Improving trajectory analyses using advection correction: Tests with a 30-m simulation of the 24 May 2011 El Reno, OK, supercell

A major limitation to the accuracy of trajectory analyses constructed from observational or numerically simulated datasets is the necessarily finite value of the temporal sampling frequency (data output frequency). Errors can be particularly large in cases of severe convective phenomena characterized by high velocities and small length scales. This study is concerned with a technique to mitigate these errors by incorporating an advection correction procedure (space to time conversion) into the integration of the trajectory equations. The advection correction is based on a procedure proposed by Gal-Chen (1982) to mitigate radar data analysis errors arising for the non-simultaneity of the data collection. Tests of the advection-corrected trajectory analyses are conducted with data from a 30 m resolution CM1 numerical model simulation of a supercell storm that impacted central Oklahoma on 24 May 2011 (the CM1 simulation is discussed by L. Orf elsewhere at these proceedings). The advection-corrected (AC) trajectory analyses are compared with traditional linear time-interpolated (LTI) trajectory analyses for a range of data output frequencies and a range of pattern translation velocity estimates corresponding to different sub-domain (box) sizes. Although the AC trajectory analyses are seen to provide substantial improvements over the LTI trajectory analyses for data output periods typically used in generation of high spatial resolution datasets (e.g., 5 or 10 sec), our presentation will focus on weaknesses as well as strengths of the procedure.