I haven't felt fully satisfied regarding how many radials are sufficient; what length is required; and whether or not elevated outperforms buried systems. Rudy Severns, N6LF, wrote an article for QST in July 2000 and it was reprinted in More Vertical Antenna Classics (2006) published by the American Radio Relay League. I recommend owning this book for your antenna reference library.
My feeling is vertical and loop antennas are gaining in popularity as a solution to HF operation within regulated communities. One blogger is currently considering one while another uses a loop antenna. Locally, a couple of us propagate radio frequency, using vertical polarization because of limited space. The vertical is a good option if wire antenna systems aren't practical within your regulated community.
I've searched for a clear answer in reference to metal beneath a vertical and Severns (2002) in Verticals, Ground Systems, and Some History helped. His article opened with a tribute to George H. Brown whom wrote several papers on the topic of vertical antenna systems in radio broadcasting. Brown's work established the Federal Communication Commission standard for broadcast antenna installation according to the author.
Severns, on page 52, discussed total loss inside given radius and remaining with my understanding; the absolute total loss is much higher for a short antenna when compared to a taller one. The author further stated that a ground screen extending at least 0.1 wavelength eliminates at least 90 percent of ground loss. Certainly, for those of us living within regulated communities, achieving optimal distance is an extremely difficult proposition.
However, let the #hamr urban legend rest in peace because for short antennas one should focus on a ground system inside a short radius according to Severns (2002). Consequently, according to the author, if space and wire are in short supply then it is better to use many short radials instead of a few long radials. He defends his short versus long argument stating that a few long radials do not pick up much current.
Accounting for the buried radial argument, Severns (2002) cited Brown's research, wherein; it was determined that ground loss is small when constructing a shallow, dense packed system of radials. On the other hand, according to the work of Burke and Miller as cited by Severns, elevated radials longer than 0.3 wavelength and few of them would significantly increase your losses. However, more metal beneath your elevated vertical the better because it reduces loss and increases efficiency, according to the author.
I feel like Severns (2002) answered my questions about metal beneath the vertical. Admittedly, my posting is an oversimplification of years of scientific research lead by the likes of George H. Brown. Likewise, I've grown weary of the #hamr urban debate and, instead; I'm seeking my answers in the literature. You might consider my ongoing postings as a review and my goal is to share my discovery process.
73 from the shackadelic near the beach.