A SINKHOLE FIELD EXPERIMENT IN THE EASTERN ALPS

Observations from a high-altitude limestone sinkhole have helped improve understanding of temperature inversion formation and minimum temperatures in basins.

Over the last 25 years, a series of studies has investigated the diurnal formation and destruction cycle of temperature inversions in valleys and basins (for a review see, e.g., Whiteman 1990). In most of the valleys and basins that have been studied to date, advective influences produced by the well-known, diurnally reversing up- and down-valley wind systems have greatly complicated the analyses (e.g., Maki and Harimaya 1988; Mori and Kobayashi 1996). Thus, in recent years some investigators have attempted to find experimental locations in mountain areas with sloping terrain where inversion evolution occurs in the presence of up- and downslope wind systems, and in the absence of the larger-scale up- and down-valley wind systems. An ideal location for these studies would be a naturally confined basin without an outlet stream or river in which a constant-elevation ridgeline completely encircles the basin. Geomorphologie analogs for such an idealized experimental environment include open pit mines, meteor craters, and sinkholes. Sinkholes, which form in many types of terrain, are most frequently found in karst topography where solution of the underlying rock results in the collapse of underground caverns or drainage channels. Idealized small sinkholes or dolines are a natural laboratory for the study of the formation, maintenance, and dissipation of temperature inversions, especially during fair-weather conditions when the sinkholes are isolated from flows aloft by the surrounding mountains and high atmospheric stability. In such conditions, sinkholes and small basins are well known in many parts of the world for producing extreme minimum temperatures, and thus for forming intense inversions.

This paper gives an overview of a set of temperature inversion experiments conducted from October 2001 to early June 2002 in a set of limestone sinkholes of different sizes and shapes on Austria's Hetzkogel Plateau (see Fig. 1). It provides a review of previous and recent experimental studies on this plateau, providing information on the design of the recent experiments, the topography, the instrumentation used, the wintertime snow cover, the synoptic weather conditions, and the initial meteorological analyses. Separate papers have already been published by the experiment participants that extend the basic research findings reported here. The observational data with a spatial and temporal resolution never achieved before allow for a much-improved quantitative investigation of the thermodynamic and dynamic processes in and above the sinkhole and represent the ideal basis for the validation of very high resolution simulation models.

REVIEW OF LITERATURE. Because literature on temperature inversions and heat loss in small basins was reviewed recently by Clements et al. (2003), we will concentrate our survey on the history of measurements on the Hetzkogel Plateau. The basis for research on this plateau was established in 1906 when Carl Kupelwieser opened the Biological Station in Lunz, Austria, (Sauberer 1952), a station where regular meteorological observations had been taken since 1898. The primary goal of the Lunz Biological Station was to explore the rivers and lakes in the area, because there are three lakes of completely different types very close to Lunz. Shortly thereafter, biology and bioclimatology investigations were extended into the mountains south of Lunz, the so-called Durrensteinstock, a limestone plateau with an altitude of about 1200-1878 m MSL. In 1927, Franz Ruttner, the long-time director of the Biological Station, started to build up a microclimatic observation network with thermographs. In cooperation with Wilhelm Schmidt and the Central Institute for Meteorology and Geodynamics [Zentralanstalt f�r Meteorologie und Geodynamik (ZAMG)] in Vienna, a measuring network containing 13 stations was installed in the mountainous area south of Lunz (Schmidt et al. 1929; Schmidt 1930, 1933; Lauscher 1937). These stations were equipped with thermohydrographs and a total of six air and ground extreme thermometers, which were mounted at different heights. The thermographs were serviced once a week either on foot or, under harsh winter conditions, by skis to change the paper charts. In a short time, many interesting results were found in this diverse area. Particularly surprising were the data from the Gruenloch Sinkhole where, in the first winter, -460C was recorded by thermometers, mounted at the wall of a wooden shelter in the deepest part of the basin. Because the reliability of the instruments was called into question at these temperatures different types of alcohol thermometers were installed. These measured even lower temperatures in the following winter. In the following years, microclimatic stations were set up at other locations in the area, including at the top of the highest mountain, Mount D�rrenstein (1878 m MSL). Because of the interesting results, minimum thermometers were operated in the Gruenloch for 14 yr in total (from 1928/29 to 1941/42). Readings of the minimum temperature were taken every 2-4 weeks, depending on weather conditions. In 8 of these 14 winters, temperatures below -5O0C were recorded, with the lowest value being -52.60C, which occurred between 19 February and 4 March 1932 (Aigner 1952). This temperature still stands as the lowest certified temperature minimum in Central Europe. Such extremely low temperatures appear to occur only when snow cover is present, because this minimizes the upward flux of heat from the ground, which normally helps to counter a net longwave loss from the sinkhole surface.

Meteorological measurements were suspended in the Gruenloch during World War II, but, during this time, the German armed forces used the sinkhole for testing engines under extreme conditions, similar to those in Siberia. After the war, Sauberer and Dirmhirn (1954,1956) conducted new measurements in the Gruenloch, following up on Aigner's research with the objective of investigating inversion structure in the (lruenloch with better instruments. Tethered balloons were used during two measuring campaigns in February/March 1953 and October 1954, and free-air measurements were compared with temperature measurements made on the basin's sidewalks. In addition, radiation balance measurements were taken and the wind distribution was studied using pilot balloons.

A tew years later (1960/61), the Gruenloch area was again a target of meteorological research. I.itschauer (1962) installed thermographs in several sinkholes of different sixes on the plateau and used the resulting temperature series as the basis for his dissertation on cold air pools in alpine valleys and basins. His main goal was to compare basins and sinkholes with different geometries to determine the factors that lead to the buildup of extreme temperature inversions. He found relations between the depth, volume, and elevation of different sinkholes and minimum temperatures.

Meteorological research in the I,un/ area diminished following Litschauer's work, because the research emphasis at the Biological Station changed to limnology. After 40 yr, a new era of research in this area has started, with cooperation between the land owner, Mr. R Kupelwieser, the University of Vienna, and the l.unz Biological Station. The scientific aim was to use modern technology to collect a new dataset to he ahle to study still-open questions with regard to cold-pool formation and breakup. Since the winter season of 1001/02, temperature measurements have been taken continuously in the Gruenloch area. On a longer term the data will allow for an ideal comparison between historic and recent temperature minima in the light of global warming, because this location was, and still is, absolutely unaffected by human activity. 1-urthermore, it becomes possible with modern observation strategies to investigate the three-dimensional energy and mass budget within a sinkhole with a never-achieved accuracy. Is the turbulent erosion on top of the cold-air pool the most important factor that leads to an early nighttime temperature equilibrium, or are seiches with quasi-periodic swashing of air over the lowest saddle the way in which this is achieved? Are there slope winds still effective after inversion formation? Why is there hardly ever fog formation observed in the sinkhole despite a significant cooling below the eveningdewpoint temperature? The instrumentation of sinkholes with different sizes and shapes, as well as the comparison to the results of field campaigns concerning sinkholes in other part of the world, promises a precise answer on how geometry, size, and climatic conditions determine the diurnal cycle of temperature. This is important to prove whether high-resolution numerical models are able to simulate the cooling and inversion breakup over complex terrain correctly.

METHODOLOGY. The Gruenloch topography. The study area ( 1-ig. 2) is on the 1 let/kogel Plateau, a limestone plateau in the eastern Alps approximately 5 km south of I.un/. The plateau contains many sinkholes, or dolines, of various sizes. The largest of the dolines is the Gruenloch (in previous literature this was called the Gstettner-Alm doline), a near-circular sinkhole of about 1-km diameter and approximately 150-m effective depth, with slope angles of 15�0 -20�. The ridgeline of this sinkhole (Fig. 3) is about 5.5 km in circumference. The lowest opening in the ridgeline, the Lechner Saddle (1324 m MSL), is 54 m above the floor (1270 m MSL) of the doline northwest of the basin center. On the far side of this saddle is the Lechnergraben [Lechner Gorge (LG)], which falls steeply into the adjacent Ybbs Valley, with a valley floor height of about 560 m MSL.

Other dolines of various sizes are found on the Hetzkogel Plateau (Table 1); in addition to the Gruenloch (DO), four of these (D1-D4) were instrumented. The next largest, the seekopfalm, is located on a pass on the southeast edge of the Gruenloch. Two others (D2 and D3) are located side by side northeast of the Gruenloch, and a fourth doline (D4) is located north-northeast of the Gruenloch.

The plateau, a remote area inhabited by wildlife, is best known for its highest mountain, the Duerrenstein (1878 m MSL, some 3 km south of the Gruenloch). In summer, the pastures on the plateau are used for grazing cattle. No permanent habitations are found on the plateau, but a mountain hut [the Ybbstaler Hut (YH)], which is used mostly in summer, provides access to the Gruenloch via a 2-km hiking trail that crosses a saddle to the southwest of the sinkhole. Access to the Gruenloch in winter is either from the Ybbs Valley up a hiking trail through the Lechner Gorge, or up a road (winter trail) to the Ybbstaler Hut.

The climate and vegetation of the Gruenloch. The eastern Alps are in a temperate climate affected by both continental and maritime influences. The climate of the Hetzkogel Plateau is well known from extensive studies conducted over many years by the Lunz Biological Station (Lauscher 1937; Sauberer 1947, 1948). While the lowest elevations in the Ybbs Valley (e.g., Lunz) typically receive less than 1750 mm of precipitation annually, the average annual precipitation on the Hetzkogel Plateau is usually in the range from 2000 to 2750 mm (Sauberer 1947). Much of this falls as snow during the winter months. The average date of first snowfall at Lunz is 1 November, and continuous snow cover typically lasts there from 20 December through 3 March (Lauscher and Roller 1952). Continuous snow cover lasts considerably longer, approximately to mid-May, at the higher elevations of the Hetzkogel Plateau.

Because snowfall and snow cover are expected to provide important influences on the inversions that form in the sinkholes on the Hetzkogel Plateau, we have estimated (Fig. 4) the daily snow cover in the Gruenloch and at the passes in the Hetzkogel Plateau for the 17 October 2001 through 4 June 2002 experimental period from snow cover and snowfall data measured at Lunz and the surrounding mountaintop observatories (Rax and Feuerkogel; Table 2), considering wind direction, wind strength, and synoptic setting. The estimates were adjusted based on field experience from the monthly data-downloading visits to the Gruenloch.

The sinkholes on the Hetzkogel Plateau are formed in limestone bedrock. In the Gruenloch, where most of our experiments were performed, a deep humus layer is found on the basin floor, but bare limestone can be seen at places on the basin slopes. The low minimum temperatures at the sinkhole floor hinder the growth of trees. Only grasses and other subalpine herbaceous plants that can be covered by a well-insulating snow cover can survive such hostile conditions in winter. Despite the humid soil on the basin floor, which would otherwise be suitable for trees, tall conifers grow only on the crests of the basin and on the higher-elevation slopes. On the lower slopes, there is a sparse cover of dwarf pine trees, which one would normally find only at higher elevations at the same latitude. On the lowest point of the basin there is a small pond (less than K) m in diameter) that contains water all year, but is snow covered in winter.

Experimental design. The experiments were designed to investigate the buildup and breakdown of temperature inversions or cold-a i r pools in a con lined topography where along-valley advection (a complicating factor) plays a minor or nonexistent role. Multiple dolines of different shapes and sizes were selected to determine the role of topography on cold-pool evolution. A dense network of meteorological instruments was used to monitor the diurnal inversion evolution over the October-June period to identify external events that affected the regular development of the inversion, including toehn winds, cloud cover, and frontal passages; and internal factors affecting the heat budget of the doline atmosphere, including soil moisture, snow cover, and tog. The spatial resolution was on the order of a few meters vertically along the slope of the dolines, with 5-min temporal resolution to detect sudden temperature changes. Special intensive observations were conducted on 2-4 June 2002 to investigate the detailed evolution of the vertical wind and temperature structure over the largest doline, to relate it to temperatures measured at the same time on the sidewalks, and to investigate the mechanisms leading to temperature inversion evolution.

Instrumentation. The meteorological instruments used in the sinkhole experiments and their operating characteristics are listed in Table 2. Locations of the instruments are shown in Fig. 2. The primary data came from a dense network of Hobo H8 Pro temperature dataloggers. These dataloggers (Hobos) are self-contained, battery-powered devices that use a thermistor sensor. Operational and laboratory tests for these dataloggers have been described by Whiteman et al. (2000). The datalogger locations and their periods of operation are shown in Table 3. Temperature dataloggers (Hobos) were installed primarily at the floors and lowest saddles of five dolines; on three main lines that ran up the sidewalls of the Gruenloch to the northwest (NW), southeast (SE), and southwest (SW) of the basin center; and on a line in the Lechner Gorge. The logger sites were named using two-digit identifiers (IDs) for the main lines (NW, SE, or SW), for shorter lines in either the LG or in smaller dolines (D1-D4), or for individual sites at the basin center (GL) or at the Ybbstaler Hut (YH). The two-digit ID is followed by a two-digit number running from the lowest to the highest site on each line. Where two instruments were at the same site but at different heights above ground a suttix indicates whether the instrument is in the upper (U) or lower (L) position. The Hobo thermistor sensors were exposed in RM Young six-plate radiation shields on wooden poles at heights of 1.4 m AGL. At several sites, however, additional Hobos (with a U suffix) were installed al 2.2-ni height on the slime poles. Data were sampled and stored at 5-min intervals. Heavy snowfall resulted in many of the Hobos being completely covered by snow for much of the winter, starting in mid-December, and some of the Hobos were buried so deeply that they could not be found for data downloading during the monthly visits to the sites. Hobos at sites SEOl, NW06U, NWlOU, SE14U, SW08, and YHOl, because of their exposures or their placement at 2.2 m, were above the snow cover for most of the winter. By mid-April, most ol'the Hobos had emerged from the snowpack and almost all of the loggers were operational during the 2-4 June intensive observational period (IOP).

Additional measurements came from three automatic weather stations (Wl, W2, and W3; see Fig. 2) that recorded temperature, relative humidity, wind speed, and wind direction at 1- or 5-min averaging intervals, depending on deployment date. The weather stations were located on the Gruenloch floor, on the l.echner Saddle, and on the Kleiner H�hnerkogel (a peak located high on the southwest sidewall of the (iruenloch). The operating characteristics of the automatic weather stations are given in Table 2. The three sites are identified in Table 4, along with the periods of record. Sensor heights were as follows: pressure at 1 m; temperature, humidity, and radiation at 1.5 m; and wind speed and direction at 2.5 m. A pyranometer was in use at one of the weather station sites. Heavy snow cover and cloudiness affected the operation of the weather stations occasionally when solar panels became covered with snow and could not recharge the station's batteries. Thus, the weather station data were intermittent throughout the winter, although they were operating well during the June IOP.

Tethersondes were operated in intensive experiments conducted on 2-4 June 2002. They recorded temperature, wet-bulb temperature, pressure, wind direction, and wind speed as the balloon ascended and descended through the basin atmosphere under operator control at speeds of typically 0.1-0.3 m s '. The operating characteristics are shown in Table 2. The locations and periods of record are provided in Table 4.

The intensive observational period. Observations during the intensive experimental period were primarily made during the nighttime periods of 2-3 and 34 lune. The intensive observations in the (lruenloch during these nights supplemented the continuous measurements from the temperature datalogger and weather station networks. In these nighttime experimcnts, two tethered balloon sounding systems (Tl and T2; see Fig. 2), three pairs of pyrradiometers, sonic anemometers, and other special supporting equipment were operated. The tethersondes made frequent (2-4 times per hour) up and down soundings of the nocturnal inversions. On the first night, one tethersonde was operated at site Tl, starting at about 1832 Central European Standard lime (CEST) and ending the next morning at 0617 CEST. On the following night, two tethersondes were operated at sites Tl and T2 starting at about 1750 CEST and ending the next morning at 0704 CEST. Other sup porting instruments used during the I OP included upward and downward-looking pairs of Schenk pyrradiometers at the basin floor, the Lechner Saddle, and the Kleiner Hiihnerkogel.

Synoptic weather conditions were somewhat different for these two experimental nights. For the 2-3 June period, the Lun/. area was under a dry high pressure ridge that extended southward into the Alps from southern Sweden. Scattered stratocumulus and cumulus dissipated early in the evening, and the remainder ot the night was undisturbed, with clear skies and weak synoptic winds, providing an ideal night lor studying an undisturbed inversion formation. In contrast, the 3-4 June experimental period was somewhat disturbed. Lun/. was in an area of weak pressure gradients, but with a weak warm front and moisten air approaching from the west. Cirrus increased in the early evening to cover 4/8 to 7/8 of the sky between 1800 and 2200 CEST. The onset of the cirrus clouds was accompanied by a strengthening of the synoptic-scale wind, which kept the inversion from building as quickly as on the previous night. During this time, however, cooling continued at the floor of the Gruenloch, wisps of fog formed, and the small pond began to steam. The cirrus band moved through, the winds in the Gruenloch weakened, and the cooling in the doline strengthened when the cirrus dissipated around 2200 CEST, leaving clear skies and light winds that persisted for the remainder of the night. At sunrise, the humidity in the cold pool was near saturation, patches of tog were present, and the entire floor of the doline was covered with rime. For reference, astronomical sunset occurred at 1946 CEST, while astronomical sunrise occurred at 0410 CEST.

RESULTS. In this section, we present a selection of basic results from analyses of the data.

Minimum temperatures. The extreme minimum teniperature recorded during the winter deployment was -40.90C at SEOl, which was then 2.5 m above the snow-covered bottom of the sinkhole on 4 January 2002 following a fresh snowfall (Fig. 4). The total snow cover in the sinkhole was about 2.5 m deep. The minimum temperature probably would have been about I0C' lower at the lowest elevation station in the sinkhole (CiLOO), but it was buried under snow during this cold episode. The synoptic situation was governed by an intense high pressure cell over central Europe with weak winds in the eastern Alps region, which was built up after a frontal passage on 2 January 2002 with some fresh snow. The temperature in the free atmosphere on 4 January 2002 was not very low (approximately -2O0C at 3000 m MSL), hut was very dry (dewpoint temperatures at 3000 m were about -350C).

Inversion strengths. Temperature inversions formed frequently in the Gruenloch throughout the experimental period and, even in midwinter, underwent regular daily buildup-breakup cycles. An indication of the relative strengths of the Gruenloch inversions throughout the experimental period is provided in Hg. 5, which shows the difference between the maximum and minimum temperatures, respectively, of sites SE�4U minus SEOl. These differences are the maximum and minimum differences determined from concurrent 5-min temperature values measured at the two sites between 1300 CEST on one day and 0900 CEST on the next. The two sites were separated by 85 m in the vertical and were above the snowpack all winter. Positive maximum values of temperature difference indicate that inversions formed on that day; negative minimum values on a given day indicate that the inversion was broken on that day.

Inversions formed in all seasons, but the highest frequency of formation in the experimental period was over snow-free ground in the tall and over fresh snow in the early winter. The strongest inversions generally produced temperature differences of 15�C or more over the 85-m elevation difference. The inversions were generally destroyed every day, except for multiday inversions that formed during a short period in early December and during a longer period in the early half of January. Fresh snow cover (Fig. 4) was present during these multiday events. In midwinter the bottom and the north-facing slopes ot the sinkhole receive very little solar insolation, so that under a clear sky, fresh snow, and calm conditions there is insufficient energy I lux to break up the inversion.

Inversion buildup of 18-19 October (ideal night). Temperature inversions built up in the Gruenloch and other sinkholes on many nights in all seasons. An example of inversion buildup in the Gruenloch during a clear, non-snow-covered period in the fall is shown in Fig. 6. Here, temperature time series are plotted for each of the Hobos from the NW line, as well as the Hobo on the basin floor at GLOOU. On the floor of the sinkhole, temperatures dropped from 17�C in the midafternoon to -9�C by sunrise, for a diurnal temperature range of 26�C. The basin floor temperature fall, which was very rapid from 1500 to 1800 CEST, moderated after 1800 CEST. Temperatures at the upper altitudes of the sinkhole were isothermal and maintained a temperature of about 9�C through most ot the night. Temperatures increased rather uniformly with altitude at about 0.15 K m ' between the floor and the upper altitudes of the Gruenloch. The very rapid breakup of the inversion after sunrise is indicated by the rapid warming at all sites and the homogenization of temperatures at all sites by about 0900 CEST. During the daytime, temperatures vary from site to site because of the varying exposures of the Hobos to sunlight and shadows from surrounding topography and trees. The basic diurnal temperature cycle in basins and sinkholes is known since 1921 when F. D. Young published his measurements from the Pomona Valley of California (Geiger et al. 2003). The high spatial and temporal resolution of the present experiments with extra measurements (Fig. 2) allows for additional insight into the microscale of the governing processes.

Inversion disturbance of 19-20 October (upper disturbance). The ideal case of an undisturbed night is relatively rare and could be observed in 8 out of 228 nights. An example for an upper disturbance is given in Fig. 7. The inversion buildup starts out quite similar to the previous night (Fig. 6); however, around midnight several temperature records show an increase of up to 6�C for about 3 h. This effect smoothes out for the lower-situated temperature probes, and finally at the basin floor the effect has almost disappeared. Also, the higher-elevated stations situated outside of the cold-air pool do not show any disturbances. We hypothesize that turbulent erosion resulting from increasing winds above the cold-air pool was responsible tor this behavior.

Inversion disturbance of 8-9 December (lower disturbance). A different type of disturbance is presented in Fig. 8. On this date the area was snow covered. A round 1700 CEST the temperature at the basin floor started to rise by 5�C. The effect dampened out for higher elevations. This effect could be observed again around 0300 CFST, but was much weaker. Temporary cloud cover might be the reason for this behavior. Further investigations of the different possible types of disturbances are under way.

Inversion breakup by subsidence, 4 June. Nocturnal inversions in the sinkholes break up after sunrise showing patterns similar to those observed in deeper valleys (Whiteman 1982; Miiller and \Vhiteinan 1988), but the inversion breakup occurs over a shorter time period. The 4 June 2002 postsunrise breakup of the nocturnal temperature inversion in the Gruenloch is depicted in Fig. 9 from a series of tethersonde soundings. At about sunrise (0420 CEST), the nocturnal inversion had developed a linear temperature increase with elevation from -4�C at the floor of the sinkhole to +10�C at the top of the inversion 100 m above basin floor (ABF). The breakup was accomplished both by a descent at the top of the inversion and a more or less uniform rate of warming within it caused by subsidence warming as a compensating response to slope wind circulations. Although the soundings were discontinued before the inversion was completely destroyed, the destruction was complete by about 0715 CEST. This pattern of inversion breakup was previously reported for deep Colorado valleys in winter when snow cover was present (Whiteman 1982). The very moist soil in the Ciruenloch in June apparently reduces the rate of release of sensible heat flux so that the summertime pattern in the Gruenloch is similar to the wintertime pattern in the Colorado valleys.

Inversion breakup by turbulent erosion, 1516 November. A single episode of inversion destruction by turbulent erosion was observed in a shallow high-altitude sinkhole, but not in a nearby deeper sinkhole. The nocturnal inversion in the 26-m-deep seekopfalm Sinkhole was destroyed by an episode ot turbulent erosion during the night of 15-16 November 2001 (Fig. 10). Shown in the figure are temperature traces from Hobos located on the floors, midsidewalls, and lowest passes ot both the seekopfalm and Gruenloch Sinkholes. In the early evening the cooling rates were similar for the pairs of Hobos located on equivalent exposures of the two sinkholes. At about 2000 CEST, however, the temperature at the seekopfalm Saddle stopped dropping relative to the Lechner Saddle as turbulent mixing produced by strong winds aloft coupled the temperatures at the seekopfalm Saddle with those aloft. The increase of the upper-level winds were caused by a frontal wave that passed the Alps to the east. This mixing eroded slowly downward into the seekopfalm temperature inversion, reaching the midslope station and causing a rapid rise in temperature at 2230 CHST, and, finally, reaching the seekopfalm floor at 2330 CEST. Hobos in the adjacent, but deeper, Gruenloch Sinkhole showed no turbulent erosion during this event. While there were no indications ot complete inversion destruction by turbulent erosion in the Gruenloch Sinkhole during the nearly 8-month experimental period, there were other isolated events in which the upper elevations at this sinkhole were subject to mixing events and warming episodes.

Minimum temperatures in sinkholes of different size. The five instrumented sinkholes covered a range of si/.es and shapes (Table 1), but, surprisingly, two ol the sinkholes at different elevations and with different drainage areas and volumes experienced nearly identical cooling curves and temperature minima on many nights. Figure 11 shows examples of this behavior for the Gruenloch and seekopfalm Sinkholes for 18-19 October 2001. The similar temperatures in the two sinkholes are also seen in Fig. 10 for 1516 November 2001. The reason for the similar cooling curves at these two sinkholes, and their differences from the other three sinkholes, are presently under investigation. Two of the hypotheses for the different behavior are the different sky-view factors and the whole-day shading ol the small, steep sink holes during winter, which results in significantly different surface energy budgets.

SUMMARY AND OUTLOOK. Meteorological experiments focused on gaining a better understanding ot cold-air pools or inversions have been conducted in a set of five sinkholes of various sixes and shapes on the Hetzkogel Plateau south ofLun/, Austria. The largest sinkhole, the Gruenloch, some K)O m deep, 1 km in diameter, and at an elevation of 1300 m MSL, is the site where the lowest surface minimum temperature in Central Europe (-52.60C) was recorded.

This paper provides an overview ot the temperature inversion experiments run in the sinkholes from 17 October 2001 through 4 June 2002, reporting on the routine, continuous observ�t ions made in the sinkholes with temperature dataloggers and automatic weather stations, as well as on the special intensive experiments using tree-air soundings in the Gruenloch with tethered balloon systems that were conducted on 2-4 June 2002. The previous meteorological research in the sinkholes and the present experimental design lias been described, along with the topography and climate ot the sinkholes and the synoptic weather situation during the intensive experiments. The types of instruments utilized, and their operating characteristics, locations, and periods of record, have been summarized in tables.

The extreme minimum temperature recorded in the Gruenloch during the experiments was 0 -40.9�C. This occurred on a clear, windless night under dry atmospheric conditions 2 days after a fresh snowfall on 4 January 2002. Temperature inversion strengths in the Gruenloch during the experimental period varied from night to night, depending on the synoptic weather conditions. An indicator of inversion strength in the Gruenloch was the temperature difference between a station located 5 m above the basin floor and a station 90 m above the floor. Temperature differences between these stations exceeded 15�C mi the strongest inversion nights, which occurred predominantly in fall and early winter. There were often sequences of nights with well-formed inversions. The nocturnal inversions broke up after sunrise on most days, but there were two periods in the winter following fresh snowfalls when the inversions persisted for multiple days. A fall case of inversion buildup in the Gruenloch showed the very rapid cooling of air in the sinkhole in the late afternoon, with the rate of cooling decreasing during the night to reach quasi equilibrium near sunrise. Inversion breakup after sunrise was investigated using a series of tethersonde soundings on 4 June 2002. The breakup was accomplished by � continuous descent of the top of the inversion and warming in the sinkhole through subsidence heating. This pattern of inversion destruction has been previously observed in other deeper valleys, but was previously seen only in snow-covered valleys in winter. The time required lor inversion breakup was shorter in the sinkhole than previously observed in other valleys. An un usual inversion destruction event was documented in the shallow seekopfalm Sinkhole on the night of 15-16 November when strong winds came in aloft. The nocturnal inversion on this night was destroyed by turbulent erosion, a process in which turbulence erodes the inversion from above, progressing deeper and deeper into the sinkhole until the inversion is completely destroyed. The process took �.5 h to break the inversion in the 26-m-deep sinkhole. Finally, a comparison of temperatures at the floors of five sinkholes of different sizes and shapes revealed that the Gruenloch and seekopfalm Sinkholes have similar cooling curves, even though they have quite different elevations, drainage areas, and volumes.

This paper has provided an overview of the Hetzkogel Plateau experiments and the initial results from analyses of the sinkhole data. The participants of the experiments have prepared additional papers focused on key phenomena, including temperature inversion breakups in the Gruenloch (VYhiteman et al. 2004e), a comparison between temperatures on the sinkhole sidewalls and those measured over the basin center at the same time (Whiteman et al. 2()04a), and a comparison of temperature inversions in sinkholes of different sixes and shapes (Whiteman et al. 2004b).

What new findings have been encountered so far? Because of tethersonde observations and smoke experiments it could be shown clearly that the air within the sinkhole is not completely decoupled trom the air above. There is a shear layer on top of the inversion that is eroded continuously by winds alott, assisted by seiche-like undulations of the cold-air pool. Despite extremely stable stratification, there are weakly pulsating downslope winds present during nighttime in the sinkhole. This means that a thermodynamic simulation of a strictly stagnant cold-air pool, for example, by an energy budget model, may lead to incorrect results. One of the big surprises was the finding that the breakup of the inversion is not being assisted by surface heating at the bottom of the sinkhole, but merely by subsiding motion as a response to upslope winds. The temporal change of the vertical temperature profile is strongly dependent on the shading and sky-view factor in accordance with the geometry ot the sinkhole. A further new result was the observation that even a lew clouds passing over the sinkhole during an otherwise clear night shows nearly instantaneously a significant temperature signal at the bottom; it reacts like an infrared thermometer.

Despite the basic understanding of inversion formation and breakup there remain several open questions that will be tackled by further experiments and a longterm observation strategy. What does the mass budget of the cold-air pool look like? There is evidence of an outflow from the doline through the I.echner Saddle. � quantification will help to transfer the Undings of sinkholes to valley basins with narrow outflows. How does the mechanism of turbulent erosion on top of the cold-air pool work? Is there a penetration of slope winds above the sinkhole into the cold-air pool? Quantifying these slope flows will help to improve the knowledge on vertical exchange in valleys during highly stable stratified episodes. Is there an internal motion within the cold-air pool? Some recent smoke experiments showed clear indications of seiches. Why is there rarely fog formation observed in the sinkhole but intense dew or hoar frost deposits? How well do very high resolution models simulate the inversion formation and its breakup? What is the relation between weather types and inversion formation? Such relations could he used to improve temperature forecasts in other Alpine basins. Is the air mass (temperaturehumidity regime) above the Alps and its characteristic downward radiative flux the determining tactor tor the minimum temperature? If so, the sinkhole could serve as a natural "thermometer," undisturbed by human activities. We could possibly find a signal of the changed greenhouse effect between the historic and recent observational data.

ACKNOWLEDGMENTS. We thank the other organi/ers ami participants ul the 2001 02 (iruenloch experiments, including Dr. M. Haute! at the University of Vienna and Mr. F. Trailer at the Central Institute for Meteorology undGeodynamic.v Students from the University of Vienna are especially thanked for their many contributions to the field program, which was conducted as part of a university praeticum. We thank the personnel at the Lunz Biological Station tor providing local weather data and other support and thank Mr. P. Kupelwieser tor providing access to the experimental area.

CDW acknowledges partial support from National Science Foundation Grant ATM-0521776 and from the U.S. Department of Hnergy's (DOK's) Vertical Transport and Mixing Program under the auspices ol the Atmospheric Sciences Program of the Office of Biological and Environmental Research at Pacific Northwest National Laboratory (PNNL). PNNL is operated for the DOF. by Battelle Memorial Institute. Two of the authors (SE and BP) wish to thank the University of Vienna, the province of Lower Austria, DOF, and PNNL for fellowships served at PX'NL in 2002 and 2003.

[Sidebar]

Because sinkholes are an excellent natural laboratory for studying processes leading to the formation, maintenance, and dissipation of temperature inversions, an extended set of meteorological field experiments was conducted in limestone sinkholes of various sizes and shapes in the eastern Alps during the period from 17 October 2001 through 4 June 2002. The experiments were conducted in an area surrounding the Gruenloch Sinkhole, which in earlier years had recorded the lowest surface minimum temperature in Central Europe, -52.6�C. A dense array of surface temperature sensors and three automatic weather stations were operated continuously during the experimental period, and special experiments enhanced with tethersondes and other equipment were conducted from 2 to 4 June 2002. An overview of the experiments is presented and first results are given. (Page 701)

[Reference]

REFERENCES

Aigner, S., 1932: Die Temperaturminima im Gstettnerboden bei I.un/ am see, Niederosterreich (The minimuni temperatures in the Gstettner basin near I.un/, Lower Austria). Wetter l.eben. Special Issue 1, 34-37.

Clements, C. B., C. IX Whitenian, and ). I). Horel, 2003: Cold air pool structure and evolution in a mountain basin: Peter Sinks, Utah. /. Appl. Meteor., 42, 752-768.

Geiger, R., R. H. Aron, and P. Todhunter, 2003: The Climate near the Ground. 6th ed. Rowman & I.ittletield Publishers, 584 pp.

Lauscher, F., 1937: Grundlagen des Strahlungsklimas der I,un/er Kleinklimastationen (Fundamentals of the radiation climate of the I.un/ microclimate stations), jahrhuch der /entralanstalt liir Meteorologie u. Geodynamik, Beiheft 4, 1-24.

____, and M. Roller, 1952: Ober die Struktur der Schneedecken/eit in einem Alpental. I.un/ am see (On the structure ot snow cover duration in an alpine valley. I.un/. am see). Wetter Lcbcii, 4, 15-20.

l.itschauer, 1)., 1962: Untersuchung der Kntwickhing von Kaltluttseen in Dolinen- und Beckenlagen (Investigation ol the development of cool air pools in sinkholes and basins). Ph.D. dissertation. University of Vienna, 129 pp.

Maki, M., and T. 1 larimaya, 1988: The effect of advection and accumulation of downslope cold air on nocturnal cooling in basins. /. Meteor. Sac. Japan, 66,581-597.

Mori, M., and I. Kobayashi, 1996: Dynamic interaction between observed nocturnal drainage winds and a cold air lake. /. Meteor. Soc. japan, 74, 247-258.

Miiller, H., and C. D. Whiteman, 1988: Breakup of a nocturnal temperature inversion in the Dischma Valley during DISKUS. /. Climate Appl. Meteor., 27, 188-194.

Sauberer, P., 1947: Kleinklimatische Niederschlagsuntersuchungen im Lunzer Gebiet (Microclimatic precipitation investigations in the Lunz area). Die Umwelt, 11,410-415.

____, 1948: Niederschlagsmessungen am Nordhang des Kleinen Het/.kogcls (Precipitation measurements on the north slope ot Kleiner Hetzkogel). Wetter Leben, 1, 297-302.

____, 1952: Kleinklimatische und bioklimatische Arbeiten an der Biologischen Station I.un/ (Microclimatic and bioclimatic studies at the I.un/, Biological Station). Wetter Leben, 4, 4-12.

____, and I. Dirmhirn, 1954: Cber die Mnlstehung der extremen Temperaturminima in der Doline Gstettner-Alm (On the occurrence of extreme temperature minimums in the Gstettner-Alm Doline). Arch. Meteor. Geophys. Rioclinuitol., 5B, 307-326.

____, and ____, 1956: Weitere Untersucluingen iiber die Kaltluttansammlungen in der Doline(islettnerAlm bei Lunz in Niederosterreich (Further investigations of the cold air buildup in the GsteltnerAlm doline near Lunz in Lower Austria). Wetter Leben,8, 187-196.

Schmidt, W., 1930: Die tietsten Minimumtemperaturen in Mitteleuropa (The lowest minimum temperatures in Central F.urope). Die Ntitiirwissenschajten, 18, 367-369.

____, 1933: Kleinklimatische Beobachtungen in Osterreich (Microclimate observations in Austria). Geogr. lahresbericht Osterreich, 16, 53-59.

____, H. (Jams, W. Kiihnelt, }. l-'urlani, and H. Miiller, 1929: Bioklimatische Untersuchungen im l.un/.er Gebiet (Bioclimatic investigations in the Lunz region). Die Naturwissenschaften, 17, 176-179.

Whiteman, C. D., 1982: Breakup of temperature inversions in deep mountain valleys. Part I: Observations. /. Appl. Meteor., 21, 270-289.

____, 1990: Observations of thermally developed wind systems in mountainous terrain. Atmospheric Processes over Complex Terrain, Meteor. Monogr., No. 45, Amer. Meteor. Soc., 5-42.

____, I. M. Hubbe, and W. J. Shaw, 2000: Evaluation of an inexpensive temperature data logger tor meteorological applications. /. Atmos. Oceanic Technol., 17,77-81.

____, S. Eisenbach, B. Pospichal, and R. Steinacker, 2004a: Comparison of vertical soundings and sideWLill air temeprature measurements in a small Alpine basin. /. Appl. Meteor., 43, 1635-1647.

____, T. Haiden, 15. Pospichal, S. Eisenbach, and R. Steinacker, 2004b: Minimum temperatures, diurnal temperature ranges and temperature inversions in limestone sinkholes of different si/.e and shape. /. Appl. Meteor., 43, 1224-1236.

____, B. Fospichal, S. Eisenbach, P. Weihs, C. B. Clements, R. Steinacker, K. Mursch-Radlgruber, and M. Dorninger, 20()4c: Inversion breakup in small Rocky Mountain and Alpine basins./. Appl. Meteor., 43, 1069-1082.

[Author Affiliation]

AFFILIATIONS: steinacker, dorninger, pospichal, eissnbach, and holzer-Department of Meteorology and Geophysics, University of Vienna, Vienna, Austria; whiteman-Meteorology Department, University of Utah, Salt Lake City, Utah; weihs and mursch-radlgruber-Institute for Meteorology and Physics, Agricultural University, Vienna, Austria; baumann-Central Institute for Meteorology and Geophysics, Vienna, Austria

CORRESPONDING AUTHOR: Reinhold Steinacker, Department of Meteorology and Geophysics, University of Vienna, Althanstrasse 14, A-1090 Vienna, Austria

E-mail: Reinhold.Steinacker@univie.ac.at

The abstract for this article can be found in this issue, following the table of contents.

DOI:IO.II7S/BAMS-88-5-70l

In final form 5 December 2006

�2007 American Meteorological Society