Sting jets in intense winter North Atlantic cyclones Suzanne Gray, Oscar Martínez-Alvarado, Laura Baker and Agnieszka Mega (Univ. of Reading), Peter Clark (Univ. Of Surrey), and Jennifer Catto (Univ. Monash) September 2012

Introduction The strongest winds in extratropical cyclones are normally associated with low-level jets along the warm and cold fronts. Some cyclones form an additional region of very strong surface winds and gusts, known as a sting jet. The sting jet is a transient mesoscale air stream that descends from the cloud head into the frontal fracture region.

Example cases: observations Mesoanalysis of the Great October storm (16 October1987) Dropsondes through cyclone Friedhelm (8 December 2011) Upper-level jet SJ SJ? Figure 9. (a) Raw data and (b) mesoanalysis of the distribution of peak surface wind gusts (m s-1) at 0130 UTC 16 October 1987, derived from anemograph traces at 20 stations. Station locations are indicated by the circles in (a) where half-hourly peak-gust measurements for each station have been converted to a series of measurements at equivalent spatial locations based on the system velocity at this time of 23 m s-1 from 217B . (The three stations enclosed by two circles indicate the stations whose anemographs are shown in Fig. 10.) Isotachs in (b) are at 5 m s-1 intervals. Some of the fine structure in (b) is based on observations extracted at finer temporal resolution than plotted in (a). Isotachs shown dashed deviate from the data in (a) to take into account French data in Fig. 8. The anomalously high underlined values at one of the stations (Portland RNAS) are discussed in the text. Four regions of strong winds, referred to in the text, are labelled A, B, C and D. The thick line in (b) represents the warm and bent-back front and the secondary cold front CF2 as derived from thermograph records. The dashed line in (b) shows the extent of dry air (RH<80%) as derived from hygrograph records. The peak-gust analysis in Fig. 9(b) draws attention to four regions of strong winds—A, B, C and D—which, with the exception of D, were all associated with the dry air: Region A: a region containing localized areas (individually labelled A in Fig. 9(b)) with gusts up to about 35 m s-1. It is shown in section 3(c)(i) that this region was associated with cumulonimbus clouds just ahead of CF2. Region B: a region containing localized areas (individually labelled B) with gusts up to about 40 m s-1. This region was associated with only shallow non-precipitating cloud in the dry slot behind CF2. Region C: a larger region of even stronger winds, with gusts from 40 to 50 m s-1. It is shown in section 3(c)(ii) that this region was located in the dry slot in the region of very shallow or no cloud close to the tip of the hooked cloud head. Region D: an extension of Region C, behind the bent-back front, but associated with moister air under the tip of the cloud head. Regions C and D are associated with the sting jet Dropsonde data for sondes released between 1130 and 1234 UTC Cold front Cyclone centre Browning (2004) The DIAMET team

Example cases: modelling SJ? System-relative wind speed at 850 hPa is shaded (25-30, 30-35, and 35-50 ms-1). Cloud head (80% RH at 500 hPa). Friedhelm – separate SJ or SJ enhancing CCB at this time – quite wrapped up. The Great Storm: 03 UTC 16 October 1987 Cyclone Friedhelm: 12UTC 8 December 2011

Theory: conceptual model The Sting jet occurs during the process of frontal fracture in Shapiro-Keyser cyclones Figure 17. Conceptual model of the near-surface flows in an extratropical cyclone. (a) Early stage of frontal wave cyclone development. L denotes low-pressure centre with direction of movement shown by thin arrow. Grey arrows show the system-relative low-level jets; WJ is the warm-conveyor-belt jet (WCB in text) and CJ the cold conveyor- belt jet (CCB in text). (b) Frontal fracture phase, when the sting jet (SJ) first appears at the surface. (c) As the cloud head wraps round further the SJ region extends. (d) Eventually the distinct SJ disappears and the dominant low-level wind in this region is due to the CJ. Positions of cross-sections shown in Fig. 18 are marked in (b). Frontal analysis based on the Shapiro and Keyser (1990) lifecycle model in which the surface cold front fractures from the surface warm front so as to leave behind a bent-back warm front (part of which is analysed here as a cold front. Secondary WCB omitted for clarity. Clark et al. (2005)

Theory: conceptual model The sting jet is distinct from the dry intrusion and cold conveyor belt jet and may be composed of multiple individual descending airstreams. Clark et al. (2005) Browning (2004)

Mechanisms: 1. conditional symmetric instability CSI is the due to the combination of inertial and conditional instability (gravitational instability) for air parcels displaced along a slantwise path. It will only be released if the atmosphere is inertially stable to horizontal displacements and conditionally stable to vertical displacements. For a region to be susceptible to the release of CSI, the relative slope of the saturated equivalent potential temperature ( ) surfaces must be steeper than the geostrophic absolute momentum (Mg) surfaces Morcrette (2004)

Mechanisms: 2. evaporative cooling Enhancement of surface winds by Intensifying the slantwise circulations and so amplifying the latent heat sources and sinks on the mesoscale. Reducing the static stability in the dry slot and/or closer to the cloud head so leading to turbulent momentum transfer or upright convection. (Browning 1994) (i) Slantwise circulations in the cloud head are associated with evaporative cooling in the descending parts, particularly where precipitation falls into them from the overlying ascending flows. The cooling is strong just to the south and south-east of the cyclone near the tip of the cloud head. This leads to the intensification of potential vorticity sheets and perhaps also to the generation or intensification of moist symmetric instability (see section 4(b)). As a result, the evaporative cooling acts to intensify the transverse slantwise circulation which in turn amplifies the latent-heat sources and sinks on the mesoscale. Evaporative cooling is most effective where there is sublimation of ice, as shown in the high-resolution modelling study of FASTEX cyclones by Forbes and Clark (2003). However, evaporation of cloud water may be important in enhancing the slantwise circulations closer to the surface. (ii) The evaporative cooling of the air leaving the tip of the cloud head reduces the static stability in the dry slot, which is a region characterized by potential instability created as described in section 2(c). It thus contributes to the triggering of the upright convection in region A where the associated downdraughts bring high-momentum air down to the surface. Closer to the tip of the cloud head, in regions B and C, the static stability, although not reduced sufficiently to permit the development of convective showers, may be reduced enough to produce very low Richardson numbers and strong turbulence capable of mixing the strong winds in the descending branches of the slantwise mesoscale circulations down to the ground. Figure 12. Pressure increase between the modelled minimum trajectory pressure and first reaching a pressure greater than 800 hPa plotted against potential-temperature change over the same period, for those trajectories which exhibit a change in wet-bulb potential temperature of less than 1 K over this time. Fig. 12 is consistent with the idea that the region of strong winds is associated with at least some air that has descended substantially (from as high as 500 hPa in the extreme case). The strong adiabatic warming associated with this descent may be offset to some extent by potential cooling (due to evaporation) and potential warming (due to mixing).There is some evidence that cooling may be by as much as 8 K in the extreme, but in parts of the SJ may be small or absent. Clark et al. (2005)

A first regional climatology of sting jet storms Published as Sting jets in intense winter North-Atlantic windstorms (2012), O. Martínez-Alvarado, S. L. Gray, J. L. Catto, and P. A. Clark, Environ. Res. Lett, 7, 024014.

to develop a regional climatology of sting jets and sting jet cyclones Aim to develop a regional climatology of sting jets and sting jet cyclones Problem observational datasets do not provide sufficient temporal resolution over the oceans to allow exhaustive identification of sting jets, and sting jets are not resolved in multi-year re-analysis datasets Analogous to convection. Global models cannot resolve convection but CAPE is still present and can be used as a precursor for where convection will occur.

Solution We have developed a diagnostic to diagnose sting jet precursor regions in re-analysis datasets (Martinez-Alvarado et al. 2011). The diagnostic detects downdraught CSI (via a metric called DSCAPE) in the moist frontal fracture zone. DSCAPE is present in cyclones that have sting jets but not present in other, equally intense, cyclones that do not have sting jets (Gray et al., 2011). Low resolution does not inhibit the accumulation of CSI, just its release to generate a sting jet. Analogy with upright convection Criteria are DSCAPE > 200 Jkg-1 RH>80% |grad theta_w| > 10^-5 Km-1 frontal zone v.Grad theta_w > 10^-4 Ks-1 cold front Where the latter two mean that the precursor could be near a frontal fracture region. Precursor regions not allowed to lie entirely in WCB region.

Methods The 100 most intense winter cyclones in ERA-Interim data (DJF 1989-2009) over the North Atlantic are identified by applying an objective feature tracking algorithm (Hodges, 1994). The sting jet precursor diagnostic is applied to each cyclone over 3 days straddling the time of maximum intensity. Verification of a subset of cases (15) is performed using high resolution (sting-jet resolving) model simulations with the Met Office Unified Model. Sting jets are identified in these simulations using established trajectory methods.

Location of precursor gridpoints Shaded area represents the average precursor region. Precursor location consistent with the origin region of SJs in previous case studies. Precursor regions not allowed to lie entirely in WCB region. Cyclone elements in composite cyclone. Cyclones rotated to a westerly travel direction Sting-jet precursor gridpoints within identified precursor regions.

Tracks 32 storms with sting jet precursor Number of storms with SJ precursors is dependent on the minimum size specified for the SJ precursor region. For minimum size thresholds yielding significant verification results between 23 and 32 storms have SJ precursors. Red dots are where SJ precursors are identified More about distribution later but note all storms with SJ precursors originated south of 50 degrees N Some storms had precursors at multiple times (31%) 32 storms with sting jet precursor 68 storms without sting jet precursor

Sting-jet precursor timing distribution Strong tendency for SJ precursors to occur in the 30h prior to the occurrence of max intensity. Consistent with conceptual model in which SJ occur in stages II and III of the SK evolution. Distribution of sting jet precursors relative to their time of maximum intensity.

Sting jet trajectories Saturated moist potential vorticty Pressure Relative humidity One ensemble of trajectories for each cyclone. Bottom row is false negative case. Ensemble mean descent rate 0.4-0.9 hPa-1; 0.5, 0.8, 1.3 hPa-1 for Gudrun, Anna and the Great storm. Descent rates achieved for 5 hrs in true positive case but just 2 hrs for false negative Sting jet trajectories dried as they descended from the midtroposphere towards the top of the boundary layer (although the depth of descent was less than in the October 1987 storm). All sting jets have at least some trajectories with –ve MPV* but static and inertial stability. Along a moist descending trajectory this implies that CSI is being released.

Contingency table for 15 cases SJ observed SJ not observed SJ Predicted 5 2 7 SJ not predicted 1 8 6 9 15 p-value = 0.035 (using Fisher’s exact test)

Environmental precursors of sting jet storms Agnieszka Mega – MSc dissertation, 2012 Results shown for 31 storms with sting jet precursors and 48 storms without sting jet precursors (so excluding storms where CSI was present but in the wrong place to satisfy the precursor criteria).

Cyclone tracks: origin

Cyclone tracks: maximum intensity

Low-level moisture Mean 276K Mean 271K Mean 290K Mean 281K 850 hPa Ɵe in the cyclone core and in the WCB within 10o of the centre

Upper-level jet crossing 97% of sting jet storms cross the upper-level jet during their intensification period but only 44% of non-sting jet storms do. Upper-level (250 hPa) windspeed (ms-1) and surface pressure analysis (hPa) for cyclone Friedhelm: 1800 UTC 7 December – 0600 UTC 8 December

Conclusions The first regional climatology of sting jets has been developed. 32 out of the 100 most intense winter cyclones during the last two decades had sting jet precursor regions. Sting-jet resolving simulations of 15 cyclones demonstrates the reliability of the sting jet precursor diagnostic. At least some sting jet trajectories are releasing conditional symmetric instability. There is little evidence for evaporative cooling. Sting jet storms have warmer cores and WBCs than non-sting jet storms (associated with their more southerly tracks). Sting jet storms are more likely to cross the upper-level jet during their intensification.

Frequency distribution No obvious time trend in the SJ or non SJ storms Time series for frequency of the 100 most intense cyclones (grey) and those with sting jet precursor regions (black).

Example of DSCAPE structure ERA-Interim MetUM - global MetUM - LAM DSCAPE: Light grey 200-500 Jkg-1 Mid grey 500-800 Jkg-1 Dark grey >800 Jkg-1 Stippled regions: cloud at 550 hPa Contours: 825 hPa-qw isolines Start of trajectories - LAM