Settling velocity of sediment grains, Part 1: Natural sediment particles are not ellipsoids

Particle settling velocity has wide application in sedimentology and engineering. The ability to predict settling velocity based on fluid and particle properties is essential for realistic sediment transport models. It is common practice to describe the size and shape of natural sediment grains with an ellipsoidal model. This paper tests the validity of the ellipsoidal model for predicting particle settling velocity for siliciclastic and carbonate sediments. The parameters associated with this model are the ellipsoidal nominal diameter (Dn = (DsDiDl)1/3) and Corey shape factor (CSF=Ds/(DiDl)1/2), calculated from the short (Ds), intermediate (Di), and long (Dl) axial diameters. An equation for the drag coefficient (i.e., nondimensional friction) expressed as a function of particle Reynolds number (i.e., nondimensional particle size) and Corey shape factor (i.e., ellipsoidity) was reconstructed for siliciclastic and carbonate grains separately, based on the largest available compilations of existing datasets (639 siliciclastic and 3 666 carbonate grains). The ellipsoidal approximation of grain size yields significant errors in settling velocity predictions, yet the equation is more accurate for siliciclastic particles (mean absolute error = 14 %) than for carbonate grains (mean absolute error = 21 %). This difference is explained by the enlarged irregularity of particles in carbonate compared to siliciclastic sediments, due to enhanced angularity and surface roughness of skeletal remains. Thus, the ellipsoidal model, using the nominal diameter and Corey shape factor, is not suitable for describing natural sediments or predicting their settling velocity, despite its widespread use. A companion paper (Part 2, Slootman et al., 2026) investigates alternative particle size and shape descriptors.