Physics and Chemistry of Frost-Flowers and Their Resultant Sea -Salt Aerosol

Lead Research Organisation: British Antarctic Survey
Department Name: Physical Sciences

Abstract

Polar regions are strong indicators of the impact of environmental factors upon global climate, as indicated by the Antarctic ozone hole, by large temperature increases on the Antarctic Peninsula, and by Arctic sea ice loss. Ice core records also contribute a powerful record of past climate changes provided they can be interpreted accurately. Traditionally, it has been assumed that greater sea salt concentrations in ice cores showed less sea ice and stronger winds, with sea salt being generated by the action of wind on open ocean water. However, it is now believed that sea salt aerosol in polar regions is generated mostly from the sea ice surface within the ice pack, including from frost flowers, rather than from open water. The aerosol causes apparent negative non-sea-salt sulphate in ice cores because of fractionation during freezing. This implies that the reality is diametrically opposed to the traditional view: increased sea salt in ice cores is associated with more sea ice, and more freshly freezing areas within the ice pack with light winds. This sea ice origin of sea salt aerosol via frost flowers would make sense of the observation that the flux of sea salt to ice cores is larger during glacial periods than warm periods. Frost flowers are also almost certainly responsible for surface ozone loss in spring by production of reactive bromine gases from the flowers or their aerosol. Our understanding of these complex processes is vital if we are to fully account for past and future polar environmental changes. We must also parameterise them so that we can model the changes, but at present there is no complete parameterisation of either the formation of frost flowers or their dispersal into the atmosphere as aerosol. Here, we will measure the physics and chemistry of frost flower growth and the dispersal of their aerosol in the laboratory, and proceed to parameterisation of the processes involved.

Publications

10 25 50