Spoke too soon. Following data shows that at the area of the plant that has been measuring levels in the 3000-4000 microsievert range in recent days (3-4 millisieverts), around 3.20pm JST the levels started to go back up again. They had fallen as low as 3345 microsieverts but as of 4pm when the data ends, had gone back up to 4485 microsieverts (4.485 millisieverts).
The next set of data came out. Levels continued to climb, reaching 5055 microsieverts (5.055 millisieverts) at that location at 5PM, and the falling back, down to 3611 microsieverts by the time the data from that location ends at 8PM. The data at another location at plant boundary, I believe the one that featured in news stories I was complaining about, where levels had dropped to 289 microsieverts, also went back up at the same sort of time, and is also now back on the decrease but is still well above the levels mentioned in earlier stories.
Its hard to know if something new happened, or if its mostly to do with wind on site. Wind direction data is supplied with these figures, but it often changes frequently without having tended to make any obvious difference to the numbers. Further away from the plant the wind direction at different points in time has made a pretty huge difference to which places have suffered the worst contamination.
There has not been any new bad news about caesium or iodine for some days now, and its probably much too early to expect to hear about how much of this stuff may be spread over an area, or how big that area really is.
Edited to add that it has now become clear where the location on the site that I quote these numbers in the 3000-4000 microsievert (3-4 millisievert) range is. Its not on the site boundary like most of the other reading sites, it is 0.5km away from reactor number 2, north west of it to be precise. So approximately twice as close to the reactors as the other monitoring points on site, which tend to give readings ten times lower than these, are. Cannot really use these numbers to estimate the level right close to the reactors, but guess it could easily involve shifting the decimal place several points.
Not much other news from the IAEA & Japanese authorities, although data on the shared fuel pool seems to now be available and it is said to be ok at the moment.