Origin and Evolution of Dry Valleys South of Ronkonkoma Moraine

Home , Ashworth Glacier, Subglacial stream

ORIGIN AND EVOLUTION OF DRY VALLEYS SOUTH OF RONKONKOMA MORAINE

A Thesis Presented

Soma Das

The Graduate School

in Partial Fulﬁlment of the

Requirements

for the Degree of

Master of Science

Geosciences

Stony Brook University

August 2007 Copyright by Soma Das August 2007 Stony Brook University The Graduate School

Soma Das

We, the thesis committee for the above candidate for the Master of Science degree, hereby recommend acceptance of this thesis.

Gilbert N. Hanson, Thesis Advisor Distinguished Service Professor, Department of Geosciences

Troy Rasbury, Chairperson of Defense Assistant Professor, Department of Geosciences

Henry Bokuniewicz Professor, Department of Marine and Atmospheric Science

This thesis is accepted by the Graduate School

Dean of the Graduate School

ii Abstract of the Thesis

Origin and Evolution of Dry Valleys south of Ronkonkoma Moraine

by Soma Das

Master of Science in Geosciences

Stony Brook University 2007

Long Island hosts a network of straight, parallel Dry Valleys with few tributaries along its southern fork, south of Ronkonkoma Moraine.

In this work I shall examine the characteristics of these Dry Valleys with the help of Digital Elevation Maps. It shall be shown that the distinguishing features of these Dry Valleys- their straight drainage patterns, few tributaries, rectangular watersheds, steep walls and ﬂat ﬂoors at some places- are similar to the valleys formed on perennially frozen ground (permafrost) and latter sapped by groundwater associated with perched water table.

I shall propose four models towards interpreting the formation of these Dry Valleys- outwash plain valleys, groundwater sapping valleys, tunnel valleys and surface runoﬀ valleys.

I shall also focus on the East Hampton region of Long Island, and discuss the special features observed here, Scuttlehole Ponds, where tributary extensions of a Dry Valley are

iii preserved across depressions that were formerly ﬁlled with buried ice.

The event of streams ﬂowing across depressions ﬁlled with buried ice will suggest the timing of the retreat of the Laurentide Ice Sheet during the Last Glacial Maximum in Long Island, around 18,000 to 20,000 before present (10Be age from boulders).

Certain characteristic features of the Dry Valleys, coupled with the evidences of permafrost on Long Island and surrounding regions of New Jersey and Connecticut in the form of fossil ice-wedge, sand-wedge and thermokarst involution, suggest that these Dry Valleys developed after the glacier retreated from Long Island - when there was desert tundra and the environment was periglacial.

iv Acknowledgements

I record my sincere thanks to my advisor Professor Gilbert N. Hanson who provided constant guidance through his wise suggestions and directions in the right track in the completion of the project. I am grateful to him for his support and help from the first day I walked into the Stony Brook campus till the last day I worked on this thesis. Special thanks are due to Mark Demitroff for acquainting me with the permafrost in the field around New Jersey region and also for providing me related valuable study materials. I am greatly thankful to Professor William Nieter for his unique gesture in lending his thesis, helping me thereby to know more about the glacial history of Long Island. My lab colleagues whose cooperation was of immense value to me deserve my heartiest thanks. I am also grateful to my parents living in India, for providing me moral encouragement. Last, but not the least, the contribution of my husband Kundan Sen, but for whose constant inspiration and encouragement and the help in giving final shape to my thesis was not possible, merits special acknowledgement. Table of Contents

Abstract ...... iii

Acknowledgements ...... v

List of Tables ...... viii

List of Figures ...... ix

1. Introduction ...... 1

2. Characteristics ...... 8

2.1. Introduction ...... 8

2.1.1. Location of study ...... 8

2.1.2. Digital Elevation Model ...... 8

2.1.3. Characteristics ...... 9

2.1.4. Hypotheses ...... 10

2.2. Description ...... 10

2.2.1. Drainage Area ...... 11

2.2.2. Sinuosity ...... 13

2.2.3. Drainage Pattern ...... 13

2.2.4. Order of streams ...... 17

2.2.5. Drainage Density ...... 17

2.2.6. Bifurcation Ratio (Rb) ...... 19

2.2.7. Slope of the valley ...... 23

2.2.8. Transverse proﬁle ...... 23

2.2.9. Seismic Study ...... 27

vi 2.3. Valleys cutting through Scuttlehole Ponds ...... 28

2.4. Hypotheses for the formation of Dry Valleys ...... 31

3. Models ...... 44

3.1. Stage 1 ...... 44

3.2. Stage 2 ...... 44

3.3. Stage 3 ...... 45

3.4. Stage 4 ...... 45

3.5. Stage 5 ...... 46

3.6. Stage 6 ...... 46

4. Conclusion ...... 50

Bibliography ...... 51

Appendix I: DEM of Long Island (pocket in back cover) ...... 57

vii List of Tables

2.1. Characteristics of some of the Dry Valleys south of Ronkonkoma Moraine in Long Island ...... 41

viii List of Figures

1.1. DEM of Long Island, from eastern Nassau in the left to East Hampton in the right...... 2

1.2. Two types of Dry Valleys in Long Island. The Cannan Dry Valley has low order of tributaries and the Setuck Dry Valley on the right has higher order of tributaries...... 4

1.3. Paleoclimate and MAT pattern ...... 7

2.1. Rectangular watershed of Cannan Dry Valley, DEM...... 12

2.2. Dendritic drainage network of Setuck Dry Valley in Long Island, DEM. . . 12

2.3. Straight, parallel dry valleys with low sinuosity. The valleys from left to right are Cannan Dry Valley, Swan Dry Valley, and Forge Dry valley, Central Long Island, DEM...... 13

2.4. Straight, Meandering and Braided channels as described by (Rosgen, 1994) . 14

2.5. Presence of Till south of Ronkonkoma moraine proves that glacier advanced beyond the moraine ...... 16

2.6. Stream Order Classification by Ojakangas,1982. The figure shows a 4thorder valley with finger tip tributary labeled as 1. Two finger tip tributaries combine to form 2nd order tributary, labeled as number 2 in the figure. Two 2nd order then combines to form a 3rd order tributary. Two 3rdorder tributaries finally form the main trunk channel which is the highest order and is labeled as number 4, Ojakangas1982 ...... 16

2.7. The ﬁgure shows the Cannan Dry Valley Patchogue, Central Long Island. It has small number of tributaries and the highest order is 3, which represents the trunk channel, DEM...... 18

ix 2.8. Drainage density classiﬁcation, (Strahler,1957)...... 19

2.9. An example of coarse drainage density. The dry valley here represents East Meadow Dry Valley Freeport, Western Long Island, NY with only 2orders of tributaries, DEM...... 20

2.10. Medium to coarse drainage network found in Setuck Dry Valley, East Moriches, Eastern LongIsland, NY, DEM...... 21

2.11. Slope of a Dry Valley is determined by the proﬁle tool of DEM software. The box on the top right shows the longitudanal proﬁle of the Cannan Dry Valley, Patchogue...... 22

2.12. Angulated Drainage of Cannan Dry Valley, Patchogue. The tributary valleys meet the trunk valley at steep angles of 40◦, hence the name. The yellow line is the line of cross-section of the tributary valley. The proﬁle indicates a V-shaped, steep walled valley, DEM...... 23

2.13. The figure represents the flat-floored profile of Cannan Dry valley in Patchogue, Central Long Island, NY. The yellow line shows the cross section path. . . . 24

2.14. The proﬁle of the tributary valley of Setuck Dry Valley in East Moriches with relatively gentle sloping walls, DEM. The yellow line shows the cross section path...... 25

2.15. Steep V-shaped transverse proﬁle of tributary valley of Setuck Dry Valley, East Moriches, Eastern Long Island. The yellow line shows the cross section path...... 26

2.16. The ﬁgure illustrates the sudden change in slope along the thalweg which is the ﬂoor of the basin. The yellow line shows the cross section path, DEM . 27

2.17. Seismic study of the Atlantic Coastal Shelf. The ﬁgure on the top shows the southern part of Long Island. Seismic study of the Atlantic coastal shelf reveals that the southward extension of the Dry Valleys join the dendritic network of the paleochannels (Foster and Scwab, 1999) ...... 28

2.18. Scuttlehole Ponds, Sag Harbour, Long Island...... 29

2.19. DEM of the ScuttleholePonds...... 30

x 2.20. Topographic map of Scuttlehole Ponds. The ponds are ﬁlled with water. . . 31

2.21. Three elliptical ponds are shown in the DEM, with the tributaries of the Hayground Cove Dry Valley system cutting through...... 32

2.22. Anastomosing pattern of Tunnel Valley in Canada...... 34

2.23. IceWedge cast in ScuttleHole Ponds...... 36

2.24. Thermokarst involution in Stony Brook Campus, (Kundic and Hanson 2005) 37

2.25. Tributaries of runoﬀ valleys preserved as extensions of groundwater sapping valleys in Swan Dry Valley in Medford ...... 38

2.26. Process of groundwater sapping and formation of streams, (Dunne, 1980). 39

3.1. Permafrost thaws due to two sources of heat, Szewczyk, 2005...... 46

3.2. Models describing formation of Dry Valleys in Long Island: (a) At the time of Ronkonkoma Glacier advance(60-20 kyr), (b)Ronkonkoma Glacier retreat(20-18 kyr), (c)During time when of permafrost until about 15kyr . . 47

3.3. Models describing formation of Dry Valleys in Long Island: (d) During time of perchedwater table on melting permafrost (15-10kyr), (e)After permafrost melted and before water table rose with post-glacial sea level rise (10kyr to ?) 48

3.4. Scuttlehole Ponds alligned in the nw-se direction, Sag Harbour, Long Island. 49

xi 1

Chapter 1

Introduction

Long Island hosts a network of straight, parallel Dry Valleys with few tributaries, and small drainage areas south of Ronkonkoma Moraine. These valleys are Dry Valleys because they do not have water ﬂow at present and are therefore dry mostly at the upstream ends which is at the northern ends of these valley system. However as we move further south along the valley lengths on the DEM, we can see water present at their lower ends, where the bottom of the valley intersects the water table. The Dry Valleys are named by the ponds and lakes that they cut through for example the Swan Dry Valley cuts through Swan Lake in Medford. There are two valleys- the Connetquot River Valley and the Carmans River Valley, which were not studied because these are broad, occupy large areas and originate north of Ronkonkoma Moraine unlike the other dry valleys of this study all of which originate south of Ronkonkoma Moraine. The valleys are visible clearly on a digital elevation map of Long Island (Figure 1.1). A larger version of the DEM of Long Island with the names of the Dry Valleys is enclosed in the back cover of the thesis.

According to Fuller (1914) these valleys are outwash channels that were formed due to the collection of outwash waters through the agency of an older topography instead of by normal conditions pertaining to outwash accumulation. He described these valleys as well-deﬁned channels with ﬂat bottoms, terraced borders and steep banks. The valleys as observed in the DEM are aligned in the North-South direction.

Such Dry Valleys are also found in England, The Netherlands, Siberia on the previously formed permafrost. Here the Dry Valleys have a subdued cross-section, and are distinct from the valleys which are currently occupied by streams. According to Schmitthenner (1926) these Dry Valleys are also called dells or dellen. He described dellen as ﬂat, elongated, 2

Figure 1.1: DEM of Long Island, from eastern Nassau in the left to East Hampton in the right. often branched depressions, with uniform gradients, the walls of which merge into each other in a gently rounded manner without a break at the bottom. Penck (1924) referred to them as corrasional valleys based on their mode of origin. Dry Valleys with flat floors have been termed cradle valleys. Shallow valleys at the head of the present streams are called rain valleys (Greenwood, 1877). Dry Valleys have been referred from several places in the world including Britain where these are preserved in the chalk of Southern England, they are also known as coombes or escarpment valleys (Reid, 1887). In Britain, Chittern Hills preserve such a network of asymmetrical Chalk Dry Valleys. Ollier and Thomasson (1957), argued that the asymmetry of these valleys are the result of frequent freeze thaw activity on the south and west facing slopes during the fluctuation of climates with colder east facing slope remaining inert (perennially frozen). Ollier and Thomasson (1957) commented that ‘it is worth stressing that by becoming dry the valleys have been left in stage of youth, relics of an early period of erosion, and their features are ”fossil” and not related to the present day conditions ’. Many of these kinds of Dry Valleys having a general asymmetry have been discovered from several other places in the world including Pleistocene gravels in Southern Germany, glacial outwash sands in the Netherlands (Gregory and Walling, 1973). Other places where these kinds of Dry Valleys are present include Wyoming (USA)(Walker, 1948) 3

and in permanently frozen subsoil in Siberia (Schostakowitsch, 1927). Some of these terms refer to the geological conditions and environments that give rise to the Dry Valleys.

There are three proposed areas of origination of Dry Valleys, as follows:

1. Areas of permeable rock and particularly limestone. However, studies in Britain revealed that these valleys can also form on other rock outcrops (Gregory, 1966).

2. Areas which experienced periglacial conditions during Pleistocene times. In western, eastern and central Europe they are called Dellen and in North America the same term applies to rounded depressions in Piedmont plains (Sharpe, 1941).

3. Areas characterized by an arid to semi-arid environment often have wadi-systems. The wadi-systems include systems which do not function during the occasional present storms and thus resemble Dry Valleys (Gregory and Walling, 1973).

All geomorphic features reﬂect the role of some geological processes and in many cases like Dry Valleys, may reﬂect the paleoclimate or the conditions in which these features developed. Thus the main interest in studying Dry Valleys of Long Island is to evaluate the role of the geological processes and also the climate in forming and shaping these Dry Valleys. To understand all these aspects, it is necessary to know the following aspects of Dry Valleys.

1. What are the characteristics of these Dry Valleys?

2. What geological agents and factors were important for their origin?

3. When and under what climatic conditions did these dry Valleys develop?

4. Why are the valleys dry at present?

Certain characteristics of Dry Valleys such as their shape (straight, meandering and braided), area occupied (watershed), number of drainage tributary valleys (order) and the cross-section of the valleys at separate elevations (transverse proﬁle) should enable us to understand the processes that were important for their evolution and development. I will test four hypotheses for the formation of these Dry Valleys, as follows: 4

Figure 1.2: Two types of Dry Valleys in Long Island. The Cannan Dry Valley has low order of tributaries and the Setuck Dry Valley on the right has higher order of tributaries.

1. Melting water from glacier forming outwash plain valleys

2. Sub-glacial meltwater forming tunnel valleys

3. Surface runoﬀ forming dendritic drainage network

4. Groundwater sapping process leading to the development of straight, steeply sloping, low order stream valleys.

Dry Valleys of Long Island mainly exhibit two types of features. At some places the valleys are straight, parallel to each other and have few tributary valleys. The Cannan Dry Valley and Swan Dry Valleys in Patchogue and Medford exhibit these features. Other Dry Valleys have dendritic drainage networks and a large number of tributary valleys, as in Setuck Dry Valley of East Moriches (Figure 1.2).

Analyzing these characteristics, along with other identifying features of Dry Valleys discussed in the next chapter, I shall discuss the role of surface runoﬀ process and groundwater sapping process in forming these particular valleys of Long Island. Impervious frozen 5

ground or permafrost and a very cold, dry climate were also important for the formation of the Dry Valleys.

The last glacial history of Long Island dates back to 25000 years ago and is brieﬂy described as follows. All ages are in calendar years and are mostly calibrated using Calib Software (Stuiver et al., 2005).

The glacial sediments underlying Long Island are a result of at least two glacial advances. These advances formed ﬁrst the Ronkonkoma Moraine in the center of Long Island and along the South Fork, then the Harbor Hill Moraine along the north shore of Long Island. The Harbor Hill moraine is considered younger because it truncates the western end of the Ronkonkoma Moraine in the vicinity of Roslyn Heights (Bennington, 2003) and because outwash channels originating at the Harbor Hill Moraine cut the Ronkonkoma Moraine. There is no direct evidence for the absolute timing of the formation of the Ronkonkoma Moraine, other than that it is younger than the last interglacial (Sangamon). The evidence for this are beach sands thrust into the Ronkonkoma Moraine that include mollusk fossils (Mercinaria, mercinaria) that give last interglacial amino acid racemization ages (Meyers, 1998). Cosmogenic exposure ages of ca 18,000 years have been obtained for boulders on both the Ronkonkoma Moraine and the Harbor Hill Moraine (Cabe et al., 2006) suggesting that the Ronkonkoma Moraine may have formed only shortly before the Harbor Hill Moraine which is thought to have formed by about 20,000 years ago (Sirkin, 1982).

Fuller (1914) and Merrill et al. (1902) suggested that the Ronkonkoma Moraine is a terminal moraine. King et al. (2003) and Schmitt (2006) however found till overlying outwash sediments south of the Ronkonkoma Moraine in Eastern Nassau and Western Suﬀolk coun- ties. Presence of this till implies that the Ronkonkoma Moraine is not a terminal moraine and that a glacier, presumably the one responsible for the formation of the Ronkonkoma Moraine, advanced south of the Ronkonkoma Moraine. It could be possible that the outwash plain valleys formed from the melting water of glacier were covered by the till and sediments left behind by the retreating glacier. This is why we do not have any relict of braided stream network characteristic of outwash plain valleys in Long Island.

After the glacier retreated from Long Island 20,000 years ago, the landscape was desert tundra, dominated by an extremely cold, dry periglacial environment. The Mean Annual 6

temperature was probably -8◦C to -10◦C (Figure 1.3). The treeless periglacial environment at that time witnessed the development of permafrost. Evidence for permafrost has been documented in Long Island (Nieter, 1975; Kundic and Hanson, 2006; Denny, 1936).

Ice-wedge cast indicative of former permafrost conditions have been reported from southern New England (Stone and Ashley, 1992) and Southern New Jersey (French et al., 2007). This implies that permafrost was present in almost all parts of Northeast USA after the glacier retreated and it extended even further south of Long Island.

Permafrost probably lasted till 14,600 years when there was sudden climatic warming known as Bolling Allerod Interstadial event (Weaver et al., 2003). However, some people, like Mayewski et al. (1993) and Logan (1983) based on their study on ammonia and methane ﬂux from the Greenland Ice Core, believed that Bolling Allerod might not have had a warming impact on continents of North America and Europe. If they are correct permafrost may have lasted until 11,500 years until the end of the Younger Dryas (12,900 to 11,500 years) (Weaver et al., 2003). Peteet et al. (1994) stated that after the Younger Dryas the temperature increased, as is supported by the appearance of thermophilous species like Pine, Oak, and Hemlock. During this phase, the permafrost could have continued melting until it disappeared completely from Long Island.

My approach in investigating the origin and evolution of these Dry Valleys will take the following steps:

1. To determine characteristics of the Dry Valleys using the Digital Elevation Model tools.

2. To compare these characteristics with the four diﬀerent models that I will consider and test.

3. To evaluate the probable time of their formation by looking at the the geological history of Long Island. 7

8 8

6 6 4 4 2 2

C) [8]Elias,1996 ˚ , 0 0 C) o

997 -2 -2 1 -4 -4 -6 -6 Long Island MAT Island Long Temperature ( -8 -8 Foster et al, ] [7 -10 -10 , -12 -12

-20 -19 -18 -17 -16 -15 -14 -13 -12 -11 -10 Time (kyr) Time

Temperature curve: MAT (in 6]Peteet,1994 ,[ d

[11]

[9] ir,

[6]

[6] ine, ine,

Fir

[9] .

ine, ine, eech, [11] P and ak, ak, B Birch

emlock , , O H

sh,

re dependant [9] A

, Pine decreases[9], [6] .

Spruce ak, [5]Kundic&Hanson,2006 Tamrak ,

O irch , eech eech and ir, ,B ood Forest, B ,F

, Pinus, e Wedge[3], Sand wedge[4], emlock [12]Allison et al,1986[12]Allison Hardw , Evidences Ic Thermokarst Involution[5],Ventifacts[3],[4] BeetleAssemblages[8]. DesertTundra Grass, pinesArctic andAlpine Sedges.[6],[7] BarrenArctic Upper in Delaware an Western Island[10]Long Alder Boreal Forest [10] Spruce H Fir, hemlock and Alder. Increasedeciduous in trees[9] Boreal Forest Spruce rise in Hardwood Forest ,oak, beech, hemlock and alder.[1],[11] Increase hickoryof pollen. (moistuHemlock species) decreased[12] Hardwood Forest Increase chestnutof pollen, hickory pollen

Paleoclimate and MAT pattern

Demitroff,2006

] [1]

[4 o , 8 [2]

15 [1] 6 to - C[6] C[6] C - o o o [11] Shuman,2004 22 [8] 16 12 4 to 0 - , : - - - :11.5

Nieter,1975 17 [6] ]

- 3 max min [

Temperatures ( Jan: July:14 [1] MAAT: T T Mean July:12 [6] Mean July: 16 Jan: Mean July: 13 to 14 Mean July: 15 to 16 Mean July 18 to 19 [6] Mean July:21 [6] Mean July: 22 [6] Jan: Figure 1.3: ,

the last deglaciation in Long Island

[1].

moist&

[10] 1977 Newman,

[2]Pewe, 1966

, , after

of pptnof 30cm of30cm pptn[1] of40cm pptn[1] - - , 1999

Climate Cold,, dry 20 Cold ,dry, 30 andWarm dry Bolling Allerod Warm, humid.[6] Cold,climate dry Younger Dryas. Warm, moist 80cm Warm, moist pptn[1]of 80 cm dry.Warm, Warm, moist 80to120cm of pptn[1]

Prentice,et al1991 ] 20 18 15 14 13 12 11 10 Paleoclimate Age (kyr) pptn: average annual precipitation; [1 [9]Kneller et al and Peteet (1999), Newman (1977), Peteet et al. (1994),Shuman et P´ew´e(1969), al. (2003), Allison et al. (1986) Kundic and Hanson (2006) Nieter (1975), Prentice et al. (1991), Elias et al. (1996), Foster et al. (1997), French et al. (2007), Kneller 8

Chapter 2

Characteristics

2.1 Introduction

2.1.1 Location of study

A digital elevation model (DEM) of Long Island is shown in Figure 1.1. The higher elevation area in orange and red on the DEM is the Ronkonkoma Moraine.

To the south of the Ronkonkoma Moraine lies a network of ‘Dry Valleys’, which are the focus of this study. In the same ﬁgure, we see a large green area between the Ronkonkoma Moraine and the Atlantic Ocean. A closer look shows straight, parallel vein-like drainage channels. These presently dry channels are called ‘Dry Valleys’.

Dry Valleys are clearly deﬁned valleys that no longer have a surface ﬂow of water (Grimes, 1995). According to (Schmitthenner, 1926), small and shallow representations of these valleys are also known as ‘dell’s. (Katasonova, 1963) refers to these dry, straight and parallel valleys as ‘delly’.

The primary objective of this study is to determine the processes responsible for the formation of these dry valleys which can help us to understand the geological history of Long Island. Before I start oﬀ with the analysis, however, we need to know a little more about a few important factors.

2.1.2 Digital Elevation Model

Historically, geologists based their work on topographic maps, which represented the natural features of the area. With the advent of computers, we now have the ability to view and analyze this information through a ‘Digital Elevation Model’. 9

‘A digital elevation model is a representation of topography in digital format. High resolution digital elevation models are available for the State of New York including Long Island. These have a horizontal resolution of 10 meters and are based on 7.5’ topographic maps. For those quadrangles with 10’ contour intervals, interpolation results in elevations with an uncertainty of about 4’. The appearance is as if one were viewing color-enhanced images of a barren terrain, for example Mars. This allows one to see much greater detail than is possible on a standard topographic map.’ (Hanson, 2005)

This higher level of detail, coupled with the availability of tools to analyze three dimen- sional profiles of arbitrary paths and areas, allows us to make interpretations that were not possible with topographic maps alone. Identification of Dry Valleys, as we did earlier in this chapter, has become significantly simpler using the digital elevation model of Long Island. All our discussions will rely heavily on these models.

Let me now move on to the next factor that will help us understand the processes responsible for the formation of Dry Valleys in Long Island - primary characteristics of these Dry Valleys.

2.1.3 Characteristics

In this section, I will introduce some important quantitative characteristics of streams and valleys :

• Drainage basin: A drainage basin is the entire area providing runoff to, and sustaining part or all of the streamflow of the main stream and tributaries (Gregory and Walling, 1973). Drainage basin is also referred to as Drainage Area, Watershed, or Catchment area. It signifies the shape and area of the basin enclosing the streams and its network of tributaries.

• Sinuosity of the valleys: Sinuosity is a measure of the degree of meandering of a stream.

• Drainage patterns: Three basic drainage patterns are present. These are straight, meandering and braided drainage network. 10

• Order of the streams: Stream order is the measure of classifying segments of a valley (Horton, 1933b). The greater the hierarchy of the tributaries, the higher the order of the streams.

• Drainage density (Dd): Drainage Density is deﬁned by (Horton, 1932) as the length of the streams per unit area.

• Bifurcation Ratio (Rb): Bifurcation Ratio is deﬁned by Horton as the ratio of the number of streams of order ‘n’ to the number of streams of the next highest order ‘n+1’. It depends on the ordering method (Horton, 1932).

• Slope of the valley: Slope of the change in the relief of the valley ﬂoor.

• Transverse profile of a valley: Transverse profile refers to the size and shape of a valley in cross-profile .

2.1.4 Hypotheses

My study considers four hypotheses for how the Dry Valleys formed.

• Outwash Plain Valleys - valleys created by the streams which are derived from the melting of the glacier on the Ronkonkoma Moraine.

• Groundwater Sapping valleys - created by headward erosion by groundwater.

• Tunnel Valleys - formed by sub-glacial meltwater.

• Surface Runoff Valleys - formed by runoffs or overland flows on a hard and impervious surface.

2.2 Description

Geomorphology of streams reﬂect their mode of origin, the controlling factors and also the role of the diﬀerent geological agents like wind, river and glaciers. These general characteristics are described as follows. The origin of the Dry Valleys can be interpreted by comparing their features with the general characteristics of streams. 11

2.2.1 Drainage Area

The area surrounding a stream and its tributaries, enclosed by drainage divides, is known as the ‘drainage area’, or ‘watershed’, of that stream. By deﬁnition, every stream has its own watershed or drainage area.

Drainage areas can have one of three shapes - rectangular, elliptical or circular. Valleys with a small number of tributaries generally have small, rectangular drainage areas. Larger valleys with a large number of tributaries have large drainage areas, which are usually circular or elliptical.

The shape and the size of a drainage basin is inﬂuenced by the amount of water yield and the rate of sediment supply. A rectangular watershed generally encloses a straight and parallel drainage network, and indicates a small amount of water yield during the formation of the streams. In contrast, an elliptical watershed usually surrounds a dendritic drainage network and suggests a higher amount of water.

The evolution of a watershed is dependent on several physical processes, surface runoffs and groundwater sapping being the primary ones. From the size and the shape of a drainage basin, we can interpret the conditions during its evolution. Easily available surface runoff water, in high volumes, coupled with low infiltration into the ground, gives rise to large, circular drainage areas with dendritic networks. However if the soil surface has high infiltration and low runoff, groundwater sapping becomes the dominating process in forming streams and valleys. Loss of water by infiltration and evapotranspiration and lower runoff results in streams with smaller drainage areas with rectangular watersheds.

To illustrate from examples on Long Island, the Cannan Dry Valley in Patchogue has a rectangular watershed and a small drainage area, as seen in Figure 2.1. On the other hand, the Setuck Dry Valley, as seen in Figure 2.2, exhibits a large drainage area and a circular watershed. As we have learnt from above, by observing the watershed patterns from a DEM model, we can interpret the roles of groundwater sapping and surface runoﬀ processes in their formation. 12

Figure 2.1: Rectangular watershed of Cannan Dry Valley, DEM.

Figure 2.2: Dendritic drainage network of Setuck Dry Valley in Long Island, DEM. 13

Figure 2.3: Straight, parallel dry valleys with low sinuosity. The valleys from left to right are Cannan Dry Valley, Swan Dry Valley, and Forge Dry valley, Central Long Island, DEM.

2.2.2 Sinuosity

Sinuosity is a measure of how straight or meandering a stream or a channel is. Sinuosity is the ratio of stream length between two points divided by the valley length between the same two points. This ratio is a speciﬁc characteristic of the stream/channel and can only be determined accurately when a stream/channel is present or preserved within a valley.

A stream/channel with a sinuosity ratio of 1.5 or greater is considered to be a meandering channel (Gregory and Walling, 1973) (Figure 2.3).

On Long Island, most of the Dry Valleys do not have preserved stream channels. Absence of channels prevents accurate determination of the sinuosity levels. However, judging by the small size of the valleys, we can interpret that the stream channels that were conﬁned within them were narrow as well. In other words, the shape of a Dry Valley conforms to the shape of the stream/channel. From the Figure2.3 we see the Dry Valleys are essentially straight and maintain minimum sinuosity. In Patchogue, for example the sinuosity of the Cannan Dry Valley is:

P = channel length/ valley length = (5.15/6) miles = 0.85

2.2.3 Drainage Pattern

Natural streams exhibit one of three diﬀerent patterns - straight, meandering or braided. Meandering stream follows a winding path along its ﬂow. Braided streams are multi-channel forms separated by bars and islands (Knighton, 1998). 14

Figure 2.4: Straight, Meandering and Braided channels as described by (Rosgen, 1994)

The most common part of the pattern continuum proceeds from the stream with no lateral migration or switching (straight and inactive sinuous ones) through actively meandering streams to braided streams (Ferguson et al., 1992).

Streams are characterized by

1. Stream power, which implies discharge at constant slope or slope at constant discharge. It is low for straight, medium for meandering and high for braided.

2. Width to depth ratio which is associated with bank erodibility and decreasing bed load. This property is low for straight stream channels, low to moderate for meandering and high for braided.

3. Amount or calibre of bed load It is low for straight, low to moderate for meandering and moderate to high for braided

The above three conditions control the development of patterns in streams as in Figure 2.4, (Rosgen, 1994). Outwash plain valleys with braided stream pattern form when the ﬂow 15

of water is high and straight or meandering streams form otherwise. A higher power of water indicates high overland ﬂow, such as the meltwater from a deglaciation process. A lower power of water may arise due to groundwater sapping.

Applying this to the Long Island Dry Valleys, we see most of the Dry Valleys have a straight drainage pattern as in Figure 2.3, with no evidence of braided patterns. To have a better understanding of the absence of braided patterns, we need to look into the glacial history of Long Island.

Glaciers advanced over Long Island several times in the past. According to Sirkin (1996) the last glacier was present on Long Island during 25000 years and were gone by 20000 years Sirkin (1982). Another source of age of the glacier comes from the ages on boulders on the Harbour Hill moraine which is derived from cosmogenic exposure which is 17000 to 18000 years. Although less constrained it can be argued that the last glacier retreated from Long Island around 20000 years (Cabe et al., 2006). At some stage of glacial retreat, the meltwater could have created outwash plain valleys with a braided appearance. However, when the glacier re-advanced over Long Island during the Last Glacial Maximum, these outwash plain valleys could be completely covered by the advancing ice, till, and outwash deposits. There are evidences of outwash deposition in Long Island as documented by Fuller (1914). ‘He described these outwash from the Ronkonkoma moraine as the great sloping plains lying south of the moraine and extending continuously except for short breaks at the Shinnecock Hills, from the west end of the island to the vicinity of Amaganset.’

King et al. (2003) provided evidence of presence of till south of Ronkonkoma Moraine from WestHampton to the east to North Amityville to the west, ﬁgure 2.5 . The till in this region is described by them as a surface layer often covered by loess and it is underlain by stratiﬁed sand and gravel or clay. Dry Valleys do not resemble the braided streams which are characteristics of outwash plain valleys. This may imply the role of till and glacial sediments in burying the outwash plain valleys. The absence of outwash plain valleys in Long Island suggests that meltwater from glaciers was not the agent that eroded the Dry Valleys of Long Island. 16

Figure 2.5: Presence of Till south of Ronkonkoma moraine proves that glacier advanced beyond the moraine

Figure 2.6: Stream Order Classification by Ojakangas,1982. The figure shows a 4thorder valley with finger tip tributary labeled as 1. Two finger tip tributaries combine to form 2nd order tributary, labeled as number 2 in the figure. Two 2nd order then combines to form a 3rd order tributary. Two 3rdorder tributaries finally form the main trunk channel which is the highest order and is labeled as number 4, Ojakangas1982 17

2.2.4 Order of streams

It is often convenient to classify streams within a drainage basin by systematically defining the network of branches. Each non-branching channel segment (smallest size) is desig- nated a first-order stream. A stream which receives only first-order segments is termed a second-order stream, for example Figure 2.6 (Ojakangas and Charles, 1982). The order of a particular drainage basin is determined by the order of the principle or largest segment, and an increasing order of network should be associated with greater stream flow values.

In addition to giving us an idea of index size and scale, stream ordering can also give us an estimate of the streamflow that can be producted by a particular drainage network (Gregory and Walling, 1973). A lower order means a smaller volume of water available for flow, possibly due to groundwater sapping. A higher order reflects a greater streamflow, such as from surface runoff processes. It might also reflect how close one is to the headwaters of the stream.

The Cannan Dry Valley in Patchogue has an order of three for the main valley running North to South (Figure 2.7). The Swan Dry Valley in Medford also has an order of three. The lower order of some streams of Long Island suggests that the streamﬂow was low at the time of formation of the streams.

2.2.5 Drainage Density

Drainage Density of a stream network is deﬁned as the length of the channel per unit area of the drainage basin. Drainage density can be mathematically expressed, as noted in (Hanson, 2006), as:

Drainage Density (Dd) = Channel Length / Basin Area

There are three types of Drainage Densities as classiﬁed by (Strahler, 1957), shown in Figure 2.8.

Coarse drainage densities, which means few tributaries of a channel are typical of areas characterized by permeable rocks and low rainfall. Medium drainage densities are found in humid parts. Fine drainage densities are found in the areas of poor vegetation cover and impermeable substrate where surface runoﬀ is eﬀective and decreases with increased plant 18

Figure 2.7: The ﬁgure shows the Cannan Dry Valley Patchogue, Central Long Island. It has small number of tributaries and the highest order is 3, which represents the trunk channel, DEM. 19

Figure 2.8: Drainage density classiﬁcation, (Strahler,1957).

cover (Melton, 1957) .

Fine drainage density is also found in semi-arid regions and drainage density changes from ﬁne to coarse in more arid regions, due to lack of precipitation.

On Long Island we observe mainly two kinds of drainage densities. It is coarse to medium in Setuck Dry Valley of Eastport and Hayground Cove Dry Valley of East Hampton (Figure 2.10). The different values of drainage densities of the Dry Valleys are listed in Table 2.1. The higher values of drainage densities here may reflect the impact of runoff over an impervious surface, such as permafrost, in which ground remains frozen and impervious for more than a year. Coarse drainage density is found further west on Long Island. Here the Dry Valleys, such as the Cannan Dry Valley in Patchogue, have coarse drainage density as in Figure 2.9. This may imply the role of subsurface process like groundwater sapping process.

2.2.6 Bifurcation Ratio (Rb)

Bifurcation Ratio(Rb) is the ratio of a number of stream segments of one order to the number of stream segments of higher order and is an important characteristic of a drainage 20

Figure 2.9: An example of coarse drainage density. The dry valley here represents East Meadow Dry Valley Freeport, Western Long Island, NY with only 2orders of tributaries, DEM. 21

Figure 2.10: Medium to coarse drainage network found in Setuck Dry Valley, East Moriches, Eastern LongIsland, NY, DEM. 22

Figure 2.11: Slope of a Dry Valley is determined by the proﬁle tool of DEM software. The box on the top right shows the longitudanal proﬁle of the Cannan Dry Valley, Patchogue.

basin.

Mathematically, the Bifuraction Ratio is represented as follows: P P WRb = [Rb(n : n + 1)x(Nn + Nn+1)]/ N where WRb is the weighted mean Bifurcation Ratio, Rb is the bifurcation ratio of n order to (n+1) order of streams, N is the number of streams.

The values for Bifurcation Ratio of Dry Valleys of Long Island are calculated and are listed in Table 2.1. Bifurcation Ratio depends on the order of streams. Certain values of Bifurcation ratio- characteristically between 3.0 and 5.0 suggest geological structure does not exercise a dominant inﬂuence on the drainage pattern (Strahler, 1964). From the table we see that most of the Dry Valleys have bifurcation ratio in the range of 3-5. A few exceptions include the Cannan Dry Valley and the Setuck Dry Valley, where these valleys have Bifurcation Ratio of 8.85 and 19.2 respectively. These two Dry Valleys may imply the structural control behind their origin while the other Dry Valleys are independent of this factor. 23

Figure 2.12: Angulated Drainage of Cannan Dry Valley, Patchogue. The tributary valleys meet the trunk valley at steep angles of 40◦, hence the name. The yellow line is the line of cross-section of the tributary valley. The proﬁle indicates a V-shaped, steep walled valley, DEM.

2.2.7 Slope of the valley

Slope of the main channel of the trunk stream is calculated from the diﬀerence between highest and lowest points in basin elevation, divided by the total distance between those two points as seen in Figure 2.11. It is also described as the basin relief along longest dimension of basin parallel to principal drainage line (Strahler, 1952). On Long Island the average slope of the Dry Valleys is between 0.0016 and 0.004. The slopes of the Dry Valleys are calculated by the proﬁle tool in DEM program of Global Mapper. The slope of the valleys may have been slightly less if the valleys formed shortly after the Last Glacial Maximum, 18000 years before present, when the land started rising due to isostatic adjustment. Since the ice would have been thicker to the north, the headwaters would have rebounded more, leading to the gradual increase in the slope. It may be noted that the distance from the headwaters to the present shore line is less than 10 miles, so the change may be imperceptible.

2.2.8 Transverse proﬁle

Looking at the digital elevation model in Figure 1.1, we see a network of straight, parallel dry valleys located around Patchogue and Medford areas. These valleys are Cannan Dry Valley 24

Figure 2.13: The figure represents the flat-floored profile of Cannan Dry valley in Patchogue, Central Long Island, NY. The yellow line shows the cross section path.

and Swan Dry Valley. These valleys are characterized by angulated patterns, implying the tributaries meet the trunk channel at higher angles of 40◦. Figure 2.12 shows us a transverse profile of one such valley, the Cannan Dry Valley, and we see that the valleys are V-shaped with steep walls. Figure 2.13 shows another tributary valley of Cannan Dry Valley in the same vicinity where the floor is flat.

Moving further east along the DEM map, a dendritic drainage pattern is observed at East Moriches (Setuck Dry Valley). Unlike the earlier valleys, here we see both gently sloping (U-shaped, Figure 2.14) as well as steep sloping (V-shaped, Figure 2.15) transverse proﬁles.

Figures 2.14 and 2.15 describe the change in the shape of a transverse profile of a dry valley in East Moriches, as we move southward along the thalweg. The white lines indicate the points where the transverse profile are drawn with the help of DEM profile tool. The dendritic drainage pattern is a characteristic of runoff streams. The upper end of the runoff tributary has a gently sloped transverse profile while the lower end of the tributary has a steeper transverse profile. Additionally we observe that there is a sharp change in slope progressing from the upper end of the runoff tributary to the lower end of the same. The steep walled transverse profile is a characteristic of groundwater sapping valleys and the 25

Figure 2.14: The proﬁle of the tributary valley of Setuck Dry Valley in East Moriches with relatively gentle sloping walls, DEM. The yellow line shows the cross section path. 26

Figure 2.15: Steep V-shaped transverse proﬁle of tributary valley of Setuck Dry Valley, East Moriches, Eastern Long Island. The yellow line shows the cross section path. 27

Figure 2.16: The ﬁgure illustrates the sudden change in slope along the thalweg which is the ﬂoor of the basin. The yellow line shows the cross section path, DEM

gently sloping valleys imply the role of surface runoﬀ process. Figure 2.16 shows a sudden change in the slope of the thalweg during traversal from the upper end of the runoﬀ tributary to the lower end of the same.

A closer look at the DEM of the Dry Valleys reveals relics of a former drainage pattern. On the DEM they appear as faint, indistinct lines preserved as extensions from the visible tributaries in Figure 2.25.

2.2.9 Seismic Study

Seismic studies of the Long Island coastline reveal paleochannels preserved beneath the ocean, Figure 2.17. Several lenticular bodies have been identiﬁed in this region as buried shallow, dendritic channel systems. Close correlations can be drawn between these buried channels and the existing Dry Valleys which are present south of Ronkonkoma moraine. The buried channels on the sea ﬂoor could be the southward extensions of the Dry Valleys 28

Figure 2.17: Seismic study of the Atlantic Coastal Shelf. The ﬁgure on the top shows the southern part of Long Island. Seismic study of the Atlantic coastal shelf reveals that the southward extension of the Dry Valleys join the dendritic network of the paleochannels (Foster and Scwab, 1999)

of Long Island. It is believed that during the last glacial maximum, when the Laurentide Ice Sheet advanced over Long Island, the global sea level was about 125 meters below today’s sea level (Fairbanks, 1989). The shore should then be lying further south of what it is at present. The valleys formed after the glacier retreated were thus preserved on the coastal shelf. After the glacier retreated roughly 18000-20000 years before present, the sea level started rising and eventually submerged the paleodrainage system which is now preserved as buried channels (Foster et al., 1999).

2.3 Valleys cutting through Scuttlehole Ponds

On Eastern Long Island we ﬁnd the extensions of the tributaries of the Hayground Cove dry valley system (Figure 3.4) across the ‘Scuttlehole ponds’. These are a series of elliptical ponds with diﬀerent surface elevations which are uniformly aligned in the NW-SE direction. They occupy the region south of an interlobate zone of Eastern Long Island (Figure 2.19).

These ponds bear a distinct relation to the adjacent Hayground Cove dry valley system. Tributaries of the valley system seem to have cut right through the Scuttlehole ponds (Figure 29

Figure 2.18: Scuttlehole Ponds, Sag Harbour, Long Island. 30

Figure 2.19: DEM of the ScuttleholePonds. 31

Figure 2.20: Topographic map of Scuttlehole Ponds. The ponds are ﬁlled with water.

2.21) - tributaries on the west of the ponds are found to have their terminal ends preserved on the eastern bank.

We see that both of the above systems have water present (Figure 2.20) only in their southern ends. This implies that the rising water table is ﬁlling up the Dry Valleys and depressions of the ponds. The upper extremities of these streams, exhibiting a lower water table, remain dry.

The uniform alignment of the Scuttlehole ponds, and their typical elliptical shape, point to their resemblance to kettle holes.

2.4 Hypotheses for the formation of Dry Valleys

The Dry Valleys of Long Island south of Ronkonkoma Moraine are characterized by two diﬀerent drainage patterns. Valleys located on the west and central part of Long Island, namely the Valley Stream Dry Valley, East Meadow Dry Valley, New Mill Pond Dry Val- ley and Cannan Dry Valley, have a straight, parallel drainage network with rectangular watershed and small drainage areas. These valleys also have small orders, low bifurcation ratios, low drainage density and steep walled transverse proﬁle. Further east for example 32

Figure 2.21: Three elliptical ponds are shown in the DEM, with the tributaries of the Hay- ground Cove Dry Valley system cutting through. 33

the Setuck Dry Valley and Hayground Cove Dry Valley are characterized by a dendritic drainage pattern, elliptical and circular watersheds with large drainage area, higher orders and large drainage density. Also valleys here have both steep and gentle transverse proﬁle for example in Setuck Dry Valley.

All the above characteristic features of these Dry Valleys give us evidence of how these valleys originated. As has been already mentioned the valleys do not resemble the braided outwash plain valleys. There are three other hypotheses that are considered to explain the origin of Dry Valleys.

1. Were they Tunnel valleys or Subglacial stream valleys?

Tunnel valleys are also called subglacial stream valleys eroded by water ﬂowing in tunnels under the ice. The remnant of a tunnel channel is a scoured dry valley (tunnel valley) which may contain outwash sand and gravel and has a coating of till left by the overlying glacier. Tunnel valleys are identiﬁed by their following characteristics (Cofaigh, 1996):

• Tunnel valleys are elongate depressions with asymmetric sides.

• These valleys have an anabranching character as opposed to the dendritic pattern as seen in Figure 2.22 from (Cofaigh, 1996).

• Tunnel valleys may include relatively straight individual segments independent of each other.

• They usually lack tributaries.

• Valleys have variable widths.

• Valleys are generally oblique to modern drainage pattern.

• Till drapes the walls and the ﬂoors of the tunnel valleys.

Another interesting feature of a tunnel valley is that the floor of the channel may go up and down in elevation (a.k.a. up-and-down long profiles), unlike a scoured river valley with steadily sloping floor (Hanson, 2005).

(Cofaigh, 1996) suggested three methods for the formation of these tunnel valleys: 34

Figure 2.22: Anastomosing pattern of Tunnel Valley in Canada.

• Subglacial stream erosion: Deformed sediments start creeping to the site of erosion which is being removed by the subglacial stream.

• Receding ice during the retreat of a glacier may form tunnel valleys along the margin of the retreating glacier.

• Catastrophic meltwater of glaciers can form these tunnel valleys as reﬂected from the large, anastomosing network of the subglacial stream valleys.

The above features of tunnel valleys are absent in case of the Dry Valleys. Unlike tunnel valleys these Dry Valleys have the following features:

• Dry Valleys are straight and essentially shallow e.g East Meadow Dry Valley, Cannan Lake Dry Valley, Swan Dry Valley.

• Some Dry Valleys possess a dendritic pattern as in Setuck Dry Valley.

• Dry Valleys do not have individual segments which are independent of each other.

• Dry valleys have tributaries as in Massapequa Dry Valley, Setuck Dry Valley, Hayground Cove Dry Valley.

• Widths of the Dry Valleys remain more or less constant throughout their lengths. 35

• Till is found at some places (King et al., 2003) south of Ronkonkoma Moraine, but it is still unclear as whether the till is draping the walls and the ﬂoors of all the Dry Valleys.

2. Were they Surface runoﬀ Valleys?

Surface runoff valleys form when the rate of precipitation exceeds the rate of infiltration. It is often called Hortonian overland flow or unsaturated overland flow (Horton, 1933a). The surface runoff valleys are dominant in the arid to semi-arid regions where the intensity of precipitation is high and the downward infiltration into ground is pre- vented by surface sealing. Surface runoff valleys are also found in periglacial regions where the subsoil is underlain by impervious permafrost (Dunne et al., 1976). This impervious permafrost thus facilitates the development of surface runoff valleys or overland flow. The valleys are formed mainly during peak snow melt in spring and summer (Woo, 1999). They thus have water only during these times and remain dry for the rest of the year due to low rainfall during the summer. Because of this particular feature, the surface runoff valleys are often referred to as ephemeral streams, or arroyos. Ephemeral streams flow only in response to wet precipitation or snow melt (Terell, 2005).

The identifying characteristics of the surface runoﬀ valleys are:

• Surface Runoﬀ valleys have a dendritic drainage network.

• Drainage density is high due to a large number of tributaries.

• Valleys are mostly U-shaped in cross-section with gently sloping walls.

Surface Runoﬀ Valleys form in three stages:

• Surface runoﬀ valleys develop very rapidly and form a master channel.

• The initial slope at start is replaced by slopes that slant towards the main drainage line.

• New slopes thus formed are then subject to erosion. This in turn results in development of tributaries and expansion of channel’s drainage area. This ﬁnally creates a dendritic drainage pattern. 36

Figure 2.23: IceWedge cast in ScuttleHole Ponds.

All of the above features of surface runoﬀ valleys are present in Long Island Dry Valleys as for example in the Setuck and Hayground Cove Dry Valleys. The process of formation of these valleys can be explained as follows:

• The Laurentide Ice Sheet retreated from Long Island some 18000-20000 years before present, leaving behind a periglacial environment and permafrost.

• Landscape was a desert tundra. Cold and dry, treeless landscape characterized by the presence of sedges and grass.

• The frozen, impervious ground inhibited infiltration of water. As a result the rate of overland flow exceeded the rate of infiltration resulting in the formation of dendritic drainage network.

Evidences for permafrost are found on Long island in the form of ice wedge on the eastern rim of Scuttlehole depression in Long Island as in Figure 2.23, (Nieter, 1975).

In Figure 2.24 we see an example of thermokarst involution on the wall of a trench in Stony Brook University in Long Island (Kundic and Hanson, 2006). It is believed that the permafrost inﬂuenced the development of the surface runoﬀ valleys and streams. The valleys were formed during peak snow melt periods, during spring and summer- 37

Figure 2.24: Thermokarst involution in Stony Brook Campus, (Kundic and Hanson 2005)

and had a dendritic drainage pattern with a large number of tributaries. Relics of these surface runoﬀ valleys are still preserved in several locations of Long Island, either as a dendritic drainage pattern as in case of Swan Dry Valley or as ghost tributaries extending from the present visible tributaries (Figure 2.25).

3. Were they Groundwater Sapping Valleys?

Valleys generated by groundwater sapping have the following characteristics (Aharon- son et al., 2002):

• Valleys have steep walls.

• Valleys possess ﬂat ﬂoors.

• Longitudinal proﬁle of a groundwater sapped valley is irregular.

• Sinuosity of the valley is low.

• Width of the valley remains relatively constant throughout width.

• Valleys trend towards the drainage basin.

• Groundwater sapping valleys have small number of tributaries, compared to the dendritic drainage of runoﬀ valleys (Kochel and Piper, 1986). 38

Figure 2.25: Tributaries of runoﬀ valleys preserved as extensions of groundwater sapping valleys in Swan Dry Valley in Medford 39

Figure 2.26: Process of groundwater sapping and formation of streams, (Dunne, 1980).

The groundwater sapping valleys form when rate of inﬁltration into the ground exceeds the rate of surface runoﬀ. However the process of forming groundwater sapping valleys involves the following steps, as outlined in Figure 2.26 (Dunne, 1980):

• A single rapid uplift brings a smooth land surface of permeable rocks above sea level (Dunne, 1980).

• Disturbance of the flow leads to the flow concentration towards a more permeable zone where erosion occurs . A spring head is excavated leading to further flow concentration at the same site (Dunne, 1980).

• Retreat of spring head increases ﬂow convergence and the potential for future seepage erosion. Water emerging along the valley sides exploits a susceptible zone to form a tributary which also undergoes headward retreat (Dunne, 1980).

• The process of repeated failure, headward retreat and branching forms a network of valleys. The pattern becomes stable when the decreasing drainage area of each spring head is no longer large enough to supply enough groundwater to cause erosion of the bedrock (Dunne, 1980).

Evaluating the characteristics of the Dry Valleys of Long Island it can be interpreted 40

that groundwater sapping process might have played a fundamental role in modifying and shaping the pre-existing surface runoﬀ valley system. The possible model that describes how groundwater sapping may have occurred in a manner slightly modiﬁed from that of (Dunne, 1980) was important, is as follows.

After the glacier retreated around 18000-20000 years ago, the warming of climate started, which resulted in the melting of permafrost. This period was called the Bolling-Allerod. The warming event started around 14500-13000 years ago , which led to much of the warming in the Northeastern United States (Ellis et al., 2004). During this warm period, the active layer of the permafrost started thawing from the surface downward and it gradually disappeared from the ground. This enhanced the increased inﬁltration of precipitation into the ground. At this point, a perched water table could have developed beneath the surface of Long Island. A similar situation has been depicted in frozen mineral substrate of Canada Carey and Woo (2001). The groundwater in this perched water table may have started sapping the Dry Valleys by headward erosion process. Thus a straight, parallel drainage network with rectangular watershed, low order of tributaries and steep sloping transverse proﬁle evolved. The Cannan Dry Valley represents such a characteristic drainage network.

As the permafrost may have continued to melt, the water table associated with the perched water table descended below the bottom of the valleys. The valleys on Long Island became dry. Finally when the water table started rising associated with the sea level rise during the Holocene period, the bottoms of the Dry Valleys which were at the same elevation as the water table filled with groundwater. This explains why there are water filled valleys (streams) along the lower reaches of the Dry Valleys. The sea level and water table are still rising and groundwater is now filling up these Dry Valleys.

The Characteristics of the Dry valleys have been determined and measured using the Digital Elevation Model and its tools. These characteristics which helped me to understand the processes responsible for their formation are listed in the Table 2.1. 41

Table 2.1: Characteristics of some of the Dry Valleys south of Ronkonkoma Moraine in Long Island

Dry Valley Details Other Features

Watershed area (A) 14.8sq.m Order 3 Valley Drainage Density (Dd) 3.33 Straight Dry Valley with Stream Bifurcation ratio (Rb) 4 Rectangular watershed. Slope of the valley 0.003 Sinuosity 0.92

Watershed area (A) 17.1sq.m Straight and angulate drainage Order 2 East pattern. Dd is less suggesting Drainage Density (Dd) 0.53 Meadow coarse drainage network. This Bifurcation ratio (Rb) 4 Brook indicates water table was low at the Slope of the basin 0.002 time of formation of the stream. Sinuosity 1

Watershed area (A) 18.5sq.m Orders (n) 3 A small dendritic pattern is New Mill Drainage Density (Dd) 0.95 present. Tributaries are longer Pond Bifurcation ratio (Rb) 4 than the main trunk channel. Slope of the basin 0.002 Sinuosity 0.38

Watershed area (A) 14.7sq.m Straight, parallel drainage network. Order 3 Lower order implies low streamﬂow Drainage Density (Dd) 0.723 velocity, and hence less power of Cannan Bifurcation ratio (Rb) 8.85 the streams. Rb>5 indicates that Slope of the basin 0.002 stream pattern may be controlled Sinuosity 0.808 by the area structure of the area. Continued on Next Page. . . 42

Table 2.1 – Continued

Dry Valley Details Other Features

Watershed area (A) 17.3sq.m Order 4 Drainage Density (Dd) 1.113 Straight angulated pattern of the Swan Bifurcation ratio (Rb) 4.3 streams. Tributaries are larger Slope of the basin 0.003 than the trunk channel. Sinuosity 0.925

Watershed area (A) 22.6sq.m Order 4 Straight drainage pattern with Drainage Density (Dd) 0.8 more number of tributaries. Forge Bifurcation ratio (Rb) 3.54 Lateral growth is more than Slope of the basin 0.002 headward growth. Sinuosity 1.036

Watershed area (A) 19.7sq.m Order 4 Circular drainage basin which Drainage Density (Dd) 1.85 encloses a dendritic drainage Setuck Bifurcation ratio (Rb) 19.2 pattern. Higher Rb implies that Slope of the basin 0.002 the local topography may play an Sinuosity 0.617 important role in its formation .

Watershed area (A) 8.6sq.m Order 3 Dendritic drainage pattern. Drainage Density (Dd) 1.9 Streams are found to cut through Hayground Bifurcation ratio (Rb) 2.69 the ponds which are believed to be Cove Slope of the basin 0.003 formed by the burying of the ice. Sinuosity 0.925 These are called kettle ponds.

Continued on Next Page. . . 43

Table 2.1 – Continued

Dry Valley Details Other Features

Watershed area (A) 11.2sq.m Order 3 Lateral growth of the tributaries is Drainage Density (Dd) 1.9 more than the headward growth Georgica Bifurcation ratio (Rb) 4 resulting in longer tributaries and Slope of the basin 0.003 shorter trunk channel. Sinuosity 0.92 44

Chapter 3

Models

In this chapter, we consider the proposed models for the formation of the Dry Valleys south of Ronkonkoma Moraine on Long Island. All the ages are in terms of calendar years. Radiocarbon ages are calibrated following Calib Radiocarbon Calibration (Stuiver et al., 2005)

3.1 Stage 1

More than 25,000 years ago (Sirkin and Mills, 1975; Sirkin, 1996), the Laurentide Ice sheet advanced over Long Island during Last Glacial Maximum. At this time the area was desert tundra (Sirkin et al., 1970). The Ronkonkoma moraine, which is Suffolk’s spine and South Fork was formed during this advance of this glacier (Stoffer and Messina, 1996). It is believed that the Ronkonkoma Moraine is a typical push moraine and was formed by the thrusting of ice against sediments (Figure 3.2). The glacier advanced beyond the Ronkonkoma Moraine to deposit ice at several places creating depression filled with buried ice, some of these being located in the eastern part of Long Island and are now known as Scuttlehole Ponds (Figure 3.4). An aerial view of the Scuttlehole depressions as seen on the DEM indicates that these ponds have circular to elliptical shapes similar to kettle holes (Figure 3.2).

3.2 Stage 2

The glacier started receding- around 20,000 years ago (Sirkin, 1982). The environment was typically periglacial characterized by very cold and dry conditions. The intense cold weather was coupled with a treeless landscape dominated by desert tundra (Sirkin, 1977; Peteet et al., 1994). The mean temperature was approximately -5 to -10◦C. Permafrost 45

was present on Long Island and from the mean annual temperature it can be suggested that the possible thickness of permafrost was a few hundred meters (300-400m) below the ground (Nathenson and Guﬀanti, 1987). Since Long Island is a wedge of Cretecous and glacial unconsolidated sediments of 300-400 m and it is about 600 m thick at the south shore Sirkin (1974) it is believed that the permafrost might have extended till the bottom of the bedrock in Long Island (Figure 3.2).

3.3 Stage 3

The frozen, impervious permafrost bearing surface of Long Island led to the generation of surface runoﬀ as shown in Figure 3.2. The surface runoﬀ stream valleys had a high drainage density and exhibited a dendritic pattern. The DEM of Long Island shows a number of dendritic Dry Valleys located to the east (Hanson, 2005). During this time the tributary streams of Hayground Cove Dry Valley, near the Scuttlehole depressions developed on the frozen sediments on top of the buried ice in the depressions. This implies that the Scuttlehole depressions or in other words the kettle lakes in Scuttlehole regions are the result of Last Glacial Maximum. If there were no buried ice in those depressions the tributary streams of the main streams of present day’s Setuck Dry Valley would not have cut across them (Figure 3.2).

3.4 Stage 4

The cold, dry, periglacial condition lasted till 14,680+/-400 years ago(from the GISP2) (Laura et al., 2005; Weaver et al., 2003) when the climate started warming (Bolling Allerod) and this resulted in the melting of permafrost from the surface. Studies have revealed that permafrost thaws both from the bottom and the top as in Figure 3.1. Earth’s heat flow is responsible for the upward thawing of permafrost and the climatic heat flow is responsible for the downward thawing of permafrost (Szewczyk, 2005). At this time the vegetation on Long Island was mainly deciduous forest characterized by an abundance of Oak and Birch (Peteet et al., 1994; Shuman et al., 2003), (Figure 1.3). The downward thawing of the permafrost allowed the infiltration of water creating a perched water table above the 46

Figure 3.1: Permafrost thaws due to two sources of heat, Szewczyk, 2005.

ice. Where the perched water table was at the same level as the bottom of the valleys, the groundwater started sapping these valleys as in (Figure 3.3) and led to the formation of groundwater sapped stream valleys at the lower ends of these valleys.

3.5 Stage 5

About 12000 years ago the climate began warming again and permafrost continued to melt. The warming trend was coupled with a further increase in thermophilus taxa like Tsuga, Betula and Oak (Prentice et al., 1991; Peteet et al., 1994). Permafrost continued to melt from the top and from below. The perched water table decreased in elevation and at some time dropped below the bottom of the stream valleys leaving them dry, Figure 3.3.

3.6 Stage 6

During the Holocene as the sea level rose the water table also rose and when it intersected the bottom of the Dry Valleys the groundwater began to reoccupy the valleys on their southern ends. 47

(a)

Interlobate zone Permafrost

Depression filled with ice (b)

Scuttlehole Depressions

Surface Runoff Streams

(c)

Figure 3.2: Models describing formation of Dry Valleys in Long Island: (a) At the time of Ronkonkoma Glacier advance(60-20 kyr), (b)Ronkonkoma Glacier retreat(20-18 kyr), (c)During time when of permafrost until about 15kyr 48

Groundwater Sapped Stream

(d)

Dry Kettle Hole

(e)

Figure 3.3: Models describing formation of Dry Valleys in Long Island: (d) During time of perchedwater table on melting permafrost (15-10kyr), (e)After permafrost melted and before water table rose with post-glacial sea level rise (10kyr to ?) 49

Figure 3.4: Scuttlehole Ponds alligned in the nw-se direction, Sag Harbour, Long Island. 50

Chapter 4

Conclusion

Long Island witnessed the development of permafrost after glacier retreated around 18000- 20000 years ago. The climate was cold and desert tundra. On this frozen impervious ground the surface runoff that prevailed during spring/summer melt period created the surface runoff streams with dendritic drainage network and a large number of tributaries. This period was then followed by warming event during the Bolling Allerod around 15000 years ago. The permafrost melted from the top and a perched water table may have developed. The groundwater of this perched water table started sapping the lower ends of the runoff tributaries while the upper ends at higher elevations became dry. These higher ends of runoff tributaries are now preserved as ghost tributaries. After the climatic warming the perched water table dropped below the bottom of the stream valleys and they became dry. Present the sea level is rising and the groundwater table is also rising. The present groundwater table is reoccupying the lower ends of the Dry Valleys. 51

Bibliography

Aharonson, O., Zuber, M.T., Rothman, D.H. and Whipple, N.S.X., 2002, Drainage basins and channel incision on Mars, Proc Natl Acad Sci U S A, vol. 99(4), pp. 1780–1783.

Allison, T., Moeller, R. and Davis, M., 1986, Pollen in Laminated Sediments Provides Evidence For a Mid-Holocene Forest Pathogen Outbreak, Ecology, vol. 67(4), pp. 1101– 1105.

Bennington, J.B., 2003, New observations on the glacial geomorphology of Long Island from a digital elevation model DEM., 10th Conference - Geology of Long Island and Metropoli-

tan New York, vol. 10, URL http://pbisotopes.ess.sunysb.edu.lig/conferences/ abstracts/bennington/.

Cabe, M., Kelly, M., Schaefer, M., Schwartz, R., Hanson, G. and Benimoﬀ, A., 2006, Dating Glacial Features in New York’s Lower Hudson Valley- The Last Deglaciation of the Eastern Laurentide Ice Sheet, GSA abstracts.

Carey, S. and Woo, M., 2001, Slope runoﬀ processes and ﬂow generation in a subarctic, subalpine catchment, Journal of Hydrology, vol. 253, pp. 110–129.

Cofaigh, O.C., 1996, Tunnel valley genesis, Progress in physical geography, vol. 20(1), pp. 1–19.

Denny, C.S., 1936, Periglacial Phenomena in Southern Connecticut, American Journal of Science, vol. 32, pp. 322–42.

Dunne, T., 1980, Formation and controls of channel networks, Progress in Physical Geog- raphy, vol. 4, pp. 211–239. 52

Dunne, T., Price, A.G. and Colbeck, S.C., 1976, The generation of runoﬀ from subarctic snowpacks, Water Resources Research, vol. 12 (4), pp. 677–685.

Elias, S., Anderson, K. and Andrews, J., 1996, Late Wisconsin climate in northeastern USA and southeastern Canada, reconstructed from fossil beetle assemblages, Journal of Quaternary Science, vol. 11(5), pp. 417–421.

Ellis, K., Mullins, H. and Patterson, W., 2004, Deglacial to middle Holocene (16,600 to 6000 calendar years BP) climate change iin the northeastern United States inferred from multi- proxy stable isotope data, Seneca Lake, New York, Journal of Paleolimnology, vol. 31(3), pp. 343–361.

Fairbanks, R., 1989, A 17,000-year glacio-eustatic sea level record: inﬂuence of glacial melting rates on the Younger Dryas event and deep-ocean circulation, Nature, vol. 342, pp. 637–642.

Ferguson, R., Ashmore, P., Ashworth, P., Paola, C. and Prestegaard, K., 1992, Measure- ments in a Braided River Chute and Lobe- Flow Pattern, Sediment Transport, and Chan- nel Change, Water Resources Research, vol. 28(7), pp. 1877–1886.

Foster, D., Swift, B., Schwab, W., Army, Survey, G., States, U. and of Engineers, C., 1999, Stratigraphic Framework Maps of the Nearshore Area of Southern Long Island from Fire Island to Montauk Point, New York, The Survey.

Foster, R., Koski, C. and Johnson, B., 1997, Tundra, Online, URL http://ths.sps.lane. edu/biomes/tundra5/tundra5.html.

French, H.M., Demitroﬀ, M., Forman, S.L. and Newell, W.L., 2007, A Chronology of Late- Pleistocene Permafrost Events in southern New Jersey, Eastern USA, Permafrost and Periglacial Processes, vol. 18, pp. 49–59.

Fuller, M.L., 1914, The Geology of Long Island, US Geological Survey Professional Paper, vol. 82, p. 231.

Greenwood, G., 1877, River Terraces, London, p. 247. 53

Gregory, K.J., 1966, Dry valleys and the composition of the drainage net, Journal of hydrology, vol. 4, pp. 327–340.

Gregory, K. and Walling, D., 1973, Drainage Basin Forms and Processes - A Geomorpho- logical Approach, New York, Wiley, 456p.

Grimes, K., 1995, The Vulcon Guidebook - Glossary, http:// home.mira.net/ gnb/ caving/ papers/ kg-svckt.html.

Hanson, G.N., 2005, Glacial Geology of the Stony Brook-Setauket-Port Jeﬀerson Area, Report., Stony Brook University, http:// pbisotopes.ess.sunysb.edu/ reports/.

Hanson, L., 2006, Fluvial processes and Landforms, Report, http://members.aol.com/ gls210/ streams.htm.

Horton, R.E., 1933a, The Horton Papers.

Horton, R., 1932, Drainage Basin Characteristics, Transactions of the American Geophysical Union, vol. 13, pp. 351–60.

Horton, R., 1933b, The role of inﬁltration in the hydrological cycle, Transactions of the American Geophysical Union, vol. 14, pp. 446–460.

Katasonova, E.G., 1963, The role of thermokarst in the development of delly- Conditions and special features of the formation of frozen horizons in Siberia and in Far-east(in Russian), Moscow, Isd-vo AN SSSR, pp. 81–100.

King, C., Mion, L., Pacholik, W. and Hanson, G., 2003, Evidence of Till South of

Ronkonkoma Moraine, Conference-abstract, URL http://www.geo.sunysb.edu/lig/ Conferences/abstracts-03/king/king-index.htm.

Kneller, M. and Peteet, D., 1999, Late-Glacial to Early Holocene Climate Changes from a Central Appalachian Pollen and Macrofossil Record, Quaternary Research, vol. 51(2), pp. 133–147.

Knighton, D., 1998, Fluvial Forms and Processess:: a New Perspective, Oxford University Press US. 54

Kochel, R. and Piper, J., 1986, Morphology of large valleys on Hawaii: Evidence for groundwater sapping and comparisons with Martian valleys, Journal of Geophysical Research, vol. 91(B13), pp. E175–E192.

Kundic, V. and Hanson, N., 2006, Evidence of Permafrost in Long Island, abstracts, De- partment of Geosciences, Stony Brook University, http://www.geo.sunysb.edu/ lig/ Con- ferences/ abstracts06/ kundic-06.pdf.

Laura, F., Adkins, J., Keigwin, L., Southon, J., Fernandez, D., Wang, S. and Scheirer, D., 2005, Radiocarbon Variability in the Western North Atlantic During the Last Deglacia- tion, Science, vol. 310(5753), pp. 1469–1473.

Logan, J., 1983, Nitrogen oxides in the troposphere-Global and regional budgets, Journal of Geophysical Research, vol. 88, pp. 10785–10807.

Mayewski, P.A., Meeker, L.D., Morrison, M.C., Twickler, M.S., S.I.Whitlow, Ferland, K.K., Meese, D.A., Legrand, M.R. and J.P.Steﬀensen, 1993, Greenland ice core ‘signal’ characteristics - An expanded view of climate change, Journal of Geophysical Research, vol. 98, pp. 839–847.

Melton, M., 1957, An Analysis of the Relations Among Elements of Climate, Surface Prop- erties, and Geomorphology, Tech. rep., Dept. of Geology, Columbia University.

Merrill, F. et al., 1902, Metamorphic crystalline rocks of the New York City quadrangle, Tech. rep., Description of the New York City district: United States Geological Survey- Geologic Atlas of the United States, New York City Folio, No. 83, 19 p.(Includes colored geologic map on a scale of 1:62,500).

Meyers, 1998, Studying exposed sections on Long Island: Stratigraphy of Ronkonkoma

Moraine exposed in Ranco Quarry (Manorville), URL http://www.geo.sunysb.edu/ lig/meyers/html.

Nathenson, M. and Guﬀanti, M., 1987, Compilation of geothermal -gradient data in the conterminous United States, Open-ﬁle report 87-592, Department of the Interior U.S. Geological Survey. 55

Newman, S.W., 1977, Late Quaternary Paleoenvironmental Reconstruction: Some contra- dictions from Northwestern Long Island, New York, Annals of the New York Academy of Science, vol. 288, pp. 545–570.

Nieter, W., 1975, A late Wisconsin loess deposit in southeastern Long Island, New York, Master’s thesis, Queens College. Department of Earth and Environmental Sciences.

Ojakangas, R. and Charles, L., 1982, Minnesota’s Geology, University of Minnesota Press.

Ollier, C. and Thomasson, A., 1957, Asymmetrical Valleys of the Chiltern Hills, The Geo- graphical Journal, vol. 123(1), pp. 71–80.

Penck, W., 1924, Morphological analysis of landforms, trans. H. CZECH and K.C.BOSWELL, p. 429.

Peteet, D., Daniels, R., Heusser, L., Vogel, J., Southon, J. and Nelson, D., 1994, Wiscon- sinan Late-glacial environmental change in southern New England: A regional synthesis, Journal of quaternary science, vol. 9(2), pp. 151–154.

P´ew´e, T., 1969, The Periglacial Environment-past and present, Montreal, McGill-Queens University Press.

Prentice, I., Bartlein, P. and Web III, T., 1991, Vegetation and Climate Change in Eastern North America Since the Last Glacial Maximum, Ecology, vol. 72(6), pp. 2038–2056.

Reid, C., 1887, On the origin of dry chalk valleys and the coombe rock, Quaterly Journal of Geological Society, vol. 43, pp. 364–73.

Rosgen, D., 1994, A Classiﬁcation of natural rivers, Catena, vol. 22, pp. 169–199.

Schmitt, K., 2006, More Evidence of Till South of the Ronkonkoma Moraine, 13th Confer-

ence - Geology of Long Island and Metropolitan New York, URL http://pbisotopes. ess.sunysb.edu.lig/conferences/abstracts/.

Schmitthenner, H., 1926, The origin of Dellen and their morphological signiﬁcance, Zeitschrift fur Geomorphologie, Band 1, Translated from the German by Franchesca Vin- cent in collaboration with C.F.Stewart Sharpe, pp. 3–28. 56

Schostakowitsch, W.B., 1927, Der ewig gefrorene Boden Sibiriens, Gesell. Erdkunde Berlin Zeistschr, pp. 394–427.

Sharpe, C., 1941, Geomorphic aspects of normal and accelerated erosion, Symposium of Dynami of Land Erosion, vol. 21, pp. 558–570.

Shuman, B., Newby, P., Huang, Y. and Webb III, T., 2003, Evidence for the close climatic control of New England Vegetation History, Ecology, vol. 85(5), pp. 1297–1310.

Sirkin, L., 1977, Late Pleistocene Vegetation and Environments in the Middle Atlantic Region, Amerinds and their Paleoenvironments in Northeastern North America. The New York Academy of Sciences, New York, pp. 207–217.

Sirkin, L., 1982, Wisconsinan glaciation of Long Island, New York to Block Island, Rhode Island, in Larson, G.J., Stone, B.D., eds. Late Wisconsinan glaciation of New England:, Kendall/Hunt.

Sirkin, L., 1974, Microﬂora in the Sangamon and younger beds of the Delmarva peninsula, Delaware, Maryland and Virginia, Geological Society of America Abstracts with Programs, vol. 5(1), p. 74.

Sirkin, L. and Mills, H., 1975, Wisconsinan glacial stratigraphy and structure of northwestern Long Island, TripB5, Guidebook to Field Excursions: New York State Geological Association, Annual Meeting, 47th, Department of Geology, Hofstra University, Hemp- stead, NewYork, vol. 327, pp. 299–327.

Sirkin, L., Owens, J., Minard, J. and Rubin, M., 1970, Palynology of some Pleistocene peat samples from the coastal plain of NewJersey, Tech. rep., U.S.Geol.Survey of Professional Paper 700-D D77-D87.

Sirkin, L., 1996, Western Long Island Geology: History, Processes & Field Trips, Book & Tackle Shop.

Stoﬀer, P. and Messina, P., 1996, The Transient Atlantic Shoreline, Tech. rep., CUNY Earth

and Environmental Science, URL http://www.geo.hunter.cuny.edu/bight/coast. html. 57

Stone, R. and Ashley, M., 1992, Ice-wedge casts, pingo scars, and the drainage of Glacial Lake Hitchcock, 84th Annual Meeting of the New England Intercollegiate Geological Con- ference, Amherst, vol. 2, pp. 305–331.

Strahler, A., 1952, Dynamic basis of geomorphology, Bulletin of the Geological Society of America, vol. 63(9), p. 923.

Strahler, A., 1957, Quantitative analysis of watershed morphology, Trans.Amer.Geophys.Union, vol. 38, pp. 913–920.

Strahler, A., 1964, Quantitative geomorphology of drainage basins and channel networks, Handbook of Applied Hydrology, pp. 40–74.

Stuiver, M., Reimer, P.J. and Reimer, R., 2005, CALIB Radiocarbon Calibration, pp. radiocarbon calibration.

Szewczyk, J., 2005, The frozen time, Tech. rep., Polish Geological Institute, URL http: //www.pgi.gov.pl.

Terell, T., 2005, Riparian Plants of New Mexico, Tech. rep., New Mexico State University, http://cahe.nmsu.edu/riparian/WHTRIPAREA.html.

Walker, E.H., 1948, Diﬀerential erosion on slopes of the northern and southern exposure in Western Wyoming, Bull. Geol.Society of America, vol. 59, p. 1360.

Weaver, J., Saenko, A., Clark, U. and Mitovica, X., 2003, Meltwater Pulse 1A from Antarc- tica as a Trigger of the Bølling-Allerød Warm Interval, Science, vol. 14, pp. 1709–1713.

Woo, K., 1999, Snow, frozen soils and permafrost hydrology in Canada, Hydrological Pro- cesses, vol. 19, pp. 215–229.