VOLUME 128 MONTHLY WEATHER REVIEW AUGUST 2000 Dynamics of a Catalina Eddy Revealed by Numerical Simulation CHRISTOPHER DAVIS AND SIMON LOW-NAM National Center for Atmospheric Research,* Boulder, Colorado CLIFFORD MASS University of Washington, Seattle, Washington (Manuscript received 27 April 1999, in ®nal form 20 December 1999) ABSTRACT Through numerical simulations with the Pennsylvania State University±NCAR Mesoscale Model the dynamics of a Catalina Eddy event that formed during the period 26±30 June 1988 off the coast of southern California is examined. A strengthening and veering of the low-level synoptic-scale winds, from climatological northwes- terlies to northerlies, results in a more pronounced effect of the coastal orography around the California bight region. In particular, relative vorticity formed by ¯ow over the coastal terrain remains offshore. Prior to the formation of a quasi-steady eddy within the bight, the northerlies are strong enough to advect anomalously high vorticity out of the region. The formation of a mature Catalina eddy relies on a rapid deceleration of the synoptic- scale northerlies on 30 June, such that vorticity, once formed, remains in the bight. The eddy is also strongly modulated by the diurnal cycle. Northwesterly ¯ow around 500 m above mean sea level impinging on the mountains north of the bight is enhanced during the late afternoon, mainly as a response to the land±sea thermal contrast. This strengthened ¯ow overlaps temporally with a minimum in low-level strati®cation due to surface heating. The result is air characterized by a relatively high Froude number, which traverses over the coastal mountains and strongly depresses the marine layer over the bight. The depression in the marine layer results in a warm anomaly and cyclonic circulation. Later at night, the incident northwesterlies weaken and the ¯ow becomes more stable, resulting in ¯ow around, rather than over, the coastal mountains. This regime transition yields a wake with little depression of the marine layer and an absence of vorticity generation on the scale of the bight region. Given strong ambient ¯ow, vorticity generated in the evening is swept southward past the bight the following day, but with weak ambient ¯ow, the eddy persists in the bight during daytime, weakening slowly. 1. Motivation and into the Los Angeles basin, an unmistakable sig- nature in visible satellite imagery (Rosenthal 1968). Pre- Disruptions in the climatological ¯ow along the coast- diction of the onset of overcast conditions and deep- line of southern California associated with the Catalina ening of the marine layer on the California coast rep- eddy have long been known to be important for air resents a signi®cant challenge to local forecasters during quality in the Los Angeles basin, as they are associated the otherwise benign warm season. with pronounced increases in the depth of the marine The climatology of eddy events presented in Mass layer (Wakimoto 1987; Thompson et al. 1997). The Cat- and Albright (1989, hereafter MA) demonstrates the im- alina eddy is a quasistationary mesoscale vortex with a portance of changes on the synoptic scale. Following horizontal scale of roughly 100 km, extending through the passage of a low pressure trough into the Paci®c a layer from the surface to between 1 and 2 km above Northwest, the winds near 850 hPa become stronger and mean sea level (MSL). Eddy events are often accom- veer to a northerly or northeasterly direction as the trail- panied by stratus spreading northward along the coast ing anticyclone moves north of the bight region. Cy- clonic circulation is produced over the bight in the lee of the San Rafael Mountains (Fig. 1), and this circulation * The National Center for Atmospheric Research is sponsored by appears to expand slowly with time. According to MA, the National Science Foundation. lee troughing reverses the alongshore pressure gradient, forcing coastally trapped southerly ¯ow from roughly the U.S.±Mexico border to the Los Angeles basin. Ob- Corresponding author address: Christopher A. Davis, National Center for Atmospheric Research, P.O.Box 3000, Boulder, CO 80307- servational case studies by Bosart (1983), Wakimoto 3000. (1987), and MA, as well as modeling studies by Ueyoshi E-mail: [email protected] and Roads (1993, hereafter UR), Ulrickson et al. (1995), q 2000 American Meteorological Society 2885 Unauthenticated | Downloaded 09/25/21 12:38 PM UTC 2886 MONTHLY WEATHER REVIEW VOLUME 128 FIG. 1. Locations of the three domains used in our simulations. Inset ®gure depicts the terrain on domain 3 (400-m contour interval) and the locations of various cross sections shown in the paper. Gray box indicates area of averaging to produce time±height cross section in Fig. 7. Gray dot marks point at which time±height cross sections is produced in Fig. 8. and Thompson et al. (1997), all describe cases that gen- is, ``What generates the lee troughing and cyclonic vor- erally ®t the climatology. ticity in the ®rst place?'' It may appear from the literature that the Catalina Although it is apparent that the eddy results from ¯ow eddy is a well-studied phenomenon, but there are some over/around complex terrain, ¯ows in the region in important issues regarding its dynamics that remain. which the eddy forms differs from the uniform ¯ows Most of the previous studies (e.g., Clark 1994) have that have been studied extensively, and about which emphasized the formation of coastally trapped south- most of our knowledge of mesoscale lee vortices and erlies as a de®ning feature of the eddy. However, these lee troughs is based (cf. Smith 1979; Boyer et al. 1987; southerlies depend in part on lee troughing and asso- Smolarkiewicz and Rotunno 1989). For instance, the ciated generation of mesoscale cyclonic vorticity in the strong baroclinicity implied by the contrast between a bight region. Therefore, perhaps the most basic question warm continent and a cool ocean implies strong low- Unauthenticated | Downloaded 09/25/21 12:38 PM UTC AUGUST 2000 DAVIS ET AL. 2887 level shear. Second, the presence of a marine inversion passing through the Paci®c Northwest, followed by ridg- layer yields a complicated pro®le of static stability. ing in the lower troposphere. As the anticyclone moves Third, the diurnal cycle amplitude over the continent is over the West Coast region, enhanced northerly ¯ow is maximum in the warm season, when eddies are most noted, particularly at Vandenberg Air Force Base (see frequent. The baroclinicity, shear, and stability vary di- Fig. 25 of MA). Just after this time a mature cyclonic urnally implying that the terrain-induced response will eddy is observed in the bight region of southern Cali- vary similarly. fornia. The formation of a quasistationary eddy seems The Catalina eddy is strongly modulated by the di- to coincide with an overall weakening of the synoptic- urnal cycle. Studies of different cases (Bosart 1983; scale northerly ¯ow as the anticyclone moves to the MA; Clark 1994; Thompson et al. 1997) all show an northeast of the region on 30 June. eddy strengthening in the early evening, maturing over- Mesoscale surface analyses (refer to Fig. 22 in MA) night, then weakening the following day. This has been depict an event that requires about four days to unfold. interpreted as an effect of the sea breeze, which disrupts Each day, on average, the eddy appears slightly stronger the eddy circulation during the day. However, UR noted and more extensive than the previous day. However, from their simulation of the 26±30 June 1988 eddy case there is a diurnal variation in the intensity, with sys- that vorticity through a layer roughly 1±2 km deep cen- tematically stronger circulation at night, at least until tered at 950 mb over the bight was systematically larger 30 June when a closed cyclonic circulation persists dur- at night. The diurnal modulation of vorticity aloft is not ing the day as well. During most of this 4-day period, likely the result of the sea-breeze circulation because the eddy appears as a trough elongated northwest to its depth and its horizontal scale (roughly 100 km) great- southeast. At night, southeasterly ¯ow prevails along ly exceed that of the classical sea-breeze circulation. the coastline and somewhat offshore. The ¯ow near and The mechanism responsible for the variation in vorticity slightly inland from the coast appears related to oro- aloft has yet to be determined. graphic blocking by the coastal mountains south of the Because of the relative sparseness of observations, Los Angeles basin. The southeasterlies farther offshore much of the present study will encompass a diagnosis are more nearly geostrophic. To the west of the trough, of the eddy evolution produced by a numerical simu- strong northwesterly ¯ow (10±15 m s21) prevails until lation. Previous modeling studies of this case with high- 30 June. Typical values of relative vorticity associated ly simpli®ed physics and relatively coarse resolution with the trough and subsequent more circular eddy are models have shown reasonable success in simulating on the order of 2±3 3 1024 s21. the general features of the Catalina eddy. Our approach The vertical structure of the eddy is largely unknown, will be one of case study, the case being the eddy of owing to the poor sounding coverage offshore. Based 26±30 June 1988, studied by MA, Ueyoshi and Roads on time±height sections of vorticity from their numerical (1993), and Ulrickson et al. (1995). This choice is eco- simulations, which selectively ®ltered out the larger- nomical from the point of presentation, since much of scale evolution, UR revealed the eddy to be about 2 km the observations have been documented in MA and will deep. Thus, the eddy itself appears to be rather shallow, hence only be brie¯y summarized here.