© 2015 American Geophysical Union. Transients in water content are well known to mobilize colloids that are retained in the vadose zone. However, there is no consensus on the proper model formulation to simulate colloid release during drainage and imbibition. We present a model that relates colloid release to changes in the air-water interfacial area (A < inf > aw < /inf > ) with transients in water content. Colloid release from the solid-water interface (SWI) is modeled in two steps. First, a fraction of the colloids on the SWI partitions to the mobile aqueous phase and air-water interface (AWI) when the A < inf > aw < /inf > increases during drainage. Second, colloids that are retained on the AWI or at the air-water-solid triple line are released during imbibition as the AWI is destroyed. The developed model was used to describe the release of Escherichia coli D21g during cycles of drainage and imbibition under various saturation conditions. Simulations provided a reasonable description of experimental D21g release results. Only two model parameters were optimized to the D21g release data: (i) the cell fraction that was released from the SWI (f < inf > r < /inf > ) and (ii) the cell fraction that partitioned from the SWI to the AWI (f < inf > awi < /inf > ). Numerical simulations indicated that cell release was proportional to f < inf > r < /inf > and the initial amount of retention on the SWI and AWI. Drainage to a lower water content enhanced cell release, especially during subsequent imbibition, because more bacteria on the SWI were partitioned to the AWI and/or aqueous phase. Imbibition to a larger water content produced greater colloid release because of higher flow rates, and more destruction of the AWI (smaller A < inf > aw < /inf > ). Variation in the value of f < inf > awi < /inf > was found to have a pronounced influence on the amount of cell release in both drainage and imbibition due to changes in the partitioning of cells from the SWI to the aqueous phase and the AWI.