The potential for a stably stratified air mass upstream of the Sierra Nevada (California) to descend as foehn into the nearly 3-km-deep Owens Valley was studied for the 2 March 2006 case with observations from sondes, weather stations, and two aircraft flights. While upstream conditions remained almost unchanged throughout the day, strong diurnal heating on the downstream side warmed the valley air mass sufficiently to permit flow through the passes to descend to the valley floor only in the late afternoon. Potential temperatures of air crossing the crest were too warm to descend past a virtual floor formed by the strong potential temperature step at the top of the valley air mass, the height of which changed throughout the day primarily due to diurnal heating in the valley. The descending stably stratified flow and its rebound with vertical velocities as high as 8 m s−1 were shaped by the underlying topography and the virtual valley floor.
1. Introduction and overview of the event
Foehn, as defined by the WMO (1992), is a “wind warmed and dried [adiabatically] by descent, in general on the lee side of a mountain.” Crucial to foehn is that the virtual potential temperature of the descending upstream air mass is at least as low as the virtual potential temperature in the downstream valley. This requirement was shown by Mayr and Armi (2008) for the Alps with data from the Mesoscale Alpine Programme (Mayr et al. 2004). As a consequence, the diurnal frequency of foehn reaches a maximum in the early afternoon, when climatologically the valley air mass is warmest through diurnal heating (e.g., Mayr et al. 2007, their Fig. 9).
Here we document the diurnal cycle of the foehn for the case of 2 March 2006 in the Sierra Nevada, California, with the unique dataset of intensive observation period 1 (IOP1) of the Terrain-Induced Rotor Experiment (T-REX; Grubišić et al. 2008). The Sierra Nevada is a quasi-2D mountain range with a crest height of approximately 4 km MSL with a few high passes (Fig. 1). On its eastern face the terrain drops steeply by about 3 km to Owens Valley but only 2.4 km from Kearsarge Pass. The White and Inyo mountain ranges enclose the valley air mass to the east. Crest-to-crest distance is about 30 km, and the valley floor width is 15 km. We will use radiosondes upstream; downstream radiosondes at Independence, California (Fig. 1); dropsondes from the High-Performance Instrumented Airborne Platform for Environmental Research (HIAPER) aircraft; and automatic weather stations (AWSs) high up on the sierra slope (Notch) and in the valley (ISS2), together with in situ measurements from the University of Wyoming King Air aircraft, to study the case.
A time series of potential temperature from AWSs up in the canyon, which runs down from Kearsarge Pass (Notch), and in the valley (ISS2) is depicted in Fig. 2 to give an overview of the event of 2–3 March 2006. Local standard time (LST), which is 8 h behind UTC, will be used herein. Southwesterly flow at crest level persisted throughout the day (cf. inset of Fig. 1). The valley atmosphere was very stably stratified in the morning and remained stable after sunrise through the afternoon. The slope station stayed warmer than the valley atmosphere (boldface squares) at the same altitude. Clouds covered the area upstream of the Sierra Nevada, and potential temperatures, based on upstream soundings, remained lower at the elevation of Kearsarge Pass than at Notch. Shortly before 1600 LST (0000 UTC), potential temperatures at Notch began to fall. Based on wind speed and direction data and their temporal change, foehn began at Notch. As the potential temperature at Notch dropped further and fell both below the value at ISS2 and also below the value over the valley, weak foehn began at ISS2. The 10-min-average wind speeds were no more than 7 m s−1. After sunset the valley bottom cooled more strongly than Notch so that the foehn no longer reached ISS2 in the valley.
2. Diurnal stratification change of the valley air
In the course of the day the stratification in the valley atmosphere changed from stable with a strong potential temperature step at middepth to one mixed up to the altitude of Kearsarge Pass. Figure 3 traces the diurnal evolution of the vertical profile of potential temperature in the valley.
At 0900 LST the valley was still stably stratified from nighttime cooling with the exception of the lowest 250 m. As is typical in complex terrain, the profile was characterized by a complicated layered structure of weakly stratified residual layers, separated by thin strongly stratified interfacial steps. Strong surface heating had begun to erode this structure from underneath. A strong step in potential temperature of 6 K over only 400 m existed at 2.6 km MSL. Above this step, the valley air had weak stable stratification to above crest height up to another interface at 4.3 km MSL. Three hours later, at 1200 LST, the lower valley atmosphere was mixed up to the interfacial step at 2.5 km. Convective mixing associated with strong sensible heat fluxes in this semiarid region had occurred. The layer was unstable with interspersed superadiabatic gradients. The layer above the interface had become slightly colder. The interface was eroded not only from below, as in undisturbed convective boundary layer (CBL) evolution, but also from above. Fifty minutes later the CBL had grown deeper and lifted the interface by about 200 m. Combined with the cooling above the interface, the interface step was only half as large as in the early morning. An hour later, at 1350 LST, the interface base had risen an additional 200 m. The cooling aloft had also continued, turning the initially sharp step into a diffuse transition zone spread over a larger depth.
At 1500 LST the base of the interface is at about the altitude of Kearsarge Pass. Shortly afterward foehn on the west side of the valley began (cf. Fig. 2). The profile is no longer unstable up to the interface but has become slightly stable, with the exception of the immediate surface layer. At 1800 LST the interface base had descended to 2.9 km MSL due to warming aloft. Underneath, the stability had increased slightly, especially near the surface through nocturnal cooling.
3. Causes of the limited descent
Changes below as well as above the stability interface had influenced temperatures in Owens Valley and thus the ability of upstream air to descend to the valley floor. A comparison of potential temperature in upstream and downstream soundings in Fig. 4 explains why the descent of upstream air was limited and did not reach the valley floor for most of the day. A novel depiction of the three-dimensional wind field from two King Air flights in Figs. 5 and 6 enables one to visualize the accompanying flow field using cones that point along the direction of the 3D wind vector, are scaled to wind speed, and are color coded with vertical velocity.
At 0900 LST only air close to the altitude of Kearsarge Pass and above could cross the sierra. Upstream air at the altitude of Kearsarge Pass could descend adiabatically only to the interface in the valley, as indicated by the arrow in Fig. 4. The upper limit of the descending layer was at about crest height, as indicated by the second arrow tracing the adiabatic descent to the valley interface. The interface step of 6 K formed a virtual floor for the incoming upstream air and limited the descent to about 1 km. Above the crest the possible drop is only about 0.5 km. At 1200 LST the depth of the upstream cold air had increased sufficiently for very shallow flow across the crest to also descend to the valley interface. The larger depth upstream and the slight lowering of the valley interface top allowed overall a larger descent than at 0900 LST.
The King Air flew the first flight approximately during the time between these radio soundings. Conditions well above the valley floor changed little during the flight, as seen in the 0900 and 1200 LST radio soundings. Up to the valley interface at about 2.6 km MSL, the flow is upvalley across Owens Valley (Fig. 5). Higher up, flow on the western slope of the sierra could descend, but the deepest descent was only to 2.6 km MSL, the altitude of the potential temperature step seen in the profiles of the downstream radio soundings at 0900 and 1200 LST radio soundings in Fig. 4. The descent remained confined to the eastern sierra slope. Farther east across the valley, the flow was upvalley and perpendicular to the descending flow to about the altitude of Kearsarge Pass. The stably stratified flow sampled with the King Air up to 8.5 km MSL descended as seen in Fig. 5. The descent and the subsequent rebound are a stratified flow response, not a free wave. The horizontal distance associated with the descent from the sierra crest to the intersection with the valley interface at 2.6 km MSL is 4 km. This sets the quarter wavelength for the rebound of the descending air. Consequently, the whole wavelength should be 16 km. The rebound is seen as upward vertical velocities (red cones in Fig. 5) up to the highest aircraft flight leg at 8.5 km. The vertical velocity maxima were aligned vertically. A weaker secondary undulation is seen downstream. The distance between descents (blue cones) is 18 km along the aircraft track, which translates to 16 km when the angle between aircraft track and wind direction is taken into account. Note that the horizontal extent of the stratified flow response of 16 km is about 3 times as long as the free-oscillation wavelength 2πU/N of 5 km (U = 15 m s−1, N = 1.9 × 10−2 s−1).
At 1500 LST, potential temperatures of the upstream air at the altitude of Kearsarge Pass were finally colder than the valley air mass, which had been convectively mixed from underneath throughout the day. Shortly afterward, foehn in the valley began (cf. Fig. 2). Upstream potential temperature at crest height, however, was too high for the air to descend by more than about 0.4 km (cf. arrow in Fig. 4). Since the foehn was confined to flow below mountain crest it was so-called shallow foehn and a gap flow (Armi and Mayr 2007).
At 1800 LST, potential temperature of upstream air at the altitude of Kearsarge Pass was still low enough to descend into the higher potential temperatures of the valley (Fig. 4). Air from crest level, however, could only descend to about 3 km MSL, since the valley interface had risen in the course of the day from the 2.6 km during the morning flight. The second King Air flight took place between the 1500 and 1800 LST soundings in the afternoon. During its second stage, shown here, the flight tracks ran farther north (cf. Fig. 1) across Sawmill Pass, the next pass north of Kearsarge Pass. Because the mixed valley air mass was deeper in the afternoon, the horizontal distance from the crest to the intersection with the valley interface at 3 km MSL is shorter: about 3 km. With this quarter wavelength the total response distance of 12 km is less than in the morning, as seen in Fig. 6. There was only one undulation. Maximum vertical velocities in the rebound were 8 m s−1.
The 3D wind field in Fig. 6 shows that flow below the crest across the gap was cold enough to plunge down beneath the valley interface, as is visualized by the downward-pointing blue wind cones. This can also be seen in the potential temperature profiles on the upstream and downstream sides at 1800 LST in Fig. 4. In the lowest flight leg at 1.8 km MSL (roughly 600 m AGL), the foehn extended about halfway across the valley, undercutting the valley air mass before the wind direction changed from westerly to a southerly upvalley direction again. At higher flight legs from 2.2 km upward, the foehn flow was confined to a narrow swath along the eastern sierra slopes. Flow across most of the valley was upvalley and thus perpendicular to it. The flight legs along the sierra slopes show that outside of the terrain fold extending down from the pass, the flow had separated higher up on the sierra slopes and did not extend down to the valley floor.1 These airstreams had to cross the higher sierra crest, and their potential temperatures were warmer than the air in the lower portion of the valley. Therefore, foehn at the valley floor was not encountered all the way along the foot of the eastern sierra slope but only emanated as foehn streaks from the gaps cut into the Sierra Nevada.
At night (2100 LST) the valley atmosphere had cooled sufficiently so that potential temperatures were too cold for the air crossing the passes to descend into Owens Valley. The arrows in the rightmost soundings in Fig. 4 show that adiabatic descent from the upstream layer at altitudes of Kearsarge Pass was limited to an altitude of 3.2 km in Owens Valley. Air crossing the crest could descend to about the same altitude. As was shown in the time series from the AWS in Fig. 2, foehn at ISS2 in the valley had already ceased at the time of these soundings.
4. Discussion and conclusions
Conditions upstream of the Sierra Nevada alone would not have been sufficient to lead to foehn in the almost 3-km-deep Owens Valley without the aid of diurnal heating of the valley atmosphere. At pass and crest level, a synoptic system (cf. inset in Fig. 1) brought colder air upstream (west) of the Sierra Nevada. Only in the afternoon could this air descend to the valley floor because of the diurnal heating of the valley air mass from underneath. This is similar in other foehn regions of the world [e.g., Inn Valley in the Alps (Mayr et al. 2007)] and leads to maxima of foehn onset in the afternoon. The gap flow portion of the 2 March situation shares this aspect with the Washoe zephyr, which occurs on the eastern slopes of the Sierra Nevada on summertime afternoons under high pressure conditions (Zhong et al. 2008a). In the case of the zephyr, the colder air at gap and crest level west of the Sierra Nevada is provided by the coastal air mass over the San Joaquin valley, which does not warm as much during the day as the air over the elevated heat source east of the Sierra Nevada. The thermal difference manifests itself in a surface pressure difference across the sierra.
Because of the large 3-km depth of the valley, the occurrence of lower potential temperatures at pass and particularly crest level than at the valley floor are expected to be rare. Indeed, Zhong et al. (2008b) found in a climatological study that high wind events in the Owens Valley are predominantly up and down valley. The only locations where they found westerly high-wind events were at Independence and Bishop, California, both of which lie downstream of passes. We expect a higher frequency of foehn events at the ISS2 location at the foot of the eastern Sierra Nevada than at the weather station 7 km farther east across the valley used in Zhong et al. (2008b) because the shallow overflow through Kearsarge Pass becomes diluted with valley air on its way across the valley. This can be seen here in Fig. 6 and in greater spatial detail in De Wekker and Mayor (2009, their Fig. 6d) using a rapidly scanning aerosol lidar. The lidar location was at the eastern foot of the sierra near ISS2. The lidar backscatter data placed the valley interface at the same altitude as the radio sounding profiles (Fig. 4) and showed a similar time of foehn onset in the valley (Fig. 2).
The 2 March event confirms the finding from the Alps that foehn occurs when potential temperatures at the downstream valley floor become at least as high as the potential temperature on the upstream side at barrier height (Mayr and Armi 2008). The strength of cross-barrier flow remained almost unchanged before and after in the Alpine cases and the event here and is thus only of secondary importance. A forecaster will need to predict the evolution of both upstream barrier height temperature and downstream valley temperature correctly in order to successfully forecast foehn.
The rebound of the air plunging down the Sierra slope to the valley interface and subsequent undulations differ from the highly idealized studies (Jiang and Doyle 2008, and references therein), which use thin uniform boundary layers across low O(100 m) mountains. The situation here involves an almost 3-km-deep valley with different air masses upstream and downstream. (cf. Fig. 4).
We have shown the diurnal heating of the 3-km-deep Owens Valley air mass with a sequence of radiosondes–dropsondes. These are compared with radiosondes from the San Joaquin valley upstream of the Sierra Nevada showing fairly constant stratification and westerly flow at crest height. Upstream remained mostly cloud covered during the day, and no CBL evolved. Because of diurnal heating the valley air mass became warm enough in the late afternoon to permit flow through the passes to descend to the valley floor. Higher up, the descent of the stratified flow across the sierra is limited by a strong potential temperature step at the top of the colder valley air mass, the height of which changes throughout the day, primarily due to diurnal heating in the valley.
This study would not have been possible without the careful planning and flying of the University of Wyoming King Air operations group. As mission scientist, LA is particularly thankful for their willingness and ability to carry out these turbulent flights close to the eastern sierra slope. We thank Stephen Mobbs and Ralph Burton and their team from the Institute for Atmospheric Science, University of Leeds, for the automatic weather station data funded by the Natural Environment Research Council (NERC), United Kingdom. Access to the ECMWF data was provided through the Zentralanstalt für Meteorologie und Geodynamik, Austria. Georg Mayr’s participation in T-REX was funded by Austrian Science Foundation Grant P18940-N10. Writing of the manuscript was supported by the Alpine Research Centre Obergurgl of the University of Innsbruck. Larry Armi’s participation was partly funded by The University of California and by Lucky Larry’s Auto Repair. We thank the reviewers and editor for their careful review of the manuscript.
Corresponding author address: Georg Mayr, Institute of Meteorology and Geophysics, University of Innsbruck, Innrain 52, 6020 Innsbruck, Austria. Email: email@example.com
This article included in the Terrain-Induced Rotor Experiment (T-Rex) special collection.
These legs are partly obstructed in the perspective shown in Fig. 6.