a)Convective-scale high impact weather systems often develop in response to conditions created by planetary and synoptic scale circulations, both in tropical and extra-tropical latitudes. We need to understand aspects of those large scale circulations that create conditions leading to hazardous weather, e.g. large scale moisture transport, and hence define their relevance to predictability on timescales of up to 15 days. 

b)Anecdotal evidence suggests that weather is less predictable than usual during some high impact weather events. We need to understand whether this is the case in general, and if so, to quantify the predictability that is achievable for high impact weather. In particular, we need to understand whether and how often this is reflected in the existence of bifurcations in the forecast trajectory (e.g. changes of track of Tropical Storms, splitting or not of severe convective storms). 

c)The impact of most high impact weather is heightened if the weather system is stationary or quasi-stationary. We need to understand what mechanisms produce quasi-stationary hazardous weather systems and how predictable they are. 

d)The dynamical structure of most high impact weather systems is influenced by diabatic heating. We need to understand the role of diabatic heating in creating the conditions for high impact weather, whether its effects can be quantified from observations and the level of complexity with which moist diabatic processes must be represented in order to represent the interaction with Rossby wave dynamics accurately.   

e)Most weather-related hazards are experienced at or near the surface and are influenced by surface and boundary layer processes, many of which are parametrized even in the finest resolution NWP models. We need to understand the role of those processes, e.g. in triggering convection, so as to enable development of parametrizations that better represent their influence, especially for urban areas and mountainous terrain. 

f)Some hazards depend on pre-conditioning (e.g. saturated ground for flooding, dry vegetation for fire, dry ground for excess heat, cold ground for snow & freezing rain). We need to understand how best to represent this pre-conditioning in prediction systems. 

g)For flooding from storm surges, we need to better understand the relationship between storm structure, bathymetry and surge response 

h)For flooding from convection, we need to explore the use of stochastic parametrizations to represent convection uncertainty. 

i)For flooding from rainfall, we need to better understand the relationship between what can be observed (e.g. by radar and satellite) and the aspects of the atmospheric state that determine its short range evolution. 

j)For wildfires, we need to understand the aspects of the wind field that are critical to determining the behavior of the fire, and the way in which the fire feeds back onto the wind field at those scales. 

k) For extreme winds, we need to understand the sensitivity of surface wind predictions to model vertical resolution and boundary layer parametrization. 

l)For winter weather, we need to understand the sensitivity of predictions of precipitation phase to different aspects of the model, including the microphysics parametrization, and the role of surface and boundary layer preconditioning. 

m)For heat and air pollution, we need to better understand the role of the urban canopy in creating dangerous conditions, and the sensitivity of forecasts to the accuracy of the boundary layer inversion and the representation of aerosol and chemistry, so as to enable the design of efficient coupled models.