There are lots of stumps, some upright on several levels in Nova Scotia. Coffin and Rupke have studied these in some detail, and I believe both of them have published their work in scientific journals. Their findings, based not on surmise, but on experimental data using the only surviving relatives of the Carboniferous flora, led them to conclude the trees could have been transported in upright and buried in situ. The roots associated with these plants, Stigmaria, by form genus name, are present in abundance in the sediments. These root-like structures were likely transported with the debris flow that entombed the trees, since it seems highly unlikely that Stigmaria could penetrate soil. Rather the roots, which spun off spirally arranged rootlets, and had somewhat the appearance of bottle brushes, appear to have been designed to penetrate peat, and to hold upright the quite large superstructures of these plants by spiraling through the vegetative substratum of a peat bog. The rootlets, about the diameter of a pencil, and blunt ended, were made up of very thin walled cells with no apparent strengthening features. The root itself grew from a terminal bud which was quite blunt and appears ideal for penetrating peat, but quite incapable of penetrating soil.
Since these are found, often disconnected from the host plant, in carboniferous sandstone, the conclusion that both trunks and Stigmaria were transported seems to me to be inevitable.
The so-called "seat earths" or "soil horizons" that underlie coals in many areas, have also been used as evidence of autochthonous origin. However, a careful analysis of the presumed soils reveals that they rarely have enough clay minerals to be typical soils, and they rarely show any degraded organic material. A number of papers have been written on this subject, and it is not a closed case.
In a different, but related context are the famous Fossil Forests of Yellowstone. These "forests" consist of upright and prostrate trees entombed in volcanic breccia (broken rock fragments resulting from volcanic flows) on many levels (some estimates are as high as 66 levels with fossil trees). As a result of careful scientific studies, and the "laboratory" studies facilitated by the eruption of Mount St. Helens, the National Park Service no longer refers to the deposits as Fossil Forests, but as fossil tree deposits. For more details of these studies see another paper at this site.
2010 Arthur V. Chadwick, Ph.D.