If it's true groundwater, of course that's not closely linked to the river level. If there's even a small tributary stream of the Thames nearby, maybe it is.

If it is groundwater, there are things that can be done, and perhaps quite quickly. Essentially that means drilling holes near or under the track, and sucking the water out. That would probably need to use surrounding land, though.

The latest forecast for the Thames at Reading, including rain that's expected, is that it will not go higher than it is now. That's roughly where it was a month ago, and just below its 2003 peak.

The levels from Mapledurham to Sonning have not moved for 24 hours, either. Does that mean the level should peak soon downstream?

I saw a BR^▸ (British Rail(ways)) engineer on site being interviewed on TV. He said there was at least six weeks of work - but also that they would first need to survey it and plan the work. His initial estimate would also need to be replaced by one based on the actual plan. So it's no surprise that the total time to completion would be nine weeks, and even that may be subject to extension due to various things - not least the weather!

I don't think that's how seawalls work. Those spectacular pictures of waves showed water being thrown upwards, not coming horizontally over the top. The bulk of the wave hits the wall lower down, and can only go upwards - so it pushes any water heading to overtop the wall upwards. Some seawalls are curved at the base to encourage this, but all do it.

Of course the water that shoots upwards has to come down, and water is very heavy. So the top of the wall takes a pounding, but from all directions. The Dawlsih wall was made of quite small blocks of masonry, mortared together - and mortar is not as strong material. So the wall can be taken down a block at a time. Big single lumps, if heavy enough, resist that. They might move a millimetre or so, but that's OK.

I think NR^» (Network Rail - home page) must have spent a lot of time remortaring blocks into the top of the wall, both after storms and routinely. Then along comes a storm big and strong enough to take the wall right down before any repair is possible. Once the embankment can be washed out, there's nothing behind the wall pushing back and it just gets pushed over.

I guess that'll be the Loddon - nearly a foot higher than its highest measured level. It has cut all the roads from Wokingham into reading bar the motorway - again. And perhaps this time also the road via Twyford, though as Sonning bridge is out that's academic.

By the way, had you noticed that Bridgwater, Athelney, and Crewkerne are all close to the Parrett?

That BBC^» (British Broadcasting Corporation - home page) story quotes this from Network Rail.
The NR^» (Network Rail - home page) item has a picture that shows the repair and the first container in place.

I guess that's another variant on the idea of blocks or gabions - whatever is to hand that's big and is, or can quickly be made, heavy.

Well, that's also true of the roads in the same area. There were two catch fences here, but 20 tons was too big for them to stop.

News coming in of a derailment near Saint-Benoit on the "train des Pignes" in the south of France - a metre-gauge local/tourist/heritage line from Nice to Digne.

First reports are that a train struck a rock on the line, and poor weather is hampering the emergency response. A carriage turned over resulting in two fatalities and seven injured.

News report with picture here. Evidently the carriage's departure from the track was much more dramatic that the early reports suggested.

The RUSs^▸ (Route Utilisation Strategy) favoured "skip-stop", which only gets you a small time gain - nothing like a true fast or semi-fast. That would need to run on the main lines, which was considered but rejected because there is no grade separated crossing onto them. The top speed would be an issue too.

But across London, the gain in time should be substantial - it's only seven miles to Canary Wharf, and with just six stops and suburban rail rather than tube timings.

Yes, I know I said "clean the ballast", and I must admit you have done that, but ...

They were filling in most of the interior of the platform, around the upright of the signal gantry. Would that be where the ground was dodgy?

They weren't using the small pump they had before, for some reason ...

The first picture also shows the track as initially aligned and sitting on top of the base ballast ... plus a wire that has to run over the wall at this point.

When I saw the cross-section of the existing wall at the break, I was struck by how insubstantial it is - more like a garden retaining wall than a proper seawall. You wonder how it has lasted so long, even with a lot a repair and maintenance each year.

As for an emergency repair, my first thought was to get some big concrete blocks and plonk them on the flat wall (path) in front of the ex-embankment. "Big" would mean big enough to stay put by their own weight - I'd guess about 20 tons or so. However, I can't find any mention of that as a standard technique, and it may not be possible to find a combination of crane size and crane siting that provides strong enough ground and adequate reach - the access road in front of the houses looks to narrow and likely to collapse. Still, I can't help feeling that there must be a method that starts with a phone call to Ainscoughs.

I can only quote this morning's Metro, alongside a slightly different picture: "The service is definitely suspended" - then repeated as "the damage to the track saw trains suspended between ... Exeter and Cornwall".

BBC^» (British Broadcasting Corporation - home page) news were reporting that Network Rail have announced their proposed plan, involving a "5-mile tunnel under Heathrow". I can't find it on their website yet, but there is this news release from NR^» (Network Rail - home page). It says the proposals have been "laid out", but not where.