The constitutive relation that describes the amount of nonwetting fluid pha
se entrapment is critical to the modeling of multiphase flow in porous medi
a, but it has received insufficient attention in the literature. We studied
this relation using both experimental and modeling approaches: we used a n
ondestructive, X-ray monitored, long-column experiment that yielded a rich
data set for two different porous media; we also used a quasi-static networ
k model to simulate the experimental data and examine mechanisms affecting
the relation. The experimental work yielded a significant data set for resi
dual nonwetting phase (NWP) saturation as a function of maximum funicular n
onwetting phase saturation. We suggest a functional form that represents th
e observed data sets accurately. Network model calibration to experimental
data yields acceptable model-data agreement and a clear understanding of co
nstraints that should be satisfied when using such models to avoid physical
ly unrealistic behavior. We found that the pore-throat size characteristics
and the snap-off process occurring in pore throats strongly influence the
manifestation of pore-body-size characteristics during imbibition and nonwe
tting fluid phase entrapment. We examined an estimation method proposed by
Wardlaw and Taylor for the residual-funicular relation. We observed that th
e method yields an unrealistic relation for each porous media in the long-c
olumn experiments, and we used network modeling to understand the criteria
that ensure a realistic estimate. (C) 2000 Elsevier Science Ltd. All rights
reserved.