feature

It’s all change for our view of the Carboniferous, as modern palaeobiology and sequence stratigraphy revolutionise our understanding of Coal Measure cycles, say Howard Falcon-Lang, Martin Gibling, Michael Rygel, John Calder, and Sarah Davies*

Fig. 2

A dance to the Ask a geologist what the Late Carboniferous world was like and you will be immediately regaled with romantic tales of Mississippi-like deltas and gigantic dragonflies. Three hundred million years ago, he or she will enthusiastically tell you, North America and Eurasia lay together on the equator and steamy coal-forming rainforests stretched in a continuous belt from Kentucky to the Urals.

This “Carboniferous Coal Age” is without doubt a favourite story in our Earth History repertoire. Indeed, as static and timeless as the coal forests that form its subject, this tale has remained essentially unchanged across a century of retelling. Evocative museum dioramas from the late 19th Century look essentially similar to modern reconstructions, even if details have been refined (Fig. 1). But should we be entirely surprised? With declining coal production in Europe and North America, is it realistic to expect that anything “new” could come out of the Carboniferous? Well it may be that an unlikely liaison between palaeobiology and sequence stratigraphy may be about to alter our view of the Carboniferous tropics for ever. Over the past decade we have been studying what Sir Charles Lyell referred to as the finest natural Upper Carboniferous exposure on Earth: the Fossil Cliffs of Joggins, Nova Scotia, Eastern Canada (Fig. 2). These spectacular sea-cliffs, hewn by the world’s highest tides, have a special place in the history of geology. In particular, the famous Joggins fossil forests, which record the repeated growth and destruction of luxuriant coal-forming vegetation, once led Lyell to remark that nowhere else had his uniformitarian ideas been so confirmed as in Nova Scotia. Yet despite being powerfully iconic of a stable Coal Age, this breathtaking natural laboratory ironically holds the key to a much more dynamic understanding of the Carboniferous tropical world.

Fig. 1

*HJ Falcon-Lang, University (Bristol University, UK); MR Gibling & MC Rygel (Dalhousie University, Canada); JH Calder, (Department of Natural Resources, Canada) and SJ Davies (University of Leicester, UK).

Fig. 3

Fig. 1. A Late Carboniferous dioramas published in Geology by Chamberlain & Salisbury, Henry Holt & Company1906. Fig. 2. Part of the Upper Carboniferous section at Joggins, where 25m cliffs have been carved by the world’s highest tides on the Bay of Fundy, Nova Scotia. Fig. 3. A typical sedimentary cycle at Joggins showing the main facies formed during transgressive-regressive events. Fig. 4. Spectacular tree trunk buried in deltaic sands during the early part of regression. Fig. 5. Sediment mound associated with an upright trunk formed as high-energy floodwaters scoured around the trunk and infilled the resultant hollow with sand. Fig. 6. Curious downturns in the red beds units at Joggins mark the position of trees that were oxidised away shortly after burial.

Fig. 4

e music of time Cyclic sedimentation

We have recently logged the entire 1433m-thick section at Joggins (for the first time since the ambitious work of Sir William Logan in 1845) and developed a sequence stratigraphic framework for the succession. As at most other palaeotropical Upper Carboniferous sites worldwide, the Joggins succession is characterised by distinctly cyclic sedimentation patterns (Fig. 3). Each of these typically 30-100m thick cycles represents a transgressive-regressive event as the sea first flooded the continental margins and then more slowly withdrew. However, because the Joggins basin was connected to the open ocean by narrow straits, a brackish embayment - rather like the present-day Baltic Sea developed at times of maximum transgression, instead of the full “marine bands” seen in the European Coal Measures. Presumably heavy tropical rainfall and high river discharge freshened these partially enclosed bays. It is now well established that many Carboniferous cycles were primarily generated by c. 75m amplitude sea-level fluctuations caused by the waxing and waning of Gondwanan ice sheets in high southern latitudes. However, this sedimentary record of glacialinterglacial climate cycles has been somewhat modified locally at Joggins because strike-slip tectonics resulted in unusually high rates of basin subsidence. Rapid subsidence had the dual effect of reinforcing the signal of sea-level rise, while suppressing the record of sea-level fall. Consequently, while Joggins cycles contain coals that are overlain by sharp-based fossiliferous limestones that mark rapid sea-level rise, they lack unequivocal corresponding evidence for sea-level fall (such as a palaeovalley incision and mature palaeosols).

Fig. 5

Fig. 6

Volume 14 • No. 4 • 5

Fig. 7. Charcoal from cordaite wildfires showing exquisite preservation under the SEM.

However, this rapid subsidence rate is also the key to understanding the importance of Joggins because it has resulted in the preservation of a remarkably complete record of environmental change. While at many other cyclic Carboniferous successions, the only surviving record of regressive phases are uninformative mature palaeosols (condensed terrestrial sections), the “expanded record” at Joggins provides a unique window on these fascinating and poorly known sections of deep time. By examining fossil assemblages within this dynamic and extraordinarily complete framework we have been able to show more clearly than ever before how Late Carboniferous terrestrial ecosystems responded to transgressive-regressive cycles.

Rise and fall of coal-forming wetlands One interesting outcome of our work at Joggins has been the recognition that coal-bearing strata developed both during transgressive phases and the early part of regressive phases, and the discovery that the vegetation ecosystems responsible for each were distinctly different. Amalgamated coal zones formed during the early part of transgression are up to 1.5m thick at Joggins. Studies of modern rates of tropical peat accumulation suggest that these coals formed slowly over a period of several thousand years as mire vegetation unsuccessfully tried to hold its ground against the encroaching brackish bays. It is hardly surprising that these coals have high sulphur contents because they were probably frequently infiltrated by saline coastal waters in this setting. Two tree-sized club mosses called Lepidodendron and Lepidopholois dominated this coal-forming environment. Most readers will be familiar with the distinctive diamondshaped bark of these bizarre trees, which are among the most abundant and distinctive plant fossils in the Carboniferous Coal Measures. The known ecological tolerance of these plants tells us that these thick coal seams were formed in long-lived, submerged, nutrientpoor peat mires. In marked contrast, coal-forming environments that existed during the early stages of regression were much more dynamic disturbed settings. The sandy deposits of small delta channels dominate these successions as coastal regions advanced to infill the deep embayments formed during sea-level rise. Coals in this delta top setting are typically only a few centimetres thick but often exhibit spectacular upright tree trunks rooted on top of them and

buried by sandy sheets (Fig. 4). Amazingly, not just the stumps of these trees are preserved but in some cases up to the lower 7m of the trunk have been buried as well. Metre deep scours and complex sediment mounds formed in the sands around the bases of many of the trees bear witness to the destructive power of sediment-laden floodwaters that periodically wiped out this vegetation (Fig. 5). These delta top mires were dominated by the tree-sized club moss, Sigillaria, together with a great diversity of other plants such as horsetails and seed-ferns. Most of the plants were especially well adapted to flood disturbance by their prolific reproductive ability and their amazing capacity to re-sprout from underground rhizomes.

Enigmatic dryland ecosystems Our sequence stratigraphic analysis of Joggins fossil assemblages has also shed light on Late Carboniferous dryland ecosystems, which existed during maximum regression. Although knowledge about wetland coal-forming communities has been steadily accumulating for two hundred years, until recently really very little has emerged about the drylands. One reason for this is that fossils in Carboniferous red beds have always been assumed to simply represent poorly preserved examples of the same kind of fossils seen in the coal-bearing strata. An extreme expression of this understanding was the view held in some circles that Carboniferous strata had been reddened by aggressive, oxidising groundwater recharged from Permian desert landscapes, but it is now clear, at Joggins at least, that the red beds and paleosols formed at the landscape surface. The Joggins dryland environments were sun-baked alluvial plains crisscrossed by seasonal river networks, and in many ways resembled the anastomosed alluvial plains seen in northern and central Australia today. Here and there fossil tree stumps are rooted within the bases of channel deposits showing that these river courses periodically dried up for long enough for trees to establish within them. Due to the oxidising conditions plant material is rarely preserved. This is most evocatively demonstrated by the occurrence of curious narrow hollows up to 2m deep, filled with slumped material, that record the position of trees that oxidised away shortly after burial (Fig. 6). Where plants are preserved, they reveal that the Joggins drylands were dominated by a type of conifer-like shrub known as a cordaite, in contrast to the club moss-dominated wetlands. Cordaites probably thrived in dry settings because they reproduced using seeds, an innovation analogous to the reptile egg that afforded them much greater drought resistance than the spore-producing club mosses. Cordaite remains are often exquisitely preserved as charcoal from ancient wildfires (Fig. 7). Although charcoal also occurs in coal-bearing deposits, it is almost an order of magnitude more common in red beds indicating that the dryland vegetation was particularly fire-prone. Estimates of fire temperature based on charcoal reflectance suggest that high intensity conflagrations in excess of 800°C were a regular feature of the drylands. continued on page 8

6

Geoscientist

Fig. 8. Cliff section showing waterhole deposits in a red bed alluvial channel complex.

An important further advance in our knowledge of dryland ecosystems has been the discovery of a diverse faunal assemblage in one river channel deposit at Joggins (Fig. 8). Detailed sedimentological studies of this deposit have shown that it represents a waterhole formed as the river partially dried up during the dry season. Gigantic freshwater Archanodon bivalves have been found in life position in the point bars of this ponded channel, where they probably lay dormant in the silty sediment during drought like many Australian bivalves today (Fig. 9).

Fig. 9. Gigantic freshwater bivalve from waterhole deposit.

The distribution of this genus across Late Palaeozoic Euramerica is strongly disjunct indicating the endemic nature of this organism and the inherent difficulties of dispersal in such isolated habitats. Another denizen of the waterhole was the earliest known land snail, Dendropupa. These remains are preserved clustered around organic detritus, and were evidently entombed in sand whilst the snails were enjoying their last meal! Finally, the skeletons of a variety of semi-terrestrial amphibians also occur, these animals perhaps being drawn to the waterhole during regional drought when surface water was scarce elsewhere (Fig. 10).

Far-off mountains

Fig. 10. Some vertebrate remains in the waterhole deposit.

8

Geoscientist

A final facet of the Late Carboniferous tropical world illuminated by our work is the question as to whether mountainous terrains were vegetated, like the lowlands. This has been a source of heated controversy for decades. More recently its importance has grown because determining the presence or absence of upland vegetation is the key to refining mathematical models of the long-term carbon cycle, and to computer climate reconstructions. However, given that uplands are sites of net erosion how, you might ask, can this debate ever be resolved?

This difficulty was actually overcome by the famous palaeobotanist, W G Chaloner, as far back as 1958. Chaloner cleverly reasoned that at times of maximum transgression (when the lowlands were drowned) most of the vegetation detritus washed into the basins would come from the upland zones that remained emergent. At Joggins, the deep brackish bay deposits formed at times of maximum transgression contain a rare and poorly preserved flora that has been entirely overlooked in previous studies. These fossil plants are often highly fragmented, having undergone significant transport; but a few specimens are more complete and readily identifiable. Interestingly, the brackish bay assemblages are dominated by a number of strange archaic progymnosperms, plants much more characteristic of Devonian or Early Carboniferous times than of the Coal Measures. Accepting Chaloner’s hypothesis, the progymnosperms, together with co-occurring conifer-like cordaites and a few seed-ferns, may provide a faint glimpse into the composition of the distant upland vegetation. These forests may have cloaked the Himalayan-scale mountains that were rising along the collisional suture between the supercontinents of Laurasia and Gondwanaland, or grown on more local intrabasinal fault blocks like the Caledonian Highlands near Joggins.

All change All these exciting ongoing discoveries at Joggins are harbingers of great change. A much more dynamic model is replacing the paradigm of a persistent uniform “Carboniferous Coal Age”. While it remains true that coal-forming ecosystems with nearly constant species composition existed throughout the Late Carboniferous interval, the emerging picture is one of repeated ecosystem fragmentation and reassembly rather than monotonous stasis. The rhythmic fall and rise of sea level, probably caused by glacial-interglacial cycles, laid down a steady global beat to which the Carboniferous tropical biomes appear to have kept unerring time. Further reading Davies, S.J. & Gibling, M.R. 2003. Architecture of coastal and alluvial deposits in an extensional basin: the Carboniferous Joggins Formation of Eastern Canada. Sedimentology, 50 (3), 415-439. Falcon-Lang, H.J. 2003. Response of Late Carboniferous tropical vegetation to transgessive-regressive rhythms, Joggins, Nova Scotia. Journal of the Geological Society, London, 160 (4), 643-648. Falcon-Lang H.J. 2003. Late Carboniferous dryland tropical vegetation, in a dryland alluvial-plain setting, Joggins, Nova Scotia, Canada. Palaios 18 (3), 197-211. Falcon-Lang, H.J. Rygel, M., Gibling, M.R. & Calder, J.H. 2004. An early Pennsylvanian waterhole deposit and its fossil biota in a dryland alluvial plain setting, Joggins, Nova Scotia. Journal of the Geological Society of London, 161 (2), in press.

Volume 14 • No. 4 • 9