Abstract. Mineral dust particles from wind-blown soils are known to act as effective ice nucleating particles in the atmosphere and are thought to play an important role in the glaciation of mixed phase clouds. Recent work suggests that feldspars are the most efficient nucleators of the minerals commonly present in atmospheric mineral dust. However, the feldspar group of minerals is complex, encompassing a range of chemical compositions and crystal structures. To further investigate the ice-nucleating properties of the feldspar group we measured the ice nucleation activities of 15 characterized feldspar samples. We show that alkali feldspars, in particular the potassium feldspars, generally nucleate ice more efficiently than feldspars in the plagioclase series which contain significant amounts of calcium. We also find that there is variability in ice nucleating ability within these groups. While five out of six potassium-rich feldspars have a similar ice nucleating ability, one potassium rich feldspar sample and one sodium-rich feldspar sample were significantly more active. The hyper-active Na-feldspar was found to lose activity with time suspended in water with a decrease in mean freezing temperature of about 16 • C over 16 months; the mean freezing temperature of the hyper-active K-feldspar decreased by 2 • C over 16 months, whereas the "standard" K-feldspar did not change activity within the uncertainty of the experiment. These results, in combination with a review of the available literature data, are consistent with the previous findings that potassium feldspars are important components of arid or fertile soil dusts for ice nucleation. However, we also show that there is the possibility that some alkali feldspars may have enhanced ice nucleating abilities, which could have implications for prediction of ice nucleating particle concentrations in the atmosphere.

Abstract. Mineral dust particles are thought to be an important type of ice-nucleating particle (INP) in the mixed-phase cloud regime around the globe. While K-rich feldspar (K-feldspar) has been identified as being a particularly important component of mineral dust for ice nucleation, it has been shown that quartz is also relatively ice-nucleation active. Given quartz typically makes up a substantial proportion of atmospheric desert dust, it could potentially be important for cloud glaciation. Here, we survey the ice-nucleating ability of 10 α-quartz samples (the most common quartz polymorph) when immersed in microlitre supercooled water droplets. Despite all samples being α-quartz, the temperature at which they induce freezing varies by around 12 ∘C for a constant active site density. We find that some quartz samples are very sensitive to ageing in both aqueous suspension and air, resulting in a loss of ice-nucleating activity, while other samples are insensitive to exposure to air and water over many months. For example, the ice-nucleation temperatures for one quartz sample shift down by ∼2 ∘C in 1 h and 12 ∘C after 16 months in water. The sensitivity to water and air is perhaps surprising, as quartz is thought of as a chemically resistant mineral, but this observation suggests that the active sites responsible for nucleation are less stable than the bulk of the mineral. We find that the quartz group of minerals is generally less active than K-feldspars by roughly 7 ∘C, although the most active quartz samples are of a similar activity to some K-feldspars with an active site density, ns(T), of 1 cm−2 at −9 ∘C. We also find that the freshly milled quartz samples are generally more active by roughly 5 ∘C than the plagioclase feldspar group of minerals and the albite end member has an intermediate activity. Using both the new and literature data, active site density parameterizations have been proposed for freshly milled quartz, K-feldspar, plagioclase and albite. Combining these parameterizations with the typical atmospheric abundance of each mineral supports previous work that suggests that K-feldspar is the most important ice-nucleating mineral in airborne mineral dust.

A minute fraction of atmospheric particles exert a disproportionate effect on the phase of mixed-phase clouds by acting as ice-nucleating particles (INPs). To understand the effects of these particles on weather and climate, both now and into the future, we must first develop a quantitative understanding of the major INP sources worldwide. Previous work has demonstrated that aerosols such as desert dusts are globally important INPs, but the role of biogenic INPs is unclear, with conflicting evidence for their importance. Here, we show that at a temperate site all INPs active above −18 °C at concentrations >0.1 L−1 are destroyed on heating, consistent with these INPs being of biological origin. Furthermore, we show that a global model of desert dust INPs dramatically underestimates the measured INP concentrations, but is consistent with the thermally-stable component. Notably, the heat sensitive INPs are active at temperatures where shallow cloud layers in Northern Europe are frequently observed to glaciate. Hence, we suggest that biogenic material is important for primary ice production in this region. The prevalence of heat sensitive, most likely biogenic, INPs in this region highlights that, as a community, we need to quantify the sources and transport of these particles as well as determine their atmospheric abundance across the globe and at cloud altitudes.

Ice-nucleating particles (INPs) have the potential to remove much of the liquid water in climatically important mid- to high-latitude shallow supercooled clouds, markedly reducing their albedo. The INP sources at these latitudes are very poorly defined, but it is known that there are substantial dust sources across the high latitudes, such as Iceland. Here, we show that Icelandic dust emissions are sporadically an important source of INPs at mid to high latitudes by combining ice-nucleating active site density measurements of aircraft-collected Icelandic dust samples with a global aerosol model. Because Iceland is only one of many high-latitude dust sources, we anticipate that the combined effect of all these sources may strongly contribute to the INP population in the mid- and high-latitude northern hemisphere. This is important because these emissions are directly relevant for the cloud-phase climate feedback and because high-latitude dust emissions are expected to increase in a warmer climate.

No abstract