Start Wikipedia excerpt.
Iodine-135 is a fission product of uranium with a yield of about 6% (counting also the iodine-135 produced almost immediately from decay of fission-produced tellurium-135).[3] This 135I decays with a 6.7 hour half-life to 135Xe. Thus, in an operating nuclear reactor, 135Xe is being continuously produced. 135Xe has a very large neutron absorption cross-section, so in the high neutron flux environment of a nuclear reactor core, the 135Xe soon absorbs a neutron and becomes stable 136Xe. Thus, in about 50 hours, the 135Xe concentration reaches equilibrium where its creation by 135I decay is balanced with its destruction by neutron absorption.
When reactor power is decreased or shut down by inserting neutron absorbing control rods, the reactor neutron flux is reduced and the equilibrium shifts initially towards higher 135Xe concentration. The 135Xe concentration peaks about 11.1 hours after reactor power is decreased. Since 135Xe has a 9.2 hour half-life, the 135Xe concentration gradually decays back to low levels over 72 hours.
The temporarily high level of 135Xe with its high neutron absorption cross-section makes it difficult to restart the reactor for several hours. The neutron absorbing 135Xe acts like a control rod reducing reactivity. The inability of a reactor to be started due to the effects of 135Xe is sometimes referred to as xenon precluded start-up, and the reactor is said to be "poisoned out"[4]. The period of time where the reactor is unable to override the effects of 135Xe is called the xenon dead time.
End Wikipedia excerpt.
My non-nuclear scientist wild guess on this would be that we may have small pockets of recriticality occurring and then shutting down. All throughout the melted down reactor cores where ever the heck they are scattered about.
From the link you provided:
“Because radioactive releases after the March 11 quake occurred almost immediately, the radioactive isotopes from Japan took about 10-11 days to reach Europe.”