The effects of organized upstream convection on downstream precipitation

As communicated by National Weather Service forecasters in the Southeast US, past events have demonstrated a weakness in the ability of numerical weather models to accurately simulate the effects of upstream convection on quantitative precipitation forecasts downstream of the convection, especially for cases with rapidly moving mesoscale convective systems (MCSs).

A number of events were examined in which either numerical model forecasts predicted heavy precipitation, and/or observational analyses revealed the occurrence of heavy precipitation. These cases demonstrated that the presence and character of organized convection upstream of the study area exerted an influence on the downstream QPF. Inspection of these cases indicates that there are at least two primary differing scenarios for upstream convection in the Southeast US: The first of these scenarios features upstream convection that moves quickly eastward with respect to the speed of the parent system. In these cases, the presence of intense upstream convection appears to reduce precipitation amount in the study region by (i) disrupting moisture transport as the convectively altered momentum field becomes more westerly, by (ii) moisture removal in the upstream air mass, or by (iii) stabilizing the upstream air mass. Analysis of several cases of this type reveals that model QPF often exhibits a strong positive bias in these instances. Alternatively, convection moving slowly with respect to the main system may enhance moisture transport via a diabatically-enhanced low-level jet (LLJ), and has been observed to thereby increase downstream precipitation.

The purposes of this research are to identify the physical mechanism(s) associated with the convection that most strongly affect downstream QPF, and also to pinpoint the mechanism(s) that are misrepresented in operational numerical forecasts in order to eventually improve predictions in upstream convection events.

Results from two detailed case studies show that NWP model inability to sufficiently resolve the near-storm convective environment in fast-moving convective lines may inhibit accurate downstream QPFs in the case of a quickly-propagating upstream convective feature. Initial results suggest that model misrepresentation of convective propagation are tied to the QPF bias. A comparison case characterized by slowly propagating upstream convection further reveals that imperfect representation of the convection itself may preclude an accurate forecast of diabatically-enhanced features (such as the LLJ) that are important to moisture transport and precipitation production in the downstream environment. In the slow convective cases, model errors were generally smaller and a negative QPF bias was noted in several cases.

Forecasting techniques with which to evaluate these model biases in an operational forecasting setting are suggested.