The ground system (described earlier) is now "complete". The shack's Single Point Ground (SPG) panel is now connected to the copper waterpipe entrance via ~30 ft of 1.5-inch copper strap. The attachment is adjacent to the AC service entrance ground clamp. It would be better to go immediately out the window by the SPG to "ground" -- only a few feet away, except that that ground is dry (under a stone walkway) and there is only 4 or 5 inches of soil on top of granite ledge.
The next improvement in this system, in my opinion, would be to lay a perimeter ground loop around the house, attached to a large grid that extends over the ledge under the soil. A lot of work for a not super-high-risk area.
This morning at 5 AM, we had the first test of the system as a lightning rich storm front passed through. There were apparently some hits in the neighborhood, but all equipment survived here.
We are left with an S9-plus source of HF noise, however. It sounds like power line arcing, and it's bad over at least 40-12 meters. The good news (?) is that it's not in our house. The beam indicates a strong maximum at about 10 degrees azimuth, which is the direction that power comes in in the neighborhood.
Your comments are appreciated!
June 29 note: The intense RFI disappeared by evening. I had predicted to the XYL that it might just "burn itself out". I did a little neighborhood sniffing and found the noise was coming from a nearby house. By its sound (very spikey 60 Hz related), it could have been a burned-out diode in a battery charger, caused by a lightning surge. (One neighbor reported their phones went out of service.) I saw a smaller version of this when the 12-volt switcher "brick" for my computer's LCD display went bad. Lots of RFI, and the brick got quite hot to the touch.
So the RFI disappeared, I did not have to complain to anyone, and the neighbor's house did not burn down.
By the way, this was the first time I got to test the Orion's noise blanker function. (At this QTH, there is rarely much impulse noise.) The hardware NB was quite good, up to 5 S-units suppression on 20 meters. The DSP NB helped some, but not nearly so much.