Longitudinal Ripples in the Upper Tunbridge Wells Delta, Kent, and their Probable Mode of Origin by F. G. BERRY Received 20 January 1960 CONTENTS I. INTRODUCTION 2. LOCATION AND GENERAL STRATIGRAPHY 3. DESCRIPTION OF THE SEQUENCE .•• 4. DESCRIPTION OF THE LONGITUDINAL RIpPLES 5. SUGGESTED MODE OF ORIGIN 6. DISCUSSION 7. CoNCLUSIONS ACKNOWLEDGMENTS REFERENCES
33 33 34 35 36
37 38 38 39
ABSTRACT: Structures interpreted as longitudinal ripples similar to those found in marine sediments are described for the first time from non-marine sediments and a hypothesis explaining their formation is developed. Ripples of this type are likely to form during the transport of small quantities of sand by relatively fast currents across mud-flats.
1. INTRODUCTION cut in fine-grained sediments (van Straaten, 1951) are distinctly rare in non-marine environments. In fact, apart from a brief reference to giant forms in rivers (Chenoweth, 1955) there appear to be no previous reports or descriptions in the literature. The occurrence here described yields details of their structure which endorse the observations of van Straaten as to their erosional origin-although it is recognised that similar forms may arise by accumulation (van Straaten, 1953). Either type may be subsequently propagated by conformable sedimentation (Kelling, 1958). LONGITUDlNAL R1PPLES
2. LOCATION AND GENERAL STRATIGRAPHY A degraded section at Sandhole Farm, five miles west of Tonbridge, Kent (509549),1 exposes the junction between a fine-grained laminated sandstone and gray silty clays above. These occur well up in the third Wealden megacyclothem (Upper Tunbridge Wells-Lower Weald Clay) postulated by Allen (1959) and confirmed locally by the writer's detailed 1
All Grid References lie within the 100 km, square 51 (TQ) 54.
33 PROC. GEOL. ASSOC., VOL. 72, PART 1, 1961
mapping. They are some distance below the abundant red-mottled silty clays which are customarily taken as the base of the Weald Clay. To the west of Sandhole the sandstone passes beneath the Eden alluvium and is not known to reappear. To the east it is breached by a small tributary of the Eden, but reappears in the vicinity of Moorden Farm (523459). Between this point and Redleaf Wood it begins to thicken and form natural bluffs. It disappears from outcrop within the next hundred yards. The available field evidence strongly suggests that the sandstone is a linear deposit of 'lenticular' cross-section, following the local strike trend where exposed and truncated at Redleaf Wood by a change in topography. The silty clays above-about thirty feet thick at Sandhole-pass vertically and laterally into flaggy siltstones interbedded with unlithified silt, silt-day grades and waxy clays, usually with, inter alia, some fine sands. Much ofthis type of sediment is currently regarded as leveealluvium. 3. DESCRIPTION OF THE SEQUENCE The passage from the sandstone to the silty clay above is associated locally with retreat structures similar to those described by Allen (1949, 1959) from the Ashdown-Wadhurst and Lower Tunbridge Wells-Grinstead junctions (Fig. 1). The basal sands (a) are followed by a series of oscillation ripple systems (b) derived from current-formed precursors, built,
UPPER TRANSVERSE RIPPL.ES
LOWER, TRANSVERSE RIPPLES
Fig. 1. The 'retreat' sequence at Sandhole Farm, Kent, with orientation histograms of the lower transverse ripples from 'b', groove-load-casts from the base of 'c', longitudinal ripples from the upper surface of 'c' and upper transverse ripples from 'd', Coarse stippling = very fine sands, fine stippling = silts, white = clay-silt grades
LONGITUDINAL RIPPLES IN TUNBRIDGE WELLS DELTA 35
near the base, of very fine sand and, higher up, of coarse and fine silts. They are separated by clayey silts. This ripple sequence is succeeded by a relatively thick, very fine sand-siltstone (c) the base of which bears a variety of groove-load-casts. At one point a diminutive piercement structure formed by the injection of the underlying clayey silts has bent these up into a near vertical position. The upper surface of this unit is furrowed by the system of longitudinal ripples. The siltstone is followed abruptly by more clayey silts (d) containing only the faintest traces of silt-crowned transverse ripples, some of which occur only two or three centimetres above the longitudinal ripples. No pebble or rootlet beds have so far been observed and it is doubtful whether these retreat structures, in the absence of thick clays of Wadhurst or Grinstead type, signal the onset of a widespread lake transgression. The orientations of oscillation ripples above and below the siltstone (c) and the top- and bottom-surface structures are shown in circular histograms in Fig. 1. It would appear from their arrangement that the wave oscillations were normal to the current trends that produced the longitudinal structures and that the strikes of the current-formed precursors have been preserved. The arithmetic means of these trends are 355, 343, 328 and 351 degrees, in ascending order. Whether currents moved predominantly to the north or south is not clear, for ripples propagate in opposed directions and groove-Ioad-cast bifurcation is also inconsistent. 4. DESCRIPTION OF THE LONGITUDINAL RIPPLES In plan the ripples show an essentially regular and parallel arrangement, with an occasional bifurcation. Fig. 2 shows a series of transverse sections through one of these bifurcations. The surface profile is smoothly rounded but less regular than that described by van Straaten (1951). The wavelengths of this surface vary between 1.5 and 7.0 centimetres (average 4.2 centimetres) with amplitudes ranging from 2.0 to 0.3 centimetres (average 0.7 centimetres). These yield ripple indices of from 18.0 to 3.5. The smooth profile is due to a coating of fine silts masking an underlying and more irregular surface (indicated in the figure by heavy lines). In all cases where the original, horizontal laminations in the siltstone can be observed it is cut by this surface, which is clearly an eroded one. The silt covering conceals a variety of slots and shallow, round or flatbottomed grooves in the troughs. The base of the infilling then usually consists of very fine, sharp sands along some part of the grooves. The fine silts above the sands generally show discordant bedding, but the laminae sometimes curve upwards at steep angles near the sides (Fig. 2c). This may be due, in part, to compaction, though some sections (a) indicate that it is
F. G. BERRY
not. Two distinct layers of sand are not infrequently present, separated by fine silts. In longitudinal section the infilling appears to be level bedded. The fact that some of the grooves have vertical walls and are occasionally undercut (a) suggests that the muds were partially consolidated before the grooves were cut. One small slump has been observed in the fine silt covering.
14 CMS. 4CMS. Fig. 2. Transverse serial sections through the bifurcation of a longitudinal ripple. (Distance between the sections is indicated at the side)
5. SUGGESTED MODE OF ORIGIN By analogy with aeolian conditions (Bagnold, 1942, 117) it seems that strong currents may favour transverse instability. Casey (in Bagnold) found that longitudinal strips of material were deposited under water only when the velocity of flow was raised above a certain level and was associated with regular transverse elements These facts apply primarily to accumulative processes and will no doubt lead to an explanation of the way in which certain types of longitudinal ripples are formed; but linked with the foregoing description an hypothesis concerning the origin of indubitable erosion-formed longitudinal ripples may be developed Van Straaten (1951) recorded the occurrence of transverse current ripples in sand adjacent to longitudinal ripples in muds Both were clearly stable under conditions of similar current strength-implying a considerable difference in the nature of the bottom flow over the two surfaces. This
LONGITUDINAL RIPPLES IN TUNBRIDGE WELLS DELTA 37
is evidently related to bottom roughness and the power of the individual gra ins to move and interact with the boundary layers (Inman, 1949). Over sands current flow is retarded (whether ripples are present or not) while over adjacent muds, current flow is, relatively speaking, accelerated. However, if small quantities of sand become washed over the muds from adjacent areas (or, alternatively, are winnowed from them) bottom flow over the muds will tend to move in preferred paths (since flow over the portions covered by the sands is retarded) so drawing streamers of sand along in the direction of the current. These will remove particles beneath them into suspension by traction and impact, assisted by greater bottom roughness, and thus promote the formation of grooves. Current speed must eventually diminish and the sand grains will accumulate then in the bottoms of the grooves-generally where they are widest (Fig. 2 b). At this point the oscillatory movements due to waves become effective and impart slight to-and-fro movements to the sand grains which locally deepen the grooves. Ultimately, the finer material begins to settle from suspension to complete the filling. At this stage a local resurgence of currents or wave oscillations may form double layers of sand and cut into the groove filling to form the curved laminae (Fig. 2a). A more general form of scour will continue so long as material remains in suspension-gently rounding the profiles of the grooves. It is clear that longitudinal ripples can form only when the supply of sand being carried over the muds remains below some (unknown) proportion, appropriate to current strength; for if it should exceed this amount, bottom flow will be retarded and transverse ripples form over an unbroken blanket of sand in which preferred paths cannot be established (this situation should not be confused with the formation of certain groove-casts and similar structures found on the base of thick-graded beds and possibly due to a similar mechanism). The supply of sand will also influence the wave-length (mean spacing)-again in relat ion to current strength, since it controls the width and thickness of the slow streams and hence their ability to resist longitudinal dissection by faster flow.
6. DISCUSSION Contrary to Kelling (1958, 129) there seems to be nothing in the presence of flat-bottomed grooves and uniform parallel ridges to invalidate a scour hypothesis operating in the manner and conditions outlined above. By analogy with aeolian conditions again (Bagnold, 1942, 232) a uniform and parallel arrangement is to be expected in generally uniform conditions of flow over level mud-flats; the uniformity is related to secondary transverse circulation or the probability th at eddy zones between fast and slow streams are of similar competence, while their parallel arrangement is due
to the absence of small bumps or other obstructions that could divert the streams. The part played by wave-height is uncertain. Van Straaten (1951) found a correlation between (a) wave-height (b) wave-length of the longitudinal ripples and (c) depth of water. That depth of water should be related to wave-length of the longitudinal ripples seems very probable since depth of water is also related by van Straaten to current velocity-an integral part of the process of erosion. It seems significant that the smaller wave-lengths he reported occurred in the region of fastest currents, where the sand streams might well be attenuated and hence split into thinner and more frequent streams, so producing longitudinal ripples closer together. Regarding the source of the currents which formed the longitudinal ripples, much depends on the correct interpretation of the adjacent sediments. The basal sands may represent, as their linearity suggests, the bar underlying a minor distributary pass and the local grey clays,accumulations within an abandoned channel receiving intermittent flood waters. If the currents are attributed to crevassing the infrequence of this type of longitudinal ripples is surprising, since the sediments requisite in their formation constitute the bulk of the Hastings Beds. It seems more probable that they are passed over as ripples of normal type. 7. CONCLUSIONS (a) Longitudinal ripples superficially similar to those of van Straaten (1951) are described from a non-marine and presumably tide-less environment. (b) The ripples were formed by erosion. (c) Evidence is reviewed which supports the hypothesis that ripples of this type are most likely to form on level mud-flats when a small quantity of sand is being transported across them by fast currents. (d) The parallel and generally uniform appearance of the ripples is no more than would be expected in the conditions of uniform flow that may occur over mud-flats. (e) Wave-length may be related to strength of current flow and quantity of sand available. ACKNOWLEDGMENTS The author gratefully acknowledges the assistance given by Dr. G. Kelling and Dr. C. Tozer in reading and constructively criticising the manuscript of this paper.
LONGITUDINAL RIPPLES IN TUNBRIDGE WELLS DELTA 39
REFERENCES ALLEN, P. 1949. Notes on Wealden Bone Beds. Proc. Geol. Ass. , Lond. , 60, 275-93. - - - . 1959. The Wealden Environment-Anglo-Paris Basin. Phil. Trans., B., 242, 283-346 . BAGNOLD, R. A. 1942. Physics of Blown Sands and Desert Dunes. London. CHENOWETH, P. A. 1955. Unusual Type of Ripple-mark. Bull. geol. Soc. A mer. (abstracts),66, 1541-2. INMAN, D. L. 1949. Sorting of Sediments in the Light of Fluid Mechanics. J. sediment. Petrol., 19 (2), 51-70. KELLTNG , G . 1958. Ripple Mark in the Rhinos of Galloway. Trans . Edinb. geol. Soc., 17 (2), 117-32. VAN STRAATEN. 1951. Longitudinal Ripple Marks in Mud and Sand. J. sediment. Petrol ., 21 (1),47-54. - - - . 1953. Rhythmic Patterns on Dutch North Sea Beaches. Geol. Mijnb, N.W., Ser. 15,31-43.
F. G. Berry 12 Malden Road S.W.IS