With this configuration and 8E20 8 GeV protons/year on target, we get a soil activation that corresponds to a concentration at the Dolamite (assumed elevation of 677') of 6500 pC/mL of tritium (limit is 20 pC/mL) and 2.3 pC/mL of Na22 (limit is .2 pC/mL). While these numbers are well above the allowed limits, they are not so far out of range that they are worrisome at this point. Additional shielding and a better simulation will probably brings these into line. However, one of the most important factors in these calculations is the depth of the Dolamite layer below the target hall. I would suggest taking core samples at three places:
As to shielding, the highest area of irradiation is right behind the dump. In the next plot, I show the star density in a linear contour plot in the soil right behind the dump. [Up is up in this plot; right is beam's right. The scale is in cm. The contours shown essentially map out the portion of the end of the decay hall not covered by the dump.] We get significant irradiation around the edges of the dump. To reduce the activation level, we must shield this area. The final plot is a logrythmic plot of the activation in the soil laterally around the decay hall. [Up is along the beam line in this plot; right is beam's right. The plot starts at the upstream shielding of the target. The contours mark out the edge of the soil outside the decay hall. The scales are in cm.] This plot shows that the upstream end of the hall is pretty well shielded, but once the shielding ends, activation begins. The highest region is at the end of the hall where the particles miss the steel dump. However, there is significant activation in the sides of the all also (more to beam's left than beam's right only because the beam is closer to the wall on the left than on the right).
The shielding studies that I would recommend (and will do) are to increase the wall thickness on the sides of the hall to get below the concentration limit, to add steel shielding to the end of the hall, and to make the hall smaller to just fit around the beamstop.
Finally, a comment on MARS vs. Casim: I have retrofitted Casim to allow us to use it to specifically answer our questions about our target hall. Unfortunately, I cannot get the same level of detail out of MARS. I plan to compare MARS to Casim on some geometries that MARS can handle and then to extrapolate to our present geometry from that extrapolation.