National Severe Storms Laboratory

On the case- and scale-dependence of the predictability of precipitation patterns by a storm-scale ensemble forecasting system

Madalina Surcel

McGill University in Montreal, Canada

03 May 2016, 4:00 PM

National Weather Center, Room 1313
120 David L. Boren Blvd.
University of Oklahoma
Norman, OK

This presentation discusses the scale- and case- dependence of the predictability of precipitation by the CAPS SSEF run during NOAA’s Hazardous Weather Testbed (HWT) Spring Experiments of 2008-2013. The effect of different types of ensemble perturbation methodologies on quantitative precipitation forecasts (QPF) is quantified as a function of spatial scale. It is found that uncertainties in the large-scale initial and boundary conditions (IC/LBC) and in the model microphysical parameterization scheme are responsible for most of the forecast error, causing a loss of predictability at scales smaller than 200 km after 24 h. Other types of ensemble perturbation methodologies, such as small-scale initial condition perturbations and physics only perturbations, were found to have a lesser impact on QPF.

In terms of the case-dependence of predictability, the analysis of individual events indicates that ensemble spread and QPF skill are better for events characterized by large precipitation coverage and convective equilibrium than for weakly-forced events. Also, in agreement with previous studies, accounting for the uncertainty in the model microphysical parameterization is more important for weakly-forced cases than for strongly-forced cases, and properly accounting for small-scale errors in the ICs is especially desirable for weakly-forced cases. However, the search for statistically significant relationships between fractional precipitation coverage or the convective-adjustment time-scale and predictability indicators for the 149 events under study yielded disappointing results. This shows that a more thorough investigation is needed to understand the relationship between convection and large-scale forcing and to determine more suitable parameters to classify between weakly- and strongly- forced events.