When tiny convective spread affects a midlatitude jet: spread sequence

TL;DR

This study investigates how tiny convective uncertainty can propagate to midlatitude jet variability using an ICON ensemble with near-intrinsic perturbations. It tests the applicability of the three-stage spread-growth framework (convective, divergent, rotational) to individual convective systems near the jet and reveals substantial case-to-case variability, with longwave radiative tendencies playing a prominent, sometimes dominant role near convective anvils. The Baltic Sea system exhibits a clear sequence with downstream Rossby dispersion, while the Alps and Germany cases show partial or localized manifestations influenced by jet geometry and wave phase. By pairing ensemble sensitivity analysis with PV-tendency diagnostics, the work provides mechanistic attribution of downstream jet spread to convective variability, offering insights into intrinsic predictability limits and the role of longwave processes in spread growth.

Abstract

We investigate the evolution of spread over three days in a numerical ensemble experiment starting from tiny initial condition uncertainty. We simulate a real event during which three mesoscale convective systems occur in close proximity to the midlatitude jet. The spread evolution is compared with an existing conceptual three-stage model. Each system follows the first stage, characterised by development of convective variability. Nevertheless, we find significant variation among the systems in their propensity to interact with the jet stream, which characterises conceptual stage 2. One exemplary convective system follows the conceptual evolution of Baumgart et al., i.e., convective uncertainty initially projects onto the jet by upper-tropospheric outflow, which further amplifies spread through nonlinear growth as it propagates downstream. Rossby-like dispersion in the downstream spread is strongly associated with the convective variability. In contrast, for another convective system, convective variability projects onto the local anticyclonic flow aloft. Subsequently, this anticyclonic perturbation hardly (if at all) projects convective uncertainty onto the particularly straight jet stream, which truncates the conceptual evolution. For the third system, negligible fingerprints of second and third stages are identified. Alongside convective heating, longwave radiation jointly dominates the spread evolution near the convective systems (as opposed to earlier studies). Longwave-radiative tendencies of convective anvils outlive the accompanied heating tendencies and extend spatially. Furthermore, we link convective variability of the exemplary system directly to longwave-radiative tendencies. Therefore, longwave radiation appears to contributes substantially to stages 1 and 2 here. Finally, we identify flow dependence of the impact of convection on the jet. (Truncated abstract)

Figures (12)

• Figure 1: Accumulated precipitation over past 6 h for 12 (top left), 24 (top right), 36 (bottom left) and 48 h (bottom right) forecasts of ICON (using the ECMWF IFS operational run's initial conditions). The PV contour of 2PVU at 250 hPa is contoured in black (coarse grained to 0.75 degrees). Labels A, B and G mark the Alps convective system, the Baltic Sea convective system and the Germany convective system.
• Figure 2: Mean precipitation rates (fill) and spread in mean precipitation rate (isolines) at approximately the time of maximum intensity for (left) the Alps System and (right) the Baltic Sea system. Spread is contoured at grid values of 0 (i.e. precipitation occurs in some ensemble member(s)), 0.01, 0.05, 0.25 and 1 mm/h. The masks (black outline) defining the extent of each convective system is also visualised at this (approximate) time of maximum intensity.
• Figure 3: Evolution of 2 and 4 PVU contours at 330 K (ensemble mean) with PV-spread at 327.5-335 K superposed (colour fill) at 3 hour intervals (initially) and 6 hour intervals (beyond 12 h).
• Figure 4: Three hour mean of the potential enstrophy tendency for two different processes at $\theta$ of 327.5-335 K, $t=6$ h: heating by the deep convection parameterisation (a), the heating by longwave radiation parameterisation (b), the differential PV-advection resulting from differences in divergent winds (c) and similarly for the rotational winds (d). Red colours indicate spread growth, whereas blue colors indicate spread reduction for each term. For this level, the dynamical tropopause (2PVU) and 4 PVU contour are also shown in dark grey/black.
• Figure 5: Potential vorticity covariance at 250 hPa within the ensemble associated with enhanced mass divergence in the upper troposphere over the convective system in the Alps between 2 h and 5 h simulation time. Top: 12 h; bottom: 24 h. Isolines indicate regions of enhanced $\sigma_{PV}$ within the ensemble (isolines at 0.06 PVU intervals between 0.06 and 0.30 PVU).