GHOSTS IN THE ESTUARY

As sea-level rises, as it has been throughout the Holocene, and now rises faster due to human-accelerated climate change (yes, MAGA-world, that's a fact), bottomland hardwood swamps near the coast are being converted to oligohaline* and brackish marshes. As water levels rise, trees that can't tolerate regular flooding and constantly saturated soils die off. Then, as saltier water intrudes, trees that can tolerate being wet all the time but cannot tolerate salinity die off. This is creating so-called "Ghost Forests" (see my earlier Ghost Treespost). Generally the last tree standing, except for a few salt-tolerant shrubs and some holdouts on isolated higher spots, is baldcypress (Taxodium distichum) which can tolerate a little bit of salinity, which swamp tupelo and water tupelo (Nyssa biflora, N. aquatica) cannot. Baldcypress wood is famously decay-resistant; and the white trunks of dead trees are called ghost cypress. 

Oligohaline marsh and ghost cypress, Upper Broad Creek (top) and Cahoogue Creek (bottom), North Carolina.

You can see this happening throughout the southeastern U.S.A. coastal plains, but the general swamp-to-marsh conversion has been going on, albeit more slowly (see How to Drown a Swamp), for a long time. The evidence is under, or just above the water--stumps and roots of bald cypress at the edge of the marshes, with no living (and often no upright dead) cypress anywhere close by.

Baldcypress stump, Hancock Creek, N.C.

Cypress wood is renowned for its decay resistance, which is attributable largely to an oil called cypressene. In a recently-deceased cypress trunk, log, or root, the bark and sapwood decompose fairly rapidly. The heartwood, however, can persist for very long periods, particularly underwater, or buried in wetlands (fungi and insects can eat away some of the heartwood, creating tree cavities and some interesting patterns in the wood, such as sought-after "pecky" cypress). 

Remnants of a cypress stump adjacent to black needlerush (Juncus romerianus) marsh, Upper Broad Creek

How long is very long? Good question. There exists a cottage industry for lumber from cypress "sinker logs." From the 1880s to about 1920, a lot of virgin and old growth cypress was logged in the south, and was typically dragged from the swamps to canals, creeks, and rivers to float in log rafts to market. Some of the logs sank (one source estimates 10%). Because sinkers were difficult and expensive to recover at the time, nobody did. But in the late 20th century it was realized that these often huge, high-commercial-quality logs--with the heartwood typically not noticeably decayed--were worth recovering. Most of this activity is in the Gulf Coast states, but at least one operation is going on the Cape Fear River, N.C. (https://www.oldgrowthriverwood.com), and another in Charleston, S.C. (https://www.facebook.com/HeartwoodSouth/). Obviously cypress wood can last for >100 years underwater.

Sinker cypress wood for sale at Heartwood South

Ancient buried cypress wood is called "subfossil" rather than fossil because the heartwood is preserved more or less as-is (or as-was), and has not been fossilized by mineral replacement. Buried subfossil cypress has been reported in the southern U.S. since the 18th century. Sand mines along the Pee Dee and Lynches Rivers, S.C. have uncovered subfossil logs roughly 12 m below the ground surface, up to 2.4 m in diameter and up to 29 m long (Stahle et al., 2005). Researchers from the University of Arkansas Tree Ring Lab who studied the logs reported that other hardwood logs were also present, but none as well preserved as the Taxodium distichum. Radiocarbon dates indicate their ages range from about 25 to 45 thousand years old. The report also reviewed other accounts of subfossil cypress, including stumps in growth position with roots and knees still attached uncovered by 20th century construction in Washington, D.C., estimated to be about 100 ka. Rooted cypress stumps and fallen logs exposed along the Intracoastal Waterway in Horry County, S.C. are believed to date to 125 to 135 thousand years BP. There's more about subfossil cypress logs and what can be learned from them in Stahle et al. (2012). 

Along eroding shorelines of the Neuse River estuary, N.C., a swamp paleosol is exposed with cypress roots and stumps in growth position. This paleosol is under the ~200 ka Flanner Beach formation and is at least that old. It overlies the James City formation, which may be up to a million years old, but is generally thought to be about 700 ka (Miller, 1986a; 1986b). 

Without radiocarbon dating the age of the underwater marsh edge cypress cannot be determined, but we can say they could be very old. My guess is less than 4 ka, which is about the time the modern estuaries of the Carolinas became established. 

It is not easy, at least if you are no better a photographer than I am, to get good pictures of the underwater cypress, but below are a few attempts. All these examples are adjacent to brackish marshes with no trees nearby. They are all from tributaries of the Neuse River estuary, not because that's the only place this is happening, but because that's where I live and therefore kayak a lot. Early 20th century records of the U.S. Army Corps of Engineers dredging operations in the Pamlico River at Washington, NC, report encountering in situ cypress stumps at the bottom of the river, for one example. 

Portion of a submerged cypress stump, Goose Creek, Pamlico County, NC

Cahoogue Creek, NC

Upper Broad Creek, NC

A synthesis of Holocene, recent, and predicted future sea-level rise for the N.C. coast was produced by Kopp et al. (2015). Over the past 11 ka, the region experienced episodes of RSL (relative sea-level) rise acceleration and deceleration, but no periods of fall or stillstand. RSL reached to <10 m below modern levels by about 2000 BCE (or about 4 ka BP), when modern estuaries became established in roughly their modern locations. My guess is that at about that time some of the tributaries to the Pamlico Sound estuary were flanked by bottomland hardwood cypress-tupelo swamps. As RSL continued and continues to rise, salinity increased, first killing off the tupelo and other salinity-intolerant trees. In addition to dying sooner, the wood of these trees is less decay resistant than cypress, the last canopy-size tree to go. Tributaries such as the ones where pictures above were taken first saw formation of ghost forests and marsh conversion, eventually reaching a point where nothing is left of the cypress but a few snags (standing dead trees), stumps, and submerged roots and stumps. That process is ongoing and will continue and probably accelerate. Whether there is a net loss of bottomland hardwood swamps will depend on the extent to which they can expand along their inland and upstream margins. In terms of marsh area, the swamp-to-marsh conversion is offset by erosion and drowning. In most areas, including the Neuse River, there has been a net loss. 

*Salinity between 0.5 and 5.0 ppt. Fresh water is <0.5; ocean water is 33 to 35 ppt. 

References cited

Kopp, R.E. & 3 others. 2015. Past and future sea level rise along the coast of North Carolina, USA. ClimaticChange 132, 693–707. 

Miller, W.M., III. 1986a, Community replacement in estuarine Pleistocene deposits of eastern North Carolina. Tulane Studies in Geology & Paleontology 19, 97-122. 

Miller, W.M., III. 1986b. Paleoecology of benthic community replacement. Lethaia 19, 225–231. 

Stahle, D.W. & 3 others. 2005. Ancient Baldcypress Forests Buried in South Carolina. Tree Ring Laboratory, Dept. of Geosciences, University of Arkansas.

Stahle, D.W. & 9 others. 2012. Tree-ring analysis of ancient baldcypress trees and subfossil wood. Quaternary Science Reviews 34, 1-15.

My own related research

Phillips, J.D. 2024. Sequential changes in coastal plain rivers affected by rising sea-level. Hydrology 11, 124. 

Phillips, J.D. 2024. Ghost cypress as indicators of sea-level rise in the Neuse River, North Carolina. Wetlands Ecology and Management 32, 287-302. 

Phillips, J.D. 2023. Landscape change and climate attribution, with an example from estuarine marshes. Geomorphology 430: 108666.

Phillips, J.D. 2018. Environmental gradients and complexity in coastal landscape response to sea level rise. Catena 169: 107-118. 

Phillips, J.D. 2018. Coastal wetlands, sea-level, and the dimensions of geomorphic resilience. Geomorphology 305: 173-184.

GAR TALK

 On recent paddle up Tranters Creek near Washington, NC, the water, particularly in certain reaches, was churning with longnose gar surfacing (often called rolling) and leaping. It was a beautiful day on the creek, and I was reporting on it to my wife Lynn, who was not able to accompany me. When I indicated that I wished she had been there, she indicated that she had no interest in paddling through "gar soup," as I called it. She finds them "icky," and referred to them as primitive, unevolved, gill-less, top-breathing bottom feeders (OK, I am paraphrasing and embellishing here). Though I have nothing against gar I did not argue as I thought it was not an inaccurate characterization.

Gar party on Tranters Creek, NC

But I was wrong.

Longnose gar (Lepisosteus osseus) are indeed often described as primitive, because they have retained some primitive features, such as a spiral valve intestine and ganoid scales. They are olive brown to green in color, and their scaly armoring, torpedo shape, and long, toothy, spear-like snout do make them look like unevolved Jurassic killing machines. It seems they have been around for at least 100 million years, which would place them in the Cretaceous (younger than Jurassic but still dinosaur times), but in the U.S. fossils date only to the Pleistocene. 

Longnose gar in the N.C. Aquarium at Fort Fisher

Unevolved, they are not; they are actually highly adapted to their habitats. They are indeed gill-less. Instead they have a swim bladder that allows them to breathe air or water, which in turn affords them success in waters with low dissolved oxygen (DO), a frequent hazard for fish in the estuaries and swamps of the southeastern U.S. Their surfacing is the air-breathing part, and when they are churning the water as they were recently on Tranters Creek (I have seen similar gar parties in other creeks and swamps) it probably indicates low DO. Longnose gar are not bottom-feeders, or at least not exclusively or primarily. They generally eat small fish, and occasionally crustaceans and insects, wherever they find them. 

Longnose gar distribution in the U.S.A.

The species is mostly freshwater, but has been found in salinities up to 31 ppt, nearly that of ocean water. They are quite common in estuaries of the Pamlico-Albemarle Sound system, where they come out of the freshwater to feed on menhaden. They are found in lakes, ponds, swamps, and slow-moving rivers and creeks. 

                                                    Longnose gar from eastern N.C.

Fun gar factoids:

•The species was first named (though the genus was later renamed) by Carl Linnaeus, the father of biological taxonomy, in 1758. The name Lepisosteus osseus is derived from lepis (Greek for scale) and osteos (Latin for bony). 

•The North Carolina record for longnose gar has been broken twice so far in 2025, first in the Intracoastal Waterway near Coinjock, and then in a quarry pond near Maysville. The world record (25.5 kg) was caught by a bowfisher on Lake Palestine in east Texas. 

•Longnose gar can reach 2 m in length. The ones I see in the swamps of the Carolinas are generally in the 0.5 to 1 m range. 

•Longnose gar are known to breed and hybridize with the even bigger and scarier looking alligator gar (Atractosteus spatula), but that only happens in the Gulf Coast and lower Mississippi/Ohio River valleys. Alligator gar are not found along the SE coast. 

87.5 kg alligator gar from the Trinity River, Texas (https://www.reddit.com/r/Fishing/comments/13n2cte/alligator_gar_193_lbs_trinity_river_texas/#lightbox)

More information at animalia.bio.

WATER TUPELO AND LOGGING LEGACIES

In an earlier post I noted that there are some swamps along the lower Neuse River, NC that are almost entirely dominated by water tupelo (Nyssa aquatica). Water tupelo is a common swamp tree in the area, and is known to dominate some stands, but usually (even when dominant) co-occurs with other species such as bald cypress (Taxodium distichum). But some areas along the Neuse and its side-channels contain almost nothing else in the overstory, canopy-tree layer in the swamp interiors (some cypress occurs along the channel margin, though water tupelo is still dominant). 

The most likely explanation is logging for cypress. Where most or all cypress is removed—which apparently was common when the trees were of good commercial quality—only tupelo (Nyssa aquatica and/or N. Biflora, swamp tupelo) among potential dominant canopy trees in frequently flooded swamps, were left behind as a seed source. Thus tupelo were left standing, and able to reproduce. Some tupelo has historically been cut (in the New Bern/Craven County Public Library I found a few early 20^th century advertisements for “swamp gum” lumber; Nyssa species are also referred to as tupelo gum, black gum, swamp gum, and sour gum)..

Though there is a clear historical record of cypress logging in the region, I have tried without success to learn anything about the specific logging and forestry history of the lower Neuse—i.e., what tracts were cut and when. To confirm the logging interpretation of the tupelo dominance, I visited these water tupelo-dominated swamps again, specifically to looked for sawed-off cypress stumps.

Sawed-off bald cypress stumps (with 2 uncut trunks in the foreground) and regrowth along the Northeast Cape Fear River 

Sawed stumps are a sure indicator of at least some tree harvesting, and as cypress wood is famously decay-resistant, you would expect at least some of all but the most ancient to remain—but no. Cypress logs at the bottom of a river or cypress stumps and roots buried by sediments can indeed last for very long periods, and cypress lumber is exceptionally durable. But cypress stumps in hot, wet swamps of the southeastern U.S. apparently don’t last more than a few decades. I have seen some sawed stumps around the area, but on my recent expedition specifically looking for them in the water tupelo swamps did not turn any up, which is not too surprising if the logging occurred in the early 20^th century. 

Stumpless tupelo swamp where cypress harvesting occurred sometime in the early 20^th century.

Neuse River swamp near Fort Barnwell, NC logged between 2012 and 2014, photographed in 2023. 

Another potential fate for cypress stumps is that they serve as “nurse” sites for tupelo or other trees. As shown below, it is not unusual for downed logs and stumps to be sites for establishment of new trees. In time, it seems possible that a water tupelo base could simply cover a stump as the latter rots away.

Sawed cypress stumps supporting new growth of water tupelo and other trees.

Apparently, however, cypress knees do not decay nearly as quickly. Throughout many of the tupelo stands cypress knees are evident even where no cypress trees are nearby. This clearly shows a cypress-to-tupelo transition, or more likely a mixed cypress-tupelo to nearly-complete tupelo domination of the canopy. 

Cypress knees in water tupelo swamp with no living cypress nearby.

THE FACTORS OF SWAMP FORMATION

 In 1807 Alexander von Humboldt published Essay on Plant Geography. Humboldt made major contributions to botany, zoology, meteorology, oceanography, and anthropology, but is in particular known at the “father of geography.” The Essay is particularly remembered for relating vegetation to climate and elevation (Humboldt studied vegetation on Chimborazo, a 6264 m mountain in Ecuador with tropical rainforest at the base and a snow-capped peak). However, Humboldt (and his collaborator Aimé Bonpland) also linked plant distributions to geology, soils, geophysical phenomena, and human impacts. Earth and ecological scientists ever since (and probably before) have used a “state factor” approach to explain the spatial variation of various phenomena and how and why the plants, soils, climate, or other phenomena in different places differ from each other. It’s a fundamentally geographic logic that says, in essence, if the factors that determine the type of, e.g., vegetation, soil, or climate in different places are the same, then the plants, soil, or climate should also be very similar. 

In the late 19^th and early 20^th century, soil scientists in Russia and a bit later in the U.S.A., formalized the concept, the best-known statement being the so-called “clorpt’ equation:

where S is soil or some soil property, cl is climate, o for organisms indicates biological effects (originally focussed almost entirely on vegetation, r (relief) indicates topography, p is parent material, and t is time, usually thought of as the time available for pedogenesis. The trailing dots indicate the possibility of other factors not encompassed in cl, o, r, p, t that influence soil formation in particular cases, but not necessarily in general (for example sea-level change or contaminants). 

Swamp soil along the Waccamaw River, SC

The state factor conceptual model is to this day the underlying framework for soil mapping and continues to be used, usually implicitly but sometimes explicitly, throughout the ecological and Earth sciences. State factor approaches have been subject to considerable debate in pedology, mainly centered on three general areas. One is the independence of the state factors—clearly they are not independent of each other or of the soil itself, but the debate concerns whether they can be treated as though they are in certain contexts, and how to disentangle their effects. Second is how to actually use factorial models for prediction, given that each state factor may itself have multiple variables to describe it. Third is the association of the state factor model and related ideas with practices in soil taxonomy, pedogenesis studies, and the general praxis of soil science (I don’t like that word because it seems so pretentious, but I didn’t want to use “practice” twice in the same sentence) that are thought to have inhibited other approaches. With respect to the latter, soil state factors have sometimes been presented in opposition to studies of soil processes. The two approaches are actually complementary, with the state factors providing the environmental context and boundary conditions within which processes operate, and processes explaining how the state factors work. 

With that out of the way, what are the factors of swamp formation? 

Let’s start with climate. Though small, isolated wetlands can occur even in deserts if there are local spots where water accumulates and wetness lingers, a full-on swamp requires a humid climate, wet enough to support trees, particularly hydrophytes. Climate is only one factor that influences water availability, illustrating right off the bat the interdependence of state factors. Temperature regimes will also influence swamps—for example, consider the distribution of swamp tupelo (Nyssa biflora) shown below. You don’t find the tree north of central Virginia, or west of the 100^th meridian (which corresponds with the straight line on the left, a general divider between humid and drier climates in the U.S.A.). The concentration on the Atlantic and Gulf of Mexico coastal plains hints at elevation and topographic-related factors—interdependence again. 

Distribution of Nyssa biflora (swamp tupelo) based on U.S. Forest Service Forest Inventory and Analysis data (https://www.fs.usda.gov/nrs/atlas/tree/694).

Another obvious candidate is topography. Swamps will occur where water can collect and persist; at lower elevations, gentler slopes, and in depressions. Closely related, but also linked to factors other than topography and climate is hydrology, including influences such as flow regimes, tides, and groundwater dynamics. 

Differences among swamps are also defined by vegetation. Plant communities are obviously related to climate, hydrology, topography, and soils, but also to other factors such as seed sources, dispersal mechanisms, disturbances (e.g., floods, storms, fires, pests, logging), and ecological interactions among plants and other biota. Perhaps animals could be lumped with plants as an organisms or biota factor, but one can also argue that the presence or absence of critters such as beavers, alligators, nutria, feral hogs, and others (not to mention Homo sapiens) justifies fauna as a separate state factor. 

Swamp along the Black River, NC

Soils are another state factor. In the swamps I have worked and played in the Carolinas and Texas/Louisiana the main differences among soils are associated with the texture (particularly sandy vs. clayey vs. organic and combinations thereof), drainage class (obviously related to hydrology), geomorphic environment (e.g., floodplains, terraces, infilled oxbows, depressions), and presence or absence of specific soil features such as argillic horizons (clay-rich subsoils). Finer distinctions are related to mineralogy and pH. 

Independently of closely related topographic and hydrologic factors, I consider geomorphology a state factor. This includes the formation and modification of specific features such as natural levees, oxbows, and point bars, and the extent to which the setting is erosional, depositional, or both/either. In the classic form of the soil state factor equation, the type of swamp or its characteristics (S_w) is a function of climate (cl), topography (r for relief), hydrology (h), vegetation (v), fauna (a for animals), soils (s), and geomorphology (g):

This is a conceptual framework, not an equation to be numerically solved. For one thing, there are any number of specific variables or indicators that could be associated with each of the factors. For another, as we have seen, all the factors are interrelated and interdependent.

I will explore this framework further in future posts. 

CAVITY SEARCH

Tree cavities are important habitat for many types of wildlife, and the swamps of the southeastern U.S.A. are no exception. Furthermore, some of the trees in those swamps have some hellacious cavities. In the Carolinas these may be home to birds such as wood ducks and other ducks, promontory warblers, chimney swifts, and pileated and other woodpeckers. The apparently extinct ivory billed woodpecker called such cavities home. Cavities in swamp cypress and tupelo trees also host some important bat species such as the big-eared and the mouse-eared bat. River otters, black bear and alligators may also den in such cavities. I have read that beavers may also den in tree cavities, but I have observed no evidence of this, even in swamps with many beaver and many tree cavities. 

This post is basically a photo album of some of the best and most interesting cavity trees I've seen in the coastal plain swamps of North and South Carolina.

Grinnell Creek, NC

Great Pee Dee River, SC

Neuse River backwaters, NC

Black River, NC

Crabtree Swamp, SC

Enterprise Creek, SC

Holly Shelter Creek, NC

Northeast Cape Fear River, NC

Core Creek, NC

Pinetree Creek, NC