Smoke-plume distributions above large-scale fires: implications for simulations of 'nuclear winter'.

Abstract

Smoke from raging fires produced in the aftermath of a major nuclear exchange has been predicted to cause large decreases in surface temperatures. However, the extent of the decrease and even the sign of the temperature change depend on how the smoke is distributed with altitude. We present a model capable of evaluating the initial distribution of lofted smoke above a massive fire. Calculations are shown for a two-dimensional slab version of the model and a full three-dimensional version. The model has been evaluated by simulating smoke heights for the Hamburg firestorm of 1943, and a smaller-scale oil fire which occurred in Long Beach in 1958. Our plume heights for these fires are compared to those predicted by the classical Morton-Taylor-Turner theory for weakly buoyant plumes. We consider the effect of the added buoyancy caused by condensation of water-laden ground level air being carried to high altitude with the convection column, as well as the effects of background wind on the calculated smoke plume heights for several fire intensities. We find that the rise height of the plume depends on the assumed background atmospheric conditions, as well as the fire intensity. Little smoke is injected into the stratosphere unless the fire is unusually intense or atmospheric conditions are more unstable than we have assumed. For intense fires, significant amounts of water vapor are condensed, raising the possibility of early scavenging of smoke particles by precipitation.-Authors