The Sundowner Winds of Santa Barbara, California (2024)

1. Introduction

Numerous observational and modeling studies have shown that mountain barriers can exert a significant influence on atmospheric flows. Various modalities in this regard exist; these include drainage flows, blocking or damming phenomena, flow funneling, induction of cyclonic circulations (e.g., Denver cyclone, Catalina eddy), and mountain wave development and associated downslope wind events. The last of these appears to be of greatest relevance to the Sundowner events discussed in the present paper.

Mountain waves are simply gravity waves forced by mountains. As noted by Durran (1990) and others, large-amplitude mountain waves can be associated with strong surface winds that blow down the lee slope of the mountain (range)—with wind gusts in excess of 50 m s⁻¹ (97 kt) in extreme events. Significant warming can also be produced on the lee side of the mountains by the downslope winds.

Although mountain wave development and associated downslope wind events have been well studied for a number of large midlatitude mountain ranges (e.g., the Front Range of the Rocky Mountains, the Sierra Nevada of California), this is not necessarily the case when these phenomena are associated with orographic features of much smaller length scale. Effects can nonetheless be dramatic and can have significant impact on the local environment. The induced change from more typical meteorological conditions can be particularly striking in those situations where the mountain range adjoins a coastline; conceptual complexity also increases as influences of the large-scale land–ocean thermal contrast and variation in boundary layer structure need to be considered, as well as the shape of the coastline itself.

Perhaps the most dramatic examples of coastal downslope wind events within the contiguous United States are those that periodically occur along a short segment of the southern California coast in the vicinity of Santa Barbara. The term “Sundowner” has long been used to refer to these downslope winds because of their typical onset in the late afternoon or early evening. Associated with a rapid rise in temperature and decrease in relative humidity, the most intense events of this sort can bring extremely warm temperatures and/or damaging winds. The region susceptible to these Sundowner winds is characterized by a mesoscale topography unique within the conterminous United States (Fig. 1): a coastline that is oriented east–west for a distance of approximately 100 km, with the adjoining narrow coastal plain bounded by the coast-parallel and steeply sloped Santa Ynez Mountains (peak elevation approximately 1300 m). [Somewhat similar warm episodes that have occurred along limited segments of the coast of southern Oregon and northern California have been examined by Mass (1987).]

In the most severe Sundowner wind events, wind speeds can be of gale force or higher and surface air temperatures on the coastal plain, and even at the coast itself, can well exceed 100°F (37.8°C). The associated fire danger can become extreme under the combination of high temperatures, low relative humidity, and strong and gusty winds—circ*mstances dramatically different than the more typical marine-influenced local weather conditions. Many of the most destructive conflagrations that have occurred in the Santa Barbara region, including the Painted Cave fire of June 1990, which was among the more devastating fires in California history (losses in public and private buildings totaled almost $250 million), have occurred during one of these wind episodes.

Thus far, however, little in the way of synoptic or climatological analysis of these events has yet been published. This paper is thus among the first to examine in some detail the meteorological conditions associated with several significant Sundowner wind events.

2. Methodology

A rigorous definition of a Sundowner wind event has not yet been provided in the literature. Finke (1990) simply characterizes the Sundowner as the sudden onset of strong desiccating winds from the north to northeast accompanied by a rapid rise in temperature and decrease in humidity. Ryan and Burch (1992) and Ryan (1996) classify warming events in the vicinity of Santa Barbara into four categories, the first being a non-Sundowner warm event and the other three representing different intensities of Sundowner events. In their classification system, Sundowner events are distinguished from other warm episodes by the strength and direction of the surface winds and by the time of day of the maximum temperature at Santa Barbara: in the Sundowner case, surface winds tend offshore even at the coast, and the maximum temperature occurs outside the normal diurnal temperature profile.

Given the coastal location of Santa Barbara and the climatologically cool sea surface temperatures of even the nearshore water (less than 20°C throughout the year), it is hard to envision the occurrence of much warmer than average temperatures except in the presence of regional-scale offshore (and thus downslope) flow, regardless of the time of day the anomalously high temperatures are observed. In addition, local correspondence at and near the coast between the strength of the offshore-directed component of the wind and the magnitude of the anomalous increase in temperature is less than robust. In some cases, weak onshore flow is observed at Santa Barbara despite the presence of very warm air temperatures, low relative humidity, and larger-scale offshore-directed lower-tropospheric flow. As shown in Blier and Ma (1997) in a high-resolution mesoscale modeling study of the 1995 New Year’s Eve Sundowner wind event, this can result from a local rotor circulation downwind of the lee side of the mountain range. There have also been cases in which downslope winds, in association with anomalously high temperatures and low relative humidity, occur in the mountains just above Santa Barbara, while typical maritime conditions are evident in Santa Barbara adjacent to the coast.

To identify the presence of a downslope wind event, then, it might seem desirable to consider observations from locations farther inland, at the base of the lee slope of the Santa Ynez Mountains, or on the mountain slope itself. Unfortunately, routine hourly observations in the vicinity of Santa Barbara were generally available to us from only the FAA site at Santa Barbara airport (SBA).² And, as noted by Ryan (1996), this can be among the last locations on the coastal strip to manifest Sundowner conditions. Observations from SBA alone may fail to indicate a Sundowner event clearly evident elsewhere in the city. It also is not clear that one would necessarily want to classify every incident of downslope flow and associated warming anywhere on the mountain slope as a Sundowner event. Rather, it would seem more appropriate, especially in consideration of the long-standing local popular use of the term, to confine its application to those cases that have a significant impact on the meteorological conditions (i.e., wind and/or temperature) on the populated coastal plain. (Conditions on the mountain slope itself would, of course, nonetheless be of critical importance to the fire weather situation.)

In the present study we investigated the synoptic-scale meteorological conditions associated with three Sundowner events that did produce significant wind and/or warming at the official FAA site. In addition to considering the degree of departure of the maximum temperature from the normal diurnal profile [following Ryan and Burch (1992) and Ryan (1996)], we also examined the hourly positive temperature difference between Santa Barbara and Santa Maria (on the coastal plain to the northwest of the Santa Ynez Mountains). Both stations are typically under the influence of thermodynamically similar cool and moist maritime air, but during a Sundowner event the temperature at Santa Barbara will increase.³

In two of the three Sundowner events considered in the present paper, the most significant manifestation in weather conditions observed at Santa Barbara was the degree of warming that occurred, while the third event was accompanied by winds of destructive force. In the July 1992 case, the maximum temperature difference between SBA and Santa Maria (SMX) was 35°F (19.4°C) at 0000 UTC 20 July (1700 PDT 19 July); the high temperatures at SBA and SMX were 106°F (41.1°C) and 76°F (24.4°C), respectively. In the November 1992 episode, the maximum temperature difference was 23°F (12.8°C) at 0000 UTC 3 November (1600 PST 2 November), while high temperatures were 98°F (36.7°C) at SBA and 84°F (28.9°C) at SMX. The third case is the New Year’s Eve 1995 windstorm, where gusts of 44 kt (22.7 m s⁻¹) were reported at SBA (though higher values may well have occurred at times of missing reports); unofficial reports and wind damage, however, suggest that elsewhere in Santa Barbara wind speeds locally exceeded 70 kt (36.0 m s⁻¹). During this evening event, the temperature difference between Santa Barbara and Santa Maria reached approximately 12°F (6.7°C) (missing reports from both sites precluded exact determination of the maximum temperature difference).

3. The 19 July 1992 Sundowner event

At 0000 UTC 20 July 1992 (1700 PDT 19 July 1992), the temperature at Santa Barbara airport was 106°F (41.1°C), the high temperature for the day, while the temperature at Santa Maria was 71°F (21.7°C). In addition to the 35°F (19.4°C) temperature difference, the high temperature at SBA clearly occurred much later in the day than average, consistent with the definition of Ryan and Burch (1992) and Ryan (1996). Hourly weather reports from both stations are presented in Table 1. These observations show that the rapid increase in temperature at SBA occurred during the 3-h period from 2100 to 0000 UTC, while little temperature change occurred at SMX during this time period. Interestingly, though, a north wind was evident at Santa Barbara only at 0000 UTC; at earlier times the wind remained southerly-southwesterly and thus was blowing from the coast despite the temperature increase that was occurring [suggestive of a low-level rotor circulation over the nearshore waters to the lee of the Santa Ynez Mountains, as was found by Blier and Ma (1997) in a high-resolution numerical simulation of the 31 December 1995 Sundowner event, or some other local circulation in which strongly heated air quickly recirculated onshore]. Thus at the location of SBA, the Sundowner (offshore) wind itself was short lived and relatively weak, but the prior dramatic temperature increase provides evidence for the presence of the associated subsided air (though the absence of a commensurate decrease in the dewpoint suggests some remaining marine influence). Thus just above the station and through the passes and along the mountain slopes, there may well have been significantly stronger and more persistent offshore flow. Such a contrast in winds was noted in the extreme Sundowner event of 27 June 1990, associated with the Painted Cave fire. Forest Service and Sheriff’s Department personnel reported winds speeds up to 40–60 mph (17.9 m s⁻¹ to 26.8 m s⁻¹) in the Santa Ynez Mountains (Gomes et al. 1993), significantly stronger than at SBA itself (though the fire-induced circulation may have also contributed to the strength of the wind).

The mesoscale data plot for 0000 UTC 20 July 1992 (Fig. 2) shows the dramatic contrast in temperature that developed over a short distance in the vicinity of Point Conception. Cool temperatures and significant flow out of the northwest are evident from the buoy reports to its west and north. Santa Maria on the coastal plain to the north of Point Conception is under the influence of this maritime air while significantly warmer temperatures are reported from stations in the hills and valleys more protected from the ocean winds. The warmest temperature, however, is at Santa Barbara on the narrow coastal plain to the lee of the Santa Ynez Mountains. Also apparent is a strong mesoscale gradient in the sea level pressure, with lowest pressures in the vicinity of Santa Barbara. An eddy circulation is evident from the wind reports to the southeast of Santa Barbara; the buoy temperature report of 73°F (22.8°C) with a west-southwesterly wind and a sea level pressure of 1010.4 mb indicates that some of the subsided air may be infiltrating this circulation. A relatively common mesoscale low-level cyclonic circulation in this region, referred to as the Catalina eddy, has long been recognized. In fact, in a study of the 26–29 May 1968 Catalina eddy event, Bosart (1983) showed that the circulation of the incipient eddy initiated on the coast near Santa Barbara, and that this developed in association with lee troughing downwind of the coastal mountains. It was therefore speculated that mountain wave activity, occurring in the presence of synoptic-scale subsidence and with the generation of a stable layer near the crest of the local (Santa Ynez) mountains, may have aided the incipient eddy formation. Data presented by Bosart are consistent with the occurrence of a Sundowner wind event in Santa Barbara concurrent with the development of the eddy (see Fig. 5 in Bosart 1983). The exact relationship between these two phenomena (Sundowner winds and Catalina eddies), however, remains to be determined. It is conceivable, for example, that a Catalina eddy could develop without the subsidence warming to the lee of the Santa Ynez Mountains reaching the surface at Santa Barbara; the approach of an upper-level trough from the west and effects of the dramatic change in orientation of the coastline in the vicinity of Point Conception on the boundary layer flow might also play a significant role in some cases of eddy development.

At 0000 UTC 20 July, weak and generally westerly winds are evident at the 500-mb level (Fig. 3) over the region of south-central California, while the 850-mb flow (Fig. 4) is more meridional—between an inverted ridge offshore and an inverted trough over the intermountain states and the desert Southwest. At Vandenberg Air Force Base (VBG), near Santa Barbara, the 500-mb wind is from the west-northwest at 25 kt (12.9 m s⁻¹), while at 850 mb the wind direction is more northwesterly and the wind speed slightly greater: 30 kt (15.4 m s⁻¹). Thus the component of the wind perpendicular to the Santa Ynez Mountains is greater at 850 mb than at 500 mb. A more substantial temperature gradient also appears at the 850-mb level than at 500 mb. While relatively uniform warm temperatures are evident at the 500-mb level (Fig. 3), there is a significant northwest-to-southeast temperature gradient at 850 mb (Fig. 4). The 850-mb temperature increases from 20°C at Oakland on the north-central coast of California to 23°C at VBG, 26°C at San Diego, and 30°C at Tucson in southern Arizona, while the 500-mb temperature at San Diego (−5°C) is only 1°C warmer than that at Oakland.

The National Centers for Environmental Prediction (NCEP, formerly known as the National Meteorological Center) surface analysis for 0000 UTC 20 July (Fig. 5) shows a significant synoptic-scale pressure gradient from west-northwest to east-southeast in the vicinity of southern California. While Thermal in the low desert reports a sea level pressure of 1004.1 mb, the pressure at VBG is more than 10 mb higher at 1015.2 mb. Even over the relatively short distance between SBA and SMX, however, there is a difference in sea level pressure of 5.4 mb at 0000 UTC 20 July (Fig. 5 and Table 1). Note that the machine-analyzed contours fail to capture the mesoscale area of low pressure in the vicinity of SBA (reporting a pressure of 1009.8 mb). The contrast between the temperature at Santa Barbara (106°F) and at other coastal locations is strikingly apparent. In fact, the temperature at SBA is only a few degrees less than those reported in the hottest inland desert locations of southern California.

The VBG sounding from 0000 UTC 20 July 1992 (Fig. 6) shows the presence of a very strong low-level marine inversion. A cool, well-mixed layer of marine air extends from approximately 1000 to 970 mb, with a temperature of approximately 14°C at the top of this layer. Between 970 and 950 mb, the temperature warms almost 13°C, reaching approximately 27°C at the top of the inversion. A thin isothermal layer is apparent from 950 to 920 mb, then the temperature decreases significantly with height. This Sundowner event thus occurs in the apparent absence of a strongly stable layer at (or just above) the level of the ridgetop of the Santa Ynez Mountains, which at 1000–1300 m would be in the approximate pressure range of 880–910 mb.

It is instructive to examine the level from which air on this sounding would have to subside dry adiabatically in order to achieve the temperature of 106°F (41.1°C) reported at SBA at 0000 UTC 20 July. This can be determined by simply following a dry adiabat on the thermodynamic diagram (Fig. 6) from the point specified by this temperature and the pressure level of measurement (which to a good approximation is simply the sea level pressure, 1009.8 mb, given that the station elevation is only 3 m) to the point of intersection with the temperature profile. Following this procedure, a pressure level of 810 mb is found. This is well above the top of the Santa Ynez Mountains and, thus, would imply the presence of a dynamical mechanism capable of bringing air from well above the mountaintop level down to sea level on the lee side (e.g., mountain wave activity). It is, however, not uncommon for the near-surface lapse rate to become superadiabatic (typically by a few °C) on a hot summer day as a consequence of strong surface heating. Thus the foregoing calculation is repeated, but now assuming that the adiabatic ascent only needs to warm the temperature to 38.1°C (100.6°F), with diabatic processes producing the additional 3°C (5.4°F) needed to reach the observed temperature. Under these conditions the air would only have to subside from approximately the 860-mb level in order to account for the high temperature reported at SBA, or a level not far above the ridgetop. It appears, then, that the high temperature observed at SBA could occur without subsidence from a level significantly above the top of the mountains.

It is also conceivable that the very warm temperature could result simply from transport of hot surface air over the Santa Ynez Mountains from the valley to its north [as suggested by Ryan and Burch (1992)]. On 19 July 1992, the maximum temperature at Lake Cachuma was 98°F (36.7°C) at an elevation of 240 m. Simple adiabatic subsidence of this air to sea level would raise its temperature to 102°F (38.9°C). It is quite possible, though, that as a result of sensible heat flux from the surface, this air would cool at less than the dry-adiabatic lapse rate as it ascends the windward slope of the mountains. Thus transport of this air over the mountains to SBA could well account for the high temperature reported. On the other hand, a location near SBA might have experienced a higher temperature than SBA itself, especially as northerly winds were reported at SBA for only a brief period of time. If so, then a mechanism capable of producing subsidence of air from well above the mountaintop level, such as mountain wave activity, would be suggested.

4. The 2 November 1992 Sundowner event

Another dramatic example of local warming in the vicinity of Santa Barbara in association with downslope winds occurred on 2 November 1992. Temperatures (°F) reported at various locations in the southern half of California at 2100 UTC (1300 PST) are shown in Fig. 7. With a temperature of 96°F (35.6°C), SBA is significantly warmer than any other reporting station, including those in the interior desert regions. The contrast in temperature is even more dramatic when SBA is compared with other coastal locations in central and southern California. The concurrent temperature at SMX (not shown) was 81°F (27.2°F), which although quite warm is still 15°F (8.3°C) lower than at SBA. One hour later, at 2200 UTC, the temperature at SMX had decreased to 73°F (22.7°C) as the wind shifted to onshore, while winds remained offshore at SBA with a temperature of 95°F (35.0°C). The high temperature at SBA on 2 November was 98°F (36.7°C), quite remarkable for a midlatitude coastal location in the month of November.

Table 2 shows the dramatic rise in temperature that occurred at SBA between 2000 and 2100 UTC 2 November. As in the previous case, however, this temperature increase occurs in the absence of north winds at the observation site itself; again SBA appears to be within a local reverse circulation situated to the lee of the mountains. Despite a local wind direction indicating flow from the nearby coast, the dewpoint temperature has decreased significantly, demonstrating the presence of much drier air; thus the air arriving at SBA from the south is likely subsided air originating over the mountains to the north of the station, rather than air with a significant overwater trajectory.

Unlike the July event, the highest hourly temperature here occurs much closer to its time of occurrence on a typical (i.e., non-Sundowner) day. Thus Sundowner events may not always be well defined on the basis of the timing of the maximum temperature as suggested by Ryan and Burch (1992).

Also in contrast to the July case are the offshore winds and warming observed at Santa Maria as the Sundowner develops in Santa Barbara. As these stations are quite near each other, the contrast in temperature between the two locations gives some measure of the added degree of warming that occurs at Santa Barbara as a result of the mountains immediately to its north. In addition, the occurrence of a Sundowner with north-northeast flow at Santa Maria indicates that onshore flow of cool marine air at Santa Maria is not a necessary condition for a Sundowner event to occur, despite its apparent presence in the vast majority of cases.

Another difference from the July case is the significant northwesterly flow at 500 mb (Fig. 8). Similar to the earlier case, however, is the presence of backing winds between the 850- (Fig. 9) and 500-mb levels, which according to the thermal wind relation is indicative of cold advection in the intermediate layer. Cold advection is also evident at the 850-mb level itself.

The surface analysis for the same time (Fig. 10) shows a strong northwest–southeast-oriented pressure gradient extending the length of California. Similar to the July 1992 case, sea level pressure to the north of Santa Barbara, both along the coast and in the central interior of the state, are significantly higher than at Santa Barbara itself. The 0000 UTC 3 November sea level pressure difference between SMX and SBA, for example, is 4.2 mb (Fig. 10 and Table 2). Unlike the July 1992 event, however, sea level pressures at SBA are still higher (albeit just slightly) than at coastal locations to the southeast (Fig. 10).

The VBG sounding for 0000 UTC 3 November 1995 is shown in Fig. 11. Although the sounding contains obvious errors (most notably the temperature and dewpoint profiles in the 550–600-mb layer; also the surface layer where an excessive superadiabatic lapse rate is shown adjacent to the ground), it is still useful in analysis of this Sundowner event. Most notably, in comparison with the July 1992 event, no significant inversion appears near the surface [with the exception of the likely erroneous thin very stable layer just above the depicted (and also likely erroneous) superadiabatic surface layer]. However, in this case a significant stable layer extends from 855 to 885 mb, which is just above the top of the Santa Ynez Mountains, with a weaker stable layer above. A wind maximum appears at the base of this stable layer, with flow out of the north-northeast at 45 kt (23.2 m s⁻¹). Thus there is much stronger flow perpendicular to the mountains than in the July case, along with a stable layer at the appropriate level for mountain wave development (Queney et al. 1960). In order to account for the high temperature of 98°F (36.7°C) at SBA, air would have to descend from approximately the 780-mb level, or well above the top of the Santa Ynez Mountains, indicating significant mountain wave activity. Unlike the July 1992 case, alternative explanations for the high temperature appear less plausible. In November, the degree of superadiabatic warming of the surface layer is expected to be small; likewise any surface air transported from the windward side of the mountains is unlikely to experience significant warming from surface sensible heat fluxes and thus little mitigation of the adiabatic cooling during its ascent to the level of the ridgetop. In fact, the high temperature at Lake Cachuma on 2 November 1995 was only 82°F (27.8°C), or 15°F (8.3°C) less than at Santa Barbara [compared to 8°F (4.4°C) less in the July case]. Thus in the present case mountain wave development alone would appear to provide a plausible explanation for the anomalously warm temperature observed at Santa Barbara.

5. The 31 December 1995 Sundowner event

In this event temperatures were significantly more moderate than in the two episodes described above, but wind speeds reached damaging levels. A statement issued by the National Weather Service Forecast Office in Oxnard at 0645 UTC 1 January 1996 (2245 PST 31 December 1995) noted that in the coastal areas of Santa Barbara County “several roads were closed due to debris cluttering the streets” and that “wind speeds of 40 to 45 mph with gusts to near 60 mph were reported by several residents in the Santa Barbara area.” An Associated Press wire report dated 1 January 1996 indicated that “hundreds of trees blew down onto roads, cars, and buildings” and that “live power lines lay in some roads.” The report goes on to say “strong winds—measured between 50 and 60 mph at Santa Barbara Harbor—caused several traffic accidents resulting in minor injuries” and, according to a local fire captain, “wind also smashed windows of shops and restaurants on Santa Barbara’s wharf.” A number of boats in Santa Barbara Harbor were reported to have broken free of their moorings, while an Amtrak train was blocked by downed trees. The author of this paper, who happened to be entering the Santa Barbara area from the east on U.S. Highway 101 at 2130 PST 31 December, encountered a number of large trees (length >12 m) down on the road obstructing traffic flow, as well as several telephone poles that were snapped in half.

At 2100 PST 31 December, the official temperature report from SBA was 68°F (20.0°C), with a dewpoint of 36°F (2.2°C), and a northwest wind of 23 kt (11.8 m s⁻¹) with gusts to 35 kt (16.0 m s⁻¹). One hour later, the reported temperature was 70°F (21°C), with a dewpoint of 32°F (0°C), and winds from the north at 29 kt (14.9 m s⁻¹) with gusts to 44 kt (22.7 m s⁻¹). Unfortunately, there were several missing surface observation reports during the period of this event from both Santa Maria and Santa Barbara. However, at 2300 PST, Santa Maria reported a temperature of 59°F (15.0°C) and a dewpoint of 42°F (5.6°C), with a wind out of the east at 13 kt (6.7 m s⁻¹). As in the November 1992 event, even in the presence of offshore flow at both locations, the temperature was significantly cooler at Santa Maria than at Santa Barbara.

Interestingly, at the times of the preceding Santa Barbara reports, sky conditions in the Santa Barbara area were reported as “partly cloudy.” Periodic personal observation by the author from approximately 2115 to 2300 PST 31 December indicated the presence of a single linear stationary lenticular cloud approximately 10 km to the lee of the Santa Ynez Mountains and oriented parallel to the mountain range (no other clouds were evident). An observation made soon after the end of the significant wind episode revealed clear skies.

The NCEP surface analysis for 0000 UTC 01 January 1996 (1600 PST 31 December 1995) (Fig. 12) shows a strong north-northwest to south-southeast pressure gradient across the state of California. Unlike the preceding two cases, the difference in sea level pressure between SBA and SMX is only 3 mb, and the pressure at SBA is significantly higher than at coastal locations farther to the south. It should be noted, however, that this Sundowner wind episode initiated several hours after the 0000 UTC analysis time, rather than significantly earlier in the day as in the other two cases. The surface analysis also shows the tail end of a “back door” cold front approaching southern California from the northeast; pressure rises are evident behind the front while pressure falls are occurring ahead of it.

The corresponding 850-mb analysis (Fig. 13) indicates strong north-northeasterly geostrophic flow in the vicinity of Santa Barbara, while Vandenberg AFB is reporting northerly winds at 40 kt (20.6 m s⁻¹). A region of significant cold advection is evident beginning several hundred kilometers upstream. Consistent with this cold advection, backing of the geostrophic flow occurs between the 850- and 500-mb levels. At Vandenberg AFB, the observed 500-mb wind direction (Fig. 14) is the same as that at 850 mb, but with the wind speed increased to 65 kt (33.5 m s⁻¹). The geopotential height of 581 dm is quite high for the location and time of year, as is the 500-mb temperature of −10°C.

Significant additional information on the vertical structure of the atmosphere just prior to the initiation of the wind event is evident from the 0000 UTC 1 January 1996 VBG sounding (Fig. 15). Three key features appear in this sounding: a strong stable layer at approximately the 850-mb level—and thus just above the top of the Santa Ynez Mountains, a lower-tropospheric jet with wind speeds reaching 75 kt (38.6 m s⁻¹) at about the 710-mb level (the maximum wind speed between the surface and the tropopause), and (albeit modest) backing of the wind above this level, resulting in further reduction of the magnitude of the cross-mountain component of the flow.

6. Discussion

A number of different dynamical mechanisms have been proposed to explain downslope windstorms. These have, however, generally been derived using simplified theory, atmospheric structures, and terrain profiles and thus their degree of applicability to the specific and more complex circ*mstances associated with a particular observed event is unclear. With respect to the Sundowner phenomenon considered here, several additional complications arise. First, lack of a robust historical database of surface observations at locations adjacent to or on the mountain slope itself limits the analysis to just consideration of those cases that are sufficiently intense to impact conditions near the coast, and even then, as noted by Ryan (1996), not necessarily where the strongest effect is typically observed. Second, there are really two somewhat different forecast concerns that arise in association with these Sundowner events: the potential for strong winds and the potential for very warm temperatures. In the July 1992 case, for example, very high temperatures occurred on the coastal plain but associated wind speeds were quite modest. In contrast, the December 1995 event brought winds of destructive force but temperatures were no more than pleasantly mild for a midwinter evening. Finally, the mountain barrier formed by the Santa Ynez Mountains does contain some significant canyons and passes and thus the potential for strong local channeling of the downslope wind exists (Ryan 1992).

Linear mountain wave theory should be reasonably applicable to those cases in which wind speeds, both upstream and aloft (e.g., in the VBG sounding) and downwind on the coastal plain, remain modest (with, again, the caveat that conditions at SBA might not always be representative of those at other locations nearby). Within this framework, three key criteria were described by Queney et al. (1960) and Klemp and Lilly (1975). First, the mountain barrier causing the event should have a steep lee slope. Significant theoretical work by Smith (1979) and Lilly and Klemp (1979) within an inviscid framework showed that asymmetric mountains, with a steep leeward slope and a gentle windward slope, functioned best at generating large-amplitude mountain waves. Subsequent numerical simulation with inclusion of surface friction, however, not only produced more realistic results in general (Richard et al. 1989) but also suggests that it is just the steepness of the lee slope that is of significance (Miller and Durran 1991). Second, throughout a deep layer of the troposphere, the wind should be directed significantly across the mountain barrier (within 30° of perpendicular to the ridge line), with the wind speed increasing with height and exceeding 7–15 m s⁻¹ at the crest (values depending on the height and shape of terrain). Third, the upstream temperature profile should show a strongly stable layer near the mountaintop level, with weaker stability above. [Weaker stability below the inversion seemed to also favor lee wave development to the east of the Sierra Nevada (Colson 1954).] Finally, Durran (1986) notes that strong downslope winds are more likely when the synoptic-scale pressure gradient is in phase with the wave-induced pressure gradient.

The July 1992 case satisfies the first two of the above criteria, but apparently not the third (assuming that the VBG sounding is representative of the upstream temperature profile). In this case, then, it would seem plausible that the very warm temperatures observed near the coast at Santa Barbara were produced by a combination of simple advective transport of hot surface air from the valley upwind of the Santa Ynez Mountains, strong surface diabatic heating, and erosion of the marine boundary layer allowing downward penetration of the very warm air above.

In the November 1992 case, however, a significant stable layer is evident just above the level of the mountain crest; here, then, all of the aforementioned criteria appear to be satisfied and thus mountain wave development is suggested. Also in contrast to the July case (and consistent with the presence of mountain waves) is the lack of a plausible mechanism to account for the very high temperatures observed at SBA other than downward transport of air from a level significantly above that of the mountain crest. As in the July 1992 event, offshore winds at SBA were of only modest magnitude.

In the December 1995 case, however, a very strong downslope wind event clearly occurred, thus indicating the likely development of large-amplitude mountain waves and therefore the need for consideration of the fully nonlinear dynamics. Durran (1990) discusses three different possible mechanisms for the production of severe downslope winds: 1) the development of supercritical flow in a hydraulic jump, 2) the generation of large-amplitude vertically propagating mountain waves, and 3) the development of a wave-induced critical layer. Durran concludes that the first of these mechanisms appears to offer the best dynamical description for downslope windstorms but, unfortunately, is not easily applied in practical forecasting due to the number of rather different physical circ*mstances that could potentially result in a transition from subcritical to supercritical flow.

Results of a recent three-dimensional high-resolution nonhydrostatic modeling study of windstorms along the western side of the Washington Cascade Mountains (Colle and Mass 1998) suggest that intense windstorms in the foothills of the lee slope of these mountains are associated with the combination of strong cross-barrier flow and an environmental critical level (i.e., a level at which there is no cross-barrier flow) or layer of reverse shear (cross-barrier winds decreasing with height) in the lower to middle troposphere. The magnitude of the simulated windstorm increased if, in addition to the above, a moderately shallow strongly stable layer existed just above the mountain crest with reduced stability at higher levels.

All of these conditions were satisfied in the December 1995 Sundowner event. However, general applicability of these findings to the Sundowner winds of the Santa Barbara region may be mitigated by significant differences in the topography, especially the comparatively much larger width of the Cascades and the presence of a significant gap in that mountain range. In addition, both previous theoretical studies and the numerical simulations of Colle and Mass (1998) more thoroughly examine the influence of a tropospheric critical level than that of a layer of reverse shear. While the former is present in most significant Cascade Mountain wind storm events, this is not the case for the Sundowner winds of Santa Barbara.

Operational Eta Model numerical output was examined for the December 1995 case (corresponding operational model output for the other two events considered in this paper is no longer readily obtainable). In particular, we examined 18-h output from the model forecast initialized at 1200 UTC 31 December 1995 (valid at 2200 PST 31 December; not shown). Although the large-scale structure of the surface, 850-, and 500-mb flow appeared well predicted, mesoscale detail of the Sundowner event was entirely absent, as would be expected given the very small horizontal scale of the Santa Ynez Mountains.

Assuming, then, that operational model output is capable of accurately forecasting the larger-scale wind and thermal structure of the atmosphere in the region in which these Sundowners occur, some guidance can be provided to the operational forecaster. During the summer season, one should be alert for the possibility of anomalously warm temperatures on the coastal plain whenever sea level pressure at Santa Barbara is lower than that in the central interior of California and significantly lower than at Santa Maria [for the latter, Ryan (1996) proposes a minimum of 1.8 mb, while Finke (1990) suggests a minimum of 2.5 mb between Santa Maria and Los Angeles with even lower sea level pressure at Santa Barbara], and the 850-mb-level flow has a significant cross-mountain (i.e., northerly) component. The shallower the marine inversion layer, the warmer the air just above, and the higher the surface air temperatures in the Santa Ynez Valley upwind of the mountains, the greater the potential for very high temperatures to occur under such circ*mstances. In essence, these cases seem to represent the migration of typical warm season interior weather conditions to the more commonly marine-air-dominated coastal plain. Although temperatures at Santa Barbara are thus dramatically higher than those usually experienced, they are generally comparable to those typically observed during the summer at low elevations in the interior of the state (as is evident from Fig. 5).

Anomalously very warm temperatures can also occur during other times of the year, as evidenced by the November 1992 event. In these cases, though, temperatures above the boundary layer and in the interior are typically much lower than in the summer months and thus subsidence from well above the mountaintop level is necessary. This implies mountain wave development, as previously discussed. Here, then, the forecaster would want to be alert for the combination of the sort of pressure gradient and 850-mb flow described above with a layer of enhanced stability in the VBG sounding at or slightly higher than the level of the mountain crest. The occurrence of a stable layer well above the typical top of the marine boundary layer in conjunction with northerly flow aloft is most likely to occur in association with an elevated frontal zone and, thus, after the passage of a surface cold front.

Infrequently, an elevated stable layer (i.e., one extending above the mountaintop level) in association with significant northerly flow might occur in one of the summer months when the temperature inland and just above the stable layer tends to be at its warmest. The potential resulting combination of a very warm air mass and mountain wave development could conceivably produce the sort of extraordinary high temperature described in the introductory section of this paper. Cursory examination suggests the 27 June 1990 Sundowner was such an event. On this day the maximum temperature at Santa Barbara airport was 109°F while the temperature at El Capitan Beach 11 miles (17.7 km) farther west reached 116°F; winds of 30 mph (13.4 m s⁻¹) were reported at the airport at 1548 PDT and wind gusts to 60 mph (26.8 m s⁻¹) occurred within the city of Santa Barbara (Ryan 1996).

Whenever there is both strong cross-mountain flow and a strong mountaintop stable layer (a combination most likely to occur during the cool season months), the possible development of a significant wind event on the coastal plain needs to be considered. It would appear that the potential for a severe windstorm event is enhanced if the high wind speeds upstream extend down to low levels, the stable layer is relatively shallow with a layer of significantly reduced stability above, and a layer of reverse shear is evident in the lower to middle troposphere.

7. Summary

A mesoscale region of the California coast in the vicinity of Santa Barbara occasionally experiences extraordinarily warm temperatures and significant downslope winds in association with lower-tropospheric flow over the abruptly rising Santa Ynez Mountains to its north. Synoptic overviews of three such cases of Sundowner wind events have been presented, along with consideration of possible physical mechanisms responsible for the observed occurrence of high temperatures and/or strong winds.

In the cases considered here, as well as in others examined in a more cursory fashion by the author, significant perpendicular flow at the ridgetop level of the Santa Ynez Mountains is present. The predominant resulting surface flow in the vicinity of Santa Barbara on the narrow coastal plain to the lee of the mountains is thus offshore, precluding the significant and more typical cool marine influence. In the July 1992 case (and perhaps to a more limited degree in the November 1992 case), some of the warming is thus likely a consequence of the greater diurnal heating enabled in the absence of a sea breeze. In order to account for the very high temperatures sometimes observed, however, other processes are necessary. In particular, significant warming can be produced by adiabatic descent from above ridgetop level associated with mountain wave development, and (in the warm season) by transport of hot surface air from the (sea-breeze protected) valley to the windward side of the Santa Ynez Mountains. The former appears to have primarily been responsible for the warming observed in the November 1992 case, while the latter may have played a significant role in the July 1992 case.

In the December 1995 case, by contrast, the wind manifestation was of far greater significance than the associated warming. In this Sundowner event, wind speeds reached destructive magnitudes and caused significant damage in the Santa Barbara area. This event also started in the early evening hours, rather than midday to early afternoon as in the other two cases.

Much further work is needed, however, on this fascinating and heretofore relatively unexamined mesoscale topographic phenomenon. In particular, acquisition of much higher resolution surface and upper-air data during a Sundowner wind event would contribute significantly to our understanding of the processes responsible for the associated very high temperatures (and sometimes accompanying strong winds). As just one example, there is at present little in the way of observational data to enable definitive examination of the possible influence of the nearby, higher, and larger-scale Sierra Madre and San Rafael Mountains on these events. It is conceivable, for example, that mountain wave development from these ranges could affect conditions in the Santa Barbara area; the barrier formed by these mountains could also act to block onshore flow and divert it to the west with the result of enhanced low-level flow over western Santa Barbara county.

Mesoscale modeling studies are also needed, both to further understanding of these Sundowner wind and warming events, and to examine their relationship to the development of the Catalina eddy. Numerical simulation of the December 1995 event has been performed and will be reported on in a subsequent paper; preliminary results can be found in Blier and Ma (1997). As extraordinarily warm temperatures, strong and sometimes damaging winds, and a number of destructive firestorms have been associated with these Sundowner winds, improvement in our understanding and forecasting capability of these events is essential.