Observed Dynamics of Coastal Flow at Cape Mendocino during Coastal Waves 1996

Linda Ström Department of Earth Sciences, Meteorology, Uppsala University, Uppsala, Sweden

Search for other papers by Linda Ström in
Current site
Google Scholar
PubMed
Close
,
Michael Tjernström Department of Meteorology, Stockholm University, Stockholm, Sweden

Search for other papers by Michael Tjernström in
Current site
Google Scholar
PubMed
Close
, and
David P. Rogers Scripps Institution of Oceanography, La Jolla, California

Search for other papers by David P. Rogers in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Airborne in situ and remote meteorological measurements from around Cape Mendocino, California, sampled during the Coastal Waves 1996 field program are analyzed for three different days: 7, 12, and 26 June 1996. Two days conformed to typical summertime conditions, with a strong northerly downcoast flow, while the wind on the third day was weaker. On the first 2 days, the flow was supercritical in the sense that the Froude number was larger than unity and these 2 days feature expansion fans in the lee of the cape. On the third day, no such phenomenon was observed. All 3 days had a strong thermal wind caused by the marine inversion sloping down toward the coast. On the first 2 days, the flow aloft was westerly or northerly, so that the thermal wind added to the background flow results in a strong jet. On the third day the flow aloft was southerly, and consequently even with the added thermal wind, the northerly flow in the marine layer was too weak to be supercritical. The main difference between the first 2 days was the fact that 7 June is cloud free, while 12 June had a stratocumulus cover.

The 3 days are analyzed as composites, and several scales are identified and described: 1) the large-scale synoptic forcing, determined from the wind aloft and synoptic conditions; 2) blocking by the coastal topography, the thermal wind balance in the near-coast zone, and nondimensional properties; 3) the hydraulic properties in the sub- and supercritical flows as they pass the change in coastline orientation at Cape Mendocino; 4) the impact of the local blocking by the terrain at the cape itself, generating a lee-wave phenomena for the high-wind days, which was the actual cause of the collapse of the marine layer in the lee of the cape; 5) the boundary layer interaction, apparently generating SST anomalies, but with little or no feedback to the wind field.

A momentum budget analysis for the two high-wind days show a significant difference between one day (7 June) when mesoscale perturbation dominated the flow and the other day (12 June), when large-scale forcing dominated and the mesoscale perturbation was smaller. The importance of the cloud layer on 12 June is illustrated, using lidar data.

Corresponding author address: Michael Tjernström, Department of Meteorology, Stockholm University, Arrhenius Laboratory, S-106 91 Stockholm, Sweden.

Email: michaelt@misu.su.se

Abstract

Airborne in situ and remote meteorological measurements from around Cape Mendocino, California, sampled during the Coastal Waves 1996 field program are analyzed for three different days: 7, 12, and 26 June 1996. Two days conformed to typical summertime conditions, with a strong northerly downcoast flow, while the wind on the third day was weaker. On the first 2 days, the flow was supercritical in the sense that the Froude number was larger than unity and these 2 days feature expansion fans in the lee of the cape. On the third day, no such phenomenon was observed. All 3 days had a strong thermal wind caused by the marine inversion sloping down toward the coast. On the first 2 days, the flow aloft was westerly or northerly, so that the thermal wind added to the background flow results in a strong jet. On the third day the flow aloft was southerly, and consequently even with the added thermal wind, the northerly flow in the marine layer was too weak to be supercritical. The main difference between the first 2 days was the fact that 7 June is cloud free, while 12 June had a stratocumulus cover.

The 3 days are analyzed as composites, and several scales are identified and described: 1) the large-scale synoptic forcing, determined from the wind aloft and synoptic conditions; 2) blocking by the coastal topography, the thermal wind balance in the near-coast zone, and nondimensional properties; 3) the hydraulic properties in the sub- and supercritical flows as they pass the change in coastline orientation at Cape Mendocino; 4) the impact of the local blocking by the terrain at the cape itself, generating a lee-wave phenomena for the high-wind days, which was the actual cause of the collapse of the marine layer in the lee of the cape; 5) the boundary layer interaction, apparently generating SST anomalies, but with little or no feedback to the wind field.

A momentum budget analysis for the two high-wind days show a significant difference between one day (7 June) when mesoscale perturbation dominated the flow and the other day (12 June), when large-scale forcing dominated and the mesoscale perturbation was smaller. The importance of the cloud layer on 12 June is illustrated, using lidar data.

Corresponding author address: Michael Tjernström, Department of Meteorology, Stockholm University, Arrhenius Laboratory, S-106 91 Stockholm, Sweden.

Email: michaelt@misu.su.se

1. Introduction

The marine atmosphere in coastal areas and its response to surface forcing, for example, a step-change in surface characteristics or the effect of coastal topography, is not well understood. Even for relatively well understood coastal weather phenomena such as the sea breeze, the offshore extent and structure are not well known (Banta et al. 1993; Banta 1995). More complex situations, such as coastal jets, expansion fans, and hydraulic jumps (e.g., Winant et al. 1988; Tjernström and Grisogono 2000), and the so-called coastal wind reversals (Bond et al. 1996; Dorman 1997), are only now becoming better understood. One main reason for the lack of insight is a lack of offshore measurements with an adequate spatial and temporal resolution. Thus, the current progress is linked to the results from recent field experiments.

The coastal atmospheric boundary layer (MBL) is generally heterogeneous. Advective effects are important and steady-state conditions seldom prevail (e.g., Beardsley et al. 1987; Brost et al. 1982a,b). The U.S. west coast is one example of a region where significant coastal mountains result in strong along-coast winds and relatively cold coastal ocean temperatures. For much of this coast, the height of the coastal mountains exceeds 400 m thus forming an almost continuous barrier from Oregon to southern California. Several experimental studies, for example, the comprehensive interdisciplinary Coastal Ocean Dynamics Experiment (CODE 1 and 2; Beardsley et al. 1987), concluded that ocean dynamics are strongly linked to the local atmospheric forcing while the latter is primarily caused by the interaction between synoptic-scale conditions set up by the North Pacific subtropical high and local topographical effects. Other similar regions, mostly located on the west side of continents at low or middle latitudes, are discussed in Winant et al. (1988).

The large-scale synoptic situation for coastal California is dominated by the Pacific high pressure system. Temporal variations in the strength and position of this system govern the annual cycle in the synoptic conditions along the U.S. west coast (Halliwell and Allen 1987). During summer, the high is more intense and situated more to the north than during the winter. This cycle leads to a higher frequency of, and a larger influence by, synoptic-scale low pressure disturbances during the winter season. In summer, a thermal low develops over the continent, causing the synoptic-scale pressure field to exhibit a significant across-coast pressure gradient. Thus, the summer MBL along much of the California coast is characterized by a strong and persistent north-northwesterly flow aligned with the coastal topography (Halliwell and Allen 1987; Kelly 1985; Dorman and Winant 1995). At local irregularities in the coastal topography, such as local headlands, mesoscale wind and pressure structures are superimposed on and interact with this background flow.

During spring and summer, the moist and cool MBL air is separated from the warm and dry free troposphere air by a sharp and well-defined subsidence inversion, which is also a manifestation of the Pacific High (e.g., Beardsley et al. 1987). Dorman and Winant (1995) found this inversion to be present more than 90% of the time, in contrast to winter conditions when it occurred only ∼50% of the time. The depth of the MBL, under the inversion, is typically lower than most of the coastal topography and the inversion thus intersects the mountain barrier. This efficiently blocks cross-coast flow, except for in gaps in the coastal terrain, and generates conditions that are favorable for coastally trapped mesoscale responses in the MBL (Gill 1977). The inversion can be persistent for long periods, maintained from above by subsidence and from below by turbulent mixing, and the potential temperature difference is typically Δϒ = 10–20 K (Dorman and Winant 1995). The density of the cooler MBL air is thus typically 5% higher than in the air above the inversion, and the flow has often been treated as a single-layer reduced-gravity flow with a free upper surface (e.g., Winant et al. 1988; Samelson 1992; Samelson and Lentz 1994). An overview of such flows past a varying sidewall can be found in Baines (1995), for example. One main feature is the sensitivity to irregularities in the coastline that appear for supercritical conditions, which is shown below. Expansion fans and hydraulic jumps due to this mechanism were encountered during CODE and have been discussed in the literature (Winant et al. 1988; Samelson 1992; Tjernström and Grisogono 2000; Tjernström 1999). Extensive and detailed experimental documentation has, however, been limited to a very few cases.

Overland (1984) argued that the coastally trapped flow is semigeostrophic; geostrophic in the cross-coast component, while in the alongcoast component the pressure gradient is balanced either by the ageostrophic acceleration, as in “gap winds,” or by surface friction. When the local coast deviates from being parallel to the pressure field, for example, in the lee of headlands, the flow often becomes entirely ageostrophic (e.g., Overland 1984). This has been confirmed by measurements (Winant et al. 1988; Zemba and Friehe 1987; Samelson and Lentz 1994) and by model simulations (Cui et al. 1998; Tjernström and Grisogono 2000).

The strong tendency for the flow to align to the coast was also investigated by Cui et al. (1998). They found that for the central coast of California, synoptic-scale flow in a sector from the northwest through north to southeast generated a consistently along-coast MBL flow from the northwest. In contrast, synoptic-scale flow between the south and the west consistently generated along-coast flow from the southeast. Channeling and pressure accelerations are also reported from other locations (Overland 1984; Lackman and Overland 1989; Smedman et al. 1996; Smedman and Bergström 1995) and seem to be inherent to the mesoscale climate in mountainous or coastal areas.

CODE results indicate that the ocean response to the mesoscale atmospheric variability is manifest in coherent sea surface temperature (SST) variations, which are due to upwelling. The mean wind stress along the coast, resulting from the down-coast equatorward flow, forces surface water offshore by Ekman transport, which is replaced by upwelling of cold nutrient-rich water from below. Local horizontal divergence of the wind stress enhances this process (Kelly 1985; Davis and Bogden 1989; Enriques and Friehe 1996; Tjernström and Grisogono 2000).

The cool surface along the coast in combination with the capping inversion often causes clouds to form in the MBL. Fog, stratus, or stratocumulus cloud cover is thus typical for the California MBL during summer, generally with stratocumulus west of 125°W and stratus or fog to the east (Bridger et al. 1993). Farther offshore, the clouds often become convective, either in cellular patterns or in horizontal rolls (Brooks and Rogers 1997). The presence of these clouds is important to the coastal MBL since they affect the depth of the MBL. In general, the increased turbulent mixing driven by cloud-top buoyancy deepens the MBL, compared to the cloud-free case (Brost et al. 1982a,b; Duynkerke and Driedonks 1987; Tjernström and Koracin 1995; Tjernström and Rogers 1996). However, the direct effect of the cloud-top convection on the surface-layer turbulence is generally small and here the main source for the turbulence is the vertical wind shear.

Until quite recently the coastal MBL dynamics had been poorly observed, mostly because of the logistic problems of performing high quality measurements over the ocean. The Coastal Waves program from 1996 (CW96) was conceived and planned to address this problem (Rogers et al. 1998). The CW96 program builds on and expands the knowledge obtained from CODE; in particular, regarding the mesoscale structure of the MBL in the lee of significant headlands. CW96 thus provides one of a very few datasets with quasi-three-dimensional data from extensive aircraft measurements. The measurements targeted the flow around several major headlands along the California coast, from Point Conception in the south to Point Blanco in the north. In particular, supercritical flow conditions were investigated. Especially useful are repeated visits to some of these locations on different days, with different conditions. This provides an opportunity to increase the understanding of these coastal flows by systematic analysis on different scales, horizontal as well as vertical.

The present paper considers the observed mesoscale features in the MBL around Cape Mendocino, which was one of the target areas for CW96. Data from flights on three different days obtained with the NCAR1 C-130 Hercules research aircraft are analyzed. Section 2 contains a brief discussion of the experiment and of the analysis while section 3 summarizes the results and section 4 and 5 include a discussion and conclusions, respectively.

2. The field experiment

The Coastal Waves 1996 program is described in detail by Rogers et al. (1998). Dorman et al. (2000) provide a climatology for the experiment, while extensive details from aircraft measurements in the vicinity of Point Sur are presented in Dorman et al. (1999) and in the vicinity of Cape Mendocino are described in Dorman and Rogers (1998, unpublished manuscript). Several model studies of the flow around Cape Mendocino and Point Sur have also be completed (Tjernström and Grisogono 2000; Burk and Haack 2000).

The CW96 program consisted of a longer effort from May to September 1996 to obtain reliable average conditions and also to capture the rare events of so-called“southerly surges” (Bond et al. 1996; Dorman 1997), with an intensive observing period during the month of June. Automated surface weather stations, drifting and fixed buoys (including the National Data Buoy Center buoy system), wind profilers, acoustic sounders, and radio soundings were available along the coast from Southern California to the Oregon coast throughout this period, with airborne measurements obtained from the UNC Piper Seneca III research aircraft (Bane et al. 1995).

During the intensive observing period, the primary measurement platform was the NCAR C-130 Hercules research aircraft. The standard suite of data from this platform includes winds, temperature, and humidity; cloud liquid water; droplet and aerosol spectra; atmospheric radiation; and radiometric surface temperature. Most data are available at a low rate (1 Hz), while wind speed, temperature, and humidity are also sampled at a higher rate (25 Hz), for calculation of turbulence moments by the eddy correlation technique. In addition, remote sensing of the MBL structure was provided by the NCAR-developed Scanning Aerosol Backscatter Lidar (SABL). See Rogers et al. (1998) for a complete list of aircraft instrumentation.

From 2 June to 1 July, the C-130 flew 11 missions, mostly focusing on major headlands: four in the vicinity of Point Sur, three at Cape Mendocino, three at Point Conception, and one at Cape Blanco. A typical research flight results in 6 h of measurements within a typical domain of 150 km × 150 km × 1 km around the cape in question. This includes several stacks of horizontal flight legs (typically at 30, 100, and 1000 m) and series of sawtooth profiles where the aircraft ascends or descends between 30 and 1000 m at 2–4 transects across and 1–2 transects along the coast.

Here we consider the observations from the three flights near Cape Mendocino, which is located at ∼40.2°N. This area is illustrated in Fig. 1, which shows the main coastal terrain in this region, that forms a solid barrier to heights greater than 1 km to the north while slowly decreasing in height farther south. Note also the terrain at the headland, which is quite significant. It consists of two relatively isolated ridges, approximately ∼600 m high, oriented west to east. In contrast to the main upstream coastal mountains, these ridges are oriented roughly perpendicular to the flow. The cape protrudes about 40–60 km out into the ocean. The three flights in this area (on 7, 12, and 26 June) are illustrated in Fig. 2, showing the flight paths on each day. Most flight tracks were flown perpendicular to the coast (three, four, and three stacks during these three flights, respectively), while a few stacks were also flown along the coast. On 7 June, extensive high-altitude flight legs at 3000 m, for simultaneous dropsonde releases and lidar measurements, limited the low-level endurance. The flight on 12 June also includes one stack extending a longer distance offshore.

One shortcoming of aircraft measurements is the inherent inability to distinguish between spatial and temporal variability. This implies that some averaging is required, which often involves an assumption of stationarity for portions of the flight, or in some cases for a whole flight. In particular, it is sometimes necessary to know the aircraft’s vertical position with respect to the inversion, to make possible an analysis of measurements made along a single flight leg at a given height within the MBL. This is the case for the cross-coast stacks here, since the MBL depth varies with distance from the coast. Since the SABL lidar could be directed to view both up and down, it was possible to detect the lidar backscatter and thus, in most cases, the position of the inversion base simultaneously with the in situ airborne measurements. Problems in extracting the height of the inversion arise from conditions when the MBL is dry (low signal to noise), when flying to close to the inversion (i.e., within 100–200 m), and when the signal becomes attenuated by reflectivity saturation in clouds. For most of the stacks, the MBL depth is retrievable using a combination of upward- and downward-looking lidar data from the lowest and highest flight leg in a stack of flight legs. The standard sampling of the lidar data in CW96 is 3.75 m in the vertical and 100 m in the horizontal; in this paper, however, the MBL depths are averaged over horizontal lengths of 10 km. The high density of sawtooth slant profiles can be used to verify instantaneous lidar-inferred inversion height estimates; however, due to their slant character, the vertical MBL structure from these profiles is sometimes contaminated by horizontal variability, in addition to slightly overrepresenting the sharpness of vertical gradients (e.g., Brost et al. 1982a). This problem is especially severe close to the coast, where this variability is large. These profiles are nevertheless useful in determining the vertical structure when combined with the lidar data.

In general, only horizontal flight legs flown at 30 and 100 m are used here for turbulence calculations, using the eddy correlation technique. Additional levels are available only for some stacks on some flights and at some locations, and some additional levels are included in the analysis. The data from the lowest level (30 m) were collected to specify the surface-layer turbulence. All data from low-level horizontal flight legs was subdivided into 50% overlapping 10-km segments, within which the data were detrended and averaged. Turbulence statistics can then be calculated for each segment by eddy correlation. Analysis of Ogives (not shown) indicates that most of the turbulent fluxes is attributable to motions on scales less than a few kilometers. Closer to the coast, however, the situation is more heterogeneous and significant variability occur on larger scales, for example, low-level jet streaks and gravity wave activity. Larger scales appear more common as the altitude increases. This is particularly clear from flight legs at 1000 m. When estimating the momentum budgets,“subgrid scale” fluxes are thus representative for motions up to 5 km in scale (2Δx in the 10-km segments), while “resolved-scale motions” are inferred from the 10-km averages but at 5-km overlapping resolution. This is a practical compromise to a problem that really has no true solution.

The SST is measured as the surface skin temperature with a Heinmann radiometer. These measurements are corrected to account for attenuation and emission of longwave radiation in the atmosphere below the aircraft and for reflection at the surface. This correction is O (0.1°C), which is comparable to the measurement accuracy (∼±0.4°C). The observed SSTs during CW96 were quite low, often below the climatological values for this area.

3. Results

a. Introduction

The results in this section are discussed in the context of geophysical fluid dynamics theory, based on the highly simplified shallow water theory, see above. In that framework, a shallow well-mixed (constant density) fluid is capped by a jump in density. The offshore influence of the coastline is limited by the Rossby radius of deformation expressed as lR = cf−1, where c is the phase speed of the external (shallow water) gravity wave, which may propagate on the density jump (the MBL inversion). The phase speed is defined as c = (gh)1/2, where g′ is the reduced gravity and h is the depth of the fluid. The reduced gravity is defined as g′ = gΔϒ/ϒ, where ϒ is the potential temperature and Δϒ is its jump over the inversion, g is the acceleration of gravity, and f is the Coriolis parameter.

In this framework, the Froude number (Fr) is defined as the ratio between the velocity of the fluid and the phase speed of the shallow water gravity wave, thus Fr = U/c = U(gh)−1/2, where U is the velocity of the fluid. Supercritical flow is defined as Fr > 1, with the physical interpretation that information in the fluid borne by these gravity waves, for example, on a local perturbation in the pressure field, cannot propagate upstream. When the coast turns away from the flow (a “widening channel”), an expansion fan will form and the flow accelerates as the MBL depth decreases inside the fan. In the opposite case, when the coast turns into the flow (a “narrowing channel”), the flow will decelerate and can transit from super- to subcritical in a shock wave, a “hydraulic jump.” The angle between the outer limit of the expansion fan and the upstream coastline, α, is determined by the relationship between the cross flow propagation of the shallow water gravity wave and the velocity of the flow so that sin(α) = Fr−1 (Winant et al. 1988; Baines 1995). The exact structure of the expansion fan is modified by surface friction and the earth’s rotation (e.g., Samelson 1992; Tjernström and Grisogono 2000). Expansion fans and hydraulic jumps were also discussed and simulated in a shallow water model by Rogerson (1999). For both Fr and lR there are counterparts for a continuously stratified fluid, thus Fr = U(Nz)−1 and lR = Nhf−1, where N(z) is the local Brunt–Väisälä frequency and z is the height. The Brunt–Väisälä frequency is the frequency at which an air parcel displaced from its original altitude would oscillate due to the stratification, while being transported downstream by the main flow.

Defined the latter way, but using the height of the terrain instead of the MBL depth, Fr may also be used as an indicator of the degree of linearity in gravity waves induced by the airflow over the terrain. Other important nondimensional numbers for flow along a mountainous coastline involve the Burger number and “Nhu.” The Burger number is defined as Bu = hmN(lmf)−1, where hm and lm defines the height and the width of the terrain, with the physical interpretation that if Bu is large, the flow will not be able to flow across the blocking terrain either due to thermal effects (if N is large) or due to rotational effects (through f). Nhu, defined as NhmU−1, also has implications for terrain-induced blocking of the flow, where large numbers of Nhu [which lacks a proper name, Baines (1995)] indicate that the flow prefers to flow around, rather than over, an isolated terrain obstacle. These nondimensional numbers apply here both to the effects of the main coastal mountain barrier and to isolated terrain features along the coast. For example, see Baines (1995) for an overview of these and other nondimensional numbers.

In the following analysis, Cape Mendocino is used as a reference for north–south location; that is, “upstream” and “downstream” will always refer to north and south of the cape. The terms “offshore” and “inshore” or “coastal” will be used in a loose sense, with the former being tens of kilometers from the coast. With a strong northerly flow, 7 and 12 June correspond to the most frequent MBL conditions, according to the categorization of summertime flows by Dorman and Winant (1995), while 26 June has weaker winds and less topographic forcing. Clouds appear on all 3 days, but to varying degree (Fig. 3). On 7 June, most of the coastal MBL as well as the inland planetary boundary layer (PBL) are free from clouds while only a thin layer of stratocumulus appears far to the southwest. On 12 June, most of the coastal area is covered by what appears to be a single layer of stratocumulus, with the exception of a larger cloudfree area south of both Cape Blanco and south of Cape Mendocino. Note the intrusions of cloud marine air in valleys north of Cape Mendocino. On 26 June, the cloud pattern is much more complex with the appearance of several layers of different cloud types that covers different areas. This affects the vertical structure of the MBL.

b. Vertical structure

Figure 4 shows the average vertical structures of wind speed and direction, potential temperature, and specific humidity for all 3 days. All aircraft data north of 38°N were block averaged with respect to radar altitude. Included, as dotted lines are plus/minus one standard deviation, calculated in a similar manner. Note that the smaller standard deviation aloft may sometimes be an artifact of the limited data there, often including only one profile usually covering only a portion of the total domain. Thus, the data from the highest levels (>1 km) mostly comes from the southern part of the area, which corresponds with profiles obtained as the aircraft transit into and out of the region. In many cases, however, the variability aloft actually decreases with height.

From this crude analysis alone, some striking differences and similarities between the days are obvious. On the first 2 days, the MBL winds are significantly higher than the winds aloft, on average by a factor of three or more, while on the last day there is very little significant change in wind speed in the vertical, except for a drop in velocity around 1.4 km where the wind also changes direction. On the first 2 days, the flow is onshore above ∼2 km, while on the third day it reverses almost completely with altitude. The change in wind speed from the free troposphere to the MBL is more abrupt over the inversion on 7 than on 12 June. In terms of the variability of the vertical temperature structure, the two first days are similar; the conditions range from a well-mixed 4–600-m-deep MBL, to continuously stable stratification throughout the whole layer. The variability is significantly smaller on the last day. No well-mixed layer exists, rather the MBL appears weakly stable and much deeper, ∼1 km. On the first 2 days, the MBL is a well-defined cool and moist layer capped by an inversion, while on the last day, this distinction is less clear.

There are also differences between 7 and 12 June. The air aloft is more homogeneous on 7 June while the transition from the MBL to the free troposphere is deeper on 12 June, with a significant wind speed gradient between the MBL top and up to ∼2 km. One particular, possibly very important feature is observed in both the temperature and moisture profiles. The inversion layer has a depth that is similar to, or on 12 June even deeper than, the MBL depth, ∼500 m and ∼700–1000 m, respectively. This seems to suggest that the concept of a two-layer fluid (or a “single-layer reduced-gravity” fluid) may be inappropriate, and that a three-layer model should be applied. Some of this variability may be attributed to the sloping inversion, which slopes downward toward the coast. However, the many (more than 20) individual profiles also show this three-layer feature, particularly close to the coast. Also, note that the depth of the MBL observed here is similar to the height of the local terrain at the headland, Cape Mendocino. This indicates that the stability of the inversion layer itself could be quite important.

The differences in stability and the character of the inversion, can be illustrated by the Brunt–Väisälä frequency, N. Wave motions at a frequency higher than N cannot be supported; however, as N and U vary with height, nonlinear wave processes (e.g., wave breaking, absorption and reflection at critical layers) occur and complicate the analysis. Here, N is calculated from flight stacks that emanate right at the cape (stack #2), since up- and downstream stacks are more affected by coastline curvature and other phenomena. The mean, maximum, and minimum N, estimated from the virtual potential temperature gradient, are shown in Fig. 5. Again, 7 and 12 June are similar with constantly low values at mid-MBL (<0.01 s−1) and a tendency for higher values close to the surface. This tendency is strongest for 7 June and is due to the lower SST close to the coast, which increases the low-level stability. Here N increases rapidly into the inversion (z > ∼400 m) with values reaching ∼0.03 s−1 in the inversion. On 26 June on the other hand, the stability increases linearly from close to zero to 0.02 s−1 between the surface and ∼1 km.

The mean and plus one standard deviation profiles of liquid water, as measured by the aircraft, are shown in Fig. 6 and confirm the inference drawn from the satellite image. On 7 June, no clouds were observed directly within the entire measurement domain. In contrast, 12 June appears to be dominated by a single stratocumulus layer, with maximum liquid water reaching several tenths of a gram per kilogram air at ∼500 m, corresponding to the mean inversion height. The structure on 26 June appears to be more complex, as also suggested by the satellite image, with indications of at least one stratocumulus layer around 1 km topped by a second thinner layer with more liquid water at ∼1300 m. However, no or little clouds were observed in the area in the lee (south) of the cape on either day (not shown).

The main characteristics of the three cases are summarized in Table 1. It is obvious that the flow is stronger on 7 and 12 June, but other factors are also important. For example, the Froude number, Fr, whether calculated from shallow water theory (using the depth of the MBL and the inversion strength to estimate the gravity wave phase speed, and its mean wind speed) or for a continuous fluid (using N in the inversion and the height to midinversion, and the MBL mean wind speed), is supercritical or close to supercritical (Fr ∼ 1 or larger) for the first 2 days, but is clearly subcritical for 26 June. Even assuming that the background conditions on 7 June were slightly subcritical, only a marginal acceleration was sufficient to make the flow supercritical around Cape Mendocino; this is sometimes called a transcritical flow (e.g., Burk et al. 1999). This implies that as the coast turns away from the flow south of the cape, the geostrophic adjustment following the change in the pressure field induced by changes in the MBL depth can only take place downstream. An expansion fan forms, which enhances the shrinking of the MBL depth in the lee of the cape and limits the horizontal extension of the largest inversion slope offshore. With a lower Fr, as on 26 June, this adjustment is expected to be smoother and gradual, affecting the whole area also upstream of the cape.

The Burger number, Bu, is large both for the coastal mountain barrier when using the stability in the inversion and for the terrain at the cape when using the stability in the MBL. First, this means that the general coastal flow is efficiently blocked by the main coastal mountain barrier, allowing no flow across the barrier. The Rossby radius of deformation, lR, is >100 km on all three days, which implies that the effect of the coast will be seen all through the measurement domain; this is verified below. Second, the latter Bu indicates that also the terrain at the cape will influence the flow. In particular for 7 and 12 June, when the MBL depth is similar to the height of this local terrain, the flow will be blocked or partially blocked by this terrain. A lee wave may then form that brings down warm, dry, and low-momentum air from aloft, further enhancing the collapse of the MBL in the lee of the cape. This will be discussed below.

c. Horizontal variation of the vertical structure

Part of the variability in the profiles shown above is attributable to the fact that different horizontal locations have different vertical structure. This is illustrated for potential temperature, specific humidity and scalar wind speed in Figs. 7–9. Here all the aircraft data were separated into three regions, shown from the top and down in each figure: upstream offshore, downstream offshore, and downstream inshore (essentially in the lee of the cape). The separation in up- and downstream areas is determined by the position of the cape. The separation of the downstream area into off- and inshore was subjective, using the structure of the potential temperature profile. This separation is very similar to the extension of possible expansion fans (Winant et al. 1988; Samelson 1992; Baines 1995). Note that the height coverage is different for measurements in different areas on different days. This reflects the differences in how the flight tracks were set up during different days. This also affects the representability of the profiles in Fig. 4.

Figure 7 shows the potential temperature profiles. The first and second profiles on 7 June are very similar showing that any mesoscale perturbations by the cape have a small lateral propagation to the mean flow. The only area with a major perturbation is found downstream of the cape, where the collapse of the MBL and the stable inversion reaches all the way to the surface making the temperature profile stable throughout the whole layer. This is also seen in the moisture profiles (Fig. 8) and in the wind speed profiles (Fig. 9). In the latter, the elevated wind speed maximum (the coastal jet, see below) is lowered from ∼300 to 500 m all the way down to the surface, with a significant increase in mean wind speed from the upstream conditions.

The profiles on 12 June are similar; however, the upstream offshore inversion is deeper and the downstream inshore stability is slightly less pronounced (Fig. 7). The downstream offshore MBL is also deeper, as could be expected for a cloud-capped MBL (see Fig. 3). The MBL in the lee of the cape still shows some of the same collapsing structure as on 7 June but the mean wind speed is significantly more well mixed—almost constant at ∼12 m s−1 but with a larger variability that increases with decreasing altitude. The acceleration downstream appears somewhat smaller. These structural differences, compared to 7 June, may be attributed to the presence of a cloud layer on 12 June, which keeps the MBL more well mixed. This counteracts the slope of the inversion toward the coast, and may thus enhance the local slope closer to the coast, where the clouds dissipate in the lee of the cape—this will be verified below.

The profiles for 26 June are strikingly different, although there is still a jetlike structure in the upstream offshore profile and the temperature profile also shows a slight shallowing of the MBL and a lowering of the wind speed maximum in the downstream inshore MBL. The upstream depth of the MBL is actually hard to determine from the temperature and moisture profiles alone. There is, however, a jetlike structure present between ∼500 m and ∼1 km, that becomes slightly enhanced and lowered downstream, in the lee of the cape. In all three cases and for all three regions, the highest wind speed is found within the MBL, with a decline in the inversion.

The combination of lidar and in situ measurements makes it possible to illustrate both some of the spatial variability in the vertical structure, simultaneous small-scale structures, and temporal variability. Figure 10 shows two cross sections of lidar backscatter intensity on 12 June, from the stack emanating at the cape (stack 2). Figure 10a shows the results from the upward-looking lidar from the 30-m flight leg while Fig. 10b shows the results from the downward-looking lidar from the (later) 1000-m leg of the same stack. Here, a composite of the scalar wind speed from the entire stack, using both the level flight legs and the sawtooth profiles is included. The completely white areas are where the lidar signal is attenuated by clouds, illustrating quite well the slope of the MBL top from ∼600 to ∼400 m along this transect.

There also seems to be a high degree of vertical correlation between properties in the MBL, as illustrated by the aerosol content indicated by the lidar returns, and the cloud field. Note in particular the region approximately between longitude −124.9° and −124.7°, in particular west of −124.8°. Here the trace seems to indicate some larger than average vertical motions in the cloud field, possibly indicating gravity wave motions. Approximately 1 h later, when looking down on the same area from 1 km (Fig. 10b), the cloud layer has started to dissolve here. This area also coincides with the highest wind speeds, the coastal jet. It is possible that wave instabilities associated with the wind shear above the speed maximum, manifested in the larger amplitude in the small-scale motions at the cloud top (Fig. 10a), are responsible for the dissipation of the clouds here. Note also the tendency for the local MBL depth to decrease rapidly when the clouds dissipate.

Figures 11–13 show composite cross sections of wind speed and potential temperature from the flight stacks upstream (stack 1) and downstream (stack 3) of the cape. For 26 June there are two cross sections downstream, flown before and after the upstream transect, respectively. On 7 June (Fig. 11), there is a weak upstream coastal jet (Umax ∼ 16 m s−1), in a ∼600 m deep rather homogeneous MBL, while the downstream MBL collapses completely to less than 100-m depth, capped by a layer of warm (ϒ ∼ 300 K), slow-moving (U < 5 m s−1) air. This may be due in part to the presence of a lee wave from the upstream cape. The location of the maximum inversion slope, coincident with a wind maximum of Umax ∼ 28 m s−1, extends westward from the same longitude as that of the upstream tip of the cape. Note also the skewed shape of the jet. The structure of the inversion slope, with one slope west of the line downstream from the cape and another slope in the lee of the cape, and the shape of the jet, both indicate that there is more than one process responsible for the collapse of the MBL.

On 12 June (Fig. 12), the increase in wind speed is less significant, but the upstream jet is more pronounced. The slope of the downwind inversion is larger and more localized, but still located at about the same longitude. It is likely that these differences to the 7 June case are due to the cloud field on this day, keeping the MBL west of the wind maximum more well mixed and thus less sloping. In the subsidence behind the cape, the clouds are dissipated completely, thus the localized slope. Note also that the downwind stack (Fig. 12b) on this day extends ∼200 km west of the coast. It is only in the very west most portion that one can possibly see the MBL top asymptotically approaching a more constant height, corroborating the discussion on lR above. Although this cross section does not extend as far to the east in the lee of the cape as on 7 June, the area of warm low-momentum air in the lee of the cape is also seen.

The cross sections on 26 June reveal more temporal variability. The upstream transect (Fig. 13a) shows a weak jet associated with a slightly sloping inversion from ∼1.3 km to ∼800 m. The two downstream transects indicate the temporal variability. The first transect, flown before the upwind stack, shows a somewhat enhanced jet, Umax ∼ 9 m s−1, along a somewhat more (than upstream) slanting inversion. The later transect, however, only shows weakly increased winds toward the coast, while most of the previous inversion slope has disappeared. The MBL on this day is also much less well mixed.

The 100-m wind speed and potential temperatures (Fig. 14) analyzed from along-coast stacks close to the coast on all 3 days (although aligned somewhat differently to the local coast, see Fig. 2), suggest the presence of a lee wave triggered by the terrain at the cape. The first 2 days both show a sharp increase in wind speed around the location of the cape, coinciding with an abrupt increase in potential temperature, by ∼6 K. On 26 June, in contrast, the temperature is almost constant or only slightly decreasing while there is a weak wind speed maximum, of a only a few meters per second, which is much weaker than on the previous days. Within shallow water theory, the increasing wind speed associated with an expansion fan should be located within the well-mixed MBL, with a roughly constant potential temperature. Here the wind speed and the potential temperature increases simultaneously, which indicates the presence of a gravity wave, possibly with a wave-breaking event, consistent with the observed increase in both wind speed and potential temperature. From measurements alone it is difficult, if not impossible, to distinguish completely between the effects of the expansion fan and a possible gravity wave, in particular in the absence of extensive measurements well above the MBL that could reveal the vertical structure of a wave. However, the evidence here, together with that from a modeling study of the 7 June case simulated with and without the terrain at the cape (Tjernström 1999), strongly suggest that a gravity wave induced by the local terrain at the cape is indeed a major reason for the observed collapse of the MBL directly in the lee of Cape Mendocino.

d. Horizontal structure

Structures in the horizontal plane are illustrated in Figs. 15–19. The first of these (lidar MBL heights) utilizes data from both 30 and 1000 m, while the rest are compiled from the 30-m flight tracks, using all the data from the entire flight. This procedure assumes quasi-stationarity, which may be somewhat less than accurate for 26 June. Figure 15 shows the MBL depth, inferred from the inversion base in cloud-free conditions and the cloud top when clouds are present, for 7 and 12 June. Both feature a significant reduction in MBL depth in a southward expanding area in the lee of the cape, to <200 and ∼300 m, respectively. The upstream inversion is also sloping, from ∼700 to ∼500 m and from 500–600 to ∼400 m, respectively. On 12 June, the deepest MBL within the domain is found to the southwest, which is consistent with the thickening cloud layer here (see Fig. 3).

From simple hydrostatic reasoning, changes in the MBL depth imply changes in the surface pressure. The observed changes in the upstream inversion height implies a horizontal pressure difference of typically ∼1–2 hPa. Taken over a distance of ∼100 km this would translate into an additional geostrophic wind of the order of ∼20 m s−1. This is the driving mechanism for the coastal jet. The larger MBL depth differences downstream of the cape, of order ∼300–500 m, translate into an even larger horizontal pressure difference of ∼2–4 hPa. These numbers, and the differences for the 2 days, agree roughly with the surface pressure distributions calculated from the 30-m flight levels in Fig. 16. The pressure gradient appears to be much more perpendicular to the coast on 12 than on 7 June. Note that not all of the increased pressure gradient will translate into wind speed increase, as the turbulent momentum flux divergence closer to the surface also becomes larger with increasing wind speed. The local perturbation in the surface pressure at (40.5°N, 124.6°W) and the corresponding changes in MBL depth (Fig. 15) may be due to a stationarity problem.

Figure 17 shows the horizontal composite of the 30-m winds speed for all 3 days. In all cases, there is an increased wind speed in lee of the cape, although significantly less on 26 June when the maximum is also much more localized. For 7 and 12 June, the pattern corresponds well with the pressure fields in Fig. 16. On 7 June, the case with the most coast-parallel low-level flow upstream, the wind speed increases from <10 m s−1 to >20 m s−1, in a wide wedge-shape maximum starting at the tip of the cape. On 12 June, the acceleration at 30 m is less pronounced mostly because of higher winds upstream, from ∼12 m s−1 to >18 m s−1. In both these cases, the patterns closely resemble that of an expansion fan (Ippen 1951; Baines 1995), but at different angles. This may be due to different Fr on these days, but also to differences in the alignment to the coast of the flow upstream. Such patterns were also found in model simulations for Cape Mendocino (Tjernström and Grisogono 2000; Tjernström 1999). On 26 June, a weak and broad acceleration of a few meters per second extends due west from the cape, but a narrow and stronger maximum to ∼8 m s−1 is located close to the lee of the cape. On this day, the low-level flow also has the largest angle to the coast at this height.

This type of flow field has implication for the forcing of the coastal ocean, as indicated by the stress vectors and the spatial pattern of the friction velocity, u∗ (Fig. 18). Rogers et al. (1998) found that estimates of surface-layer fluxes using eddy correlation and bulk formulations often differ significantly. The momentum fluxes here were thus estimated by eddy correlation from the lowest aircraft legs. It is not clear, if this problem is due to inadequate measurements, or to the underlying theory for the bulk aerodynamic formulations, specifically the validity of the assumptions on stationarity and horizontal homogeneity. On 7 June there is an increase in the offshore surface stress from u∗ ∼ 0.1 to ∼0.7 m s−1, but there is also a band of lower stress aligned with the coast south of the cape. The pattern is different on 12 June, in that the upstream stress is larger, u∗ ∼0.4 m s−1, possibly because of the more well-mixed conditions, and that here there is a broad minimum downstream of the cape, probably associated with the minimum in SST, see below. The stress increases again farther south, probably as an effect of the establishment of a new balance in the MBL. On 26 June the stress pattern is well correlated with the wind speed, but the stress is quite weak.

The curl of the stress vector is also important for additional upwelling (e.g., Gill 1982; Enriques and Friehe 1997). Tjernström and Grisogono (2000) correlated the curl of the stress vector calculated from model results with the observed depression of the SST for 7 June. They found that although SST depressions did occur in areas with low curl, high values of the curl were always associated with a large SST depression. They interpreted this as a “background” SST depression, associated to the down-coast flow, with an additional SST depression generated by high values of the stress-vector curl. Figure 19 shows the SST distributions for all 3 days. Although an accurate calculation of the stress-vector curl is difficult from measured data, a subjective comparison with the largest stress gradients in Fig. 18 supports this idea. Although the exact values of the SST are obviously dependent on many other processes, for example, transport processes in the coastal ocean and the bottom topography (see, e.g., the dashed line in Fig. 19b), it seems plausible that the wind forcing on these days are important for the large depressions in SST, about ∼5°C, or larger, close to the coast.

However, while the SST depression appears closely linked to the spatial pattern in momentum transfer from the atmosphere to the ocean, it is difficult to see any clear feedback to the atmosphere, for example, to the wind field. While the atmospheric forcing appears vital for the state of the ocean surface, the feedback to the atmosphere via the SST, which alters the local stability, appears weak. This conclusion is supported by model simulations by Tjernström and Grisogono (2000) and Tjernström (1999). They found that only minor improvements in the agreement between simulation and observation resulted from using observed SST rather than a constant background value in their model. It may be expected, however, that the turbulent characteristics of the flow in the expansion fan are affected and will appear highly atypical.

e. Momentum balance

The conservation of momentum written in Reynolds-average terms in an Eulerian framework are governed by the equations:
i1520-0469-58-9-953-e1
where u, υ, and w are motions in x, y, and z directions, and ρ is density and p is pressure. Terms I–IV are the ageostrophic acceleration, or the Lagrangian time-rate-of-change, and the balance between terms V (the Coriolis term) and VI (the pressure gradient) yields the geostrophic balance. The last term (VII) is the vertical momentum-flux divergence due to turbulence. While being relatively simple to calculate from model results (e.g., Cui et al. 1998; Burk et al. 1999; Tjernström and Grisogono 2000), many terms in these equations are quite difficult to estimate from experimental data directly.

Samelson and Lentz (1994) estimated some of the terms using data from the CODE experiment from a limited array of buoys around Point Arena. They found that the across-coast momentum was usually in geostrophic balance, that is, the dominating balance in the equation for the across-coast wind speed component was between the across-coast pressure gradient and the Coriolis force associated with the along-coast wind. In the along-coast momentum equation there was a balance between the pressure gradient and the surface friction and/or the ageostrophic advection. They were, however, forced to ignore term IV, the vertical advection (a minor omission since their measurements were at nominally 10 m), and to make a rather crude assumption about the last term, VII. They simply assume that this could be estimated as a difference between the surface flux, estimated from bulk formulas, and a zero flux at an assumed (fixed) MBL top, thus assuming a linear momentum flux profile. Thus, their residual term was quite substantial. Nevertheless, their results have been broadly verified by later model calculations (e.g., Cui et al. 1998).

An estimation of terms II–VI from horizontally gridded aircraft data taken from an entire flight implies that the first term is zero or at least negligible. For 7 June, this is probably a reasonable assumption. The observations appear quite steady and model simulations of that day (Tjernström and Grisogono 2000) also seem to confirm this. On 12 June, it was shown above that some of the cloud field dissipates even during the course of one stack, and that this has immediate consequences for the MBL depth. Considered over a longer time, comparing MBL heights estimated by lidar at the same locations but at different times, there is a temporal variability that may affect the dynamics. The case of 26 June is the one here that varies the most in time (Fig. 13). Calculating spatial gradients from measured data is usually accompanied by a large variability that can be interpreted as noise. Calculating these gradients from block-averaged data, which was interpolated to a rectangular grid, is believed to increase the accuracy while limiting the resolution.

Calculation of the vertical advection requires an estimate of the vertical wind, which is not directly available from aircraft measurements. This is due to inadequate accuracy—the turbulent variations in vertical wind around a mean value can be measured but not the mean value itself. Instead, the mean vertical wind speed was estimated from continuity considerations, integrating the continuity equation from a zero vertical wind at the surface and upward, using the gridded horizontal winds fields from each (entire) flight. Figure 20 shows the resulting vertical wind speed at 375 m on 7 and 12 June, respectively. The quite small offshore values, varying only marginally in space, indicate that the method is working. The area of mostly negative values more or less covers an area similar in shape to an expansion fan. This subsidence is responsible for dissipating the clouds here on 12 June. Even at moderate heights, the vertical winds become quite substantial, w ∼ −0.1–−0.3 m s−1. An order-of-magnitude estimate for the vertical wind at inversion height, just from observed gross wind speed differences (e.g., Fig. 17) along a sloping inversion, for the 3 days gives:
i1520-0469-58-9-953-eq1
In reality, the depression of the MBL depth was never as large as the whole upstream MBL depth, and the influence of entrainment is not estimated here. However, this analysis is only meant to check the order of magnitude values of the vertical velocity, and the results are in rough agreement with the values in Fig. 20.
The problem of how to estimate the momentum flux divergence (term VII) arises from lack of vertical resolution and coverage in the measurements. In principal, accurate measurements of the turbulent momentum flux at two levels, reasonably close in space is needed to calculate this term at one intermediate level. In Samelson and Lentz (1994), this term was estimated as a bulk value for the entire MBL as
i1520-0469-58-9-953-e2
where the zero subscript refers to the surface and the flux at the height H, by Samelson and Lentz given a fixed value, is assumed to be zero. With this approach, two problems follow: first, there is no reason to assume that H, which for an ideal horizontally homogeneous well-mixed boundary layer would be the PBL depth, is constant, either in space or in time. This is obvious from results in the present paper. Second, in this complex MBL, the momentum flux profiles are rarely linear. The structure of the momentum flux associated to the jet in the wind speed profile is illustrated in Fig. 21 from different sections of the along-coast transect on 12 June. As the jet is intensified and lowered, a secondary maximum in momentum transport appears below the wind speed maximum. The momentum flux profile thus becomes far from linear from the surface to the MBL top;this was also found in Smedman et al. (1993). A simple extrapolation of the momentum flux linearly upward from the surface, using the flight legs at 30 and 100 m on 7 June, crosses zero on average at zi/3 (not shown). Without extensive coverage and resolution with many flight legs in the vertical, exact estimates of this term thus become impossible. Here we simply assume that most of the residual in the budgets are due to this term, realizing that both temporal variability and accumulated errors in the remaining terms will also contribute to the residual.
The budgets are thus estimated as follows: the measurement domain was subdivided into a three-dimensional grid with a horizontal resolution of 5 km and a vertical resolution of 50 m. All mean variables were averaged onto this grid, in the horizontal plane over 10 km (overlapping). Subdomains that were not covered by flight legs or sawtooth profiles were assigned values by linear interpolation in space. From this data, all terms in Eq. (1) except for terms I and VII are estimated. It is then assumed that the main contributor to the residual is term VII. This assumption was checked at locations where direct estimates of turbulent momentum flux divergence were possible, and was found reasonable. Results from 7 and 12 June are summarized in Figs. 22 and 23 averaged for downstream cross sections coinciding with the southmost west–east transects on both days (corresponding to stacks 3 and 4, respectively). For both days, the pressure gradient term is the single leading term in the cross-coast momentum equation, offshore as well as inshore (Figs. 22b and 23b). Although similar in their appearance, the budget reveals differences between these days. On 12 June, the cross-coast momentum appears to be in a better geostrophic balance, at least above 200 m. Below this height the residual becomes more important, which is to be expected if this term does indeed represent the turbulence. In contrast, on 7 June there is a substantial contribution by term III, the alongshore advection of cross-shore momentum, in particular in the inshore region. In addition, on 12 June the across-shore pressure gradient is relatively constant with height, while on 7 June it increases dramatically in the inshore MBL. The conclusion from this analysis may thus be that the fields on 7 June are due to a mesoscale perturbation on the pressure field, to a larger extent than on 12 June. Examining the alongshore momentum budgets is more complicated. At least for the offshore budget on 7 June (Fig. 22d), it appears that there is no geostrophic balance. The pressure gradient term is balanced by the sum of the residual (including the turbulence term) and the alongshore advection of alongshore momentum. The former is more important in the MBL (as expected), while the latter is almost height independent. The Coriolis term is small. Offshore on 12 June, all terms contribute, but the pressure gradient is surprisingly small. At least close to the surface (below 100 m) the largest terms appear to be the pressure gradient, the Coriolis, and the residual terms, which would indicate an almost geostrophic balance (modified by surface friction) here. In the inshore along-coast momentum budgets, all terms are about equally important, with a significant scatter. Two conclusions are possible. That the residual is a large term, which is assumed to indicate the importance of the turbulence, and that the flow is highly ageostrophic. The curvature terms (term II) are usually small, except for in the inshore cross-shore momentum below 100 m on 7 June and in the inshore along-coast momentum. The vertical advection is large only in the inshore budgets on 7 June, which is expected from the strong lee-wave influence on that day. In summary, the flow offshore (but still well within a Rossby radius of deformation) may be characterized as
i1520-0469-58-9-953-e3
for 7 June and
i1520-0469-58-9-953-e4
for 12 June, respectively, although with significant scatter. It is interesting to note that the difference between 7 and 12 June only becomes obvious when investigating the budgets although these days appear relatively similar from the analysis of mean conditions.

4. Processes and scales

It is clear that there is an intricate scale interaction involving several scales, governing the dynamics of the flow in these three cases, which are believed to be typical for this location in the summer. Below we will try to address these scales in a descending order.

On the largest scale, the general flow in the lower troposphere is governed by the subtropical pacific high pressure system. In general, its position and strength set up the north-northwesterly flow that is typical for this region during summer. The difference in this pressure field is also what causes 7 and 12 June to be different from 26 June. On the latter day, the large-scale flow was anomalous for this time of the year, from south, which is clearly seen in the observed upper-level winds on this day.

The next scale down is governed by the presence of the main coastal topography. A useful measure of the horizontal extent of its influence is the Rossby radius of deformation, which also indicates that this scale is governed by quasigeostrophic motions and the geostrophic adjustment process. Within this distance from the coast, the low-level flow is blocked by the mountain range and the interaction with the coastline causes the boundary layer inversion to slope down toward the coast. This general slope of the inversion, here observed upstream of the cape, implies a horizontal pressure gradient, so that an inversion height difference of 100 m translates into horizontal pressure differences of ∼1.2 hPa. Viewed from a thermal-wind perspective, this is the cause of the coastal jet. Integrating the thermal-wind law downward from the relatively low wind speed aloft, the northerly component of the flow must increase downward until it becomes balanced by the surface friction. Using the wind speeds aloft as a measure of the background flow, the increase in the northerly wind speed component on 7 and 12 June is about 10–15 m s−1 at the most. Also on 26 June this increase is significant, at about ∼10 m s−1. However, the main difference is that on 26 June this thermal wind is added to a southerly background flow, up the coast. Thus the total down-coast flow in the boundary layer becomes weak compared to the two previous days. In terms of synoptic-scale geostrophic flow, all three cases are thus clearly supergeostrophic; however, from the meso α scale O(10–100 km) perspective, the flows are more quasigeostrophic, at least upstream of the cape. The deviation from geostrophy is illustrated from the offshore momentum-budget analysis, as somewhat representative for upstream conditions. A rough estimate can also be obtained by comparison of the actual mean wind in the upstream boundary layer to the surface pressure gradient, estimated directly from the measurements. The estimated values of the geostrophic wind were 12.5, 9.0, and 3.0 m s−1, while the actual mean winds were 14.8, 17.5, and 3.7 m s−1, respectively. From this, the flow upstream of the cape on 12 June is more supergeostrophic while that on 7 and 26 June are relatively close to geostrophy. This is one of two main differences between 7 and 12 June. One possible explanation for this difference is the fact that upstream and offshore MBL on 12 June was capped by a stratocumulus cover. It was observed that these clouds tended to reduce the tilt of the inversion on the Rossby radius scale, while it was significantly enhanced locally, where the clouds were dissipated by the inversion slope. It must be noted, however, that while the effect of the cloud edge on the inversion height could be detected at high resolution in the composite of the sawtooth profiles, the geostrophic wind is estimated from the low-level flight leg with significantly more spatial smoothing. Also, the cloud layer did not appear to be stationary and when the clouds were locally dissipated, the boundary layer depth was immediately decreased. Such an event was observed by Sundararajan and Tjernström (1999) to be a likely cause of a supergeostrophic oscillation in the wind at a coastal location in southern Sweden.

The third scale down is governed by local changes in the coastline orientation. As the semitrapped flow passes around the cape, the “channel” widens and because of mass continuity, the near-coast MBL must become shallower; the inversion slope increases. This represents a spatial change in the local pressure field. In a Lagrangian sense, however, the flow will experience this as a temporally changing pressure field and the response will be a geostrophic adjustment. If Fr > 1, the gravity waves that ensure the mass transfer toward a new equilibrium cannot propagate upstream and only the downstream MBL will be affected; an expansion fan forms. Still from hydrostatic considerations, a MBL depth depression of 650 m would here generate the quite substantial pressure perturbation of ∼7.8 hPa. It is here that the main difference between the two high-wind cases and the low-wind case becomes most obvious. Although having about the same thermal-wind increase along the flow, the Fr supercriticality is governed by the absolute wind speed. Thus 7 and 12 June are supercritical (or for 7 June at least transcritical) while 26 June is clearly subcritical. Although, there is a local acceleration in the flow past the cape in all three cases, as a response to the locally enhanced tilt in the inversion when the channel widens, only the first 2 days reveal an expansion fan.

Calculated from the stability in the inversion Fr > 1 on both days, and using sin(α) = Fr−1, where α is the angle of the expansion fan to the flow, the 2 days would have expansion-fan angles of ∼40° and ∼30°, respectively. Noting that the upstream flow on 12 June is slightly from northwest, this appears to be in rough agreement, for example, from the wind speed isoline pattern in Fig. 17. In the expansion fan, there is a considerable subsidence (also see below), which causes the clouds on 12 June to dissolve locally. In the inferred (from continuity) vertical winds on this day, there is also a hint of a hydraulic jump farther downstream (see Fig. 20). However, since no flight tracks parallel to the wind were flown here, it is not possible to directly verify this. It is also in the expansion fan that the next difference between 7 and 12 June is apparent. Previous measurements (Samelson and Lentz 1994) and model simulations of 7 June (Tjernström and Grisogono 2000) have revealed that the flow in general, and on 7 June in particular, is semigeostrophic. Indeed this analysis of 7 June also roughly indicates a semigeostrophic balance upstream and offshore. The cross-coast wind component is roughly a balance between the pressure gradient, Coriolis, and turbulence terms with contribution also from along-coast advection. The Coriolis term is very small in the along-coast wind and the pressure gradient is instead balanced by advection and friction. However, 12 June is different in that it is much closer to geostrophic balance, at least in the outer portions of the expansion fan. This can also be illustrated by the increase in wind speed, comparing the upstream jet to that in the expansion fan; 17.3 to 28.0 and 19.0 to 22.1 m s−1, respectively. It is also borne out by the increase in the pressure gradient in the expansion fan on 7 June.

In summary, it would appear that on 7 June, the high wind speeds south of the cape are due to a strong mesoscale ageostrophic perturbation in the flow within the expansion fan, while the upstream flow is closer to a geostrophic balance. On 12 June, in contrast, the high winds are more due both to a mesoscale pressure pattern upstream, in combination with a flow in quasigeostrophic balance, while the expansion-fan dynamics is less important. The higher Fr contributes in narrowing the expansion fan, as does the slightly more onshore-directed flow upstream. The presence of the clouds on 12 June must also have played a role. The clouds offset the shallowing of the MBL in the expansion fan, which also causes the flow to be less perturbed south of the cape. The weak gradients on 26 June prohibit an estimation of the momentum budget, but it is clear that from the smooth character of all fields that it is roughly in mesoscale geostrophic balance, and no expansion fan can be detected.

The fourth scale on the descending order of scales is the impact of the local terrain at the cape itself. This is difficult to analyze since this effect will be superimposed on the effects of the expansion fan dynamics and could be expected to act in the same manner. Both would cause a depression of the MBL depth in the lee of the cape. In addition, no flight legs were flown directly across the cape and the few north–south flight legs that were flown were all at lower altitude. However, from simple scale analysis it is probable that this terrain could cause a partial blocking of the upstream flow. In theory, this could produce a gravity wave event giving a downslope flow enhancement. Since the crosswind extent of the terrain is finite, one would only expect to see some of the signs of this in the measurements. However, the strongest wind speed acceleration is coincident with a drastic increase in temperature (Fig. 14) and also capped by a sharp reduction in wind speed aloft (Figs. 11, 12). Both these features are consistent with bringing down low-momentum warm air from aloft. This was also the conclusion from the modeling study by Tjernström and Grisogono (2000), while Tjernström (1999) confirmed this hypothesis by simply removing this terrain in a sensitivity simulation. In the latter, the expansion fan remained but the wave component disappeared. This was particularly clear in the flow aloft, where unfortunately there were no measurements taken on straight and level flight legs. On 26 June, the upstream boundary layer was deeper and the flow speed was weaker. There was a weak acceleration but the absence of a significant change in temperature is consistent with the absence of such a wave feature.

Finally, the smallest scale is obviously the local impact of boundary layer turbulence. While this is clearly important for the wind field, as borne out by the momentum budgets, and for the forcing of the upper ocean, as borne out by the SST fields, it would appear here that there is little feedback from this interaction between the flow and the surface. This is also consistent with model simulations (Tjernström and Grisogono 2000; Tjernström 1999; S. Burk 2000, personal communication) where simulations were performed with constant SST or an interpolated SST from the measurements. It is likely that the observed highly variable SST fields would be more important in weaker wind conditions. However, it is also possible that the same variability would be less prominent on such days, due to the less vigorous atmospheric forcing of the ocean surface. The apparent insensitivity of the models also indicated here is probably quite sensitive to the scales involved.

It was also shown here that the simplifying mixed layer assumptions for the profiles of turbulent fluxes were in general not applicable. It may indeed be speculated if well-established similarity theories, such as the Monin–Obukhov theory, are at all applicable here, since the vertical scale of the flow is so strongly imposed by the expansion fan dynamics and the wind speed jet (Smedman et al. 1997). This was also indicated by the failure of the bulk-flux parameterization estimates for the turbulent fluxes in stable conditions in Rogers et al. (1998). The failure of this theory could be linked to its fundamental assumptions of stationarity and horizontal homogeneity. While the former can be satisfied here, the latter is violated in large areas, in particular in the expansion fans and for the large SST depression.

5. Conclusions

One objective of the Coastal Waves 1996 field experiment was to make detailed observations in the marine atmospheric boundary layer in the vicinity of major headlands. The purpose was in particular to study the flow during so-called supercritical conditions. In this paper, observations from airborne missions to the coastal marine boundary layer in the vicinity of Cape Mendocino, California have been described and analyzed. Flight-level in situ meteorological measurements are combined with some lidar data, to relate the observed wind fields to the boundary layer topography. The observations were taken from three different days. These days include 2 days with a strong northerly flow, typical for summertime conditions along this coast (7 and 12 June) and 1 day with weak, more atypical flow (26 June).

On a first inspection it would seem that the two high-wind cases are similar and should be compared to the low-wind case. A closer analysis reveals that although there are similarities between the cases from 7 and 12 June that are indeed in contrast to the low-wind case on 26 June there are also significant differences between the two high-wind cases. As always with observational data gathered by aircraft, it will be difficult to distinguish between temporal and spatial variability thus leaving room for speculation. It is certain that modeling studies of these cases will shed more light on hypotheses that arise from this study.

The two high-wind days are supercritical using the simple shallow water form of Fr, although for the first day, 7 June, the upstream Fr calculated in this manner is close enough to unity that this day may have been transcritical. However, both are clearly supercritical locally downstream of the cape. For all cases, the local perturbation of the flow falls within a Rossby radius of deformation from the coast. On June 7, the upstream flow is in approximate thermal wind balance while the expansion fan induces a large local pressure perturbation and the flow becomes strongly ageostrophic. On 12 June, the presence of the cloud layer maintains a relatively uniform upstream MBL depth. The coastal jet appears only very localized close to the coast and appears to be more supergeostrophic than on 7 June. The clouds on 12 June also counteract the depression of the MBL depth in the expansion fan. Consequently, the mesoscale perturbation on the wind field south of the cape is much less prominent on 12 June.

One common and presumably important factor for these flows is the effect of the local terrain at the cape. These measurements suggest that a large fraction of the MBL depression south of the cape is not due to the expansion fan but to a lee wave behind this terrain. This is supported by high temperatures coinciding with the increase in the wind speed and that the most shallow MBL is capped by quite low winds. This is entirely consistent with bringing down warm and low-momentum air from above the MBL, and confirms hypotheses in Tjernström and Grisogono (2000) and Tjernström (1999), who observed this behavior in model simulations.

It is worth noting that the mesoscale perturbation on the wind speed is almost as large on 26 June as it is on the other 2 days. However, the thermal wind, caused by the interaction between the flow and the coastal mountain barrier, is here added to a background upcoast flow. Thus, the total downcoast wind is not sufficient to generate a supercritical flow. Note, however, that since the mesoscale perturbation in the wind speed appears to be O(10 m s−1), the inversion strength is typically 10°–20°C and a typical MBL depth is 200–400 m, supercritical flow may appear even with very light background flow. Since the predominant flow is from north or northwest, it may be expected that supercritical flows are very common.

Acknowledgments

This study was sponsored by the Office of Naval Research through Grant N00014-96-1-0002 and by the American Scandinavian Foundation. The authors are grateful to all participants in the Coastal Waves 1996 project, in particular to the NCAR flight crew. The authors are grateful to Clive Dorman, Stephen Burk, and Branko Grisogono for many valuable discussions.

REFERENCES

  • Baines, P. G., 1995: Topographic Effects in Stratified Flow. Cambridge University Press, 482 pp.

  • Bane, J. M., S. M., Haines, L. Armi, and M. H. Sessions, 1995: The California Coastal marine layer: Wind and thermodynamics. June 1994 Aircraft Measurement Program, University of North Carolina Tech. Rep. CSM-95-1, 289 pp.

  • Banta, R. M., 1995: Sea breeze shallow and deep on the California coast. Mon. Wea. Rev.,123, 3614–3622.

  • ——, L. D. Olivier, and D. H. Levinson, 1993: Evolution of the Monterey Bay sea-breeze layer as observed by pulsed Doppler Lidar. J. Atmos. Sci.,50, 3959–3982.

  • Beardsley, R. C., C. E. Dorman, C. A. Friehe, L. K. Rosenfeld, and C. D. Winant, 1987: Local atmospheric forcing during the coastal ocean dynamics experiment. 1. A description of the marine boundary-layer and atmospheric conditions over a northern California upwelling region. J. Geophys. Res.,92, 1467–1488.

  • Bond, N. A., C. F. Mass, and J. E. Overland, 1996: Coastally trapped wind reversals along the United States west coast during the warm season. Part I: Climatology and temporal evolution. Mon. Wea. Rev.,124, 430–445.

  • Bridger, A. F. C., W. C. Brick, and P. F. Lester, 1993: The structure of the marine inversion layer off the central California coast: Mesoscale conditions. Mon. Wea. Rev.,121, 335–351.

  • Brooks, I. M., and D. P. Rogers, 1997: Aircraft observations of boundary layer rolls of the coast of California. J. Atmos. Sci.,54, 1834–1849.

  • Brost, R. A., J. C. Wyngaard, and D. H. Lenschow, 1982a: Marine stratocumulus layers. Part I: Mean conditions. J. Atmos. Sci.,39, 800–817.

  • ——, ——, and ——, 1982b: Marine stratocumulus layers. Part II: Turbulence budgets. J. Atmos. Sci.,39, 818–836.

  • Burk, S. D., and T. Haack, 2000: The dynamics of wave clouds upwind of coastal orography. Mon. Wea. Rev.,128, 1438–1455.

  • ——, ——, and R. M. Samelson, 1999: Mesoscale simulation of supercritical, subcritical and transcritical flow along coastal topography. J. Atmos. Sci.,56, 2780–2795.

  • Cui, Z., M. Tjernström, and B. Grisogono, 1998: Idealized simulations of atmospheric coastal flow along the central coast of California. J. Appl. Meteor.,37, 1332–1336.

  • Davis, R. E., and P. S. Bogden, 1989: Variability on the California shelf forced by local and remote winds during the Coastal Ocean Dynamics Project. J. Geophys. Res.,94, 4763–4783.

  • Dorman, C. E., 1997: Comments on “Coastally trapped wind reversals along the United States west coast during the warms season. Part II: Synoptic evolution.” Mon. Wea. Rev.,125, 1692–1694.

  • ——, and C. D. Winant, 1995: Buoy observations of the atmosphere along the west-coast of the United States, 1981–1990. J. Geophys. Res.,100, 16 029–16 044.

  • ——, D. P. Rogers, W. Nuss, and W. T. Thompson, 1999: Adjustment of the summer marine boundary layer around Pt. Sur, California. Mon. Wea. Rev.,127, 2143–2159.

  • ——, T. Holt, D. P. Rogers, and K. Edwards, 2000: Large scale structure of the June–July 1996 marine boundary layer along California and Oregon. Mon. Wea. Rev.,128, 1632–1652.

  • Duynkerke, P. G., and A. G. M. Driedonks, 1987: A model for the turbulent structure of the stratocumulus-topped atmospheric boundary layer. J. Atmos. Sci.,44, 43–64.

  • Enriques, A. G., and C. A. Friehe, 1996: Parameterization of momentum, heat, and moisture fluxes over a coastal upwelling area. J. Geophys. Res.,102, 5781–5798.

  • Gill, A. E., 1977: Coastally trapped waves in the atmosphere. Quart. J. Roy. Meteor. Soc.,103, 431–440.

  • ——, 1982: Atmosphere–Ocean Dynamics. Academic Press, 662 pp.

  • Halliwell, G. R., and J. S. Allen, 1987: The large-scale coastal wind field along the west coast of North America. J. Geophys. Res.,92, 1861–1884.

  • Ippen, A. T., 1951: Mechanics of supercritical flow. Trans. Amer. Soc. Civ. Eng.,116, 268–295.

  • Kelly, K. A., 1985: The influence of winds and topography on the sea-surface temperature patterns over the northern California slope. J. Geophys. Res.,90, 1783–1778.

  • Lackman, G. M., and J. E. Overland, 1989: Atmospheric structure and momentum balance during a gap-wind event in Shelikof Strait, Alaska. Mon. Wea. Rev.,117, 1818–1833.

  • Overland, J. E., 1984: Scale analysis of marine winds in straits and along mountainous coasts. Mon. Wea. Rev.,112, 2530–2536.

  • Rogers, D., and Coauthors, 1998: Highlights of Coastal Waves 1996. Bull. Amer. Meteor. Soc.,79, 1307–1326.

  • Rogerson, A. M., 1999: Transcritical flows in the coastal marine atmospheric boundary layer. J. Atmos. Sci.,56, 2761–2779.

  • Samelson, R. M., 1992: Super-critical marine-layer flow along a smoothly varying coastline. J. Atmos. Sci.,49, 1571–1584.

  • ——, and S. J. Lentz, 1994: The horizontal momentum balance in the marine atmospheric boundary layer during CODE-2. J. Atmos. Sci.,51, 3745–3757.

  • Smedman, A.-S., and H. Bergström, 1995: An experimental study of stably stratified flow in the lee of high mountains. Mon. Wea. Rev.,123, 2319–2333.

  • ——, M. Tjernström, and U. Högström, 1993: Analysis of the turbulence structure of a marine low-level jet. Bound.-Layer Meteor.,66, 105–126.

  • ——, H. Bergström, and U. Högström, 1996: Measured and modeled local wind fields over a frozen lake in a mountainous area. Beitr. Phys. Atmos.,69, 501–516.

  • ——, U. Högström, and H. Bergström, 1997: The turbulence regime of a very stable marine air-flow with quasi-frictional decoupling. J. Geophys. Res.,102, 21 049–21 059.

  • Sundararajan, R., and M. Tjernström, 1999: Observations and simulations of a non-stationary coastal atmospheric boundary layer. Quart. J. Roy. Meteor. Soc.,126, 445–476.

  • Tjernström, M., 1999: The sensitivity of supercritical atmospheric boundary-layer flow along a coastal mountain barrier. Tellus,51A, 880–901.

  • ——, and D. Koracin, 1995: Modeling the impact of stratocumulus on boundary layer structure. J. Atmos. Sci.,52, 863–878.

  • ——, and D. P. Rogers, 1996: Turbulence structure in decoupled marine stratocumulus: A case study from the Astex field experiment. J. Atmos. Sci.,53, 598–619.

  • ——, and B. Grisogono, 2000: Simulations of super-critical flow around points and capes in a coastal atmosphere. J. Atmos. Sci.,57, 108–135.

  • Winant, C. D., C. E. Dorman, C. A. Friehe, and R. C. Beardsley, 1988: The marine layer off northern California—An example of supercritical channel flow. J. Atmos. Sci.,45, 3588–3605.

  • Zemba, J., and C. A. Friehe, 1987: The marine atmospheric boundary-layer jet in the coastal ocean dynamics experiment. J. Geophys. Res.,92, 1489–1496.

Fig. 1.
Fig. 1.

Surface plot showing a schematic of the terrain around Cape Mendocino. Note the mountain barrier stretching up to ∼1 km and the two ridges ∼500 m high stretching into the ocean at an angle to the northern coastline.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 2.
Fig. 2.

Flight tracks for the research missions to Cape Mendocino during Coastal Waves 1996: (a) 7 Jun, (b) 12 Jun, and (c) 26 Jun. The cross-coast tracks are numbered from north to south. The land is shaded gray.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 3.
Fig. 3.

GOES satellite images for (a) 16 LST 7 Jun, (b) 10 LST 12 Jun, and (c) 16 LST 26 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 4.
Fig. 4.

Mean (solid) and ± one standard deviation (dotted) profiles of wind speed (WS, m s−1), wind direction (WD, °), potential temperature (ϒ, °C), and specific humidity (Q, g kg−1), block-averaging all data from each entire flight according to radar altitude. The plots show (a) 7 Jun, (b) 12 Jun, and (c) 26 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 5.
Fig. 5.

Mean (solid) and min or max (diamonds) profiles of Brunt–Väisälä frequency for (a) 7 Jun, (b) 12 Jun, and (c) 26 Jun 1996. The data are taken at transect number two on each day, emanating at Cape Mendocino.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 6.
Fig. 6.

Mean (solid) and plus one standard deviation (dotted) profiles of cloud liquid water(g kg−1) for (a) 7 Jun, (b) 12 Jun, and (c) 26 Jun 1996. The data was averaged as in Fig. 4.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 7.
Fig. 7.

Mean and ± one standard deviation profiles of potential temperature (°C), as in Fig. 4 but each flight is subdivided into three regions: (a) upstream offshore, (b) downstream offshore, and (c) downstream inshore. See the text for a discussion.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 8.
Fig. 8.

Same as Fig. 7 but for specific humidity (g kg−1).

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 9.
Fig. 9.

Same as Fig. 7 but for the wind speed (m s−1).

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 10.
Fig. 10.

Cross sections of lidar backscatter return from stack 2 on 12 Jun, showing (a) the backscatter from the upward directed lidar at the first (lowest, 30 m) flight leg and (b) the from the last (highest, 1 km) flight leg. (b) Contour lines show the composite wind speed (m s−1) field derived using the entire stack.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

 Fig. 11.
Fig. 11.

Cross-coast cross sections of wind speed (gray scale, m s−1) and potential temperature (solid lines, °C) from 7 Jun 1996, for locations (a) upstream and (b) downstream of Cape Mendocino.Fig. 12. Same as Fig. 11 but for 12 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 13.
Fig. 13.

Same as Fig. 11 but for 26 Jun 1996, showing two cross sections downstream: (b) before and (c) after the upstream cross section (a).

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 14.
Fig. 14.

Composite analysis of (a) potential temperature (K) and (b) wind speed (m s−1) at 100 m for along-coast transects past Cape Mendocino for (solid) 7, (dashed) 12, and (dashed–dotted) 26 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 15.
Fig. 15.

Horizontal cross sections showing boundary layer heights (m) inferred from lidar data on (a) 7 and (b) 12 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 16.
Fig. 16.

Horizontal cross sections showing surface-pressure- inferred (hPa) 30-m flight legs on (a) 7 and (b) 12 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 17.
Fig. 17.

Horizontal cross sections of 30-m wind speed (solid, m s−1) and wind direction (arrows) on (a) 7, (b) 12, and (c) 26 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 18.
Fig. 18.

Same as Fig. 17 but for friction velocity u∗ (solid, m s−1) and stress direction (vectors).

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 19.
Fig. 19.

Same as Fig. 17 but for SST (°C). (b) Dashed line shows the bottom topography of the shelf.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 20.
Fig. 20.

Horizontal cross sections of vertical wind speed (m s−1) at 375 m as inferred by continuity considerations, on (a) 7 and (b) 12 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 21.
Fig. 21.

Vertical profiles of (a) total vertical turbulent momentum flux (m2 s−2) and (b) wind speed (m s−1) along segments of an along-coast transect on 12 Jun 1996. The lines 1–5 progress from north to south.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 22.
Fig. 22.

Momentum budgets for (a) and (b) across-coast and (c) and (d) along-coast momentum downstream of Cape Mendocino on 7 Jun 1996. (a) and (c) Close to the coast and (b) and (d) offshore. The different lines indicated in (a) refer to Eq. (1) in the text. The dividing line between off- and inshore is at −127.75° longitude.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Fig. 23.
Fig. 23.

As in Fig. 22 but for 12 Jun 1996.

Citation: Journal of the Atmospheric Sciences 58, 9; 10.1175/1520-0469(2001)058<0953:ODOCFA>2.0.CO;2

Table 1.

Summary of some characteristic conditions of the vertical structure of the MBL on 7, 12, and 27 Jun 1996. Here Umean refers to the MBL average, Umax refers to the maximum in the vertical of the mean wind profile for each region, Ni is the Brunt–Väisälä frequency in the capping inversion, zi is the height to the inversion base, and Δzi is the thickness of the inversion. Also, Δϒ is the temperature jump over the inversion; lR is the Rossby radius of deformation, estimated from the shallow water flow concept; while FrSW is the corresponding Froude number. Using inversion values of N and z, Frc is calculated for a continuously stratified atmosphere. Nhu and the Burger number, Bu, are calculated with two sets of terrain parameters: for the upstream coastal mountains (index 1, hm = 1500 m and lm = 70 km and Ni) and one for the terrain at the cape (index 2, hm = 500 m and lm = 15 km and NMBL).

Table 1.

1

NCAR is supported by the National Science Foundation.

Save
  • Baines, P. G., 1995: Topographic Effects in Stratified Flow. Cambridge University Press, 482 pp.

  • Bane, J. M., S. M., Haines, L. Armi, and M. H. Sessions, 1995: The California Coastal marine layer: Wind and thermodynamics. June 1994 Aircraft Measurement Program, University of North Carolina Tech. Rep. CSM-95-1, 289 pp.

  • Banta, R. M., 1995: Sea breeze shallow and deep on the California coast. Mon. Wea. Rev.,123, 3614–3622.

  • ——, L. D. Olivier, and D. H. Levinson, 1993: Evolution of the Monterey Bay sea-breeze layer as observed by pulsed Doppler Lidar. J. Atmos. Sci.,50, 3959–3982.

  • Beardsley, R. C., C. E. Dorman, C. A. Friehe, L. K. Rosenfeld, and C. D. Winant, 1987: Local atmospheric forcing during the coastal ocean dynamics experiment. 1. A description of the marine boundary-layer and atmospheric conditions over a northern California upwelling region. J. Geophys. Res.,92, 1467–1488.

  • Bond, N. A., C. F. Mass, and J. E. Overland, 1996: Coastally trapped wind reversals along the United States west coast during the warm season. Part I: Climatology and temporal evolution. Mon. Wea. Rev.,124, 430–445.

  • Bridger, A. F. C., W. C. Brick, and P. F. Lester, 1993: The structure of the marine inversion layer off the central California coast: Mesoscale conditions. Mon. Wea. Rev.,121, 335–351.

  • Brooks, I. M., and D. P. Rogers, 1997: Aircraft observations of boundary layer rolls of the coast of California. J. Atmos. Sci.,54, 1834–1849.

  • Brost, R. A., J. C. Wyngaard, and D. H. Lenschow, 1982a: Marine stratocumulus layers. Part I: Mean conditions. J. Atmos. Sci.,39, 800–817.

  • ——, ——, and ——, 1982b: Marine stratocumulus layers. Part II: Turbulence budgets. J. Atmos. Sci.,39, 818–836.

  • Burk, S. D., and T. Haack, 2000: The dynamics of wave clouds upwind of coastal orography. Mon. Wea. Rev.,128, 1438–1455.

  • ——, ——, and R. M. Samelson, 1999: Mesoscale simulation of supercritical, subcritical and transcritical flow along coastal topography. J. Atmos. Sci.,56, 2780–2795.

  • Cui, Z., M. Tjernström, and B. Grisogono, 1998: Idealized simulations of atmospheric coastal flow along the central coast of California. J. Appl. Meteor.,37, 1332–1336.

  • Davis, R. E., and P. S. Bogden, 1989: Variability on the California shelf forced by local and remote winds during the Coastal Ocean Dynamics Project. J. Geophys. Res.,94, 4763–4783.

  • Dorman, C. E., 1997: Comments on “Coastally trapped wind reversals along the United States west coast during the warms season. Part II: Synoptic evolution.” Mon. Wea. Rev.,125, 1692–1694.

  • ——, and C. D. Winant, 1995: Buoy observations of the atmosphere along the west-coast of the United States, 1981–1990. J. Geophys. Res.,100, 16 029–16 044.

  • ——, D. P. Rogers, W. Nuss, and W. T. Thompson, 1999: Adjustment of the summer marine boundary layer around Pt. Sur, California. Mon. Wea. Rev.,127, 2143–2159.

  • ——, T. Holt, D. P. Rogers, and K. Edwards, 2000: Large scale structure of the June–July 1996 marine boundary layer along California and Oregon. Mon. Wea. Rev.,128, 1632–1652.

  • Duynkerke, P. G., and A. G. M. Driedonks, 1987: A model for the turbulent structure of the stratocumulus-topped atmospheric boundary layer. J. Atmos. Sci.,44, 43–64.

  • Enriques, A. G., and C. A. Friehe, 1996: Parameterization of momentum, heat, and moisture fluxes over a coastal upwelling area. J. Geophys. Res.,102, 5781–5798.

  • Gill, A. E., 1977: Coastally trapped waves in the atmosphere. Quart. J. Roy. Meteor. Soc.,103, 431–440.

  • ——, 1982: Atmosphere–Ocean Dynamics. Academic Press, 662 pp.

  • Halliwell, G. R., and J. S. Allen, 1987: The large-scale coastal wind field along the west coast of North America. J. Geophys. Res.,92, 1861–1884.

  • Ippen, A. T., 1951: Mechanics of supercritical flow. Trans. Amer. Soc. Civ. Eng.,116, 268–295.

  • Kelly, K. A., 1985: The influence of winds and topography on the sea-surface temperature patterns over the northern California slope. J. Geophys. Res.,90, 1783–1778.

  • Lackman, G. M., and J. E. Overland, 1989: Atmospheric structure and momentum balance during a gap-wind event in Shelikof Strait, Alaska. Mon. Wea. Rev.,117, 1818–1833.

  • Overland, J. E., 1984: Scale analysis of marine winds in straits and along mountainous coasts. Mon. Wea. Rev.,112, 2530–2536.

  • Rogers, D., and Coauthors, 1998: Highlights of Coastal Waves 1996. Bull. Amer. Meteor. Soc.,79, 1307–1326.

  • Rogerson, A. M., 1999: Transcritical flows in the coastal marine atmospheric boundary layer. J. Atmos. Sci.,56, 2761–2779.

  • Samelson, R. M., 1992: Super-critical marine-layer flow along a smoothly varying coastline. J. Atmos. Sci.,49, 1571–1584.

  • ——, and S. J. Lentz, 1994: The horizontal momentum balance in the marine atmospheric boundary layer during CODE-2. J. Atmos. Sci.,51, 3745–3757.

  • Smedman, A.-S., and H. Bergström, 1995: An experimental study of stably stratified flow in the lee of high mountains. Mon. Wea. Rev.,123, 2319–2333.

  • ——, M. Tjernström, and U. Högström, 1993: Analysis of the turbulence structure of a marine low-level jet. Bound.-Layer Meteor.,66, 105–126.

  • ——, H. Bergström, and U. Högström, 1996: Measured and modeled local wind fields over a frozen lake in a mountainous area. Beitr. Phys. Atmos.,69, 501–516.

  • ——, U. Högström, and H. Bergström, 1997: The turbulence regime of a very stable marine air-flow with quasi-frictional decoupling. J. Geophys. Res.,102, 21 049–21 059.

  • Sundararajan, R., and M. Tjernström, 1999: Observations and simulations of a non-stationary coastal atmospheric boundary layer. Quart. J. Roy. Meteor. Soc.,126, 445–476.

  • Tjernström, M., 1999: The sensitivity of supercritical atmospheric boundary-layer flow along a coastal mountain barrier. Tellus,51A, 880–901.

  • ——, and D. Koracin, 1995: Modeling the impact of stratocumulus on boundary layer structure. J. Atmos. Sci.,52, 863–878.

  • ——, and D. P. Rogers, 1996: Turbulence structure in decoupled marine stratocumulus: A case study from the Astex field experiment. J. Atmos. Sci.,53, 598–619.

  • ——, and B. Grisogono, 2000: Simulations of super-critical flow around points and capes in a coastal atmosphere. J. Atmos. Sci.,57, 108–135.

  • Winant, C. D., C. E. Dorman, C. A. Friehe, and R. C. Beardsley, 1988: The marine layer off northern California—An example of supercritical channel flow. J. Atmos. Sci.,45, 3588–3605.

  • Zemba, J., and C. A. Friehe, 1987: The marine atmospheric boundary-layer jet in the coastal ocean dynamics experiment. J. Geophys. Res.,92, 1489–1496.

  • Fig. 1.

    Surface plot showing a schematic of the terrain around Cape Mendocino. Note the mountain barrier stretching up to ∼1 km and the two ridges ∼500 m high stretching into the ocean at an angle to the northern coastline.

  • Fig. 2.

    Flight tracks for the research missions to Cape Mendocino during Coastal Waves 1996: (a) 7 Jun, (b) 12 Jun, and (c) 26 Jun. The cross-coast tracks are numbered from north to south. The land is shaded gray.

  • Fig. 3.

    GOES satellite images for (a) 16 LST 7 Jun, (b) 10 LST 12 Jun, and (c) 16 LST 26 Jun 1996.

  • Fig. 4.

    Mean (solid) and ± one standard deviation (dotted) profiles of wind speed (WS, m s−1), wind direction (WD, °), potential temperature (ϒ, °C), and specific humidity (Q, g kg−1), block-averaging all data from each entire flight according to radar altitude. The plots show (a) 7 Jun, (b) 12 Jun, and (c) 26 Jun 1996.

  • Fig. 5.

    Mean (solid) and min or max (diamonds) profiles of Brunt–Väisälä frequency for (a) 7 Jun, (b) 12 Jun, and (c) 26 Jun 1996. The data are taken at transect number two on each day, emanating at Cape Mendocino.

  • Fig. 6.

    Mean (solid) and plus one standard deviation (dotted) profiles of cloud liquid water(g kg−1) for (a) 7 Jun, (b) 12 Jun, and (c) 26 Jun 1996. The data was averaged as in Fig. 4.

  • Fig. 7.

    Mean and ± one standard deviation profiles of potential temperature (°C), as in Fig. 4 but each flight is subdivided into three regions: (a) upstream offshore, (b) downstream offshore, and (c) downstream inshore. See the text for a discussion.

  • Fig. 8.

    Same as Fig. 7 but for specific humidity (g kg−1).

  • Fig. 9.

    Same as Fig. 7 but for the wind speed (m s−1).

  • Fig. 10.

    Cross sections of lidar backscatter return from stack 2 on 12 Jun, showing (a) the backscatter from the upward directed lidar at the first (lowest, 30 m) flight leg and (b) the from the last (highest, 1 km) flight leg. (b) Contour lines show the composite wind speed (m s−1) field derived using the entire stack.

  • Fig. 11.

    Cross-coast cross sections of wind speed (gray scale, m s−1) and potential temperature (solid lines, °C) from 7 Jun 1996, for locations (a) upstream and (b) downstream of Cape Mendocino.Fig. 12. Same as Fig. 11 but for 12 Jun 1996.

  • Fig. 13.

    Same as Fig. 11 but for 26 Jun 1996, showing two cross sections downstream: (b) before and (c) after the upstream cross section (a).

  • Fig. 14.

    Composite analysis of (a) potential temperature (K) and (b) wind speed (m s−1) at 100 m for along-coast transects past Cape Mendocino for (solid) 7, (dashed) 12, and (dashed–dotted) 26 Jun 1996.

  • Fig. 15.

    Horizontal cross sections showing boundary layer heights (m) inferred from lidar data on (a) 7 and (b) 12 Jun 1996.

  • Fig. 16.

    Horizontal cross sections showing surface-pressure- inferred (hPa) 30-m flight legs on (a) 7 and (b) 12 Jun 1996.

  • Fig. 17.

    Horizontal cross sections of 30-m wind speed (solid, m s−1) and wind direction (arrows) on (a) 7, (b) 12, and (c) 26 Jun 1996.

  • Fig. 18.

    Same as Fig. 17 but for friction velocity u∗ (solid, m s−1) and stress direction (vectors).

  • Fig. 19.

    Same as Fig. 17 but for SST (°C). (b) Dashed line shows the bottom topography of the shelf.

  • Fig. 20.

    Horizontal cross sections of vertical wind speed (m s−1) at 375 m as inferred by continuity considerations, on (a) 7 and (b) 12 Jun 1996.

  • Fig. 21.

    Vertical profiles of (a) total vertical turbulent momentum flux (m2 s−2) and (b) wind speed (m s−1) along segments of an along-coast transect on 12 Jun 1996. The lines 1–5 progress from north to south.

  • Fig. 22.

    Momentum budgets for (a) and (b) across-coast and (c) and (d) along-coast momentum downstream of Cape Mendocino on 7 Jun 1996. (a) and (c) Close to the coast and (b) and (d) offshore. The different lines indicated in (a) refer to Eq. (1) in the text. The dividing line between off- and inshore is at −127.75° longitude.

  • Fig. 23.

    As in Fig. 22 but for 12 Jun 1996.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 254 56 10
PDF Downloads 71 27 4