Analysis of Fluvial Depositional Systems

By Andrew D. Miall


Consisting of 12 chapters, this publication contains information on sediment types and transport modes, channel morphology, methods of facies analysis, bars, facies models for alluvial fans and braided rivers, facies models for meandering rivers, facies models for anastomosed rivers, recognition of large rivers, large scale fluvial cycles, basin architecture and tectonic settings, and fuels and minerals in fluvial deposits.

  1. Page 2

    Fluvial deposits are predominantly clastic, and range in grain size from giant boulders to the finest clay. Coal is commonly associated with fluvial sediments, and some sequences contain carbonate concretions representing fossil soil horizons.

    Sediment is transported in three ways:

  2. Page 4

    The most important control on the geometry of fluvial deposits is channel morphology. This is defined using two parameters, sinuosity and a braiding parameter (Rust, 1978a).

    Sinuosity is defined as the ratio of thalweg length to valley length. The thalweg is the line of deepest channel. Valley length is the straight line distance down that part of the valley over which the thalweg length is measured. Rivers are divided arbitrarily into those of high sinuosity (>1.5) and low sinuosity (<1.5).

    The braiding parameter (B.P.) measures the number of bars or islands in the channel, and thus also defines the channel multiplicity. Rust (1978a) suggested that each braid or island be defined by using the mid-line of the channels surrounding it. Theoretically this avoids the problem that the number of bars or islands increases as water level drops, although in practice it is difficult to define channels or submerged islands if a river is in full flood. The braiding parameter is defined as the number of bars or islands per meander wavelength. Single channel rivers are those with a braiding parameter less than 1, and multiple channel rivers are those with a braiding parameter greater than 1.

    These two indices define four basic channel types (Table 1), as illustrated in Figure 1.

    Braided rivers are characterized by high width/depth ratios, normally greater than 40 and commonly exceeding 300. Leopold and Wolman (1957) found that they have steeper slopes than other types

  3. Page 10

    Systematic study of fluvial deposits requires at least three main steps. First the major and minor lithofacies components must be identified and described. Secondly, these should be studied to determine important lithofacies associations and internal relationships; and thirdly the geometry and orientation of the depositional units should be studied. A series of facies models is available with which the results of these analyses can be compared in order to arrive at a satisfactory interpretation. These steps are taken one at a time in the following discussion:

  4. Page 20

    Fluvial channel systems consist of active and inactive channels, bars and stable (commonly vegetated) islands. Bars represent areas of net sediment accumulation and are therefore of considerable interest to sedimentologists.

    There is a great deal of confusion in the fluvial literature about what constitutes a bar, and over thirty terms have been used to name bars with specific shapes, orientations and positions within a channel (Smith, 1978). A convenient definition of a bar is that it is a bedform of a size comparable in magnitude to that of the channel in which it occurs. Smith (1978) suggests that bars should also be defined as those bedforms which have a non-periodic (i.e. irregular) occurrence within a channel, to distinguish them from large periodic bedforms such as sand waves. Bars may be the product of both depositional and erosional events, and preserved bars in the ancient record commonly consist only of erosional remnants which have had a complex history before final burial.

    Bars may be classified into three broad groups (Table 8). Such a classification is liable to offend the fluvial geomorphologist, who can recognize many variations in flow pattern and sediment response, but to the geologist dealing mainly with limited vertical exposure and little lateral control a simplified classification is the most useful. Some of the main bar types are illustrated in Figure 6.

    Small bars have been termed “mesoforms” by Jackson (1975). Their development depends mainly on individual dynamic events

  5. Page 23

    Because of their relatively steep slope and abundant coarse bedload the proximal reaches of rivers tend to be of low-sinuosity, multiple-channel type. These are usually called braided rivers, although this term has also been used for certain types of high-sinuosity, multiple-channel, bedload rivers. Rust (1978a) attempted to clarify the definitions of channel patterns.

    In alluvial basins the river network, may be either of distributary type, in which case the individual distributary systems are commonly referred to as alluvial fans, or they may be contributary, in which case the term alluvial plain may be used (Rust, 1979). The distinction is a morphological one which may be difficult to make in the ancient record. Alluvial fans tend to occur where rivers in confined mountain valleys emerge onto an alluvial basin or trunk river. The flow expands laterally and may become shallower; it separates into more than one channel, and there is a loss in competency which results in deposition of the coarser bedload. Therefore alluvial fans may be recognized commonly by the presence of coarse facies banked against the basin margin (Figure 7). However, such an association does not prove the prior existence of a fan distributary network because similar deposits can occur in any proximal river setting. The term alluvial fan is therefore a difficult one to use with precision when working with ancient rocks.

  6. Page 32

    The fining-upward point bar model was one of the first facies models ever to be erected. It was developed more or less simultaneously by Allen (1963b) and by Bernard et al. (1962) and Bernard and Major (1963). In later papers Allen (1964, 1965, 1970) documented the model in detail and it became so widely used that other styles of fluvial sedimentation were for a time almost ignored (e.g. see review by Visher, 1972). Largely as a result of critical work by Jackson (1978) it is now realized that the classic Allen model is only one of several applicable to meandering river deposits. Jackson (1978) proposed five models based mainly on observations in modern rivers. Not all of these are equally well documented. Variations between the models relate mainly to differences in sediment calibre.

  7. Page 38

    Not all high sinuosity rivers can be categorized in terms of the meandering models described above. Smith and Smith (1980) and Smith and Putnam (1980) defined a new facies model for rivers consisting of a low energy complex of several interconnected channels of variable sinuosity, typically >1.5. In humid environments wetlands, peatbogs and floodplain ponds are common, but this fluvial style also occurs in arid environments, e.g. Lake Eyre Basin, Australia. Channel gradients are low and overbank floods are frequent. Channel banks are stabilized by vegetation. This type of river pattern seems to develop where the river basin is subsiding relative to downstream base level or where the base level itself is rising. An example of an anastomosing belt of the Saskatchewan River is shown in Figure 16.

    Lateral channel migration and point bar accretion is not characteristic of anastomosed rivers. Instead they accrete vertically, as shown in the facies model diagram (Figure 17). The most distinctive feature of the resulting deposits is the near-vertical facies contacts. Bore holes penetrating such a succession would encounter anomalously thick channel and overbank units and correlation between boreholes would be impossible. Smith and Smith (1980) and Smith and Putnam (1980) described modern and interpreted ancient examples of anastomosed rivers. Recognition of subsurface examples is difficult because of the requirement for very tight well control.

  8. Page 40

    Most of the facies models described above have been developed for rivers of moderate dimensions, with channels typically only a few metres deep. Recognition of the deposits of major rivers comparable to the giants of today (e.g. Mississippi, Amazon, Ganges, Brahmaputra) requires very large outcrop or good well control.

    Mossop (1980) described fining-upward fluvial sequences in the order of 50m thick containing epsilon cross stratified units up to 25m thick (Figure 18). Deep channels might be expected to contain giant bedforms, and these have been recorded in a few cases. Coleman (1969) described sand waves in the Brahmaputra 8 to 15m high which migrated as much as 600m in 24 hours during flood episodes (Figure 19). Internally they comprise extremely large scale sets of trough and planar crossbedding. A possible ancient analogue was described by Conaghan and Jones (1975) . Jones and McCabe (1980) described crossbed sets 40m thick containing a complex pattern of internal erosion surfaces. They interpreted them as the deposits of alternate bars in a low sinuosity delta distributary channel.

  9. Page 40

    The facies models we have been considering up to this point are based largely on the concept of cyclicity. Processes such as channel migration, and flood events, generate certain lithofacies in a predictable sequence causing a stratigraphic repetition throughout a fluvial succession. However, we have dealt with only one type of cyclicity, that which Beerbower (1964) termed autocyclic, arising from energy distribution within the depositional basin. Allocyclic processes are those which originate outside the depositional basin through tectonic or climatic causes. They can also generate cyclic sequences, and there is potential for confusion between the various cycle types. At present we do not Have all-encompassing models to incorporate this complexity and, in fact, to attempt this would require a considerable broadening of the facies model concept. This problem was discussed in detail by Miall (1980), and a summary of part of that paper follows.

  10. Page 46

    Under this heading are considered some additional sedimentary effects of age, tectonics and climate which complicate the accumulation of fluvial deposits but which, if interpreted correctly, may add considerably to the overall geological synthesis.

  11. Page 52

    The architecture of a basin can be described as the geometry and relative arrangement of the major facies assemblages in the basin fill. It reflects basin shape, subsidence rates and the position of major sediment sources. Most of these parameters are governed by tectonics.

    Miall (1981b) showed that in most alluvial basins the drainage network has an orthogonal relationship to local tectonic grain. The main drainage from surrounding source areas descends a paleoslope oriented perpendicular to the main controlling tectonic elements.

    These transverse rivers may drain into a lake at the basin centre or into the sea. Alternatively the centre of the basin may be occupied by a trunk river flowing along the basin axis as a longitudinal river.

    Transverse rivers may include a belt of alluvial fans (bajada) at the basin margin. They form alluvial plains or piedmonts tens to a few hundreds of kilometres in width. Longitudinal drivers may be hundreds to thousands of kilometres long. All the world's major rivers (Amazon, Nile, Mississippi, Ganges, Brahmaputra etc.) are longitudinal.

    Rivers end as estuaries, or as deltas, or as ephemeral channels dying out on lake margins or tidal flats. Deltas may conveniently by classified into three types, reflecting the predominance of either fluvial, wave or tidal energy as the main sediment dispersal mechanism on the delta front (Galloway, 1975). River dominated deltas build out a birdsfoot or lobate pattern unimpeded by marine dispersal processes. In areas of high wave energy the river-borne sediment is carried

  12. Page 56

    Three types of economic deposit are associated with fluvial sediments:

    1. Primary deposits accumulated by hydraulic sorting and concentration of detrital grains in fluvial channels. These are placer deposits, including gold, platinum, uranium, diamonds, titanium (rutile, ilmenite), zirconium (zircon) cesium (monazite) and tin (cassiterite).

    2. Primary deposits formed as an integral part of the depositional system. Coal is the most important of these. Sideritic iron ores were also important at one time but are no longer mined.

    3. Secondary deposits in which the porous sand body acts as a conduit for low-temperature fluid migration and as a host for the ultimate deposit. Oil, gas, heavy oils, copper and uranium are the principle examples.

    The following notes discuss the application of fluvial sedimentology to exploration or interpretation of these deposits.

  13. Page 68

Purchase Chapters

Recommended Reading