AbstractResults from a one‐dimensional mixed‐layer model with explicit microphysics and a discrete ice‐spectrum were compared to in situ data provided by the Met Office C‐130 aircraft. Bimodal ice‐spectra were observed. The first mode was found at diameters near the lower limit of the instrumentation resolution (25 µm) at a roughly constant concentration. The second mode was found initially at diameters exceeding 200 µm but it shifted to larger sizes and smaller peak concentrations as depths below the cloud top increased.The objectives of the paper are threefold: validation of the model with real cloud data; an analysis of parameters controlling the rate of aggregation, and to show that a secondary ice‐production process is needed to maintain the first mode and produce a second mode which is sensible.A reasonable first mode could only be maintained by producing a constant supply of small ice particles at a rate of at least 10 m−3s−1. Primary nucleation was unable to supply this rate of ice‐particle production at temperatures above −20°C. In the model, it was achieved by invoking a secondary ice‐particle production process, such as the break‐up of evaporating ice crystals. The rate of production required is in broad agreement with laboratory investigations of secondary ice‐particle production during evaporation.The relative importance of the processes of aggregation, deposition and sedimentation to the production of the bimodal spectra is also studied. While it is possible to grow ice particles to large sizes by deposition alone, the number concentration observed in these clouds produces so much competition for water vapour that deposition growth is very limited. The aggregation process is the strongest mechanism for the production of the larger ice‐particles, which then precipitate. Consequently, its strength controls the rate of small ice‐particle production needed, and so is mainly responsible for the second mode. Copyright © Royal Meteorological Society, 2003. P. R. Field's contribution is Crown copyright.
Read full abstract