It's a bit confusing to compare concentrations of cesium 137 to radon, in cubic meters vs liters. Yes, both are forms of radiation, but are they comparable in this manner?
The most useful information pertains to the practical significance of these concentrations, such as the incidence of cancer and birth defects, and comparisons to "normal" background levels.
Comparing radon and cesium might seem off to you, but as it happens the sources of the radiation are less important than the fact that both are radiation. The radiation is question is discrete particles; the measurement unit "becquerel" (bq) is the occurrence of a single particle from a decay event per second, regardless of the source. The radioactivity from radon is mainly alpha radiation, while the cesium is beta; these are fairly comparable in that alphas are rated as 20 to 1 harsher in effect than beta; see sievert
WP. Thus my comparison of a harsher radiation type, higher in concentration (both numerically and by a factor of 1000 in a smaller volume [there should be no confusion over liter vs cubic meter]) was meant to qualitatively point to a lack of impression on my part with the amount of radiotides in the ocean water off SF.
Twenty years ago such small concentrations of radiotides would have been unmeasurable; they are today unmeasurable in chemical terms; they are only detected because of their radioactivity. Today they are scare headlines from the anti-nuclear power media.
I agree that it would be appropriate to put the dosage in more human terms. Unfortunately, the arguments as to how exactly to do that are still wrapped in controversy and as yet unexplored data. That started when studies of the Japanese exposed to radiation in Hiroshima/Nagasaki showed that for the data available, it appeared that harm caused by radiation was roughly linear to the amount of radiation dose; this is the "Linear No-Threshold" theory. Unfortunately, the data was very sparse in the mini-dose region. Further data from Chernobyl in particular seems by many researchers to deny the LNT model in low intensity situations, but rather support a non-linear "hormesis" theory, in which the body's immune system takes a hand in helping cells recover from radiation damage. Since this has an obvious dimension in the nuke/anti-nuke power debate, don't expect quick resolution.
Scientists have taken a step in that direction, though, in declaring a unit of radioactivity which quantifies the biological action of various radiation types and their targets. The "sievert" (SI unit of equivalence to the older unit REM, or radiation equivalent man) does sort out the types of radiation and how they interact with the body. 4.50 sieverts (450 REM) is the amount of absorbed radiation which has a 50/50 chance of killing the recipient human, regardless of source (alpha, beta, fission event) or target (whole body, head, heart, hand). That leaves only the manner in which objective measurements (such as mentioned for Cs137 or radon above) can be converted to seiverts. That is difficult. For a look at important numbers in this realm, see
http://xkcd.com/radiation/ .