Skewness for sand‐sized particles by laser diffraction: Looking inside the bias

ABSTRACT

Laser diffraction (LD) is progressively replacing sieving as the standard technique for particle size analysis (PSA) in sedimentology, a transition that raises concerns about data continuity for long‐term geological and environmental monitoring. However, the comparability of statistical parameters describing the grain‐size distribution remains an open question, particularly those that define the distribution's asymmetry, such as skewness (SK). SK is one of the most relevant parameters from a geoscience perspective, and the transition from traditional dry sieving to LD poses a challenge to the continuity of geoscience data sets. This study, by analysing paired data sets, evaluates the feasibility of correlating SK values derived from traditional sieving with those obtained via LD. Moving beyond known instrumental biases, this research employs a component‐by‐component deconstruction of Folk & Ward's SK formula—separating the central portion ( component, involving φ 16 and φ 84) from the distribution tails ( component, involving φ 5 and φ 95)—alongside the application of categorical metrics to demonstrate that no viable mathematical or classification‐based correction exists to align LD‐derived SK with sieving standards. Moreover, data analysis reveals that LD overestimates specific percentiles in a nonlinear way (with extreme variance driven by outliers), an error that the formula's mathematical structure inherently amplifies. This artificially forces naturally skewed distributions to become more symmetric, thereby misrepresenting the asymmetry of the distributions. As a result, skewness values calculated from LD data using traditional formulas lack both direct geological significance and equivalence to legacy data. Consequently, the use of these values poses a high risk of misinterpretation; therefore, researchers should not only abandon attempts to correct LD skewness to match sieving data but also refrain from using this parameter entirely if derived from LD analysis. This conclusion highlights a significant gap between modern LD methods and nearly a century of established knowledge regarding grain‐size distributions and depositional processes.