Rock avalanches, debris flows, and related phenomena consist of grain-fluid
mixtures that move across three-dimensional terrain. In all these phenomen
a the same basic forces govern motion, but differing mixture compositions,
initial conditions, and boundary conditions yield varied dynamics and depos
its. To predict motion of diverse grain-fluid masses from initiation to dep
osition, we develop a depth-averaged, three-dimensional mathematical model
that accounts explicitly for solid- and fluid-phase forces and interactions
. Model input consists of initial conditions, path topography, basal and in
ternal friction angles of solid grains, viscosity of pore fluid, mixture de
nsity, and a mixture diffusivity that controls pore pressure dissipation. B
ecause these properties are constrained by independent measurements, the mo
del requires little or no calibration and yields readily testable predictio
ns. In the limit of vanishing Coulomb friction due to persistent high fluid
pressure the model equations describe motion of viscous floods, and in the
limit of vanishing fluid stress they describe one-phase granular avalanche
s. Analysis of intermediate phenomena such as debris flows and pyroclastic
flows requires use of the full mixture equations, which can simulate intera
ction of high-friction surge fronts with more-fluid debris that follows. Sp
ecial numerical methods (described in the companion paper) are necessary to
solve the full equations, but exact analytical solutions of simplified equ
ations provide critical insight. An analytical solution for translational m
otion of a Coulomb mixture accelerating from rest and descending a uniform
slope demonstrates that steady flow can occur only asymptotically. A soluti
on for the asymptotic limit of steady flow in a rectangular channel explain
s why shear may be concentrated in narrow marginal bands that border a plug
of translating debris. Solutions for static equilibrium of source areas de
scribe conditions of incipient slope instability, and other static solution
s show that nonuniform distributions of pore fluid pressure produce bluntly
tapered vertical profiles at the margins of deposits. Simplified equations
and solutions may apply in additional situations identified by a scaling a
nalysis. Assessment of dimensionless scaling parameters also reveals that m
iniature laboratory experiments poorly simulate the dynamics of full-scale
flows in which fluid effects are significant. Therefore large geophysical f
lows can exhibit dynamics not evident at laboratory scales.