This whole issue illustrates a complete misunderstanding and/or willful distortion of sedimentary processes by creationists.
One of the first principles that is extremely important to understand is that sedimentation in any one location is a fundamentally discontinuous process. In order to be able to deposit sediment, you need three things: 1) a source of sediment; 2) a means of transporting that sediment from its source to a given location; and 3) space to deposit the transported sediment. The variability of these three items in any one location will determine the completeness of the sedimentary record there.
For example, let’s say you can have a source for sediment and a means to transport it, but there’s no space available for a period of 500 years in the spot you’re interested in. The sediment will then be carried on to another location (a process called sedimentary bypass), and your spot will have no sedimentary record of that 500-year span, i.e., it experiences a period of non-deposition. Assuming that there is space available for a sedimentary deposit, then fluctuations in the amount of source material available and/or transport means will lead to more or less sediment being deposited at different points in time - that is, the rate of sedimentation does not stay constant over time.
And of course, erosion also plays a major role in the development of the stratigraphic record. Sediments or sedimentary rocks can themselves become sources of sediment for younger deposits through the erosional process, and thus play an important part of the rock recycling process in the Earth’s crust. In the process, you can’t help but erase the sedimentary record in one location while creating a record in the new. Erosion can also make space for those new deposits to accumulate - canyons, valleys, lakes, and other depressions can become filled over time.
Now, if you have consecutive sedimentary layers that are undeformed, it can become rather difficult to determine where erosion has affected the completeness of the rock record. That is particularly true in very fine-grained rocks like the Hermit Shale, because they do not have readily visible beds (more likely laminae just a few millimeters thick at most) that would make the results of regional-scale erosion obvious. That’s the point at which dating methods such as biostratigraphy come in handy for constraining the age of the rock. The 10 million year gap between the Hermit Shale and the Coconino Sandstone would have almost certainly been determined that way.
W/r/t the issue of fossils sticking up through younger layers: There is a process called differential erosion, which results in a rock being weathered and losing material in an uneven fashion. This typically happens when the composition of the rock is not uniform, either in terms of mineralogy (different minerals have differing levels of resistance to erosion) or grain size (smaller grain sizes tend to be more easily eroded). In the case of the “polystrate” fossils, all you need is for the fossil to have been replaced by a mineral that is more resistant to weathering than the material it is surrounded by. (The process of fossil mineralization can be complex and gets into details of chemistry that I don’t have space to describe here.)
With petrified trees, for example, the tree has commonly been replaced by silica (quartz) that is more resistant to erosion than the volcanic ash deposits around it, so it is possible to exhume the fossilized tree at least partially. At any rate, the end result is that you have a fossil tree sticking up out of its original deposit, and in a position now to be buried by younger sediment. How much younger that new sediment is turns out to be depends entirely upon the three criteria I listed above: sediment source, means of transport, space for deposit. It is in fact entirely possible to have a landscape that experiences a rate of new sedimentation of a meter or less over a total of millions of years - vast portions of Australia are a perfect example of a landscape that has hardly changed over the last several million years. If you now cover the fossil tree with more sediment - gradually over a period of years or through a flash flood, doesn’t matter - you will have something several million years old sticking up into much more recent sediment. There is no problem reconciling the age of the fossil tree with the younger sediment that it protrudes into.
Now, just imagine that the process I’ve just described happened 20 million years ago, so that even the younger sediment cover over the tree has been lithified (turned into rock). Why would you now interpret this process to mean that there is something wrong with the age determination of the rocks with respect to the fossil tree?