An Examination of the 22 May 2014 Duanesburg, New York, Unexpected Tornadic Supercell

Around 1900 UTC on 22 May 2014, a long-lived supercell dropped up to 4.00” hail in Amsterdam, New York and spawned an EF3 tornado in the towns of Delanson and Duanesburg, New York. The storm was remarkable in that the morning antecedent convective available potential energy (CAPE) and vertical shear values did not suggest the possibility of supercells. We examine radar data, lightning data, in-situ observations, and model analyses in the hours leading up to and during the event to identify factors that led to the initiation and maintenance of the supercell, including the large hail and tornado.

The supercell formed between 1500 and 1600 UTC at a triple point, and then moved south along a north-south oriented stationary front that was maintained by strong differential shortwave heating. High equivalent potential temperatures in the Mohawk Valley at the intersection of stronger heating to the east and moisture pooling along the stationary front increased the CAPE to 2500 J/kg locally. This, combined with the low wet bulb freezing levels around 2500 m and mesocylonic circulation, led to the formation of large hail as the storm passed through the valley around 1900 UTC. Falling surface pressures to the west of the boundary led to increased backing of surface winds over the Hudson and extreme eastern Mohawk Valleys, providing an enhanced region of >10 m/s 0–1 km and >20 m/s 0–6 km vertical wind shear that created a localized environment favorable for the reinvigoration of the mesocyclone and formation of the tornado between 1900–2000 UTC. Interaction of the mesoscale flows, such as the rear flank downdraft, with the terrain may have also played a crucial role in the storm evolution.

This unexpected case highlights the challenge of severe convective forecasting in the northeastern United States, particularly due to the complex interaction of synoptic, mesoscale, and terrain features that may not be initialized or well-resolved by the current generation of convective-permitting models.