Non-buoyant Regime

What about cases where not enough air is entrained into the column? What happens when a column collapses?

A simple analogy is what happens when one stands in the backyard with a garden hose pointed up -- the hose fountains. In the case of a volcanic eruption, the volcano "fountains." but instead of sedning streams of water down the sides of the mountain, clouds of hot rock and gas, pyroclastic density currents (PDC's), are shed instead.

These currents can travel at speeds exceeding 50 m/s (Burgisser and Bergantz, 2002), and present substantial hazard to life and property. Though PDC's can be generated by dome collapse as well as column collaspe, this page will only discuss the latter. However, the discussion below is broadly applicable to either method of formation.

Surges versus Flows

PDC's can broadly be classified into two groups, pyroclastic surges and pyroclastic flows. The former are comprised of dense granular flow at the base and dilute flow at the top, whereas the latter are entirely dilute. Additionally, flows tend to follow topographic lows, such as valleys, whereas surges can go virtually anywhere -- in an eruption of Fisher Caldera, in the Aleutian arc, surges overtopped a 1500 m mountain range north of the caldera (Gardner, personal communication).

Pyroclastic flows have produced extensive deposits in volcanic regions. In many polaces flows can form deposits 100's of meters thick, such as at Katmai, where up to 300 m of flow deposits are present near the Novarupta dome.

These deposits can be recognized in the field by their lack of internal stratification and poor sorting. Both of these characteristics indicate that PF's are highly turbulent flows lacking a high degree of internal structure.

Pyroclastic surges are far more dilute than PF's and behave quite differently. Where flows lack internal structure and deposit massive units, surges can be well sorted and often deposit finely laminated beds or even anti-dune beds, indicative of deposition from a very high speed, low concentration flow.

It should, however, be noted that surges often occur as valley-overtopping portions of pyroclastic flows.

Surges can also, and often do, elutriate a coignimbrite plume off their tops. this plume behaves very similar to plumes discussed earlier, however, instead of having a single virtually-point source, the plume is sourced from the length of the surge.

Competer simulation of a collapsing volcanic column and the associated generation of a coignimbrite plume, from Dobran et al., 1993. The left-most column shows particle fraction, the center column shows gas temperature, and the right-most column shows the mixture density.

Cartoon from Fujii and Nakada (1999) showing a model for coignimbrite plume generation and also how flows may segregate a valley-overtopping surge.