon is in the Full, or Changes, then it is High Water at or near Three o'Clock following; and it is likewise High Water there at Eight o'Clock, or within a few minutes, after the Moon enters its First or Last Quarters, and you are only to add --- Minutes to each 24 hours (or Days) if your Enquiry hapens after each said Quarters of the Moon, to the Hour here given.

In the original it simply says "add --- Minutes to each ...". I think this should probably read "add 48 minutes to each ..."
Note that in the above account the obvious Spring/Neap variation has been entirely ignored.

1911: Probably more than you wish to know about tides

source: Figure 11.5 of http://www.es.flinders.edu.au/~mattom/IntroOc/lecture11.html reproduced by permission

Matthias Tomczak's Oceanography Lecture Notes, Lecture 11 -

Tides in the North Sea as derived from observations. Red lines are co-phase lines of the M2 tide, labelled in hours after the moon's transit through the meridian of Greenwich. (There are thus only 25 minutes between the co-phase lines labelled 12 and 0.) Blue lines give the mean tidal range at spring tide (co-range lines of the sum of M2 and S2). The progress of the tidal wave from the Atlantic Ocean into the North Sea is clearly demonstrated by the co-phase lines. The wave enters from the north and propagates along the British coast; it then proceeds around two amphidromic points along the Dutch, German and Danish coastline. Another wave enters from the south west, through the English Channel.
The influence of the Coriolis force is demonstrated by the co-range lines, which show large tidal range along the British coast and small tidal range along the German, Danish and Norwegian coast. The same effect (amplification on the right side of the wave) is seen in the English Channel, where the tidal range along the French coast is as high as 11 m compared with 3 m on the English coast.

Quoted by A Tour on the Banks of the Thames from London to Oxford, in the Autumn of 1829 By A. Walton -

Oft as the changing moon the ocean wide
Impels, our Thames receives the changing tide ;
When in mid Heaven fair Cynthia glorious rides,
By her directed, onward rush the tides ;
When, on the other side, she wears in wane,
The tides, attendant, hasten back again,
By force acquired, the exulting river swell'd,
Rolls on, and cries "to me all rivers yield",
Save the twin-brother floods of Elbe and Scheld.
With such true tides no river can be found
In all the realms that Europe's empire bound.

Height profile of the Thames. If the horizontal and vertical scales were the same, the profile would be a horizontal line without a single pixcel variation. As it is the vertical scale is highly magnified.