An Observational History of Small-Scale Katabatic Winds in Mid

Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x

An Observational History of Small-Scale Katabatic Winds in Mid-Latitudes Greg Poulos1 and Shiyuan (Sharon) Zhong2* 1Clipper Wind Power Development Inc 2Department of Geography, Michigan State University.

Abstract Katabatic winds have been the subject of investigation since about the 1840s. These winds, which flow down the topographic gradient as a result of surface cooling, provide a major transport and dispersion mechanism in mountainous regions and affect the energy exchange between the earth’s surface and the atmosphere. Various theories of their structure, evolution, and fundamental dynamics have been proposed. Initial interest in katabatic winds, which was prompted by field observations, has been followed by a long history of observational studies. This article reviews observational work undertaken on small-scale katabatic winds in mid-latitudes, with an emphasis on the historical background, and recent work on the causes of their variation.

1 Introduction From its Greek origins the term katabatic, where ‘kata’ means downward and ‘batos’ means moving beyond, can be interpreted to refer to essentially any flow that moves downward. This broad interpretation is not exactly consistent with the general use of the word, however, which typically refers to winds that flow down the topographic gradient or out of a valley due to surface cooling that gives this air a greater density than the free atmospheric air. This cooling of slope surfaces, which is due primarily to a net negative surface radiative balance, produces a temperature difference between the air adjacent to the slope and the ambient air away from the slope. Winds then accelerate from the slope toward the ambient air, where gravity forces the dense flow to follow the sloping surface (Figures 1 and 2). In a valley terrain configuration, the katabatic flows on the slopes of the valley sidewalls coagulate in the valley base where they will continue their course out valley (or down valley if the valley is sloped; Figure 1). This katabatic flow out of a valley is often called mountain wind or mountain breeze or down-valley flow (with sloped valley floor). Numerous authors have offered their opinions on which terminology is best used in

Fig. 1. Illustration of slope and valley katabatic winds driven by the differential cooling and the resulting temperature differences between the air adjacent to the slopes and away from the slopes in the valley and between the air inside and outside the valley. various circumstances (Atkinson 1981; Barry 1992; Jeffreys 1922; Talman 1911; Whiteman 1990), but in the following the term katabatic flow or katabatic wind will be used to generally represent surface-based flow caused by surface cooling on either a slope or in a valley, where the context (valley or slope) will be obvious from the topic at hand. The terms ‘downslope wind’ and ‘mountain wind’ may also be used where the former refer to katabatic flow on a slope as opposed to the latter term which refers to katabatic flow out of a valley. Downslope wind in this context should not be confused with the same terminology used in the meteorological literature to describe wind storms in the lee of a mountain barrier. Although katabatic winds have been observed over sloping terrain of different scales all over the world, including Antarctica (Ball 1956, 1957; Bromwich and Parish 1998; Mawson 1915; Rees 1991; Renfrew and Anderson 2002), Greenland (Broeke et al. 1994; Heinemann 2002; Loewe 1935), Europe (Defant 1951; Ekhart 1934; Oerlemans et al. 1999; Smeets et al. 1998; Tollner 1931), North America (Buettner and Thyer 1965; Clements et al. 1989; Doran et al. 2002; Haiden and Whiteman 2005; Horst and Doran 1986; Moni et al. 2002; Princevac et al. 2005; Tower 1903), and the Mediterranean (Martinez et al. 2006), this review will focus only on small-scale katabatic winds that occur in mountainous regions in mid- latitudes. There is a large amount of literatures on large-scale katabatic winds over Polar ice sheets, which will not be reviewed here. Note also that this article will not describe anabatic flow, the closely related daytime analogue to katabatic flow, that moves upslope or up valley due to surface heating. The observations of small-scale katabatic flows and the development of theories describing these flows have a long history whose origins began in the early 1800s. In recent decades, numerical modeling studies of katabatic

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd Observing mid-latitude katabatic flows 3 flows have proliferated, providing insights into the forcing mechanisms of the katabatic flows and their interactions with synoptic-scale winds and with the ambient environments. Due to the page limitation and the vast amount of literatures on this subject, this review article will concentrate primarily on the observational studies of the katabatic winds with an emphasis on an understanding of the historical work and the more recent work on causes of variations in such flows. Other reviews of downslope flows and mountain breezes can be found in Atkinson (1981), Barry (1992), Defant (1951), Geiger (1975), Mahrt (1982), Poulos (1996), Sturman (1987), Thyer (1966), Vergeiner and Dreiseitl (1987), and Whiteman (1990). The most thorough analysis and interpretation of the first-century (1800s through 1945) katabatic flow investigations can be found in Hawkes (1947).

2 Katabatic Flow Observations in Mountain Valleys Katabatic flows were first observed in valleys because of the natural tendency for mankind to dwell in the base of valleys where water sources are readily available. Although knowledge of the diurnal nature of valley winds had been known to the agricultural community for many years before, the first scientifically reported observations began in Europe near more populated valleys beginning with Fournet (1840) in the Savoie region of France and later in the Alps by Ekhart (1932a,b, 1934) and in England by Heywood (1933) among others. Tower (1903) was one of the first recorded observations of these mountain winds within several valleys in the northern Colorado Rockies of the United States. These observations made certain features of the katabatic winds clear: (i) they develop near sunset when the surface starts to cool, (ii) generally clear skies and quiescent synoptic conditions are most conducive to their development, (iii) they are from tens of meters to a few hundred meters deep, (iv) they form within a developing temperature inversion, (v) their wind speed generally increases from the ground to 2–6 m/sec at some fraction of the inversion height and decreases from this jet to the top of the inversion, and (vi) they are common and frequent around the world. Some of these observed general characteristics of katabatic flows are illustrated in Figure 2. Using katabatic valley flow observations taken during an ASCOT (Atmospheric Studies of Complex Terrain) field campaign in the Brush Creek Valley of Colorado, Clements et al. (1989) show an approximately 7 °C nocturnal temperature inversion in the lowest 300–400 m of the valley atmosphere with near isothermal structure above that to the ridgeline. They show that the depths of the katabatic valley flows correspond rather well with the depth of the temperature inversion. The katabatic flow jet of 6 m/sec lies at approximately 100 m above ground level (AGL) or 0.25 of inversion depth, agreeing well with the early theoretical predictions of Prandtl (1942, 1952). Davidson and Rao (1963) found this jet height to be 0.4–0.5 of inversion depth in a valley near Manchester, Vermont, and

Downslope wind

Fig. 2. Illustration of typical wind and temperature profiles in katabatic wind layer (adapted from Whiteman 2000).

suggest that the ratio becomes smaller at steeper slope angles. The observations in the Brush Creek Valley exhibit some evidence of the return flow above the katabatic flow, the so-called ‘anti-wind’, compensation current, or return flow. This feature was routinely observed by Buettner and Thyer (1965) in a valley wind system near Mt. Rainier, Washington, over four consecutive summer seasons, which is consistent with the theoretical explanation of katabatic flow (Defant 1933) based on the circulation theorem (Bjerknes 1902; Kelvin 1866). They find that the ‘anti-wind’ layer has the same depth as the down valley katabatic flow but weaker. Anti-winds are not always found in katabatic flow observational studies (Reiter et al. 1983), and for that reason the theoretical necessity of their existence is questioned (Davidson and Rao 1963; Ekhart 1932a,b). Most likely is that an anti-wind is simply difficult to observe in many cases, because synoptic or regional-scale influences generally dominate above the strongly forced near-surface katabatic flow (Defant 1951). Anti-winds are also likely to be difficult to observe in sinuous portions of a canyon. The time evolution of valley katabatic flow has been well observed by a number of studies, but generalization is difficult due to differences in the valley configurations, observational locations, and measurement techniques. For instance, the measurements of Buettner and Thyer (1965) showed a katabatic flow peaking just before sunrise, whereas Neff and King (1987) reported on a valley configuration where pooling in the lower valley

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd Observing mid-latitude katabatic flows 5 slowed katabatic flow much before sunrise. Clearly, valley katabatic flow can peak in velocity at a variety of times relative to sunrise and sunset, depending on when the along-valley pressure gradient is maximized at a particular location. Furthermore, McKee and O’Neal (1989) show that valley geometry can determine the local strength of along-valley pressure gradient. Using the ratio of valley width to cross sectional area, they find that widening and narrowing of valleys along their course can strongly influence, and even dominate, the along-valley pressure gradient generated by radiative cooling. Davidson and Rao (1963) find that the flow in a smaller valley can be dominated by the valley wind system of a larger valley in which it resides, which greatly influences the time of onset of the smaller valley’s wind system. A similar effect was suggested by Defant (1951) for the Maloja wind in Switzerland where the wind in a valley blows in the same direction day and night, and was also observed by Buettner and Thyer (1965). The evening transition from daytime upslope and up-valley flow to nighttime downslope and down-valley flow has generally been found to occur quickly – over a period of about an hour (Defant 1951; Orgill and Schreck 1985; Urfer-Henneberger 1967; Vergeiner and Dreiseitl 1987). Buettner and Thyer (1965) found that the transition occurs at approximately the same time everywhere in the valley, about 1 hr after sunset and that both the katabatic slope flow and the katabatic valley winds begin simultaneously. However, this suggestion is contrary to their finding that a spiraling wind regime can develop near sunset as one slope is anabatic while the opposite slope is katabatic. Various other studies in different valleys (e.g., Eckman 1998 in the Tennessee Valley and Banta et al. 2004 in the Salt Lake Valley) have shown this transition to be prior to, at, or well after sunset, suggesting that orientation, time of year, and terrain shading also determine when the surface heat flux will reverse (Whiteman et al. 1989a,b). North-south oriented valleys are likely to have downslope winds on the west side while the east side, which is last to be affected by decreasing in insolation toward sunset, continues to have an upslope flow (Reiter et al. 1983). In this case, it is also seen that the transition to the mountain wind from up valley flow will take at least as long as the along-valley pressure gradient takes to reverse and, for a short period, up-valley anabatic winds can exist over the top of lower-level down-valley flow. Larger valleys have been found to develop along-valley pressure gradients that are more difficult to reverse than that of smaller valleys (Davidson and Rao 1963; Wagner 1938). In a recent field study, Clements et al. (2007) documented a rapid transition from upslope to downslope flow on the west sidewall in Arizona’s Meteor Crater, a near circular basin approximately 165 m in depth and 1.2 km in diameter at the rim. With the aid of smoke grenade, a downslope flow was found to develop as soon as the west sidewall became shaded around 4 p.m. when upslope winds still prevailed on the east sidewall. The transition occurred in less than 30 min and the downslope winds during the transition

Fig. 3. Time lapse photography of smoke showing the development of a katabatic flow as on the west sidewall of the Arizona’s Meteor Crater during the late afternoon transition period on 20 October 2006. Times listed in each photo are MST. Photos by Craig Clements. period were intermittent, changing back to upslope repeatedly before becoming steady. As the smoke was transported down the slope, entrainment of ambient air was observed and the smoke disperses vertically, indicating that the downslope flow grew in depth with distance from the top (Figure 3). Valley width and depth also have important influence on the characteristics of katabatic valley winds. Buettner and Thyer (1965) found that shallow valleys are rather easily influenced by overlying flow during daytime, making the observation of up-valley flows difficult, but the down-valley katabatic flow can still be well developed. Because anabatic flows are caused by surface heating, the turbulent eddies thus created during solar heating are conducive to the mixing downward of ambient flow and the masking of the anabatic mechanism. Banta and Cotton (1981) found that daytime mixing was a regular occurrence in North Park, Colorado, but that upslope flow did occur for a time prior to the mixing event. Toward evening, when the stabilization of the atmosphere prevented turbulent mixing, katabatic flow was easily observed. McHattie (1968) found similar behavior in the Kananaskis Valley near Calgary, Canada. Even valleys so broad that the terrain appears nearly flat to humans, such as near Newark, New Jersey, have been found to be dominated climatologically by a katabatic flow regime (Davidson 1963a,b). For a deeper valley in Vermont, however, it was found that both the anabatic and katabatic flows largely control wind direction at Rutland Municipal Airport (although the anabatic portion

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd Observing mid-latitude katabatic flows 7 was frequently influenced by overlying flow). Davidson and Rao (1963) found that katabatic flow can develop even when the prevailing flow was 10–12 m/sec, whereas the anabatic flows were undetectable with similar flow aloft. Although not explicitly ignored, the cross-valley circulations crucial to down-valley katabatic flow development have not sparked the interest of nearly as many researchers as their down-valley brethren. However, cross- valley flows play an essential role in valley dynamics, particularly with respect to vertical motion, dispersion, and thermal distribution (Tang 1976; Wagner 1938). Broder et al. (1981) observed that cross-valley flow occurred as slope flows moved cross canyon toward the main valley flow. During katabatic flow, this, by mass continuity arguments, leads to upward vertical motion and cooling in the valley center. They found that the cross-valley component was important for determining the distribution and destruction rate of ozone (O3) in the valley atmosphere. Hennemuth and Schmidt (1985) found significant cross-valley circulations throughout the course of a day in a small alpine valley that were crucial to the valley heat balance. Echoing this sentiment was Vergeiner et al. (1987) who found that without the effect of cross-valley flow, the momentum budget for down-valley winds requires vertical motions on the order of 0.10 m/sec to replace lost mass in the valley center. This rate of descent corresponds to an adiabatic warming rate of 2.5 K/hr, which must then be balanced by radiative cooling and cold air advection to maintain the down-valley flow as observed; they postulate that their neglect of cross-valley flow caused an overestimate of the volume flux divergence and, by extension, the vertical motion at valley center. Adding a further complication, Buettner and Thyer (1965) found that they could measure no significant vertical component in the valley center despite conditions sufficiently quiescent for the observation of elusive return flow aloft. Using Doppler LIDAR (Light Detection and Ranging) observations in the Salt Lake Valley during the Vertical Transport and Mixing experiment (Doran et al. 2002), Banta et al. (2004) showed the occurrence of cross-valley flows, resulting from sloped drainage and canyon outflows. They showed that these cross-valley winds interact with the down- valley jet at night to generate localized regions of convergence and divergence in the Salt Lake Valley that produce vertical motions and mixing.

3 Katabatic Flow Observations over Isolated Mountain Slopes While katabatic valley winds are the most easily observed due to the amplifying effect of their confining topographic configuration, katabatic slope flow is a more basic dynamical element of valley katabatic flows. Despite the early observational emphasis on valley flows, theoretical development focused on steady-state slope flows because of the complicating factor of three-dimensionality in valleys. Sufficient observations of natural slope flows themselves lagged behind the theoretical development, however,

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd 8 Observing mid-latitude katabatic flows because of the difficulty of their observation. As is described below, relative to the long history of katabatic flow interest, katabatic slope flow observations are a recent occurrence. Perhaps the most well known of these studies is that of Manins and Sawford (1979b), who present a case study of katabatic flow in the Jeeralang Hills in southeast Australia. Although not a perfectly formed slope (a slight valley shape, 1–2 km in width with ridges of less than 50 m), the authors argue that their slope represents a typical slope in nature. The slope drops vertically 330 m at an inclination of 4.5° from the horizontal with a roughness length of about 0.02 m. The katabatic flow jet appears at 0.50 to 0.60 of inversion depth, which is relatively high compared with typical katabatic flow measurements in valleys. Comparisons made by the authors with the one-dimensional theory of Petkovsek and Hocevar (1971) were unsuccessful, because the simple theory requires an improper amount of sensible heat flux and/or radiative flux divergence in the katabatic layer to account for the observed profiles. The authors argue that there must be some influence from side flows (i.e., three-dimensional effects). An important conclusion they draw is that the effects of turbulent mixing in the shear zone above the katabatic jet are more significant than those at the surface due to the strong stability at the surface and that at the jet level the Richardson number, Ri (Richardson 1920), which represents the dimensionless ratio of buoyant suppression of turbulence to shear generation of turbulence, is at its maximum. Mahrt et al. (1979) corroborates high values of Ri at the jet level due to a minimum in vertical wind shear at that level. Another series of well-known slope flow papers were published by Doran and Horst (1983) and Horst and Doran (1986, 1988) based on a series of observations over simple slopes and theoretical comparisons. Horst and Doran (1986) find that slope flow is generated consistently when the temperature deficit of the slope falls to a sufficiently low value as to overcome any existing ambient pressure gradient. They also conclude that for simple slopes of 15–21° slope angle, the katabatic flow depth increases with distance down the slope and can be approximated as 5% of the drop in elevation from ridge top. This ‘rule of thumb’ is supported by observations made by Clements and Nappo (1983), but it does not seem to fit all observations especially those over lower angle slopes. For example, Manins and Sawford (1979b) observed a katabatic flow on a 4.5° slope that reached a depth of 80 m at where the elevation dropped by 330 m from the top, leading to a ratio of 0.24. The role of radiative flux divergence in the development of the katabatic flows was investigated by Manins (1992) using measurements from a rapid succession of tethered balloon soundings. He found that radiative flux divergence is important to both the generation and retardation of the flow. Radiative flux divergence of up to 15 °C per day at the top of the katabatic flow layer where wind shear is great helps lessen the temperature

gradient (by cooling the top of the inversion) and induce turbulence. The turbulence then can act to reduce the katabatic jet speed and inversion strength. Radiative flux divergence at the ground up to 20 °C per day was a significant factor in the generation of katabatic flow. Being a very small-scale feature, Manins (1992) suggests that no modeling study of katabatic flow has shown a significant radiative flux divergence due to their generally poor vertical grid spacing and/or insufficient radiation parameterization. A study of katabatic flows on lower angle slopes was conducted by Mahrt and Larsen (1990) who analyzed data on land that has an average slope of only 2° toward the sea near Riso, Denmark. Without ambient flow they find that katabatic flow can continue to increase after initiation for a number of hours. Often the initial katabatic surge of cold air can overcome an opposing ambient wind. This surge is initially deep but later becomes shallower and more steady with occasional and smaller pulses. As the surge, regarded as a pressure head, proceeds downslope, it slows and raises the opposing flow due to the adverse pressure gradient. The induced rising motion adiabatically cools the layer of ambient flow. Depending on the depth of the forced rising motion, this mechanism has the potential to cool a significant portion of the overlying air. The most comprehensive study on small-scale katabatic flows over gentle slopes was conducted recently in the Salt Lake Valley, Utah, as part of the Department of Energy’s Vertical Transport and Mixing Program (Doran et al. 2002). The experiment employed a line of four tethered balloons running down the topographic gradient and separated by about 1 km on a low-angle slope (~1.6°) at the foothill of the Oquirrh Mountains in the southwest part of the Salt Lake Valley (Figure 4). By analyzing the tethersonde data, Whiteman and Zhong (2008) show that downslope flows over this low-angle slope are deeper and stronger than reported previously by other investigators, who generally investigated steeper slopes and, in many cases, slopes on the sidewalls of isolated mountains where the downslope flows are not subject to the influence of nighttime buildup of ambient stability within valleys. They reported that typically on clear, undisturbed October nights, a 25-m-deep temperature deficit of 7 °C and a 100- to 150-m-deep downslope flow with a jet maximum speed of 5– 6 m/sec at 10–15 m AGL develop over the slope during the first 2 hr following sunset (Figure 5). The temperature deficit, the strength of the downslope jet and its height above the slope surface, and downslope volume flux increase with downslope distance. The deepening of the katabatic flow layer in the downslope direction is found to be small, but the increase of wind speed with the downslope distance is large. The downslope flows weaken in the late evening as the stronger down-valley flows expand to take up more of the valley atmosphere and as ambient stability increases in the lower valley with the buildup of a nocturnal temperature inversion. Using the same set of tethersonde observations, Haiden and Whiteman (2005) evaluated the layer-averaged budget terms

Fig. 4. Photograph of the low-angle (~1.6°) slope on the west side of the floor of Utah’s Salt Lake Valley below the Oquirrh Mountains where a slope flow experiment was conducted in October of 2000. TS1, TS2, TS3, and TS4 indicate the locations of each of the four tethered- balloon sounding sites over the slope; the distance between two sites was about 1 km. (Whiteman and Zhong 2008). in the momentum and heat budget equations to identify the dominant forcing in comparison with what is assumed in simplified models. They found that the katabatic downslope flow observed on undisturbed nights with a strong temperature inversion is in near local equilibrium at each site, in which buoyancy is nearly balanced by friction while radiative cooling is nearly balanced by advection of background temperature field. However, with less stable ambient stratification, the downslope acceleration is systematically stronger than what would be supported by the observed buoyancy. They attribute the stronger acceleration to confluence of flows due to cross slope heterogeneity of the sloping terrain. The studies described thus far have all dealt with katabatic flows on the scale of a few kilometers or a few tens of kilometers. Over large-scale mountain slopes such as those in western United States, katabatic winds are expected to reach scales of hundreds of kilometers. A recent meteorological

Fig. 5. Selected soundings of potential temperature (K) and downslope wind speed component (m/sec) from the four tethersonde sites TS1 (black), TS2 (dark gray), TS3 (light gray) and TS4 (dashed black) at the times (MST) indicated on 8–9 October 2000. The abscissas on the downslope wind component sub-figures run from −3 to +5 m/sec (Whiteman and Zhong 2008). field experiment revealed the frequent presence of a katabatic flow over 100 km scale on the gentle slopes going down from the Colorado Plateau to the Little Colorado River in Northern Arizona (Savage et al. 2008). Using several instrumented towers, a Doppler SODAR (Sound Detection and Ranging) with RASS (Radio Acoustic Sounding System), and frequent

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd 12 Observing mid-latitude katabatic flows radiosonde launches at night, the study shows that a downslope flow, which occurred on almost all clear nights, can be as deep as 250 m above ground, but the maximum wind speed occurs within the lowest 50 m above ground and the peak speed of 4–6 m/sec falls within the range of typical katabatic flows over gently sloped terrain.

4 Internal Variability Katabatic flows, particularly those in valleys, are subject to numerous potential sources of variability aside from those evident in previous sections. Whereas some sources of variability in katabatic flow are external (such as mountain waves and larger-scale pressure gradients), processes within katabatic flows themselves and their particular terrain configuration can also create variability. Tyson (1968) and Doran and Horst (1981) showed that this variability appears to create energy within the portion of the atmospheric energy spectrum associated with oscillations between 0.1 and 5.0 hr, thereby filling the spectral gap in atmospheric timescales observed by Panofsky and van der Hoven (1955) and van der Hoven (1957). While not always easy to delineate from external influence, numerous studies in relatively quiescent atmospheric conditions (limited external influence) have noted katabatic flow variability on timescales of less than an hour. While offering no explanation, Buettner and Thyer (1965) observed fluctuations in katabatic flow speed 100 m above the valley floor from 1.5 to 6.5 m/sec at 20–30 min intervals, which are comparable to the 5–30 min episodes found by Weber and Kurzeja (1991). Manins and Sawford’s (1979a,b) fluctuations in the ‘skin-flow’ (a few meters deep) were of the order 15 min. Start et al. (1975), during a tracer release (SF6) and oil fog photographic session in Huntington Canyon, Utah, noted that pulsations in the katabatic flow appeared to be associated with tributary canyon interaction. Citing the Fleagle (1950) mechanism (oscillatory katabatic flow fluctuations during development) as a possible cause, Tyson (1968) found that cold air pulses occurred in approximately 1 hr intervals, and also found turbulent bursts with 20 min intervals near Pietermaritzburg, South Africa. Observational statistical analysis by Doran and Horst (1981) showed fluctuations of about 1.5 hr. Nappo (1991) investigates the occurrence of sporadic breakdowns in the nocturnal stable boundary layer using observations of wind speed and temperature and their covariance. By band pass filtering at 5 and 30 min intervals, he extracts the deviation of the 5- min structure from the 30-min average structure. Based on the frequency of bursts (10–20 per night over both flat and complex terrain, more frequent over complex terrain), he speculates that a significant portion of the nocturnal fluxes are contained within these occasional bursts concurring with the conclusions of Ruscher and Mahrt (1989). Davidson and Rao (1963) also observed sporadic turbulent events as evidenced by sudden sub-jet mixing and above-jet mixing in shear zones.

Katabatic flows converging, accumulating, or combining from different source regions can also cause self-induced variability. Erasmus et al. (1993) present observations near Greeley, Colorado, of multiple scales of drainage flow interacting at that location (katabatic flows from the Cheyenne Ridge to the north, Poudre River to the west–northwest and South Platte River to the southwest). In more severe topography, Neff and King (1987, 1989) observe cold air accumulation in mountain valleys, elevated inversions, jets, and complicated drainage flow structure due to the merging of numerous katabatic flows. As a katabatic slope flow descends into a more stable cold pool near the surface, variability can be produced by the resulting gravity waves. Gryning and Lyck (1983) document this effect for a coastal draining valley on the west coast of Greenland. In this study, occasional surges of the valley floor cold air up the slope interact with the down flowing katabatic flow. The period of these surges, about 45 min, is consistent with the calculated period for gravity waves in this valley. Gravity waves in the stable surging valley bottom air were found to be amplified, as they impacted the sudden change in slope angle (where they met the katabatic flow) much as ocean waves react when they encounter the shoreline (Perkin and Lewis 1978; Stigebrandt 1976). While far less internally turbulent than their daytime counterparts (anabatic flows in convective boundary layer), katabatic flows cannot often be idealized or characterized as mostly laminar, even in quiescent atmospheric conditions. In fact, Mahrt and Gamage (1987) dissect aircraft observations of stratified flow during the Second European Stratospheric Arctic and Mid-latitude Experiment and the Alpine Experiment and find intermittent turbulence driven by shear, continuous turbulence characterized by sharp boundaries, and weaker continuous turbulence. Stratification acts to decrease the slope of the spectral energy density to values smaller than predicted by inertial subrange theory. Weak continuous turbulence seemed to be decaying, perhaps as daytime eddies weakened. The primary generation of horizontal velocity fluctuations was in the direction of the shear, but observations taken in the mountains show longitudinal modes. Self-induced katabatic flow variability was also observed by Blumen (1984) as a cold drainage density current passed the Boulder Atmospheric Observatory (BAO) on two consecutive days. Instabilities, probably gravity waves, set off by the passage of the density current, are observed to alter the vertical motion field to heights at least as high as the highest observation at 300 m AGL. Kelvin-Helmholtz waves are observed in the katabatic flow following the drainage front and, as the katabatic flow following the front continues past the BAO, the gradual breakdown of the waves into turbulence – and then to less and less turbulence – was observed. These processes involve the entrainment of ambient air into the katabatic flow. The rate of entrainment may be most easily parameterized using the Richardson number, because measurements of mean vertical motion are very difficult in such variable regimes (Blumen 1984).

Neff and King (1989), using SODAR profiles and topographic information, investigated the source of variability in katabatic wind profiles. Topographic constrictions and subsequent pooling, an effect detailed by McKee and O’Neal (1989), could elevate katabatic flows. Davidson (1961) concludes that turbulence in katabatic flows is highly sensitive to the position relative to local topographic features and the wind direction. The alternate narrowing and widening of a valley, its turns, and merging with larger valleys, are found by McKee and O’Neal (1989) to possibly cause pressure gradients that oppose the down-valley katabatic flow gradient. Utilizing the fact that the cooling rate of a particular valley cross-section depends on its volume (the topographic amplification factor), they show that certain valley configurations, such as that where the ratio of valley width to valley area decreases with down-valley distance, should drain more effectively than others. A configuration where that ratio increases at some point along the valley is conducive to accumulation and pooling because of a greater cooling rate at that point than upstream. Consequently, a pressure gradient adverse to that conducive to katabatic flow can be generated.

5 External Influence From the previous section, we understand that katabatic flows, simply due to the environment they develop within, can be far more variable than their generally idealized treatment. External meteorological influences, such as mesoscale and synoptic-scale pressure gradients, wind fields, clouds, mountain waves, and temperature gradients, are also of considerable importance to their evolution and variability (Banta et al. 2004; Orgill et al. 1992; Poulos et al. 2000; Tower 1903; Vergeiner and Dreiseitl 1987). As described by Wagner (1938), the PhD dissertation of Bondy (1935) investigated conditions of strong gradient winds and katabatic flow for the Inn Valley. Bondy (1935) found that during foehn conditions, katabatic flow was non-existent but that katabatic forcing could be found in a nighttime wind direction shift to more down valley. During the day, a wind shift in the opposite sense occurred. Other studies verify this effect (Kanitscheider 1936) and even suggest that a katabatic westerly flow will develop in the Inn Valley underneath a westerly foehn. In agreement with this concept, Banta and Cotton (1981) have shown that katabatic winds can develop underneath overlying westerlies (at least in westerlies less than 10 m/sec) as temperatures drop and an inversion develops in South Park, Colorado. Wagner (1938) suggests that it is conceivable that the pressure distribution forced by the foehn enhances the down-valley pressure gradient. Regardless of the wind phenomena in the Inn Valley, a valley-plain pressure difference that affects the flow can be detected. In deep valleys, Wagner (1938) finds that, with the exception of strong winds, the gradient wind will not generally influence the katabatic down-valley wind. He argues that the pressure gradients causing the katabatic flow are so large over such

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd Observing mid-latitude katabatic flows 15 relatively short distances that they overwhelm the synoptic pressure gradient. A climatological study of katabatic flow in Denmark (Mahrt and Larsen 1990) confirms that katabatic flows can develop underneath ambient winds from any direction, even for a slope as small as 1°. Regardless of whether the ambient flow is strong enough to prevent the downslope flow from occurring, the effect of katabatic forcing on the flow is still evident. As part of the ASCOT 1980 investigation near the Geysers area of Northern California, Neff and King (1987) found that the depth of the Pacific marine inversion strongly modulated the initiation and evolution of katabatic slope winds. They observed that buoyancy forces were able to overcome mesoscale and synoptic pressure gradients if topographic slope was steep enough (5–6°) and surface cooling large enough. Orgill et al. (1981) analyzed the observations taken during ASCOT 1980 near the Geysers area of California in the Anderson Creek drainage. They found that warm surges occur on 1–4 hr time scales within downslope flow. These surges are inconclusively attached to oscillations of elevated inversions. Other surges may occur when the marine air penetrates inland in its diurnal oscillation. Orgill and Schreck (1985) extend the analysis to find cases where migrating mesoscale and synoptic systems disrupt the katabatic flow. Modified cool marine air was found to overcome the local katabatic flows regularly during a normal diurnal cycle. Ambient flow opposing katabatic slope flow can severely impact katabatic flow development although, as found by Wagner (1938), eventually katabatic forcing due to surface cooling dominates. Balloon soundings indicate that when the gradient wind opposes the down-valley katabatic flow, a wind minimum occurs at the height where the katabatic pressure forcing decreases to that of the synoptic pressure gradient (Ekhart 1934). As described by Hawkes (1947), an investigation over small hills in England by Cornfeld (1938) showed that an opposing wind in the lee could create a calm zone along the hillside, where, without ambient flow, a katabatic flow would otherwise reside. Katabatic flow was found to exist below this calm zone but ambient flow dominated to the top of the hill. Orgill et al. (1992) use observations in Brush Creek Valley, Colorado, from the 1984 ASCOT experiment to conclude that the main factors contributing to the erosion of valley downslope wind depth are the magnitude of the opposing wind, valley stability, and the acceleration of above valley opposing flow. An acceleration of opposing ambient flow of at least 1.44 m/sec/hr is found to be conducive to turbulent erosion of katabatic winds to at least half its undisturbed depth. Otherwise, an ambient opposing flow of 5 m/sec is required to significantly decrease katabatic layer depth. A 5 m/sec wind threshold is also found by Gudiksen et al. (1992) to strongly interfere with the development of katabatic flows at Grand Mesa, Colorado. They found that at this threshold turbulent mixing of near surface temperatures did not allow enough cooling for the development of katabatic flows. Similar threshold value is also reported by Savage et al. (2008) for the beginning

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd 16 Observing mid-latitude katabatic flows of strong interference of large-scale opposing flow with downslope winds developed over the western slopes of the Little Colorado River Valley in northern Arizona. Observations of katabatic flow in opposing flow by Neff and King (1989) show a reduction in valley katabatic wind depth, an effect that is corroborated by Mursch-Radlgruber (1995) and Savage et al. (2008). Although shallower, Mursch-Radlgruber’s analyses of tethersonde data indicate, however, that the wind speed in the remaining katabatic flow layer is not reduced significantly in opposing ambient wind conditions as compared to the undisturbed conditions, indicating that katabatic forcing is actually increased relative to quiescent conditions. The stronger thermal forcing in the shallow katabatic flow layer under opposing ambient wind condition is consistent with the results from a numerical study by Arritt and Pielke (1986) on the interaction of ambient winds with katabatic flows. For a gradient wind in the same direction as the katabatic wind, Wagner (1938) finds that a minimum still occurs in the vertical profile of horizontal wind, but it is located where the sum of the ambient and katabatic pressure gradients minimizes. Hawkes (1947) claims that the strong stability of katabatic flow prevents aiding ambient flow from having a very deep effect. Cornfeld’s (1938) observations of katabatic flow with supporting ambient winds indicated that a recirculatory eddy on the lee side of a hill could manifest itself either immediately near the hill top preventing katabatic flow there or further down the hill side creating a calm at that point (where katabatic slope flow exists between hilltop and the calm). Such an effect has been observed in a much more sophisticated observational fashion by Mursch-Radlgruber (1995), which suggests that the dynamic pressure gradient caused by flow over topography in a high mountain Austrian valley opposes the katabatic flow producing a calm or even flow reversal. In valleys, the effect of supporting winds on katabatic flows can be variable. For Coal Creek Canyon, Colorado, Coulter and Gudiksen (1995), based on analyses of 3 years of observational data from a SODAR at the mouth of Coal Creek Canyon, report that the katabatic flow depth (250 m) and jet height (70 m) maximize with an external aiding wind of 3 m/sec, but the katabatic flow depth decreases when the ambient wind drops below this value. The katabatic flow speed maximum increases as external aiding flow increases above 3 m/sec. Neff and King (1987; 1989) show that ambient winds can have varying effects when intruding into a tributary versus the main canyon depending on orientation. In this case, as aiding upper-level winds increased to 7 m/sec, there is a halving of the katabatic flow speed in the main valley, similar to the Mursch-Radlgruber (1995) observations. In a side-valley perpendicular to the overlying flow, katabatic flow is nearly eliminated. In addition to ambient wind speed and direction, other external factors, such as clouds, vegetation, soil type and soil moisture, and snow cover, may also influence the properties of katabatic flows. Because katabatic flows are

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd Observing mid-latitude katabatic flows 17 forced primarily by the cooling of the slope surface, these factors can have significant impacts on katabatic flow. Barr and Orgill (1989) observed katabatic canyon flows that decreased to 25% of their typical depth during low cloud ceiling conditions. Due to the extreme difficulty of separating the influence of each factor in an observational setting, few observational studies have actually examined physiographic effects on katabatic flows. Conversely, atmospheric boundary layer and mesoscale numerical models (Banta and Gannon 1995; Savage et al. 2008) have proven to be efficient tools in understanding the relationship between properties of katabatic flow and the physiographic conditions of the slopes on which they develop.

6 Current Limitations and Future Work As the review above shows, katabatic flow observations have come a long way. Prior to the 1980s, the observations on katabatic flows had been ‘almost all confined to wind fields and only rarely was information about their spatial and temporal variation included. Practically no temperature data were available’ (Manins and Sawford 1979b). Since the 1980s, significant advances have been made as new and improved observational tools, especially remote sensing tools, such as Doppler SODARs and LIDARs and RASS, become available. These new instruments provided much-needed continuous wind measurements above the ground and temperature information that have furthered our understanding of the vertical structure and evolution of the katabatic flows and their interactions with temperature inversion. The Doppler LIDAR’s ability to scan a large area has enabled detailed observations of larger-scale katabatic flows in wide valleys and how they interact with katabatic flows on valley sidewalls (Banta et al. 1999, 2004). The observations of katabatic flows have reached a stage where we can now describe in great details the characteristics of katabatic flows and their interactions with ambient stability and large-scale winds. The increased use of resources, however, makes the planning of field observations logistically quite complicated, involving coordination of multiple investigators. The outcome has been comprehensive data sets for a few geographical locations over short-time periods of a few days to one month. These data sets have been used, together with numerical modeling, to gain insights into the physical processes responsible for the formation, maintenance, and dissipation of katabatic winds. They are, however, limited to a few topographic settings, vegetation covers, and weather and climate conditions. Future research should benefit from long-term observations on slopes in a variety of topographical and environmental settings. The observations need to be made not just near the slope surfaces, but also the entire katabatic wind layer with sufficient vertical resolution. This may be achieved by a combination of in situ measurement platforms, such as instrumented towers and remote sensors like Doppler LIDAR and SODAR and RASS. In addition to observations of wind, temperature, and humidity, enhanced

© 2008 The Authors Geography Compass 2 (2008): 10.1111/j.1749-8198.2008.00166.x Journal Compilation © 2008 Blackwell Publishing Ltd 18 Observing mid-latitude katabatic flows measurements of energy budget components over slope surfaces are also desirable to help answer some important questions regarding katabatic flow development and variation. Emphasis should also be placed on detailed measurements of longwave radiative flux divergence and turbulence in the katabatic flow layer, because relatively little is known about their roles in the maintenance and dissipation of katabatic winds and in causing the internal variability of the flow. In addition to wind speed, direction, and stability, the impact of other external factors, such as vegetation and soil moisture, deserves further attention. Finally, there is a growing reorganization of combining observations with fine-scale numerical modeling to achieve a full understanding of the internal and external dynamics of the katabatic winds. Smith et al. (1997) has pointed to the consequences of thermally driven circulations, including the katabatic flows, as a major obstacle to transpor- tation, land use planning, and air pollution management within mountain regions. In Salt Lake City, Utah, a thermally driven downslope flow was found to be correlated with fluctuations of aerosol particles less than 10 microns in diameter (Alexandrova et al. 2003). High ozone concentrations in Mexico City were also found to be a result of nighttime katabatic flow, where high levels of pollutants would drain toward the populated urban areas (Raga et al. 1999). Beyond air pollution, katabatic winds may impact emergency planning or other geographically sensitive activities. A study in China has examined the observed downslope flow around a nuclear power plant for its potential role in nuclear fall out dispersion (Sang et al. 1999). Ecologists are beginning to take notice of the effect of katabatic flow, as well. Insect migration is a major concern for mountainous regions in malaria-troubled Africa, and understanding thermally driven circulations can aid in tracking small insects movements (Burt and Pedgley 1997). The enhanced knowledge of the circulation has helped researchers to introduce new efficient methods for monitoring the ecosystem. Pypker et al. (2007) has utilized the high climatological frequency of katabatic flows within a region as a cost-effective means of monitoring CO2 concentrations as air parcels exit the base of the slope. This method was proven to work, although vertical gradients of CO2 were often dependent on the varying characteristics of the downslope flow. Continued research and understanding of katabatic winds will undoubtedly enhance understanding of these positive and negative impacts on environments and ecosystems.

Acknowledgements We would like to acknowledge all authors whose work was cited in this review article. We also thank L. Crosby Savage, III, and Wenqing Yao for their help in preparing the illustrations. Finally, we are grateful to the two anonymous reviewers for their helpful comments and suggestions. Partial support for the time spent on this review article was provided by US National Science Foundation by Grants ATM 0646206 and 0646299.

Short Biographies Gregory S. Poulos received a BS in Meteorology from Cornell University in 1989 and an MS in Atmospheric Science from Colorado State University in 1991. He began work at Los Alamos National Laboratory in 1992 and completed a PhD in Atmospheric Science from Colorado State University in 1996. His research in this period concerned atmospheric flows in complex terrain, including a detailed literature review of katabatic wind study, high-resolution mesoscale modeling and large-eddy simulations of near-surface winds, and atmospheric observations using a variety of instrument platforms such as towers, SODAR, LIDAR, and profiling atmospheric radar. In 1996, he joined Colorado Research Associates, where, among other numerical and observational work, he led the CASES-99 tall tower and remote sensing field experiment studying the winds and turbulence in stable boundary layer over the plains of Kansas. This field study has resulted in over 100 scientific publications and over 10 doctoral and master’s theses. In 1999, he co-founded the operational numerical weather prediction company Foresight Weather, LLC. In 2003, he joined the National Center for Atmospheric Research as manager for Research Technology Facility whose purpose was to develop and deploy in situ and remote sensing observational instruments for atmospheric science research worldwide. He joined Clipper Windpower in July 2007 and retains an affiliate faculty position in the Meteorology Department at the University of Utah. Shiyuan (Sharon) Zhong is an associate professor in the Department of Geography at Michigan State University, Michigan, USA. Prior to joining Michigan State University in 2006, she was an associate professor in the Geosciences Department at the University of Houston, Texas, USA. Before that, she was a senior research scientist at the US Department of Energy’s Pacific Northwest National Laboratory. She received her PhD in Atmospheric Science from Iowa State University, Iowa, USA. Dr. Zhong’s research encompasses many areas of atmospheric sciences, including complex terrain meteorology, land-atmosphere interactions, boundary layer and mesoscale processes, atmospheric transport and dispersion, fire-atmosphere interactions, and regional climate studies. Her research involves both field observations and numerical modeling. She has authored or co-authored more than 50 peer-reviewed articles in these areas in a variety of professional journals. She has been serving as an editor for Journal of Applied Meteorology and Climatology published by the American Meteorological Society.

Note * Correspondence address: Sharon Zhong, Department of Geography, Michigan State University, 116 Geography Building, East Lansing, MI 48824, USA. E-mail: [email protected].

References Alexandrova, O. A., et al. (2003). The influence of thermally driven circulation on PM10 concentration in Salt Lake Valley. Atmospheric Environment 37. pp. 421–437. Arritt, R. W., and Pielke, R. A. (1986). Interactions of nocturnal slope flows with ambient winds. Boundary Layer Meteorology 37, pp. 183–195. Atkinson, B. W. (1981). Meso-scale atmospheric circulations. New York: Academic Press, p. 495. Ball, F. K. (1956). The theory of strong katabatic winds. Australian Journal of Physics 9, pp. 373–386. ——. (1957). The katabatic wind of Adelie Land and King George V Land. Tellus 9, pp. 201– 208. Banta, R. M., and Cotton, W. R. (1981). An analysis of the structure of local wind systems in a broad mountain basin. Journal of Applied Meteorology 20, pp. 1255–1266. Banta, R. M., and Gannon, P. T. Sr. (1995). Influence of soil moisture on simulations of katabatic flow. Theoretical and Applied Climatolology 52, pp. 85–94. Banta, R. M., et al. (1999). Wind flow patterns in the Grand Canyon as revealed by Doppler LIDAR. Journal of Applied Meteorology 38, pp. 1069–1083. ——. (2004). Nocturnal low-level Jet in a mountain basin complex. Part I: evolution and effects on local flows. Journal of Applied Meteorology 43, pp. 1348–1365. Barr, S., and Orgill, M. M. (1989). Influence of external meteorology on nocturnal valley drainage winds. Journal of Applied Meteorology 28, pp. 497–517. Barry, R. G. (1992). Mountain weather and climate, 2nd edn. New York: Routledge. Bjerknes, V. (1902). Zirkulation relativ zu der Erde. Meteorologische Zeitschrift 19, pp. 97–108. Blumen, W. (1984). An observational study of instability and turbulence in nighttime drainage winds. Boundary Layer Meteorology 28, pp. 245–269. Bondy, F. (1935). Über beziehungen zwischen periodischen talwinden und verschiedenen meteorologischen faktoren [Relationships between periodic valley winds and various meteorological factors]. PhD Dissertation, University of Innsbruck, Institute of Meteorology and Geophysics, Innrain 52, A-6020, Innsbruck, Austria. Broder, B., Dutsch, H. U., and Graber, W. (1981). Ozone fluxes in the nocturnal planetary boundary layer over hilly terrain. Atmospheric Environment 15, pp. 1195–1199. Broeke, M. R. van den Duynkerke, P. G., and Henneken, E. A. C. (1994). Heat, momentum and moisture budgets of the katabatic layer over the melting zone of the West Greenland ice sheet in summer. Boundary Layer Meteorology 71, pp. 393–413. Bromwich, D. H., and Parish, T. R. (1998). Meteorology of the Antarctic. In: Karoly, D. J. and Vincent, D. G. (eds) Meteorology of the southern hemisphere: meteorological monographs, Vol. 27, Issue 49. Boston, MA: American Meteorological Society, pp. 175–200. Buettner, K. J. K., and Thyer, N. (1965). Valley winds in the Mount Rainier area. Archiv fiir Meteorologie Geophysik und Bioklimatologie, Series B 14 (H. 2), pp. 125–147. Burt, P. J. A., and Pedgley, D. E. (1997). Nocturnal inset migration: Effects of local winds. Advances in Ecological Research 27, pp. 61–92. Clements, C. B., et al. (2007). Slope flows observed during METCRAX. The 29th International Conference on Alpine Meteorology, Chambéry, France, 4–8 June 2007. Clements, W. E., Archuleta, J. A., and Hoard, D. E. (1989). Mean structure of the nocturnal drainage flow in a deep valley. Journal of Applied Meteorology 28, 457–462. Clements, W. E., and Nappo, C. J. (1983). Observations of a drainage flow event on a high- altitude simple slope. Journal of Applied Meteorology 22, pp. 331–335. Cornfeld, C. E. (1938). Katabatic winds and the prevention of frost damage. Quarterly Journal of Royal Meteorological Society 64, pp. 553–584. Coulter, R. L., and Gudiksen, P. (1995). The dependence of canyon winds and surface cooling and external forcing in Colorado’s Front Range. Journal of Applied Meteorology 34, pp. 1419– 1429. Davidson, B. (1961). Valley wind phenomena and air pollution problems. Journal of the Air Pollution Control Association 11, pp. 364–383. ——. (1963a). Reflections on some recent observations of local wind systems. Transactions of the New York Academy of Sciences, Series II 25, pp. 663–673. ——. (1963b). Some turbulence and wind variability observations in the lee of mountain ridges. Journal of Applied Meteorology 2, 463–472.

Davidson, B., and Rao, P. K. (1963). Experimental studies of the valley-plain wind. Air and Water Pollution 7, pp. 907–923. Defant, A. (1933). Der Abfluss schwerer Luftmassen auf geneigten Böden, nebst einigen Bemerkungen zur Theorie stationärer Luftströme. [The flow of heavy air masses down slopes along with some remarks on the theory of stationary air currents] Sitzungsberichte der Preuβischen Akademie der Wissenschaften, Physikalisch-Mathematische Klasse 18, pp. 624–635. Defant, F. (1951). Local winds. In: Malone, T. F. (ed.) Compendium of meteorology. Boston, MA: American Meteorological Society, pp. 655–672. Doran, J. C., and Horst, T. W. (1981). Velocity and temperature oscillations in drainage winds. Journal of Applied Meteorology 20, pp. 361–364. ——. (1983). Observations and models of simple nocturnal slope flows. Journal of Atmospheric Sciences 40, pp. 708–717. Doran, J. C., Fast, J. D., and Horel, J. (2002). The VTMX 2000 campaign. Bulletin of American Meteorological Society 83, pp. 1233–1247. Eckman, R. (1998). Observations and numerical simulations of winds within a broad forested valley. Journal of Applied Meteorology 37, pp. 206–219 Ekhart, E. (1932a). Zur aerologie des berg und talwinds, ergebnisse von pilotballonaufstiegen in Innsbruck. Beitrage zur Physik der freien Atmosphare 18, pp. 1–26. ——. (1932b). Weitere beiträge zum problem des berg und talwindes. Beitrage zur Physik der freien Atmosphare 18, pp. 242–252. ——. (1934). Neuere untersuchungen zur aerologie der talwinde. Beitrage zur Physik der freien Atmosphare 21, pp. 245–268. Erasmus, D. A., et al. (1993). Observations of the nocturnal stable boundary layer in the Greeley area. Faculty Research and Publications Board, University of Northern Colorado, Denver, CO. Fleagle, R. G. (1950). A theory of air drainage. Journal of Meteorology 7, pp. 227–232. Fournet, M. J. (1840). Des brises de jour et de nuit autour des montagnes. [Daytime and nighttime winds around mountains]. Annales de chimie et de physique 74, pp. 337–401. Geiger, R. (1975). The climate near the ground. Cambridge, MA: Harvard University Press, 5th printing, p. 611. Gryning, S. E. (1984). Oscillating natural slope flow in a coastal valley in Greenland. Polarfront 39, pp. 8–12. Gryning, S. E., and Lyck, E. (1983). A tracer investigation of atmospheric dispersion in the Drynaes valley, Greenland. RISO-R-148, RISO National Laboratory, Roskilde, Denmark. Gudiksen, P. H., et al. (1992). Measurements and modeling of the effects of ambient meteorology on nocturnal drainage flows. Journal of Applied Meteorology 31, pp. 1023–1032. Haiden, T., and Whiteman, C. D. (2005). Katabatic flow mechanisms on a low-angle slope. Journal of Applied Meteorology and Climatology 44, pp. 113–126. Hawkes, H. B. (1947). Mountain and valley winds with special reference to the diurnal mountain winds of the Great Salt Lake Region. PhD dissertation, Ohio State University, Columbus, OH, p. 312. Heinemann, G. (2002). Aircraft-based measurements of turbulence structure in the katabatic flow over Greenland. Boundary Layer Meteorology 103, pp. 49–81. Hennemuth, B., and Schmidt, H. (1985). Wind phenomena in the Dischma Valley during DISKUS. Archiv fiir Meteorologie Geophysik und Bioklimatologie, Series B 35, pp. 361–387. Heywood, G. S. P. (1933). Katabatic winds in a valley. Quarterly Journal Royal Meteorological Society 59, pp. 47–58. Horst, T. W., and Doran, J. C. (1986). Nocturnal drainage flow on simple slopes. Boundary Layer Meteorology 3, pp. 263–286. ——. (1988). The turbulence structure of nocturnal slope flow. Journal of Atmospheric Sciences 45, pp. 605–616. Jeffreys, H. (1922). On the dynamics of wind. Quarterly Journal Royal Meteorological Society 48, 29–48. Kanitscheider, R. (1936). Beiträege zur mechanik des föehns [Contributions to the mechanics of the foehn]. Beitrage zur Physik der freien Atmosphare 18, pp. 27–35. Kelvin, L. (1866). On stationary waves in flowing water. Philosophical Magazine 5 (22), pp. 353– 357, 445–452, 517–530. Loewe, F. (1935). Das klima des grönländischen inlandeises. Handbuch der Klimatologie 2, pp. 67–101.

Mahrt, L. (1982). Momentum balance of gravity flows. Journal of Atmospheric Sciences 39, pp. 2701–2711. Mahrt, L., and Gamage, N. (1987). Observations of turbulence in stratified flow. Journal of the Atmospheric Sciences 44, 1106–1121. Mahrt, L., et al. (1979). An observational study of the structure of the nocturnal boundary layer. Boundary Layer Meteorology 17, pp. 247–264. Mahrt, L., and Larsen, S. (1990). Relation of slope winds to the ambient flow over gentle terrain. Boundary Layer Meteorology 53, pp. 93–102. Manins, P. C. (1992). Vertical fluxes in katabatic flows. Boundary Layer Meteorology 60, pp. 169– 178. Manins, P. C., and Sawford, B. L. (1979a). A model of katabatic winds. Journal of Atmospheric Sciences 36, pp. 619–630. ——. (1979b). Katabatic winds: a field case study. Quarterly Journal Royal Meteorological Society 105, pp. 1011–1025. Martinez, D., Cuxart, J., and Jimenez, M. A. (2006). Katabatic wind over gentle slope on the Majoarca island. 17th Symposium on Boundary Layers and Turbulence, 22–27. May, San Diego, CA. Mawson, D. (1915). Home of the Blizzard. London: Heineman. McHattie, L. B. (1968). Kananaskis Valley winds in summer. Journal of Applied Meteorology 7, pp. 348–352. McKee, T. B., and O’Neal, R. D. (1989). The role of valley geometry and energy budget in the formation of nocturnal valley winds. Journal of Applied Meteorology 28, pp. 445–456. Moni, P., et al. (2002). Observations of flow and turbulence in the nocturnal boundary layer over a slope. Journal of the Atmospheric Sciences 59, pp. 2513–2534. Mursch-Radlgruber, E. (1995). Observations of flow structure in a small forested valley system. Theoretical and Applied Climatology 52, pp. 3–17. Nappo, C. J. (1991). Sporadic breakdowns of stability in the PBL over simple and complex terrain. Boundary Layer Meteorology 54, pp. 69–87. Neff, W. D., and King, C. W. (1987). Observations of complex terrain flows using acoustic sounders: experiments, topography and winds. Boundary Layer Meteorology 40, pp. 363– 392. ——. (1989). Observations of complex terrain flows using acoustic sounders: drainage flow structure and evolution. Boundary Layer Meteorology 43, pp. 15–41. Oerlemans, J., et al. (1999). Glacio-meteorological investigations on Vatnajokull, Iceland, summer 1996: an overview. Boundary Layer Meteorology 92, pp. 3–26. Orgill, M. M., Kincheloe, J. D., and Sutherland, R. A. (1992). Mesoscale influences on nocturnal valley drainage winds in Western Colorado valleys. Journal of Applied Meteorology 31, pp. 121–141. Orgill, M. M., and Schreck, R. I. (1985). An overview of the ASCOT multi-laboratory field experiments in relation to drainage winds and ambient flow. Bulletin of American Meteorological Society 66, pp. 1263–1277. Orgill, M. M., Schreck, R. I., and Whiteman, C. D. (1981). Synoptically driven downslope winds and their effects on local nocturnal drainage air flow in the Geysers geothermal resource area. Preprints, Second Conference on Mountain Meteorology, November 9–12. Panofsky, H. A., and van der Hoven, I. (1955). Spectra and cross-spectra of velocity components in the mesometeorological range. Quarterly Journal of Royal Meteorological Society 81, pp. 603–606. Perkin, R. G., and Lewis, E. L. (1978). Mixing in an arctic fjord. Journal of Physical Oceanography 8, pp. 873–880. Petkovsek, Z., and Hocevar, A. (1971). Night drainage winds. Archiv für Meteorologie, Geophysik und Bioklimatologie, Serie B A20, pp. 353–360. Poulos, G. S. (1996). The interaction of katabatic winds and mountain waves. Ph.D. dissertation. Atmospheric Sciences Department: Colorado State University. Poulos, G. S., et al. (2000). The interaction of katabatic flow and mountain waves. Part I: observations and idealized simulations. Journal of Atmospheric Sciences 57, 1919–1936. Prandtl, L. (1942). Führer durch die strömungslehre. Braunschweig, Germany: Viewig und Sohn, pp. 367–375. ——. (1952). Essentials of fluid dynamics. Hafner, New York: Hafner Publishing Company.

Princevac, M, Fernando, H. J. S., and Whiteman, C. D. (2005). Turbulent entrainment into nocturnal gravity-driven flow. Journal of Fluid Mechanics 533, pp. 259–268. Pypker, T. G., et al. (2007). Using nocturnal cold air drainage flow to monitor ecosystem processes in complex terrain. Ecological Applications 17, pp. 702–714. Raga, G. B., et al. (1999). Some aspect of boundary layer evolution in Mexico City. Atmospheric Environment 33, pp. 5013–5021. Rees, J. M. (1991). On the characteristics of eddies in the stable atmospheric boundary layer. Boundary Layer Meteorology 55, pp. 325–343. Reiter, R., et al. (1983). Aerological investigations of diurnal mountain winds with special consideration of wind fields in the valley cross section. Meteorologische Rundschau 36, pp. 225– 242. Renfrew, I. A., and Anderson, P. S. (2002). The surface climatology of an ordinary katabatic wind regime in coats land, Antarctica. Te ll u s 54A, pp. 463–484. Richardson, L. F. (1920). Some measurements of atmospheric turbulence. Philosophical Translation. The Royal Society. London Series A 221, pp. 1–28. Ruscher, P., and Mahrt, L. (1989). Coherent structures in the very stable atmospheric boundary layer. Boundary Layer Meteorology 47, pp. 41–54. Sang, J., Guanming, L., and Zhang, B. (1999). Numerical modeling for emergency response of nuclear accident. Journal of Wind Engineering and Industrial Aerodynamics 81, pp. 221–235. Savage, L. C., et al. (2008). An observational and numerical study of regional scale downslope flow in Northern Arizona. Journal of Geophysical Research 113, D14114, pp. 1–17. Smeets, C. J. P. P., Duyankerke, P. G., and Vugts, H. F. (1998). Turbulence characteristics of the stable boundary layer over a mid-latitude glacier. Part I: a combination of katabatic and large-scale forcing. Boundary Layer Meteorology 87, pp. 117–145. Smith, R., and Co-authors (1997). Local and remote effects of mountains on weather: research needs and opportunities. Bulletin of American Meteorological Society 78, pp. 877–892. Start, G. E., Dickson, C. R., and Wendell, L. L. (1975). Diffusion in a canyon within rough mountainous terrain. Journal of Applied Meteorology 14, pp. 333–346. Stigebrandt, A. (1976). Vertical diffusion driven by internal waves in a sill fjord. Journal of Physical Oceanography 6, pp. 486–495. Sturman, A. P. (1987). Thermal influences on airflow in mountainous terrain. Progress in Physical Geography 11, pp. 183–206. Talman, C. F. (1911). Some mountain winds and their names. Scientific American 71 (Suppl. 1831), pp. 80–86. Tang, W. (1976). Theoretical study of cross-valley wind circulation. Archiv für Meteorologie, Geophysik und Bioklimatologie, Serie B A25, pp. 1–18. Thyer, N. H. (1966). A theoretical explanation of mountain and valley winds by a numerical method. Archiv für Meteorologie, Geophysik und Bioklimatologie, Serie B A15, pp. 318–348. Tollner, H. (1931). Gletscherwinde in den ostalpen. Meteorologische Zeitschrift 48, pp. 414– 421. Tower, W. S. (1903). Mountain and valley breezes. Monthly Weather Review 31, pp. 528–529. Tyson, P. D. (1968). Velocity fluctuations in the mountain wind. Journal of Atmospheric Sciences 25, pp. 381–384. Urfer-Henneberger, C. (1967). Zeitliche Gesetzmässigkeiten des Berg- und Talwindes. Veröffentlichungen der Schweizerischen Meteorologischen Zentralanstalt 4, pp. 245–252. van der Hoven, I. (1957). Power spectrum of horizontal wind speed in the frequency range 0.0007 to 900 cycles per hour. Journal of Meteorology 14, pp. 160–164. Vergeiner, I., and Dreiseitl, E. (1987). Valley winds and slope winds – Observations and elementary thoughts. Meteorology and Atmospheric Physics 36, pp. 264–286. Vergeiner, I., Dreiseitl, E., and Whiteman, C. D. (1987). Dynamics of katabatic winds in Colorado’s Brush Creek Valley. Journal of Atmospheric Science 44, pp. 148–157. Wagner, A. (1938). Theorie und Beobachtung der periodischen Gebirgswinde [Theory and observation of periodic mountain winds]. Beiträge zur Geophysik 52, pp. 408–449. Weber, A. H., and Kurzeja, R. J. (1991). Nocturnal planetary boundary layer structure and turbulence episodes during the Project STABLE field program. Journal of Applied Meteorology 30, pp. 1117–1133.

Whiteman, C. D. (1990). Observations of thermally developed wind systems in mountainous terrain. In: Blumen, W. (ed.) Meteorological monograph, atmospheric processes over complex terrain, 23, no. 45. Boston, MA: American Meteorological Society, pp. 5–42. Whiteman, C. D. (2000). Mountain meteorology: fundamentals and applications. New York: Oxford University Press. Whiteman, C. D., Allwine, J. K., Fritschen, L. J., Orgill, M. M., and Simpson, J. R. (1989a). Deep valley radiation and surface energy budget microclimates. Part II: energy budget. Journal of Applied Meteorology 28, pp. 427–437. ——. (1989b). Deep valley radiation and surface energy budget microclimates. Part I: radiation. Journal of Applied Meteorology 28, pp. 414–426. Whiteman, C. D., and Zhong, S. (2008). Downslope flows on a low-angle slope and their interactions with valley inversion. Part I: observations. Journal of Applied Meteorology and Climatology 47, pp. 2023–2038.