Tuesday, November 8, 2016

Describing surface stratigraphic sections

Methods of measuring and recording the data

Vertical stratigraphic sections 

The simplest way to record the details of a surface outcrop is by measuring and describing a vertical stratigraphic section. Ideally, the location of the section should be chosen to include important stratigraphic features, such as formation contacts, but, in practice, the location is commonly determined by accessibility-the presence of bars or beaches allowing us to walk along a river cut or a negotiable gully cutting through a cliff section. Only those geologists who use their profession as an excuse to practice their favorite sport of mountaineering will be able to apply sound geological principles to the choice of section. The rest of us take what we can reach. 
In reconnaissance work, rapid measurement and description techniques are acceptable. For example, a hand-held altimeter (aneroid barometer) may be used in conjunction with dip measurements to reconstruct stratigraphic thicknesses using simple trigonometry. Another method that is commonly described in field handbooks is the pace-and-compass technique, suitable for estimating thicknesses across relatively level ground, given accurate stratigraphic dip. The same distances may be measured from maps or air photographs. Long experience with these methods has shown that they are not very reliable; errors of up to 50% can be expected.
 Fig. 1. A "pogo stick" for measuring surface sections, showing sighting bar and clinometer on upper 50 em of rod.
Fig. 2. Use of pogo stick in section measurement.
By far the simplest and most accurate method for measuring a section is the use of a Jacob's staff or "pogo stick." The stick is constructed of a 1.5-m wooden rod, with a clinometer and sighting bar (Fig. 1). The clinometer is preset at the measured structural dip and can then be used to measure stratigraphic thickness as fast as the geologist can write down descriptive notes (Fig. 2). The best technique is to use two persons. The senior geologist observes the rocks and makes notes, while the junior (who can be an inexperienced student) "pogo's" his or her way up the section recording increments of 1.5 m on a tally counter and collecting samples. The length 1.5 m is convenient for all but the tallest or shortest persons, although it can be awkward to manipulate on steep slopes. The only skill required by the pogo operator is the ability to visualize the dip of the strata in three dimensions across whatever terrain the geologist may wish to traverse. This is important so that the pogo can at all times be positioned perpendicular to bedding with the line of sight extending from the sighting bar parallel to bedding. 
It is far preferable to measure up a stratigraphic section rather than down, even though this often means an arduous climb up steep slopes. Many geologists working by helicopter in rugged terrain have made their traverses physically easier by working downhill wherever possible. But, not only is it difficult to manipulate a pogo stick downward in a section, it makes it more difficult for the geologist to comprehend the order of events he or she is observing in the outcrops. 
The geologist should search for the cleanest face on which to make observations. Normally, this should be weathered and free of vegetation, talus, or rain wash. Most sedimentary features show up best where they have been etched out by wind or water erosion, or where a face is kept continuously clean and polished by running water, as in a river bed or an intertidal outcrop. Such features rarely show up better on fresh fracture surfaces, so a hammer should only be used for taking samples. Carbonates may benefit from etching with dilute acid. The geologist should methodically examine both vertical cuts and the topside and underside of bedding planes; all may have something to reveal. It is also useful, on larger outcrops, to walk back and examine them from a distance, even from a low-flying helicopter or from a boat offshore, as this may reveal large-scale channels, facies changes, and many other features of interest. A different technique may be used to document such large outcrops, which may conveniently be termed lateral profiling to distinguish it from vertical profiling. We discuss this in the next section. 
In the interests of maximum efficiency, the geologist obviously should ensure that all the necessary measurements, observations, and sample collections are made during the first visit to a section. It may be useful to carry along a checklist, for reference each time a lithologic change in the section requires a new bed description. Many geologists have attempted to carry this process one step further by designing the checklist in the form of a computer processible data card, or they record the data in the form of a computer code. There are two problems with this: 
  1.  If attempts are made to record every piece of field information on the computer file, the resulting file is likely to be very large and cumbersome. Storage and retrieval programming may consume far more time than the original field work, unless the geologist can draw on some preexisting program package. This leads to the second problem. 
  2. If data are to be coded in the field or if a preexisting program system is to be used, it means that decisions will already have been made about what data are to be recorded and how they are to be recorded before the geologist goes into the field. If the geologist knows in advance what is likely to be found, as a result of some previous descriptions or reconnaissance work, this may be satisfactory, but in the case of isolated field areas such foreknowledge may not be available. In this case, there is a certain risk involved in having the observation system designed in advance. 
Individual geologists vary in their interests and in the observations they make and, of course, the rocks are highly variable, so that it is not possible to design a single, all-purpose, section-measuring software package. Specialized systems have to be designed for specific projects, and although this leads to expense in programming and debugging, it means that the program can be designed for the specific type of output required. It should not be forgotten, though, that programming, as such, is a technical, not a scientific, skill. Students and other workers who write programs for their research get few points for producing a workable program, only for the interesting scientific ideas their programming allows them to test. For example, an obvious application of computerized stratigraphic section data is in the statistical study of facies associations and cyclic sedimentation.
Fig. 3. A form for field recording of stratigraphic sections that can be processed by a computer. Heavy horizontal lines define the input for each descriptive unit. Items are entered by making a pencil mark in the dashed line tracks.
An example of a successful computerized section-description system is that devised by P.F. Friend and his students for studying Devonian nonmarine clastic sequences in East Greenland (Alexander-Marrack et aI., 1970; Friend et aI., 1976). They designed a form with spaces to be filled in by the geologist in the field (Fig. 3). Marks were made in the appropriate tracks with a soft pencil, corresponding to the observed properties. These cards could be fed directly into a computer using an optical mark page reader. The data form provides a choice of items to be filled in within each small block outlined by a solid line. Areas of the card delimited by heavy horizontal lines across the full width of the page correspond to individual bedding units recognized in the field. Items encoded on the card include thickness of unit, exposed/covered choice, color, grain size, pebbles (present/ absent), nature of basal contact, presence of carbonate, internal structures, and paleocurrent measurements. The range of possible observations is limited by the card. For example, the method would not be suitable if the section was found to contain thick limestone intervals because, although carbonate content can be indicated, there are no spaces to record typical carbonate sedimentary features, such as ooliths, birdseye structures, stromatolites, etc. Also, the method is suitable only for recording vertical stratigraphic changes. The presence of lateral facies changes or the exposure of a large channel could not be recorded in this system. 
The particular system designed by Friend worked very well for a team of about a dozen geologists working over three field seasons, but it is unlikely that it could be applied anywhere else without modification. Unless a geologist is committed to doing much the same type of field geology for a long period, the investment of time and resources into designing and setting up such a system may not be worth the effort. In some cases, the field records must, in any case, be modified following laboratory work. This is particularly the case with carbonate sediments, which commonly are best studied in polished hand specimens or thin sections. In this case, much of the effort of the field geologist must be devoted to ensuring that a sufficiently rigorous sample collection routine is observed. In the long run, it may be simpler to stay with an oldfashioned field notebook for field observations. Computer files can be built at a later date from corrected data, using a file structure designed for a specific purpose. For example, structural data, paleocurrent data, and vertical facies succession data lend themselves readily to processing in special-purpose computer files, which can exploit one of the main advantages of the computer to the full, namely, its ability to carry out complex or repetitive numerical manipulations with great rapidity. 

The construction of lateral profiles 

Some stratigraphic units are essentially tabular at the scale of the outcrop and can be quickly and accurately documented using vertical profiles, in the way outlined in the previous section. However, some types of sedimentary assemblage contain complex facies changes, which may be at a small enough scale to observe in individual outcrops, especially in large outcrops. For example, a reef core, with its reef-front talus slope and back-reef lagoonal deposits, or a large fluvial or submarine-fan channel, with its fill of complex bar deposits, may be spectacularly displayed in a road cut or mountainside. The measurement of a few vertical sections across such an outcrop is a quite inadequate way to document the wealth of facies detail that may be available. 
An alternative field method is to construct lateral profiles, long sections that encompass the full vertical stratigraphic height of the outcrop, and also extend along the strike as far as possible, to illustrate the facies changes. These may be constructed by careful surveying, but a much quicker method is to make use of photographic mosaics of the outcrop. The geologist moves well back from the exposed face, perhaps as much as several hundred meters in the case of a very large outcrop, and carries out a traverse parallel to the face, taking a series of overlapping frames until he or she has covered the entire outcrop. By taking care to remain at the same distance from the face, each frame will be at approximately the same scale. The same end can be achieved by taking photographs from a boat, or even from an aircraft flying low over the outcrop. 
The prints are carefully overlapped and mounted in the office in order to construct a mosaic, which can then be used in the field as a kind of topographic map base on which to enter stratigraphic and sedimentologic detail. Some scale distortions inevitably arise. Outcrops are rarely flat, and projections and gullies will not fit together precisely in the mosaic because of the differing perspectives of adjacent frames. Such distortions are trivial relative to the immense amount of detail that can be shown on such profiles, and if accurate measurements of individual features are required, they should, in any case, be made in the field and not from the photograph. 

Types of field observation 

Subdivision of the section into descriptive units

This is a subjective operation based on the rock types present, the quality and accessibility of the exposure, and the amount of detail required in the description. Very detailed descriptions may require subdivision into units containing (for example) a single mudstone lens or crossbed set, and will therefore be on the order of a few centimeters or tens of centimeters thick. Thicker units can be defined by grouping similar rock types, but sedimentologically useful detail may be lost thereby. For each unit, the kinds of observations listed in the succeeding paragraphs are made where appropriate.
Fig. 4. An example of a lateral profile. Above is the outcrop photo-mosaic, and below is the interpretation made from photo and outcrop examination. This example illustrates a succession of compound bars deposited in a fluvial system. A major internal bounding surface (labelled B) indicates a pause in deposition and the initiation of a new channel. Beds above this surface downlap on to it, and are interpreted as lateral accretion surface.
In the case of lateral profiles (Fig. 4), among the most valuable kinds of observation that can be made is the documentation and classification of the various kinds of bounding surface that separate stratigraphic units. These range from the simple bedding plane surfaces that separate individual crossbed sets through the surfaces that bound channels and bars to the major (usually horizontal) surfaces that delimit mappable stratigraphic units (formations, members, etc.).

Lithology and grain size

Table 1. Standard grain-size scales for carbonate and clastic sediments.
Fig. 5. A grain-size comparison chart, for use in the field or for logging well cuttings or cores. 
Lithologic classification of clastic rocks can usually be done satisfactorily by visual observation in the field, without the necessity of follow-up laboratory work. Classification is based on grain size (Table 1), which is easily measured on the outcrop. For sand grade rocks, it is useful to take into the field a grain size chart (Fig. 5) or a set of sand samples each representing one phi class interval through the sand size ranges. These are used for comparison purposes and permit recognition of the main sand grade subdivisions: very fine, fine, medium, coarse, and very coarse. Many tests by the author and others have shown that such observations provide adequate, accurate information on the modal size range of the sandstones. The description should be modified by appropriate adjectives if sorting characteristics require it; for example, pebbly coarse-grained sandstone, silty mudstone, etc. For the purpose of regional facies analysis, this is usually the only kind of grain size information required. Skewness, kurtosis, and other statistical data are not needed.
Siltstone and mudstone can be distinguished in a hand specimen by the presence or absence of a gritty texture, as felt by the fingers or the tongue. This is, of course, a crude method and should be checked by making thin sections of selected samples. However, field identifications of this type commonly are adequate for the purpose of facies analysis. 
Fine-grained rocks, including those consisting of a mixture of sandstone, siltstone, and mudstone, are difficult to classify and describe. Dean et al. (1985) discussed the methods used in the Deep Sea Drilling Project based on smear slides of soft sediments made on board ship. 
For conglomerates, maximum clast size is often a useful parameter to measure. Typically, this is estimated by taking the average of the 10 largest clasts visible within a specified region of an outcrop, such as a given area of a certain bedding plane. In thick conglomerate units, it may be useful to repeat such measurements over regular vertical intervals of the section. It is also important to note the degree of sorting, clast shape and roundness, matrix content, and fabric of conglomerate beds. For example, does the conglomerate consist predominantly of very well rounded clasts of approximately the same size, or is it composed of angular fragments of varying size and shape (breccia)? Do the clasts "float" in abundant matrix, a rock type termed matrix-supported conglomerate, or do the clasts rest on each other with minor amounts of matrix filling the interstices-clast-supported conglomerate?
Carbonate rocks commonly cannot be described adequately or accurately in outcrops and require description from thin sections or polished sections observed under a low-power microscope. Among the reasons for this are the ready susceptibility of carbonate rocks to finescale diagenetic change, and the fact that weathering behavior in many cases obscures rather than amplifies such changes, as seen in outcrops. Another important reason for not relying on outcrop observation is that some of the types of information required for carbonate facies analysis are simply too small to be seen properly with the naked eye. These include mud content, certain sedimentary textures, and biogenic features. 
Field geologists traditionally take a dropper bottle of 10% hydrochloric acid with them to test for carbonate content and to aid in distinguishing limestone from dolomite (on the basis of "fizziness"). However, for research purposes, the test is quite unsatisfactory, and the geologist is advised to abandon the acid bottle (and stop worrying about leakage corroding packsacks). Dolomite commonly can be distinguished from limestone by its yellowish weathering color in the field, but a better field test is to use alizarin red-S in weak acid solution. This reagent stains calcite bright pink but leaves dolomite unstained. In both hand specimens and thin sections, use of this reagent can reveal patterns of dolomitization on a microscopic scale.
Because of the problem with carbonate rocks discussed previously, the geologist is advised not to rely on field notes for facies analysis of these rocks, but to carry out a rigorous sampling program and supplement (and correct) the field notes using observations made on polished slabs or thin sections. Sampling plans are discussed later in this section. Laboratory techniques for studying carbonates are described by Wilson (1975).
Evaporites are difficult to study in surface outcrops. They are soft and recessive and commonly poorly exposed, except in arid environments. Like carbonates, they are highly susceptible to diagenetic change, so that field observations must be supplemented by careful laboratory analysis. Only recently have sedimentological methods been applied to the study of evaporites (observations of grain size, textures, bedding, structures, etc.) and this subject offers hope for exciting developments.
Mixed carbonate-clastic sediments are common and are typically dealt with as if they were carbonate or clastic, which may not be the most effective way to emphasize subtle lithologic characteristics. Mount (1985) discussed the problems of classifying these rocks and suggested some methodological approaches. 

Porosity

Porosity and permeability are of particular interest if the rocks are being studied for their petroleum potential. Observations in surface outcrops may be of questionable value because of the effects of surface weathering on texture and composition, but the geologist should always break off a fresh piece of the rock and examine the fracture surface because such observations commonly constitute the only ones made. The geologist should distinguish the various types of porosity, such as intergranular (in detrital rocks), intercrystalline (in chemical rocks), and larger pores, such as vugs, birdseye texture, moulds of allochems, such as oolites or pellets, fossil moulds, fracture porosity, etc. More accurate observations may be made from thin sections, and samples may be submitted to a commerical laboratory for flow tests if required. Porosity types should be reported in terms of the estimated percentage they occupy in the bulk volume of the sample. 

Color

Color mayor may not be an important parameter in basin analysis. Individual lithologic units may display a very distinctive color, which aids in recognition and mapping. Sometimes it even permits a formation to be mapped almost entirely using helicopter observations from the air, with a minimum of ground checking. However, the sedimentological meaning and interpretation of color may be difficult to resolve. 
Some colors are easily interpreted-sandstones and conglomerates commonly take on the combined color of their detrital components, pale greys and white for quartzose sediments and darker colors for lithic rocks. As noted, limestones and dolomites may also be distinguished using color variations. However, color is strongly affected by depositional conditions and diagenesis, particularly the oxidation-reduction balance. Reduced sediments may contain organically derived carbon and Fe2+ compounds, such as sulphides, imparting green or drab gray colors. Oxidized sediments may be stained various shades of red, yellow, or brown by the presence of FeH compounds, such as hematite and limonite. However, local reducing environments, such as those created around decaying organisms, may create localized areas or spots of reduction color. Color can change shortly after deposition, as shown, for example, by Walker (1967) and Folk (1976). Moberly and Klein (1976) found that oxidation and bacterial action caused permanent color changes when fresh sediments, such as deep sea cores, are exposed to the air. 
Thus, the problem is to decide how much time to devote to recording color in the field. Ideally, each descriptive unit in the stratigraphic section should be studied for color using a fresh rock-fracture surface and comparisons to some standard color scheme, such as the U.S. National Research Council RockColor Chart (Goddard et aI., 1948). In practice, for the purpose of facies and basin analysis, such precision is not required. Simple verbal descriptions, such as pale gray, dark redbrown, etc., are adequate. More precise descriptions may be useful if detailed studies of diagenetic changes are to be undertaken, but recent work has shown that such studies may give misleading results if carried out exclusively on surface exposures because of the effects of recent weathering (Taylor, 1978). 

Bedding

Table 2. Scale of stratification thickness.
An important type of observation, particularly in clastic rocks, is the thickness of bedding units. Thickness relates to rate of environmental change and to depositional energy. In some cases, bed thickness and maximum grain size are correlated, indicating that both are controlled by the capacity and competency of single depositional events. Bed thickness changes may be an important indicator of cyclic changes in the environment and sedimentologists frequently refer to thinning upward and fining upward or coarsening and thickening upward cycles. It is important to distinguish bedding from weathering characteristics. For example, a unit may split into large blocks or slabs upon weathering, but close examination may reveal faint internal bedding or lamination not emphasized by weathering. Bedding can be measured and recorded numerically, or it can be described in field notes semiquantitatively using the descriptive classification given in Table 2.

Sedimentary structures produced by hydrodynamic molding of the bed

Sedimentary structures include a wide variety of primary and postdepositional features. All individually yield useful information regarding depositional or diagenetic events in the rocks, and all should be meticulously recorded and described in the context of the lithology and grain size of the bed in which they occur. The assemblage of structures and, in some cases their orientation, can yield vital paleogeographic information. 
Table 3. Classification of inorganic sedimentary structures. 
Fig. 6. Various types of graded bedding 
Inorganic sedimentary structures can be divided into three main genetic classes, as shown in Table 3. 
Sediment carried in turbulent suspension by mass gravity transport processes, such as debris flows and turbidity currents, is subjected to internal sorting processes. When the flow slows and ceases, the sorting may be preserved as a distinct texture termed graded bedding. Grading commonly consists of an upward decrease in grain size, as illustrated in Fig. 6; this is termed normal grading. However, certain sedimentary processes result in an upward decrease in grain size, termed inverse grading
Clastic grains can be divided into two classes on the basis of their interactive behavior. Cohesive grains are those that are small enough that they tend to be bound by electrostatic forces and thus resist erosion once deposited on a bed. This includes the clay minerals and fine silt particles. A range of erosional sedimentary structures is present in such rocks (Table 3). Larger clastic grains, including siliciclastic, evaporite, and carbonate fragments, of silt to cobble size are noncohesive. They are moved by flowing water or wind as a traction carpet along the bed or by intermittent suspension. The dynamics of movement causes the grains to be moulded into a variety of bedforms, which are preserved as crossbedding within the rock. 
There are three main classes of bedforms and crossbedding found in ancient rocks: 
  1. Those formed from unidirectional water currents, such as are found in rivers and deltas, and oceanic circulation currents in marine shelves and the deep sea. 
  2. Those formed by oscillatory water-currents, including both wave- and tide-generated features. Although the time scale of current-reversal is, of course, quite different, there are comparable features between the structures generated in these different ways. 
  3. Those formed by air currents. Such currents may be highly variable, and the structure of the resulting deposits will be correspondingly complex. However, examination of ancient wind-formed (eolian) rocks indicates some consistent and surprisingly simple patterns.
Recognition in the outcrop or in the core of the diagnostic features of these crossbedding classes is an invaluable aid to environmental interpretation, and therefore crossbedding structures must be examined and described with great care wherever they are found. 
Fig. 7. Terminology of crossbedding
Fig. 8. Growth offoresets, advance of a fiat-topped bar and development of size sorting by migration of ripples along the bar top. Note increase in foreset grain-size from (a) to (b) as ripple crest reaches avalanche face.
The components of crossbedding are illustrated in Fig. 7. A foreset represents an avalanche face, down which grains roll or slump or are swept down by air or water currents. Fig. 8 shows one of many types of bedform morphology illustrating the way grains advance up the up-current (stoss) side of the bedform and are fed to a continuously advancing or prograding down-current foreset (lee) surface. In this case, the grains are transported by ripples, which contain internal foreset structure. Continuous deposition produces repeated foreset bedding or lamination as the bedform accretes laterally, resulting in a crossbed set (McKee and Weir, 1953). A coset is defined as a sedimentary unit made up of two or more sets of strata or crossbedding separated from other cosets by surfaces of erosion, nondeposition, or abrupt change in character (McKee and Weir, 1953). Note that a coset can contain more than one type of bedding. 
Fig. 9. Criteria used in the description and definition of crossbedding types. 
Fig. 10.  Internal structures of crossbedding.
When describing crossbedding, attention must be paid to seven attributes, as illustrated in Figs. 9 and 10. All but the last of these were first described in detail in an important paper on crossbedding classifiction by Allen (1963). In the field, crossbeds are classified first according to whether they are solitary or grouped. Solitary sets are bounded by other types of bedding or crossbedding, grouped sets are co sets consisting entirely of one crossbed type. Scale is the next important attribute. In water-laid strata, it is found that a bedform amplitude of about 5 cm is of hydrodynamic significance and, accordingly, this amplitude is used to subdivide crossbeds into small- and largescale forms. An assumption is made that little or none of the top of a bedform is lost to erosion prior to burial; generally, the amount lost seems to increase in approximate proportion to the scale of the bedform or the thickness of the crossbed structure. Forms thinner than 5 cm are termed ripples, whereas forms larger than 5 cm are given a variety of names, reflecting in part a diversity of hydrodynamic causes and in part a considerable terminological confusion. 
Most crossbed sets contain foresets that terminate at the base of the set, in which case the foresets are said to be discordant. In rare cases where the crossbeds are parallel to the lower bounding surface, as occurs in some sets with curved lower surfaces, the crossbeds are described as concordant. 
The crossbeds may show either homogeneous or heterogeneous lithology. Homogeneous cross beds are those composed offoresets whose mean grain size varies by less than two phi classes. Heterogeneous crossbeds may contain laminae of widely varying grain size, including interbedded sand and mud or sand and gravel, possibly even including carbonaceous lenses. 
The minor internal structures within crossbeds are highly diagnostic of their origin (Fig. 10). The dip angle of the fore set relative to the bounding surface is of considerable dynamic significance. Are the foresets curved, linear, or irregular in sections parallel to the dip? Is the direction of dip constant, or are there wide variations or reversals of dip within a set or between sets? Do the sets contain smaller scale hydrodynamic sedimentary structures on the foresets, and if so, what is the dip orientation of their foresets relative to that of the larger structure? What is the small-scale internal geometry of the foresets-are they tabular, lens, or wedge shaped? Do they display other kinds of sedimentary structures, such as trace fossils, synsedimentary faults, or slumps? Are reactivation surfaces present? These represent minor erosion surfaces on bedforms that were abandoned by a decrease in flow strength and then reactivated at some later time (Collinson, 1970). 
Fig. 11. The 15 crossbedding types defined by Allen (1963), distinguished using the criteria set out in Fig. 9. 
Fig. 12. Types of crossbedding, as observed in outcrops. A. Climbing ripple cross-lamination. Note increase in angle of climb, a common feature that indicates decrease in flow strength and increase in detritus settling from suspension. Fluvioglacial, Pleistocene, Ottawa. B. Ladderback ripples. An older ripple set is visible as short segments of ripple crest preserved in the troughs of a younger set formed in a perpendicular direction. Fluvial, Karoo Supergroup, Beaufort West, South Africa. C. A coset of planar cross-strata (alpha type of Fig. 11). Note consistency of orientation. Fluvial, Cretaceous, Banks Island, Arctic Canada. D. Planar crossbed set in a clastic limestone. Foresets are emphasized by chert lenses. Shallow marine, Mississippian, northeast Alaska. E. Oblique section through a solitary trough crossbed set with trace fossils. Shallow marine, Cretaceous, Banks Island, Arctic Canada. F. A coset of trough cross strata (pi type of Fig. 11). Fluvial, Devonian, Somerset Island, Arctic Canada. G. Bedding plane view of a trough crossbed. The curved foresets dip toward the left. Fluvial, Huronian, (Proterozoic), near Elliot Lake, Ontario. H. Bedding plane exposure of straight-crested megaripples. Shallow marine, Karoo Supergroup, near Durban, South Africa. I. Herringbone crossbedding: five planar crossbed sets showing reversals of flow direction. Shallow marine, Permian, Ellesmere Island, Arctic Canada. J. Low-angle, curved crossbedding, possibly hummocky cross-stratification, with the trace fossil Ophiomorpha. Shallow marine, Eocene, Banks Island, Arctic Canada. K. Large-scale crossbedding (epsilon type of Fig. 11). Fluvial point bar, Eocene, Ellesmere Island, Arctic Canada. L. Large scale crossbedding (epsilon type of Fig. 11). Fluvial point bar, Carboniferous, Alabama. M. Large-scale eolian crossbedding, Jurassic Navajo Formation, Zion National Park, Utah. N. Talus slope of a reef composed of coarse, crinoidal calcarenite and calcirudite, Devonian, Princess Royal Islands, Arctic Canada.
These seven attributes can be used to classify crossbed sets in the field. It is time-consuming to observe every attribute of every set, but it is usually possible to define a limited range of cross bed types that occur repeatedly within a given stratigraphic unit. These can then be assigned some kind of local unique descriptor, enabling repeated observations to be recorded rapidly in the field notebook. One way to classify cross bed sets is to use a published classification system, of which that erected by Allen (1963) is the most complete. This is illustrated in Fig. 11. However, this classification does not encompass any of the internal structures illustrated in Fig. 10. Allen's classifiction has become widely used, particularly in descriptions of nonmarine rocks; some of his Greek letter names are well known among sedimentologists and are used as a convenient shorthand. However, care should be taken to supplement this classification with observations of the internal structures. Examples of crossbedding in outcrops are illustrated in Fig. 12.
A vital component of basin analysis is an investigation of sedimentological trends, such as determining the shape and orientation of porous rock units. Paleocurrent analysis is one of several techniques for investigating sedimentary trends based, among other things, on studying the size, orientation, and relative arrangement of crossbedding structures. Therefore, when describing outcrop sections, it is essential to record the orientation of crossbed sets.
Fig. 13. Clast imbrication in a modern river, indicating flow from left to right. Ellesmere Island, Arctic Canada. 
Crossbedding represents a macroscopic orientation feature, but each clastic grain is individually affected by a flow system and may take up a specific orientation within a deposit in response to flow dynamics. The longest dimension of elongated particles tends to assume a preferred position parallel or perpendicular to the direction of movement and is commonly inclined upflow, producing an imbricated or shingled fabric. This fabric may be present in sandsized grains and can be measured optically, in thin section (Martini, 1971) or using bulk properties, such as dielectric or acoustic anisotropy (Sippel, 1971). In recent years, paleomagnetic data have also been recognized to contain much useful information relating to primary sedimentary fabrics. Eyles et al. (1987) discussed magnetic orientation and anisotropy data with reference to the depositional processes of till and till-like diamict deposits. Oriented specimens must be collected in the field for such an analysis. In conglomerates, an imbrication fabric commonly is visible in an outcrop and can be readily measured by a visual approximation of average orientation or by laborious individual measurements of clasts. Fig. 13 illustrates imbrication in a modern river bed. It has been found that in nonmarine deposits, in which imbrication is most common, the structure is one of the most accurate of paleocurrent indicators (Rust, 1975).
Fig. 14.  A. Plane bedding in sandstone containing dark, sand-sized comminuted carbonaceous debris, Eocene, Ellesmere Island, Arctic Canada. B. Bedding plane exposure of a sandstone such as that in photo A, showing parting lineation, Upper Proterozoic, Banks Island, Arctic Canada.
Grain sorting is responsible for generating another type of fabric in sand grade material, which is also an excellent paleocurrent indicator. This is primary current lineation, also termed parting lineation because it occurs on bedding plane surfaces of sandstones that are flat bedded and usually readily split along bedding planes. An example is illustrated in Fig. 14. Primary current lineation is the product of a specific style of water turbulence above a bed of cohesionless grains, as are the various bedforms that give rise to crossbedding. It therefore has a specific hydrodynamic meaning and is useful in facies analysis as well as paleocurrent analysis. Rather than the bed itself, objects such as plant fragments, bones, or shells may be oriented on a bedding plane.

Sedimentary structures produced by hydrodynamic erosion of the bed

A wide variety of erosional features is produced by water erosion of newly deposited sediment. These result from changes in water level or water energy in response to floods, storms, tides, or winddriven waves and currents. They can also result from evolutionary change in a system under steady equilibrium conditions. These processes result in the development of various types of bounding surfaces in the rocks. Recognition and plotting of these features in outcrop sections are important components of facies analysis, and with adequate exposure, orientation studies may contribute significantly to the analysis of depositional trends. 
These features range in size up to major river and tidal channels, submarine canyons, and distributary channels several kilometres across and tens or hundreds of meters deep, but large features such as this can rarely be detected in the average small outcrop. They may be visible in large outcrop sections, where they can be documented using lateral profiles and on seismic sections, and it may be possible to reconstruct them by careful lithostratigraphic correlation and facies analysis of scattered outcrops, but this is beyond the scope of our immediate discussion. At the outcrop scale, there are two types of small-scale erosional features to discuss, those that truncate one or more bedding units and those that scour or pit the bedding plane without significantly disrupting it. 
Fig. 15. Macroscopic erosional features. A. Abandoned fluvial channel showing levees and fine-grained fill, Carboniferous, Kentucky. B. Tidal channel filled with calcarenite, cut into finer grained tidal flat deposits, Carboniferous, Kentucky. C. A fluvial cutbank, Tertiary, Bylot Island, Arctic Canada. D. Close up of a stepped scour surface at the base of a fluvial channel, Permian, southern Poland. E. Intertidal carbonate mudflat deposits containing several scour surfaces covered by bioclastic debris (above) and a desiccated layer (below) showing incipient brecciation. Compare the latter with photo G. Silurian, Somerset Island, Arctic Canada. F. Discontinuity surface with manganiferous nodules and crusts, Lower Cretaceous, Provence, France. G. Intraformational breccia produced by breakup and reworking of a desiccated dolomite bed on a tidal flat. Ordovician, Somerset Island, Arctic Canada. H. Rill marks produced by water seepage from an exposed fluvial bar. Karoo Supergroup, Beaufort West, South Africa. I. Lowermost bed is a caliche formed by subaerial weathering of a shallow marine carbonate bed. It is followed by a coarse-grained lag deposit formed by a storm (center of view, below overhanging bed). Carboniferous, Kentucky. J. A caliche breccia, Cretaceous, Provence, France. K. A slide surface overlain by a large slumped mass of carbonate sediments. Cretaceous, Provence, France.
The first type includes channels, scours, lowrelief erosion surfaces, and rill markings, in decreasing order of scale. Typical examples are illustrated in Fig. 15. These may be classified as macroscopic erosion features (Table 3). Channels and scours are usually filled by sediment that is distinctly different in grain size and bedding characteristics from that into which the channel is cut. Almost invariably the channel fill is coarser than the eroded strata indicating, as might be expected, that the generation of the channel was caused by a local increase in energy level. Figure 15A illustrates an exception to this, where a channel was abandoned and subsequently filled by fine sediment and coal under low-energy conditions. 
It is a common error to confuse trough crossbedding with channels. Troughs are formed by the migration of trains of dunes or sinuous-crested megaripples. They rest on curved scour surfaces, but these are not channels. The scours are formed by vortex erosion in front of the advancing dunes and are filled with sediment almost immediately. Channels, on the other hand, may not be filled with sediment for periods ranging from hours to thousands of years after the erosion surface is cut, and so the cutting and filling of the channel are quite separate events. 
Erosion surfaces may exhibit little erosional relief, which may belie their importance. In the nonmarine environment, sheet erosion, wind deflation, and pedimentation can generate virtually planar erosion sufaces. In subaqueous environments, oceanic currents in sediment-starved areas, particularly in abyssal depths, can have the same result. Exposed carbonate terrains may develop karst surfaces, with the formation of extensive cave systems. At the outcrop scale, careful examination of erosion surfaces may reveal a small-scale relief and the presence of features such as infilled desiccation cracks, basal intraformational or extrabasinal lag gravels (Fig. 15E), fissures filled with sediment from the overlying bed, zones of bioclastic debris, etc. In some subaerial environments, soil or weathering profiles may have developed, including the development of caliche or calcrete (Fig. 151 and J) and the presence of surfaces of nondeposition. In carbonate environments surfaces of nondeposition commonly develop subaqueously. Hardgrounds are organically bored surfaces that may be encrusted with fossils in growth positions. Alternatively, they may be discolored by oxidation, giving a red stain, or blackened by decayed algal matter. Long-continued winnowing of a surface of nondeposition may leave lag concentrates of larger particles, blackened by algal decay, and possibly including abundant phosphatized fossil material (Fig. 15F) (Wilson, 1975, p. 80-81). In continental-slope deposits, giant slumps and slides are common and are particularly well exposed as intraformational truncation surfaces in deep water carbonate sediments. An example is illustrated in Fig. 15K. 
It may be difficult to assess the length of time missing at erosion surfaces-some may even represent major time breaks detectable by biostratigraphic zonation. In any case, a careful search for and description of such features in the field is an important part of section description. 
Mesoscopic erosional features fall mainly into a class of structure termed sole markings. These are features seen on the underside of bedding planes, usually in sandstones, and they represent the natural casts of erosional features cut into the bed below, which is typically siltstone or mudstone. They attest to the erosive power of the depositional event that formed the sandstone bed, but, beyond this, most have little facies or environmental significance. However, they can be invaluable paleocurrent indicators.
Fig. 16. Sole markings A. Flute markings. Note vortex flow lines on large flute near center of view. Cretaceous, Provence,France. B. Tool markings, mainly groove casts and load casts. Cretaceous, Provence, France. 
Flute markings are formed by vortex erosion, typically at the base of turbidity currents. Erosion is deepest at the up-current end of the scour and decreases down current, so that in a flute cast the high-relief nose of the cast points up current. In rare examples, vortex flow lines may be perceived in the walls of the flute (Fig. 16A). Flutes generally are in the order of a few centimeters deep. 
Tool markings are a class of sole structure formed by erosional impact of large objects entrained in the flow, including pebbles, plant fragments, bone, or shell material. The many varieties that have been observed have been assigned names that indicate the interpreted mode of origin. They include groove, drag, bounce, prod, skip, brush, and roll markings. A few examples are illustrated in Fig. 16B, which show the strongly linear pattern on the bed, providing excellent paleocurrent indicators. 

Liquijication, load, and fluid loss structures

Clay deposits saturated with water are characterized by a property termed thixotropy: when subjected to a sudden vibration, such as that generated by an earthquake, they tend to liquify and loose all internal strength. This behavior is responsible for generating a variety of structures in clastic rocks. Clay beds commonly are interbedded with sand or silt and, when liquified, the coarser beds have a higher density than the clay and tend to founder under gravity. This mayor may not result in the disruption of the sand units. Where complete disruption does not take place, the sand forms bulbous shapes projecting into the underlying clay, termed load structures. These are usually best seen in ancient rocks by examining the underside of a sandstone bed (Fig. 15B). They are therefore a class of sole structure, though one produced without current movement. Clay wisps squeezed up between the load masses form pointed shapes termed flame structures because of a resemblance to the shape of flames. These are best seen in cross section. Occasionally, loading may take place under a moving current, such as a turbidity flow, and load structures may then be stretched, possibly by shear effects, into linear shapes paralleling the direction of movement.
Fig. 17. Sedimentary structures produced by liquefaction. load or fluid loss. A. Dish structures in a grain flow deposit. Cretaceous, Provence, France. B. Pipe structure produced by penecontemporaneous vertical water escape in a fluvial sandstone, Devonian, northeast Scotland. C. A set of small water escape pipes in a lacustrine sandstone. Devonian, northeast Scotland. D. Small sand volcano created by water escape at the depositional surface, Eocene, Banks Island, Arctic Canada. E. A clastic dyke filled with sandstone and conglomerate. Huronian (Proterozoic), Espanola, Ontario. F. Desiccation cracks in a lacustrine deposit, Devonian, northeast Scotland. G. Synaeresis markings. Silurian, Somerset Island, Arctic Canada. H. Ball and pillow structures in a turbidite unit, Cretaceous, Provence, France. (. Penecontemporaneous slump produced by failure on a delta front slope, Eocene, Banks Island, Arctic Canada. J. Load and drape structures caused by presence of massive crinoidallimestone in thin bedded limestones and mudstones. Devonian, Princess Royal Islands, Arctic Canada. K. Overturned crossbedding caused by fluid shear of a saturated sandstone. Proterozoic, Banks Island, Arctic Canada. L. Ice crystal casts in a fluvial floodplain pond. Karoo Supergroup, Beaufort West, South Africa.
Commonly, load masses become completely disrupted. The sand bed may break up into a series of ovate or spheroidal masses that sink into the underlying bed and become surrounded by mud. Lamination in the sand is usually preserved in the form of concave-up folds truncated at the sides or tops of each sand mass, attesting to the fragmentation and sinking of the original layer. Various names have been given to these features, including ball, pillow, or pseudonodule structures (Fig. 17H). These structures rarely have a preferred shape or orientation and are not to be confused with slump structures, which are produced primarily by lateral rather than vertical movement. 
In many environments, sediment is deposited on a sloping rather than a flat surface (largescale examples of this are visible on seismic reflection surveys and are called clinoformssee), for example, the subaqueous front of a delta, which is something like a very large-scale foreset built into a standing body of water. The difference is that once deposited such material usually does not move again as individual grains because the angle of the slope is too low. Typically, large-scale submarine fans and deltas exhibit slopes of less than 2°. However, the sediment in such environments is water saturated and has little cohesive strength. Slopes may therefore become oversteepened, and masses of material may be induced to slump and slide downslope by shock-induced failure. Undoubtedly, the thixotropic effects described previously facilitate this process. The result is the production of internal shear or glide surfaces and deformed masses of sediment, termed slump structures. Failure surfaces may be preserved as syndepositional faults. Small-scale examples of these features are commonly seen in outcrops, as illustrated in Figs. 171 and J. Some examples of convolute bedding may be produced this way, although others are the result of water escape, as discussed below.
Structures produced by failure and lateral movement commonly retain an internal orientation with a simple geometric relationship to the orientation of the depositional slope. This could include the elongation of slump masses and the orientation (strike) of slide surfaces, both parallel to depositional strike, or the asymmetry, even overturning, of folds in convolute beds. Recognition of these geometric properties in an outcrop is important because it helps distinguish the structures from those of different origin, and orientation characteristics obviously have potential as paleo slope indicators.
Deformed or overturned crossbedding (Fig. 17K) is developed in saturated sand beds by the shearing action of water or turbid flow across the top of the bedform. The upper few centimeters of the crossbedded unit move down current by a process of intragranular shear, and fore set lamination is overturned as a result, producing an up-current dip. Obviously, to produce this structure, the shearing current must have a similar orientation to that of the current that generated the crossbedding. Deformed crossbedding is common in fluvial and deltaic environments.
As additional sediment is laid on top of saturated deposits, grains within the substrate begin to settle and pack more tightly. Pore waters are expelled in this process and move upward or laterally to regions of lower hydrostatic pressure. Eventually, they may escape to the surface. This process may take place slowly if sediment is being deposited grain by grain and the fluid movement leaves little or no impression on the sediments. However, if loading is rapid, a much more energetic process of fluid loss takes place, and the sediment itself will be moved around in the process. The result, in sand-grade deposits, is a group of features called dish and pillar structures (Fig. 17A). Dishes are produced by escaping water breaking upward through a lamination and turning up the edges; pillars record the vertical path of water flow moving to the surface. Dishes may be up to 50 em in diameter. These structures are particularly common in the deposits of sediment gravity flows, such as fluidized flows, in which sediment emplacement is rapid. They are produced by water escape as a flow ceases movementthe loss of the lubricating effect of water itself being one of the main reasons why the flow stops. However, dish and pillar structures have also been observed in fluvial and other deposits and are therefore not environmentally diagnostic. Obviously, they can only be seen in deposits containing lamination and will not be present if the sand is uniform in texture. 
Fluid movement within a bed is an additional cause of convolute lamination. In this case, the laminations are folded by internal shear and may occasionally be broken through by pipes (Fig. 17B,C). At the sediment-water interface, escaping water may bubble out as a small spring, building up a miniature sand volcano (Fig. 17D). 
The emplacement of relatively more dense material over a lower density layer was discussed previously as the cause of load casts. In addition to the downward movement of the denser material, this situation can be the cause of upward movement of lighter sediment, which is injected, together with contained pore fluids, into the overlying rocks. This can occur on a large scale, producing diapirs of evaporite or mud, both of which flow readily under the overburden weight of a few hundred meters of sediment. These diapirs may be several kilometers across and may extend up for several kilometers through overlying deposits. Evaporite diapirs commonly develop on continental margins. Mud diapirs are a characteristic feature of deltas, where coarser deltaic sediment is dumped rapidly on marine mud deposits.
On a smaller scale, the same injection process can generate clastic dykes, consisting of sheets of sandstone or conglomerate (siliciclastic or carbonate) cutting through overlying or underlying beds (Fig. 17E). The host rocks usually are sharply truncated and not internally deformed, indicating that they were at least partially lithified prior to intrusion. Some dykes intrude along fault planes. Some are intensely folded, suggesting deformation by compaction and further dewatering after injection. 
Desiccation cracks are readily recognized by even the untrained eye (Fig. 17F). They are one of the best and most common indicators of subaerial exposure in the rock record. They may penetrate as deep as a few meters into the underlying rocks (although a few centimeters is more typical) and are normally filled by sediment from the overlying bed. Teepee structures are a variety of large desiccation cracks caused by limestone or evaporite expansion on tidal flats. Desiccated beds on tidal flats may peel or curl as they dry, and disrupted fragments commonly are redeposited nearby as an intraclast breccia (Fig. 15G). A subtly different kind of shrinkage feature, termed a synaeresis structure, has been recognized in recent years (Fig. 17G; Donovan and Foster, 1972). These are formed by volume changes in response to salinity variations in the ambient waters and superficially resemble desiccation cracks. They may be distinguished by two principal differences: unlike desiccation cracks, they do not normally form continuous polygonal networks across bedding planes, but may appear as a loose assemblage of small worm-like relief markings on a bedding surface; second, they do not show deep penetration into the substrate, but appear to rest on the bedding plane in which they are found. Synaeresis structures are formed subaqueously and are common in such environments as lakes, lagoons, and tidal pools, where salinity changes may be frequent. 
Evaporation and freezing may cause the development of crystals of evaporite salts and water, respectively, on the depositional surface, particularly on alluvial floodplains, supratidal flats, and lake margins. Evaporite crystals and nodules may be preserved in the rock record and, of course, major evaporite deposits are common; but individual crystals commonly are replaced by pseudomorphs, or are dissolved and the cavity filled with silt or sand, forming a crystal cast. Such structures are a useful indicator of subaerial exposure and desiccation, but do not necessarily imply long-term aridity. Gypsum and halite are the two commonest minerals to leave such traces. Gypsum forms bladeshaped casts and halite characteristic cubic or "hopper-shaped" structures. Ice casts may be formed in soft sediments during periods of freezing, but have a low preservation potential (Fig.17L). 
Particularly distinctive evaporite structures on supratidal (sabkha) flats are termed ptygmatic, enterolithic, and chicken-wire structures. These are caused by in-place crystal growth, expansion, and consequent compression of evaporite nodules, possibly aided by slight overburden pressures. The enterolithic structure is so named for a resemblance to intestines; the chicken-wire structure is caused by squeezing of carbonate films between the nodules; ptygmatic folds may be caused largely by overburden pressures (Maiklem et al., 1969). 
Last, gas bnbble escape structures should be mentioned. These are produced by carbon dioxide, hydrogen sulphide, or methane escaping from buried, decaying organic matter. Gas passes up through wet, unconsolidated sediment and forms bubbles at the sediment-water interface, leaving small pits on the bedding surface. These structures have been confused with rain imprints, but form subaqueously, as may be apparent from associated features preserved in the rocks. True rain imprints probably are very rare. 

Fossils

Body fossils are obviously among the most powerful environmental indicators to be found in sedimentary rocks and should be observed and identified with care. Paleontology and paleoecology are specialized subjects, a detailed discussion of which is beyond the scope of this book. However, those engaged in describing outcrop sections for basin-analysis purposes should be able to make use of such information as they can gather. A complete and thorough paleontological-paleoecological examination of an outcrop section may take several hours, days, or even weeks of work, involving the systematic examination of loose talus and breaking open fresh material or sieving unconsolidated sediment in the search for a complete suite of fossil types. Extensive suites of palynomorphs or microfossils may be extracted by laboratory processing of field samples. Many apparently unfossiliferous or sparsely fossiliferous stratigraphic intervals have been found to contain a rich and varied fauna or flora by work of this kind, but it is the kind of research for which few basin analysts have the time or inclination. 
Weare concerned in this book with reconstructing depositional environments and paleogeography. Fossils can be preserved in three different ways that yield useful environmental information. 
Fig. 18. In-place fossils (life assemblages). A. Biostrome containing Disphyllid and Thamnopora corals, Devonian, northern Spain. B. Coral biostrome, Devonian, near Norman Wells, N. W.T., Canada. C. Typical reef facies. Boundstone composed of calcisponges and encrusting stromatolites, Capitan reef, Permian, Guadalupe Mountains, Texas. D. Tree in growth position, Pennsylvanian, Kentucky. E. Domal stromatolites, bedding plane exposure, Proterozoic, Dismal Lakes, N .W.T., Canada. F. Domal stromatolites capped by finger stromatolites, Silurian, Somerset Island, N.W.T., Canada, (photos A and C courtesy of B. Pratt). 
Fig 19. Preservation of delicate structures, including soft parts. A-D are from the Middle Cambrian Burgess Shale, Yoho National Park, British Columbia; E-F are from the Upper Jurassic Solnhofen Limestone, Bayern, Germany. Bar scales are all 5 mm in length. A. Vauxia gracilenta, Walcott, a sponge. B. Burgessochaeta setigera (Walcott), a polychaete worm. C. MarreLla splendens Walcott, an arthropod. D. Canadaspis perfecta (Walcott), a crustacean; E. Antrimpos speciosus Munster, a shrimp. F. Aeschnogomphus intermedius, a dragonfly. All photos courtesy of D. Rudkin and the Royal Ontario Museum. 
Fig. 20. Fossil assemblages. All but H are death assemblages. A. Modern shell beach, Sanibel Island, Florida. B. Pelecypods in shallow-marine sandstone, Cretaceous, Banks Island, N.W.T., Canada. C. Trilobites and brachiopods on a bedding plane, Silurian, southern Ontario. D. Oyster beach deposit, Pleistocene, Mallorca. E. Crinoidal reef debris·, Devonian, Princess Royal island, N.W.T., Canada. F. Brachiopods, corals, ostracodes and other bioclastic debris in vertical section, Devonian lagoonal limestone, southern Ontario. G. Starfish in shallow marine sandstones of the Molasse Marine, Miocene, near Digne, France. H. Plant fragments are typically preserved as fragments in a death assemblage, but this is an exception. The photograph shows in-place rootlets of Stigmaria sp., Pennsylvanian, Kentucky. (photo C courtesy of R. Ludvigsen; photos D, F courtesy of B. Pratt). 
  1. In-place life assemblages include invertebrate forms attached to the sea bottom, such as corals, archaeocyathids, rudists, some brachiopods and pelecypods in growth positions, some bryozoa, stromatoporoids, stromatolites, and trees. In-place preservation usually is easy to recognize by the upright position ofthe fossil and presence of roots, if originally part of the organism. This type of preservation is the easiest to interpret because it permits the drawing of close analogies with similar modern forms, in the knowledge that the fauna or flora almost certainly is an accurate indicator of the environment in which the rocks now enclosing it were formed. Examples are illustrated in Fig. 18. 
  2. Environmental indicators almost as good as in-place life assemblages are examples of soft-bodied or delicately articulated body fossils preserved intact (Fig. 19). These indicate very little transport or agitation after death and preservation in quiet waters, such as shallow lakes, lagoons, abandoned river channels, or deep oceans. The Cambrian Burgess Shale in the Rocky Mountains near Field, British Columbia, contains one of the most famous examples of such a fossil assemblage, including impressions of soft, nonskeletal parts of many organisms, such as sponges, that are not found anywhere else. The Jurassic Solnhofen Limestone of Germany is another good example. The bones of vertebrate animals tend to disarticulate after death because of the decay of muscle and cartilage and the destructive effects of predators. Nevertheless, entire skeletons of bony fish, reptiles, and mammals are commonly found in certain rock units, indicating rapid burial under quiet conditions. The Solnhofen Limestone is a wellknown example, containing, among other fossils, the entire skeleton and feather impressions of a primitive bird. Such fossil assemblages must, nevertheless, be interpreted with care because a limited amount of transportation is possible from the life environment to the site of eventual burial. Presumably the bird of the Solnhofen Limestone lived in the air, not at the sediment-water interface where it was deposited! 
  3. Much more common than either of the foregoing are death assemblages of fossils that may have been transported a significant distance, perhaps many miles from their life environment. These commonly occur as lag concentrates of shelly debris such as gastropod, pelecypod, brachiopod, or trilobite fragments and fish bones or scales (Fig. 20). Such concentrations may be abundant enough to be locally rock forming, for example, the famous Silurian Ludlow Bone Bed of the Welsh borderlands. They normally indicate a channel-floor lag concentration or the product of wave winnowing and can usually be readily recognized by the fact that fragment grain size tends to be relatively uniform. 
Transported body fossils may not occur as concentrations but as scattered, individual occurrences, in which case each find must be interpreted with care. Did it live where it is now found or was it transported a significnt distance following death? The environmental deduction resulting from such an analysis may be quite different depending on which interpretation is chosen. Evidence of transportation may be obvious and should be sought; for example, broken or abraded fossils may have traveled significant distances. Overturned corals, rolled stromatolites (including oncolites), and uprooted tree trunks, are all obviously transported.
These problems are particularly acute in the case of microfossils and palynomorphs that, on account of their size, are particularly susceptible to being transported long distances from their life environment. For example, modern foraminiferal tests are blown tens of kilometers across the supratidal desert flats of India and Arabia, and shallow water marine forms are commonly carried into the deep sea by sediment gravity flows, such as turbidity currents. Environmental interpretations based on such fossil types may therefore be very difficult, though it may still be possible if the analysis is carried out in conjunction with the examination of other sedimentary features. In fact, it was the occurrence of sandstones containing shallowwater foraminifera interbedded with mudstones containing deep-water forms in the Cenozoic of the Ventura Basin, California, that was one of the principal clues leading to the development of the turbidity current theory for the origin of deep water sandstones. 

Biogenic sedimentary structures

Fig. 21. Trace fossils as exposed on bedding plane surfaces. A. Interconnected burrows in Ordovician limestone, Ottawa. B. Short burrows in Ordovician sandstone, Boothia Peninsula, Arctic Canada. C. Cross sections of vertical, tubular Skolithus burrows, Karoo Supergroup, near Durban, South Africa (see Fig. 22A). D. Casts of bird footprints, Cretaceous flysch, Provence, France. E. Markings of a resting or swimming crustacean, Karoo Supergroup, Beaufort West, South Africa. 
Fig. 22. Trace fossils as exposed in cross-section. A. Skolithus, Karoo Supergroup, near Durban, South Africa (see Fig. 21C). B. Dipiocraterion, Cretaceous, Provence, France. C. Sloping lungfish burrows, Devonian, northeast Scotland. D. Mottling produced by bioturbation in a dolomite, plus individual small burrows. Ordovician, Boothia Peninsula, Arctic Canada. 
Footprints, burrows, resting, crawling, or grazing trails, and escape burrows are abundant in some rock units, particularly shallow marine deposits; all may yield useful environmental information, including water depth, rate of sedimentation, and degree of agitation. Even the nondescript structure "bioturbation," which is ubiquitous in many shallow marine rocks, can be interpreted usefully by the sedimentologist. Footprints are, of course, best seen on bedding plane surfaces, as are many types of feeding trails and crawling traces (Fig. 21). Burrows are better examined in vertical cross section and are most visible either in very clean, fresh, wetted rock surfaces or in wind- or wateretched weathered outcrops (Fig. 22). Stromatolites are a distinctive component of many carbonate successions, particularly those of Precambrian age and have been the subject of much detailed study.
Table 4. Classification of trace fossils.
Information to record on trace fossils includes morphology, size, attitude, abundance, internal structures, and lithofacies associations. Farrow (1975) provided a useful review of field and laboratory techniques for studying trace fossils. Table 4 is a classification of some common invertebrate trace fossils.

Credits: Andrew D. Miall (Principles of Sedimentary Basin Analysis)