Skip to main content



Submitted by jdp on Wed, 09/28/2022 - 06:00 pm

Tupelo gum and bald cypress (Nyssa aquatica and Taxodium distichum) are the main trees you are likely to find in perennially flooded deepwater swamps and stream channels in the southeastern U.S.A. Once established, they can grow in sites that are always inundated. However, they cannot germinate from seeds and establish seedlings in standing water. The substrate must be exposed at least once, at the right time of year, for that to happen. 

So how come you can find cypress and tupelo growing in perennially flooded areas, in standing water, that apparently never dry out? How do they get a start on the bottom of a stream or lake?

Pinetree Creek, an anabranch of the lower Neuse River, N.C.

The obvious answer is that they (or their parents, if they grew from stump sprouts or nurse logs) got their start when the site was not flooded, or always inundated, or that they got their start on some raised spot within the water that is no longer evident.

Here’s the problem, as I see it. I see these underwater-base trees in situations where there appears to be no evidence, or reason to believe, that they have ever dried out, at least within the time span of even two or three generations of trees. They also sometimes occur far from any banks or higher spots, in deep water where there is no extant evidence of channel change. In many cases these occur in systems where no mid-channel bars occur, and though I look for them, I have not yet seen any examples of cypress or tupelo growing on nurse logs or stumps of other species (though Taxodium and Nyssa stumps do frequently survive as germination sites for other species, and stump sprouting does occur, though only occasionally with cypress). Just because I haven’t seen something does not mean it hasn’t happened, of course, but I feel a lot better when there exist empirical examples of proposed or supposed phenomena. 

Lower Neuse River

My observations are based mainly on field observations in the lower Neuse River, North Carolina, and other nearby rivers and swamps, and the lower Waccamaw River, South Carolina. I’ve been in a lot of other low-gradient rivers and swamps, particularly in Texas, Louisiana, and the Carolinas, but not when I had begun thinking about this problem, so I was not on the lookout for evidence pertaining thereto. 

Once a tree becomes established at a site, stump sprouting or germination on emergent (above the water line) nurse logs or stumps is possible. Establishment on exposed rootwads of uprooted trees is also possible (though again, I haven’t seen that for the case of tupelo or cypress). But how does the initial drowned-base tree get there?

Backwater of the Waccamaw River, SC.

Let’s walk through some possibilities.

1. Water drawdown or drying allows seedling establishment, followed by reflooding. This could happen during extreme droughts, due to flow diversions, or due to temporary upstream damming (e.g., by beavers or logjams). I am skeptical of the drought/climate explanation in my field sites, as there is no record of any drought severe enough to dry out some of these features, or of any Holocene climate drying that could account for it. The diversion and damming are plausible; the lower Neuse is teeming with beavers at present. 

2. Initial establishment in discontinuously flooded sites, followed by increased inflow or water level rise. The flooding could occur due to channel avulsions, beaver dams, or logjams. 

Core Creek near Cove City, NC; a Neuse River tributary.

3. Initial establishment on near-bank environments high enough to dry out during dry periods, followed by channel change or migration so that the initially near-bank, shallow water patch becomes a perpetually flooded location. This definitely happens, though it cannot account for some deepwater trees far from any banks. 

4. Shallowing due to sediment deposition, so that substrates become exposed at low water. The trees become established, and this is later followed by erosional flushing, leaving the trees in a permanently flooded condition. I have seen germination, particularly of cypress, on fresh sediment deposits in swamps, particularly after Hurricane Florence. However, I have no evidence of subsequent flushing. 

Mid-channel complex including a large cypress stump and younger trees.

5. Initial establishment on mid-channel bars or islands that are subsequently eroded or drowned. Such islands and bars are rare in the areas I frequent, except those that already support mature trees. 

6. Large woody debris in mid-channel. This could provide nurse logs, or stimulate deposition of bars or islands. I have seen many such features, but none supporting cypress or tupelo recruits. 

Nurse stumps, Core Creek. 

Bald cypress stump supporting a resprouted cypress (left) and serving as a nurse site for a willow tree (right).

Some of these mechanisms definitely occur, and all of them probably occur at least occasionally, though I would feel much better with a few clear field examples. This problem is no doubt relevant to foresters, botanists, and wetlands ecologists, but to me the main scientific fascination is that the presence of these drowned-base trees indicates that some significant geomorphic or hydrologic change or event has occurred. These deepwater trees could give clues or indicators about these changes or events, or at the very least show us where to start looking for other evidence. 


Submitted by jdp on Sat, 09/10/2022 - 12:50 pm

Last year, I wrote about how the warnings about human-accelerated climate change we’ve been hearing (and those of us in the business have been sending) for decades are, unfortunately, coming true. Almost daily, our news feeds remind us of this, or provide new evidence that Earth’s climate, and the environmental systems affected by it, are approaching unknown territory. We are seeing ocean temperatures, ice loss from the great Antarctic and Greenland ice sheets, storm and flood regimes, heat waves, and fires that are unprecedented in human history and in some cases unprecedented in Earth history, period. 

National Weather Service heat warnings for California


Engineering design, insurance, land use planning, and economic forecasting, among other things, are grounded on statistical analyses and risk assessments based on data from the past—the “from the past” would ordinarily be unnecessary, as there are no data from the future. But I wish to emphasize that the data record is becoming increasingly irrelevant to the present and future. What used to be rare and extreme heat waves, tropical cyclones, fire seasons, etc., are becoming commonplace. The hundred-year storm or flood concepts, for instance, essentially apply to the 20th century. Beyond analysis of recorded data, we often rely on deeper historical information such as paleoclimate or paleoecological evidence to guide us in understanding environmental change. 

Flooding in Hyberabad, Pakistan in early September, 2022 caused by a record monsoon attributed to climate warming (Reuters photo—Yasir Rajput)

All this clearly gives useful knowledge and clues about ongoing and future change. But we must temper our interpretations with the knowledge that we are experiencing “new normals” and that the Earth system and its components ain’t what they used to be. I’m comfortable saying that we are in the Anthropocene Epoch. As usual, I am loath to engage in scientific, political, cultural, or semantic debates about terminology (and as there are many others willing to do so, I am happy to let them do it), so feel free to call it something else. But it is real, and it is here. 

If Earth history (at geological to contemporary time scales) as a whole is not representative of what is happening now and likely to happen in the future, what are we to do? First, let me acknowledge that historical understanding has intrinsic value and is therefore important, irrespective of the extent to which it is relevant to our currently changing planet. But here we are concerned with responding to the current crisis. 

A potential answer is to identify and focus on hot spots and hot moments (not necessarily, though sometimes, in a literal thermal sense) where and when specific landscapes are or were responding to changes such as those that are now occurring. It’s a simple concept, and not new, but worth revisiting. 

In studying impacts of heat waves on cities, for example, look for the ones that have exhibited the most pronounced urban heat island effects or that have historically endured extreme heat waves. Focus on ecosystems that were already fire-prone or frequently burned to gain insight into how more frequent fires will affect other ecosystems, or what adaptations nature has come up with that we might mimic. Continue to look at events such as Meltwater Pulse 1A about 14,600 years ago, when rapidly melting ice raised sea levels about 18 feet in 500 years (an average rate more than 10 times greater than at present). Whatever we can learn about coastal responses around 14.5 to 15.1 ka could be a better clue to future responses than other situations. 

Earlier this year I published a study examining the effects of Hurricane Florence in 2018 on the Neuse River estuary, North Carolina. The storm included unprecedented river discharges from upstream and storm surges from downstream. Yet, in the lower Neuse River fluvial-estuarine transition zone (FETZ), geomorphic impacts were minimal, in sharp contrast to other landforms and ecosystems in the region. The reason is that the Neuse FETZ is a complex of various types of channels, wetlands, and water bodies that is perfectly adapted to handle large volumes of water from upstream and/or downstream. As I explored further in another paper, this is because the FETZ has developed under the influence of rising sea-level during the Holocene, constantly subjected to stream discharge coming downstream and sea-level gradually encroaching upstream. 



One of many subchannels in the Neuse FETZ 

The traits of the Neuse FETZ that enabled it to absorb these impacts (and those of previous Hurricanes and floods) include extensive wetlands, very high channel-wetland connectivity, various “spillways” for exporting and storage areas for storing excess water, and water flows and exchanges between these components that can move in multiple directions, depending on circumstances. For the case of river, wetland, and coastal environments and their response to more and more powerful storms and floods and to sea-level rise, this points to the importance of preserving, protecting, and perhaps restoring or rehabilitating wetlands and hydrogeomorphic features that facilitate the key traits and dynamics. It also suggests the importance of multiple degrees of freedom or ways to respond to changes. 

The Neuse example suggests the importance of examining, if you will, cold spots (not at all in the literal thermal sense) where impacts of changes and disturbances likely in the future are readily absorbed. 

By examining the hot spots of climate-driven change and the cold spots of resistance or resilience to climate change we can, hopefully, gain insight as to what to expect and what to do as the climate change shit continues to hit the fan. 



Submitted by jdp on Sat, 08/27/2022 - 03:16 pm

As much as we’d like to think otherwise, the facts (data, analyses, results, observations) do not speak for themselves. As scientists and educators, we are obliged to explain and interpret the facts; to attach meaning to them. As things have come to pass in the scientific world, we are obliged to speak for the facts in English. 

This post was inspired by a discussion posted on by Alejandro Bortolus of the Centro Nacional Patagonico (Argentina): Is the use of English in scientific articles a real need for an international working language, or a sign of long-lasting Colonialism? The lively discussion can be accessed here.

You can’t rely on me for a comprehensive and coherent summary of the comments and reactions, but some key themes are:

•The (obvious) advantages of having a single lingua franca to support global scientific communication. 

•The (obvious) advantages of respecting and preserving local languages and multilingualism, and allowing authors and scientists to communicate at their best, which is usually in our native languages.

•The professional demands that scientific publication be solely or primarily in English, and the adverse impacts thereof (see the bullet point above).

•What could or should be done?

I thank Bortolus for bringing this question to the fore. As I work quite a bit—as collaborator, coauthor, and reviewer—with colleagues who are not native English speakers, it is something I have thought about a lot. 

First, note that I comment from a position of privilege and unearned good fortune. Born, raised, and living my entire life in the USA, I am a native English speaker and writer. I am terrible at languages (somehow, I managed two years of high school French with passing grades but without any fluency whatsoever), and due to the good fortune of my birth and residence (linguistically, anyway), I have been able to dodge the need to learn any other languages. In college, I exploited a loophole that existed in some places in the 1970s that allowed one to substitute a computer programming language for an actual language. Accordingly, I was once fluent is what is now a dead language (FORTRAN). As a scientist and writer, I would be completely helpless if I had to communicate in any language other than English. 

I am strongly sympathetic to non-NES (native English speakers). Knowing how much I often wrestle with the details and nuances of wording to get my points across, and how small, subtle variations in superficially similar phrasings can make a big difference, I worry about what I/we may be missing from non-NES scholars more-or-less forced to publish in English, while recognizing that if the publications were only in Mandarin, Russian, Czech, or Thai, I would miss the whole damn thing. 

I dismiss the issues of colonialism, and of pushing back on the establishment of English as the standard language of science. I dismiss these not because they are not important and legitimate—they certainly are! On colonial legacies, however, I have nothing to say that hasn’t been said before, mainly by people with more expertise than I. And I believe the linguistic hegemony of English in science to be a done deal that we cannot do anything about in the near future. 

Bortolus himself has some recommendations that I agree with:

(1) Non-NES scientists must exercise their legitimate right to write and communicate their ideas in their own language without negative feedback.

(2) International scientific editorials should help non-NES scientists to counteract the loss of valuable local literature, historically considered disposable gray literature, by encouraging their citation and soliciting (through the ‘‘Guide for Authors’’) electronic reprints to archive them as supporting material with open access (a win–win situation). 

(3) Local non-NES scientific institutions and editorials should support more, and explicitly, the publication of books and review papers in local languages to make this information more accessible to laypeople and to promote the engagement of young non-NES scientists in modern local schools of thought.

(4) Leading non-NES scientific journals and editorials must pursue the creation of experienced and attractive editorial boards willing to achieve the highest possible standard ofpublication based on international counterparts. There is no point in favoring publication in local languages if the quality of the resulting papers will be mediocre. 

(5) Balancing the number of publications in English with those in local languages must be on the agenda of all non-NES nations that aim to achieve the sustainable development of local science in communion with society. 

I would add a couple of items. In addition to items 2 and 3, journals should allow electronic archiving not only of background materials, but also of non-English versions of published articles—that is, a published English version could be coupled to a version in another language. 

Second, as referees, reviewers, and editors, we privileged NES need to cut others some slack. Sometimes that may mean lowering the bar a bit with respect to the literary (not the scientific!) quality of a manuscript. Sometimes it means being understanding when a perfectly acceptable but non-traditional term is used (for instance, underground instead of subsurface). Often it means taking more time to get through a manuscript to evaluate its scientific value, rather than recommending rejection because the writing is poor (though sometimes the writing is so poor that the scientific value cannot be reliably assessed). 

Finally, consider taking the time to help non-NES authors correct and polish their English through detailed editing. I know--those of us who review a lot of papers can’t do this every time, and often there is not sufficient time to do this even if you want to. But every now and then, for a piece of work that you consider promising, do it. 


Submitted by jdp on Tue, 07/26/2022 - 04:43 pm

Spoiler alert--the answer is: maybe, but I’m not sure.

Argillic horizons are subsoil layers that are enriched in silicate clays. I have long been interested in soil morphology as it relates to argillic horizons. First, it was with respect to soil erosion. As these horizons are by definition formed below the surface, their exposure at or near the ground surface indicates removal of overlying soil. To the extent soils have a characteristic depth, or range of depths, to the top of the argillic horizon, then variations in DTA (depth to argillic) can indicate erosion or deposition. I used this to study soil erosion in the North Carolina coastal plain and piedmont in the late 1980s and 1990s, and in the Ouachita Mountains of Arkansas in the 2000s and 2010s.

Multiple argillic horizons in a Kandiustult in Zambia (source:

Soils with argillic horizons, by implication, are vertical texture contrast (VTC) soils, where coarser surficial horizons overlie finer-textured subsurface horizons. Such soils, also called duplex soils, are globally common. Along with the erosion work, my studies of soil geomorphology, geography, and spatial variation piqued my interest in how and why argillic (often designated as Bt) horizons and VTC soils develop.

The conventional explanation is that they form due to vertical translocation by percolating water. This water physically washes smaller particles out from between the larger ones (often sand grains) and moves them downward, concentrating smaller, clay-size material in the subsoil. This process is called lessivage or argilluviation. The water can also dissolve material from the upper layers which precipitates in the subsoil.

Vertical translocation by water most definitely occurs, and is in my opinion the single most important process for creating VTCs. But it is far from the only one! Preferential erosion of finer and/or deposition of coarser sediment at the surface can get the job done. Bioturbation often plays a key role, and texture contrasts can by partly or wholly inherited from parent material layering. In-place weathering and clay synthesis can produce silicate clays in subsoils, and upward movement of groundwater can lead to precipitation in the B horizon.

Vertical texture contrast Hapludult from the Ouachita Mountains, Arkansas. The argillic horizon is labelled Bt.

Three things in particular spurred my interest. One was my own findings of variations in DTA over short distances and small areas that could not be explained by any measurable variation in soil forming factors. This suggested to me that either something else in addition to vertical translocation by water was going on, and/or that the translocation process is characterized by complex, perhaps even deterministically chaotic, dynamics. Second was a series of papers from 1987 on into the 2000s by Don Johnson and colleagues showing the strong but often, at least traditionally, overlooked role of organisms in creating and changing soil morphology (e.g., Johnson, 1990). The third was the book Soils: A New Global View by T.R. Paton, Geoff Humphries, and P.B. Mitchell (Yale University Press, 1995).  In this book they challenged the prevailing approach to pedology, favoring a more geological and geomorphological viewpoint. They also specifically challenged the notion of vertical translocation by water as a ubiquitous factor in soil formation. I did not, then or now, buy all their arguments, but they definitely made a strong case for rethinking the conventional wisdom.

So, from 1993 to nearly the present I was involved in, and published a number of papers on, how VTC soils and weathering profiles are formed, complexity in pedogenesis, spatial variation of soils, and coevolution of soils, landforms, and ecosystems (shameless plug: See my book Landscape Evolution. Landforms, Ecosystems, and Soils, Elsevier, 2021).

Soil profile from the North Carolina coastal plain (from Phillips, 2004).

Now cut to the 2010s. I come across an article by William Verboom and John Pate (2013) on ecosystem engineering of soils by eucalyptus trees in western Australia—including the synthesis of clays in the root zone! This led me to some of their earlier work showing that vertical redistribution of water and minerals dissolved therein resulted in formation of dense clay layers. The lateral root systems of the trees were, in essence, forming a subsoil clay layer by bringing the raw materials for clay synthesis into contact with each other. Because these clay-rich horizons benefit the trees via their water and nutrient storage, this is an example of positive ecosystem engineering. Thus, in at least one environment, trees and woodlands, if not forests sensu stricto, can build claypans and argillic horizons. Could this be a more general phenomenon?

Around the same time, I recalled something I had read earlier in Greg Retallack’s masterful Soils of the Past (3rd ed. 2019, John Wiley)—that soils with argillic horizons (Alfisols and Ultisols in the U.S. Soil Taxonomy) do not appear in the paleosol (fossil soil) record until forests appeared in the Devonian. Coincidence? Unlikely. Deeper rooting depths of trees and effectiveness of weathering under forests likely play a role, but Retallack also noted the strongly tapering geometry of tree roots. These create large pathways for water movement in upper parts of the soil that taper down to nearly nothing at their tips, allowing translocated material to move down, but to start building up at the end of the root line, so to speak. This is broadly consistent with my own work, which showed that tree roots (and root paths following death and decay) and insect and other faunal burrows are important in maintaining translocation which might otherwise be reduced to negligible levels as low permeability clays accumulate and soil pores are blocked.

Key events in terrestrial plant evolution and root-soil interactions (from Pawlik et al. 2016).

Admitting that much of the biochemical, geochemical, and mineralogical details are at the edge of or beyond my expertise, I have not found any work other than Verboom and Pate’s and some of their coworkers that directly address the question posed in the title. However, I did come across Pierre Velde and Pierre Barre’s book: Soils, Plants, and Clay Minerals. Mineral and Biologic Interactions (Springer, 2010). Velde and Barre come from a school of thought that I was not much aware of before, that soil clay minerals are fundamentally different from those formed below the soil or prior to soil formation by primarily geochemical processes.

Independently of this perspective, however, they show the role of plants and vegetation-based soil organic matter in the formation and retention of phyllosilicate clays, both directly and via their bacterial and fungal symbionts. Particularly important is plant uplift (via water intake) of silica and potassium, key building blocks of silicate clays. An important point is that clay minerals are a necessary source of the critical nutrient K (potassium), so facilitating the formation of clays that can retain K is of great benefit to the vegetation.

Without plants, Velde and Barre assert, there would be no clay accumulation in surface layers. However, their book does not directly address the “can trees make argillic horizons” question. First, they are concerned with plants in general, and grasses may be more effective clay-formers than trees, at least when it comes to clay in A horizons. Second, much of their work is indeed concerned with A horizon clay—it is in these layers that most root mass occurs, after all. They are less concerned with clay migration to the subsoil, though they do note that loss of clays from the surface layer is highly probable in forest soils. Finally, the book has more to say about retention of clays and clay minerals than the (flora-assisted) formation thereof.

In subtropical savannas of south Texas, there’s a body of work on vegetation relationships to soils that doesn’t quite fit into the trees-make-argillic horizons framework. Large woody patches occur on soils without argillic horizons, whereas smaller patches with more herbaceous vegetation is found in adjacent sites in the same landscape where an argillic horizon is present (Midwood et al., 1998; Zhou et al., 2017). This could occur because the clayey layers restrict root penetration, favoring more shallow rooted grasses and shrubs rather than trees. One study showed that shrubs on argillic soils had less aboveground and greater belowground root mass than those on non-argillic soils. Root biomass and density on argillic soils was elevated at shallow (< 0.4 m) depths, whereas root density of the same species on non-argillic soils were skewed to depths >0.4 m (Zhou et al., 2019). Obvious relationships exist between the presence or absence of argillic horizons, root depth and biomass, vegetation-driven water use (including hydraulic lift), and soil hydrological properties (Zou et al., 2005), and similar results have been obtained in Australia (Yunusa et al., 2002). However, this body of work does not make it clear (at least to me) whether soil morphology is driving vegetation distributions, or vice-versa—or both, via reciprocal interactions. It is also possible, I think, that the argillic horizons may be inherited from earlier, moister climates (this is sometimes the case even in desert Aridisols), and poorly related to contemporary pedo-ecological dynamics with respect to argillic formation.

So, can trees & forests build argillic horizons? Yes, but . . . .

Yes, they have been shown to do so in specific situations, but this has not been demonstrated as a widespread, general phenomenon in forests.

Yes, plant uplift of water, nutrients, and Si can help retain or even form silicate clays. But, this is not restricted to trees or woody vegetation, and does not necessarily result in subsoil clay concentrations.

Yes, argillic horizons are strongly associated with forest cover, or with environments where the natural vegetation cover is mainly forest. But, VTC soils do occur in other environments, and clay synthesis is not the only mechanism by which forest cover could facilitate argillic horizon formation. Also, while forest soils (other than Histosols or other wetland soils) typically show evidence of vertical translocation, argillic horizons (even incipient ones) are not always found—sandy Spodosols or podzols are a prominent example.

Forest soils from three forest preserves in the Czech Republic (from Šamonil et al., 2020).


Yes, VTC soils with argillic horizons were rare or absent before the Devonian advent of trees. But, clay synthesis by the trees may not have played a major role.

Yes, forest cover clearly facilitates formation of argillic horizons. But, if sufficient clay-size material is present in parent material, it can become concentrated in a Bt horizon without any clay synthesis (by trees or otherwise) in the soil—some of my own work in coastal plain soils showed this (Phillips, 2007).

Yes, formation of VTC soils with argillic horizons plays a role in the formation of store-and-pour soil hydrology structures that are advantageous to plants. Any role of plants in creating such morphology would by positive ecosystem engineering and niche construction or reinforcement. But, this is something I am still working on and thus somewhat speculative at the moment.

Do forests build vertical texture contrast soils with argillic horizons? Evidence strongly supports the possibility. But we still need some specific case studies to show that it indeed happens, and to shed more light on how. I’m betting the answer is yes, and look forward to others proving me right—or wrong.


Johnson, D.L., 1990. Biomantle evolution and the redistribution of earth materials and artifacts. Soil Science 149, 84 – 102.

Midwood, A.J., Boutton, T.W., Archer, S.R., Watts, S.E. 1998. Water use by woody plants on contrasting soils in a savanna parkland: assessment with delta H-2 and delta O-18. Plant and Soil 205, 13-24.

Pawlik, L., Phillips, J.D., Šamonil, P., 2016. Roots, rock, and regolith: biomechanical and biochemical weathering by trees and its impact on hillslopes - A critical literature review. Earth-Science Reviews 159: 142-159.

Phillips, J.D. 2004. Geogenesis, pedogenesis and multiple causality in the formation of texture-contrast soils. Catena 58: 275-295.

Phillips, J.D., 2007. Formation of texture contrast soils by a combination of bioturbation and translocation. Catena 70: 92-104.

Šamonil, P., Phillips, J.D., *Danĕk, P., Beneš, V., Pawlik, Ł. 2020. Soil, regolith, and weathered rock: Theoretical concepts and evolution in old-growth temperate forests, central Europe. Geoderma 368, 114261.

Verboom, W.H., Pate, J.S., 2006. Bioengineering of soil profiles in semiarid ecosystems: the “phytotarium” concept. A review. Plant and Soil 289, 71e102,

Verboom, W.H., Pate, J.S., 2013. Exploring the biological dimensions to pedogenesis with emphasis on the ecosystems, soils, and landscapes of southwestern Australia.Geoderma 211–212, 154–183.

Verboom, W.H., Pate, J.S., Abdelfattah, M.A., Shahid, S.A., 2013. Effects of plants on soil-forming processes: case studies from arid environments. In: Shahid, S.A. (Ed.), Developments in Soil Classification, Land Use Planning and Policy Implications: Innovative Thinking of Soil Inventory for Land Use Planning and Management of Land Resources. Springer, Dordrecht, pp. 329e344.

Yunusa, I.A.M., Mele, P.M., Rab, M.A., et al. 2002. Priming of soil structural and hydrological properties by native woody species, annual crops, and a permanent pasture. Australian J. Soil Research 40, 207-219.

Zhou, Y., Boutton, T.W., Xu, X.B., Yang, C.H. 2017.  Spatial heterogeneity of subsurface soil texture drives landscape-scale patterns of woody patches in a subtropical savanna. Landscape Ecology 32, 915-929.

Zhou, Y., Watts, S.E., Boutton, T.W., Archer, S.R. 2019. Root density distribution and biomass allocation of co-occurring woody plants on contrasting soils in a subtropical savanna parkland. Plant and Soil 438, 263-269.

Zou, C.B., Barnes, P.W., Archer, S., McMurtry, C.R.  2005. Soil moisture redistribution as a mechanism of facilitation in Savanna tree-shrub clusters. Oecologia 145, 32-40.

Questions or comments:



Submitted by jdp on Tue, 07/12/2022 - 04:05 pm

Some incomplete thoughts and notes on Earth surface system (ESS) evolutionary pathways, focusing on how to think about the enormous variety and large number of possibilities.


ESS encompasses geomorphic and soil landscapes, hydrological systems, and ecosystems. There exists a huge variety of them on our planet. Assuming we could ever inventory or even estimate them all, we can define NESS  as the number of ESS. For each of these multiple possible evolutionary pathways exist. So we define

Ni(p) = number of possible evolutionary pathways for each of i = 1, 2, . . . , NESS.

Image credit:

At any given point in history there were multiple potential evolutionary possibilities, such that Nglobal(p) = number of total possible pathways = Σ Ni(p). However, only one history has occurred for each individual ESS, so that the number of actual past pathways now manifest = NESS.

Canalization is the phenomenon whereby the occurrence of a specific evolutionary trajectory or event constrains (and thus channels or canalizes) future possibilities. For instance, prevention or non-occurrence of fire forecloses post-fire forms of ecological succession unless or until a fire eventually occurs. Incision of stream channels in certain locations restricts or prevents their formation at other locations within the system, and directs fluxes of water, sediment, and nutrients. If Ni(f) is the number of possible future evolutionary pathways, Ni(f) < Ni(p) due to canalization. We could use this to define a canalization ratio, and even conceptualize a global ratio.

Canalization could result in either problems or opportunities by constraining and guiding evolution.



Submitted by jdp on Fri, 07/08/2022 - 09:11 am

In ecological systems, structural redundancy refers to the extent to which more than one species (or taxanomic group) can perform a given function or play a given role in the system. Microbial communities or ecosystems, for instance, tend to have high structural redundancy at the species level, as there usually exists multiple bacteria or other microbes that can, say, break down specific forms of organic matter, reduce iron, precipitate calcium, or what have you. Systems with a single keystone species have low redundancy, at least with respect to whatever the keystone organism does (if something else could perform the same function, then it would not be a keystone). Redundancy tends to be inversely correlated to the degree of biotic specialization, and directly related to ecosystem resilience. 

Tasnuba Jerin (Jerin, 2021; Jerin and Phillips, 2020) linked redundancy to biogeomorphic ecosystem engineering via the concept of biogeomorphic keystones and equivalents. Biogeomorphic keystones perform a unique role such that their removal from or addition to a geomorphic system results in a transformation or state change. Equivalents are different taxa that can do the same biogeomorphological job, and indicate redundancy. Many different kinds of grass or tree, for instance, can provide more or less similar functions with regard to erosion resistance or substrate stabilization. However, no other species performs the same biogeomorphic role as the biogeomorphic keystone species Castor canadensis (beaver). The same species may be a keystone in some settings but not others—for instance, in Tasnuba’s work Platanus occidentalis (sycamore) is a keystone in some bedrock streams where it forms unique pools, but generally not otherwise. 

Adorable baby beavers on their lodge in the lower Neuse River floodplain, North Carolina.

But let’s take a look at structural redundancy in geomorphology more generally. In some cases it is not very fruitful to think about redundancy. For example, aeolian sediment transport is accomplished only by wind, and fluvial erosion only by flowing water, and the disappearance of those processes (or their initiation where once absent) would be part of a larger story that is not enriched by thinking about redundancy. One could, of course, think about, say, sediment transport and identify multiple processes that can accomplish that—mass wasting, flowing water, waves, wind, ice. But while the transitions among sediment transport process dominance and the possible combinations thereof are interesting and important, I can’t see how thinking in terms of structural redundancy helps our cause (but I am prepared to be convinced otherwise!). 

A possible exception is weathering processes. Carbonate dissolution in karst, for example, has high biogeomorphic structural redundancy. Hydrocarbonate dissolution requires a CO2 subsidy from soil and ground-dwelling biota. But any respiring plant, microbe, or animal can accomplish that. While the amount and rate of CO2 subsidy matters, it makes no difference which organisms provide it. The same goes for bacteria that process iron (oxidation or reduction), as multiple species can perform that function (biogeomorphic equivalents or functional groups). 

Sticking with karst, in central Kentucky an important process is woody root penetration into limestone joints and fractures, where a combination of microbial activity in the rhizosphere, formation of organic acids, funneling of water into the rock, and pressure exerted by root growth helps break down the rock. Only two species (in that region) seem to do this ubiquitously—Quercus muehlenbergii (chinquapin oak) and Platanus occidentalis. This process thus seems to have limited structural redundancy, as root penetration of rock by other plants is less frequently observed and their root growth within the rock is less extensive (Phillips, 2016 and this).

Chinkapin oak displacing limestone in central Kentucky.

Sycamore root in bedrock joints, central Kentucky. 

This issue may become more urgent as climate changes, and biogeography along with it. In the U.S. Atlantic and Gulf Coastal Plains, for instance, bald cypress (Taxodium distichum) fulfills several geomorphic roles (in addition to its ecological roles) that few or no other species can fulfill. Evidence suggests that sea-level rise is resulting in gradual replacement of cypress with tupelo gum (Nyssa aquatica) as salinity and tidal conditions encroach further upstream (e.g., Peterson and Li, 2015; Tallie et al., 2019). The trees are similar in some respects--both are capable of growing, once established, in perpetually standing water but both require non-inundated conditions for seed germination and seedling survival; both typically develop wide buttressed trunks. But they differ in geomorphically significant ways. Taxodium produces above-ground roots (cypress knees) that are very effective for erosion protection in some settings, and is more prone to uprooting. Nyssa decays faster, is more likely to sprout from stumps or form coppiced trunks, and is more prone to breakage (as opposed to uprooting), all of which are important with respect to large wood and organic matter dynamics. Biogeomorphic structural redundancy is thus quite low in lower coastal plain floodplain swamps. 

Tupelo gum swamp, Turkey Quarter Creek, North Carolina


Jerin, T. 2021. Scale associated coupling between channel morphology and riparian vegetation in a bedrock-controlled stream.Geomorphology 375, 107562.

Jerin, T., Phillips, J.D. 2020. Biogeomorphic Keystones and Equivalents: Examples from a Bedrock StreamEarth Surface Processes and Landforms 45, 1877-1894.

Peterson, A.T. & Li, X. 2015. Niche-based projections of wetlands shifts with marine intrusion from sea level rise: An example analysis for North Carolina. Environmental Earth Sciences 73, 1479–1490. 

Phillips, J.D., 2016. Biogeomorphology and contingent ecosystem engineering in karst landscapes. Progress in Physical Geography 40: 503-526.

Taillie, P.J.,  et al. 2019. Decadal-scale vegetation change driven by salinity at the leading edge of rising sea level. Ecosystems 22, 1918–1930. 

Comments or questions?


Submitted by jdp on Tue, 06/28/2022 - 02:41 pm

Just published in Earth Surface Processes and Landforms (vol. 47, p. 2044-2061): Geomorphology of the fluvial-estuarine transition zone, lower Neuse River, North Carolina. The abstract is below and the article is attached. 



Neuse FETZ .pdf (31.27 MB)


Submitted by jdp on Sun, 05/22/2022 - 05:58 pm

Just published in Catena: Store and Pour: Evolution of Flow Systems in Landscapes (vol. 216, paper number 106357). The abstract is below, and the article is attached.

This continues my effort to figure out why certain "optimal" configurations appear recurrently in nature, despite the fact that most environmental entities have no intentionality, and that these must be emergent phenomena--accidental agency, if you will. This recognizes some similarities in the development of flow systems in terms of dual-porosity in soil and groundwater (preferential flow patterns and matrix), dendritic fluvial channel networks, and other hydrological (and related geomorphological and ecological) phenomena. It's really pretty simple, as reflected in the figure below (Fig. 4 from the published paper).


store&pour.pdf (12.41 MB)


Submitted by jdp on Tue, 05/10/2022 - 05:43 pm

Juncus romerianus or black needlerush is a graminoid plant that grows in coastal marshes from Virginia down the southeastern coast and around the Gulf of Mexico to Texas. It has high salinity tolerance and is often found in salt marshes, but can grow in near-about fresh water and everything in between. It has no direct, consumptive economic use that I know of (though in pre-industrial times it was used as a needle; hence the common name). However, anecdotal evidence from my neck of the woods suggests that it is highly resistant to erosion and perhaps a good candidate for “living shoreline” erosion control and wetland restoration. 

Juncus roemerianus near my house.


A quick-and-dirty literature search didn’t turn up anything on marsh fringe erosion or erosion resistance focused on needlerush, though there is plenty of field and experimental evidence of its efficacy in trapping sediment and promoting deposition. I have seen it eroded away where it becomes undermined, and the surface layer it is rooted in collapses. 

However, anecdotal observational evidence shows that it is quite resistant to erosion, often persisting even as adjacent coastal landforms and vegetation is eroded away. 

The evidence comes from the Neuse River estuary, North Carolina. One example is at Fisher’s Landing near New Bern. The three photos below show three small patches of Juncus romerianus before, immediately following, and about a year and a half after Hurricane Florence in September, 2018. In the first of the sequence the Juncus patches (which according to Google EarthTM images had been there for at least 20 years previously) are highlighted by rectangles. Note the end of a granite boulder rip-rap at the bottom. The second is from a few days after Hurricane Florence. While the approximately 10 m tall bluffs have retreated 10 m or more laterally, the needlerush patches remain (the middle, small one is hard to see in this image, but it is there). And they are still there in the later image. Note that you can’t infer too much about the beach widths shown. Water levels here are highly sensitive to wind, with water levels going way out during strong SW winds, and coming way up in strong NE winds, so visible changes don’t necessarily reflect any erosion or accretion. The middle image does show a temporary increase in beach width due to onshore transport of sand during Florence, and sand added from erosion of the adjacent bluffs. The new wide beach was gone in two years. 

Fisher’s Landing, February 2017 (Google EarthTM image)

Fisher’s Landing, September 2018 (Google EarthTM image)

Fisher’s Landing, March 2019 (Google EarthTM image)

Now we go downriver a few km to a patch of Juncus near my house. It has been there since at least 1985 (the earliest photo I could find with sufficient resolution). I have personally observed it since 1990, and it never appeared to casual observation to significantly increase or decrease in size. I recently measured it with a survey tape; its area was right at 26 m2. I am not confident comparing field measurements to areal measurements from imagery, but if anything, the size of the patch has increased. Estimates using Google Earth’s polygon measuring tool on images from 1993 to 2019 give results of about 15 to 22 m2

Juncus patch, showing the obvious sediment-trapping effects.

Part of my elite field research team.

The needlerush patches survived Florence, but during the phenomenal 4 m storm surges then the Juncus was totally submerged. But they survived battering during the rise and fall of the storm surge, and at least seven other hurricanes, and countless northeasters. Nearby (within 30 m) dense patches of sawgrass (Cladium Jamaicense) were wiped out by Florence, and the common reed (Phragmites australis)—a famously, some would say notoriously, resistant marsh plant--that replaced it has been eroded back in subsequent northeasters, As the needlerush can obviously resist erosion, it raises for me the question of why the patches have not been able to expand to any great extent.

Juncus patch at low water.

My shoreline fringe wetland observation station. 






Submitted by jdp on Thu, 05/05/2022 - 08:17 am

In 2013 I developed and published something called the flow-channel fitness model (FCF; Phillips, 2013; attached). Fitness refers to the fit between channel size or conveyance capacity—yes, it’s a problematic concept, but a venerable one in hydrology and geomorphology. Underfit channels are “too large” for the range of flows they typically convey. They often occur where large channels and valleys were formed during previously wetter climates, or by megafloods or glaciers, with those big ‘ol channels now occupied by smaller streams that rarely overflow their banks and can’t do much to reshape the channel. Overfit channels are “too small.” They frequently can’t hold all the discharge that comes their way and flood frequently. Fit channels, at least as conventionally conceived for alluvial channels in humid climates, have a reasonably good match. They flood (on average and according to the conventional wisdom) every year or two but otherwise hold their water.

The FCF model is a conceptual and practical model for predicting the qualitative response of alluvial channels to modifications of flow regimes—that is, whether channels experience aggradation, degradation or relative stability, and whether aggradation or degradation is dominated by width or depth. The model is based on transitions among seven possible fitness states, triggered by key thresholds of sediment supply versus transport capacity and shear stress versus shear strength, and requires that potential changes in sediment supply and water surface or energy-grade slope also be accounted for. The summary table and diagram from that paper are shown below. 

I stand by the FCF model, at least in the context for which it was devised, and recognizing that while criteria are simple (fitness, sediment supply vs. transport capacity, shear stress thresholds) they are hardly easy to deal with in terms of measurement or change over time and space. I decided to revisit it, however, to ignore the problematic fitness concept (in part because I’ve dealt with so many lower coastal plain streams that flood frequently under any circumstances and for which overfit and fit are irrelevant). While the FCF is essentially a state transition model, recognizing and (ideally predicting) changes among states, I wanted to think a little more about some of the interrelationships within the system and possible changes. 

The first iteration is shown below. Where sediment supply exceeds transport capacity, the key question is whether there exist opportunities for alluvial storage in channels, subchannels or abandoned channels (oxbows, sloughs, etc.), or floodplains. If so, aggradation occurs.

Flow-channel fitness model (see attached article)

Channel evolution model 1.0

Infilling sub-channel, Navasota River, Texas

If insufficient storage is available, the channel begins to clog with sediment. This can be the first step of infilling, or can trigger feedbacks such as avulsions, cutoffs, or situations where increasingly large flows are confined within ever-higher banks until thresholds are exceeded and erosional stripping occurs. If sediment supply is less than transport capacity, the available sediment is transported through the reach, either with or without channel erosion, depending on whether the mean boundary shear stress of flow exceeds the critical stress for channel and bank erosion (this also applies to aggradation with or without erosion, but I didn’t want to make the flow chart too busy at this point). 

True steady state in alluvial rivers is rare and transient, but a situation with no significant net erosion or aggradation can appear to be in steady state. 

Eroding channel, Board Camp Creek, Arkansas.

Apparent steady state (no net erosion or aggradation), Guadalupe River, Texas. 

Now let’s think about what aspects of the system are most likely to change as a result of external (to the fluvial system) factors such as climate change, sea-level rise, dams, flow diversions, and modifications to watershed runoff and erosion due to climate, land use, or vegetation change. These are highlighted below.


Channel evolution model 1.1, with factors most vulnerable to external change highlighted.

But wait. Aggradation can occur without or with channel erosion (for example, point bar/cutbank pairs on meandering alluvial rivers. 

Aggradation with erosion, Sabine River, Louisiana/Texas.

Aggradation without erosion, also Sabine River. 

Channel evolution model 2.0, including aggradation with or without erosion. 

But wait. Again. We have not considered the internal feedbacks in the system. The version below includes some of these. Eroding channels as well as cutoffs and avulsions affect sediment supply. Aggradation influences sediment storage capacity, and channel erosion affects the shear stress ratio on both sides—by modifying channel geometry and thereby affecting shear stress, and by potentially exposing materials of differing resistance (and also removing or modifying bank or channel vegetation or debris). This is reflected below. 

Channel evolution model 3.0. Dashed lines show internal feedbacks. 

So what now? Where do we go from here? I don’t know, at least not yet. I am putting this out there in hopes that it may help others think this through, in case I don’t get around to pressing it forward. 


rra2602.pdf (1.13 MB)