Assessing the skill of high-impact weather forecasts in southern South America: a study on cut-off lows

Abstract. Cut-off lows (COLs) are mid-tropospheric cyclonic systems that frequently form over southern South America, where they can cause high-impact precipitation events. However, their prediction remains a challenging task, even in state-of-the-art numerical weather prediction systems. In this study, we assess the skill of the Global Ensemble Forecast System (GEFS) in predicting COL formation and evolution over the South American region where the highest frequency and highest intensity of such events are observed. The target season is austral autumn (March to May), in which the frequency of these events is at its maximum. Results show that GEFS is skillful in predicting the onset of COLs up to 3 d ahead, even though forecasts initialized up to 7 d ahead may provide hints of COL formation. We also find that as the lead time increases, GEFS is affected by a systematic bias in which the forecast tracks lie to the west of their observed positions. Analysis of two case studies provides useful information on the mechanisms explaining the documented errors. These are mainly related to inaccuracies in forecasting the vertical structure of COLs, including their cold core and associated low-level circulation. These inaccuracies potentially affect thermodynamic instability patterns (thus shaping precipitation downstream) as well as the horizontal thermal advection which can act to reinforce or weaken the COLs. These results are expected to provide not only further insight into the physical processes at play in these forecasts, but also useful tools for the operational forecasting of these high-impact weather events over southern South America.