Cohesive sediment transport is important for many geophysical and engineering applications, for example, the rapid siltation in navigation channels and harbors (Hayter & Mehta, 1986;Winterwerp et al., 2000), cohesive sediment transport in salt marsh (Graham & Manning, 2007) and long-term delta morphology (Edmonds & Slingerland, 2010). Cohesive sediment can absorb pollutants (e.g., heavy metals, pesticides and nutrients) and is a concern for water quality (Ongley et al., 1992). Modeling suspended sediment dynamics is also essential for the understanding of biogeochemical cycles. Transport of estuarine and nearshore cohesive sediment therefore plays an important role in coastal processes and the functioning of healthy ecosystems (Grabowski et al., 2011;Deegan et al., 2012). Settling velocity is one of the key parameters in sediment transport. Cohesive sediment can bind together to form flocs and large flocs can breakup by shear. A floc size distribution often develops in sediment suspension and flocs of different size will settle at different velocities. The settling velocity of cohesive sediment strongly depends on the flocculation dynamics (Winterwerp & Van Kesteren, 2004). However, due to complex surface physico-chemical and biological properties of cohesive sediment, cohesive sediment dynamics are complex and empirically parameterized (Winterwerp et al., 2006;Kuprenas et al., 2018). It is fundamentally important to understand the cohesive sediment dynamics in natural environment, in particular how flow turbulence affects the flocculation dynamics, and hence the floc size distribution and the settling velocity.Flocculation plays an important role in cohesive sediment transport. To represent flocculation of cohesive sediment more realistically, models based on either characteristic floc diameter (Winterwerp, 1998;Son & Hsu, 2008) or size-classes (McAnally & Mehta, 2002;Maerz et al., 2011) have been developed. Models based on the characteristic floc diameter simply assume flocs to have a single equilibrium size, which can either be derived from a single-class population balance equation or empirical equations (Maggi, 2009). Kuprenas et al. (2018) improved the Winterwerp model (Winterwerp, 1998) to consider the upper limit on floc size due to turbulent shear. However, bimodal floc size distribution (FSD) with the first peak around microflocs by direct contact