Somehow I've overlooked this! It is two weeks old; it is a survey by a team of academics, hosted on Kyoto University's Reactor Research Institute:
Soil analysis of 8 isotopes:
Note the 137Cs levels on the order of ~1 MBq/m2. This validates previous measurements (see MEXT page for their latest data). MEXT's daily 137Cs readings in this area have been fluctuating by almost two orders of magnitude -- maybe (?) they are measuring in different locations, moving in and out of hot spots (note the units are different here, Bq/kg of soil):
("Island Soil" is a mistake for "Inland Soil"). As a common reference point: the Kyoto University report describes the flowerbed soil sample as averaging 21 kBq/kg 137Cs, and 749 kBq/m2 in terms of area. Maybe this suggests a conversion factor (it's consistent with the earlier media reports).
They also extrapolate back from soil samples to estimate past dose rates -- they infer up to 200 μSv/hour on March 15, or 5 mSv/day, in Magata (where they measured their highest fallout levels). Based on their calculation, the accumulated dose there would be around 50 mSv to date, going up to 95 mSv over 90 days . Note these are outdoor doses (standing above soil); they say being inside a wooden framed house would reduce this by 1/2.
These are the radiation dose equivalents of multiple CT scans. I would definitely not stick around here for extended periods -- and indeed, the Japanese government is planning evacuations for this area.
Note that this survey is geographically limited (map below). The highest dose rates there (colored red) are 18 - 20 μSv/hour; apparently they extend higher further south (into the "stay indoors" zone). As being reported by Kyodo News and others, the NSC estimates the accumulated dose in the 20 - 30 km ring to be <50 mSv, and over 100 mSv in some areas of the (evacuated) 20 km zone.
For comparison, here's the map from the NNSA overflights in approximately the same time period. Not clear to me if "R" is short for "roentgen", or if they've misabbreviated "rem".