The Relationship of Localized Severe Convection to Local and Non-Local Irrigation

Numerous studies have indicated a sensitivity of high-impact weather events (e.g., precipitation) to local soil moisture. Here we attempt to understand how soil moisture, both locally and non-locally, can influence specific severe convective elements. On May 20, 2013 a tornado struck the town of Moore, OK, killing 24 people and causing substantial damage to property and the environment. Since the strength and type of such localized storms like the one that caused the Moore tornado can be modulated by the character of both their inflow and the surrounding environmental state, it is reasonable to expect lower atmospheric perturbations caused by irrigation to play a role in the storm's development. Initial efforts at perturbing the atmospheric state with a convection-allowing mesoscale model (the Weather Research and Forecasting, or WRF model) have proved difficult in quantifying the effects of irrigation due to the apparent unrealistic ultra-high sensitivity of moist physics to perturbations. In turn, here we use an ensemble approach to assess the relationships between the character of the convection and soil moisture perturbations, potentially mitigating these unrealistic effects. This approach employs ensemble sensitivity, which uses a linear regression between aspects of the storm and different, perturbed soil moisture states to understand the extent to which irrigation plays a role in the nature of localized severe storms. WRF ensemble 36-hr forecasts are used at 3-km grid spacing over the western two-thirds of the U.S. to investigate these relationships.