U.S. Deparment of the Interior - U.S. Geological Survey

U.S. GEOLOGICAL SURVEY OPEN-FILE REPORT 00-358

CHAPTER 1: GRAIN-SIZE ANALYSIS OF MARINE SEDIMENTS: METHODOLOGY AND DATA PROCESSING

Poppe, L.J.¹, Eliason, A.H.², Fredericks, J.J.³, Rendigs, R.R.¹, Blackwood D.¹ and Polloni, C.F.¹

¹Coastal and Marine Geology Program, USGS, Woods Hole, MA 02543
² Eliason Data Services, 230 Meetinghouse Road, Mashpee, MA 02649
³Woods Hole Oceanographic Institution, Woods Hole, MA 02543

INTRODUCTION

Grain size is the most fundamental physical property of sediment. Geologists and sedimentologists use information on sediment grain size to study trends in surface processes related to the dynamic conditions of transportation and deposition; engineers use grain size to study sample permeability and stability under load; geochemists use grain size to study kinetic reactions and the affinities of fine-grained particles and contaminants; and hydrologists use it when studying the movement of subsurface fluids (Blatt and others, 1972; McCave and Syvitski, 1991). Therefore, with these reasons in mind, the objectives of a grain-size analysis are to accurately measure individual particle sizes or hydraulic equivalents, to determine their frequency distribution, and to calculate a statistical description that adequately characterizes the sample.

The techniques and equipment used for particle-size analysis must be fast, accurate, and yield highly reproducible results. The accuracy of these measurements is limited by sampling techniques, storage conditions, analytical methods, equipment, and, especially, the capability of the operator. Care and attention to detail must be exercised to achieve the best possible results. As with most types of sedimentological analyses there is no ultimate technique or procedure that will produce the most desirable grain size data for all cases. Several types of analyses have been developed over the years to accommodate the different types and sizes of samples and the reasons for conducting the analysis.

For many years, the size and distribution of the sand and gravel fractions were determined solely by sieve analyses, and silt and clay fractions were determined by pipette or hydrometer methods. Later, rapid sediment analyzers (RSA; Fig. 1; Ziegler and others, 1960; and Schlee, 1966) and electro-resistance multichannel particle-size analyzers (EMPSA; Fig. 2), such as the Coulter Counter, automated the analysis and removed much of the tedium from grain-size analyses. About this same time Formula Translation (FORTRAN) programs for the calculation of statistical parameters on geologic data also were developed (Kane and Hubert, 1962; Schlee and Webster, 1967). Other particle-size statistical programs have since been written in Algorithmic Language (ALGOL; Jones and Simpkin, 1970), Beginners All-Purpose Symbolic Instruction Code (BASIC; Sawyer, 1977), and even for use with hand-held calculators (Benson, 1981). Early attempts to integrate computers with particle-size analysis equipment began with the settling tube (Ziegler and others, 1964; Rigler and others, 1981), and hardware and software packages are now available for EMPSA units (Muerdter and others, 1981; Poppe and others, 1985; Coulter Electronics, 1989).

The commercial availability of inexpensive personal computers allows sedimentologists to construct complete computerized particle-size analysis systems (Poppe and others, 1985). The major advantage of using these systems is the time-, labor-, and cost-saving functions they afford. Most of the laboratory equipment and procedures, and all of the data processing described herein may be adapted to IBM-compatible personal computers.

The purpose of this chapter is to describe some of the laboratory methods, equipment, computer hardware, and data-acquisition and data-processing software employed in the sedimentation laboratory at the Woods Hole Field Center of the Coastal and Marine Geology Program of the U.S. Geological Survey. The recommendations and laboratory procedures given below are detailed, but are by no means complete. Serious users are strongly encouraged to consult the original references and product manuals.

FIELD METHODS

Sampling Recommendations

The results of grain-size analyses are sensitive to the manner in which the original samples are collected, handled and stored. For example, it is clear that the introduction of foreign particulate matter into the sample through improper care or cleaning of equipment, or through improper processing, can alter the texture. The growth of authigenic minerals because of improper storage can have a similar effect and, although physical disturbance may not generally have a direct effect on subsequent analyses, the trends of those data may be altered if sediment layering has been disrupted. In sum, analyses that follow improper field work may be meaningless.

To keep sampling artifacts and variation in data to a minimum, certain precautions should be taken in the field during sampling and subsequent storage:

* Records on sampling, including field measurements, should be painstakingly maintained. The appropriate field measurements and any information peculiar to the sample should be supplied to the laboratory along with the sample.

* Particularly in the case where the top of the sediment section is of interest, care must be exercised during sampling. Over penetration by sampling device and jarring or bumping during retrieval should be avoided. Surficial samples for grain size analyses are typically defined as having been collected from 0-2 cm below the sediment/water interface.

* If the samples are in cores with liners that need to be split, the device chosen for cutting should not add plastic shards to the samples since contamination from these shards affects the textural analysis. Core halves should be stored in well-sealed D-tubes with a piece of moist Oasis or sponge to retard evaporation. If the cores are split, it is recommended that one half should be held in reserve as an archive until all analyses are complete and data quality can be assured.

* The subsamples should be stored in containers that are appropriate to the individual analyses. Inert air-tight containers such as plastic jars, boxes, or bags are preferable to metal cans that could rust and contaminate the sample, or cloth bags that allow evaporation and the loss of fines.

* All sampling utensils and containers must be clean and free of contamination. If splits of the samples are to be used for geochemical analyses (e.g. trace metals), the sampling gear should be plastic or Teflon coated and the containers acid washed.

* Care must be taken to obtain a representative split when sampling in the field. Be aware of lateral and vertical variability in grab samples. Collect larger samples from poorly sorted sediment; smaller samples from well sorted sediment.

* Careful labeling is critical. Labels must be legible and permanent. Test the permanency of markers prior to fieldwork.

* To prevent geochemical reactions, the growth of organics, or evaporation from occurring within a sample, refrigeration or freezing is necessary. Such reactions may change the mineralogy and grain-size distributions, complicate laboratory analyses, and increase costs. For example, authigenic minerals such as gypsum may form with individual gypsum crystals growing up to sand and even gravel-sized particles. Similarly, excessive evaporation must also be avoided, especially if the samples are marine and it is necessary to correct for salt content.

* All analyses should be performed within a reasonable (<2 months) amount of time. The longer a sample is stored, the greater the opportunity for storage-related alteration.

Gallery of Common Sediment Sampling Devices

Grab Samplers

Cambell Grab Emery and Schlee, 1963

Clamshell Grab Lay, 1969

Orange Peel Grab Lay, 1969

Petersen Grab Pettersson, 1928

SEABOSS Grab Blackwood and Parolski, 2000

Smith-McIntyre Grab Smith and McIntyre, 1954

Shipek Grab North Carolina State University, 1999

Van Veen Grab Lay, 1969

Corers

Alpine Vibracorer Sea Surveyor Inc, 2000

Electric Vibracorer PVL Technologies Inc, 2000

Gravity Corer Lee and Clausner, 1979

Hydraulically-Dampened Corer Bothner and others, 1997

Lamont Box Corer Lee and Clausner, 1979

Phleger Corer ` Phleger, 1951

Piston Corer Lee and Clausner, 1979

Spade Box Corer Rosfelder and Marshall, 1967

Figure 3 outlines the laboratory techniques and software sequence of operation in a generalized flow diagram for those procedures used in the sedimentation laboratory at the Woods Hole Field Center. Because the size data become part of a data base and as it is necessary to archive the raw unprocessed data, there must be a formal system for recording sample identifiers. Appropriate identifiers include: requestor, cruise identification (id), project id, work requested, purpose of investigation, sample id, latitude, longitude, water depth, top-depth, bottom-depth, sample device, sampling area, and what analyses have been completed on each sample. To this end, the use of standardized “request for analysis” and “sample identification” forms (Figs. 4 and 5) are recommended as organizational aids. The completed forms should be included when samples are submitted for analysis. At this time a unique lab number should be assigned to each sample. At the Woods Hole Field Center, this number consists of two letters followed by three numbers (e.g. AA795). These identifiers are manually entered into a computer using the program GSANV (Appendices A and C).

If the whole sample is not to be used, the sample is typically split using one of three methods: a micro-splitter, cone and quartering, or random bulk (Krumbein and Pettijohn, 1938). Size of sample, its homogeneity, and the degree of accuracy that is required govern which method is selected for a given size analysis. Generally, the larger the analyzed sample the more accurate the grain size analysis. However, unless a sample contains gravel, a 50-g split is usually sufficient. Larger samples are more time consuming to analyze and do not provide significantly more reliable results. If gravel is present, the sample must contain enough material to give adequate representation of the largest sizes present. For example, approximately 250 g would be needed to texturally analyze a sample containing -2 phi material.

Place the split wet sample in a pre-weighed 100-ml beaker, weigh, and record the weight on a “grain-size analysis” form (Fig. 6). The beakers should be scribed with a unique number and, therefore, should require no additional labels, which would alter the weight of the beaker. Gross wet-sample weight minus the weight of the beaker gives the net wet-sample weight. Dry these samples in a convection oven at, or slightly below, 100 C (Fig. 7). This temperature will only drive off unbound water and should not affect the grain size. When the samples are dry (overnight is usually sufficient), place them in a desiccator (Fig. 7) to cool. Weigh the samples as they are removed from the desiccator or the sample will absorb water from the air and their weights will be less accurate. Gross dry-sample weight minus the weight of the beaker gives the net dry sample weight; net wet sample weight minus the net dry sample weight gives the water weight. The weight of salt can be calculated from the salinity and the weight water. The net dry-sample weight minus the weight salt gives the corrected sample weight. Inasmuch as the weight of fines are typically determined by subtracting the coarse weight from the net sample weight, the salt in marine samples must be determined to prevent an overestimation in the amount of fines present.

Whole or fragmented calcite secreting micro- and macro-organisms can bias the grain size distribution if they occur in significantly high concentrations. Because biogenic carbonates commonly form in situ, they usually are not considered to be hydraulically representative of the depositional environment from a textural standpoint (Reineck and Singh, 1980). Their presence alters the textural data and complicates interpretation. If pervasive throughout the sand fraction, biogenic calcite or aragonite may be selectively dissolved from the bulk sample using cold, dilute (10%) hydrochloric acid (HC1) or a 5-pH Na0Ac buffer (Jackson, 1956; see section on insoluble residues below). After the carbonate has been dissolved, the HC1 or Na0Ac can be removed with multiple decantations or centrifugations using distilled water to remove the salts and acid and (or) buffer. If limited to the gravel fraction, it is often easier to manually remove the fragments of bivalve shells and other biogenic carbonate debris (see below) rather than treating the whole sample. Care must be taken to use dissolution techniques only when detrital carbonate is absent.

The removal of organic matter may be necessary to achieve complete dispersion of the clay and, in sediments with an elevated organic content (>3%), to prevent the organics from being counted as part of the sample, which would bias the grain-size distribution. The sample is placed in a 600-ml beaker and a small volume (~10 ml) of 30% hydrogen peroxide is added. The sample is stirred and, if necessary, water is added to slow the reaction down and prevent bubbling over. More hydrogen peroxide is added until the dark color of the organic matter has largely disappeared; then the sample is washed three times with a NaOAc buffer of pH 5 and once with methanol to remove the remaining released cations (Jackson, 1956).

If the samples have been treated with acid or hydrogen peroxide, they should be re-dried and re-weighed to recalculate the post-treatment net sample weight. Subtracting this value from the original sample weight will determine the weight of the removed constituent.

Soluble salts and exchangable polyvalent cations can be removed by decantation or centrifugation with distilled water to prevent flocculation of the sediment and to give effective dispersion during pipette analysis. However, this decantation may result in partial loss of the colloidal-clay fraction.

Wet sieve (Fig. 8) the sample through a 62-micron (American Society for Testing and Materials (A.S.T.M). Number 230) sieve to separate the sample into coarse (sand and gravel which are retained on the sieve) and fine (clay and silt) fractions (Fig. 9; Tab. 1). Distilled water (Fig. 10) is used

during wet sieving if the fine fraction comprises greater than 5% of the sample. Samples will disaggregate more easily if allowed to soak in a small amount of distilled water or electrolyte solution prior to wet sieving. An approximately 3-5% solution of sodium hexametaphosphate (800 g of purified (NaPO₃)₆, 80 ml of formaldehyde to retard biogenic growth, and 20 liters of distilled water) is recommended for wet sieving if an EMPSA (e.g. Coulter Counter or Elzone) will be used in the fine-fraction analysis and only small amounts of fines are present. This solution acts as an electrolyte, which is necessary when EMPSAs are used.

Seawater may also be used for sample soaking and sieving purposes and as an electrolyte for the Coulter Counter as dictated by the objectives of the specific project. A 0.5% sodium hexametaphosphate solution should be used if the fine fraction is to be analyzed by pipette; at this concentration, the sodium hexametaphosphate acts as a dispersant. In any case, the solution must be filtered to 0.2 microns; a sequential submicron filtration system combining 5, 0.45, and 0.2 micron filters works well (Fig. 11). A rubber policeman and a squeeze bottle of distilled water or the solution provide the best sieving results. The coarser sand and gravel fractions are retained on the screen while the finer silt and clay fractions are collected in a catch pan or mason jar. Wash the coarse fraction into a pre-weighed 100-ml beaker and seal the fine fraction in the mason jar. Wet sieve using only enough liquid to fit in a single 32-ounce mason jar to prevent any possible fractional biasing of the fine fraction.

Dry the coarse fractions in a convection oven at, or slightly under 100 C. When the samples are dry (overnight is usually sufficient), place them in a desiccator to cool before weighing them. Gross coarse weight minus the weight of the beaker gives net coarse weight.

Much of the biogenic calcite found in a sample is commonly in the form of coarse pelecypod or gastropod shell fragments. Rather than treating the bulk sample with acid, it is often easier and less time consuming to manually remove this carbonate from the washed coarse fraction and to re-weigh the decalcified portion to determine the new net coarse weight. The weight of the carbonate must be subtracted from the net sample weight before entering the data into the computer.

The grain-size distribution within the coarse fraction can be determined by use of a settling tube (Figs. 1, 12; Schlee, 1966). However, if this fraction weighs less than 10 g or if it contains greater than about 2% foraminifera, which will not settle properly in a sedimentation column, then wire-mesh sieves must be employed.

If the coarse-fraction grain-size distribution is to be determined by sieving, assemble a bank of sieves, which generally consists of a cover, the -1, 0, 1, 2, and 3 phi sieves, and a catch pan (Table 1; Figs. 9, 13). The bank of sieves is agitated in a shaker (Fig. 13) for a minimum of at least 15 minutes. Although Mizutani (1963) and McManus (1963, 1965) recommend a sieving time of 35 minutes for 8-inch diameter sieves, Royse (1970) suggests 10 minutes as an adequate sieving time. After sieving, weigh the individual sand and gravel phi fractions and record the weights from each phi class. Net coarse weight minus the gravel weight gives the sand weight. If material coarser than 2 mm is present, assemble a bank of gravel-fraction sieves (the cover, the -5 thru -2 phi sieves, and a catch pan); sieve to separate, weigh each phi class, and record the weights. Calculate the relative percentages of the sand-fraction phi classes and the relative percentages of the gravel-fraction phi classes. The sieve data are normally entered directly into a computer using the program RSAM (Appendices A, and C) to obtain a complete grain-size distribution.

Sieving efficiency increases with reduction in load (i.e. smaller sample size) and the maximum sieve load diminishes with mesh size, but it is the quantity of finer near-mesh size particles that actually determines sieving efficiency (Royse, 1970). Twenhofel and Tyler (1941) recommended 40 g of sandy sediment as the maximum load for 8-inch diameter sieves. When selecting a procedure, it is important to remember that sieves sort material not only according to size, but also according to the shape and roundness of the particles (Sahu, 1965; Kennedy and others, 1985).

The settling tube, also called a rapid sediment analyzer (RSA, Figs. 1, 12; Zeigler and others, 1960; Zeigler and others, 1964; Schlee, 1966, Syvitski and others, 1991) provides a means for efficient analysis of sand-sized material by settling the grains through a column of water. One common settling tube design is based on using the pressure differential between two columns of water that have a common head. The pressure change caused by the introduction of sediment within one of the columns is measured by a Validyne DP-103 variable-reluctance wet-wet pressure transducer and amplified by a Validyne Model CD-23 frequency demodulator. Results are relayed to a Dell Dimension personal computer equipped with an ADAC Corporation 5500HR-1 analog-to-digital data acquisition board and associated ADLIBWIN and ADLIBPC data acquisition software drivers. As the sediment settles past the pressure transducer port, the pressure differential decreases with time. Because the sedimentation rate, in accordance with Stoke's Law (Fig. 9), is a function of grain size, one can interpret the sand-fraction grain size distribution from the variation in pressure differential. In addition to the settling tubes that use pressure transducers, other devices incorporate a balance, and measure changing weight with time as the sediment settles.

If the sand-fraction grain-size distribution is to be determined by settling tube and the program RSA2000 (Appendices A and C), the gravel-fraction distribution, if present, must still be determined by sieve analysis. To use a settling tube with a pressure transducer, a warm-up and stabilization time of about an hour is required for the transducer and associated electronics. During warm-up the operator should check the system pressure lines for air bubbles and the settling column water level and turn on the computer and initialize the program RSA2000. Remove all air bubbles from the lines; raise the water level to 3-5 mm above the actual settling tube (Figs. 1, 12). The operator has the option of selecting either of two modes of sample introduction: an automated flipper or manually with a tablespoon (this is generally true for most automated settling tubes). Both methods require practice to achieve reproducibility, but the results of both methods are equally accurate and comparable.

A micro-splitter or cone-and-quartering is used to obtain a sample of about 10 g. Smaller samples generate too weak a signal from the pressure transducer; a weak signal may be diluted by electrical and mechanical noise causing greater error in the reproducibility of the data. Samples larger than 10 g are more apt to form density currents down the sides of the settling tube or, upon introduction, to form clumps of finer sediment which act as much larger particles.

Whether the automated flipper or manual spoon technique is employed, the sample must be dampened with just enough water to afford enough inter-grain cohesion to insure simultaneous introduction of the sample, and to prevent air from becoming trapped in the sample, which may slow the settling rate of the particles.

If the automatic flipper technique is selected, place the sample on the flipper using a tablespoon, flatten the top of the sample to <1 cm in height, and dampen the sample with an eyedropper. When the switch that controls the flipper is shifted, the sample will be introduced into the settling tube. When the flipper is halfway between the vertical and contact with the settling tube, move the switch back to its original position. This will prevent the flipper from crashing into the settling tube and introducing noise into the system.

The manual spoon technique simply involves placing the sample split into a tablespoon, dampening the sample with an eyedropper, and dumping the sample in one smooth, rapid motion from a height of about 3 cm down the center of the settling tube. If the operator misses the center and introduces the sample near the wall of the tube, a density current that flows down the side of the tube will form and the sample will appear to be coarser than its true distribution. The operator is encouraged to run a few standards to check for equipment problems and practice sample introduction. This exercise will help the operator to produce more accurate and reproducible analytical results. After the sample has settled to the bottom of the tube and the pressure differential returns to approximately zero, a hard copy of the sand-fraction grain-size distribution can be generated (Fig. 14), and the data can be saved to disk.

The fine-grained fraction of a sediment is defined as silt (particles with diameters less than 62 microns down to 4 microns; Fig. 9) plus clay (particles with diameters less than 4 microns, with colloidal clay being less than 0.1 microns). Because of their small size, fine-fraction particles are difficult to measure by sieving. Therefore, the classical techniques to analyze the fine-grained portion of grain-size distributions are dominated by sedimentation methods like the hydrometer and pipette methods (Krumbein and Pettijohn, 1938; Milner, 1962; Folk, 1974). These methods involve preparing a dispersed, homogeneous suspension of the fine fraction, dilution with a dispersant solution (usually 0.5% sodium hexametaphosphate) to 1000 ml, and allowing the particles to settle in a graduated cylinder. As the settling of sediment particles continues according to Stoke’s Law, samples are either withdrawn (pipette) or measurements made (hydrometer) of the suspension at preset time intervals. However, many sedimentation laboratories have discontinued or limited the use of pipette and hydrometer techniques because of inherent problems with settling (e.g. Brownian motion), thermal convection, irregular particle shape, mass settling in density currents, rounding errors due to large multiplication factors, and, especially, the time necessary to extend an analysis down into the clay range. For example, over 24 hours are required to extend a pipette analysis down to 11 phi at 20^oC, room temperature (Krumbein and Pettijohn, 1938; Milner, 1962).

Although originally designed to count and size blood cells (Coulter, 1957; Berg, 1958), EMPSAs have found other applications because of the short time and small amount of material required for analysis (Figs. 2, 15). These devices, which are currently manufactured by Coulter/Beckman (Coulter Counter) and Particle Data (Elzone, now affiliated with Micrometrics), are also used in industry, in the biological sciences (Sheldon and Parsons, 1967), and in geology, where they are used to measure the grain size of sediments by measuring differences in electrical resistivity produced by sediment particles (McCave and Jarvis, 1973; Shideler, 1976; Kranck and Milligan, 1979; Muerdter and others, 1981; Milligan and Kranck, 1991). Because the aperture tubes used by EMPSAs can size particles of only 2-40% of the aperture diameter, at least two aperture tubes with overlapping ranges are required to determine the size distribution within the fine fraction of a marine sediment. For example, during a standard multi-aperture analysis using a Coulter Counter, a 200-micron aperture tube would be used to resolve the size distribution between 64 to 8 microns, and a 30-micron aperture tube would be needed to resolve the size distribution between 12 microns down to approximately 0.5-0.7 microns. These individual analyses or passes through the respective aperture tubes are then mathematically combined to produce the fine-fraction (silt plus clay) distribution. The fine fraction is, in turn, combined with the coarse fraction data to generate a complete grain size distribution.

An obvious limitation of this method is that an EMPSA, even if calibrated to resolve down to about 0.6 microns (10.75 phi), can only resolve a portion of the clay size distribution. The fine clay in the 0.6 to 0.1 micron range (some of the 11 phi and all of the 12 phi and 13 phi fractions that extend down to the colloidal clay boundary) is not detected. Although an analysis performed down to 0.6 microns is adequate for most freshwater, estuarine, and shelf environments, the sediments from many deeper water marine environments (e.g. rise and abyssal plain) may contain significant material in the fine clay fraction. By combining a double-point pipette analysis with the data from an EMPSA analysis details of the fine-fraction can be extended through the fine clay range, but this solution is less desirable because of the problems with pipette analyses cited above and because the resultant data are no longer at whole-phi intervals.

Earlier efforts to extrapolate grain size data beyond the limits of the analysis used graphical methods (Schlee and Webster, 1967; Schlee, 1973). For example, if Schlee determined that more than 5% of the fine-grained sediment from a sample was less than 1 micron, additional data points were estimated for the finer sizes at whole-phi intervals by projecting the grain-size curve as plotted on probability paper and following the slope of the line. While this is a practical solution, it is much more labor intensive than the computer program CLAYES2K utilized as part of this system and provided herein.

The fine fraction, which has been stored in mason jars with distilled water or sodium hexametaphosphate solution, may be analyzed by pipette or EMPSA. The EMPSA determines particle volume; the pipette method measures settling rates. The reason for performing the analyses should determine which method is used (see Comments section).

The pipette method consists of preparing a dispersed (by sonic probe), homogeneous (by vigorous stirring) suspension of the fine fraction, dilution with a 0.5% sodium hexametaphosphate solution to 1000 ml, and allowing the particles to settle in a graduated cylinder (Royse, 1970; Folk, 1974). The optimum sample is approximately 20 cm³ (about 15 g dry weight). With more sample, the grains interfere with each other during settling; with less sample, the effects of weighing errors on the analysis increase. Subsamples of 20 ml are withdrawn from the suspension at levels at 5, 10, or 20 cm below the water surface at standard intervals of time (Table 2A). Because all particles of' a given equivalent diameter (based on calculations using Stoke's Law; Fig. 9) will have settled below that level after a standard interval of time, the samples should contain only finer particles. These aliquots are either suction-mounted on pre-weighed filters and rinsed in distilled water or placed in pre-weighed 100-ml beakers. If pre-weighed beakers are used, the operator must remember to account for the weight of the dispersant when calculating the phi-fraction weights. The pipette is rinsed with distilled water after each extraction and the rinse water is also passed through the filter or drained into the pre-weighed beaker with the aliquot. To begin the analysis, start the timer as soon as the stirring rod emerges for the last time. At the end of 20 seconds, insert the pipette to a depth of 20 cm and withdraw the precise subsample (exactly 20 ml). Inasmuch as the subsequent analysis is based on this subsample, this is the most important single step. If excess liquid is accidently drawn into the pipette, do not attempt to return it to the cylinder. Remove the pipette and discard the excess.

Continue the withdrawals at the specified time intervals and depths. The aliquots are dried and weighed, the weights are recorded on an analysis form (Fig. 16), and the size distribution is calculated from the weight of sediment. The principle behind the computation is this: if the fine sediment is uniformly distributed throughout the entire 1000-ml column by stirring, and exactly 20 ml is drawn at each of the stated times, then the amount of mud in each withdrawal is equal to 1/50 of the total amount of mud remaining suspended in the column at that given time and at that given depth (i.e., the amount of mud finer than the given diameter; all particles coarser than the given diameter will have settled past the point of withdrawal). The first withdrawal is made so quickly after stirring and at such a depth that particles of all sizes are present in suspension. Therefore, if we multiply the weight of the first withdrawal by 50, we will obtain the weight of the entire amount of mud in the cylinder. Then if we withdraw a sample at a settling time corresponding to a diameter of 6 phi, and multiply by 50, then we know that the product represents the number of grams of mud still in suspension at this new time, therefore the weight of mud finer than 6 phi. Similarly, we can compute the weight percent at any size and obtain an entire distribution.

Because of the length of time required to complete a whole-phi interval analysis (16 hours, 24 minutes to 10 phi at 20^oC; Table 2A) and because Brownian motion interferes with the settling of particles less than 10 phi, the pipette method is now used 1) to determine the silt/clay boundary in percent gravel-sand-silt-clay analyses and 2) when a sample contains a significant amount of material smaller than 0.6 microns in diameter. This latter condition is important to consider because an EMPSA determines a size distribution based solely on the grain-size range it has been calibrated to analyze (e.g. 0.062 mm >x>0.6 microns) and many low-energy and deepwater environments contain significant amounts of very fine clay and colloidal clay-sized material that are undetected by EMPSAs. The time needed to perform a pipette analysis may be reduced by placing the settling cylinders in a constant temperature bath and lowering the viscosity of the settling, medium by increasing its temperature (Table 2B, C).

The data from a pipette analysis can be converted into frequency or cumulative frequency percentages. These percentages can subsequently be combined with the coarse-fraction data using the programs ENTRY to generate a complete grain-size distribution, statistics, and database files with the program GSTATM (Appendices B and C).

To perform this silt/clay boundary analysis, aliquots are collected at 20 cm depth after 20 seconds elapsed time to determine the concentration of sample in suspension, and at a depth of 10 cm after exactly 2 hours and 3 minutes to determine the concentration of clay in solution. Because only two aliquots are collected, the size of these aliquots are usually enlarged to 50 ml to minimize analytical error. The percentages of silt and clay can be determined from the weight of sediment in the aliquots.

The Coulter Counter, a widely used EMPSA (Figs. 2, 15), permits easy, fast, and accurate analysis of fine-fraction size distributions. The instrument's optimum precision, as with all the above analyses, is realized only if the operator is conscientious in attention to detail and consistently follows established procedures. The following procedure is by no means complete and is intended solely as a set of guide lines. All operators are encouraged to familiarize themselves with the complete instrument manual to achieve the best results.

The Coulter Counter Multisizer IIe EMPSA is activated and operated according to the following steps:

1) The main Coulter Counter unit, sample stand, computer, and printer are turned on and allowed to warm up for 15 minutes. Initialize the computer program CLTRMS2K (Appendices A and C).

2) Using the menu select arrows and the numeric key pad enter the proper information. The main setup menu should read: ORIFICE DIAMETER - 200 microns, SETUP - manual, ANALYSIS - Sample, CALIBRATION - Recall, and SIZE UNITS - microns. The ORIFICE LENGTH, Kd, and SIZE are set by the Multisizer.

3) Step through the other setup screens, ensuring that the proper settings are selected. The ANALYSIS SETUP-1 menu selections should read: CURRENT AND GAIN - Automatic, APERTURE CURRENT - 1600 micro-Amps, GAIN - 1 or 2, POLARITY - Alternate +/-, INSTRUMENT CONTROL - Manual, and TIME - seconds. The ANALYSIS SETUP-2 menu should read: CHANNELS - 256, AUTOSCALING - On, EDIT - Off, COINCIDENCE CORRECTION - Off, ANALYTICAL VOLUME - 0 micro-liters, PARTICLE RELATIVE DENSITY - 1.00, DIFFERENTIAL VALUES - %, and END TONE - On. The COMMUNICATIONS SETUP screen should read BAUD RATE - 9600, DEVICE - Computer, AUTO OUTPUT - No, CHANNEL DATA - No, ANALOG PLOT - No, SCREEN DUMP - Yes, OVERLAY MODE - No, FORMAT - Standard, SEND STX/ETX - Yes, END FIELD CHARACTER - 59, END LINE CHARACTER 0 CR+LF, and LEADING ZEROS - Yes.

4) When the menu check is complete, ensure that the outside of the aperture tube has been rinsed and immersed in a beaker of clean electrolyte, and that the sample stand door is closed. The X-axis should be set to linear diameter; the Y-axis should be set to number differential. Turn the Reset/Count Control clockwise to Reset, applying a vacuum to the aperture tube system, and, when the light in the sample stand turn on, press the FULL key on the Multisizer to display the selected range of particle analysis.

5) Perform a noise check using blank electrolyte. Acceptable cumulative background counts for both the 200- and 30-micron tubes are less than 500 counts per 10 seconds.

6) Shake the fine-fraction portion of the sample in the mason jar vigorously, sonify the bulk sample for about 2-4 minutes (probes typically work much better than baths; Fig. 10), and use a stirrer to completely suspend the sample. While the sample is stirring, transfer a sub-sample, drawn in a sweeping motion from the top, middle, and bottom of the mason jar, to the clean beaker of filtered electrolyte in the sample stand using a disposable pipette. Never allow the tip of the pipette to touch the side or bottom of the mason jar while sub-sampling because it may break. Set stirring motor to adequately mix the sample without producing bubbles, close the door to the sampling stand, turn the Reset/Count Control to reset, and press Full.

7) Ensure that the concentration index meter reads 5-10%. If the concentration is above 10%, add more electrolyte to dilute. If the concentration is below this, add more sample to the beaker until 5-10% is reached. Press the Reset button, wait 5 seconds, and then Start Control buttons to check that the concentration is less than the maximum for coincidence (<10,000 cts/10 s). Coincidence is the occurrence of two or more particles in the sensing zone at one time (Fig. 2). Exceeding coincidence limits will cause data loss in the smaller size classes, thereby causing the observed grain-size distribution to be coarser (Milligan and Kranck, 1991).

8) When the concentration is within the correct limits, continue to accumulate data for about 100 s, then press the Stop button. If using the program CLTRMS2K (Appendices A and C) to acquire the data, press the OUTPUT/PRINTER button at the program’s “Please Send Raw Data” prompt. Data will be sent to the computer and a hardcopy will be generated (Fig. 17).

9) Check the output and, if acceptable, save the data to disk. To proceed to the next sample remove the beaker from the sample stand, cover it with cellophane to avoid contamination, and reserve the sample for the 30-micron aperture tube analysis. It is advisable to perform all of the 200-micron analyses before proceeding to the 30-micron analyses. Complete only enough 200-micron aperture tube analyses, however, that will allow adequate time to perform the corresponding 30-micron aperture tube analyses on the same day.

10) When the 200-micron aperture tube analyses are complete, the computer program and the Multisizer needs to be reset for the 30-micron aperture tube analyses. Change the aperture tube, change the Setup menu to the proper settings, place a clean beaker of electrolyte on the sample stand, and check for acceptable cumulative background count. Do not use the stirrer motor during 30-micron aperture tube analyses.

11) Pour sample saved from 200-micron aperture tube analysis through a clean 20-micron sieve into a clean electrolyte-rinsed beaker. When the sample has passed through the sieve, pass the sample back through the sieve into the original beaker, and then through the sieve back into the new beaker. This re-suspends the sample and ensures that all of the sample has been scalped by the sieve.

12) Immediately place beaker in the sample stand, close the sample-stand door, and turn the Reset/Count Control clockwise to reset. Push the Start button, accumulate data for 100 s, push the Stop button, and push the Print button to send the data to the computer.

a. Make sure bubbles are initially cleared from the aperture tube by opening both stopcocks. This will reduce system noise.

b. If the tube clogs, brush the aperture opening with a sweeping motion. If the tube is still clogged, clear tube in a distilled-water ultrasonic bath. If this fails, replace the distilled water with 50 % nitric acid.

c. Always have aperture current off when not running an analysis or when changing tubes.

d. After the analyses are completed, the sample stand, Faraday cage, vacuum and electrolyte reservoir flasks, and counter top should be rinsed in distilled water and dried with paper towels.

e. When the analyses are complete remove the 30-micron aperture tube, place the 200-micron aperture tube on the sample stand, and set all switches to the proper settings for the 200-micron analysis.

f. Keep the aperture tubes in a cleaning solution (e.g. Coulter Clenz) when not in use.

g. With proficiency, an operator can speed up the analytical procedure by sonifying (Fig. 10) the next sample while entering pertinent identifiers into the computer for a sample that is being analyzed by the EMPSA.

Raw grain-size data are typically in the form of weight percentages of sediment in various size classes. Because many analyses and

commonly more than one data set are involved in geological studies, the method by which these data are presented and compared becomes important when determining how the distributions differ and the magnitude of these differences (Royse, 1970). Both graphical and statistical methods are available. Of the graphical techniques, one of the more common methods is to plot the basic sand, silt and clay percentages on equilateral triangular diagrams (Shepard, 1954; Folk, 1956). This means of data presentation is simple and facilitates rapid classification of sediments and comparison of samples. The USGS sediment lab at the Woods Hole Field Center uses a modification of Shepard’s (1954) ternary classification system (Fig. 18).

The statistical measures of size distributions used by sedimentologists are most commonly based on quartile measures (Trask, 1932), phi notation (Inman, 1952; Folk and Ward, 1957), or the method of moments (Kane and Hubert, 1962; Folk, 1974; Sawyer, 1977). Inclusive graphics statistics, the phi notation method utilized by the programs provided herein (Folk, 1974), are graphical in that the data are read directly from a computer-drawn cumulative-frequency curve at five points (the 5, 16, 50, 84, and 95 percentiles). Verbal equivalents for standard deviation, skewness, and kurtosis based on the inclusive graphics statistics. The percentages of gravel, sand, silt, and clay, and the modified frequency percentages for size distributions ranging from 11 ph (or 13 phi using the program CLAYES2K) to -5 phi are also computed. Method of moments statistics, which involve more complicated arithmetic calculations that account for every grain in the distribution, have gained popularity with the increasing availability of fast, inexpensive computers. The method of moments statistics generated with the programs provided herein include modal classes, modal frequencies, arithmetic mean, median, and standard deviation, skewness, and kurtosis. However, in open-ended distributions (truncated distributions or where a lot of material of unkown size occurs in a pan fraction), the method of moments is probably not justified (Folk, 1974).

The insoluble residue of a sediment is that portion which remains after digestion in hydrochloric acid (HCL) of a prescribed strength under specific conditions. Because carbonate rocks may contain significant amounts of chert, anhydrite, and siliciclastic sand, silt, and argillaceous material, the determination of insoluble residues is useful for comparing sediments from different localities and for expressing their degree of purity or contamination by detrital and authigenic constituents (Milner, 1962).

To perform a typical analysis, approximately 15 g of wet sediment are placed in a pre-weighed 600-ml beaker and dried at 100^oC. The beaker is placed in a desiccator to cool and re-weighed to determine water content. Cool dilute HCL (10%) is slowly added in stages till all effervescence has stopped. After effervescence has ceased, the excess acid is carefully decanted. The insoluble residue is washed at least five times and each washing is decanted. Alternately, the insoluble residue can be washed into a labeled pre-weighed paper filter and throughly rinsed. In any case, the insoluble residue is dried at 100^oC and is calculated by weight as a percentage of the original bulk sample. The difference between 100% and the value obtained is the percent calcium carbonate. Grain-size analyses may be performed on the residues, or they may be examined petrographically.

This procedure will remove calcium carbonate. If significant amounts of dolomite are present, 6N HCL should be used and the suspension must be heated to near boiling to dissolve the dolomite. The water content need be determined only if the samples are marine and a salt correction must be applied, or if appreciable amounts of hygroscopic water are present. Although commercial grade HCL (muriatic acid) may be used, gypsum formation due to the presence of sulfate ions is a concern (Krumbein and Pettijohn, 1938). The acid digestion portion of this analysis should be performed under a fume hood; gloves and safety glasses should be worn.

The purpose for performing the analyses must be considered when selecting which method will be used. For example, the Coulter Counter determines particle volume as opposed to the pipette method which measures settling rates. Therefore, if a researcher is studying flow regimes and hydraulic equivalents, the pipette method might produce more applicable data. Other settling-velocity analysis methods used for fine-grain sizes are the hydrometer (Buoyocoz, 1928) and decantation methods. However, these techniques are more difficult and less accurate (Folk, 1974) and thus are generally not recommended. Numerous journal articles have compared the various techniques used for size analyses of fine-grained suspended sediments. All operators contemplating the use of any of these methods are strongly encouraged to familiarize themselves with this literature. For example, Shideler (1976) and Behrens (1978) compared the Coulter Counter with pipette techniques. Hydrophotometers have been compared with the pipette method (Jordan and others, 1971; Singer and others, 1988) and with the Coulter Counter (Swift and others, 1972). The above fine-fraction considerations also apply to the coarse fraction. The coarse fraction is usually determined by settling-tube analysis. However, if this fraction weighs less than 10 g or contains greater than 2-5% foraminifera which will not settle properly in a sedimentation column, the operator must utilize sieves, which will measure nominal diameter. These and other particle-size methods and considerations are discussed in Barth (1984) and Syvitski (1991).

Aggregates of clay-sized particles, which are difficult to break up, commonly form when samples are dried. Because these aggregates or “bricks” cause the size distribution to appear slightly (about 2-4%) coarser, care should be exercised to sonify samples with appreciable clay-sized material for at least 4 minutes prior to EMPSA analysis to minimize this effect. Other techniques to determine the weight percentages of the coarse and fine fractions that do not involve drying the fine fraction include pipette analyses and splitting the sample. In the latter technique where the sample is split, half is dried to determine the percentages of the coarse and fine fractions and the fine fraction from the other undried half is analyzed by EMPSA. Unfortunately, getting representative splits is difficult because of intra-sample heterogeneity and the splitting process can commonly introduce an error much greater than the original 2-4 percent. For those interested in a determination of the entire clay fraction with the greatest accuracy, pipette analyses or the use of the computer program CLAYES2K (Appendices B and C) are strongly recommended.

Extensive textural analyses performed on standards have shown that the above methods, with the possible exception of the hydrometer, will produce results with an accuracy of better than plus or minus 10% (electro-resistance multichannel particle size analyzer: about 3%; sieves: about 5%; pipette and sand-fraction settling tube: about 5%). However, the optimum precision for the above analyses is realized only if the operator is conscientious in attention to detail and consistently follows established procedure. To monitor precision, the operator should run replicates on at least every tenth sample.

The representativeness of any textural data set depends on the methods used to obtain the data. These methods must be chosen in accordance with the original purpose of the study. If the importance of knowing the actual size distribution outweighs the hydraulic equivalence, the sieve and the particle-counter analyses should be performed rather than the settling-tube and pipette analyses.

The comparability problems between sets of data generated by different devices are not as great as one might expect. For example, the data produced by sieve and settling tube analyses would be remarkably similar for a given sample because a settling tube is usually calibrated using natural sediments sieved to known fractions. On the other hand, pipette and hydrometer analyses will usually produce data indicating a given sample is slightly finer than the data produced on an EMPSA. This result occurs because fine-fraction particles are generally plate shaped and do not settle as fast as the quartz spheres upon which Stoke's Law, and therefore the pipette and hydrometer settling analyses, is based. And again, an EMPSA can not typically be calibrated to analyze the very fine clay (less than 0.6 microns) and colloidal-clay portions of the size distribution.

Six videos with voice commentary were produced as a guide for laboratory technicians and students in the earth-science and oceanographic communities who would like to use the above described methods. They are not intended to be detailed presentations of the procedures, but to serve only as supplements to help familiarize users with the basic principles. Users are strongly encouraged to consult the original references cited above for more detailed analytical descriptions.

The free multi-media software, RealPlayer, must be installed on the user’s computer with a web browser before the videos, which are stored as compressed streaming files, can be viewed. Once installed, if using Internet Explorer, the user can launch the demonstration videos below by clicking on the respective links. If the user's browser does not connect properly to the links below, the user should get out of the browser, go directly to the videos directory of chapter1 and open the respective .rm files with RealPlayer.

The basic lithologic symbols shown in figure 19 are provided as guidelines. Symbols similar to these are generally acceptable for use as patterns both on sediment distribution maps and in lithologic columnar sections, such as on core description forms. Although these are the most commonly used symbols for each of the respective lithologies, the patterns must still be explained separately in a key. As with most graphics, users should also be mindful of the eventual publication scale when choosing line weights.

Figure 20 outlines the software sequence of operation in a generalized flow diagram for those programs used on the computers in the sedimentation laboratory at the Woods Hole Field Center. The data-acquisition programs RSA2000, CLTRMS2K, and SEDITY2K, and the data-processing program CLAYES2K, run in DOS 3.0 or later and in DOS under Windows 3.1,/95/98. The data-acquisition programs GSANV, RSAM, CLTRM, and ENTRY, and the data-processing programs JSORT, GSTAT, and GSTATM, run under Windows 95/98 in the “Born-Again-Shell” (bash). This shell, which is part of the GNU operating system (Hagerty and others, 1998) distributed by the Free Software Foundation, creates a Unix-like environment under Microsoft Windows 95/98.

Raw data records are created on IBM-compatible computers interfaced with an RSA and EMPSA by using the programs RSA2000 and CLTRMS2K, respectively (Appendices A and C). The user is prompted for all the necessary identifiers, and, upon completion of a satisfactory analysis, the data and identifiers can be written to both an output file and a printer. Three raw data records are generated for each sample: an RSA record, a 200-micron aperture tube EMPSA record, and a 30-micron aperture tube EMPSA record. These records each contain a lab number, equipment type, sample identification, project identification, requestor, operator, and analysis date. The RSA data records include sample weight, coarse weight, sand weight, and the relative percentages of the-5 to -1 phi gravel and 0 to 4 phi sand fractions. The EMPSA data records include aperture diameter and tabulated micrometer diameters and the corresponding relative-frequency percentages from the size-fraction channels collected by the Coulter Counter Multisizer IIe. Although the Multisizer IIe generates 256 channels of data, this accuracy is unnecessary when generating whole-phi grain-size distributions. The program CLTRMS2K combines and reduces the data into a more manageable 16-channel format that is equivalent to the data originally produced by the Coulter Counter Model TAII.

The raw RSA and EMPSA data are archived into master files by using the program SEDITY2K (Appendices A and C). This program also contains utilities to inspect, edit, and print hard copies of the data generated by the RSA2000 and CLTRMS2K programs.

Occasionally, distribution and statistics must be generated for samples which were not analyzed or only partially analyzed with the RSA2000 and CLTRMS2K programs. For example, many fine-grained samples do not contain enough sand (< 10 g) to perform settling-tube analyses, and the entire coarse fraction (>0.062 mm) must be determined by sieving or approximated by visual estimates. When this occurs, a raw data master file may be created or updated by keying the raw sediment data directly into a file using the programs RSAM and CLTRM (Appendices A and C). These programs prompt the user for all necessary identifiers and data. The data can then be processed as output from the programs RSA2000 or CLTRMS2K.

The field and navigation parameters are entered by using the program GSANV (Appendix A and C). These parameters and their coding are described in the Request For Analysis forms which have been completed by all personnel who submit samples for particle-size analysis. The field and navigation parameters include lab number, latitude, longitude, area, sampling device, water depth, depth in section to the top of the sample, and depth in section to the bottom of the sample.

The data-processing program JSORT (Appendices B and C) is used to sort the multi-line records according to lab numbers and to output the records that are complete (all RSA and EMPSA data, and field and navigation parameters) to an output file for further processing. The program GSTAT (Appendices B and C) is used to retrieve data from the complete, sorted, raw data files created by JSORT and to compute the modified frequency percentages, method of moments statistics, textural nomenclature and classification (Wentworth, 1929; Shepard, 1954; Figs. 9, 18), and percentages of gravel, sand, silt, and clay. The modified frequency percentages are given for size distributions up to and including 11 phi to -5 phi. The method of moments statistics includes arithmetic mean, median, standard deviation, skewness, kurtosis, modal classes, and modal frequencies. If requested by the user, a hardcopy of this output (Fig. 21), which also includes inclusive graphics statistics and verbal equivalents for standard deviation, skewness, and kurtosis (Folk, 1974), can be produced. The frequency percentages for the corresponding phi classes are computed by the subroutines MPVC, SUMRY, and WTFP (Appendix C). The cumulative frequency-percent curve used in computing the inclusive graphics statistics is approximated by using the International Mathematical and Statistical Library Inc. (IMSL) routines IQHSCU and ICSEVU (Appendix C).

The programs ENTRY and GSTATM (Appendices B and C) allow the user to generate statistics for pre-existent grain-size data produced outside the software system described above. The program ENTRY permits input of sediment data as relative frequency percentages at whole-phi intervals; the program GSTATM performs the functions of the program GSTAT on these data. To correct the errors introduced by the assumption that each value within a phi class is centered at the midpoint of that class, Shepard's Correction (Kenny and Keeping, 1951) has been applied to the second and fourth moments about the mean in the statistics generated by the programs GSTAT and GSTATM.

When creation of a database file is specified in either GSTAT or GSTATM, the cumulative frequency percentages, phi classes, method of moments statistics, and the appropriate identifiers are written to a “pre-database” file in a comma-delimited free-field format. Once the data are in a database, multi-variate analyses such as factor analysis or cluster analysis may be applied to retrieved data to help interpret complicated sedimentological phenomena.

Because an EMPSA can detect only those particles for which it is calibrated, the tail of the fine fraction produced by this instrumentation (below about 0.6-0.7 microns) is commonly truncated. The program CLAYES2K (Poppe and Eliason, 1999; Appendices B and C) allows the user to extrapolate the grain-size distribution contained in the “pre-database” file generated by the program GSTAT to the colloidal-clay boundary. The user may select solutions based on linear or exponential extrapolations, or the mean of both.

Figure 22 presents a stylized plot showing that portion of the grain-size distribution typically truncated by EMPSAs. Also shown in this figure are examples of

linear, exponential, and averaged extrapolations of the clay fraction to 0.1 microns, and the clay-colloidal clay boundary. Errors inherent in the linear and exponential extrapolations have been exaggerated within this figure to more clearly demonstrate their effect on the grain-size distribution. The linear extrapolation tends to slightly overestimate the amount of clay present in a typical distribution and may be used by operators to account for some of the material within the colloidal fraction. The exponential extrapolation tends to slightly underestimate the amount of clay present in a typical distribution and may be used by operators who want data that represents the minimal amount of clay present. Although an operator may select linear or exponential extrapolation, the mean of both extrapolations usually provides the most accurate estimate and is, therefore, the recommended solution for most applications.

RSA2000, CLTRMS2K, SEDITY2K, and CLAYES2K were written and compiled in Microsoft QuickBASIC version 4.5; RSAM, CLTRM, GSANV, JSORT, GSTAT, ENTRY, and GSTATM were originally written in RATFOR (Poppe and others, 1985), but have been rewritten in “C” and compiled with the DJGPP Development System v. 2.01 (Walnut Creek CDROM).

This report is preliminary and has not been reviewed for conformity with U.S. Geological Survey editorial standards (or with the North American Stratigraphic Code). Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government. The software presented herein is provided as a service; no warranty is given or implied. Images on the title page and table of contents are clipart from Corel XARA (v. 1.5). We thank A. Robinson, J. Commeau, and D. Walsh, the sediment lab analysts who assisted during preparation of this report.

Barth, N.G., 1984, Modern methods of particle size analysis, John Wiley and Sons, 309 p.

Behrens, E.W., 1978, Further comparisons of grain size distributions determined by electronic particle counting and pipette techniques: Journal Sedimentary Petrology, v. 48, no. 1, p. 121.3-1218.

Benson, D.J., 1981, Textural analyses with Texas Instruments 59 programmable calculator: Journal Sedimentary Petrology, v. 51, no. 2, p. 61-62.

Berg, R.H., 1958, Electronic size analysis of sub-sieve particles by flowing through a small liquid resistor: ASTM Special Technical Publication 234, p. 245-258.

Blackwood, D.B. and Parolski, K.F., 2000, "SEABOSS": an elegantly simple image and sampling system: Proceedings of the Sixth International Conference on Remote Sensing for Marine and Coastal Environments, Veridan ERIM International, v. 1, p. I-99-I-106.

Blatt, Harvey, Middleton, Gerard, and Murray, Raymond, 1972, Origin of Sedimentary Rocks: Englewood Cliffs, New Jersey, Prentice-Hall, 634 p.

Bothner, M.H., Gill, P.W., Boothman, W.S., Taylor, B.B., and Karl, H.A., 1997, Chemical and textural characteristics of sediments at the EPA reference site for dredged material on the continental slope SW of the Farallon Islands: U.S. Geological Survey Open-File Report 97-87, 50 p.

Buoyocoz, G.J., 1928, The hydrometer method for studying soils. Soil Science, v. 25, p. 365-369.

Coulter, W.H., 1957, High speed automatic blood cell counter and cell size analyzer: proceedings of the National Electronics Conference, p. 1034-1042.

Coulter Electronics, Inc., 1988, Fine particle application notes: Instruction manual, Hialeah, Florida 72 p.

Coulter Electronics, Inc., 1989, Coulter Multisizer AccuComp color software: Reference Manual, Hialeah, Florida, 188 p.

Emery, K.O. and Schlee, J.S., 1963, The Atlantic Continental Shelf and Slope, A Program for Study: U.S. Geological Survey Circular 481, 11 p.

Emmitt, W.W., 1980, A field calibration of the sediment-trapping characteristics of the Helley-Smith Bedload Sampler: U.S. Geological Survey Professional Paper 1139, 44 p.

Folk, R.L., 1956, A review of grain-size parameters: Sedimentology, v. 6, p. 77-93.

Folk, R.L., and Ward, W.C., 1957, Brazos River bar - a study in the significance of grain size parameters: Journal Sedimentary petrology, v. 27, p. 3-26.

Folk, R.L., 1974, The petrology of sedimentary rocks: Austin, Tex., Hemphill Publishing Co., 182 p.

Hagerty, Dan, Weisshaus, Melissa, and Zaretskii, Eli, 1998, GNU software for Ms-Windows and Ms-DOS: Boston, MA, Free Software Foundation, 108 p.

Honjo, S., Spencer, D., and Gardner, W., 1992, A sediment intercomparison experiment in the Panama Basin: Deep-Sea Research, v. 39, p. 333-358.

Inman, D.L., 1952, Measures for describing size of sediments: Journal Sedimentary Petrology, v. 19, p. 51-70.

Jackson, M.L., 1956, Soil chemical analysis: advanced course: Published by the author,. Madison, Wis., 895 p.

Jones, S.B., and Simpkin, P., 1970, A computer program for the calculation of hydrometer size analyses: Marine Geology, v. 9, p. M23-M29.

Jordan, C.F., Jr., Fryer, G.E., and Hemmen, E.H., 1971, Size analysis of silt and clay by hydrophotometer- Journal Sedimentary Petrology, v. 41, no. 2, p. 489-4 96.

Kane, W.T., and Hubert, J.F., 1962, FORTRAN program for the calculation of grain-size textural parameters on the IBM 1620 computer: Sedimentology, no. 2, p. 87-90.

Kranck, K. and Milligan, T.G., 1979, The use of the Coulter Counter in studies of particle-size distributions in aquatic environments: Bedford Institute Report Series B1-R, v. 79, p. 7.

Kennedy, S.K., Meloy, T.P., and Durney, T.E., 1985, Sieve data - size and shape information: Journal Sedimentary Petrology, v. 55, p. 356-360.

Kenny, J.F., and Keeping, E.S., 1954. Mathematics of statistics, Part 1: Princeton, N.J., D. Van Nostrand Company, Inc., 348 p.

Krumbein, W.C., and Pettijohn, F.J., 1938, Manual of Sedimentary Petrology, New York, Appleton, Century, and Crofts, 549 p.

Lay, E.P., 1969, Core sampling techniques, Oceans Magazine, v. 1, no. 3, p. 27-29.

Lee, H.J., and Clausner, J.E., 1979, Sea-floor sampling and geotechnical parameter determination - handbook: Technical Report R873, Civil Engineering Laboratory, Naval Construction Battalion Center, Port Hueneme, CA, 127 p.

McCave, I.N., and Jarvis, J., 1973, Use of the Model T Coulter Counter in size analysis of fine to coarse sand: Sedimentology, v. 20, p. 305-315.

McCave, I.N., and Syvitski, J.P.M., 1991, Principles and methods of particle size analysis, In J.P.M. Syvitski (ed.), Principles, Methods, and Applications of Particle Size Analysis, New York, Cambridge University Press, p. 3-21.

McLane Research Laboratories, 2000, Mark 78G-21 Sediment Trap: http://www.mclanelabs.com/

McManus, D.A., 1963, Sieve calibration: Journal Sedimentary Petrology, v. 33, p. 953-954.

McManus, D.A., 1965, A study of maximum load for small diameter sieves,: Journal Sedimentary Petrology, v. 35, p. 792-796.

Milligan, T.G., and Kranck, Kate, 1991, Electro-resistance particle size analyzers, In J.P.M. Syvitski (ed.), Principles, Methods, and Applications of Particle Size Analysis, New York, Cambridge University Press, p. 109-118.

Milner, H.B., 1962, Sedimentary Petrography: New York, Macmillan Company, 643 p.

Mizutani, S., 1963, A theoretical and experimental consideration on the accuracy of sieving analysis: Journal Earth Science, v. 11, 27 p.

Muerdter, D.R., Dauphin, J.P., and Steele, G., 1981, An interactive computerized system for grain size analysis of silt using electro-resistance: Journal of Sedimentary Petrology, v. 51, p. 647-650.

North Carolina State University, 1999, Shipek grab sampler: Web Site, http://www.ncsu.edu/coast/research/shipek/shipek5.jpg

PetterssonO., 1928, A new appartus for taking bottom samples: Svensk Hydrological Biological Komm Skrift, v. 6, p. 1-7.

Phleger, F.B., 1951, Ecology of foraminifera, northwest Gulf of Mexico: Geological Society America Memoir 46, Part 1, 88 p.

Poppe, L.J., Eliason, A.H., and Fredericks, J.J., 1985, APSAS: an automated particle-size analysis system: U.S. Geological Survey Circular 963, 77 p.

Poppe, L.J., and Elisaon, A.H., 1999, An interactive computer program to extrapolate the clay fraction distributions of truncated grain-size data: U.S. Geological Survey Open-File Report 99-27, 44 p.

Reineck, H.E., and Singh, I.B., 1980, Depositional sedimentary environments with references to terrigenous clastics: 2d ed., Berlin and New York, Springer-Verlag, 549 p.

Rickly Hydrological Company, 2000, US BLM-84 --USGS-designed bedload samplers: http://www.rickly.com/

Rigler, J.K., Collins, M.B., and Williams, S.J., 1981, A high-precision digital-recording sedimentation tower for sands: Journal of Sedimentary Petrology, v. 51, p. 642-644.

Rosfelder, A.M., and Marshall, N.F., 1967, Obtaining large, undisturbed and oriented samples in deep water, In A.F. Richards (ed.) Marine Geotechnique, Proceedings of 1966 International Conference in Marine Geotechnique, University of Illinois Press, p. 243-262.

Royse, C.F., 1970, An introduction to sediment analysis: Tempe, Arizona, Arizona State University, 180 p.

Sahu, B.K., 1965, Theory of sieving: Journal Sedimentary Petrology, v. 35, p. 750-753.

Sawyer, M.B., 1977, Computer program for the calculation of grain-size statistics by the method of moments: U.S. Geological Survey Open-File Report 77-580, 15 p.

Schlee, J., 1966, A modified Woods Hole Rapid Sediment Analyzer: Journal Sedimentary Petrology, v. 30, p. 403-413.

Schlee, J., and Webster, J., 1967, A computer program for grain-size data: Sedimentology, v. 8, p. 45-54.

Schlee, J., 1973, Atlantic continental shelf and slope of the United States -- sediment texture of the northeastern part: U.S. Geological Survey Professional Paper 529-L, 64 p.

Sheldon, R.W., and Parsons, T.R., 1967, A practical manual on the use of the Coulter Counter in marine research: Toronto, Coulter Electronics, 66 p.

Shepard, F.P., 1954, Nomenclature based on sand-silt-clay ratios: Journal Sedimentary Petrology, v. 24, p. 151-158.

Shideler, G.L., 1976, A comparison of electronic particle counting and pipette techniques in routine mud analysis: Journal Sedimentary Petrology, v. 46, no. 4, p. 1017-1025.

Singer, J.K., Anderson, J.B., Ledbetter, M.T., McCave, I.N., Jones, K.P.N., and Wright, R., 1988, An assessment of analytical techniques for size analysis of fine-grained sediments: Journal Sedimentary Petrology, v. 58, p. 534-543.

Smith, W., and McIntyre, A.D., 1954, A spring-loaded bottom sampler: Journal Marine Biological Association, v. 33, p. 257-264.

Swift, D.J.P., Schubel, J.R., and Sheldon, R.W., 1972, Size analysis of fine-grained suspended Sediments--a review: Journal Sedimentary Petrology, v. 42, no. 1, p. 122-134.

Syvitski, J.P.M., Asprey, K.W., and Clattenburg, D.A., 1991, Principles, design, and calibration of settling tubes: In J.P.M. Syvitski (ed.), Principles, Methods, and Applications of Particle Size Analysis, New York, Cambridge University Press, p. 3-21.

Trask, P.D., 1932, Origin and environment of source sediments of petroleum: Houston, TX, Gulf Publication Company, 323 p.

Twenhofel, W.H., and Tyler, S.A., 1941, Methods of study of sediments: McGraw Hill, New York, 183 p.

Wentworth, C.K., 1929, Method of computing mechanical composition of sediments: Geological Society of America Bulletin, v. 40, p. 771-790.

Zeigler, J.M., Whitney, G.G., Jr., and Hayes, C.R., 1960, Woods Hole Rapid Sediment Analyzer: Journal of Sedimentary Petrology, v. 30, p. 490-495.

Zeigler, J.M., Hayes, C.R., and Webb, D.C., 1964, Direct readout of sediment analyses by settling tube for computer processing: Science, v. 145, p. 51.

Figure 1. Schematic diagram showing the basic design of a rapid sediment analyzer (RSA; Schlee, 1966), a settling tube.

Figure 2. Basic design of an electro-resistance multichannel particle size analyzer (EMPSA).

Figure 3. Generalized flow diagram outlining the laboratory techniques and the computer software used in the sedimentation laboratory at the Woods Hole Field Center.

Figure 4. A form used at the USGS Woods Hole Field Center to request grain-size analyses. Form supplies sample identifiers and defines the purpose of the project and the laboratory procedures to be employed.

Figure 5. Form used to record sample identifiers and assign the corresponding lab numbers. Information supplied includes: submitter, cruise id, project id, latitude, longitude, water depth in meters, sampling device area, top depth within the sediment column in centimeters, and bottom depth within the sediment column in centimeters. Information from this form is entered into a computer using the programs GSANV or ENTRY.

Figure 6. Form used to record and calculate analytical data during grain-size and carbonate analyses. Grain size data from this form are entered into a computer using the programs RSAM and RSA2000.

Figure 7. Photograph showing a typical convection oven used to dry samples and a desiccator used to store samples prior to weighing.

Figure 9. Correlation chart showing the relationships between phi sizes, millimeter diameters, size classifications (Wentworth, 1929), and ASTM and Tyler sieve sizes. Chart also shows the corresponding intermediate diameters, grains per milligram, settling velocities, and threshold velocities for traction.

Figure 10. Photograph showing a typical distillation system used to produce distilled water and a sonic probe used to disaggregate particles prior to EMPSA and pipette analysis.

Figure 11. Basic filtration system used to remove submicron-sized particles from the electrolyte/dispersant used during EMPSA and pipette analyses.

Figure 12. Photograph of a rapid sediment analyzer, a settling tube. This instrument (Schlee, 1966) uses a pressure transducer to monitor the settling of sediment with time. An amplifier demodulator on the shelf next to the settling tube relays electrical information to and analog-to-digital converter in the computer.

Figure 13. Photograph of a bank of sieves mounted in a sieve shaker. The shaker sits in a sound-proof cabinet.

Figure 14. Hard copy of a coarse-fraction raw data record produced by the computer program RSA2000. Plot is stylized.

Figure 15. Photograph of a Coulter Counter Multisizer IIe and associated computer hardware.

Figure 16. Form used to record and calculate sample identifiers and data generated during a pipette analysis. Data from pipette analyses can be entered into a computer using the program ENTRY.

Figure 17. Hard copy of a fine-fraction raw data record produced by the computer program CLTRMS2K.

Figure 18. Sediment classification scheme modified from Shepard (1954) used by the programs GSTAT and GSTATM.

Figure 19. Basic lithologic symbols commonly used on sediment distribution maps and in lithologic columnar sections.

Figure 20. Generalized flow diagram showing the available grain-size analysis software and its sequence of operation.

Figure 21. Hard copy of a data record produced by the computer programs GSTAT and GSTATM. Hard copy shows sample identifiers, size distribution, method of moments and inclusive graphics statistics, and verbal equivalents.

Figure 22. Stylized plot showing that portion of the clay fraction typically truncated by electro-resistance particle size analyzers. Also shown are examples of linear, exponential, and average of the linear and exponential extrapolations to 0.1 microns, the clay-colloidal boundary by the program CLAYES2K. Errors inherent in the linear and exponential solutions have been exaggerated to show their effect.

Figure 24. Raw coarse-fraction data records from the Rapid Sediment Analyzer produced by the computer program RSA2000 and output into a file by the program SEDITY2K. The first line, which is the master file name assigned by the operator with the program SEDITY2K, must be deleted before further processing.

Figure 25. Raw fine-fraction data records from the Coulter Counter (200- and 30-micron analyses) produced by the computer program CLTRMS2K and output into a file by the program SEDITY2K. The first line, which is the master file name assigned by the operator with the program SEDITY2K, must be deleted before further processing.

Figure 26. Data file produced by the computer programs GSTAT and GSTATM and used as input for the computer program CLAYES2K. Fields are delimited by commas; records are delimited by dollar signs. A header file has been inserted as the first record to show the field attributes and their order.

Figure 27. Hard copy of a data record produced by the computer program CLAYES2K. The sample identifiers and the original and revised data are shown.

Figure 28. Data file produced by the computer program CLAYES2K. Fields are delimited by commas; records are delimited by dollar signs. A header file has been inserted on the first line to show the field attributes.

4.) RSA2000

NAME: RSA2000

TYPE: Main program

PURPOSE: To create raw rapid sediment analyzer (RSA; Figs. 1, 12) data records

OPERATING SYSTEM: DOS version 3.0 or later, or DOS under Windows 3.1/95/98

SOURCE LANGUAGE: Microsoft QuickBASIC

SOURCE CODE: RSA2000.BAS

COMPILED SOFTWARE: RSA2000.EXE

SEE ALSO: This program requires and/or generates the following associated data files. The files contain only ASCII (viewable) data and reside in the same directories (folders) as the associated executable.

ADAC.DAT -- Initialization file required to set up ADAC a/d board. An example file is included using IRQ 4. The IRQ must be reserved for the card. Please see ADAC documentation for more information.

RAWINDEX.DAT -- Required to run program. May be set initially to just contain ASCII “0” (character zero).

RawDat*.DAT -- *= 1,2,3 Intermediate data generated by the program during successive runs.

RSA.DAT -- Reduced data generated by the program. No initial file is needed but the file contains data from all saved runs. It is maintained by the program SEDITY2K.EXE.

RSA.NDX -- Index file to RSA.DAT. An initial file will be created upon first run.

MIXED_AD.LIB -- Library file needed in the development environment (QuickBASIC) to compile the program

MIXED_AD.QLB -- Quick library file needed in the development environment to compile the program.

RSA2000.MAK -- A MAK file that uses WINDOWS.BAS, BITS.BAS, BIOSCALL.BAS, MOUSSUBS.BAS, and KEYS.BAS. These files are included in Microsoft QuickBASIC Programmer’s Toolbox and are also needed in the development environment to compile the program.

INPUT: At the beginning of each run, the operator will be prompted for the plotter scale, operator, requestor, cruise id, project id, sample id, lab number, sample weight, coarse weight, sand weight, and relative percentages of the phi classes from the gravel fraction.

OUTPUT: A raw data record stored to disk and a hard copy (Fig. 14)

USAGE: To use the program from DOS, change to the working directory and type: RSA2000

To use the program from Windows 95/98, access it from the file manager or the Run command

The user is prompted for all necessary identifiers and data. For each run enter the plotter scale (program defaults to full scale). For each analysis enter:

a) Operator

b) Requestor

c)Cruise id

d) Project id

e) Sample id

f) Lab Number

g) Sample weight

h) Coarse weight

i) Sand weight

j) Relative percentages of the phi classes from the gravel fraction

The sample is introduced into the settling tube when the “Waiting For Sample Introduction” prompt is given.

A copy of the identifiers, the data, and the cumulative frequency plot of the pressure change versus time are generated on the printer (Fig. 14).

The operator may then save , reprint, smooth, or, if the analysis was no good, purge the data. A default saves the data. Upon responding to a choice, the program is recycled.

A default (carriage return) to any of the prompted identifiers other than sample id or Lab Number will enter the identifier from the previous analysis.

RESTRICTIONS: Operator, requestor, cruise id, and project id may be up to 19 characters and have no imbedded spaces or commas. Speed requirements necessitate the BASIC source code be compiled into machine language.

AUTHOR/MAINTENANCE: Eliason Data Services, Mashpee, MA/L. Poppe, USGS, Woods Hole, MA

5.) CLTRMS2K

NAME: CLTRMS2K

TYPE: Main program

PURPOSE: To create raw EMPSA (Coulter Counter; Figs. 2, 15) data records

OPERATING SYSTEM: DOS version 3.0 or later, or DOS under Windows 3.1/95/98

SOURCE LANGUAGE: Microsoft QuickBASIC

SOURCE CODE: CLTRMS2K.BAS

COMPILED SOFTWARE: CLTRMS2K.EXE

CLTR.NDX -- Index file to CLTR.DAT. An initial file will be created in the working directory upon first run.

CLTR.DAT -- Reduced data generated by the program. The file contains data from all saved runs. It is maintained by the program SEDITY2K.EXE.

MS_LABEL.DAT -- Required. The filename is spelled that way. This is a list of parameter names for the Multisizer II currently in use.

MSII.DAT -- Required in order to use the test option (3) on the input port option menu (MSII=msii in lowercase).

INPUT: At the beginning of the first run, the operator will be prompted for diagnostics, printed output, and printer setup. Subsequently and for each sample thereafter the operator will be prompted for operator, requestor, cruise id, project id, sample id, lab number, and tube diameter.

OUTPUT: A raw data record stored to disk and a hard copy (Fig. 17)

USAGE: To use the program from DOS, change to the working directory and type: CLTRMS2K

To use the program from Windows 95/98, access it from the file manager or the Run command

For each analysis enter:

a) Operator

b) Requestor

c)Cruise id

d) Project id

e) Sample id

f) Lab Number

g) Tube diameter

h) Initial micron diameter

Press the Print/Plot button on the EMPSA after the analysis is complete and the “Waiting For Sample” prompt is given.

A copy of the identifiers and raw data are generated on the printer.

The operator may then save, reprint, or, if the data are no good, purge the data. Upon responding to any of the three choices, the program is recycled.

A default (carriage return) to any of the prompted identifiers other than sample id or Lab Number will enter the identifier from the previous analysis.

RESTRICTIONS: This program runs only with output from a Beckman-Coulter Electronics Multisizer IIe EMPSA. Operator, requestor, cruise id, and project id may be up to 19 characters and have no imbedded spaces or commas. Speed requirements necessitate the BASIC source code be compiled into machine language.

AUTHOR/MAINTENANCE: Eliason Data Services, Mashpee, MA/L. Poppe, USGS, Woods Hole, MA

6.) SEDITY2K

NAME: SEDITY2K

TYPE: Main program

PURPOSE: To archive the raw RSA and EMPSA data into master files subsequently used by the processing programs. SEDITY2K also examines, edits, and generates additional hard copies of these records.

OPERATING SYSTEM: DOS version 3.0 or later, or DOS under Windows 3.1/95/98

SOURCE LANGUAGE: Microsoft QuickBASIC

SOURCE CODE: SEDITY2K.BAS

COMPILED SOFTWARE: SEDITY2K.EXE

CLTR.DAT -- The user will specify the data file name & location.

CLTR.NDX -- The program will modify the index file as appropriate to the tasks performed.

RSA.DAT -- The user will specify the data file name & location.

RSA.DAT -- The program will modify the index file as appropriate to the tasks performed.

SEDXFER.DAT -- The program will create a new data transfer file.

SEDXFER.OLD -- Before rewriting the transfer file, the program will archive the last transfer file in case of problems.

INPUT: RSA and EMPSA raw data records, which are stored to disk during operation of the RSA2000 and CLTRMS2K programs, are input into files accessed by SEDITY2K

OUTPUT: Master files of raw RSA (Fig. 24) and EMPSA (Fig. 25) data records stored to disk or a hard copy of the raw-data records. Note: the first line of each master file is a user-supplied project identifier that must be removed before the sorting routines are applied.

USAGE: To use the program from DOS, change to the working directory and type: SEDITY2K

To use the program from Windows 95/98, access it from the file manager or the Run command

User is prompted for which raw data records (RSA, EMPSA, or both) are to be accessed. Operations performed by the program are listed in and selected from the main menu (Table 3). When a number 1 through 7 is entered, the operator can use that utility to perform the corresponding task.

In the Display or Print mode, the user can scroll through the list of logged records. This list is composed of Lab Numbers and letters (O: open, D: deleted, or A: assigned) designating the records present status.

When options 3 or 4 are selected, the Assign and Edit/Display Mode menu (Table 4) appears and the data records scroll up from the bottom. The records may be scroll as in the Display or Print modes, but, when O, D, or A commands are given, the indicated record is tagged for that respective operation.

When the Edit command (E) is given, the raw data record for the corresponding Lab Number appears. The user is then prompted for editing commands.

When all raw data records in the SEDITY2K files have been written to Master files or deleted, the K key will permanently remove these records from disk.

AUTHOR/MAINTENANCE: Eliason Data Services, Mashpee, MA/L. Poppe, USGS, Woods Hole, MA

APPENDIX B. SOFTWARE DOCUMENTATION AND CODE FOR DATA-PROCESSING SOFTWARE

1.) JSORT 11/3/98

NAME: jsort - to sort sediment analyses samples by lab number and retrieve complete sets which will be appended to the complete data set file.

SYNOPSIS

jsort file_name

DESCRIPTION: This version of the program, which was written in “c” and compiled with DJGPP (v. 2.01), will run under Windows 95/98 in the Born-Again-Shell (Bash). The program jsort expects the input file to have a file-name suffix with “.dat”. The “.dat” is not entered on the run line, but is used solely for file management purposes. The input file is produced by combining or concatenating the raw data records from the GSANV, RSAM, CLTRM, and, through SEDITY2K, the RSA2000 and CLTRMS2K programs. The input file is reorganized by 'pack' to contain one analysis type on only one physical record. This 'packed' file is piped into a) a temporary file "jsortnn" where nn is the process_id; and, b) into the program 'dirct'. A directory of which analyses were entered into the input file for each lab number is listed on the terminal. The listing is piped through 'more'. The original data file “file_name.dat’ is left unaltered.

Upon completion of the directory listing, the user is queried as to whether the file should be split. This refers to retrieving complete data sets from the sorted temporary file. If the user responds 'y' to the prompt "Enter 'y' to split file:" the complete samples will be copied to a file 'file_name.fin’. If incomplete samples remain, an output file ('file_name.inc') will be created with only the incomplete samples. If the file exists, the data are appended to the file. The samples will be sorted and contain only one physical record per analysis.

The temporary file “jsortnn" will be purged from the system.

FILES

sedlab.h

Wet Sieving video	Dry Sieving video
Rapid Sediment Analyzer video	Coulter Counter video
Pipette video	Constant Temperature Bath video