An Object-Based Forecast Verification Tool for Synoptic-Scale Rossby Waveguides

Abstract

An accurate representation of synoptic-scale Rossby waves in numerical weather forecast models is very important as these waves are closely linked to weather formation at the surface. Enhanced potential vorticity (PV) gradients at the tropopause levels act as waveguides for synoptic-scale Rossby waves, so spatial errors in the waveguides imply errors in the amplification and propagation of Rossby waves. This paper focuses on evaluating the forecast representation of these waveguides and presents an object-based forecast verification tool. In both forecast and the verification data, Rossby waveguide objects are defined based on enhanced PV gradient fields on isentropic surfaces. The tool automatically pairs the complex objects, compares their properties, and assesses the number of objects without a matching partner in either the forecast or the reanalysis. In the last step, error measures are calculated for the area and the location of the objects. As proofof-concept application of the method for the year 2008, five lead times of the Integrated Forecast System (IFS) from the ECMWF are compared with the ECMWF reanalysis dataset. The majority of the waveguide objects are found to be in the correct position, and there are no systematic positional errors; however, the forecast objects and hence the areas of enhanced PV gradients are smaller.

An atmospheric dynamics perspective on the amplification and propagation of forecast error in numerical weather prediction models: A case study

“…WCB air parcels in the 312.5-317.5 K isentropic layer indicate regions where WCB activity (particularly outflow) might impact the wave guide (compare with Grams et al, 2011;Joos and Wernli, 2012). The PV-Θ perspective thus helps in focusing on the error evolution in the upper-level Rossby wave pattern, which is most relevant for the large-scale midlatitude flow (Davies and Didone, 2013;Giannakaki and Martius, 2016). We discuss this with the wave guide in the analysis and ensemble mean (black and red 2 PVU contours in Figure 4, respectively), analysis WCB intersection points (black crosses, Figure 4), and ensemble standard deviation of PV at the 315 K isentropic surface (shading, Figure 4).…”

“…They demonstrate further that error in the representation of potential temperature and PV evolution in the WCB can be responsible for error in the Rossby wave pattern. Several studies showed that, once such a forecast error for the large-scale midlatitude flow is established, it often grows and propagates along the upper-tropospheric Rossby wave guide like a dry-dynamic baroclinic disturbance (for example, Dirren et al, 2003;Hakim, 2005;Davies and Didone, 2013;Rodwell et al, 2013;Giannakaki and Martius, 2016).…”

Representation of Synoptic‐Scale Rossby Wave Packets and Blocking in the S2S Prediction Project Database

“…However, state‐of‐the‐art global NWP models struggle to adequately represent the structure of the RWPs. This is due to a systematic reduction in Rossby wave amplitude (e.g., Gray et al, ; Martínez‐Alvarado et al, ) and systematic errors in the area and strength of the waveguide along which they propagate (Giannakaki & Martius, ). Although the latter studies focused on the verification of the structure of RWPs in NWP models, this study verifies the duration, propagation, onset, and decay of RWPs, as well as their relation to atmospheric blocking in a set of subseasonal forecasting systems.…”

Atmospheric blocking and upper‐level Rossby‐wave forecast skill dependence on model configuration

“…Harvey et al (2016) have also shown that the reduction of the isentropic PV gradient can be linked to the slower eastward propagation of Rossby waves and to a reduction in Rossby-wave amplitude. The correspondence between errors in the structure of Rossby waves and those in the structure of Rossby waveguides (Giannakaki and Martius, 2016), as well as the relationship between Rossby waves and warm conveyor belts (Martínez-Alvarado et al, 2016), suggests that these errors have a direct impact on the synoptic variability of models. However, their impact on the representation of stationary waves is not known.…”

“…Through analysis of seven winters from 2006/07–2012/13 in three EPSs, Gray et al () found systematic forecast errors in the structure of Rossby waves in terms of a reduction in Rossby‐wave amplitude and tropopause sharpness with lead time. In agreement with those results, Giannakaki and Martius () found systematic errors in the area and strength of Rossby waveguides, defined as long and narrow bands of strong isentropic PV gradient, in ECMWF forecasts as compared to reanalyses. There is a causal relationship between diabatic processes and ridge development (Davis et al , ).…”