Some of the most extensive coal reserves in the world are found right here in the United States. Lateral beds spread out for thousands of cubic kilometers and individual beds can be as thick as 900 feet (Volkov 2003). Often, these beds are stacked one atop the other intercalated with limestones, silts, and sandstones, and can reach impressive overall thicknesses in the thousands of feet.

According to conventional wisdom, these coal reserves formed from the growth and subsequent burial of peat swamp forests over millions of years. Simplistically, organic and plant material falls to the forest floor and accumulates as peat over thousands of years. Over time, either sea-level rises or the land subsides. Either way, you get what’s called a marine transgression as the sea moves over the land. Now underwater, the peat is slowly covered by limestone as small, shelled marine animals die and fall to the seafloor. Eventually, sea-level drops and the land is exposed as it was before, allowing a new forest to grow over the same spot as the previous forest. This cycle is repeated, perhaps dozens of times.

This interpretation, however, hasn’t always dominated geology. In the late 1800s, geologists were divided over the origin of coal, with some arguing for an allochthonous (transported) interpretation. Geologist William S. Gresley, for example, the discoverer of coal balls, and often hailed as the father of coal petrology in North America, argued against coal formation in swamps by pointing to coal composition, coal underclays, coal roof-bed architecture (which was typically planar or box-like) and coal parting structures (Austin and Sanders 2018).

Based on his observations, Gresley was convinced that coal beds were produced when, “…vegetation of such character as thrived in luxuriant profusion upon the surface of the water…living afloat and dying and decaying, falling through the water” was deposited on the seafloor (Quoted by Gresley in Austin and Sanders 2018, p. 280). In other words, Gresley believed in an allochthonous model whereby coal reserves were sourced from large floating mats of vegetation, and not in situ (in place) swamps.

The in situ swamp theory eventually won out over the floating mat model due to a combination of conflicting data and aggressive uniformitarianism. Catastrophism, as a geological paradigm, was rapidly disappearing from mainstream geology, replaced instead with the slow and gradual processes espoused by Charles Lyell in his groundbreaking work, Principles of Geology. In situ (grown in place) vertical tree stumps found within many of these coals along with “rooted underclays,” in conjunction with the shifting tide towards slow and gradual processes were enough to push the consensus against a floating mat model.

The floating mat model for the origin of coal was revived by creationists Steve Austin (1979) and Joachim Scheven (1981). Since then, several other creationists have contributed to what is now known within creationism as the floating mat model for the origin of coal. (See top image below.)

The floating mat model has much to commend it. Not only does it expand and build upon the work of earlier geologists, it also better explains the presence of the conflicting data originally used to push the swamp model in the first place—the presence of “in situ” tree stumps and “rooted underclays” (explained below).

Perhaps the greatest challenge to the peat swamp theory is the presence of wide-spread clay partings that separate one coal bench from the next. These thin clay layers of sediment, often just a few inches in thickness, are currently interpreted as local flood deposits. The Kentucky Geological Survey says:

Partings are laminae or beds of noncoal rock, usually mudstone or claystone, in coal beds. They form from sedimentation during flooding of the peat that ultimately became a coal bed. Some coal beds have many partings, some do not. Some partings are regional extent, but most are local.

Essentially, the peat swamp is flooded when a local river bursts it banks. The sediment from the river covers the peat in a thin layer of mud and silt. Once the flood subsides, the peat swamp reactivates, and growth continues over the newly deposited shale parting.

The problem with this interpretation, however, becomes evident when one considers the thickness of the partings in conjunction with their lateral extent. One shale parting within the Pittsburgh Coal Bed, for example, is three-eighths-inch thick and covers an area of about 15,000 square miles (Austin and Sanders 2018, p. 280). Shale partings of this thickness and lateral scale are typical of subaqueous (under water) depositionary environments, not subaerial (above water) ones. The problem becomes insurmountable when we consider the ecology of these ancient coals.

Carboniferous coals, although variant in plant type, are primarily composed of arborescent lycopods (e.g., Lepidodendron) and tree ferns (Eble 2001). Some of these trees could grow as high as 160 feet. How can a three-eighths-inch thick mud layer be uniformly deposited over a veritable forest of 160 foot-high lycopods? That this incredibly thin layer was deposited over thousands of square miles was enough to convince many early coal workers that these coal seams had a subaqueous origin (Austin and Sanders 2018).

Even John Stevenson, an early geologist who staunchly opposed the allochthonous model was aware of such examples. He describes the areal extent of a thin coal bed (“3 to 6 inches”) sandwiched between two half-inch clay partings:

“The [clay] partings between benches of coal beds are usually extremely variable but in some beds they show amazing persistence. The bearing-in bench of the Pittsburgh bed is from 3 to 6 inches thick and is bounded by partings which rarely exceed one half inch; yet these are present under more than 2,000 square miles, changing little in thickness or in composition” (Stevenson 1913, p. 58).

Remarkably, these incredibly thin partings are largely intact and unbroken (except for the occasional “rootlets.” Discussed below). Yet in the peat swamp theory, we are told that reactivated growth of the swamp continued once the local flood subsided. Shouldn’t a three-eighths-inch thick mud layer be destroyed by the downward-growing roots of the overlying vegetation? This is especially significant since the composition of modern peat swamps is dominated by extensive root systems. Thin partings simply should not exist.

The overall dimensions of Carboniferous coal seams are also problematic for the peat swamp model. Modern peat swamps, although large in lateral extent (Staub 1991), only have an average depth of about 5-10 meters (Volkov 2003), and represent what might one day become a single coal seam. Present examples do not have swamps stacked one atop the other.

In contrast, some Carboniferous coal seams cover areas greater than 10,000 square miles with thicknesses of up to 900 feet for a single seam (Stevenson 1913; Volkov 2003; Austin and Sanders 2018). The overall coal-bearing deposit can contain as many as 100 separate seams intercalated with sandstones, shales, and muds that cumulatively have thicknesses on the order of 20,000 feet! Simply put, modern peat swamps are not good analogues for Carboniferous coals since modern processes cannot account for their deposition.

What about “in situ” stumps and “underclay rootlets”? As stated above, these two pieces of information weighed heavily on the overall decision to abandon allochthonous models of coal deposition. But is there a better interpretation?

Let’s deal with the stumps first. Back in the early 1980s, creationist Steve Austin followed up his PhD dissertation by documenting the deposition of thousands of floating logs that had been catastrophically dumped into Spirit Lake, Oregon. Austin discovered that water-logged trees and stumps sink vertically within the water column. As a log takes on water, one end of the log sinks before the other end, orientating the log vertically in the water column.

Eventually the log will sit on the lake bottom in a vertical position (see image, bottom right below). Over time, the “bottom” end of the log becomes buried by bark and other detrital material sourced from the lake’s surface. If another log sinks to the lake bottom later, it’s “bottom” end will sit stratigraphically higher than that of the first log due to the continuing deposition of bark and detrital sediments. The result is a “grove” of upright logs buried at differing levels within the lake-bottom sediments (again, see image below). If the lake was drained and a cross-section cut through this little “grove,” one might interpret the logs as having grown in place (in situ).

Given a floating mat model, lycopod trees calving from the mat and sinking to the ocean bottom provide an adequate explanation for their presence within Carboniferous coal seams. This explanation best explains the occurrence of inverted stumps in the coals. Although rarely described in the literature, creationist researchers have documented dozens of them. Conventional scientists explain their presence in terms of incoming storms, but some examples defy such a simplistic interpretation. Click here ( to view examples.

One of the latter examples has a photo of an upright stump (the bottom of which supposedly represents the “soil” horizon) sitting on top of an inverted stump. If the inverted stump resulted from a storm, shouldn’t the stump be sitting above the upright one? Why is it below the “in situ” stump?

A simplistic “in situ” explanation also doesn’t explain the nature of some “soil” horizons. It is true, some stumps are found within the coal proper, but many are found within detrital sandstones and shales where no “soil” horizon exists (Gastaldo 1986). If these were real “soil” horizons, where is the rest of the carbonaceous sediment and/or coal representative of the soil? (See image bottom left below. Notice the rootless trees standing vertically on sand horizons. Notice also the similarity of this drawing to the image on the right representative of the Spirit Lake scenario.)

Other problematic observations include eighteen-foot-high stumps filled with sandstone, the absence of non-lycopod roots in “underclays” (Gastaldo 1986, pp. 205-206), and trees that protrude through subsequent coal beds (Gastaldo 1986, figure 3). Regarding the first observation, one must ask how almost pure sand managed to “climb up” into an eighteen-foot-high stump? The second is problematic because lycopods, lacking “true” wood (except for a thin core) and being mostly hollow, would have decomposed rather quickly. How did these trees stay erect during a marine incursion, and then mange to remain erect for thousands of years while another peat swamp grew around it? The latter observation is pertinent given the poly-specific nature of these “swamps.” Why is it that almost all upright stumps and associated “rooted underclays” have a lycopod origin? Where are all the other “roots”?

That leaves one final piece of information: the rootlets. Sediments interpreted as underclays (stratigraphically sitting below fossil stumps) often contain stigmarian roots. Lycopods are an extinct tree that flourished in the pre-Flood world. Although they could grow to heights of 160 feet, however, they were not comparable to modern trees. Carboniferous lycopods lacked “true” wood, were mostly hollow, and bore hollow “roots” called stigmaria.

Emanating from the stigmarian roots were even smaller rootlets that splayed out in a pattern not unlike the bristles on your toilet bowl brush! It is not uncommon to find stigmarian roots lying on a “soil” horizon with the rootlets protruding down into the “soil.” Often, however, only the rootlets are found; the stigmarian root not being preserved.

As with the stumps (image bottom left), stigmaria and associated rootlets are sometimes found within sandstones, and not within carbonaceous soils. I propose a similar scenario as that described above for the stumps. As stigmarian roots fell away from the floating mat, they sank to the ocean floor where they were buried by terrigenous sand (sand sourced from land-based river systems) and/or other mat-derived debris. Roots, stumps, and other mat debris buried en masse became a coal seam. During times of minimal agitation, “sprinklings” of stigmaria sparsely covered the ocean floor where they were buried by sand and shale. These pseudo-horizons of buried stigmaria also preserved the rootlets which are now interpreted as “penetrating” structures. Often, only the fossil rootlets are present without the stigmaria root. I content that the hollow stigmaria were eventually flattened by the overburden, leaving just the rootlets “penetrating” the substrate.

This interpretation might solve the almost monotypic nature of “rootlets” in Carboniferous coals (Gastaldo 1986, p. 206). The 360° radiating nature of the rootlets, stemming at right angles to the stigmaria root, guarantee that the rootlets will sit on the ocean floor in a vertical orientation. Most of the other Carboniferous plant root systems lack this potentiality.

Floating mat model

Floating mats of vegetation are not uncommon today. Krusi and Wein (1988, p. 61), for example, document the existence of a Phragmites (a species of grass) mat that covers hundreds of square kilometers floating on the Danube River in Rumania.

Most of the plants associated with Carboniferous coal reserves were semi-aquatic and/or aquatic (Gastaldo 1986) and lacked “true” wood. Lycopod trees, such as Lepidodendron, along with the stigmarian roots were also hollow. Unlike most tree roots that are generally orientated “down” into the ground, Lycopod roots “splayed out” from the main trunk at near right angles. In other words, the stigmarian roots wouldn’t have efficiently penetrated a true soil. These observations led many early scientists to postulate a Carboniferous floating mat. Botanist Otto Kuntze even sketched such a mat back in 1884 depicting arborescent lycopods with stigmarian roots splayed out over the surface of the water (Austin and Sanders 2018, figure 2).

In a floating mat model, lycopod stigmarian roots would intertwine with the roots of neighboring trees. Understory vegetation would fill in the gaps, and a fresh-water table within the mat would provide sustenance for flora and fauna ill-equipped to life in brackish or salty environments (Austin and Sanders 2018).

One of the greatest challenges to the floating mat model is the overall architecture of Carboniferous cyclothems. As the name suggests, a cyclothem is a genetic package of vertical strata that repeats itself. These cyclothems have been well-studied and typically consist of a marine limestone, followed by a shallowing-upwards sequence of shales and sandstones, then an “underclay,” and finally capped by a coal seam. Some Carboniferous sequences have dozens of such cyclothems so the creationist researcher must take them into consideration (Hampson et al. 1999).

Although cyclothems often do not follow the exact sequence stated above (Hampson et al. 1999), they nevertheless are suggestive of a genuine depositional cycle.

As a first order approximation, I suggest that continental-scale, floating mats existed in the pre-Flood world (Wise 2003) (Top image below). These mats may have been anchored to continents, or they may have been free-floating (Wise 2003). I propose that mat destruction likely began before the onset of the Flood proper, as exponentially increasing geologic activity set the stage for the events recorded in Genesis 6-9. This means that the mat, designed as it was for calm settings, slowly began to break up as conditions became progressively tumultuous.

As a result of these escalating conditions, sea-level rose and fell in cyclic fashion. At the onset of the Flood proper, these cyclic conditions continued within an overall sea-level rise. I propose that the limestones were deposited during sea-level high-stands with an influx of marine sediment. As sea-level dropped, land-based river systems dumped sand, silt, and mud into the ocean beneath the mat. The sediments that travelled farther from the land were the silts and muds, and so these were deposited on the ocean floor “first.” Following these fine sediments came the sands, much like what occurs today in modern deltas.

During this same phase, rip currents tore the mat to shreds from the outside in. Since the water was agitated, the non-woody lycopod trees became entrained in the silts and sands, allowing debris to be settle at various rates and entombing it within these sediments. This accounts for sand and/or shale that almost always is found within lycopod stumps and roots, as well as for stumps that are found sitting on silt and sand horizons.

Between this low-stand and the next high-stand, water became less agitated allowing the remaining vegetative debris to coalesce into a future coal seam on the ocean floor. Entrained mud could also flocculate (coalesce and thus act like fast-falling sand grains) and be deposited as continuous, lateral layers extending over very large areas of thousands of square miles. This explains thin clay partings within coal seams. The high-stand would signal a return to the limestone phase.

The regional scale of these depositional processes accounts for the very flat bottoms and tops of extant coal seams, the large lateral extent of coal seams, and the very thin lenses of detrital sediments found within these seams.

Anyway, just a thought!

Austin, S.A., and Sanders, R.W. 2018. Historical survey of the floating mat model for the origin of Carboniferous coal beds. In Proceedings of the Eighth International Conference on Creationism, ed. J.H. Whitmore, pp. 277–286. Pittsburgh, Pennsylvania: Creation Science Fellowship.

Dawson, J.W. 1868. Acadian Geology. The Geological Structure, Organic Remains, and Mineral Resources of Nova Scotia, New Brunswick, and Prince Edward Island, 2nd edition. MacMillan and Co.: London, 694pp

Eble, C.F., Greb, S.F., and Williams, D.A. 2001. The geology and palynology of Lower and Middle Pennsylvanian strata in the Western Kentucky Coal Field, International Journal of Coal Geology, Vol, 47, pp. 189-206

Gastaldo, R.A. 1986. Implications on the Paleoecology of Autochthonous Lycopods in Clastic Sedimentary Environments of the Early Pennsylvanian of Alabama, Palaeogeography, Palaeoclimatology, Palaeoecology, Vol 53, pp. 191-212.

Hampson, G., Stollhofen, H., and Flint, S. 1999. “A sequence stratigraphic model for the Lower Coal Measures (Upper Carboniferous) of the Ruhr district, north-west Germany,” Sedimentology, Vol 46, pp. 1199-1231.

Krusi, B.O, and Wein, R.W. 1988. Experimental Studies on the Resiliency of Floating Typha Mats in a Freshwater Marsh. Journal of Ecology, Vol 76, pp. 60-72.

Scheven, J. 1981. “Floating forests on firm grounds: Advances in Carboniferous research.” Biblical Creation, Vol 3, No 9, pp. 36-43.

Staub, J.R. 1991. “Comparisons of central Appalachian Carboniferous coal beds by benches and a raised Holocene peat deposit,” International Journal of Coal Geology, Vol 18, pp. 45-69.

Stevenson, J.J. 1913. “Proceedings of the American Philosophical Society,” Vol. 52, No. 208, pp. 31-162.

The Kentucky Geological Survey, accessed 6/9/20,…/coal/coal-mining-geology-partings.php

Volkov, V.N. 2003. “Phenomenon of the Formation of Very Thick Coal Beds,” Lithology and Mineral Resources, Vol. 38, No. 3, pp. 223-232.

Wise, K.P. 2003. The pre-Flood floating forest: A study in paleontological pattern recognition. In Proceedings of the Fifth International Conference on Creationism, ed., R.L. Ivey, Jr., pp. 371-382. Pittburgh, Pennsylvania: Creation Science Fellowship.