A categorical assessment of forecast skill, uncertainty and biases in extended-range ensemble forecast of stratospheric regime changes

“Relevance to NOAA’s MAPP competition and long-term climate goal: The proposed work is highly relevant to the MAPP completion: Research to Advance Prediction of Subseaonal to Seasonal Phenomena (ID: 2542967) as it examines the ability of operational NWP systems to represent underlying predictability sources, both physical and dynamical, that influence the subseasonal phenomena of stratospheric regime changes. During the winter season troposphere-stratosphere coupling provides a dynamic mechanism for the stratosphere to influence the troposphere on subseasonal timescales. Stratosphere-troposphere coupling can manifest itself as a stratospheric regime change or, in extreme cases, a sudden stratospheric warming (SSW). On subseasonal timescales, SSWs have been linked to extreme weather and climate, such as cold air outbreaks and negative Arctic Oscillation (AO), which can impact the U.S., Europe and Asia.

When troposphere-stratosphere coupling is skillfully resolved in a forecast, it can provide valuable information to public and private sector forecasters and end users in industry. However, there is a gap in our understanding of the skill, uncertainty and biases in forecasts of stratospheric regime changes that is currently a source of confusion for decision makers. In addressing NOAA’s long term climate goal (e.g., NGSP section I.A.i), this research aims to close the gap in our understanding such that end-users of subseasonal forecast data will be better informed and have higher confidence in planning and decision-making.

The increase in our understanding of the two-way coupling between the troposphere and stratosphere has led many operational centers to raise the top of their NWP models to more fully resolve the stratosphere. Raising the top in a model was the first step in the goal to improve forecast skill beyond the medium range. The next steps in reaching this goal are to understand when stratospheric forecasts of wave coupling events are skillful and to increase our understanding of the sources of uncertainty in the forecast of troposphere-stratosphere wave coupling events. This research addresses these next steps by conducting research to assess stratospheric forecast skill and assess sources of uncertainty in operational NWP models during wave coupling events that produce a stratospheric regime change.

The proposed work explicitly examines how model physics, model horizontal resolution, model vertical resolution, and model top level impact the forecast skill and uncertainty of the physical (i.e., diabatic processes) and dynamical (i.e., wave activity) forcing for stratospheric regime changes at several forecast lead times. For the analysis, the physical forcing is represented by metrics that quantify the diabatic processes in the mid- and upper troposphere and the dynamical forcing is represented by metrics that quantify the troposphere-stratosphere wave coupling. Composite forecasts from the ensemble members that produce the top and bottom quartile of metrics in a forecast will serve as a means to investigate forecast biases. The model biases associated with the physical and dynamical metrics will be quantified in terms of their manifestations in important flow features in the troposphere and stratosphere. Biases in the location and amplitude of the tropospheric precursor blocking as well as biases in the location, shape and strength of the stratospheric vortex best-fit ellipse will be calculated for each composite group. The main scientific objective of this work is to evaluate whether a categorical assessment of errors and uncertainty associated with physical forcing mechanisms lead to an improved understanding of sources of biases in the stratospheric forecasts initialized prior to a stratospheric regime changes. The results of the proposed work will be available online for immediate use by the forecasting community as they become available.”