One objective of both the Convective Storm Initiation Project (CSIP, which took place in southern England in 2005) and the 2002 International H2O Project (IHOP_2002, which took place in the southern Great Plains of the USA) was to generate high quality observations of the initiation of deep convection. These are being used to both improve our understanding of this process and also evaluate the processes that are represented by the new generation of convection permitting numerical weather prediction models. Forecasting convective weather remains a challenge for numerical weather prediction, particularly for elevated nocturnal storms.

In this talk I will present three contrasting case-studies from IHOP and CSIP, two of which were elevated.  In the first CSIP case secondary initiation took place due  to both a cold pool outflow and trapped convectively generated gravity  waves. The second case from CSIP is the only case of elevated  convection from the CSIP project. An elevated mesoscale convective system (MCS) was  observed, which two rear-inflow jets.  Downdrafts from the elevated MCS  could not penetrate the stable air  below,  but instead led to a wave in the stable undercurrent, which  propagated with the  MCS, and was located just ahead of where the  rear-inflow jet impacted  on the undercurrent. Finally, I will  discuss ongoing work on a case  of nocturnal convection from IHOP_2002 (13th June). In this case elevated initiation generated an  MCS, which gave a surface cold pool outflow, with the surface properties of the outflows depending on the strength of the nocturnal stable layer. Both bores and waves were common, due to the low-level stable air, and led to secondary initiation. A squall-line structure was  observed to form once the cold pool was able to lift surface air to its level of free convection.