Sensitivity of Dryline Convection Forecasts to Upstream Forecast Errors for Two Weakly Forced MPEX Cases

Forecast sensitivity analysis provides a quantitative method of evaluating how a particular forecast metric responds to uncertainty in earlier-time forecast fields or processes.  Recent studies have primarily focused on applying these methods to cases with clear synoptic-scale forcing for convective initiation.  These results suggest that uncertainty in mid and upper tropospheric features, which may or may not have been directly involved in convective triggering, subsequently modulate the position of near-surface boundaries and hence convective initiation in specific locations.  By contrast, cases with less obvious synoptic forcing for convection might be of greater interest because these cases might exhibit more sensitivity to the location of particular features, or specific processes at earlier times.  Two of the Intensive Observing Period (IOP) days from Mesoscale Predictability Experiment (MPEX) exhibited these qualities, namely convective initiation in the Texas Panhandle and western Kansas within broad southwesterly flow and no clear synoptic feature to focus convection.  This study applies the ensemble-based sensitivity analysis to 3-km resolution Weather Research and Forecasting (WRF) ensemble model forecasts from MPEX for 27 and 28 May.  Preliminary analysis suggests that convection in the Texas Panhandle on 28 May is sensitive to the upstream mid-tropospheric equivalent potential temperature and to the net effects of overnight convection over west Texas, which modified the strength of the southerly return flow during the next day.  The relationship between the regions of sensitivity and MPEX dropsondes will also be discussed.