Visual descriptions of sediment texture can be converted into complete particle-size distributions by the Particle-Size toolbox by one of two methods. Which method is used depends on the quality threshold used.


Here the quality threshold is simply the cumulative sum of visually described component percentages. Those that sum closer to 100% are considered higher quality. Records above the quality threshold are simulated using Method 1, while records below the threshold are integrated using Method 2.
  • Sediment samples are often described by estimates of their relative proportions of gravel, sand silt and clay. These can be converted to cumulative frequency distributions by summing these values together, starting with clay and working towards gravel.

Zp: sand, Zm: sand median, Sp: silt, Lp: clay.

Estimates of median particle size for specific size fractions can also be used if it is assumed that they lie halfway between the cumulative percentage frequency values of adjacent size categories. For example, the cumulative percentage value of the sand median is given as: zm = sp+((zp-sp)/2).

  • In many cases the relative proportions of gravel, sand, silt and clay only contribute to a very small proportion of the overall sample mass (i.e. the sample is organic) or these data are missing from a sample description.

  • If a sediment sample has been assigned a lithological classification, then an alternative solution is to statistically model a particle-size distribution on the basis of measured particle-size distributions from the same lithology type.

  • The descriptive statistics of measured particle-size distributions used for modelling are:
    • mean and standard deviation of prinicpal component scores of log-ratio particle-size distributions.

  • The mean and standard deviation of each component are used to randomly simulate a set of statistically probably principal component scores, which are then converted back into particle-size distributions.

Above: 52 measured particle-size distributions for clay samples in geology type NAWA. Below: 10 simulated distributions.