Detailed monitoring at Mount St. Helens since 1980 has enabled prediction of the intermittent eruptive activity (mostly dome growth) with unprecedented success. During 1981 and 1982, accelerating deformation of the crater floor around the vent (including radial cracks, thrust faults, and ground tilt) was the earliest indicator of impending activity. Numerical experiments using the finite element method to model the mechanics of the crater floor show that all deformation features can be explained by a uniform shear-stress boundary condition along the conduit wall. The magnitude of the shear stress required to match observed displacements (1--7 MPa) is inversely proportional to the conduit diameter (estimated to be 25--100 m). The most probable source of this shear stress is the flow of viscous magma up the conduit and into the lava dome.

We propose a model in which the accelerating deformation, beginning as much as 4 weeks before extrusions, is caused by the increasing velocity of ascending magma in the conduit. This model is examined by using deformation data of the dome before four extrusions in 1981 and 1982 to estimate the volumetric flow rate through the conduit. This flow rate and an estimate of the effective viscosity of the magma enable calculation of an ascent velocity and an applied shear stress that, again, depend on the conduit diameter. The results of these calculations are consistent with the finite element experiments and show that the proposed model is feasible. In light of this model, events observed just before or near the time extrusions began, such as reversals of ground tilt direction from outward to inward and the sudden decrease in the number of shallow earthquakes, may indicate an abrupt decrease of shear stress in the conduit. This could be explained by a decrease in either the ascent velocity, or the effective viscosity of the magma ascending through the shallow conduit, or both, near the time of extrusion. Precursory deformation like that measured at Mount St. Helens should be observable at similar volcanoes elsewhere because it is caused by fundamental process of magma ascent. ¿ American Geophysical Union 1988