,
Link to CT DEEP Home LONG ISLAND SOUND RESOURCE CENTER
A Connecticut Department of Energy & Environmental Protection and University of Connecticut Partnership
*
Barn Island Wildlife Management Area - Sentinel Monitoring

`

Ditched Marsh Module - Eight Decades of Change.

The natural marsh module reveals a landscape and hydrology that is very different from the ditched marsh and that module establishes the reference marsh against which to measure post grid ditching changes. The reverted marsh module shows the capacity of ditched marshes to return to the natural marsh and "revert to a natural equilibrium" (Miller and Egler 1950). Also, the colonial ditches with their ditch side levee-ditch plugs are the earliest evidence that drained marshes revert to the levee and basin topography.

Much is written about the effects of ditches upon tidal marshes (Tonjes 2013) but the long term effects are not adequately described in the literature. Today the tendency is to attribute changes in marsh vegetation soley to accelerating sea level rise and almost nowhere is the role of aggrading ditches in marsh change recognized or studied. Described here are a series of changes caused by ditching not described in the literature. Grid ditching, with spacing of ~100 feet, connects all of the pannes to the tides. The flooding pattern switches from over levee flooding at the creekbanks to flooding the marsh first from the upstream end or head of the ditches and then the flooding advances in a creekward direction (Niering and Warren 1993). The combined volume of water needed to fill the basins (pannes) is enormous and it causes a drop in the height of high water in the tidal creeks such that the levees cease to flood shrink over 5 or more decades. No literature could be found that measures the impact of ditching upon the tidal elevation. The basis for concluding that there is a drop in tidal elevation is deduced from indirect changes that can only be explained by a drop of high water.

Most studies of the ditched marshes show that the water table position is largely unaffected by the rise and fall of the tides except along the ditch margins. Yet the study by Stearns et al. (1940) revealed a continuous decrease of the water table position over 3 years of a 3 year study. In this same study, the authors observe a loss of 2 cm of elevation over a two year period.

From inception of grid ditching to the stage of reversion to natural marsh, 3 stages of change are recognized:

  • draining phase - the earliest stage where the basins/pannes are drained and the soils become drier allowing for the colonization of the pannes by Spartina patens which becomes the dominant plant community; lowering of high tide in the creeks allows Juncus gerardii to replace S. patens on the levees; shrinkage of the wide and continuous levees begins with the initiation of ditching; draining phase is estimated at a duration of ~2 decades
  • aggrading ditch phase - the ditches begin to aggrade at the inception of ditching and gradually accumulate sediment (from the landward end towards the creek) but this process continues for many decades; S. patens gradually reverts to the panne sequence; levees continue to contract in a creekward direction and former levee is replaced by stunted S. alterniflora or forb panne; duration ~5+ decades; Juncus on the levee is replaced by S. patens
  • ditch plugging phase - as the aggrading ditch gets shorter, draining of the former basin ceases and over creek and ditch flooding become dominant and there is new levee growth; levee growth occurs along the creekward end of the ditches and on an aerial the ditch plug resembles an arrowhead just like those associated with the colonial ditches.

History of Ditching

Barn Island was grid ditched in 1931 to 1932. Post construction, the state mosquito control program inspected the ditches in the town of Stonington and accepted the responsbility to maintain the ditches under state statutes for mosquito control. The initial ditching disposed of at least some of the excavated peat upon the ditch banks as Miller & Egler (1950) report the presence of 'fence rows' of marsh elder (Iva frutescens). Thirty plus years later, Coleman (1978) still finds some remant fence rows - a testament to how slow the marsh changes. Some minor ditch mainteance by hand was observated by Coleman (1978) on the Palmer Neck Marsh. In 1979, the Mosquito Control Section of the Department of Health Services, took possession of a low ground pressure amphibious, track driven rotary ditching machine. This machine pulls a ditching device through the ditch that operates like a snowthrower. The peat is broadcast over a wide area of marsh which is quickly incorporated into the marsh with no apparent change in vegetation. Most of the ditches were open and functioning and so maintenance ditching would not have caused the impacts that the initial ditching caused. This likely delayed the reverting stage. The Mosquito Control program abandoned maintenance ditching in 1985 in favor of using non-tidal ditch and pond technology called open marsh water management. In the case of the reverted Sassafras Marsh, the ditching did not reclean the ditch where newly formed levee occurred on the western shore. This likely explains why the western levee is wider than the eastern levee.

Hydrological Changes:

Redfield (1972) describes the effect of ditching as causing overdrainage of the marsh. In an experimental manipulation of a tidal marsh through the intentional creation of mosquito ditches, Stearns et al. 1940 observed the loss of two cm of elevation in two years and the continuous drop of the water table over the course of the 3-year study. At some point in time the effect of tidal changes in the ditches ceases to cause changes in the water table on the high marsh. How many years or decades did it take create this condition?

Tidal Height:

No literature has been found that describes the pre- and post-ditching tide conditions of internal tidal creeks and there are no long term tidal observations related to how, if any, the tides change within a marsh complex as ditches aggrade, a process that occurs over many decades (Wilson et al. 2014). The tides at Barn Island are semi-diurnal and microtidal. The main valley marsh areas are flooded by four primary creeks, and each probably has a unique tidal signature as they flood and drain marshes of different sizes. The small valley marsh 4, is drained by a secondary creek associated the primary creek flooding valley marsh 3 (see figure 1 in the natural marsh module).

Five observations at Barn Island point to a lowering of the elevation of high tide post ditching within the primary and secondary tidal creeks:

  • the wide creekbank levees (up to 55 meters in width) shrink in a creekward direction over 5 or more decades,
  • Miller and Egler (1950) find the levees in 1947 supporting a grassland dominated by Juncus gerardii instead of S. patens; Juncus is intolerant of flooding with polyhaline water; new levee growth at Barn Island supports nearly pure S. patens,
  • Miller and Egler (1950) state that the levee height ranges from 30 - 45 cm (do the post ditching levees appear higher as a result of a lower high tide?), much higher than the 13 cm reported at Barnstable Marsh (Redfield 1972); Temmerman et al (2004) model shows the equilibrium height is 20 to 30 cm,
  • Miller and Egler (1950) describe an unusally high tide below dike 1 in 1947 and the entire marsh is flooded except for creekbank levees that project above the water, and
  • the bayfront levee marsh known as Wequetequock Point support S. patens before and after ditching; bayfront levees and not affected by changes to the tides in the creeks (see core#5 in figure 3 of Niering and Warren 1993).

Figure 9 in the Natural Marsh Module, shows the flooding frequency of the ditched marsh versus natural marshes with levees (Mamacoke Marsh and Wequetequock Point). The low frequency of flooding of the natural marsh is a function of the levee height which is ~25 centimeters higher than that of the former basins. So the tides need to exceed the height of the levees before flooding can occur. In the grid ditched marsh, the tides flow from the creek into the ditches and then the water flows upstream until the lowest elevation of the marsh is reached (often near the upland) and then the basin begins to flood and flood in the creekward direction toward the levees (Niering and Warren 1993).

Hydroperiod:

LeMay (2007) observed in the Rowley marshes in Massachusetts that the back marshes of the ditched marsh is 10 cm lower than the same habitat in the natural marsh. In what appear to be the levee-basin (?) marshes, Wilson (2014) found that these marshes are tracking sea level rise. Over a period of 7 decades, total ditch length has decreased 33% has decreased in the marsh and natural creeks have lengthed. Pools have also increased and when a creek incises into a pool, the pool is drained and quickly colonized by S. alterniflora. Wilson (2014) found the sedimentation rates in the former pools to be 5 mm/yr, otherwise these marshes with reverted hydrology are tracking sea level rise.

The Mamacoke Marsh in Waterford, CT, with levee and basin topography is tracking sea level (Carey et al, 2015) but the ditched Barn Island marshes are not tracking sea level rise. Many ditched marshes are not keeping pace with sea level rise, but if the high marsh areas of ditched marshes are reverting back to basins, then while it is true that surface elevation growth is not keeping pace with sea level rise but that is not the entire story. There are no hydroperiod studies at Mamacoke or Barnstable unditched marshes, these are priority areas for the study of natural marsh hydroperiods as well as reverting marshes. The other critical elements to these investigations is surface elevation changes, sedimentation rates and understanding how the basins flood. Are there gaps or tiny creeks that drain and flood levee and basin marshes (Temmerman 2004)?

Niering (1961) decribes an event that causes the prolonged flooding of the basin at Mamacoke. The flooding was attributed to dredging but it may have been an extreme flood caused by a nor'easter in March of that year. The basin was flooded, evapotranspiration increased the salt content and about 25 of the stunted died. Evidently the basin does not drain and this might create the longest hydroperiod, wherein most if not all of the sediment transported into the basin remains on the marsh. Carey et al. (2015) suggest that the tropical storms and nor'easter of 2012 and 2014 introduced sediment to allow the marsh to catch up with sea level rise. On the other hand, the overflooding of the basin may have created the sediment trap needed to catch up to sea level rise. The Thames River is a tidal estuary and lacks the tidal river floods and TSS present on rivers likes the Connecticut and Housatonic Rivers.

Miller & Egler (1950) told us that ditches are "causing far more fundamental and deep seated changes by affecting deposition and erosion, and thus regulating level changes in every square foot of the marsh".

Creekbank Levee Shrinkage

Figure 1 contains snapshots from a time series of aerial photographs that show the long-term shrinkage of the creekbank levees. In these photographs the lightest pattern corresponds to the highest elevations which include the levees and wetland adjacent to the upland. In the 1976 vegetation map, the black arrow points to color pattern that corresponds to levee. The pattern landward of the levee is Disticlis spicata with bare spots. This corresponds to former levee that is getting wetter. By 1987, Distichlis becomes forb panne. This link is to a photograph taken by Dr. Niering from the creekbank of lower Palmer Neck marsh looking westward. Forb panne is in the foreground and stunted Spartina alterniflora is in the background. The vegetation map also shows isolated patches of former levee habitat incside the Distichlis with bare spot vegetation.This is an example of fragmentation and degradation of the levee habitat.

There is virtually no levee growth post ditching but note in 2012 image, just below the dike 1 is a colonial ditch showing new levee growth supporting S. patens. This accelerated growth (?) may be the result of the jetting action of the culverts suspending sediment.The wider levees at the bottom of the 2012 image likely reflects sediment input from the bay.

Figure 1. Lower Palmer Neck marsh. Photos top row - L to R: 1932 and 1951; bottom row - L to R: 1976 and 2012. The 1976 image is the plant community mapping of Coleman (1978) - the arrow points to the color which corresponds to the levee. In the aerial images, the black line corresponds to the present day boundary of tidal marsh. The yellow lines show the location of mosquito ditches. All of the images shown here were georeferenced. The clipping process is imperfect and thus the scales of these images are are similar but not identical.

In the table below, the width of the levee is measured perpendicular to the creekbank at a point midway between ditches. The ditches form the north and south boundaries of a rectangular segment of marsh called a ditch panel. The panels are numbered south (1) to north (8; downstream to upstream). Blank cells in the table reflect locations where the entire panel is dominated by levee panel 1 or in the case of panel 8 the levees are associated with a colonial ditch, not quite perpendicular to the tidal creek. Map errors were not computed and so the values shown here are estimates. Levee loss between 1934 and 1976 ranges from 50 to nearly 100%. By 2012, there levee growth but it is quite limited.

Table 1. Width of the creekbank levees (meter) over time.
Ditch Panel # 1934 1951 1976 2012
8
colonial ditch
7
37.6
5.8
9.3
8.8
6
37.1
6.4
1.0
13.5
5
39.8
7.4
6.8
6.1
4
46.5
36.0
3.9
10.9
3
52.8
32.8
25.9
26.9
2
61.8
38.5
15.1
22.6
1
all levee

The widths of the levees panels 1 to 3 are very wide and this likely reflects a contribution of sediment from the bay.

Salt Marsh Vegetation Change:

The focus of this section is about the vegetation changes in a salt marsh complex, where the dominant vegetation contains plant communities associated with soils whose water chemistry is polyhaline (i.e., halinity is greater than 18 ppt). A convention like this was was used by Nichols (1920) to describe the commities of the salt marsh series. When ephemeral J. gerardii reforms along the upland border, it represents a brackish meadow community due to grow water seepage. Here the soils range in halinity from oligohaline to mesohaline.

A primary objective of Miller & Egler (1950) is to describe the plant communities of the salt marsh and used the vegetation of the Headquarters Marsh subsection to achieve this by tossing a 10-foot diameter hoop into seven 'core' community. This is a non-random methodology to avoid having the hoop lie across two or more community types. Four communities are identified as associated with a theoretical profile (after simplying the landscape by removal ditches, creeks, boulders, islands of upland, ponds and pannes). The communities at the border of the marsh (i.e., upland or water) use "border" and a modifier and the communities in between use "slope" as a modifier. The communities are:

  • Spartina alterniflora lower border (SA) - today this is commonly referred to as low marsh for it is flooded twice a day by the tides and the dominant plant is S. alterniflora.
  • Spartina patens lower slope (SP) - part of the high marsh complex with is flooded infrequently and the dominant species is the short meadow grass S. patens; this community contains a number of species of the forb panne community which suggests it is a transitional hybrid community.
  • Juncus upper slope (JG) - part of the high marsh complex and the dominant species is the grass-like rush J. gerardii.
  • Panicum virgatum upper border (PV) - a tall grassland dominated by the grass P. virgatum and the shrub Baccharis halimifolia.

The remaining communities are associated with depressions - specifically the panne sequence:

  • salt panne - areas that are unvegetated with shallow water that is intermittently dry.
  • Stunted Spartina alterniflora Community (SAS) - a community dominated by the stunted form of S. alterniflora.
  • Forb Pannes - occur at an elevation slightly higher than the previous one and is a mix of species especially Triglochin maritima, Plantago maritima, Limonium carolinianum and Gerardia maritima.

These communities are described in detail in Miller (1948) and Miller and Egler (1950). Coleman (1976) resamples these communities on the Headquarters Marsh and finds the same coverage and frequency. The communities noted above will be referred to in the discussion below as to how the vegetation changes post ditching. The abbreviations following some of the community names will be used. An example of another convention used is SAS/SP. This list the dominant species in decending order of abundance. S. alterniflora stunted is dominant and S. patens coverage is less.

Figure 2. Peat core from the Headquarters Marsh (3 and 4) and the bayfront marsh Wequetequock Cove (5). Source: Niering and Warren 1993.

In the natural marsh, the basins are dominated by the panne sequence and the levees are dominated by SP. Miller & Egler (1950) describes the natural marsh as having "as broad swales lying beteen the upland on one hand, and the natural levees of the bayfront or estuary." This pattern changes dramatically on the grid ditched marsh that is overdrained. Two major vegetation changes occur in the initial drainage phase:

  • overdrainage of the marsh allows SP to become the dominant species throughout the basins, reducing the areal extent of all of the panne sequence communities; Core 3 in figure 2 represents how the peat sequence changes as a result of draining; this conversion is precisely the reason farmers drained the basins and is recorded in the literature (Britton 1915; Headlee 1935); Egler (1974) warns that "one should not assume that the 1947 was the end-stage of stability"
  • SP on the levee is replaced by JG as noted in the section above description levee changes; this pattern is represented by Core 4 in figure 2 - note that JG peat overlies SP peat - that horizon corresponds to a depth where a sand layer from the Hurricane of 1938 is shown in Core 5 and his core was collected near the bayfront or creekbank (Niering and Warren 1993)

According to Egler, the SP community covered 65 to 80 percent of the Headquarters Marsh.

In the aggrading ditch phase, three major vegetation changes occur:

  • SAS begins displacing SP at least as early as 1965 on the Brucker Marsh; Gross (1966) describes this area as predominantly supporting the SAS community - on closer inspection of his data, it would appear to be composed of SAS/SP which is considered by Niering (1987) and Coleman (1978) as the transition vegetation from SP dominant to SAS dominant; the 1947 Brucker Marsh was SP and JG (Niering 1974); in 1983, Niering describes this area as "dominantly SAS with occ. forb area and SAS/SP.

    The conversion of SP to SAS at the landward side of lower Palmer Neck includes the formation of FP/SAS in 1976 and some of this marsh also supports SAS; in 1979 the ditches of Barn Island are recleaned and this would have increased drainage and perhaps favor some expansion of SP;
  • Coleman's vegetation map clearly shows that much of the former levee had been converted to DS/bare (it may be that as flooding increases and the backside of the level, JG erodes as it does along the upland margins - creating habitat for DS as also happens on the upland margins); 1947 levee supported JG/DS and in 1976 supports DS/JG; this map also shows how the DS/JG vegetation becomes fragmented islands with a matrix of DS/bare; by 1987 the DS/bare becomes forb panne (this Niering photograph is from the southern side of Palmer Neck standing near the levee and shooting across the marsh towards the upland, the foreground is forb panne and the background is SAS with SP along the upland border); this same pattern is similar to present day conditions except that salt pannes have significantly increased in association with SAS and the levee support SP with some JG patches.

    Coleman (1976) notes that on the Headquarters Marsh, Juncus along the backside of the levee has been converted to forb panne.
  • New and significant levee growth has occurred at Sassafras Marsh (see reverted marsh module) and in fact this location has reverted to the levee and basin topography. Also at the colonial ditch below dike 1, there is significant levee growth (see reverted marsh module). The new levee growth is pure SP and lacks the forb associates of Miller and Egler (1950) Spartina patens lower slope community. At Sassafras Marsh - the eastern levee has some remnant JG patches, no such patches were observed on the western levee. Must of the marsh at Barn Island appears to be in an 'incipient' state of levee growth. In places this is merely sparse SP growing in SA tall. Photostations are being established to monitor future levee growth.

Figure 3. Palmer Neck Marsh vegetation map - 1976. Legend: DS - Distichlis spicata, FP - forb panne, IF - Iva frutescens, JG - Juncus gerardii, PV - Panicum virgatum, SA - Spartina alterniflora, SAS - S. alterniflora stunted, SP - Spartina patens.

Summary:

The 1934 aerial photographs reveal the changes that followed from draining basins with ditched by colonial farmers. At some point in time, levees form along the downstream ends of these ditches and new levees are form, thus restoring the continuous nature of the levee and reestablishing the natural flooding pattern of the levee and basin topography. This occurred to all seven colonial ditches. The natural marsh module reveals that the pre-ditching marsh had the levee and basin topography. A reverted marsh has been found east of Barn Island proper - a marsh called Sassafras Marsh. Marshes like this are present at numerous sites throughout New England. The levee and basin topography appears to represent the equilibrium state. Carey et al (2015) note how the vegetation of Mamacoke Marsh, mapped in 1957 by Dr. Niering, remains the same today and this marsh is tracking sea level rise. Mamacoke has the levee and basin topography. The bayfront marsh at Wequetequock Point at Barn Island is a large levee and small basin marsh that has not been ditched. The levee has supported SP since before the Hurricane of 1938. Wilson et. al (2014) note how the Plum Island LTER in Massachusetts has hydrology that has reverted to the pre-ditching condition and this marsh is tracking sea level rise. From aerial photography, the study sites of Wilson appear to be returning to the levee and basin topography. LeMay's studies at this LTER also show that the back marsh area of unditched areas is 10 cm higher than the comparable habitat in the ditched marsh.

Miller and Egler (1950) could not predict the "ultimate vegetation "lacking sufficient data. They go on to say "by ultimate vegetation is meant that which would develop in the absence of those factors called abnormal, such as fire, mowing and ditching". In this module it is deduced from long-term shrinkage of the levees that ditching lowered the height of high water which also allowed the levees to see the replacement of SP and JG. Connecting the basins directly to the tides creates a huge prism to fill with water before the levees can flood. The post ditching increase of S. patens into the basins is a predictable outcome and the 1947 vegetation should not be considered a stable enpoint (Egler 1974).

It is highly likely that the ditches play and important role in marsh change. Two obvious changes post ditching is no levee building and aggrading ditches. So the ditches may be a sink for sediment. Initial ditching likely caused a loss of elevation and a drop in the water table as described by Stearns et al (1940). The shrinkage of the levees over the decay likely involved some subsurface processes caused by aeration and reduced flooding such as decomposition of peat that leads to subsidence and soil compaction. LeMay (2007) reveals and elevation difference of 10 cm between ditched and unditched marshes. No information could be found about how the few unditched marshes in the northeast flood and drain. If they have no outlet, then they likely have a longer hydroperiod than the ditched marshes.

If the ditched marshes are reverting the levee and basin topography, then it follows that the basins will return to the panne sequence at the expense of SP, which did not occupy the basins but the high ground between the panne sequence and the upland and on the levee. Accompanying the new levee growth at Barn Island and elsewhere, is the increase in SP. Figure 7 in the natural marsh module shows the mesotidal Great Meadows marsh with a high drainage density of creeks and near the mouth of Lewis Gut the levees create a huge area that is likely nearly pure SP. This is similar to the the downstream section of Palmer Cove nearest the bay where entire marsh panels are SP (see figure 1 above; 2012 photography).

Preliminary List of Priority Studies and Monitoring.

Tide Studies - If marshes are reverting and the tides change in response to ditch filling, then gauging the four primary tidal creeks over time may reveal the some of the effects of ditching upon tidal hydrology.

Tide studies inside Sassafras Marsh and Mamacoke Marsh will aid in the understanding of tidal flooding and hydroperiod.

Elevation changes in Sassafras Marsh can be compared to the SET network to compare and contrast a ditched marsh to a levee and basin marsh.

Develop a monitoring scheme for the levees to determine how rapdily they grow and monitoring how the edge vegetation changes in response to levee growth.

Determine how the hydrology switches from basin drainage to a hdyrology that promotes levee growth.

Develop a sediment budget for the marsh.

Monitor the changes that follow as marshes revert. Some New England reverted marshes appear to become very wet and form shallow pools.

Collect peat cores at strategic locations to attempt to define the pre-ditching vegetation and post-ditching vegetation changes.

 

Literature Cited

 

Britton, W.E. 1915. Changes in the vegetation of salt marshes resulting from ditching. CT Exp. Station Rpt. 172-179.

Carey, J.C., K.B. Raposa, C. Wigand and R.S. Warren 2015. Contrasting decadal-scale changes in elevation and vegetation in two Long Island Sound salt marshes. Estuaries and Coasts pp. 1-11.

Coleman, W. B. 1978. Vegetation of the Wequetequock-Pawcatuck Marshes Stonington, Connecticut - A Comparative Study 1948 and 1976. Smith College, MA. 130 p.

Gross, A.C. 1966. Vegetation of the Brucker Marsh and the Barn Island Natural Area, Stonington, CT. Masters Thesis, Connecticut College. 103 p. and appendices.

Headlee, T.J., 1935. Summary of symposium on the relationship of mosquito control in New Jersey to wild life on the salt marshes. Proc. of the twenty-second meeting, NJ Mosquito Control Association. p144-147.

LeMay, L.E. 2007. The impact of drainage ditches on salt marsh flow patterns, sedimentation and morphology: Rowley River. School of Marine Science at the College of William and Mary, Masters of Science, 230 p.

Miller, W.R. 1948. Aspects of waterfowl management for the Barn Island public shooting area. MS Thesis at University of Connecticut. 291 pp.

Miller, W., and F. E. Egler. 1950. Vegetation of the Wequetequock-Pawcatuck tidal marshes, Stonington, Connecticut. Ecological Monographs 20:143-172.

Nichols, G.E. 1920. The vegetation of Connecticut. VII. The associations of depositing areas along the seacoast. Torrey Bot. Club Bull. 47:511-548.

Niering, W.A. 1961. Tidal marshes: Their use is scientific research. p.3-7. In Connecticut's coastal marshes, A vanishing resource. R.H. Goodwin.

Niering, W.A. 1974. Field trip notes.

Niering, W.A. 1983. Field trip notes.

Niering, W.A. 1987. Field trip notes.

Niering, W.A. and R.S. Warren. 1993. Vegetation change on a Northeast tidal marsh: Interaction of sea-level rise and marsh accretion. Ecology 74:96-103.

Redfield, A.C. 1972. Development of a New England salt marsh. Ecol. Monographs 42:201-237.

Stearns, L.A., D. MacCreary and F.H. Daigh 1940. Effect of ditching for mosquito control on the muskrat population of a Delaware tidewater marsh.University of Delaware Agriculture Experiment Station, Bulletin 225, 55p.

Temmerman, S., G. Govers, P. Meire, and S. Wartel, 2004. Simulating the long-term development of levee-basin topography on tidal marshes. Geomorphology 63:39-55.

Tonjes, D.J. 2013. Impacts from ditching salt marshes in the mid-Atlantic and northeastern United States. Environmental Review, 21:116-126.

Wilson, C.A., Z.J. Hughes, D.M. FitzGerald, C.S. Hopkinson, V. Valentine, and A.S. Kolker. 2014. Saltmarsh pool and tidal creek morphodynamics: Dynamic equilibrium of northern latitude saltmarshes. Geomorphology 213:99-115.

 

Published October 3, 2016.

Ron Rozsa.

 

 


*