VOLUME 18 WEATHER AND FORECASTING DECEMBER 2003 Cold Pools and MCS Propagation: Forecasting the Motion of Downwind-Developing MCSs STEPHEN F. C ORFIDI NOAA/NWS/NCEP/Storm Prediction Center, Norman, Oklahoma (Manuscript received 13 March 2003, in ®nal form 20 May 2003) ABSTRACT The primary factors that affect the direction of propagation and overall movement of surface-based mesoscale convective systems (MCSs) are discussed. It is shown that although propagation is indeed related to the strength and direction of the low-level jet as previous studies have shown, it is more speci®cally dependent upon the degree of cold-pool-relative ¯ow and to the distribution of conditional instability present along a system's gust front. An updated technique that may be used to forecast the short-term (3±6 h) motion of MCS centroids based on these concepts is introduced. The procedure builds on the long-established observation that MCS motion is a function of 1) the advection of existing cells by the mean wind and 2) the propagation of new convection relative to existing storms. Observed wind and thermodynamic data, in conjunction with anticipated cold-pool motion and orientation, are used to assess the speed and direction of cell propagation, that is, whether propagation will be upwind, downwind, or some combination of the two. The technique ultimately yields an estimate of overall system movement and has application regardless of scale, season, or synoptic regime. 1. Introduction 1980). They were among the ®rst to recognize that though no true ``steering level'' exists for MCC motion, Thunderstorms are frequently organized in lines or such systems typically move approximately parallel to clusters known as mesoscale convective systems the contours of the 1000±500-hPa thickness. They also (MCSs). The term MCS is generally reserved for en- noted that although most convective systems move sembles of storms that satisfy certain spatial or temporal criteria (see, e.g., Houze 1993, p. 334; Parker and John- downshear along the contours, others inexplicably move son 2000). In a less restrictive sense, however, any me- upshear. The speed of MCC motion was found to be soalpha- or mesobeta-scale (Orlanski 1975) area of modulated by the location of maximum low-level mois- moist convection might be considered an MCS (Ray ture convergence relative to existing convection. 1990). Newton and Katz (1958) and Chappell (1986), among Because MCSs produce a disproportionate share of others, showed that the motion of a convective system signi®cant convective weather (high winds, ¯ash ¯ood- can be thought of as being the vector sum of 1) an ing, etc.) and because their evolution is often not pre- advective component approximated by the direction and dicted well by operational numerical guidance, fore- magnitude of the mean cloud-layer wind and 2) a prop- casting MCS motion is of considerable importance to agation component governed by the rate and location operational meteorologists. Forecasts of MCS motion of new cell formation relative to existing convection. are dependent upon anticipation of the predominant Building on this idea, and extending the work of Merritt propagational mode or modes likely during an event. In and Fritsch, Cor®di et al. (1996) showed that the prop- particular, it is important to distinguish between MCS agation component is, in many cases, directly propor- environments conducive to upwind propagation and tional but opposite in direction to the low-level jet. This those that exhibit downwind storm development and ®nding is somewhat surprising given that MCS prop- sometimes evolve into derecho-producing squall lines. agation can be in¯uenced by a myriad of factors such Attempts to forecast MCS movement have met with as the distribution of convective available potential en- mixed results. Merritt and Fritsch (1984) examined the ergy (CAPE), convective inhibition, gravity waves, out- motion of more than 100 MCSs, most of which were ¯ow boundaries, and orographic effects. mesoscale convective complexes (MCCs; Maddox This paper discusses MCS motion, with emphasis on those factors related to a system's cold pool that most in¯uence cell propagation and, ultimately, overall sys- Corresponding author address: Stephen F. Cor®di, Storm Predic- tion Center, 1313 Halley Circle, Norman, OK 73069. tem movement. Based on this presentation, a vector- E-mail: stephen.cor®[email protected] based forecast technique is developed for predicting the 997 998 WEATHER AND FORECASTING VOLUME 18 vection and propagation, can affect MCS movement. Therefore, motion estimates must be updated frequently when the wind ®eld exhibits signi®cant spatial or tem- poral variability. Second, there is no accounting for the in¯uence of terrain on convective development and low- level ¯ow. As a result, the concept is more dif®cult to FIG. 1. Schematic of the original vector technique, with MCS core apply in cases of orographically forced convection (e.g., motion (thick dotted arrow) expressed as the vector sum of 1) ad- vection of cells by the mean cloud-layer wind (arrow pointing to Pontrelli et al. 1999). upper right) and 2) cell propagation directed into the low-level jet A more serious shortcoming of the original vector (arrow pointing to bottom of page). MCS centroid is depicted by the approach follows from its assumption that new cell de- cross symbol (after Cor®di et al. 1996). velopment and, therefore, system propagation always occur in the direction opposite that of the low-level jet, motion of MCSs characterized by downwind propaga- or, more generally, the low-level ¯ow. To be sure, many tion. warm-season MCSs over the central United States in- deed do exhibit propagation in that direction (see, e.g., Moore et al. 1993; and Junker et al. 1999) at a rate 2. Background approximated by the speed of the jet.2 It is clear, how- a. The original vector technique ever, that this is not always the case. For example, de- recho-producing squall lines often move at a substantial Cor®di et al. (1996) developed a simple technique to angle to the low-level ¯ow, especially during their ini- predict the short-term (3±6 h) motion of the mesobeta- tiation (Johns et al. 1990). Radar data reveal that, al- scale cores or ``centroids'' of MCSs using the low-level though propagation is largely responsible for the ob- jet to estimate the direction and rate of storm propa- served motion of these systems, new cell development gation. Forecast centroid motion is taken to be the sum is not necessarily directed into the low-level ¯ow but of 1) a vector that represents cell advection by the mean rather occurs on the leading (downwind) edge of the cloud-layer wind (with ``cloud layer'' taken to be the system cold pool. For this reason, to be more universally 850±300-hPa layer)1 and 2) a vector that represents applicable, the vector concept must be modi®ed to ac- storm propagation, that is, new cell development, equal count for the presence of cold pools and the potential in magnitude but directed opposite to the low-level jet for propagation away from the low-level ¯ow. (Fig. 1). In practice, the 850-hPa wind is used to ap- proximate the low-level jet, although it is recognized that this approach may not identify the true jet in all b. Cold-pool and gust-front motion cases. In the absence of a distinct low-level speed max- One of the more distinctive features of a well-orga- imum in the vertical direction, the strongest wind in the nized MCS is the cold pool that develops at lower levels lowest 5000 ft (1.5 km) is generally used, in accordance beneath or just behind the strongest convection. Cold with Bonner (1968). pools represent the collective out¯ow of individual con- The vector technique is applicable in any kind of vective cells and the negative buoyancy of parcels with- environmental wind regime and requires knowledge of in or beneath the convection. Sublimation and/or melt- only the 850-hPa and mean cloud-layer winds. The pro- ing and evaporation of precipitation falling through un- cedure is especially useful in identifying those kinematic saturated air, precipitation drag, and vertical perturba- situations conducive to the development of quasi-sta- tion pressure gradients are all factors that may enhance tionary and ``back building'' MCSs (Bluestein and Jain downdraft development and cold-pool strength. 1985). Quasi-stationary systems arise when the wind The periphery of a cold pool, that is, the gust front pro®le is unidirectional and cell advection is exactly or out¯ow boundary, is marked by low-level conver- offset by cell propagation. Back building occurs under gence and ascent (Purdom 1973; Charba 1974; Goff similar conditions but when propagation exceeds ad- 1976). As a result, gust fronts are often the site of new vection, resulting in overall upwind motion. Identi®- cell development. Such activity typically is not distrib- cation of such events is important because they fre- uted evenly along such boundaries. Instead, storm ini- quently are associated with excessive rainfall (Chappell tiation tends to occur in discrete zones, within which 1986). kinematic and/or thermodynamic factors are most fa- The original vector concept, while useful, is never- vorable for development. Observation suggests that new theless subject to several limitations. First, the scheme cell development occurs most readily where the ambient, does not account for spatial and temporal changes in the environmental wind that, in altering both cell ad- 2 The original dataset used by Cor®di et al. (1996) was composed primarily of nocturnal MCCs that were associated with well-de®ned 1 The speed and direction of the mean cloud-layer wind are cal- low-level jets east of the Rockies.