Temporal and spatial variability of nocturnal cooling in a complex of small valleys in the Kanto Plain during the winter

Transcription

1 Short Paper J. Agric. Meteorol. 9 (3): 191-, 13 Temporal and spatial variability of nocturnal cooling in a complex of small valleys in the Kanto Plain during the winter Shohei KONNO*,, Tomoko NAKANO**, and Hideo TAKAHASHI* *Graduate School of Urban Environmental Sciences, Tokyo Metropolitan University, 1-1 Minamiosawa, Hachioji, Tokyo, , Japan **Faculty of Economics, Chuo University, 7-1 Higashinakano, Hachioji, Tokyo, , Japan Abstract We investigated the temporal and spatial variability of nighttime temperatures in a complex of small valleys in the western Kanto Plain, Saitama Prefecture, Japan, by conducting horizontal and vertical temperature measurements throughout one winter season. The magnitude of nocturnal cooling in the valley was greater in March than during the mid-winter period. The horizontal distribution of mean nighttime temperatures and nocturnal cooling under clear and calm conditions indicated that temperatures in the middle or lower areas of the valleys tended to be lower than those in the upper areas even though the elevation differences were small. Vertical profiles of air temperature measured by a kytoon system on clear nights revealed the development of a temperature inversion above the valley floor at night; the top of the inversion exceeded the height of the ridge adjacent to the valley. Under these conditions, down-valley flows of cold air were a significant phenomenon below the inversion layer. Our results suggest that local nocturnal drainage flow below the temperature inversion could contribute to nocturnal cooling in these valleys. Key words: Cold air flow, Kanto Plain, Nocturnal cooling, Small valley, Temperature inversion. 1. Introduction In recent decades, nocturnal cooling in basins or valleys has been investigated in numerous studies. Based on the results of these studies, nocturnal cooling in a basin or valley can be described as follows: (1) Under clear and calm conditions, air that is cooled by radiative cooling at ground level flows into the basin or valley from the surrounding mountain slopes, and its accumulation produces a strong inversion layer at lower elevations (Yoshino et al., 191; Mori et al., 193). () Formation of the inversion layer reduces downward longwave radiation toward the floor of the basin or valley, and promotes radiational cooling (Magono et al., 19; Maki and Harimaya, 19). (3) The surrounding mountains block the ambient wind and enhance nocturnal cooling by decreasing turbulent exchange (Saito et al., 19; Thompson, 19). Received; May 9, 1. Accepted; February 1, 13. Corresponding Author: konno-shohei@ed.tmu.ac.jp Most of these studies have been conducted in mountainous regions. For a small valley (here, defined as a valley of less than 1 km long and with a depth of approximately m) located in a plains area, the processes that govern nocturnal cooling are likely to differ from those in a mountainous area. Instead, the valley may be affected by a ground-level inversion that formed in the plains area. Previous studies of small valleys in a plains area have reported that in situ cooling and a wind-sheltering effect were the dominant factors that governed nocturnal cooling, with flows of cold air playing a lesser role (Tanaka et al., 193; Matsuoka et al., 197). However, there were insufficient temperature measurements in these studies. For example, horizontal temperature measurements were carried out at only one observation point in a valley (Matsuoka et al., 197), or vertical measurements were obtained at heights below approximately 3 m, which was lower than the top of the temperature inversion (Tanaka et al., 193). Although the results of these studies have improved our understanding of nocturnal cooling in small valleys

2 J. Agric. Meteorol. 9 (3), 13 in plains areas, some aspects of this process remain unclear because of a lack of studies or insufficient detail in existing studies. In the present study, we provided some of the missing information by investigating the temporal and spatial variability of nighttime air temperatures in a complex of small valleys in the Kanto Plain region of Japan. In this research, we obtained horizontal and vertical temperature measurements in the valley throughout one winter season. Our goals were to clarify the interseasonal variability of nocturnal cooling and to provide a detailed description of the horizontal and vertical distribution of air temperatures in the valleys.. Study Site and Methods.1 Study site The study area was located in the hilly Hiki region of Saitama Prefecture, Japan, where several gently sloping valleys stretch from east to west and from north to south (Fig. 1). The three east-west valleys were designated as V1, V, and V3, and the northsouth valley was designated as V. The mean slope gradients in valleys V1, V, V3, and V were approximately 1/11, 1/15, 1/15, and 1/, respectively. The altitudes of the valley floor decreased from a maximum of m above sea level (ASL) in valley V1 and m ASL in valleys V, V3, and V, to a minimum altitude of,, 5, and 35 m ASL, respectively, with the minimum occurring where valleys V1 to V3 joined valley V. The altitudes of the ridges surrounding the valleys ranged from to 13 m ASL. Based on the difference in altitude between the ridges that separate the valleys and the bottoms of the valleys (Kondo and Kuwagata, 19), the mean depth of the valleys was approximately m. The Hatoyama Automated Meteorological Data Acquisition System (AMeDAS) station, which is administered by the Japan Meteorological Agency (JMA), is located in the central area of the four valleys, near the point at which V1 joins V. In the present study, meteorological data, including temperatures at a height of 1.5 m above ground level (AGL) and the wind speed and wind Kumagaya Study area 3º N Y 35º N V S1 S1 V3 T S S S13 S1 S11 Mt. Monomi S1 S9 S V S19 S7 S S T T5 S5 S3 V1 S1 Hatoyama AMeDAS S17 T3 S T T1 S1 S1 139º E 1º E Horizontal measurements Multiple height measurements Kytoon measurement S15 X km Altitude (m ASL) X Kytoon measurement S17 Hatoyama AMeDAS S Distance (km) Y Fig. 1. Upper: Location and topography of the study area, and the distribution of the sites used for horizontal measurements (circles), measurements at multiple heights (diamonds), and kytoon measurements (star). Lower: Cross-section of the topography along the X-Y line shown in the upper figure

3 S. Konno et al.:temporal and spatial variability of nocturnal cooling in a complex of small valleys direction at a height of.5 m AGL, were obtained from this station.. Horizontal distribution measurements To investigate the spatio-temporal characteristics of the nocturnal temperature variation in the valleys, we measured the air temperature at.5 m AGL with automatic-logging thermometers (33, HIOKI E.E. Corp., Nagano, Japan) at points (sites S1 to S) in and around the valleys (Fig. 1). The temperature was recorded at 1-min intervals from 11 December 7 to 31 March. Sensors were kept in naturally ventilated radiation shelters to prevent them from being affected by solar radiation and rainfall. To clarify the characteristic features of the vertical profile of air temperature near the ground, measurements were conducted at multiple heights from six disaster-prevention administrative radio towers managed by the Hatoyama Town Office (sites T1, T, T3, T, T5, and T in Fig. 1). On each tower, we attached automaticlogging thermometers (RTR-5, T&D Corp., Nagano, Japan) at heights of.5, 5., 7.5, and 1. m AGL..3 Vertical profile measurements Vertical temperature profiles, wind speeds, and wind directions were measured with a kytoon system (a tethered balloon) from the night of 9 February to the morning of 1 March, and from the night of 7 March to the morning of March, at the playground of Hatoyama Junior High School, where the Hatoyama AMeDAS station is located. A kytoon is an aerodynamic helium-filled balloon that is tethered to a winch on the ground. Instead of being blown downwind, the balloon s shape allows it to soar upward like a kite and remain above the same position on the ground. Here, the tethered kytoon was allowed to rise up to m AGL, and thermometers (33, HIOKI E.E. Corp., Nagano, Japan) were attached to the tether at.5-m intervals from ground level to a height of 1 m, at 5-m intervals from 1 to m, at 1-m intervals from to m, and at -m intervals at heights above m, for a total of measurement points. Wind speed and direction were measured at heights of.5 and m AGL with sondes equipped with anemometers. All of the data were recorded at 1-s intervals, and 1-min average values were used for the analysis. The thermometers were calibrated before use, and fluctuation in kytoon height was periodically corrected with data from a pressure sensor at m AGL. Vertical temperature profiles up to 1 m were also obtained at sites T1, T, T3, and T to investigate the potential temperature structure near the ground along valley V1 (Section 3.3). The potential temperature, θ (K), is given by the following equation from Kondo (): θ = T + Γd z (1) where z is the height above the standard ground surface (m), T is the temperature (K) at height z (m), and Γ d is the dry adiabatic lapse rate (=.97 K m -1 ). The present study defined the ground surface at one valley site (S17, near the point at which V1 joins V) as the standard level.. Calculation of nocturnal cooling The magnitude of the nocturnal cooling (T noc ) at each point of measurement was calculated as follows: Tnoc = () Teve Tmin where T eve represents the evening air temperature, recorded 3 min before sunset (Kondo and Mori, 19), and T min represents the minimum air temperature recorded during the night and morning hours. In the present study, we defined a day as the period from 1: (noon) on one day until 1: on the following day, using Japan Standard Time (JST). 3. Results and Discussion 3.1 Interseasonal variability of nocturnal cooling Figure shows the interseasonal variability of T eve, T min, and T noc at site S17 and on the mountain ridge (S). We chose S17 as a representative point for the valley sites because it was located at the center of the bottom of the valley. The altitudes at S17 and S were 3 and 119 m ASL, respectively, and the elevation difference between these stations was thus 7 m. Nighttime weather conditions reported by the Kumagaya Local Meteorological Observatory of the JMA, which is located approximately km from our study site, are also shown in Fig.. The total number of clear, cloudy, and rainy days were 7,, and, respectively. Strong nocturnal cooling occurred on clear days, and weak nocturnal cooling occurred on rainy days at the valley and ridge sites. Mean nocturnal cooling values from 11 December 7 to 31 March at sites S17 and S were 11.5 and 7., respectively, on clear days,. and.1 on cloudy days, and 5. and

4 J. Agric. Meteorol. 9 (3), 13 Temperature (ºC) Clear Cloudy Rainy Tnoc_S17 Tnoc_S Teve_S17 Teve_S Tmin_S17 Tmin_S Nocturnal cooling (ºC) December 7 Temperature (ºC) Nocturnal cooling (ºC) January Temperature (ºC) Nocturnal cooling (ºC) February Temperature (ºC) Nocturnal cooling (ºC) March Fig.. Interseasonal changes in the evening temperature (T eve ), daily minimum temperature (T min ), and nocturnal cooling (T noc =T eve -T min ) at sites S17 (valley bottom) and S (mountain peak). Site locations are shown in Fig. 1. The symbol below each date describes the weather conditions on that date as reported by the Kumagaya Local Meteorological Observatory of the JMA. Arrows indicate clear and calm (wind speed of < 1 m s -1 ) days

5 S. Konno et al.:temporal and spatial variability of nocturnal cooling in a complex of small valleys on rainy days. On clear days, we observed that T eve at S17 (mean=. ) was slightly higher than that at S (7. ), and T min at S17 (-3.3 ) was lower than that at S (. ). In contrast, on rainy days, both T eve and T min at S17 (means=7. and., respectively) were higher than and similar to, respectively, those at S (.9 and.9, respectively). Iijima and Shinoda () reported that mean nocturnal cooling in the western part of the Kanto Plain in August was approximately to 7. They investigated atmospheric conditions that produced summertime nocturnal cooling in high mountain areas of central Japan, and determined the distribution of nocturnal cooling over Eastern Japan. Based on our results, the mean magnitude of nocturnal cooling in winter can reach up to twice that in summer, although differences between the present study area and those of Iijima and Shinoda () make a direct comparison difficult. Some days with strong nocturnal cooling occurred under rainy or cloudy weather conditions. In these cases, strong nocturnal cooling occurred late at night, when the weather changed from rainy or cloudy to clear. Furthermore, little nocturnal cooling occurred even in clear weather when certain conditions occurred. In these cases, the strong winter monsoon winds that occur at night inhibited the development of an inversion layer, and this suppressed nocturnal cooling. This hypothesis is supported by the relationship between T noc and nighttime mean wind speed during clear weather observed from 1: to : JST at the Hatoyama AMeDAS station (Fig. 3). At the beginning of spring, and particularly in March, the magnitude of T noc on clear days was larger than that during the mid-winter period (from December to mid-february). The maximum T noc at S17 reached 1.9 on March. The magnitude of T noc is generally controlled by net radiation at the ground surface (Geiger et al., 9). However, further studies are required to confirm the relationship between net radiation and nocturnal cooling through direct measurement of net radiation. Although the absolute values of nocturnal cooling differed among the measurement points, the interseasonal variability of nocturnal cooling at other sites during the observation period was similar to that at S17 (data not shown) Nighttime mean wind speed (m s -1 ) Fig. 3. Relationship between nighttime mean wind speed from 1: to : JST at a height of.5 m AGL and nocturnal cooling (T noc ) measured at the Hatoyama AMeDAS station under clear weather conditions. Nocturnal cooling (ºC) 3. Horizontal pattern of nocturnal cooling under clear and calm atmospheric conditions Figure shows the horizontal distribution of the mean T eve, T min, and T noc on clear and calm nights (the dates indicated by the arrows in Fig. ). In this analysis, a clear night was defined based on the description recorded at the Kumagaya Local Meteorological Observatory, and included nights described as clear or fine. A calm night was defined as one during which the mean wind speed from 1: to : JST at the Hatoyama AMeDAS was lower than 1. m s -1. Nights that satisfied both conditions were defined as clear and calm nights. A total of 1 nights met these criteria, from a total of 11 observation nights. High evening temperatures (9. T eve <1. ) were found in the middle or lower areas of the valleys (Fig. a), and low evening temperatures (T eve <. ) were found at T3 and T in valley V1 and at S7 in valley V. The other locations had temperatures in the range from. to 9.. Nighttime minimum temperatures in the valley areas ranged from - to -5 and were to 5 lower than those recorded in the ridge areas (Fig. b). Very low temperatures (-5. <T min -. ) were found at sites S1 and T1 in valley V1, at T5 in V, and at S and S9 in V3. With the exception of S, all these points were in low-lying areas, where the gradient was moderate. S was at the confluence of two

6 J. Agric. Meteorol. 9 (3), 13 a S T S1 V3 S13 S1 S11 S1 S9 T S V S7 S T5 S3 V1 S5 T3 S T T1 S1 Evening temperature b V3 S13 S1 S11 S1 S9 T S V S7 S T5 S3 V1 S5 T3 S T T1 S1 Minimum temperature c S S T T S1 S1 Nocturnal cooling V3 S13 S1 S11 S1 S9 T S V S7 S T5 S3 V1 S5 T3 S T T1 S1 V V V S19 S1 S17 S19 S1 S17 S19 S1 S17 Fig.. Horizontal distribution of mean (a) evening temperature (T eve ), (b) daily minimum temperature (T min ), and (c) nocturnal cooling (T noc =T eve -T min ) under clear and calm conditions. S1 S S1 S S1 S S1 S1 S1 S15 S15 S15 S S S Mt. Monomi (ºC) Mt. Monomi (ºC) km km Mt. Monomi (ºC) km valleys. The strongest nocturnal cooling (13. T noc < 1. ) was observed at sites S1 and T1 in valley V1, at S and S9 in V3, and at S15, S17, S1, S19, and S in V. These points were in the lower area of each valley or at the confluence between two valleys. T noc at sites in the upper areas of each valley ranged from 1. to 13. and was smaller than the cooling at the lowest sites. T noc at the mountain ridge site (S) was. to 9., which was at least lower than those in the valleys (Fig. c). If in situ radiative cooling dominates the nocturnal cycle in the valley, the temperature decrease from the evening to early morning in the upper and lower areas should be nearly equivalent at all sites within a valley (Tanaka et al., 193). Our results indicate that the nocturnal cooling at the middle or lower sites in the valley tended to be larger than that at the upper sites, suggesting that the occurrence of downhill flows of cold air and their accumulation act in addition to in situ radiative cooling and are important for nocturnal cooling in this area, even though the elevation differences that would drive these flows were small between the upper and the lower sites. 3.3 Surface wind characteristics, air temperature profiles, and potential temperature structures Figure 5 shows the wind speeds and directions measured at the Hatoyama AMeDAS station, and Figs a, b and 7a, b show the resulting profiles of air temperature as a function of height at site S17 on the nights of 9 February and 7 March, respectively. On the night of 9 February, the conditions were clear and calm throughout the night, with a wind speed of 1 m s -1 (Fig. 5). The wind did not persist but frequently discontinued during the night, showing intermittent characteristics. During this period, the top of the temperature inversion had already reached m AGL at 1: JST (Fig. a), and this height remained constant throughout the night (Fig. b). The temperature difference between the top of the inversion and the bottom of the valley was approximately 7 at 1: JST, but had increased to 1 by 5: JST (Fig. b). In comparison with the results from a previous study in which vertical temperature measurements were conducted in a nearby plains area in Saitama Prefecture (Tohsha, 1953; Akita, 1955), the height of the inversion in a valley coincided with the height of the ground inversion in the plains area; the temperature difference for the inversion in the valley was approximately larger than that of the ground inversion in the plains area. Figures c and d show the potential temperature structures along the V1 valley at 1: and 5: JST during the same observation night. Isothermal lines for potential temperature were linearly interpolated based on the potential temperatures calculated by using Eq. (1). At 1: JST, the near-surface isothermal lines for potential temperature (7 and 77 K) could be drawn parallel to the valley slope (Fig. c). On the following morning (Fig. d), the isothermal lines for potential temperature (7 and 71 K) intersected the valley slope in the middle and lower areas of the valley (from S to S17). The higher isothermal lines (7 and 73 K) could be drawn parallel to the valley slope in the upper areas of the valley, but changed to horizontal lines in the middle and lower areas of the valley. The wind direction in the valley measured at the Hatoyama AMeDAS station ranged between north and west

7 S. Konno et al.:temporal and spatial variability of nocturnal cooling in a complex of small valleys Wind speed (m s -1 ) Wind speed_9 Feb. Wind direction_9 Feb. Wind speed_7 Mar. Wind direction_7 Mar. 3 N N Time (JST) Fig. 5. Temporal variation of wind speed and direction measured at the Hatoyama AMeDAS station on the nights of 9 February and 7 March. 7 W 1 S 9 E Wind direction Height (m AGL) Height (m AGL) (a) 1 (b) 9 Feb. 1: JST Air temperature (ºC) 1 1 Mar. 5: JST Air temperature (ºC) Altitude (m ASL) Altitude (m ASL) (c) (d) 9 Feb. 1: JST Kytoon ( m AGL) Temp: 7.9 ºC Wind: Calm Hatoyama AMeDAS Temp: 1. ºC Wind: NNW 1 m s 1 7 S T S3 T3 S T T S1 S Distance (km) Mar. 5: JST Kytoon ( m AGL) Temp:. ºC Wind: Calm Hatoyama AMeDAS Temp: 3.9 ºC Wind: N 1 m s 1 S T S3 T3 S T T1 7 S1 S Distance (km) Fig.. (a, b) Vertical profiles of air temperature measured by the kytoon system at site S17 and (c, d) vertical cross-sections showing the isothermal distribution of potential temperature along valley V1 on the night of 9 February and early morning of 1 March

8 J. Agric. Meteorol. 9 (3), 13 during the night, which corresponded to the orientations of valleys V and V1 (Fig. 5). Because the station is located at the point at which V1 joins V, the area could have been affected by the down-valley flows of cold air from V1 and V. On the night of 7 March strong winds of 3 to m s -1 were recorded at the Hatoyama AMeDAS station until 1: JST, which then decreased to 1 m s -1 (with slight fluctuation) thereafter (Fig. 5). The nocturnal wind also blew intermittently during the night. At 1: JST, a weak temperature inversion, with a temperature difference of approximately 1 between the surface and a height of m, formed at S17 (Fig. 7a). However, at 5: JST, the top of the temperature inversion reached approximately 1 to 1 m above the ground surface, and the temperature difference between the top of the inversion and the ground surface was approximately 9 (Fig. 7b). Within the valley, the potential temperature was nearly homogeneous both horizontally and vertically at 1: JST (Fig. 7c). However, at 5: JST on the following morning, isothermal lines for potential temperature (9 and 7 K) intersected the valley slope (Fig. 7d). Wind measured at the Hatoyama AMeDAS station Height (m AGL) Height (m AGL) (a) 1 (b) 1 7 Mar. 1: JST Air temperature (ºC) Mar. 5: JST Air temperature (ºC) Altitude (m ASL) Altitude (m ASL) (c) (d) 7 Mar. 1: JST Kytoon ( m AGL) Temp: 3. ºC Wind: N 7 m/s Hatoyama AMeDAS Temp: 3.9 ºC Wind: WNW m s 1 S T S3 T3 S T 77 T1 S1 S Distance (km) Mar. 5: JST Kytoon ( m AGL) Temp:.1 ºC Wind: N 9 m/s Hatoyama AMeDAS Temp: 5. ºC Wind: Calm S T S3 T3 S T 71 T1 7 9 S1 S Distance (km) Fig. 7. (a, b) Vertical profiles of air temperature measured by the kytoon system at site S17 and (c, d) vertical cross-sections showing the isothermal distribution of potential temperature along valley V1 on the night of 7 March and early morning of March

9 S. Konno et al.:temporal and spatial variability of nocturnal cooling in a complex of small valleys ranged between north and northwest during the night, which was similar to the wind direction on the night of 9 February (Fig. 5). Wind measurements recorded at the Hatoyama AMeDAS station indicate that a cold air current flowed from the upper to the lower areas of the valley. Note that the cold air current occurred below the temperature inversion level. Mori and Kobayashi (199) conducted meteorological observations on the slopes of an intermountain basin in Western Japan and found that two types of nocturnal drainage wind (NDW) developed on the same slope. In their study, the NDW that contributed to the development of a pool of cold air and that flowed down into the surface layer of the pool was defined as a NDW I type, whereas the NDW that periodically occurred below the pool of cold air was defined as a NDW II type. In the present study, the cold air flow was observed below the temperature inversion level, so the cold air current can be classified as a NDW II type. The Kanto Plain is a region with a number of small valleys distributed around its margin. Furthermore, it is known that a large-scale ground inversion forms above the Kanto Plain on clear and calm winter nights (Kondo, 1995; Ueda et al., 11). Therefore, it is possible that the NDW II type promotes nocturnal cooling in these valleys. As a future subject of study, the temporal and spatial characteristics of these flows and their relationships with the development of temperature inversions should be confirmed in the valleys around the Kanto Plain. Acknowledgments The stationary sampling sites were supported by the Hatoyama Town Office, Hatoyama Junior High School, and Higashimatsuyama City Office, Saitama Prefecture, Japan. We thank Prof. Takehiko Mikami of Teikyo University and Dr. Hiroharu Tanaka of the Nagano Environmental Conservation Research Institute for their useful advice on this study. We also thank the students of Tokyo Metropolitan University for their assistance during our observations. References Akita, I., 1955: Temperature and wind velocity in the first 3 m during night at Honjo. J. Meteorol. Res., 7, 9-97 (in Japanese with English summary). Geiger, R., Aron, R. H., and Todhunter, P., 9: The climate near the ground (7th ed.). Rowman & Littlefield Publishers, Maryland, pp. Iijima, Y., and Shinoda, M., : Effects of largescale atmospheric conditions on summer nocturnal cooling in high mountains in Central Japan. Geogr. Rev. Jpn., 77, (in Japanese with English summary). Kondo, H., 1995: The thermally induced local wind and surface inversion over the Kanto Plain on calm winter nights. J. Appl. Meteorol. 3, Kondo, J., : Atmospheric science near the ground surface. University of Tokyo Press, Tokyo, 33 pp. (in Japanese). Kondo, J., and Kuwagata, T., 19: A relation between the depth of nocturnal stable layer (cold-air-pool) and the topographical depth of the basin. Tenki, 31, (in Japanese). Kondo, J., and Mori, Y., 19: Analysis on nocturnal cooling at the regional meteorological stations (Automated Meteorological Data Acquisition System - AMeDAS), Part 1. Tenki, 9, (in Japanese). Magono, C., Nakamura, C., and Yoshida, Y., 19: Nocturnal cooling of the Moshiri Basin, Hokkaido in midwinter. J. Meteorol. Soc. Jpn.,, Maki, M., and Harimaya, T., 19: The effect of advection and accumulation of downslope cold air on nocturnal cooling in basins. J. Meteorol. Soc. Jpn.,, Matsuoka, N., Horiguchi, I., and Tani, H., 197: Topographic effects on destruction of stable layer during nighttime (1) Relationship between dropping of night air temperature and topography at Hayakita and Chitose. J. Agric. Meteorol.,, (in Japanese with English summary). Mori, M., and Kobayashi, T., 199: Dynamic interaction between observed nocturnal drainage winds and a cold air lake. J. Meteorol. Soc. Jpn., 7, 7-5. Mori, Y., Kondo, J., Shoji, K., Sato, T., Yasuda, N., Haginoya, S., Miura, A., Yamazawa, H., Kawanaka, A., Takahira, S., and Abe, E., 193: Nocturnal cooling and heat balance at the mountainous district. Tenki, 3, 59-7 (in Japanese). Saito, T., Tomari, I., Hayashi, T., and Mihara, Y., 19: An explanation of severe low temperature in a depressed ground at night. J. Agric. Meteorol., 1, 13-1 (in Japanese). Tanaka, Y., Fujiwara, K., and Kobayashi, D., 193: Nocturnal cooling in the shallow valley and on the plateau () topographic dependence of cooling pro

10 J. Agric. Meteorol. 9 (3), 13 cess. J. Agric. Meteorol., 39, (in Japanese with English summary). Thompson, B. W., 19: Small-scale katabatics and cold hollows. Weather, 1, Tohsha, M., 1953: Temperature inversion in the lower atmosphere. J. Meteorol. Res., 5, 9-5 (in Japanese with English summary). Ueda, H., Kobanawa, Y., Ohba, M., Inoue, T., Kamae, Y., Ikegami, H., Takeuchi, A., and Ishii, N., 11: Annual observation at the four cardinal slopes on Mt. Tsukuba-with focus on the thermal belt-. Tenki, 5, (in Japanese). Yoshino, M. M., Tanaka, M., and Nakamura, K., 191: Formation of a cold air lake and its effects on agriculture. J. Nat. Disaster Sci., 3,