Particulate matter suspended in the River Severn (Shropshire, UK) cons
ists chiefly of clay-sized mineral particles, together with living and
dead micro-organisms (algae and bacteria). Its concentration depends
strongly on discharge, but the particle size distribution shows no sys
tematic variability. For most samples, the particle volume is log-norm
ally distributed with respect to diameter, the mean diameter being ca.
9 mum. The particles are mainly aggregates, including some with linea
r dimensions of the order of tens or hundreds of micrometres. Particle
density depends appreciably on size, decreasing from ca. 2.5 x 10(6)
g m-3 at a diameter of 2.5 mum to ca. 1.3 x 10(6) g m-3 at 20 mum. The
collision efficiency factor for particle aggregation is estimated to
be 0.01-0.03. At low discharge, the 'dead zone' in the River Severn at
Leighton is a well defined region of stagnant water behind a gravel b
ar. The rate of deposition of fine particles on its bed is of the orde
r of tens of grams per square metre per day. Resuspension requires a c
ritical bed shear velocity of 0.03-0.04 m s-1, which occurs at main ri
ver discharges greater than about 150 m3 s-1. Under such conditions th
e gravel bar is underwater and the dead zone is a region of highly tur
bulent return flow. A simple mechanistic model of particle dynamics in
the dead zone accounts reasonably well for particle accumulation rate
s when run with parameter values based on measured particle and hydrau
lic properties. Calculations with the model suggest that most of the s
edimentation flux to the dead zone bed is due to particles with equiva
lent sphere diameters in the range 30-240 mum. Simulations indicate th
at deposition proceeded continuously during spring and summer, whereas
repeated deposition and resuspension occurred in autumn and winter.