Experimental observations from coastal and island wind profilers, aircraft, and other sensors deployed during the California Land-falling Jets Experiment of 1997/98 and the Pacific Land-falling Jets Experiment of 2000/01–2003/04 were combined with observations from operational networks to document the regular occurrence and characteristic structure of shallow (∼400–500 m deep), cold airstreams flowing westward through California’s Petaluma Gap from the Central Valley to the coast during the winter months. The Petaluma Gap, which is the only major air shed outlet from the Central Valley, is ∼35–50 km wide and has walls extending, at most, a modest 600–900 m above the valley floor. Based on this geometry, together with winter meteorological conditions typical of the region (e.g., cold air pooled in the Central Valley and approaching extratropical cyclones), this gap is predisposed to generating westward-directed ageostrophic flows driven by along-gap pressure differences. Two case studies and a five-winter composite analysis of 62 gap-flow cases are presented here to show that flows through the Petaluma Gap significantly impact local distributions of wind, temperature, precipitation, and atmospheric pollutants. These gap flows preferentially occur in pre-cold-frontal conditions, largely because sea level pressure decreases westward along the gap in a stably stratified atmosphere in advance of approaching cold-frontal pressure troughs. Airstreams exiting the Petaluma Gap are only several hundred meters deep and characterized by relatively cold, easterly flow capped by a layer of enhanced static stability and directional vertical wind shear. Airborne air-chemistry observations collected offshore by the NOAA P-3 aircraft illustrate the fact that gap-flow events can transport pollutants from inland to the coast, and that they can contribute to coastally blocked airstreams. The strongest gap-flow cases occur when comparatively deep midtropospheric troughs approach the coast, while the weak cases are tied to anticyclonic conditions aloft. Low-level cold-frontal pressure troughs approaching the coast are stronger and possess a greater along-gap pressure gradient for the strong gap-flow cases. These synoptic characteristics are dynamically consistent with coastal wind profiler observations of stronger low-level gap flow and winds aloft, and greater rainfall, during the strong gap-flow events. However, gap flow, on average, inhibits rainfall at the coast.
Mountains can profoundly modify stably stratified, lower-tropospheric airstreams, resulting in shallow and relatively cold blocked flows trapped against the high terrain. Blocking deflects the low-level flow upstream and below the top of mountain ranges (e.g., Bell and Bosart 1988; Colle and Mass 1995) and results in the formation of a barrier jet below the mountain top that parallels the long axis of quasi-two-dimensional high terrain (e.g., Parish 1982; Marwitz 1983; Overland and Bond 1995; Doyle 1997). It can also locally redistribute precipitation, inhibit the forward motion of approaching fronts, and produce strong winds as a result of the superposition of a barrier jet and prefrontal low-level jet (LLJ; e.g., Dunn 1987; Peterson et al. 1991; Marwitz and Toth 1993; Doyle 1997; Sinclair et al. 1997; Braun et al. 1999; Doyle and Bond 2001; Yu and Smull 2000; Yu and Bond 2002; Neiman et al. 2002, 2004; Medina and Houze 2003). Wintertime examples of blocked flow along California’s mountainous coastal zone are shown in Ralph et al. (1999a) using data collected from a unique array of coastal and island wind profilers deployed by the National Oceanic and Atmospheric Administration/Environmental Technology Laboratory (NOAA/ETL) and from a NOAA P-3 research aircraft during the California Land-falling Jets Experiment (CALJET; Ralph et al. 1999b). The analyses highlight the deflection of the low-level flow at the coast (i.e., this flow is aligned along the mean barrier orientation) relative to the ambient southerly flow offshore. Above, a layer of directional vertical wind shear marks the interface between the lower layer of blocked flow and the upper layer in which the winds closely match the flow direction observed offshore. The top of this interface scales with the altitude of the nearby coastal mountains.
A host of physical mechanisms can contribute to the generation of shallow, stably stratified air trapped along the coast during the cool season. For example, a remnant cold pool can remain trapped against the mountains following the passage of a warm front aloft (e.g., Bjerknes and Solberg 1921; Garner 1999). Orographically forced ascent of a stably stratified airstream yields a dome of adiabatically cooled air and windward ridging (e.g., Smith 1979). Diabatic cooling via evaporation of rainfall in subsaturated conditions provides an effective mechanism for creating shallow pooling of cold air (e.g., Carbone et al. 1995; Wang and Chen 1998). The melting of hydrometeors offers another diabatic process by which shallow cold air can be created (e.g., Marwitz and Toth 1993; Bousquet and Smull 2003; Steiner et al. 2003). Another mechanism, which is the focus of the present study, is the extrusion or channeling of cold continental air through gaps in mountainous terrain. Gap flows can act as a source of cold, dry air, especially in temperate coastal zones such as along the west coast of North America (e.g., Overland and Walter 1981; Bond et al. 1997; Loescher et al. 2006; Colle et al. 2006). The study presented here extends earlier work by using a unique combination of coastal and island boundary layer wind profilers deployed for five winter seasons, as well as research aircraft and other sensors, to clearly document characteristics of flow through a key gap in California’s coast ranges and to illustrate the potential importance of gap flow in contributing to coastally trapped southerly airstreams in winter along the California coast.
To a first order, gap flow occurs in response to a gap-parallel pressure gradient when the width of the gap is less than the Rossby radius (e.g., Colman and Dierking 1992),
where f is the Coriolis parameter, g is gravitational acceleration, H is the depth of the cold air in the gap, Δθ is the potential temperature gradient between the shallow cold air and warm air above, and θ0 is the temperature of the warm air (Glickman 2000). In this regime, the gap flow attains an approximate ageostrophic balance between inertial forces and the gap-parallel pressure gradient (Overland and Walter 1981). An integrated form of the steady, frictionless force–balance relationship yields the Bernoulli equation:
where U2 and U1 are the gap-parallel wind speeds at two points in the gap, Δp is the pressure difference between these points, and ρ is the mean air density. This relationship demonstrates that the gap-parallel wind speed scales to the gap-parallel pressure gradient, and it implies that the gap flow attains its peak magnitude at the exit of the gap (assuming a given Δp exists for the length of the gap). Gap winds have been observed entering North America’s west coastal zone through numerous outlets from the interior. Perhaps the most widely studied is flow through the Strait of Juan de Fuca (e.g., Overland and Walter 1981; Bond et al. 1997; Colle and Mass 2000), separating Washington’s Olympic Peninsula and British Columbia’s Vancouver Island. The source region of the cold continental air exiting the strait initially funnels through the Fraser River Valley (Mass et al. 1995) and the fjords emptying into Howe Sound near Vancouver, British Columbia (Jackson and Steyn 1994). Other geographically favored locations for gap flow include Cook Inlet and Shelikof Strait southwest of Anchorage, Alaska (e.g., Lackmann and Overland 1989; Macklin et al. 1990; Bond and Stabeno 1998), Yakutat Bay in Alaska, and other nearby outlets (Loescher et al. 2006), the fjords near Juneau, Alaska (Colman and Dierking 1992), and the Columbia River Gorge near Portland, Oregon (Sharp 2002). Evidence of gap flow was recently reported farther south in the Los Angeles area (Neiman et al. 2004) and even in southern Mexico (Steenburgh et al. 1998).
Our observational study presents a five-winter analysis of offshore-directed flows through the Petaluma Gap along California’s central coast region near San Francisco. Figure 1a shows the domain of interest, including baselines AA′ and BB′ across the Petaluma Gap and CC′ along the gap. The ridges flanking the gap (hereafter referred to as “gap walls”) extend upward to 600–900 m MSL at its initial constriction west of the Central Valley (AA′), and the gap width here is 35–50 km (Fig. 1b). The downstream end of the gap near the coast (BB′) is shallower and less distinct. The floor along the length of the gap (CC′) does not exceed 180 m MSL (Fig. 1c). Cold air exiting the gap is typically ∼400–500 m deep (as will be shown in detail later), the potential temperature above the cold air is on the order of 285 K, and the thermal contrast is ∼5 K (based on case study observations). These values yield a Rossby radius of ∼100 km, or at least twice that of the gap width. Hence, the Petaluma Gap is predisposed to generating flows driven by along-gap pressure differences.
Gap flows here (defined as flows directed offshore from the Central Valley to the coastal zone) are generally less pronounced than those observed farther north during winter, in part because this gap is poorly defined by comparison. Petaluma Gap flows are also weaker because the source of cold air is California’s Central Valley rather than the northern interior of North America where arctic air is ubiquitous. Specifically, the Sierra Nevada Mountains act as a continuous 3–4-km-tall barrier that prevents very cold air in the continent’s interior from regularly entering the Central Valley. Rather, cold air in the Central Valley usually forms in situ when air within the valley stagnates beneath a capping subsidence inversion and radiatively cools for an extended period of time, often producing widespread fog or stratus that inhibits diurnal heating (e.g., Gilliam 1970; Holets and Swanson 1981; Underwood et al. 2004). Nocturnal katabatic flows from the Sierras can also contribute to the generation and maintenance of the cold pool. An illustrative example of a persistent Central Valley cold pool was captured recently by a NOAA/ETL wind profiler and radio acoustic sounding system (RASS; described in section 2) at Chowchilla, California (CCL; Fig. 2). Shallow cold air formed in situ on 11 December 2004 after the low-level flow weakened to <3.5 m s−1. The cold air persisted in the stagnant flow for two weeks and was capped by a layer of enhanced static stability whose base ranged in altitude from 300 to 1100 m. The wind shear across the cap was weak, thus preventing the downward mixing of warmer air from aloft. The temperature at the surface remained remarkably constant during this event (not shown), which was also characterized by widespread fog in the Central Valley. Systematic wind-direction shifts aloft marked the passage of weak synoptic-scale features in the free troposphere. The cold pool ultimately eroded shortly after 0000 UTC 24 December when a major cyclone impacted the region (not shown).
The 1–2-km-tall coast mountain ranges to the west of the Central Valley contain several passes as low as 500–1000 m MSL, which allow a small volume of the cold pool to exit the Central Valley if the stagnant air mass is sufficiently deep. However, the only major air shed outlet from the Central Valley is the Petaluma Gap. Despite the comparatively weak character of flows though the Petaluma Gap relative to its northern counterparts, these flows can significantly impact the local distributions of wind, temperature, precipitation, and atmospheric pollutants along the coast, as the forthcoming case studies and multiwinter analyses will reveal. This paper shall also describe how gap flows can contribute to coastally trapped airstreams in California, as suggested in Fig. 3 and described in Loescher et al. (2006) and Colle et al. (2006) for the Alaskan coast. In Fig. 3, a virtual potential temperature θυ analysis at 2100 UTC 28 January 1998 (Fig. 3a) shows cold air extruding westward through the Petaluma Gap from the Central Valley to the coast; the cold air turned northwestward up the coast when it encountered a southerly component flow. At about this time, a NOAA P-3 research aircraft, flying at 150–300 m MSL, recorded a distinct radar fine-line boundary separating ambient southerly flow conditions well offshore and coast-parallel, blocked flow in the coastal zone (Fig. 3b). A comparison of wind profiler data at Point Arena, California (PAA) and in situ P-3 data well offshore (Fig. 3c) reveals that the blocked flow was confined to below the tops of the nearby coastal mountains.
2. Observing systems
During the CALJET experiment of 1997/98 and the four follow-up wintertime field campaigns between 2000 and 2004 during the Pacific Land-falling Jets Experiment (PACJET; White et al. 2002; Neiman et al. 2005), numerous research observing platforms were deployed across the West Coast of the United States, and they complemented the existing operational observing networks. Those observing systems that are critical to this study are described below.
Research data collected from 915-MHz wind profilers (e.g., Ecklund et al. 1988; Carter et al. 1995) operated in California by NOAA/ETL were utilized in this study, including those whose locations are shown in Fig. 1a [Farallon Islands (FNI) at 10 m MSL; Bodega Bay (BBY) at 12 m MSL; Cazadero (CZD) at 475 m MSL; and PAA at 18 m MSL).1 The profilers provided hourly averaged vertical profiles of horizontal wind velocity from ∼0.1 to 4.0 km AGL with ≤100-m vertical resolution in clear, cloudy, and precipitating conditions. The profiler winds were objectively edited using the vertical–temporal continuity method of Weber et al. (1993), and an extra level of quality control was performed by visual inspection. A RASS (May et al. 1989) at BBY provided vertical profiles of virtual temperature, which we later transformed into profiles of virtual potential temperature using the method described in Neiman et al. (1992). The upper range of these profiles was ∼500–1000 m AGL. At each profiler site a tipping-bucket rain gauge measured rainfall with 0.01-in. (∼0.25 mm) resolution, while observations of wind, temperature, moisture, and pressure were collected from a 10-m tower, all with 2-min sampling. This surface sampling strategy was implemented at CZD for all five winters, even though the profiler was deployed here only during CALJET.
We utilized data from two NOAA P-3 research flights offshore of central California between 2045 UTC 28 January and 0400 UTC 29 January 1998 during CALJET and between 2224 UTC 20 January and 0649 UTC 21 January 2001 during PACJET. Each flight had similar dual missions: to document orographically modulated flows immediately offshore of BBY several hundred kilometers ahead of an advancing polar cold front, and to probe the cold front and its prefrontal LLJ environment beyond the reach of topographic influences. The aircraft measured standard meteorological parameters every second, and it also recorded observations of precipitation elements with its tail-mounted and lower-fuselage (LF) radars (e.g., Jorgensen and Smull 1993; Jorgensen et al. 1996). Although both radars successfully gathered data, only LF C-band (∼5.6-cm wavelength) radar data are presented in this study. During the flight of 20–21 January 2001, the P-3 aircraft carried a special airborne air-chemistry-observing system deployed by the NOAA/Aeronomy Laboratory (AL). This instrument package recorded 1-s-averaged carbon monoxide (CO) with a precision of about one part per billion by volume (ppbv) and an accuracy of better than 5% using a vacuum ultraviolet fluorescence system described by Holloway et al. (2000).
Hourly surface meteorological data were obtained from operational METAR sites (i.e., routine airport-based weather observations) and moored and drifting buoys (including 10-min resolution wind data from many moored buoys). The METAR site at Sacramento, California (SAC), and the buoy offshore of San Francisco’s Golden Gate Gap (46026) are shown in Fig. 1a. Three-hourly meteorological data were acquired from ships, and operational rawinsondes provided 12-h sampling of the troposphere. Geostationary Operational Environmental Satellites (GOES) provided nearly continuous spatiotemporal observations with their infrared, visible, and upper-tropospheric water vapor sensors. The GOES also yielded 3-hourly measurements of tropospheric feature-tracked winds during CALJET (Velden et al. 1997; Nieman et al. 1997) and hourly measurements with higher spatial resolution during the GOES Rapid-Scan Winds Experiment specifically designed for PACJET (Velden and Stettner 2001).
3. Illustrative case studies of gap flow
During its multiwinter deployment, the BBY wind profiler/RASS frequently observed shallow, stably stratified, easterly gap flow emanating from the Petaluma Gap. This crucial remote sensor, in tandem with the previously mentioned observing systems, documented important characteristics associated with gap flow. This section describes two illustrative examples of gap flow where unique supplementary data platforms were utilized: the island wind profiler at FNI and the airborne CO system on the P-3.
a. 5–10 January 1998: Coastal and island-mounted wind-profiler comparison
The gap-flow event of 5–10 January 1998 was one of the most prominent and longest lived observed by the BBY profiler, and it occurred primarily during weak and relatively steady pre-cold-frontal synoptic conditions. This case had the added advantage of a properly functioning wind profiler on the ideally situated FNI. To characterize the steady large-scale conditions, synoptic composite analyses during the core period of gap flow (i.e., between 6–9 January inclusive) were derived using the coarse (∼2.5° latitude × ∼2.5° longitude) daily global gridded dataset from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis project (Kalnay et al. 1996). The composite sea level pressure (SLP) analysis (Fig. 4a) depicts a broad region of southerly component geostrophic flow between an anticyclone centered over Nevada and a cyclone over the Pacific. Aloft at 500 mb, a low-amplitude ridge–trough couplet resided over California and the central Pacific, respectively (not shown). The mean SLP gradient along the Petaluma Gap in the inset domain in Fig. 4a was directed offshore. A representative θυ analysis in this domain at 1800 UTC 7 January (Fig. 4b) shows cold air extruding westward through the Petaluma Gap from the Central Valley. The cold air also funneled through the tight constriction of the Golden Gate Gap at the western end of San Francisco Bay. The θυ analysis also shows evidence of cold air exiting the Salinas Valley into Monterey Bay in the southeast portion of the domain.
Wind profiler data from FNI and BBY (Fig. 5) highlight key differences offshore and at the coast during this prolonged event. Most notably, persistent easterly gap flow was observed in a cold, stably stratified air mass at BBY between 0000 UTC 6 January and 0000 UTC 10 January below ∼600 m MSL, whereas the easterlies at FNI were confined primarily to the surface between 1200 UTC 7 January and 0000 UTC 10 January (Figs. 5a,b). This contrasting behavior is illustrated further using time series traces of the layer-mean zonal wind speed between 250–550 m MSL (i.e., in the layer between the undulating gap floor and the top of the gap walls). These time series (Fig. 5c) show persistent easterly gap flow above the surface at BBY, but not at FNI, during the core period between 6 and 9 January. Key differences in frontal behavior were also observed at these sites. For example, enhanced southerly component flow at FNI shifted briefly to westerly during a frontal passage at ∼0600 UTC 7 January. A second cold front traversed FNI at ∼0000 UTC 10 January and was also accompanied by a distinct wind-direction shift. Unlike at FNI, these fronts did not penetrate downward to the surface at BBY, quite likely because the stable stratification within the cold gap flow limited the intensity of vertical turbulent-scale mixing (e.g., Panofsky and Dutton 1984), thus decoupling the frontal zones aloft from the surface (e.g., Neiman et al. 1990, 1998, 2004). The gap flow subsequently eroded to the surface by 0900 UTC 10 January. Note that a diurnal θυ cycle is evident in Fig. 5b.
A more detailed portrayal of the second frontal passage is provided in Fig. 6. An analysis of zonal isotachs from the island wind profiler at FNI reveals a sharp, nearly vertical transition between 2200 UTC 9 January and 0000 UTC 10 January separating pre-cold-frontal southeasterlies from postfrontal southwesterlies. A skin layer of ocean-modified prefrontal easterly gap flow was observed at the surface, which was considerably warmer and more moist than that observed near the source region of cold, dry air at BBY, but colder and drier than the underlying sea surface. After the front removed the gap flow at FNI, the surface temperature and water vapor mixing ratio within the maritime postfrontal air mass increased by 2.0°C and 1.5 g kg−1 to values approximating the ocean surface. Intense rainfall totaling 43 mm accompanied the frontal passage. Farther east at BBY, the frontal wind shift occurred 2 h later, but only above the ∼600-m-deep layer of cold, dry easterly gap flow; a prominent pressure trough accompanied the frontal passage aloft. Little rain fell at BBY during the frontal passage aloft, quite likely because the gap flow significantly reduced the available low-level moisture content ahead of the approaching front and may have also resulted in evaporation of rain falling from above. Modest rainfall (11 mm) and a weak pressure trough at BBY eventually accompanied the surface transition from gap to maritime postfrontal flow 8 h later, during which time the surface temperature and moisture content increased to values approaching the ocean-surface values.
Mean BBY profiles of wind direction, zonal wind speed, θυ, and Brunt–Väisälä frequency composing an 84-h period within the core of the gap-flow event between 0900 UTC 6 January and 2100 UTC 9 January are shown in Fig. 7. These profiles document easterly component flow below ∼1 km MSL and the strongest vertical directional wind shear through the upper half of the gap between 300–600 m MSL. A ∼10 m s−1 easterly jet was centered at ∼240 m MSL within the confines of the gap, and vertical zonal wind shear was maximized (∼25 m s−1 km−1) atop this gap jet. Significantly, the gap jet was absent at FNI, consequently yielding the greatest difference between the FNI and BBY zonal wind profiles beneath the top of the Petaluma Gap walls. Mean thermodynamic conditions during the gap-flow event were statically stable, but the stability was maximized below ∼500 m MSL in the vicinity of the cold gap outflow. Clearly, the gap-flow signatures of enhanced easterly flow and static stability scaled to the height of the Petaluma Gap. The mean surface temperature at FNI during the 84-h period was 3.2 K warmer than at BBY (Fig. 7b), reflecting the fact that FNI was within the surface remnants of the gap flow for only 75% of this period versus 100% at BBY, and that FNI was surrounded by water and was farther from the cold air source than at BBY (i.e., airmass modification by the sea surface likely contributed to warmer surface temperatures during the remnant gap flow at FNI).
The prolonged character of this gap-flow event provided an opportunity to quantitatively explore the relationship between the pressure difference along the Petaluma Gap and the flow through the gap via a correlation analysis (Fig. 8). The outflow of cold air from the Central Valley was approximated using hourly measurements of zonal wind speed from the BBY wind profiler, while coincident hourly measurements of the surface pressure difference were taken between SAC and buoy 46026 (both located near sea level; see Fig. 4b for locations). The altitude range of the profiler data utilized in the analysis is 250–550 m MSL, that is, below the top of the gap walls. The usage of wind data within this range is more appropriate than at the surface since the surface data should be influenced more strongly by friction and local drainage flows, especially given the stably stratified character of the gap flow. The correlation analysis assumes that the flow in the cold-air source region (i.e., the Central Valley) did not change appreciably during the 5.75-day period of interest; that is, the flow through the Petaluma Gap was dominated by gap-flow dynamics rather than by evolving synoptic conditions. In fact, the synoptic conditions were generally steady, and the flow in the Central Valley weak, during most of the analysis period until 0900 UTC 10 January when the cold gap flow departed BBY following the cold-frontal passage aloft 8 h earlier.
The scatterplot of observed data (Fig. 8) reveals a cluster of 15 data points separated from the majority of points. These 15 points correspond to the period following the cessation of cold gap flow at BBY, which represents a very different regime than that characterizing the remaining 123 points. These remaining points exhibit a strong correlation of r 2 = 0.86. This analysis clearly demonstrates that the offshore-directed gap flow responded nearly linearly to the pressure gradient along the gap, despite the fact that one should expect this flow to vary with the square root of the pressure difference in an inviscid framework described by the Bernoulli relationship in Eq. (2). Furthermore, the observed gap flow at BBY was roughly half that expected from Eq. (2) for along-gap pressure differences greater than 1 mb, assuming that the along–gap flow at the entrance of the gap was zero [i.e., U1 = 0 in Eq. (2)] and that the mean density was 1.25 kg m−3 (these assumptions are supported by surface observations). These results strongly suggest that friction contributed to the observed characteristics of the low-level flow through the gap.
A variant of the Bernoulli equation that considers friction (from Mass et al. 1995) can be expressed as
where U1 and U2 are the along-gap wind speed at the entrance and exit of the gap, H is the mean depth of the boundary layer in the gap flow, ρ is the mean air density, C is the drag coefficient, and Δp/Δx is the pressure gradient between the entrance and exit of the gap. As before, U1 is assumed to be negligible. Based on Fig. 7, H = 438 m MSL, above which a capping stable layer and the strongest vertical shear were observed. The distance between SAC and BBY, and between SAC and 46026, are nearly the same at Δx = 142 km. A moderate land roughness yields a C of 7.5 × 10−3, as in Mass et al. (1995). The resulting relationship between Δp and U2 (shown in Fig. 8) produces a superior relationship relative to the observations than its inviscid counterpart. Clearly, friction retarded the flow through the Petaluma Gap to an altitude above the gap jet. However, frictional effects were likely less significant here than at the surface, given that the observations show larger wind speeds for a given Δp (for Δp > ∼1 mb) than Eq. (3) would suggest.
In addition to friction, the downward entrainment of weaker zonal flow from aloft could have also contributed to the mismatch between the observations and the theoretical curve based on Eq. (2). However, gradient Richardson number (Ri) calculations within the strong shear layer above the gap jet yielded Ri values no smaller than order unity (given the strong stability measured by RASS in this shear zone), thus suggesting that turbulence and entrainment did not play a major role in modulating the magnitude of the gap flow. Finally, venting of cold air from the Petaluma Gap into San Francisco Bay and out the Golden Gate Gap may have contributed to the discrepancy between observations and inviscid theory.
b. 20–21 January 2001: Using airborne air-chemistry data as a tracer for gap flow
During the gap-flow event of 20–21 January 2001, which was typical in its behavior, unique P-3 observations of CO over the coastal waters offshore of BBY provided supplementary insight regarding the source region of the gap flow. During the airborne mission, the P-3 initially flew offshore to intercept the cold-frontal comma-shaped cloud tail of an approaching extratropical cyclone (Fig. 9a). The aircraft then headed eastward to record data in the pre-cold-frontal environment over the coastal waters near BBY between 0445 and 0630 UTC 21 January (Figs. 9a,b). Large concentrations of CO measured by the P-3 were confined to shallow southeasterly flow in the lowest 250 m MSL within 50 km of the coast (east of the dashed line in Fig. 9b); the details of these observations will be presented below. Offshore-directed surface flow at BBY and buoy 46026 at 0530 UTC 21 January (Fig. 9b) reveal the occurrence of gap flow.
The wind profiler at BBY captured key characteristics of the gap flow at the coast (Fig. 10), including during the period when the P-3 was nearby. Between 0200 and 0300 UTC 21 January, weak easterly flow at the surface intensified and deepened to ∼220 m MSL, marking the onset of a cold and stably stratified airstream exiting the Central Valley via the Petaluma Gap. The persistent, cold, gap flow deepened further to ∼300 m MSL by 1500 UTC. Aloft, the flow transitioning to ambient southerly above ∼600–750 m MSL. Buoy 46026, located west of the Golden Gate Gap, documented offshore-directed flow similar to that observed at the surface at BBY (Fig. 10b), although the flow was rotated clockwise by ∼15° during the period of nearby P-3 observations. The buoy observations corroborate plan-view analyses (not shown) that portray cold air from the Central Valley also venting through the Golden Gate Gap, comparable to that observed during the 5–10 January 1998 gap-flow event.
The P-3 aircraft ascended or descended 7 times over the coastal waters between 0451 and 0614 UTC 21 January (Fig. 9b), 3 times in the high-CO airstream, and 4 times to its west. Representative flight-level profiles of CO, water vapor mixing ratio, θυ, and wind direction recorded during three ascents (0529, 0549, and 0614 UTC) along the northeast–southwest flight leg nearest BBY are presented in Fig. 11. The CO profiles contain background levels of 120–140 ppbv aloft, while the profile closest to the coast at 0529 UTC portrays large concentrations exceeding 200 ppbv below ∼250 m MSL (similar to that observed in the nearshore profiles at 0500 and 0533 UTC). The shallow layer of enhanced CO coincides with relatively dry water vapor mixing ratios ≤5.5 g kg−1, similar to values of ∼5 g kg−1 observed at BBY (Fig. 11b) and in the Central Valley (not shown). In contrast, moist conditions were found above the sharp CO gradient at 0529 UTC (≥6 g kg−1), at low levels farther offshore at 0549 and 0614 UTC (6.0–7.5 g kg−1), and at the sea surface (8.2 g kg−1). The concurrence of low moisture and high CO in the nearshore profile at 0529 UTC points to the recent continental origin of this shallow airstream prior to its westward extrusion through the Petaluma and/or Golden Gate Gaps. It is reasonable to assume that the primary source of CO loading was from a multitude of combustion sources (i.e., vehicles, factories, etc.; e.g. Graedel et al. 1986) in the Central Valley and San Francisco Bay area. The shallow, polluted, dry air also contained the most easterly or offshore-directed wind direction of the three airborne profiles (Fig. 11d) and lends support to the hypothesis of its Central Valley/San Francisco Bay origin. Enhanced static stability at ∼225 m MSL capped the high-CO air (Fig. 11c). This stable layer was also observed over BBY, and it extended westward to the unpolluted 0549 UTC profile (an elevated stable layer and embedded moisture gradient not related to the gap flow sloped upward to the west from 550–850 m MSL toward the approaching cold front). The polluted P-3 profile at 0529 UTC was warmer and slightly more moist than the BBY profile below ∼200 m, likely because the sea surface modified the cold, dry gap flow after it exited BBY (note that the surface easterly flow at BBY was much colder and drier than the adjacent sea surface). The four wind-direction profiles show the low-level flow rotating systematically from gap easterly (90°–120°) along the coast at BBY, to modified gap southeasterly (125°–140°) at 0529 UTC, to blocked southeasterly (140°–165°) at 0549 UTC, to ambient southerly (170°–180°) farthest offshore at 0614 UTC. The clockwise turning of the gap airstream as it encountered the southerly component flow in the coastal zone, in conjunction with the observations of polluted air in the shallow terrain-parallel southeasterly nearshore flow, indicate that gap flow quite likely did contribute to blocking along the coast. A similar rightward deflection of gap airstreams by ambient flow over the open ocean, and the subsequent contribution of this deflected flow to blocking along the coast, was documented previously in coastal Alaska by Loescher et al. (2006) and Colle et al. (2006).
The P-3 penetrated the offshore boundary of the high-CO airstream during a southwestward-directed flight leg at 220 m MSL between 0536 and 0547 UTC 21 January (Fig. 9b). Inspection of the flight-level time series trace of CO (Fig. 12a) reveals a gradual 20-km-wide gradient across this boundary, increasing from a background level of ∼120 ppbv farthest offshore to a mean value exceeding 200 ppbv in the polluted airstream originating from the interior. Companion traces of water vapor mixing ratio, θυ, zonal wind, vertical motion, and cloud liquid water (Figs. 12b–f) show no significant trend across the CO gradient zone, although comparatively dry mixing ratios with large easterly component flow and cloud-free conditions in the polluted air further highlight its continental origin. Significant undulations in CO on the shoreward side of the boundary are anticorrelated with θυ, thus implying the occurrence of vertical mixing within the capping stable layer at ∼225 m MSL. Gradient Ri calculations within this layer (based on the nearby P-3 profiles at 0549 and 0556 UTC) yielded values <0.25, which support this interpretation. The collocation of a weak vertical-motion couplet and a small spike in cloud liquid water on the offshore side of the boundary is indicative of weak convergence, although the other meteorological traces do not show prominent signatures indicative of convergence here or within the gradual CO gradient zone.
4. A multiwinter analysis of gap flow
In an effort to further understand the characteristics of flows through the Petaluma Gap and to assess factors that control these flows, a five-winter composite of gap-flow events is presented using data from the coastal wind profiler site at BBY and the nearby coastal-mountain surface site at CZD. Unlike the profiler data, the shallower RASS observations at BBY did not consistently extend upward through the gap flow. Hence, RASS was excluded from this analysis. Table 1 shows the operating periods at BBY, totaling 559 days during the five winters between November 1997 and March 2004. The CZD surface site collected data during 85% of those days.
a. Composite gap-flow characteristics from the BBY wind profiler
Using the case studies as a guide, we defined gap-flow events based on BBY data meeting the following criteria for at least six consecutive hours: 1) a surface wind direction of 50°–120°, 2) a layer-mean wind direction of 60°–130° between 100–350 m MSL (i.e., in the lowest three profiler range gates), and 3) an easterly component of at least 5 m s−1 in the same 100–350-m layer. A total of 62 events were observed (Table 2), ranging between 6 and 73 h in length, with a mean duration of ∼15 h. Each case exhibited an above-surface maximum of low-level easterly component flow (hereafter referred to as a gap jet) beneath the top of the gap walls, with the mean altitude of this maximum residing at 226 m MSL. The mean magnitude of the gap jet is 9.1 m s−1 and is bounded by cases as strong as 16.2 m s−1 and as weak as 6.0 m s−1. To ascertain the mean vertical wind shear characteristics above the gap jet, the strength of the zonal shear for each case was first calculated across the two adjacent profiler range gates surrounding the altitude at which the easterly component was 75% of the value of the gap jet beneath. The mean altitude and magnitude of the capping zonal shear layer is 556 m MSL and 16.2 m s−1 km−1, respectively. The altitude of the strong shear above the gap jet approximately coincides with the top of the shallow cold air exiting the gap in the case studies. It also generally matches the mean depth of the cold air trapped in the Central Valley during the long-lived Central Valley cold-pool event summarized in Fig. 2. During that event, the variation of the top of the cold air approximately matched the altitude range of the capping shear layer for the 62 cases (i.e., between 256 and 1119 m MSL). Most of the 62 cases were rain free at both BBY and CZD. Also, only 2.5% of the five-winter rainfall at BBY occurred during gap-flow cases, even though these cases occurred, on average, during 7.0% of the five-winter operating period (Table 1; similar behavior was observed at CZD). Hence, the composite results support conclusions drawn from the case studies that gap flow inhibits rainfall near its exit at the coast.
The mean vertical structure of the flow through (and above) the Petaluma Gap was evaluated by averaging all 937 hourly wind profiles from the 62 gap-flow cases into composite profiles of total wind speed, wind direction, and zonal and meridional wind components (Fig. 13). The composite wind speed and direction profiles portray a gap jet (9.6 ± 2.8 m s−1) directed from the east-southeast (i.e., from the Petaluma Gap) centered in the lowest two range gates (i.e., 125–225 m MSL). The magnitude of the flow decreases significantly across a prominent directional shear layer at ∼0.5 km MSL and remains weaker than the gap jet upward through 2 km MSL. Clockwise turning of winds with height from southeasterly to southerly above the gap flow suggests the presence of pre-cold-frontal warm advection, consistent with the case-study findings. The 5.7 m s−1 easterly surface flow at BBY is twice as strong as that observed in the nearby coastal mountains at CZD. Inspection of the zonal wind profile reveals a well-defined easterly gap-jet signature (9.4 ± 2.2 m s−1) in the lowest two range gates. In contrast, the meridional wind profile shows a steady increase with height of the magnitude of the southerly component flow (from nearly 0 to more than 8 m s−1) in the lowest 1.5 km MSL.
The 62 gap-flow events were stratified into two groups based on the strength of the easterly component flow in the core of the gap jet: ≥10 m s−1 and <8 m s−1. This simple categorization yields 17 strong cases and 18 weak cases, which corresponds roughly to the top and bottom quartiles of the 62-case sample (Table 2). The mean duration of the strong cases is 23.4 h, fully 2.5 times longer than the weak cases. This difference is statistically significant at a >99% confidence level, based on a two-sided t test (Mendenhall et al. 1981). The east-southeasterly flow in the gap jet is 8.5° more southerly for the strong cases and is statistically significant with >92% confidence. The altitude of this jet is somewhat higher for the strong cases (263 versus 217 m MSL), although the statistical significance of this difference is only 68.6%. The difference in altitude of the capping shear layer between strong and weak cases is even less substantial: 594 versus 562 m MSL, with a statistical significance of 33.8%. The magnitude of the shear within this layer is 17% larger for the strong cases (16.6 versus 14.2 m s−1 km−1), and the statistical significance of this difference is a modest 75.2%. These results indicate that the depth of the gap flow is not modulated greatly by its strength. Interestingly, however, the strong gap-flow events are associated with considerably more rainfall and larger rain rates at BBY than the weak events, with the acknowledgment that even the strong cases produce only modest rain. This difference in rain behavior (statistically significant with >95% confidence) will be addressed in section 4b. The relationship between gap jet strength at BBY and rainfall at CZD is less robust.
Composite wind profiles were constructed from the 17 strongest and 18 weakest gap-flow cases (Fig. 13). A total of 397 and 166 hourly wind profiles compose the strong and weak composites, respectively. Wind speeds in the strong composite are uniformly larger (by 2–8 m s−1) through 2 km MSL, thus implying that stronger gap-flow events are associated with stronger synoptic-scale flow patterns, both within and above the influence of the Petaluma Gap. Wind directions in the strong composite are uniformly more southerly below 1.5 km MSL. The easterly surface flow at BBY is stronger than that observed in the nearby coastal mountains at CZD for both strong and weak events. The primary difference in the composite zonal wind speed profiles between the strong and weak cases resides below ∼800 m MSL (i.e., within the confines of the gap), where the strong composite possesses easterly component flow as much as 4 m s−1 greater than its weak counterpart. In contrast, the chief difference in the composite meridional profiles is above 500 m MSL (i.e., above the gap), where the southerly component is as much as 7 m s−1 larger for the strong composite. The zonal and meridional differences will be placed into synoptic-scale context in section 4b.
b. NCEP–NCAR composite synoptic-scale analysis
The mean large-scale conditions responsible for generating the 62 gap-flow events at BBY were gauged by constructing composite synoptic-scale analyses using the daily gridded NCEP–NCAR reanalysis dataset. Daily gridded data for the day of each gap-flow event were included in the synoptic composites, so long as the gap flow occurred for at least 6 h on that calendar day. This methodology yielded 69 dates that contributed to the synoptic composites (Fig. 14). At 500 mb (Fig. 14a), a ridge axis is situated over Nevada and a trough is approaching the West Coast. This ridge–trough couplet is progressive, as revealed by a sequence of daily composite analyses spanning 2 days on each side of the gap-flow day (not shown). On the day preceding the gap flow, the ridge resided over California, which would be conducive to generating light-wind conditions, clear skies, and a subsequent pooling of cold air in the Central Valley through nocturnal radiational processes. Two days later, the trough made landfall across California (not shown). The companion sea level pressure analysis during gap-flow conditions (Fig. 14b) shows a trough approaching the West Coast, an anticyclone over the Great Basin, and a westward-directed pressure gradient along the Petaluma Gap (consistent with the case-study results). Offshore, the southerly geostrophic flow extends northward from California to British Columbia, and this large-scale flow should divert the cold air exiting the Petaluma gap northward up the coast in the form of a blocked airstream trapped against the coastal mountains [comparable to that observed along the Alaska coast by Loescher et al. (2006) and Colle et al. (2006)]. The composite 925-mb temperature analysis (Fig. 14c) reveals the presence of cold conditions over the Great Basin, a tongue of offshore warm air paralleling the coast, and enhanced southwest–northeast-oriented baroclinicity in the vicinity of the land-falling trough that is suggestive of an approaching cold front (consistent with the case-study results).
To explore the relationship between the strength of the gap-flow events observed at BBY and synoptic conditions, composite synoptic NCEP–NCAR analyses were created using the dates when the 17 strongest and 18 weakest gap-flow events occurred, so long as the gap flow occurred for a minimum of 6 h on that calendar day. A total of 23 and 17 dates were included in the synoptic composites for strong and weak gap-flow events, respectively. The synoptic composite mean fields are accompanied by their anomaly counterparts, which are obtained by subtracting the long-term (1968–96) averages composed of specified dates from the composite mean fields based on those same dates.
The composite 500-mb geopotential height analyses for both the strong and weak gap-flow cases (Fig. 15) portray a land-falling trough and an inland ridge. However, the offshore trough is considerably deeper in the strong gap-flow composites while the inland ridge is more amplified in the weak composites. Furthermore, the mean and anomaly values over the Petaluma Gap are much lower for the strong cases than the weak cases (5617 and −30 m versus 5685 and +38 m, respectively). These results demonstrate that strong gap-flow cases are associated with comparatively deep midtropospheric troughs approaching the coast, while the weak cases are tied to anticyclonic conditions aloft. They also provide an explanation for the much weaker winds observed aloft at the BBY wind profiler during weak gap-flow events (Fig. 13).
Companion composites of sea level pressure (Fig. 16) depict a much deeper land-falling trough situated closer to the coast during the strong gap-flow cases and a modestly stronger ridge over the Great Basin for the weak cases. In both regimes the anomaly at the Petaluma Gap is negative but much greater for the strong than weak cases (−5.2 versus −1.0 mb, respectively). The westward-directed pressure gradient is at least twice as large for the strong gap-flow cases, thus revealing a dynamically consist signal with the mesoscale wind profiler observations. Similarly, stronger southerly geostrophic flow at the surface and aloft (upward to at least 850 mb, not shown) near the Petaluma Gap for the strong cases is dynamically consistent with the stronger southerly flow observed above the gap by the BBY wind profiler for that regime (Fig. 13d).
The 925-mb temperature composites (Fig. 17) show a north–south thermal ridge axis and a positive anomaly of 2°–3°C just offshore in both regimes. However, the air farther east is colder than normal across the entire Intermountain West only for the strong gap-flow cases. This would suggest that the temperature of the continental-origin air modulates the strength of the gap flow in a mean sense, particularly if the cold air is sufficiently deep to spill over the Sierras into the Central Valley. Offshore, the southwest–northeast-oriented baroclinic zone is stronger and more organized for the strong gap-flow cases, implying that these cases are associated with the approach of stronger maritime cold fronts.
Composite rainfall analyses (Fig. 18) exhibit remarkable consistency with the rainfall observed at BBY; namely, substantially more rain falls during strong gap-flow events. In both regimes, the core region of rain is situated offshore in the vicinity of the advancing cold front, but the rain reaches the coast at the Petaluma Gap only during the strong gap-flow events. In addition, this core region possesses a larger value for the strong cases, consistent with the interpretation that the strong gap-flow events are associated with the approach of stronger maritime cold fronts. On the day after the gap-flow events, rainfall continues over the Petaluma Gap for the strong cases but remains north of the area for the weak cases. In effect, strong gap-flow cases, dominated by cyclonic conditions, tend to be wet relative to weak gap-flow cases, which are dominated by anticyclonic conditions.
Data collected from experimental observing platforms deployed during the CALJET winter of 1997/98 and the PACJET winters of 2000/01 through 2003/04 were combined with observations from operational networks to document the regular occurrence and characteristic structure of shallow (400–500 m deep), cold airstreams flowing westward through California’s Petaluma Gap from the Central Valley to the coast. The NOAA/ETL 915-MHz coastal wind profiler at BBY was the cornerstone observing system used in this study. Two unique supplementary data platforms were also utilized: the NOAA/ETL 915-MHz island wind profiler at FNI and the NOAA/AL carbon monoxide measuring system on the NOAA P-3 research aircraft.
The Petaluma Gap is ∼35–50 km wide and has walls extending, at most, a modest 600–900 m above the valley floor (Fig. 1). Based on this geometry, together with meteorological conditions typical of the region, the Petaluma Gap is predisposed to generating ageostrophic flows driven by along-gap pressure differences. Gap flows here are generally less pronounced than those observed farther north along the coast (i.e., Oregon and northward) during winter, largely because the primary source region of cold air is California’s Central Valley rather than the northern interior of North America where arctic air is ubiquitous, and also because the Petaluma Gap is not as well defined as its northern counterparts. Nevertheless, we have shown with case studies and a multiwinter analysis that modest gap flows significantly impact local distributions of wind, temperature, precipitation, and atmospheric pollutants, and that they can contribute to the generation of coastally trapped airstreams.
Two case studies and a five-winter composite analysis of 62 gap-flow cases were used to create plan-view and cross-section conceptual representations of gap flow exiting the Central Valley during the cool season (Fig. 19). The gap airstream was characterized by cold, easterly flow and an embedded jet capped by a layer of enhanced static stability and vertical wind shear below the top of the gap walls across AA′ (Fig. 19a), thus suggesting the importance of this constriction in generating the gap flow. Based on the 62-case composite of wind profiles from BBY, the mean strength of the gap jet exceeded 9 m s−1, resided at a mean altitude of ∼225 m MSL, and varied in altitude (from case to case) between 105 and 741 m MSL. The mean altitude of the capping shear layer (which averaged 16.2 m s−1 km−1) was 556 m MSL, which is beneath the top of the gap walls. Case-study results reveal that frictional effects extended upward to at least 500 m MSL, which is above the mean altitude of the gap jet. Surface observations and P-3 flight-level data documented comparatively dry, polluted characteristics associated with the gap airstream in the easterly flow below ∼250 m MSL, further demonstrating the continental origin of this shallow airstream. Hence, it is not surprising that the 62 gap-flow cases were, on average, nearly 3 times drier than that expected at BBY independent of the flow regime.
The 62 gap-flow events observed at BBY were stratified into two groups based on the strength of the easterly component flow in the core of the gap jet: ≥10 m s−1 and <8 m s−1. This yielded 17 strong cases and 18 weak cases. The strong cases persisted 2.5 times longer than the weak ones. The altitude of the gap jet and the capping shear layer were only modestly higher (by 46 and 32 m MSL, respectively) for strong cases, thus revealing that the depth of the gap flow was not greatly modulated by its strength. Nevertheless, the strong gap-flow events were associated with considerably more rainfall and larger rain rates at BBY than the weak events, with the acknowledgment that even the strong cases produced only modest rain. Composite profiles of the total and meridional wind speed were uniformly larger (by 2–8 m s−1) through 2 km MSL for the strong cases, thus implying that stronger gap-flow events were linked to stronger synoptic-scale flow patterns, including above the influence of the Petaluma Gap. Composite synoptic analyses for strong and weak gap-flow events were constructed from the NCEP–NCAR global gridded reanalysis dataset to explore the relationship between the strength of the gap flow at BBY and synoptic conditions. Strong gap-flow cases are associated with comparatively deep midtropospheric troughs approaching the coast, while the weak cases are tied to anticyclonic conditions aloft. Similarly, low-level cold-frontal pressure troughs approaching the coast are stronger and possess a greater along-gap pressure gradient for the strong gap-flow cases. These results provide dynamical context for the stronger low-level gap flow and winds aloft, and the greater rainfall, observed at BBY during strong gap-flow events.
Because the Petaluma Gap is the only major air shed outlet from the Central Valley, it is possible to perform a simple mass budget calculation of the shallow cold pool based on the conceptual results portrayed in Fig. 19b and the geometry of the Central Valley. The Central Valley is approximately 80 km wide by 600 km long, and a typical cold-pool depth is 500 m. Hence, the total volume of the trapped cold pool is 24 000 km3. The gap flow exiting the initial constriction through AA′ in Fig. 19b is ∼8 m s−1 in a 400-m-deep and ∼40-km-wide layer, which yields a mass flux of ∼11 000 km3 day−1. Therefore, it would take more than 2 days to evacuate the shallow cold pool through the Petaluma Gap, assuming vertical mixing is negligible. Because the cold air is replenished nightly through radiational cooling and katabatic flows (assuming steady-state synoptic conditions), the gap flow by itself cannot entirely remove the cold pool from the Central Valley, at least during the winter when the cold pool is generated and maintained locally through the aforementioned radiational processes. This simple mass budget calculation supports the assertion by Holets and Swanson (1981) that the trapped cold pool will persist until a major change in the synoptic pattern occurs.
The plan-view conceptual perspective (Fig. 19a) portrays the occurrence of gap flow through the Petaluma Gap and adjacent Golden Gate Gap during pre-cold-frontal conditions. The gap flows were tied to the pre-cold-frontal environment, largely because sea level pressure decreased westward along the gap in a stably stratified atmosphere in advance of the approaching cold-frontal pressure trough. Airborne air-chemistry observations offshore of BBY from the NOAA P-3 aircraft documented high concentrations of carbon monoxide in the gap flow, thus highlighting the fact that this airstream was influenced by the polluted environments of the Central Valley and San Francisco Bay area. These air-chemistry measurements, in tandem with the standard suite of meteorological observations, also provided evidence that the gap flow was deflected northwestward along the coast, and that this deflected airstream contributed to downstream blocking in this case.
The Petaluma Gap is one of several natural conduits that allow cold and/or dry continental air to enter California’s coastal zone from the state’s interior (Fig. 19a). This study also documented the occurrence of a cold gap flow exiting the Salinas Valley south of the San Francisco Bay area. The southernmost, multifingered, gap-flow airstream portrayed in Fig. 19a is based on results from a mesoscale observational study of a strong land-falling storm impacting southern California (Neiman et al. 2004), which revealed that the gap flow commenced in pre-cold-frontal conditions and contributed to downstream blocking, similar to that shown in the current study farther north. An analogous link between gap flows and coastally trapped barrier jets was documented along the Alaskan coast by Loescher et al. (2006) and Colle et al. (2006). Finally, a recent study by Ralph et al. (2005) presented composite dropsonde analyses that provided documentation of the conditions in the pre-cold-frontal environment of extratropical cyclones over the northeastern Pacific Ocean (i.e., offshore of California), including moist neutral stratification in the lowest ∼2.8 km MSL. An important implication of the moist neutrality is that the flow should readily ascend California’s relatively shallow coastal ranges without inducing blocking. However, the fact that blocking is commonly observed along the California coast (e.g., Neiman et al. 2002, 2004) indicates that an alternative process must be active, such as offshore advection of cool continental air through gaps in the coastal ranges.
This study has focused on the structural characteristics of flow through the Petaluma Gap in central California, and it has provided an observational foundation by which weather forecasters can expect gap flow to occur here and to anticipate its local impacts. It has also shown a connection between gap flows and coastal blocking. However, because coastal blocking arguably has an even more profound and widespread impact than gap flows on modulating precipitation distributions and land-falling frontal evolutions, it is important to more precisely assess the relationship between these terrain-modulated flows for forecasting considerations. Through the combined future efforts of numerical modeling and additional data gathering and analysis, this relationship can be deciphered unambiguously.
The successful implementation of the CALJET and PACJET programs was made possible by the dedicated participation of many individuals from numerous organizations. Special thanks are given to Clark King and the rest of the NOAA/ETL talented field and engineering teams for deploying and maintaining the instrumentation. We are grateful to the University of California, Davis, Bodega Marine Laboratory and to Robert Mann for granting us permission to deploy and operate the ETL instrumentation on a long-term basis at Bodega Bay and Cazadero, respectively. Likewise, we are indebted to the staff at the Farallon National Wildlife Refuge for providing access for our wind-profiler deployment at this unique location, and to the U.S. Coast Guard for transporting our equipment and personnel to and from this site. Ola Persson of NOAA/ETL provided the airborne radar analysis in Fig. 3b. We appreciate the thoughtful input from David Kingsmill and Bob Zamora of NOAA/ETL, and two anonymous reviewers. Their efforts improved the scope and quality of this manuscript. Jim Adams furnished top-quality drafting services. This research was funded with the support of the Climate–Weather Connection Program at NOAA. The experimental data collection was supported, in part, by the U.S. Weather Research Program.
Corresponding author address: Paul J. Neiman, NOAA/Environmental Technology Laboratory, Mail Code R/ET7, 325 Broadway, Boulder, CO 80305. Email: email@example.com
The profiler at the ideally situated, but difficult to reach, FNI was deployed in December 1996. Unfortunately, this profiler failed during the third week of January 1998. The profilers at CZD and PAA recorded data during the entire winter of 1997/98 (CALJET), while the BBY profiler operated during all five winters.